Introduction

The plasma membrane integrity is the corner stone of cell biology. Based upon, , current models further propose that cell division should also occur with uninterrupted membrane integrity and the invagination of this intact membrane, driven by a contractile ring of actin filaments, myosin II, and regulatory proteins (Pollard and O’Shaughnessy, 2019), with vesicle fusion supplying new membrane to accommodate shape change (Gerien and Wu, 2018), progressively deepens until pinching the cell into two daughter cells (Figure 1A). Critically, the foundational premise of uninterrupted membrane integrity during cytokinesis has never been rigorously tested until our unexpected discovery in C. elegans embryos.

Figure 1.

After fertilization, the single cell C. elegans embryo begins a series of highly stereotyped cell divisions. During this phase, cells undergo rapid division (15-30 minutes per cycle), with each cycle alternating between S (DNA synthesis) and M (mitosis and cytokinesis) phases while bypassing G1and G2 gap phases until the 28-cell stage (Edgar and McGhee, 1988; Deppe et al., 1978; Sulston et al., 1983; Bao et al., 2008; Pintard and Bowerman, 2019). Around this stage, the worm embryos are laid by the mother. Subsequent development continues outside the mother, encapsulated within a protective, multi-layer extracellular matrix known as the eggshell (Stein, 2018). This impermeable eggshell is established earlier, by the end of anaphase II of meiosis, and undergoes its final modifications during the embryo’s first few mitotic divisions (Bembenek et al., 2007; Zhang et al., 2005).

During routine ultrastructural analysis (Zhu et al., 2022), we found that rather than maintaining membrane continuity, the dividing C. elegans embryonic cells routinely exhibit plasma membrane discontinuities. Further analysis revealed that the plasma membrane of the embryonic cells ruptures as early as metaphase and sister membrane primarily extend independent from each other to achieve cytoplasmic partition (Figure 1B), a mechanism never before documented. This discovery challenges the canonical view of cytokinesis at its core, revealing an unanticipated plasticity in cell division mechanisms and prompting a re-evaluation of how membrane dynamics are regulated across species.

Results

Widespread membrane discontinuities in dividing C. elegans embryonic cells

Our study was initially prompted by the curious observation that the plasma membrane of embryonic cells in C. elegans does not appear to be continuously intact. We refer to these membrane discontinuities as membrane openings (MOs). They are particularly evident at the embryo’s periphery. As shown in Figure 1C, the plasma membrane is absent in the region marked by the white dashed line. Membrane interruptions are also be observed in the embryo’s central regions. For instance, Figure 1D shows that Cell 3 contains a MO facing both Cell 1 and Cell 2. Additionally, we noted instances where a single plasma membrane (SPM) appears to traverse the cytoplasm, seemingly partitioning it into separate compartments across adjacent cells (Figure 1D) (yellow arrows). According to the conventional view, the plasma membrane should remain intact throughout the cell cycle—whether during division or interphase—and each cell should be enclosed by its own membrane. Consequently, between two adjacent cells, we would expect to observe two distinct plasma membranes. The presence of a single continuous membrane, however, suggests the existence of regions lacking a complete boundary. For instance, Cell 4 in Figure 1D clearly contains such membrane discontinuities. Collectively, these findings challenge the established notion that the plasma membrane remains intact over the entire lifetime of a cell.

To rule out potential artifacts inherent to electron microscopy (EM) techniques, we performed a series of control experiments. First, we prepared four independent sample batches. Membrane discontinuities were consistently observed in all batches, indicating that the phenomenon is not specific to a single preparation. Second, we examined a publicly available historical collection of C. elegans embryo EM images (Wormatlas) and identified MOs in several instances, particularly at the embryonic periphery. This confirmation in an independent dataset suggests that MOs are genuine cellular structures rather than procedural artifacts. Third, we analyzed embryos across various developmental stages. MO-containing cells were observed in both intrauterine and already laid embryos at early stages (Figure S1). However, as embryos reached the comma stage (Figure S2)—a period characterized by the onset of elongation and a decline in active cell proliferation—the incidence of MOs dropped dramatically (observed in 0/13, 0/17, 0/30, and 0/30 cells examined, respectively). These results demonstrate that plasma membrane discontinuities are associated with specific stages of C. elegans embryogenesis and are not a product of random membrane tearing.

The early stage of C. elegans embryonic cells is under rapid and extensive cell division. Within approximately 100 minutes, a two-cell embryo develops into an embryo containing nearly 30 cells (Bao et al., 2008). Given this dynamic proliferation, we suspected that the formation of membrane discontinuities might be associated with cell division events. Therefore, to systematically examine the formation of membrane discontinuities, we collected EM images of 60 early embryos from 9 gravid adults. By examining the entire circumference of 425 cells in single high-pressure freezing transmission electron microscopy (HPF-TEM) sections (Figure S2) (Modern Electron Microscopy Methods for C. elegans, 2012), we found that MOs in 159 cells distributed across 57 embryos, representing 95% of the total sampled embryos (Figure 1E). Among these 159 affected cells, 89% (141/159) exhibited a single membrane discontinuity, while 13 cells had two openings and 5 cells had three.

We analyzed the 159 cells exhibiting MOs based on nuclear and cytoplasmic morphology (Figure 1E). Of these, 64 cells appeared in EM sections containing only cytoplasmic regions—9 of which completely lacked discernible plasma membranes. The remaining 95 cells contained one or more nuclei or condensed chromosomes. Among them, 10 cells were clearly in metaphase stage (defined by a single chromosomal cluster). As shown in Figure 2A, a representative metaphase cell (MC) displayed a MO (dashed white line) spanning roughly one-third of the cell’s circumference. High-magnification imaging revealed that in contrast to the wrinkled plasma membranes (Figure 2A, enlarged boxed area displayed in 1), the large MO exhibited a smoother surface morphology, and the internal cytoplasm remained enclosed within the cell (Figure 2A, enlarged boxed areas displayed in 2 and 3). MOs were also observed in anaphase cells (AC) (n=5; characterized by <10 chromosomal clumps; Figure 2B, enlarged in 2) and telophase cells (n=5; defined by two nuclei sharing cytoplasm). Intriguingly, one telophase-like cell contained three nuclei within a common cytoplasm (Figure S2, marked by &).

Figure 2.

Because EM sectioning captures cells at random stages of the cell cycle, we sought to systematically assess the relationship between mitosis and MOs. To do so, we analyzed 20 cells with unambiguous division features, selected based on nuclear morphology—such as aligned chromosomes (metaphase) or discrete clusters (anaphase)—and cytoplasmic sharing, indicated by two or more nuclei within a common cytoplasm (telophase). Strikingly, 19 out of 20 cells (95%) revealed MOs upon tracing the plasma membrane in single EM sections (Figure S3). This near-ubiquitous association strongly suggests that membrane discontinuities are an integral feature of cell division in early C. elegans embryos.

Membrane rupture in cells at metaphase

The observed membrane discontinuities suggest that the plasma membrane of early C. elegans embryos may rupture during cell division. However, because TEM provides only static, two-dimensional images, we could not determine the three-dimensional nature of these ruptures. To better understand this phenomenon, we acquired continuous serial sections and performed 3D reconstruction.

Since membrane discontinuities appear at metaphase—well before nuclear or cytoplasmic division—we performed continuous EM sectioning to trace the entire plasma membrane of a metaphase cell (Figures 2C). Through 3D reconstruction, we identified a continuous plasma membrane rupture extending over 8,940 nm, which is more than two-thirds of the cell’s circumference (Figure 2D). Notably, one end of the membrane rupture appeared long and narrow, while the other was wide and exhibited cytoplasmic leakage (Figure 2D and Video S1). Further analysis revealed that the leaked cytoplasm remained confined to a relatively localized region (Video S1). During metaphase, chromosomes align at the equatorial plane, a key region for subsequent segregation. To explore the spatial relationship between membrane rupture and chromosomal organization, we further reconstructed the entire condensed chromosome region (Figures 2C). We found that this large-scale membrane rupture largely encircled the chromosomal disc (Figure 2D and Video S1), spatially aligning with the future segregation zone.

The detachment of sister membranes

According to the conventional model, cytokinesis in animal cells begins with the formation of a membrane invagination. As this invagination deepens, it gives rise to two back-to-back membrane layers, ultimately leading to the physical separation of the two daughter cells. In the context of membrane rupture, however, plasma membrane disruption results in the separation of two putative sister membranes. We then asked how the subsequent cytokinesis process proceeds with detached sister membranes. To address this question, we examined 51 dividing cells. In each of these cells, two well-separated nuclei were observed, indicating that they were at telophase. We found that while a minority of cells (6/51) exhibited classical membrane invagination (Figure 3A), including one embryo transitioning from the one-cell to the two-cell stage, the vast majority (45/51) displayed a striking detached membrane configuration (Figure 3B). In these cases, sister membrane tips were completely separated from each other (Figures 3B and 3C, marked with *), forming distinct, non-contiguous structures that defy conventional models of cytokinesis. This predominant pattern suggests that membrane detachment—rather than continuous invagination—may represent the primary mode of cytoplasmic separation in C. elegans embryos. Of note, the membrane rupture simultaneously generates individual sister membranes with severed edges. Indeed, as shown in Figure 3B, the tips of both free-ended edges (*) of Cell 1 are visible in the EM section.

Figure 3.

In the detached membrane configuration, cytoplasmic components—such as glycogen granules and ribosomal particles—were clearly visible surrounding the severed edges of the two sister membranes (Figure 3C). This observation is particularly significant because the classical invagination model cannot generate single plasma membrane barriers between adjacent cytoplasmic domains. Instead, independent extension of severed sister membranes could explain the formation of cytoplasm-immersed membranes. To further test this hypothesis, we traced 12 cytoplasm-embedded membranes, including those separating Cell 1 from Cells 2 and 3, and those separating Cell 4 from 3 (displayed on Figure 3D, enlarged in 1 and 2). In all cases, the cytoplasmic regions flanking these membranes belonged to cytosolically interconnected sister cells. Together, these results demonstrate that C. elegans embryonic cytokinesis relies on detached sister membrane extension rather than invagination mediated by intact membranes.

Tracing from the severed edges back to their origins, we found that the sister membranes were frequently anchored side-by-side near the cytosol-immersed region (Figures 4A-4D). Within this region, gap junction plaques were often observed (57 in 94 cases) at membrane contact sites (Figure 4B), though their potential role in sister membrane cohesion is unclear. These contact sites alternated with regions where the two sister membranes separated, revealing lightly stained intercellular spaces in EM images. Beyond the anchored region, the separation between sister membranes became evident and each membrane exhibited the characteristic features of a typical plasma membrane—with one side facing the cytosol and the other exposed to the extracellular environment (Figure 4D).

Figure 4.

Cellular vesicles were observed near free membrane edges (Figures 4A-4C). Some cytosol-immersed membrane tips appeared connected to vesicles (Figure 4A), with instances of direct contact (Figure 4B) and possible fusion events (Figures 4B and 4C). Vesicles also clustered along the anchoring area, as well as the separated sister plasma membranes (Figure 4D). EM analysis revealed no morphological differences between vesicles associated with the cytosol-immersed region and those near conventional regions of the sister membranes (Figures 4B and 4D).

Analysis of a tri-nucleated cell during cytokinesis

To gain more comprehensive insight into membrane organization during cytokinesis, we further applied the AutoCUTS-SEM (automatic collector of ultrathin sections scanning electron microscopy) workflow (Li et al., 2017) to acquire serial EM sections of a telophase cell. Through continuous tracing, we discovered that this cell actually contained three independent nuclei. We color-coded each membrane region according to its associated nucleus, with relative sizes estimated from EM sections to ensure proportional representation (Figure 5A). The 3D model demonstrated that the division plane between the yellow and blue cells was roughly perpendicular to that between the purple cell and the other two (Figure 5A). The presence of three nuclei within a shared cytoplasm suggests that one of the daughter cells entered the next round of cell division without completing the previous one. This phenomenon—concurrent divisions—appears to be common in early C. elegans embryos, as our EM analysis identified more than 10 cells with three or four nuclei. For example, Figure 3D shows an EM section of a cell containing four nuclei.

Figure 5.

Using consecutive EM sections, we traced membrane integrity in this tri-nucleated cell. We found that this tri-nucleated cell was enveloped by three distinct plasma membranes, each containing a single rupture site. These three ruptures converged to form a “Y”-shaped exposed cytoplasmic region spanning >351 sections (21,060 nm; Figures 5A and 5B and Video S2). Throughout these 351 consecutive sections, we detected no continuity between adjacent membranes, indicating that membrane independence had been established by this stage. Consecutive tracing further showed that the MOs—frequently observed in single EM sections (Figures 1 and 2)—could arise from either the separation between sister membranes (Figure 5C) or from separated edges of the plasma membrane within the same cell (Figure 5D). Additionally, multiple openings within a single cell might represent different cross-sectional planes of a common membrane rupture (Figure 5E).

In all three cells (where cytoplasmic separation was incomplete), the free membrane edges were in close proximity (marked by *; Figures 5F and 5G), suggesting they were in late cytokinesis and that membrane extension mediated by detached membranes could persist into this stage. Taken together, these findings demonstrate that this tri-nucleated cell was undergoing concurrent cytokinesis, with three detached membranes independently extending toward a common cytoplasmic region to achieve cytoplasm individualization (Figure 5H). A similar mechanism may operate in four-nucleated cells (Figure 3D, for instance), in which two detached daughter membranes—without completing the last round of cell division—undergo further rupture, forming four granddaughter membranes. These membranes then extend independently to complete cell division (Figure 5H).

Capturing the membrane rupture and independent extension of detached sister membranes in living embryos

While high resolution HPF-TEM analysis revealed widespread membrane discontinuities in early C. elegans embryos, this static imaging approach cannot capture dynamic membrane behavior. To investigate whether optical microscopy can visualize membrane ruptures and the movement of severed sister membranes, we employed a fluorescence-labeled membrane marker, GFP::PH(PLC1δ1) (see Methods and Materials for details). Our live imaging detected both putative membrane ruptures (Figure 6A) and free-ended sister membrane structures (Figures 6B and 6C), providing additional evidence that membrane rupture and independent extension of detached sister membranes underlie cytokinesis in C. elegans embryos (see Discussion).

Figure 6.

Sealing the exposed cytosol by neighboring membranes and extracellular matrix layers

The integrity of the plasma membrane is essential for cell survival, as its disruption can lead to cytoplasmic leakage and, if severe, result in cell death. However, widespread cytoplasmic leakage has not been observed in wild-type C. elegans embryos. Based on our EM observations, we propose two mechanisms through which dividing embryonic cells may seal or cover exposed cytosolic regions.

First, the sharing of plasma membrane with neighboring cells may protect the exposed cytosol. As repeatedly shown in EM images in this study (Figures 1, 3 and 4), a single elongated free-ended membrane from one cell consistently helps define boundaries with adjacent cells. These neighboring cells may be direct sister cells derived from the same mother cell, or cousin cells originating from mother cells that were once sisters. In cases where three nuclei share a common cytosol, the adjacent cell could even be an aunt or nephew cell.

In addition, eggshell components may seal cytoplasmic regions lacking membrane coverage. The C. elegans eggshell is composed of six distinct extracellular matrix layers (Olson et al., 2012; Benenati et al., 2009; Rappleye et al., 1999; Schierenberg and Junkersdorf, 1992; Stein, 2018). Among these, the fifth permeability barrier layer (PBL), a lipid-rich structure, serves as the osmotic and permeability barrier for the developing embryo (Olson et al., 2012; Benenati et al., 2009). Normally, the PBL encloses the peri-embryonic layer (PEL), which separates it from the embryonic plasma membrane (Olson et al., 2012; Benenati et al., 2009) (Figures 7A and 7B). In contrast, in regions where the plasma membrane was absent, we observed the PBL adhering directly to the exposed cytoplasm (red triangles and dashed red lines, Figures 7B and 7C), bypassing the PEL. This phenomenon was also evident in other EM images (Figures 1C and Figure 2, box 1). PBL-mediated sealing is not always complete. When the PBL was not accessible or only loosely covered membrane-free regions, minor cytoplasmic leakage occurred (Figure 7D). In cases where PBL sealing was ineffective or absent, cytoplasmic material appeared to be retained within a confined region, likely by the PEL (Figure 7D). However, the mechanisms by which extracellular matrix components protect the exposed cytoplasmic region remain unknown.

Figure 7.

Discussion

In this study, we discovered that the plasma membrane of C. elegans embryonic cells undergo rupture at metaphase. Following this rupture, the detached sister membranes extend rather independently from each other to complete cytokinesis. To date, the ultrastructural dynamics of plasma membrane integrity during cell division remained unexplored in other species. We anticipate that our findings will catalyze comparative studies across different species or under various pathological conditions, to determine whether this mechanism represents a fundamental paradigm or a context-specific strategy. Our work established the first mechanistic framework to guide such future investigations.

The advantage of membrane rupture and independent sister membrane extension

Cell division involves both nuclear and cytoplasmic division. However, due to the difficulty of visualizing cell boundaries in living embryos, previous studies in C. elegans have largely emphasized nuclear events. Here, we observed that early C. elegans embryonic cells can contain three or four nuclei within a shared cytoplasm, indicating that cytoplasmic division may be uncoupled from nuclear division. This phenomenon is not unique to C. elegans—it has also been reported in other species such as Drosophila, where the early embryo undergoes 13 rapid nuclear divisions without cytokinesis before cellularization occurs. It has been proposed that by minimizing or delaying cytokinesis, embryos can accelerate nuclear multiplication, thereby supporting the rapid pace of early development. In this context, the embryonic cells of C. elegans appear to adopt a notably rapid—though seemingly risky—strategy for cell proliferation, yet one with clear developmental benefits.

First, membrane rupture may help alleviate the rise in intracellular pressure during division (Stewart et al., 2011; Hiramoto, 1963). We observed that the size of membrane discontinuities appeared to peak during metaphase, primarily around the equatorial plane-consistent with localized pressure surges at this stage. Notably, ruptures were also detected in tri-nucleated cell clusters, where all three cells exhibited membrane discontinuities during late cytokinesis. This suggests that even though the relative size of the ruptures decreases as mitosis proceeds, membrane discontinuities may persist throughout the cell cycle. Maintaining an opening in the plasma membrane during division likely reduces internal pressure, thereby conserving energy and resources that would otherwise be expended on resisting membrane tension. These conserved resources could then be redirected toward essential processes such as chromosome segregation and membrane synthesis—a particular advantage for rapidly dividing embryonic cells.

Of note, 3D reconstruction revealed that the metaphase rupture does not fully bisect the membrane (Figure 2). However, given the limitations of our EM sampling—neither random nor continuous imaging provides complete coverage—we cannot exclude the possibility that some ruptures completely sever the membrane. Regardless of the extent of separation, the rupture produces two distinct membrane domains. Since this occurs as early as metaphase, before nuclear or cytoplasmic division, the physical independence of the plasma membranes may serve as an early marker of nascent daughter cells, preceding other cellular landmarks.

Following membrane rupture, the prospective sister membranes become separated, raising the question of how cytokinesis proceeds with severed membranes. Analysis of single EM sections revealed classical membrane invagination in only a few cases (6 out of 51; Figure 3A). Notably, one such instance occurred during the transition from a one-cell to a two-cell embryo. This suggests that while a subset of C. elegans embryonic cells employs the membrane invagination and extension model, the majority do not. Interestingly, membrane discontinuities were observed in one-cell embryos (Figure S1), indicating that membrane rupture can co-occur with invagination. Alternatively, in dividing cells, the back-to-back invagination may be transient and may subsequently transition to free-ended extension. At present, distinguishing between these hypotheses remains challenging. A definitive resolution will likely require advanced imaging techniques capable of high-resolution, real-time visualization at the EM level.

Unlike a continuous bilayer, a free-ended membrane offers greater flexibility during cytokinesis. Without being physically constrained by its sister, each membrane can extend in different directions as needed. This independence may be particularly advantageous in accommodating asynchrony between nuclear and cytoplasmic division, as seen in early C. elegans embryos where multinucleated cells form through incomplete cytokinesis. In such cases, independent membrane growth from two directions facilitates cytoplasmic separation more effectively than a single continuous membrane would allow. Such autonomous movement may also explain the divergent orientations of sister membranes frequently observed in EM sections.

The expansion of the plasma membrane is generally driven by vesicle fusion. In this study, vesicles seem observed randomly distributed beneath the extending sister membranes. EM analysis further showed no morphological differences between vesicles near cytosol-immersed regions and those at conventional membrane sites. Since vesicles can fuse with the membrane at various locations, we speculate that membrane growth may be enhanced under these conditions—potentially accelerating cell cleavage during early embryogenesis. However, this remains speculative and requires further experimental validation.

Finally, when the separated membrane edges meet within daughter cells, full cytoplasmic separation is achieved. In the classical invagination-extension model, cytokinesis requires both membrane scission and fusion. In the detached-membrane mode described here, however, cytokinesis proceeds without a continuous membrane structure. This raises important questions: Is membrane scission dispensable in this context? Could bypassing scission accelerate cell separation? If so, might such a mechanism offer a selective advantage during rapid embryonic development? These questions remain open and merit further investigation.

Limitations and perspectives

In this study, the application of HPF-TEM has enabled unprecedented ultrastructural visualization of dividing cell surfaces. Our findings reveal a cytoplasmic division mechanism that challenges current models and raises important new questions regarding the plasticity of cytokinesis.

HPF-TEM offers a significant advantage for membrane studies by capturing cellular structures in a near-native state. Through ultra-rapid freezing under high pressure, it suppresses ice crystal formation and preserves water in a vitreous state, thereby minimizing chemical artifacts and maintaining the integrity of lipids and proteins. This method provides a faithful snapshot of membrane architecture and dynamics(Moor et al., 1980). In our experiments, the consistent observation of limited number of membrane discontinuities-generally one per dividing cell-that are closely associated with cell division strongly suggests these are genuine biological structures rather than preparation-induced artifacts. While one might argue that such features could reflect EM-related artifacts specific to the plasma membrane during division, the correlative evidence presented here supports their biological relevance.

Using fluorescence-labeled membrane markers, we identified membrane ruptures and free-ended sister membrane structures in live embryos (Figure 6). Similar “abnormal” or “fuzzy” membrane signals have in fact been reported previously(Harrell and Goldstein, 2011). However, interpretating these structures has remained challenging for both technical and conceptual reasons. Technically, the rapid kinetics of membrane extension-about 10-30 seconds during metaphase and less than 3 minutes during cytokinesis-combined with cell motility, introduce inherent spatiotemporal ambiguities. Conceptually, the longstanding dominance of the membrane invagination model has likely shaped—and potentially constrained—interpretations of membrane dynamics. This paradigm may have led researchers to prioritize observations consistent with invagination, while dismissing ambiguous structures such as free-ended extensions as artifacts, rather than considering them as potential elements of alternative mechanisms. Furthermore, theoretical and computational models of cytokinesis have largely been built on invagination-centric assumptions, further narrowing the conceptual framework.

A key limitation is our current inability to capture 3D dynamics of these membrane structures in real time, a challenge that awaits imaging technologies with sufficient spatiotemporal resolution at near -EM levels.

Therefore, while our findings propose a transformative model for cytokinesis, they open a new field of inquiry. Fully validating this mechanism and exploring its conservation and implications will require an integrated approach combining advanced live-cell imaging, biophysical modeling, and genetic tools.

Methods and Materials

Examination of C. elegans embryonic cells

Wild-type N2 worms were cultured on OP50-seeded NGM plates at 22°C (Brenner, 1974). To obtain images of early embryos, gravid hermaphrodites were collected and sectioned. At low magnification (×4000), individual embryonic cells were identified by the presence of a nucleus. In sections where the nucleus was not visible, cytoplasmic regions—whether covered or not covered by plasma membrane—were counted as individual cells if they were separated from surrounding cells. A total of 425 cells were examined (summarized in Figure 1C), with one EM section analyzed per cell. For embryos with fewer than 10 cells (one or two rows of cells), every cell in the given section was examined. For embryos with three or more rows of cells, only cells at the periphery were analyzed. At higher magnification (×30,000), the entire surface of each cell was examined, revealing 159 cells with one or more membrane openings. Higher magnification improved the resolution of cytoplasmic continuity, reducing potential misidentification of multinucleated cells or interconnected cytoplasmic regions that might have been miscounted at lower magnification. A total of 60 early embryos were serially sectioned as completely as possible, with each embryo yielding at least 20 sections. While Figure 1C presents summary data, images from additional sections were also used for further analyses. Since these additional sections may capture different cells, the total number of embryonic cells examined likely exceeds 425.

Additionally, we examined eight embryos laid outside the worm. Among these, four pre-comma stage embryos each contained at least one MO-positive cell. In another pre-comma stage embryo, no MOs were detected after examining 17 cells. Analysis of one comma-stage embryo (19 cells examined) and two 1.5-fold-stage embryos (>30 cells analyzed each) revealed no MOs.

Identification of MOs

The entire cell surface of individual embryonic cells was examined using electron microscopy at 30,000× magnification. The isolated cytoplasmic regions without plasma membrane coverage were defined as MOs. Sites where the plasma membrane terminated and reappeared were marked as putative MOs. In periphery region of worm embryos, in the non-plasma-membrane-covering regions, the PEL was generally absent, while the PBL typically extended over the exposed cytoplasmic surface. Plasma membranes often appeared wrinkled, whereas cytoplasmic regions covered by the PBL exhibited a smoother appearance. Smooth, non-plasma-membrane-covered regions overlaid by the PBL are readily identifiable as MOs. When exposed cytoplasm was not tightly covered by the PBL, cytoplasmic leakage—still confined to a specific area by the PEL—could also indicate a MO. Given the variable morphology and frequent cytoplasmic leakage, membrane discontinuity size could not serve as a reliable criterion for defining MOs. However, to ensure robustness in our analysis, we considered only linear membrane discontinuities exceeding 200 nm as genuine openings. Each MO was examined independently at least three times. When necessary, adjacent sections were also reviewed to conform an MO’s presence, and identification was finalized only after all three assessments were positive.

High-pressure freezing and transmission electron microscopy

Basic protocols for high-pressure freezing of a range of organisms, including C. elegans, have been described (Weimer, 2006; McDonald, 2007; Manning and Richmond, 2015). Briefly, adult worms were loaded into carriers and cryofixed using a Leica Microsystems EM ICE high-pressure freezer, followed by automatic cooling in liquid nitrogen. The carriers used for freezing (Leica Microsystems, Germany; catalog nos. 16770153 and 16770152) were coated with a non-stick layer of 1-hexadecene (Sigma-Aldrich). After high-pressure freezing (HPF), the samples were transferred under liquid nitrogen to a Leica AFS-2 freeze-substitution unit and incubated at −90°C for 72 h in a freeze-substitution solution consisting of acetone with 2% (wt/vol) osmium tetroxide. The temperature was then raised at a rate of 8°C/h for 4 h, held at −60°C for 12 h, increased again at 5°C/h for 6 h, held at −30°C for 10 h, and finally raised to 10°C at a rate of 4°C/h for 10 h. Samples were washed four times in acetone, stained with 1% uranyl acetate for 1 h, and rinsed three times with pure acetone. They were then infiltrated stepwise in increasing concentrations of SPI-Chem™ resin (2:1 resin:acetone for 3 h, 1:1 for 5 h), followed by two incubations in 100% fresh resin (4 h each). The samples were placed in fresh resin in embedding molds and polymerized in a 60°C oven for 3 days. Ultrathin sections (60 nm) were cut using a diamond knife (Diatome) on an ultramicrotome (Leica Ultracut UC7) and collected on slot copper grids (EMS). The sections were visualized using a Hitachi HT7700 transmission electron microscope operating at 80 kV. Images were recorded with a Gatan 832 (4k × 2.7k) CCD camera. Data from four independent preparations were analyzed in this study.

3D reconstruction

The metaphase 3D model was prepared using the following steps. First, resin blocks containing gravid adults were prepared via high-pressure freezing and freeze substitution. Continuous ultrathin sections (60 nm thick) were manually collected using a Leica UC-7 ultramicrotome and mounted on slot grids for TEM imaging. A total of 112 serial sections were collected, though 3 sections were lost due to obstruction, resulting in 822 TEM images. Among these, 105 sections contained metaphase-stage cell information. To examine the plasma membrane integrity, high-magnification imaging was performed. Since a single high-magnification image could not cover the entire cell, multiple overlapping images were captured per section and stitched together using ImageJ to generate a complete composite image of the cellular morphology at each section. In total, 292 images were stitched. The 105 stitched images were then aligned using ImageJ, and 99 of these aligned images were used for 3D reconstruction using Imaris software. About a quarter of the cell lacked membrane discontinuities. For this portion, we estimated its size and shape based on EM sections to ensure proportional representation. Of note, single MOs smaller than 200 nm that appeared continuously across more than 3 consecutive sections were considered the endpoint of the membrane rupture. Based on this, the actual length of the membrane rupture could be underestimated at one end, while the other end was clearly defined.

The ultrastructural 3D study of the tri-nucleated cell was performed using an AutoCUTS device (Li et al., 2017) and a Helios Nanolab 600i dual-beam scanning electron microscope (FEI). Adult worms were treated with high-pressure freezing, freeze-substitution, and embedding as described above. Subsequently, 351 serial sections (60 nm thick) were collected using the AutoCUTS device, and scanning electron microscopy imaging was performed. All datasets were analyzed using IMARIS 8 software. After Z-stacks were aligned, the membrane rupture was manually traced and reconstructed.

Live imaging of plasma membrane dynamics using GFP::PH

The plasma membrane was labeled using GFP::PH(PLC1δ1). The strain carrying this marker was obtained from the Caenorhabditis Genetics Center and was originally generated in Dr. Karen Faye Oegema’s laboratory. Gravid adult hermaphrodites expressing this marker were collected and washed with M9 buffer. Embryos were then released from the gravid adult worms using a 1 mL syringe needle and transferred onto a 2% agarose pad. Fluorescence imaging was performed on a Zeiss LSM-980 microscope with a 63×/1.42 oil objective. Only one embryo at a time was imaged, starting at the 2- to 4-cell stage. Images were acquired at 4 or 8-second intervals, with total imaging duration typically kept under 30 minutes to ensure embryo viability.

Supplementary Figures

Figure S1.

Figure S2.

Figure S3.

Data availability

The current manuscript is mainly based on EM images and all data are avaliable in the manuscript.

Acknowledgements

We thank the Center for Biological Imaging at the Institute of Biophysics, Chinese Academy of Sciences (CAS) for technical support. We are grateful to Drs. John White, Andrew Chisholm, and Zhuo Du for their encouragement and insightful suggestions. This work was supported by grants from the National Key R&D Program of China (2024YFA1803401, 2021YFA0805802, and 2024YFA1306100) and the Natural Science Foundation of China (32170790 and 32321004).

Additional information

Author contributions

Study concept and design: MD, JL. EM analyses: JL, TH, MD. Drafting and major editing of original manuscript: MD, JL, TH, XH. All authors reviewed, revised, and approved the final version of the paper.

Funding

MOST | National Key Research and Development Program of China (NKPs) (2024YFA1803401)

• Mei Ding

MOST | National Key Research and Development Program of China (NKPs) (2021YFA0805802)

• Mei Ding

Additional files

Video S1. 3D reconstruction of the metaphase cell in Figure 2. The plasma membrane is shown in red, the condensed chromosome is in blue, and exposed cytoplasm is in green.

Video S2. 3D reconstruction of the tri-nucleated cell in Figure 5. 3D reconstruction of membrane ruptures and adjacent plasma membranes. Individual membranes, each associated with a distinct nucleus, are color-coded (purple, yellow, or blue).