Emerin induces nuclear breakage in Xenopus extract and early embryos

Abstract

Emerin is an inner nuclear membrane protein often mutated in Emery–Dreifuss muscular dystrophy. Because emerin has diverse roles in nuclear mechanics, cytoskeletal organization, and gene expression, it has been difficult to elucidate its contribution to nuclear structure and disease pathology. In this study, we investigated emerin’s impact on nuclei assembled in Xenopus laevis egg extract, a simplified biochemical system that lacks potentially confounding cellular factors and activities. Notably, these extracts are transcriptionally inert and lack endogenous emerin and filamentous actin. Strikingly, emerin caused rupture of egg extract nuclei, dependent on the application of shear force. In egg extract, emerin localized to nonnuclear cytoplasmic membranes, and nuclear rupture was rescued by targeting emerin to the nucleus, disrupting its membrane association, or assembling nuclei with lamin A. Furthermore, emerin induced breakage of nuclei in early-stage X. laevis embryo extracts, and embryos microinjected with emerin were inviable, with ruptured nuclei. We propose that cytoplasmic membrane localization of emerin leads to rupture of nuclei that are more sensitive to mechanical perturbation, findings that may be relevant to early development and certain laminopathies.

INTRODUCTION

Mutations in proteins of the nuclear envelope (NE) are increasingly being implicated in a host of diseases, generally termed nuclear envelopathies or laminopathies (Prokocimer et al., 2009; Worman et al., 2010; Burke and Stewart, 2014; Davidson and Lammerding, 2014; Janin et al., 2017). Emerin (gene symbol EMD) is an important member of this group, as its mutation accounts for ∼60% of cases of Emery–Dreifuss muscular dystrophy (EDMD) (Bione et al., 1994; Helbling-Leclerc et al., 2002; Koch and Holaska, 2014; Janin et al., 2017). Emerin is an integral membrane protein found at the inner nuclear membrane where it binds directly to A- and B-type nuclear lamins (Gruenbaum et al., 2005; Barton et al., 2015). Emerin belongs to the LEM domain (Lap2, emerin, Man1 domain) protein family, whose members share a common sequence motif that mediates indirect interactions with chromatin, and emerin interacts with other LEM domain proteins as well as with itself (Haraguchi et al., 2001; Lee et al., 2001; Segura-Totten et al., 2002; Shimi et al., 2004; Bengtsson and Wilson, 2006; Holaska and Wilson, 2007; Wilson and Foisner, 2010; Berk et al., 2013, 2014). While emerin and the nuclear lamina are implicated in many functions, ranging from chromatin organization and transcription regulation to nuclear mechanics and signaling, important questions remain about how emerin controls nuclear structure and how emerin mutations give rise to disease phenotypes.

Emerin plays diverse roles in nuclear structure, actin dynamics, mechanosensation, and gene expression. Emerin knockdown in cardiomyocytes decreased nuclear invaginations and increased nuclear size (Shimojima et al., 2017). In mouse embryonic fibroblasts, emerin deletion resulted in abnormal nuclear shape (Lammerding et al., 2005), and altered actin dynamics caused by deletion or mutation of lamin A could be rescued by ectopic emerin expression (Ho et al., 2013). Further evidence for emerin’s impact on the actin cytoskeleton came from in vitro studies showing that emerin can increase actin polymerization, potentially impacting the dynamics of the intranuclear actin network (Holaska et al., 2004). Emerin also functions in actin-mediated nuclear positioning and cell migration (Chang et al., 2013). At the inner nuclear membrane, emerin interacts with SUN domain proteins, components of the LINC (linkers of the nucleoskeleton to the cytoskeleton) complexes that transduce mechanical signals across the NE (Razafsky and Hodzic, 2009; Haque et al., 2010; Tapley and Starr, 2013; Zuela and Gruenbaum, 2016). Consistent with emerin functioning in mechanosensation, emerin-deficient mouse embryonic fibroblasts showed reduced expression of mechanosensitive genes under mechanical strain, ultimately resulting in apoptosis (Lammerding et al., 2005). In addition, nuclei lacking emerin exhibited altered NE elasticity and mechanical response to applied force, which might explain the increased fragility of EDMD nuclei (Rowat et al., 2006; Guilluy et al., 2014). Emerin also plays important roles in regulating gene expression and chromatin organization (Haraguchi et al., 2004; Holaska et al., 2006; Markiewicz et al., 2006; Demmerle et al., 2012; Le et al., 2016; Willer and Carroll, 2017).

Here we investigate emerin’s impact on nuclear structure, using Xenopus laevis egg extract, a simplified system with many advantages. First, Xenopus eggs and early embryos do not express emerin (Gareiss et al., 2005), allowing us to test the effect of ectopic emerin addition. Second, it is simple to manipulate Xenopus egg extract biochemically; for instance, a recombinant protein of interest, such as emerin, can be added directly. Third, extracts are composed of stockpiled egg proteins and membrane that support de novo nuclear assembly upon addition of demembranated Xenopus sperm chromatin, and these nuclei can be subjected to typical studies of nuclear structure and function (Chen and Levy, 2018). Fourth, these extracts lack filamentous actin and are transcriptionally inert, eliminating potentially confounding effects of emerin on actin dynamics and transcription, respectively. Furthermore, complementary in vivo studies using Xenopus embryos are possible. While Xenopus egg extracts have been used to investigate the mitotic phosphorylation of emerin (Hirano et al., 2005) and the localization of emerin’s LEM domain (Dreger et al., 2002), in this study we investigate the impact of full-length emerin on both in vitro assembled nuclei and embryonic nuclei.

RESULTS

Emerin induces breakage of nuclei assembled in Xenopus egg extract

Emerin protein expression is not detectable in X. laevis until stage 43, when swimming tadpoles have already developed (Gareiss et al., 2005). As Xenopus eggs and early embryos do not contain endogenous emerin protein, something we confirmed by Western blot (Supplemental Figure S1A), we first asked how ectopic addition of emerin to Xenopus egg extract would influence nuclear structure and morphology. To our surprise, addition of recombinant emerin to nuclei assembled in Xenopus egg extract resulted in apparent nuclear rupture. Control nuclei treated with the dialysis buffer used in the final step of emerin purification exhibited Hoechst-stained DNA encased in a spherical or ellipsoid nucleus, as expected. When nuclei were treated with emerin, the DNA was converted into elongated strings and/or puncta that spread across the entire field of view (Figure 1, A and B). As a control, we heat-treated emerin at two different temperatures. Emerin inactivated at 100°C for 5 min was no longer capable of inducing the stringy DNA phenotype, while 60°C-treated emerin was still active in our assay, suggesting some level of heat resistance (Supplemental Figure S1B). We also tested a range of emerin concentrations and incubation times, observing nuclear rupture with 0.6–5 nM emerin and as little as 15 min of incubation (Supplemental Figure S1, C and D).

FIGURE 1:

Emerin induces breakage of X. laevis egg extract nuclei. (A) Experimental approach. See Materials and Methods for further details. (B) Representative images of nuclei treated with dialysis buffer or 5 nM recombinant emerin. (C) Quantification of the nuclear breakage phenotype. The intact GFP⁺ nuclei were counted for buffer- or emerin-treated nuclei in eight different X. laevis egg extracts. Above each pair of bars is the percentage of intact emerin-treated nuclei compared with the buffer control. Above the graph is the average of all data.

While the stringy/punctate DNA phenotype was striking, we also noted some apparently intact nuclei after emerin treatment. To quantify this observation, we incubated in vitro–assembled nuclei with the import substrate GST-GFP-NLS prior to emerin addition. We then counted the number of intact nuclei that still retained intranuclear GFP-NLS signal. On the basis of experiments with eight different egg extracts, we observed that more than 50% of emerin-treated nuclei were disrupted (Figure 1C) and only 18% of nuclei lost GFP-NLS signal upon treatment with heat-inactivated emerin (unpublished data). To further confirm nuclear breakage, we showed that nuclei and DNA in emerin-treated samples accumulated GFP-cGAS (Supplemental Figure S1E), a cytoplasmic DNA binding protein that is excluded from intact nuclei (Raab et al., 2016). Furthermore, stringy DNA from ruptured nuclei showed little to no staining for membrane (Figure 2A), nucleoporins (Supplemental Figure S2, A and B), or Man1 (Supplemental Figure S2C), reflecting complete dissolution of the NE. Man1 is a LEM domain protein that is present in X. laevis eggs and early embryos (Reil and Dabauvalle, 2013). Man1 staining was unaffected in intact emerin-treated nuclei (Supplemental Figure S2C), suggesting that excess emerin is not negatively influencing the localization of other LEM domain proteins.

FIGURE 2:

TRC40 rescues emerin’s nuclear localization. (A) Experiments were performed as in Figure 1, except that the extract was supplemented with 1 µM DiI and 0.4 µM SNAP-emerin conjugated to Alexa Fluor 488. Representative images of small and large emerin puncta are shown. (B) Nuclei assembled in X. laevis egg extract as shown in Figure 1 were supplemented with 8 nM TRC40-EMD or an equivalent volume of dialysis buffer. After a 30-min incubation, nuclei were fixed, spun down onto coverslips, and stained with an anti-emerin antibody and Hoechst. Emerin images were acquired with the same exposure time. Representative images are shown.

Our data suggested that some amount of force was required to induce rupture of emerin-treated nuclei. In our typical assay, nuclei in extract were pipetted onto a glass slide and overlaid with a glass coverslip for imaging. The force imparted by the coverslip was sufficient to rupture emerin-treated nuclei. On the other hand, if emerin was carefully added to a drop of extract on a glass coverslip, nuclear breakage was not immediately evident until the extract was mixed (Supplemental Video 1). To further address this issue, we utilized a microfluidic device that allowed nuclei in extract to be visualized immediately upon mixing with emerin. Time-lapse imaging of this mixing process clearly showed generation of elongated DNA strings after apparent nuclear rupture, and DNA aggregation frequently clogged the device (Supplemental Video 2). Emerin-induced nuclear breakage was also evident when nuclei in extract were encapsulated in droplets (Supplemental Figure S2D).

Video S1.

Download video file ^{(1,014.8KB, mp4)}

Emerin‐induced breakage of X. laevis egg extract nuclei upon mixing. Nuclei assembled in egg extract and labeled with 10 μg/ml Hoechst were applied to a coverslip and imaged on an inverted confocal microscope. Emerin was added at a concentration of 5 nM. Initially nuclei remained intact as shown at the beginning of the video. Imaging was initiated and then the extract droplet was mixed using a pipet tip, corresponding to the middle frames in the video when the Hoechst‐stained DNA goes out of focus. After mixing emerin into the reaction, stringy and punctate DNA is apparent, similar to what is shown in Figure 1 for emerin‐treated nuclei sandwiched between a glass slide and coverslip. The length of the movie is 30 seconds.

Video S2.

Download video file ^{(2.4MB, mp4)}

Emerin‐induced breakage of X. laevis egg extract nuclei under flow in a microfluidic device. Nuclei assembled in egg extract and labeled with Sytox green are flowing from the left. Emerin at a concentration of 400 nM is flowing from the top. XB buffer is flowing from the bottom. Note that ruptured nuclei clog the device at the junction, leading to backflow of extract into the other channels. This explains why nuclei are periodically seen to flow from the top and bottom channels. Imaging was performed on an inverted fluorescence microscope, and the length of the movie is 40 seconds.

Rescue of emerin-induced nuclear breakage

To begin to address the mechanism of emerin-induced nuclear rupture, we tested whether nuclear localization of emerin was required. Nuclei were pretreated with wheat germ agglutinin (WGA) to occlude nuclear pore complexes and prevent nuclear import. WGA-treated nuclei still ruptured upon emerin addition, suggesting that breakage was not induced by intranuclear emerin (Supplemental Figure S3A). This prompted us to examine the localization of ectopically added emerin using a fluorescently tagged version of emerin that still retained the ability to induce nuclear rupture. We primarily observed small emerin puncta that colocalized with DiI-stained membrane, and in some cases we observed larger emerin aggregates that were also membrane-localized. Notably, we did not detect any emerin associated with in vitro–assembled nuclei (Figure 2A). Consistent with these data, the recombinant emerin LEM domain also did not localize to nuclei assembled in Xenopus egg extract (Dreger et al., 2002). Because emerin is associated with cytoplasmic membrane and does not have to be intranuclear to induce breakage, we surmise that emerin is exerting its effect on nuclear integrity from outside the nucleus in the cytoplasm.

We next tested whether nuclear breakage could be rescued by altering emerin’s localization. Emerin lacking its transmembrane domain (amino acids 228–242) did not induce nuclear rupture (Figure 3), nor did the LEM domain alone (Dreger et al., 2002). Emerin purified in a complex with TRC40 (also known as ASNA1), a protein responsible for targeting emerin to the inner nuclear membrane (Yamamoto and Sakisaka, 2012; Pfaff et al., 2016), was properly targeted to nuclei in extract (Figure 2B), and nuclear breakage was no longer observed (Figure 3). It is worth noting that TRC40 is up-regulated in X. laevis embryos only after the midblastula transition (MBT; Peshkin et al., 2015). To test whether emerin has a deleterious effect on nuclei because its transmembrane domain is unchaperoned, we tested a different unchaperoned tail-anchored membrane protein, Sec61β, finding that it failed to induce nuclear breakage (Figure 3). Consistent with emerin having an effect specific to nuclear integrity, emerin did not qualitatively disturb or alter nonnuclear membrane structures visualized with DiI (Supplemental Figure S3B).

FIGURE 3:

Rescue of nuclear breakage. Experiments and quantification were performed as in Figure 1. Data are plotted as the mean and SD from multiple independent experiments. The wild-type emerin data are the same as presented in Figure 1C (n = 8 extracts). TRC40-emerin was added at 5 nM (n = 6 extracts). For the “emerin + lamin A experiment,” recombinant lamin A was added during nuclear assembly at 1 nM before the addition of 5 nM emerin (n = 5 extracts). The emerin-ΔTM protein (deletion of transmembrane domain amino acids 228–242) was added at 5 nM (n = 4 extracts). The Sec61β protein was added at 5 nM (n = 2 extracts), and 50 nM Sec61β also did not induce nuclear breakage (unpublished data). Across all experiments, the average number of intact nuclei for buffer controls was 240. Statistical analysis was performed relative to the buffer control, which was normalized to 100% intact nuclei (bold horizontal line). ***p < 0.005; NS, not significant.

We next preassembled nuclei with lamin A, an important emerin-interacting partner not present in egg extract (Wolin et al., 1987; Gareiss et al., 2005; Jevtic et al., 2015), and found that these nuclei were resistant to emerin-induced breakage (Figure 3), possibly because lamin A increases NE stiffness (Lammerding et al., 2006; Pajerowski et al., 2007; Neelam et al., 2015; Stephens et al., 2017). In addition, nuclei isolated from extract and resuspended in buffer failed to break upon exposure to emerin, showing that extract, and likely membrane within extract, is required for emerin-induced nuclear rupture (Supplemental Figure S3C). It is worth noting that when we added emerin to extract before nuclear assembly, we observed some emerin localized to the DNA (Supplemental Figure S3D) and some nuclei formed, albeit with altered nuclear shape (unpublished data). Thus, nuclear rupture can be rescued by incorporating emerin or lamin A into the NE or by disrupting emerin’s membrane association.

Emerin induces nuclear breakage and inviability in Xenopus embryos

We next asked whether emerin would induce rupture of in vivo–assembled nuclei. Our first approach was to fertilize Xenopus eggs, allow them to develop to different embryonic stages, and then prepare embryo extracts containing endogenous embryonic nuclei. Nuclei in early-stage embryo extracts (stages 5–8) were susceptible to emerin-induced breakage, similar to what was observed with in vitro–assembled nuclei. Of interest, nuclei in later stages 10–11 remained intact upon emerin addition (Figure 4A). Because Xenopus nuclei scale smaller as development progresses, we wondered whether emerin failed to induce rupture of these nuclei because of their smaller size. To test this idea, we assembled nuclei in egg extract but only allowed them to grow for a brief period of time, reaching an average cross-sectional area similar to that of stage 9–10 nuclei. These small egg extract nuclei still broke upon emerin addition (Figure 4B), indicating that susceptibility to nuclear breakage is likely not dictated by nuclear size.

FIGURE 4:

Early stage X. laevis embryonic nuclei are susceptible to emerin-induced breakage. (A) Embryo extracts containing endogenous embryonic nuclei were prepared from different-stage X. laevis embryos. At least 30 embryos were used per extract. Extracts were supplemented with 5 nM emerin or an equivalent volume of dialysis buffer as a control and incubated at room temperature for 30 min. Nuclei were stained with Hoechst and visualized as in Figure 1. Representative images from one experiment out of three are shown. Average nuclear cross-sectional area at each developmental stage is indicated in µm² (Jevtic and Levy, 2015). (B) Experiments were performed with X. laevis egg extract as in Figure 1, except that the lengths of nuclear assembly and emerin incubation were varied as indicated, so that the nuclei were different sizes when emerin was added. Representative images from one experiment out of three are shown. Average nuclear cross-sectional area for each condition is indicated in µm².

Using a second approach, we microinjected fertilized one cell–stage embryos with recombinant emerin or dialysis buffer as a control. Buffer-injected embryos developed normally, while emerin-injected embryos began to show a loss of viability at stage 8, a time when cell motility is acquired in some regions of the embryo. The percentage of embryos that had ceased development increased at stage 10 and beyond, coinciding with gastrulation (Figure 5, A and B). While less than 20% of emerin-injected embryos reached stage 12 of development, more than 90% of embryos microinjected with the TRC40–emerin complex survived (unpublished data). Imaging of emerin-injected embryos revealed many apparently broken nuclei with a stringy DNA appearance similar to that observed for emerin-treated egg and embryo extract nuclei (Figure 5C).

FIGURE 5:

Microinjected emerin protein induces nuclear breakage and death in X. laevis embryos. (A) One-cell embryos were microinjected with emerin to achieve a final concentration of 0.5 nM within the embryo or an equivalent volume of dialysis buffer and allowed to develop at room temperature. Representative images at different developmental stages are shown. (B) At the indicated developmental stages, the viable embryos were counted. Inviable embryos were those that had stopped dividing or exhibited a puffy white appearance indicative of apoptosis (Johnson et al., 2010; Du Pasquier et al., 2011; Tokmakov et al., 2011; Gillespie et al., 2012; Willis et al., 2012; Iguchi et al., 2013; Broadus et al., 2015). The number of viable embryos was normalized to the buffer control. From 30 to 60 embryos were analyzed per condition per experiment. Averages from three independent experiments are shown. Error bars represent SD. Statistical analysis was performed relative to the stage 5–6 embryos. ***p < 0.005; NS, not significant. (C) Nuclei in buffer- or emerin-microinjected embryos were visualized with Hoechst at stage 12. Representative images from one experiment out of three are shown.

To test whether emerin synthesized within the embryo would have a similar effect, we microinjected mRNA encoding emerin. As with protein microinjections, emerin mRNA induced a loss of embryo viability (Figure 6A and Supplemental Video 3). The effect was concentration–dependent, with 500, 1000, and 1500 pg of microinjected mRNA resulting in 64%, 25%, and 0% of embryos reaching stage 12, respectively (Figure 6B). To examine nuclear integrity in these embryos, we comicroinjected mRNA encoding GFP-NLS and emerin. Many cells in emerin-expressing embryos exhibited stringy DNA and a general loss of the nuclear GFP-NLS signal (Figure 6C). To quantify the latter effect, we isolated nuclei from microinjected embryos and counted the nuclei with intranuclear GFP-NLS. There was a more than 50% reduction in the number of GFP⁺ nuclei in emerin-expressing embryos (Figure 6D), suggesting that in vivo–produced emerin leads to a loss of nuclear integrity. We also microinjected embryos with mRNA encoding GFP-emerin. At stages 8 and 10, GFP-emerin often appeared punctate or diffusely cytoplasmic (Supplemental Figure S4), similar to what was observed in vitro (Figure 2A). This was particularly evident in cells with elongated, stringy DNA masses where nuclei had presumably ruptured (Supplemental Figure S4). Nuclear breakage and lysis in a subset of cells are likely sufficient to arrest development, leading to embryo inviability. A similar phenotype was observed in some stage 12 cells, while in other cells, GFP-emerin was properly localized to the nuclear periphery, as well as to the cytoplasm and cell cortex (Supplemental Figure S4). Taken together, these data show that cytoplasmic localization of emerin in early stage embryos is associated with nuclear rupture and embryo inviability.

FIGURE 6:

Microinjected emerin mRNA induces loss of nuclear integrity and death in X. laevis embryos. (A) One-cell embryos were microinjected with 1500 pg of emerin mRNA (dissolved in water) or an equivalent volume of water as a control and allowed to develop. Representative images at different developmental stages are shown. (B) One-cell embryos were microinjected with the indicated amounts of emerin mRNA or an equivalent volume of water as a control and allowed to develop. At the indicated developmental stages, the viable embryos were counted. Inviable embryos were those that had stopped dividing or exhibited a puffy white appearance indicative of apoptosis (Johnson et al., 2010; Du Pasquier et al., 2011; Tokmakov et al., 2011; Gillespie et al., 2012; Willis et al., 2012; Iguchi et al., 2013; Broadus et al., 2015). The number of viable embryos was normalized to water-microinjected controls. From 11 to 89 embryos were analyzed per condition per experiment (43 embryos on the average). Averages from three independent experiments are shown. Error bars represent SD. Statistical analysis was performed relative to the water-microinjected control embryos. ***p < 0.005; NS, not significant. Injection with 100 pg emerin mRNA had no effect on embryo viability, and 2000 pg emerin mRNA exerted an effect similar to 1500 pg (unpublished data). (C) One-cell embryos were microinjected with 1000 pg GFP-NLS mRNA with or without 1500 pg emerin mRNA and allowed to develop. When control embryos reached stage 12, nuclei in microinjected embryos were visualized with Hoechst and GFP-NLS. Representative images from one experiment out of two are shown. (D) When control embryos reached stage 12, extracts were prepared from equivalent numbers of microinjected embryos described in C. Equivalent volumes of extract were supplemented with Hoechst, applied to a slide, overlaid with a coverslip, and incubated for 15 min. Images were acquired and the intact GFP⁺ nuclei per ∼660 × 660 µm field were counted. Nuclei from at least six fields were counted per experiment and condition. Averages from two independent experiments are shown. Error bars represent SD. ***p < 0.005.

Video S3.

Download video file ^{(3.2MB, mp4)}

Embryos microinjected with emerin mRNA cease development. One‐cell embryos were microinjected with 1500 pg emerin mRNA (2 embryos on the right) or an equivalent volume of water as a control (1 embryo on the left) and allowed to develop. Bright‐field imaging was performed at room temperature starting around stage 8. Images were acquired every 90 seconds for a total of ∼19 hours.

DISCUSSION

We find that emerin added to Xenopus egg extract and early embryos is localized to the cytoplasm and induces nuclear breakage that can be rescued by disrupting emerin’s membrane localization or by properly targeting emerin or lamin A to the nucleus. Why might emerin localized away from the nucleus induce nuclear rupture? One possibility is that cytoplasmic emerin sequesters other nuclear proteins, causing nuclei to become fragile and susceptible to breakage upon the application of force. This seems unlikely, given that nuclear breakage was observed after very brief incubation periods (Figure 4B, Supplemental Figure S1D, and Supplemental Video 1), and there was no evidence of nucleoporin or Man1 aggregates in the cytoplasm (Supplemental Figure S2, A–C). Another possibility is that emerin disrupts the incorporation of nucleoporin-containing annulate lamellae into the NE, a process that has been described in early Drosophila development (Hampoelz et al., 2016) and likely also occurs in Xenopus given its vast maternal stores of annulate lamellae (Dabauvalle et al., 1991; Cordes et al., 1995; Beckhelling et al., 2003). Arguing against this idea is the fact that intact emerin-treated nuclei exhibited continuous nucleoporin staining (Supplemental Figure S2, A and B), suggesting that the permeability barrier of the NE remained intact. Another model is that egg extract nuclei are more susceptible to rupture, particularly when emerin is not properly integrated into the membrane, as in our heterologous system. In that scenario, adding lamin A or properly targeting emerin to the nucleus might generate a more rigid NE that is resistant to breakage (Lammerding et al., 2006; Rowat et al., 2006; Pajerowski et al., 2007; Guilluy et al., 2014; Neelam et al., 2015; Stephens et al., 2017). It is also possible that emerin–emerin interactions might aberrantly tether disparate membrane structures within the cytoplasm (Berk et al., 2014; Herrada et al., 2015; Samson et al., 2017). Shear force applied to this network of interactions might lead to membrane tearing that propagates to the nucleus through ER–NE contacts, leading to nuclear breakage. Future work will address the mechanism of emerin-induced nuclear rupture.

While egg extract and early-stage embryonic nuclei rupture upon emerin addition, nuclei isolated from later-stage embryos are resistant to emerin-induced breakage, possibly because the biochemical composition of these nuclei has changed. Zygotic transcription is massively up-regulated at stage 8, coinciding with the MBT (Newport and Kirschner, 1982; Collart et al., 2014). In particular, post-MBT embryos express lamin A (Wolin et al., 1987; Gareiss et al., 2005; Jevtic et al., 2015) and up-regulate TRC40 transcription (Peshkin et al., 2015), potentially explaining why post-MBT nuclei no longer rupture upon emerin addition. Consistent with this idea, GFP-emerin exhibited normal NE localization in some stage 12 cells (Supplemental Figure S4). While emerin induces rupture of early-stage embryonic nuclei, embryo death only manifests later in development, when cell motility and gastrulation movements may impart sufficient force for nuclear rupture (Chien et al., 2015). It is also possible that nuclear rupture in emerin-injected embryos occurs before the onset of bulk MBT transcription, potentially explaining why emerin-injected embryos exhibit a post-MBT loss in viability while post-MBT nuclei isolated from uninjected embryos are resistant to emerin-induced breakage.

Because emerin is not normally present in Xenopus eggs and early-stage embryos, it is possible that ectopically added emerin is not present in its native state due to improper posttranslational modification or membrane insertion. Given the heterologous nature of this system, what might be the physiological relevance of our findings? First, the phenotype we observe is striking and suggests that emerin possesses an inherent ability to induce nuclear rupture, even if this activity is not evident in most cellular settings. Second, emerin is not expressed in early embryogenesis, and our results indicate that the presence of emerin would be deleterious for development. This suggests that nuclei in early embryos possess unique structural and/or mechanical properties that change later in development when emerin is expressed. Regulated changes in nuclear structure/mechanics might be important for normal development; for instance, nuclear structure may be simplified in early development due to rapid cell divisions, and as a result these nuclei may be weaker and more susceptible to breakage. Such developmental changes in nuclear mechanics have been suggested in Drosophila (Hampoelz et al., 2016). There is a growing literature on NE rupture and repair (Vargas et al., 2012; Denais et al., 2016; Raab et al., 2016; Hatch, 2018), and our emerin findings might be relevant in nondevelopmental contexts where nuclei may be more deformable, for instance, in lamin-deficient cells, motile cancer cells, and certain immune cells (Sullivan et al., 1999; Lammerding et al., 2004, 2005; Denais and Lammerding, 2014; Denais et al., 2016; Raab et al., 2016). Consistent with this idea, we found that lamin A, which rigidifies the NE (Lammerding et al., 2006; Pajerowski et al., 2007; Neelam et al., 2015; Stephens et al., 2017), abrogated emerin-induced nuclear breakage. Third, while a cytoplasmic pool of emerin likely plays important roles in normal cells (Salpingidou et al., 2007), emerin often exhibits a shift toward the endoplasmic reticulum in EDMD, as well as in some laminopathies caused by mutated lamin A (Ellis et al., 1998; Fairley et al., 1999, 2002). Perhaps, when emerin’s distribution between the nucleus and cytoplasm is altered, nuclear and ER stability are impacted and emerin’s nuclear rupturing activity becomes physiologically relevant, potentially contributing to disease. We can speculate that EDMD disease phenotypes manifest in muscle cells because these cells experience forces capable of disrupting nuclear structure. Changes in NE protein expression levels and localizations are also frequently observed in cancers (Chow et al., 2012; Denais and Lammerding, 2014; Jevtic and Levy, 2014). Finally, given the surprising impact that emerin can have on nuclear structure and embryo viability, new therapeutic strategies might involve altering the localization of NE structural proteins.

MATERIALS AND METHODS

Plasmids

The human emerin gene was cloned into pDEST17 using the Gateway LR Clonase II kit (ThermoFisher 11791-020) and donor plasmid HsCD00081619 (DNASU) to generate pDL87. The SNAP tag was amplified from pSNAP (T7)-2 (NEB) by PCR and cloned into pET30a at NcoI and SacI to generate pDL96. Human emerin was amplified from pDL87 by PCR and cloned into pDL96 at SalI and NotI to generate SNAP-emerin pET30a (pDL97). The TRC40-EMD expression plasmid was a gift from Ralph Kehlenbach (Georg-August-Universität, Göttingen, Germany; Pfaff et al., 2016). The transmembrane domain of emerin (amino acids 228–242) was deleted from pDL87 using site-directed mutagenesis to generate pDL104 (Berk et al., 2014). The fragment corresponding to EGFP-FLAG-cGAS was removed from Addgene plasmid #86675 (pTRIP-CMV-GFP-FLAG-cGAS; Raab et al., 2016) with KpnI/SalI and cloned into pRSETA at KpnI/XhoI to generate pRSETA EGFP-FLAG-cGAS (pDL105). Human emerin was amplified from pDL87 by PCR and cloned at BamHI/XhoI into a pET30b vector containing the GFP sequence at EcoRV/BamHI to produce a GFP–emerin fusion (pDL98). The BamHI/XhoI fragment from pDL98 was ligated into pCS2+ to generate emerin pCS2+ (pDL113). The EcoRV/XhoI fragment from pDL98 was ligated into pCS2+ to generate GFP-emerin pCS2+ (pDL114). The GFP-NLS pCS2+ plasmid (pDL20) was described previously (Edens and Levy, 2014). The H2B-RFP pCS107 plasmid (pRH199) was a gift from Rebecca Heald (University of California, Berkeley).

Proteins

For emerin expression and purification, pDL87, pDL97, and pDL104 were transformed into BL21(DE3)RIL+ cells (Agilent). Cultures (1 l) were grown at 37°C to OD_{600 nm} 0.5–0.7 and induced with 1 mM isopropyl β-d-1-thiogalactopyranoside at 37°C for 4 h, or in some cases overnight. Cells were harvested and stored at –80°C. Cell pellets were thawed on ice, resuspended in 40 ml cell lysis buffer (100 mM NaH₂PO₄, 10 mM Tris-Cl, 150 mM NaCl, 1 mM dithiothreitol, 10 µg/ml leupeptin, 10 µg/ml pepstatin, 10 µg/ml chymostatin, 1 mM phenylmethylsulfonyl fluoride, Sigma S8830 protease inhibitor tablet, 1.9 mg/ml lysozyme, pH 6.0), and incubated on ice for 15 min. Cell lysates were subjected to sonication (Branson, 20% power, 4 min total, 10 s on/20 s off cycles). Lysates were centrifuged at 15,000 rpm in a JA-20 rotor for 15 min at 4°C. The supernatant was discarded, and pellets were resuspended in 20 ml cell lysis buffer supplemented with 1% Triton X-100. After another round of centrifugation, the supernatant was again discarded and pellets were resuspended in 20 ml denaturation buffer (100 mM NaH₂PO₄, 10 mM Tris-Cl, 8 M urea, 3 mM 2-mercaptoethanol, pH 8.0) using a Dounce homogenizer. After another round of centrifugation, the supernatant was collected and dialyzed against dialysis buffer (10 mM HEPES, 300 mM KCl, 50 mM sucrose, 5% glycerol, pH 7.8) using Slide-A-Lyzer dialysis cassettes (10,000 MWCO, Thermo 66380) at 4°C. The first round of dialysis was with 1 l of dialysis buffer for 4 h, followed by a second round of dialysis using another 1 l of dialysis buffer overnight. Protein aliquots were stored at –80°C. To approximate protein concentrations, emerin stock solutions and BSA standards were separated on SDS–PAGE gels and stained with Coomassie. Band intensities were measured using ImageStudio (Li-Cor). In some cases the denaturation buffer was supplemented with 10 mM deoxycholic acid and the dialysis buffer consisted of 50 mM Tris-Cl (pH 8), 150 mM KCl, 5% glycerol, and 10 mM deoxycholic acid. In other instances, dialyzed emerin was subjected to Ni-NTA affinity chromatography and an additional round of dialysis. Emerin purified using each of these different schemes induced similar frequencies of nuclear breakage in extract. SNAP-emerin generated from pDL97 was labeled with Janelia Fluor 549 (Grimm et al., 2015) or Alexa Fluor 488 according to the manufacturer’s protocol (S9129, NEB). The TRC40-EMD complex was purified as previously described (Pfaff et al., 2016). GST-GFP-NLS (pMD49) was purified as previously described (Levy and Heald, 2010). Lamin A was expressed and purified from pDL36 as previously described (Jevtic et al., 2015). EGFP-cGAS was expressed from pDL105 and purified by Ni-NTA affinity chromatography as previously described (Kranzusch et al., 2013). Recombinant human Sec61β protein was obtained from Abcam (ab161213).

Western blots

Western blots were performed as previously described (Vukovic et al., 2016). The primary antibody against emerin (MANEM1 5D10; Wolfson Centre for Inherited Diseases) was used at 1:100. IRDye 680RD anti-mouse (Li-Cor 925-68070) secondary antibody was used at 1:20,000. Blots were scanned on a Li-Cor Odyssey CLx instrument, and band quantification was performed with ImageStudio.

Xenopus egg extract experiments

In vitro nuclear assembly using X. laevis egg extract and demembranated X. laevis sperm nuclei was performed as previously described (Edens and Levy, 2014; Chen and Levy, 2018). Briefly, reactions consisted of 100–500 µl of metaphase-arrested egg extract supplemented with energy mix, 0.15 mg/ml cycloheximide, 0.5 mM CaCl₂, and demembranated X. laevis sperm nuclei at 1000/µl. After 1 h of nuclear assembly at room temperature, nuclear functionality was assessed by adding 40 µg/ml recombinant GST-GFP-NLS and confirming its nuclear import. Hoechst was added at 10 µg/ml to visualize the DNA. To preassembled nuclei, varying concentrations of recombinant emerin protein were then added (not to exceed 10% of the reaction volume). Unless otherwise noted, the usual final concentration of emerin added was 5 nM. Control reactions were supplemented with equal volumes of dialysis buffer. After a 30-min incubation at room temperature, 5 µl of each reaction was pipetted onto a glass slide, overlaid with an 18 × 18 mm square coverslip, sealed with Valap (equal parts petroleum jelly, lanolin, and paraffin), and imaged immediately. Nuclear breakage was assessed by loss of nuclear GFP-NLS signal and/or the appearance of stringy/punctate Hoechst-stained DNA. For some experiments, nuclei were preassembled with 1 nM recombinant lamin A protein before addition of emerin or dialysis buffer. Where indicated, extract was supplemented with 0.2 mg/ml WGA (Sigma L9640) or 1 µM lipophilic dye CM-DiI (ThermoFisher C7001). To test the effect of emerin on smaller egg extract nuclei, the assay was performed in the same way, except that nuclear assembly and growth were allowed to proceed for only 30 min before addition of emerin or dialysis buffer. Where indicated, spindowns were performed as previously described: nuclei were fixed, spun down onto coverslips, and processed for immunofluorescence (Edens and Levy, 2014, 2016). Primary antibodies used for spindowns included α-emerin (mouse; MANEM1 5D10; Wolfson Centre for Inherited Diseases; 1:50), mAb414 (mouse; BioLegend #902901; 1:1000), and α-Man1 (rabbit; ThermoFisher Scientific #PA5-72734; 1:100). Secondary antibodies were used at 1:1000 and included Alexa Fluor 568 conjugated anti-mouse immunoglobulin G (IgG) and Alexa Fluor 488 conjugated anti-rabbit IgG. Where indicated, nuclei in extract were incubated for 10 min with 10 µg/ml mAb414 conjugated to Alexa Fluor 594 (BioLegend #682202). All Xenopus procedures and studies were conducted in compliance with the U.S. Department of Health and Human Services Guide for the Care and Use of Laboratory Animals. Protocols were approved by the University of Wyoming Institutional Animal Care and Use Committee (Assurance #A-3216-01).

Xenopus embryo experiments

X. laevis embryos and embryo extracts were prepared as previously described (Edens and Levy, 2014, 2016). Different-stage embryo extracts containing endogenous embryonic nuclei were supplemented with recombinant emerin protein or dialysis buffer. Nuclear breakage was assessed by appearance of stringy/punctate Hoechst-stained DNA. Embryo microinjection was performed as previously described (Jevtic and Levy, 2015). Recombinant proteins were microinjected at the one-cell stage. Microinjection volumes were 10 nl, and emerin stock concentrations were adjusted so that the final concentration in the embryo was ∼0.5 nM. Control embryos were microinjected with equal volumes of dialysis buffer. Emerin, GFP-emerin, GFP-NLS, and H2B-RFP mRNA were synthesized from pDL113, pDL114, pDL20, and pRH199, respectively, by first linearizing the plasmids with NotI and then expressing the mRNA from the SP6 promoter using the mMessage mMachine kit (Ambion). Purified mRNAs were resuspended in water. One-cell embryos were microinjected with 10 nl of mRNA, and the stock concentrations were adjusted to vary the amount of mRNA delivered. Water microinjections were used as a control. Embryos were allowed to develop to different stages, and embryos that stopped cleaving or appeared white and puffy were scored as apoptotic/inviable (Johnson et al., 2010; Du Pasquier et al., 2011; Tokmakov et al., 2011; Gillespie et al., 2012; Willis et al., 2012; Iguchi et al., 2013; Broadus et al., 2015). To visualize nuclei in microinjected embryos, single embryos were placed on a glass slide with 5 µl nuclear fix solution (Chen and Levy, 2018) and squashed under a coverslip.

Microscopy and image quantification

Wide-field microscopy was performed using an Olympus BX63 upright wide-field epifluorescence microscope. This system is equipped to perform multimode time-lapse imaging using an X-Cite 120LED illumination system. Image acquisition was with a high-resolution Hamamatsu ORCA-Flash4.0 digital CMOS camera at room temperature. Objectives included the UPLFLN20X and UPLSAPO40XS (Olympus). X–Y and Z positions were controlled by a fully motorized Olympus stage. Acquisition and automation were controlled by Olympus cellSens imaging software, and image analysis was performed using Metamorph software. Images for measuring fluorescence intensity were acquired using the same exposure times. For publication, images were cropped using ImageJ, but were otherwise unaltered. To quantify the number of intact nuclei with intranuclear GFP-NLS signal, 25- or 100-mm² images were acquired by image stitching using a UPLFLN20X objective (Olympus). Images were analyzed using Stream View (Olympus), and intact nuclei were counted over the entire scan area. Intact nuclei were identified as being both GFP⁺ and Hoechst⁺. Embryo images and movies were acquired with an Olympus SZX16 research fluorescence stereomicroscope, equipped with Olympus DP72 camera, 11.5× zoom microscope body, and SDFPLAPO1XPF objective. Where indicated, confocal imaging was performed on a spinning-disk confocal microscope based on an Olympus IX71 microscope stand equipped with a five-line LMM5 laser launch (Spectral Applied Research) and switchable two-fiber output to facilitate imaging through either a Yokogawa CSU-X1 spinning-disk head or a TIRF illuminator. Confocal images were acquired with an EM-CCD camera (ImagEM; Hamamatsu). Z-axis focus was controlled using a piezo Pi-Foc (Physik Instrumente), and multiposition imaging was achieved using a motorized Ludl stage. An Olympus UPLFLN 40× (NA 0.75, air) objective was used. Image acquisition and all system components were controlled using Metamorph software.

Microfluidic design, fabrication, and operation

Microfluidic devices were designed to expose flowing nuclei to and incubate them within an emerin solution via the hydrodynamic focusing of discrete, coflowing inlet streams (Knight et al., 1998; Kauffmann et al., 2001; Stiles et al., 2005). To generate slow, steady flows in which nuclei could be imaged during elongation, fluids were passively pumped using a combination of degassed poly(dimeth­ylsiloxane) (PDMS) initial filling, permeation-driven flow, and hydrostatic pressure at each fluid inlet (Zhou et al., 2007; Gnyawali et al., 2017; Komeya et al., 2017). Hydrodynamic-focusing microfluidic devices were designed to include three inlet reservoirs feeding channels that merge at a cross-junction and flow to a large, high–surface area fluid reservoir with a single outlet. Devices were fabricated via well-documented soft lithography replication protocols (Xia and Whitesides, 1998). Briefly, a 50-µm layer of negative photoresist (SU8 50; MicroChem, Westborough, MA) was spin-coated upon a 100-mm silicon wafer and patterned by exposure to 365-nm UV light through a shadow mask (CAD/ART Services, Bandon, OR). Following a developer soak to dissolve unexposed resist, photoresist-on-silicon masters were buried in PDMS, degassed, and baked for 4 h at 70°C to cure them. Individual devices were cut and trimmed from the PDMS slab, inlet/outlet holes were punched, and the devices were bonded to a cover slip. Prior to filling, devices were placed in a vacuum chamber under house vacuum for ∼1 h and removed, and each inlet reservoir was filled with an aliquot of XB buffer (100 mM KCl, 0.1 mM CaCl₂, 1 mM MgCl₂, 50 mM sucrose, 10 mM HEPES-KOH, pH 7.8). The flux of air from channels into the degassed PDMS matrix filled the channel network in a rapid, bubble-free manner. After initial device filling, excess buffer was removed from the inlets by dabbing with a Kimwipe, and each process fluid (emerin or in vitro–assembled nuclei in egg extract) was introduced to its respective inlet reservoir by the insertion of filled Tygon tubing. Nuclei in extract were labeled by adding Sytox Green at 1:1000 (Life Technologies; S7020, 5 mM stock). Flow was maintained by the continued permeation of fluid into the PDMS matrix and balanced by hydrostatic pressure achieved by controlling the height of each fluid in the tubing.

Statistics

Averaging and statistical analysis were performed for independently repeated experiments. Two-tailed Student’s t tests assuming equal variances were performed in Excel (Microsoft) to evaluate statistical significance. The p values, numbers of independent experiments, and error bars are given in the figure legends.

Abbreviations used:

EDMD

Emery–Dreifuss muscular dystrophy

MBT

midblastula transition

NE

nuclear envelope

PDMS

poly(dimethylsiloxane)

WGA

wheat germ agglutinin.