Asymmetric nuclear division in neural stem cells generates sibling nuclei that differ in size, envelope composition, and chromatin organization

Summary

Although nuclei are the defining features of eukaryotes, we still do not fully understand how the nuclear compartment is duplicated and partitioned during division. This is especially the case for organisms that do not completely disassemble their nuclear envelope upon entry into mitosis. In studying this process in Drosophila neural stem cells, which undergo asymmetric divisions, we find that the nuclear compartment boundary persists during mitosis thanks to the maintenance of a supporting nuclear lamina. This mitotic nuclear envelope is then asymmetrically remodeled and partitioned to give rise to two daughter nuclei that differ in envelope composition and exhibit a >30-fold difference in volume. The striking difference in nuclear size was found to depend on two consecutive processes: asymmetric nuclear envelope resealing at mitotic exit at sites defined by the central spindle, and differential nuclear growth that appears to depend on the available local reservoir of ER/nuclear membranes, which is asymmetrically partitioned between the two daughter cells. Importantly, these asymmetries in size and composition of the daughter nuclei, and the associated asymmetries in chromatin organization, all become apparent long before the cortical release and the nuclear import of cell fates determinants. Thus, asymmetric nuclear remodeling during stem cell divisions may contribute to the generation of cellular diversity by initiating distinct transcriptional programs in sibling nuclei that contribute to later changes in daughter cell identity and fate.

Graphical abstract

Highlights

• •Neural stem cells maintain a nuclear lamina and envelope throughout mitosis
• •Asymmetric nuclear division forms sibling nuclei that differ in size and composition
• •The nuclear asymmetry depends on an asymmetric spindle and differential growth
• •Asymmetric nuclear division likely contributes to the generation of cell diversity

Roubinet et al. show that fly neural stem cells maintain a mitotic nuclear compartment throughout mitosis. This mitotic nuclear envelope is remodeled in a two-step process to give rise to sibling nuclei that differ in size and envelope composition prior to the nuclear import of cell fate determinants, potentially biasing daughter cell fates.

Introduction

Nuclei are the defining feature of the eukaryotic cell. Although nuclear division—the process by which a nucleus divides into two—was one of the first subjects studied by Walther Flemming and other pioneers of cell biology, it remains one of the most poorly understood aspects of the division process. While many eukaryotic cells exhibit a so-called “open mitosis,” in which the nuclear envelope is disassembled at mitotic entry before being reformed anew around chromosomes at mitotic exit,^{1, 2, 3, 4, 5} other eukaryotes undergo a “closed mitosis” in which they maintain an intact nuclear compartment and a diffusion barrier throughout.^3,6 At first sight, the mechanisms of open and closed nuclear division appear completely different. However, we recently identified a role for nuclear pore disassembly in the final steps of bridge scission during nuclear division in fission yeast—a classic model of closed mitosis.^7,8 This finding suggests that the mechanisms of open and closed mitosis differ, at least in part, as a simple consequence of shifts in the timing and position of nuclear pore disassembly. In line with open and closed mitosis sitting at either ends of a spectrum, the mechanism of nuclear division in many organisms such as flies and worms appears midway between these two extremes and is often described as a “semi-open” or “semi-closed” mitosis.^{9, 10, 11} In addition, the way nuclei divide depends on cell type. Thus, the nuclear envelope is disassembled during mitosis in Drosophila cells in culture,¹² but is maintained in the syncytial fly embryo.^9,13,14 To further investigate the mechanism of nuclear division in an intermediate “semi-closed mitosis,” we sought to study the process in Drosophila neural stem cells (neuroblasts), which divide asymmetrically to generate a large self-renewing neuroblast containing a large nucleus and a smaller differentiating ganglion mother cell (GMC) with a small nucleus.^{15, 16, 17} However, it is not known how the nuclear compartment is partitioned during these stem cell divisions, nor how nuclear division might contribute to the formation of sibling nuclei with different transcriptional profiles/identities.^18,19 In this paper we show that the nuclear envelope of neuroblasts undergoes an asymmetric remodeling during mitosis, so that the genome is partitioned into daughter nuclei that differ by more than 30-fold in volume and in envelope composition. This analysis also reveals the emergence of differences in chromatin organization between the two daughter nuclei that precede the cortical release and nuclear import of cell fate determinants. This suggests an important role for nuclear division in the generation of cellular diversity.

Results

The nuclear envelope of fly neuroblasts is maintained during mitosis and remodeled to generate two sibling nuclei that differ in size and envelope composition

To characterize nuclear division in Drosophila neuroblasts, we used markers to label plasma and nuclear membranes (CD8),²⁰ the INM (Klaroid), ONM (Klarsicht) and the ER (Bip-SfGFP::HDEL). This identified a continuous membrane boundary enveloping the spindle throughout mitosis (Figures 1A and 1B; Videos S1 and S2), which underwent asymmetric remodeling at mitotic exit to give rise to two nuclei that differ in size (yellow versus white stars). Despite the presence of this bounding membrane, the nuclear-cytoplasmic diffusion barrier was lost at prometaphase before being re-established in the two daughter nuclei in late telophase, as measured by the distribution of nuclear (NLS::GFP) and cytoplasmic (RpS13::GFP) markers (Figures 1C and S1A–S1D and Videos S3 and S4). In line with this, nucleoporins (Nup58::GFP and Nup107::GFP) were lost from the nuclear membrane upon entry into mitosis, then recruited back in an asymmetric fashion to the newly formed nuclei at mitotic exit (Figure 1D, yellow arrows; Figure S1E) prior to the cortical release of polarity proteins such as Miranda (Figure S1F). These changes to the nuclear envelope were confirmed by electron microscopy. Thus, this ultrastructural analysis revealed the formation of highly fenestrated and multilayered mitotic nuclear envelope that lacks NPCs (Figures 1E and 1F). Nevertheless, while ribosomes were seen in both compartments during mitosis (Figure 1C, bottom panel, and 1E), larger structures such as vesicles and mitochondria were effectively excluded from the mitotic nuclear compartment (Figures 1F, S1G, and S1H). These data show that Drosophila neuroblasts undergo a semi-closed mitosis, characterized by the complete disassembly of NPCs and the reorganization of the nuclear envelope into multilayered fenestrated sheets.

Figure 1.

The nuclear envelope of fly neuroblasts is maintained during mitosis and remodeled to generate two sibling nuclei that differ in size and envelope composition

(A–D) Mitotic neuroblasts expressing Cherry::Jupiter in combination with either (A) CD8::GFP, (B) Klaroid::GFP, Klarsicht::GFP or BIP-SfGFP::HDEL, (C) NLS::5GFP or Rps13::GFP, or (D) Nup107::GFP or Nup58::GFP. Yellow arrows indicate higher nucleoporin density.

(E and F) Electron microscopy performed at various cell cycle stages: (E) magnified regions are marked in yellow in (F); yellow arrowheads show nuclear pores; white arrowheads show holes in the nuclear envelope; N, nuclear compartment; C, cytoplasm. Scale bars are 5 μm.

See also Figure S1 and Videos S1, S2, S3, and S4.

Video S1. Drosophila neural stem cells maintain a nuclear compartment throughout mitosis, related to Figure 1

The nuclear envelope of neuroblasts persists during mitosis and is remodelled to generate sibling nuclei with different size. Neuroblast expressing the plasma and nuclear membranes marker CD8::GFP (white on the left panel, green on the merge) and the spindle marker Cherry::Jupiter (white on the middle panel, red on the merge). Scale bar is 5 μm. Stopwatch shows time in minutes and seconds in regards to anaphase onset.

Video S2. The mitotic nuclear envelope is labeled by the ER probe BIP::SfGFP::HDEL, related to Figure 1

The nuclear envelope of neuroblasts retains elements of its identity as neuroblasts pass through mitosis. Neuroblast expressing the Endoplasmic Reticulum luminal reporter protein BIP::SfGFP::HDEL (white on the left panel, green on the merge) and the spindle marker Cherry::Jupiter (white on the middle panel, red on the merge). Scale bar is 5 μm. Stopwatch shows time in minutes and seconds in regards to anaphase onset.

Video S3. Distribution of the nuclear marker NLS::GFP throughout mitosis, related to Figure 1

The nuclear/cytoplasmic diffusion barrier is lost upon entry into mitosis. Neuroblast expressing NLS::5GFP (white on the left panel, blue on the merge) and the spindle marker Cherry::Jupiter (white on the middle panel, red on the merge). NLS::5GFP is leaving the nucleus as neuroblast enters prometaphase. Scale bar is 5 μm. Stopwatch shows time in minutes and seconds in regards to anaphase onset.

Video S4. Distribution of the cytoplasmic marker RpS13::GFP throughout mitosis related to Figure 1

The nuclear/cytoplasmic diffusion barrier is lost upon entry into mitosis. Neuroblast expressing the ribosomal subunit RpS13::GFP (white on the left panel, blue on the merge) and the spindle marker Cherry::Jupiter (white on the middle panel, red on the merge). RpS13::GFP is entering the nucleus as neuroblast enters prometaphase. Scale bar is 5 μm. Stopwatch shows time in minutes and seconds in regards to anaphase onset.

A nuclear lamina is required for nuclear envelope maintenance during passage through mitosis

In the search for structural elements that might distinguish this type of division from a classic open mitosis,²¹ we imaged cells expressing GFP-LamDm0, the sole B-type Lamin in the Drosophila genome. This revealed that, while the levels of LamDm0 present at the nuclear envelope were visibly reduced by late mitosis (1.8-fold), a lamina was maintained throughout (Figures 2A–2C and Video S5)—an observation confirmed using antibodies raised against endogenous LamDm0 (Figure 2D) or a form of endogenous LamDm0 that lacks phosphorylation on Serine 25 (Figure 2E).

Figure 2.

A nuclear lamina is required for nuclear envelope maintenance during passage through mitosis

(A) Mitotic neuroblast expressing LamDm0::GFP with Cherry::Jupiter. Yellow arrows point the higher LamDm0 density.

(B) LamDm0::GFP mean intensity along the nuclear membrane and in the cytoplasm, throughout the cell cycle.

(C) LamDm0 mean intensity ratio between the nuclear membrane and the cytoplasm is measured at 5 different stages for 25 cells. Grey area corresponds to a ratio of 1 ± 0.25. Bars indicate mean ± standard deviation. Asterisks denote statistical significance, derived from unpaired t tests: ^∗p ≤ 0.05, ^∗∗∗p ≤ 0.001, and ^∗∗∗∗p ≤ 0.0001.

(D) Immunostaining for LamDm0, dMoesin, and DAPI.

(E) Immunostaining for non-phosphorylated LamDm0 on serine 25 and DAPI.

(F) Representative time lapse of neuroblast expressing Klaroid::GFP, Cherry::Jupiter, and an RNAi against LamDm0. Yellow stars indicate presence of intact prophase nuclear compartment.

(G and H) CLEM showing progressive dispersion of nuclear membranes into the cytoplasm (G) and a loss of nuclear compartment integrity (H) after LamDm0 RNAi expression.

(I) Quantification of mitotic neuroblasts expressing an RNAi against LamDm0 and displaying normal or abnormal nuclear division. Number of analyzed cells = 123.

(J) Representative time-lapse images of LamDm0 RNAi neuroblast expressing Klaroid::GFP and Cherry::Jupiter, in which nuclear membranes disperse and nuclei fail to reform at mitotic exit. At least 3 independent experiments were performed.

Scale bars are 5 μm. See also Figure S2 and Video S5.

Video S5. A lamina structure is maintained during mitosis of fly neural stem cells, related to Figure 2

A lamina persists throughout cell division. Neuroblast expressing LamDm0::GFP (white on the left panel, green on the merge) and Cherry::Jupiter (red on the merge). While the levels of LamDm0::GFP present at the interphase nuclear envelope were visibly reduced during mitosis, some remained at the nuclear envelope throughout. Scale bar is 5 μm. Stopwatch shows time in minutes and seconds in regards to anaphase onset.

When LamDm0 was silenced in neuroblasts and their progeny via RNAi, we found that mild reductions in LamDm0 levels were accompanied by discontinuities in the mitotic nuclear lamina (Figure S2A, middle panel, and S2B); whereas more complete silencing (Figure S2A, bottom panel, and S2B) led to the separation of nuclear membrane layers that fan out into the cytoplasm (Figures 2F and S2C). Similar nuclear membrane spreading was observed using correlative light and electron microscopy (CLEM) (Figures 2G and 2H, yellow arrows), which was associated with large discontinuities in the nuclear boundary (green arrows). Furthermore, more than 80% of mitotic neuroblasts depleted in LamDm0 underwent an abnormal nuclear division (Figure 2I), which included a dramatic delay in the onset of anaphase, the formation of defective spindles, a failure in chromosome segregation and cell division (Figure 2J), and an inability to reform nuclei at mitotic exit once nuclear membrane layers dispersed into the cytoplasm (Figure 2J). Together, these data show that LamDm0 depletion is sufficient to convert a semi-closed mitosis into an open one.

Asymmetric nuclear division depends on asymmetric nuclear sealing at sites dictated by the spindle

While live cell imaging of the nuclear envelope and plasma membrane during mitosis suggested that their remodeling might be causally linked—since both compartments developed a similar apico-basal asymmetry (Figures 3A and 3B), elongated at a similar rate (Figure 3C), and narrowed at the same position prior to cytokinesis (Figure 3D)—we were able to uncouple cell and nuclear division by perturbing cytokinesis using the actin-poisons Latrunculin A and Cytochalasin D, or by culturing isolated neuroblasts on Concanavalin A-coated dishes which interferes with furrow contraction.²² Although many cells under these conditions ultimately failed in cytokinesis, all were able to divide their nucleus into two (n = 242; Figures 3E and S3A and Video S6), even after successive cytokinesis failures (Figure S3B). These observations rule out a model in which nuclear division is driven by local contraction of the cleavage furrow. Instead, high temporal resolution movies revealed that nuclear membranes elongated at both extremities of the central spindle, to seal the nuclear compartment to form two daughter nuclei (Figures 3F and 4A and Video S6). To test whether the central spindle is required for nuclear sealing, as suggested by this analysis, we used Binuclein 2 and RNAi to interfere with the function of the Aurora B kinase, a key regulator of the central spindle assembly.²³ In these experiments, two daughter nuclei were formed when a central spindle was visible (Figure S3C), leading to the generation of binucleated cells after cleavage furrow regression, while the lack of a central spindle was associated with a failure of both cell and nuclear division (Figures S3D, S3E, 3G, and 3H). Similarly, following partial depletion of Pavarotti or Tumbleweed, two core components of the centralspindlin complex, nuclear division proceeded in cells with a central spindle (Figures 3G, 3I, and S3F), but failed when central spindle formation was compromised (Figures 3G, 3J, and S3G).

Figure 3.

Nuclear division is a sealing-dependent process

(A) Dividing neuroblasts expressing Cherry::Jupiter and CD8::GFP. The red dashed line represents the site of cortical furrowing.

(B) Graph showing the cell and nuclear width from the apical to the basal cortex, for 7 time points.

(C) Graph showing the cell and nuclear length along the apico-basal axis for ten cells throughout mitosis. Error bars represent standard deviation and the dashed line indicates anaphase onset.

(D) Graph showing the cell and nuclear width at the cleavage furrow region throughout mitosis, averaged from ten cells. Error bars represent standard deviation, dashed line indicates anaphase onset, and blue and green arrowheads indicate start of cortical and nuclear furrowing, respectively.

(E) Representative time lapse sequences of mitotic neuroblasts expressing CD8::GFP and Cherry::Jupiter, and failing in cytokinesis after LatrunculinA, CytochalasinD, or ConcanavalinA treatment.

(F) Time-lapse images of a Latrunculin-treated neuroblast expressing CD8::GFP and Cherry::Jupiter. The yellow square indicates area used on the kymograph and the yellow arrowheads represent sites of nuclear sealing.

(G) Quantification of dividing neuroblasts, neuroblasts failing in cell division but not in nuclear division, or failing in both cell and nuclear division after expression of an RNAi against AuroraB, Pavarotti, or Tumbleweed. The number of cells analyzed for each condition is indicated.

(H) Representative time-lapse images of neuroblast expressing an RNAi against AuroraB, CD8::GFP, and Cherry::Jupiter, failing in both cell and nuclear division. Orange arrow shows the absence of central spindle.

(I and J) Representative time-lapse images of neuroblasts expressing an RNAi against Tumbleweed, CD8::GFP, and Cherry::Jupiter, failing cell division (I), or failing in both cell and nuclear division (J). Orange arrows show the presence (I) or absence (J) of central spindle. For each experiment, n ≥ 3.

Scale bars are 5 μm. See also Figure S3 and Video S6.

Figure 4.

Asymmetric nuclear division depends on nuclear sealing at sites dictated by the spindle

(A) Dividing neuroblasts expressing Klaroid::GFP and Cherry::Jupiter. Orange and yellow arrows indicate envelope sealing events.

(B) Airyscan pictures of neuroblasts expressing CD8::GFP and Cherry::Jupiter. Yellow arrowheads show sealing sites. Kymographs are done along the apico-basal axis.

(C) Graph showing the spindle asymmetry ratio and the nuclear diameter ratio (n = 33). Asterisks denote statistical significance, derived from unpaired t tests: n.s., not significant.

(D) Plot showing correlation between spindle and nuclear asymmetries (n = 68).

(E) Representative time-lapse images of neuroblast expressing CD8::GFP, Cherry::Jupiter, and Galphai. Quantification shows the asymmetric ratio for the cell, nucleus, and spindle at the sealing time.

(F) Plot showing correlation between spindle and nuclear asymmetries (on the left) or between cell and nuclear asymmetries (on the right); (n = 8).

(G) Image of neuroblast at the nuclear sealing time, showing the asymmetry in the central spindle and nuclear size (green dashed line).

(H) Representative time-lapse images of neuroblast expressing Cherry::Jupiter, Patronin::GFP, and GBP-PonLD. Green dashed lines correspond to the nuclei.

(I) Graph showing the spindle asymmetry (left) and the nuclear asymmetry (right) for control neuroblasts and after expression of Patronin::GFP and GBP-PonLD (n = 25 and n = 24). Bars indicate mean ± standard deviation. Asterisks denote statistical significance, derived from unpaired t tests: ^∗∗∗∗p ≤ 0.0001. For each experiment, n ≥ 3 experiments.

Scale bars are 5 μm. See also Figure S4.

Video S6. Nuclear division in Drosophila neuroblasts is a sealing-dependent process, related to Figure 3

Nuclear division is a sealing-dependent process. Latrunculin-treated neuroblast expressing CD8::GFP (white on the left panel, green on the merge) and the spindle marker Cherry::Jupiter (red on the merge). Nuclear membranes expand to progressively close the nuclear compartment generating daughter nuclei. Scale bar is 5 μm. Stopwatch shows time in minutes and seconds in regards to anaphase onset.

If nuclei are generated via a process of “sealing,” how is the asymmetry in the size of the nascent nuclei generated? A first clue came from a quantitative analysis of live imaging data, which revealed a strong correlation between the asymmetry in the spindle and the sites of nuclear sealing (Figures 4B–4D; Pearson coefficient of 0.9878). However, to rule out a role for the asymmetry in daughter cell size in this process, we needed to uncouple the spindle asymmetry from the cell division asymmetry.²⁴ To do so, we overexpressed the polarity protein Gαi. While most of these cells divided symmetrically (Figure S4A), a subset of Gαi-expressing cells divided asymmetrically despite having near symmetric spindles (Figure 4E; cell division asymmetry ratio of 1.67 and spindle asymmetry ratio of 1.18). Strikingly, nuclear division in these cells was close to symmetric (Figure 4E; nuclear asymmetry ratio of 1.16). Thus, the size of daughter nuclei immediately after nuclear sealing appeared defined by the asymmetry of the spindle (Figure 4F, left graph; r = 0.9893), rather than by the asymmetry of cell division (Figure 4F, right graph; r = 0.6384). For the converse experiment, we selectively perturbed the normal asymmetry in central spindle length by co-expressing the microtubule-binding protein Patronin::GFP together with a nanobody fused to the basal localization domain of Pon (GBP-PonLD) and Cherry::Jupiter. While Patronin::GFP was uniformly distributed along the central spindle when expressed alone (Figure S4B, top, t357″ insert), the co-expression of GBP-PonLD was sufficient to redistribute Patronin::GFP to the basal cortex (Figure S4C) and to the basal part of the central spindle (Figure S4B, bottom, t300″ insert and Figure S4D). This led to a local increase in microtubule density along the basal part of the central spindle (apical/basal Jupiter intensity ratio: 0.82 ± 0.09 versus 0.95 ± 0.06 for control cells) (Figure S4E) and induced a 1.5-fold change in the length of the central spindle relative to that found in the control (apical/basal length ratio: 1.54 ± 0.21 versus 1 ± 0.32 for control cells) (Figure S4F). This led to an increase of spindle asymmetry (Figure 4I, left graph) and a bias in the sites of nuclear sealing (Figure 4G versus 4H, green dashed lines; Figure 4I, right graph). As a result, the asymmetry of the nuclear division was increased without changing the position of the cytokinetic furrow, altering the cell/nuclear size ratio (Figure S4G). Together, these data show that it is the asymmetry in the spindle that biases re-sealing of the nuclear compartment to generate nascent sibling nuclei with different size.

Final nuclear sizes are achieved via differential growth of the two daughter nuclei

This initial bias in nuclear size was dramatically increased during telophase/early G1. Thus, neuroblast nuclei had a volume ∼4.5 times higher than GMC nuclei at the time of nuclear sealing but were ∼37-fold larger by the time nuclear size stabilized (Figure 5A). In fact, while the size of GMC nuclei remained approximately unchanged following nuclear sealing (Figure 5B, steps I and II), neuroblast nuclei grew extensively (Figure 5B, step II) until they reached a stable size (Figure 5B, step III). Interestingly, sealing events occurring on the lateral sides of nuclei (Figure 4A, orange arrows), appeared to fold in a cytoplasmic reservoir of membrane labeled with both INM and ER markers (Klaroid::GFP and Bip-SfGFP::HDEL) (Figures 4A, 5C, 5D, S5A, and S5B). These membranes were 20 times more abundant in neuroblasts relative to the GMC right at the time of nuclear sealing (Figure 5E). Furthermore, this pool of membrane was progressively depleted as neuroblast nuclei grew and was near absent by the time nuclear size was stabilized in late telophase, as visualized by both light microscopy and CLEM (Figures 5F and S5C versus S5D; Video S7). Interestingly, the cytokinesis furrow appears to play a role in enforcing asymmetric segregation of this nuclear membrane reservoir, since both daughter nuclei grew with similar rates in cells that experienced a delay in cytokinesis as a result of having been plated on Concanavalin A-coated dishes²² (Figure 5G, time points 239″ and 345″; Figure 5H, step II). Once the furrow narrowed to a point at which full access to bulk cytoplasmic ER was visibly restricted, growth of the GMC nucleus was inhibited, while the neuroblast nucleus continued to expand (Figure 5G, time point 818″; Figure 5H, step III). As a consequence, cells that failed division at very late stages of cytokinesis yielded binucleated cells containing one large and one small nucleus. By contrast, when Latrunculin A was used to prevent cleavage furrow initiation, both nuclei (which are born with an identical size as the result of a symmetrical spindle) grew at similar rates (Figures 5I and 5J, step II), before reaching a size plateau at the same time (step III) (Figures 3F, 5I, and 5J, step I). Thus, depending on spindle asymmetry, cleavage furrow dynamics, and the timing of cytokinesis failure, it is possible to generate binucleate cells with sibling nuclei that are similar or very different in size (Figures 5K and 5L). Based on these data, we propose a model in which differences in the final size of sibling nuclei depend on two processes that occur in sequence: (1) asymmetric nuclear sealing guided by asymmetries in the spindle and (2) differential nuclear growth based on ER/nuclear membrane availability.

Figure 5.

Final nuclear sizes are achieved via differential growth of the two daughter nuclei

(A) Scatterplots showing nuclear diameter of neuroblast and GMC right after nuclear sealing or during G1. Bars indicate mean ± standard deviation. Asterisks denote statistical significance, derived from unpaired t tests: ^∗∗∗∗p ≤ 0.0001.

(B) Graph showing the diameter of a neuroblast and GMC nucleus, from nuclear sealing until G1.

(C) Neuroblasts expressing Klaroid::GFP and Cherry::Jupiter; yellow arrows show cytoplasmic reservoir of nuclear membrane.

(D) Representative time-lapse images of neuroblast expressing the ER marker Bip-SfGFP::HDEL and Cherry::Jupiter, taken with Airyscan. Yellow arrows show cytoplasmic reservoir of nuclear membrane.

(E) Graph showing the asymmetric ER segregation at sealing time, using two different probes. The sum intensity is measured in the entire neuroblast and GMC volume, and the ratio Nb/GMC is plotted. Bars indicate mean ± standard deviation. Asterisks denote statistical significance, derived from unpaired t tests: n.s., not significant.

(F) Electron micrographs. ER/nuclear membranes are highlighted in pink.

(G) Representative time-lapse images of a neuroblast cultured on ConcanavalinA expressing CD8::GFP and Cherry::Jupiter.

(H) Graph corresponding to the neuroblast presented in (G), showing the size of both nuclei throughout the cell cycle.

(I) Representative time-lapse images of a Latrunculin-treated neuroblast expressing CD8::GFP. Yellow arrowheads indicate reservoir of available nuclear membrane.

(J) Graph from a representative Latrunculin-treated neuroblast showing the size of both nuclei throughout the cell cycle.

(K) Representative pictures showing wild-type and binucleated neuroblasts expressing CD8::GFP and Cherry::Jupiter, containing two nuclei with different or similar size.

(L) Scatterplot showing the nuclear area ratio between sibling nuclei for wild-type, DMSO-, Latrunculin-, Cytochalasin-, or Concanavalin-treated neuroblasts. Grey area on the graph represents a nuclear area ratio equal to 1 ± 0.25. Number of analyzed cells = 148. Asterisks denote statistical significance, derived from unpaired t tests: n.s., not significant, ^∗∗∗∗p ≤ 0.0001. For each experiment, n ≥ 3.

Scale bars are 5 μm. See also Figure S5 and Video S7.

Video S7. Electron microscopy of a mitotic neuroblast, right after nuclear sealing, related to Figure 5

ER/nuclear membrane reservoir segregates asymmetrically between the two daughter cells. Electron microscopy performed on mitotic neuroblast in late anaphase/early telophase, within intact larval brain.^34–36

Asymmetric nuclear division affects chromatin organization

As a result of these processes, nuclear division in neuroblasts generates daughter nuclei that differ profoundly in size despite their carrying the same amount of DNA, leading to an asymmetry in chromatin packing. This was first visible in anaphase/telophase, when the His2A signal was seen becoming weaker in the neuroblast nucleus as its genome decondensed to fill the larger nuclear compartment (Figures 6A, 6B, S6A, and S6B). This difference in packing was maintained after late cytokinesis failure (Figure S6C, left graph), in cells that displayed two nuclei that differed in size, but was abolished in binucleated cells containing two nuclei of similar size (Figure S6C, right graph).

Figure 6.

Asymmetric nuclear division affects chromatin organization

(A) Representative time-lapse images of a neuroblast expressing the DNA marker His2A::RFP and the cortical marker Sqh::GFP.

(B) Graph showing His2A::RFP mean intensity over time for GMC and neuroblast nuclei, n = 5. Error bars represent standard deviation.

(C) Wild-type neuroblasts fixed and stained for DAPI, H3K4me2, Tubulin, and Phalloidin. Yellow arrows indicate similar H3K4me2 intensity on the two pools of sister chromatids.

(D) Wild-type neuroblasts fixed and stained for DAPI, H3K4me2, and Prospero. White dashed lines show the nuclear area during and after nuclear growth.

(E) For cells from early anaphase to late telophase, graph (left) shows nuclear size ratio on the x axis versus ratio (neuroblast/GMC) of sum intensity for DAPI across the entire nuclear volume on the y axis, and (right) the equivalent for H3K4me2 on the y axis.

(F) Binucleated neuroblast stained for DAPI and H3K4me2. The graph shows the nuclear area ratio (left) and the DAPI or H3K4me2 sum intensity ratio (right) for binucleated neuroblasts displaying two nuclei with different size (n = 10). Bars indicate mean ± standard deviation. Asterisks denote statistical significance, derived from unpaired t tests: ^∗∗p ≤ 0.01. For each experiment, n ≥ 3.

Scale bars are 5 μm. See also Figure S6.

The asymmetry in chromatin packing between the two nascent nuclei was accompanied by differences in H3K4me2 (a histone mark associated with neuronal differentiation in the mouse brain²⁵), which arose during the process of nuclear division. The segregation of H3K4me2 was symmetric until late anaphase/early telophase (Figure 6C), but progression through telophase was associated with the development of a striking asymmetry in the distribution of H3K4me2 between the two daughter nuclei (Figures 6D and 6E, purple circles; from 1.057 ± 0.04 in anaphase/early telophase to 0.65 ± 0.05 in late telophase), which could not be explained by a different density due to the physical constraints of a small nuclear size (the control DAPI ratio remained around 1 throughout mitosis: Figure 6E, blue circles; 1.01 ± 0.06). As a result, following division, the H3K4me2 signal was high in GMC nuclei, but was barely detected in the neuroblast (Figure S6D). Furthermore, a similar asymmetry in H3K4me2 staining (0.77 ± 0.09, relative to 1.074 ± 0.12 for the DAPI channel) was observed in binucleated cells that possess a large and small nucleus within the same cytoplasm (Figure 6F). In line with these results, in control neuroblasts, the H3K4me2 signal was visibly asymmetric long before the cortical release and nuclear import of the cell fate determinant Prospero (Figure 6D, yellow channel; Figure 7A; 16.34 ± 2.34 min after anaphase onset). Taken together, these data reveal a role for asymmetric nuclear division alongside the polarized distribution of cell fate determinants in the generation of cell diversity in this system (Figure 7B).

Figure 7.

Asymmetric nuclear division generates two daughter nuclei that differ in size, envelope composition, and chromatin organization before the nuclear import of cell fate determinants

(A) Graph showing the timing of the successive steps related to nuclear division, cytokinesis, and nuclear import of Prospero.

(B) Diagram representing different steps of the process of asymmetric nuclear division and cell fate determinant release. NPCs disassemble 330 s before anaphase onset. The nuclear envelope is then asymmetrically sealed in a spindle-dependent manner 200 s after anaphase onset to generate two daughter nuclei with moderate differences in size. This asymmetry is increased by differential nuclear growth during telophase, until nuclear size stabilization 480 s after anaphase onset. This is associated with a differential packing of the chromatin and an asymmetric distribution of histone marks between the new neuroblast and GMC nuclei, before the nuclear import of cell fate determinants (980 s after anaphase onset).

Discussion

Functional consequences of maintaining a nuclear compartment throughout mitosis

In this paper, by studying the nuclear envelope remodeling and partitioning during asymmetric divisions of Drosophila neural stem cells, we show that the nuclear compartment is not dissolved and reformed during passage through mitosis. Instead, a nuclear envelope persists, supported by an underlying mitotic lamina. Because NPCs are disassembled upon entry into mitosis, as nuclear membranes become fenestrated and multi-layered (as reported for C. elegans embryos¹¹), this envelope fails to function as a diffusion barrier. While, in an open mitosis, dispersed ER/nuclear membrane is recruited around the chromatin to re-establish the nuclear compartment anew as cells exit mitosis, in this system Lamin RNAi cells fail to reassemble their nuclei after division. Thus, fly neural stem cells appear to rely on the maintenance, remodeling, and inheritance of a nuclear compartment to generate daughter nuclei.

Forming two nuclei of different size

Importantly, while nuclear division in Drosophila neuroblasts is an autonomous process that is independent of cytokinesis, it is clear from our analysis that the asymmetry in nuclear division depends on the spindle and the position of the cytokinetic furrow. Asymmetric division is achieved in two steps. First, the asymmetric central spindle defines the sites at which the nuclear compartment is sealed, in a process that depends on Aurora B, Tumbleweed, and Pavarotti, leading to the formation of two nascent daughter nuclei that are moderately different in size. Second, as they grow, these nuclei appear to compete for the available reservoir of ER and nuclear membrane, as has been reported to be the case in sea urchin and Xenopus embryos.²⁶ Importantly, access to this membrane reservoir appears limited by partial closure of the cytokinesis furrow, since pinching of the plasma membrane was sufficient to limit nuclear growth before abscission, and since the early inhibition of furrow formation enabled the growth of the GMC nucleus. These data suggest that the cytokinesis furrow in fly neuroblasts forms a barrier that functions like the neck of dividing budding yeast cells^27,28 to restrict the movement of material between the two daughter cells prior to division. As a result, the asymmetry in nuclear size in this system is not the result of differences in cell volume, as has been suggested in other model systems.^{29, 30, 31, 32}

Functional implications of asymmetric nuclear division

As shown through our analysis, the process of asymmetric nuclear division generates two daughter nuclei that differ 37-fold in size. This large-scale physical asymmetry and the accompanying molecular asymmetries in nuclear envelope composition generated by the differential enrichment of NPCs in sibling nuclei (Figures 1D, 2A, and S1E) give rise to differences in the decompaction of chromatin in the two daughter nuclei, and are associated with asymmetries in H3K4me2 that precede the nuclear import of cell fate determinants like Prospero (Figures 6D, 7A, and 7B). Thus, these nuclear asymmetries are not the consequence of the action of cell fate determinants. Instead, they are likely to be factors that influence the ability of these transcription factors to execute their functions. This is in line with a recent study identifying a role for Ankle2, a nuclear envelope protein, in the asymmetric segregation of cell fate determinants in mitotic neural stems cells (as well as on nuclear remodeling and spindle positioning).³³

One might ask why the system is set up in this way? While further work is required to answer this question, our analysis hints at this being a way to define nuclear identity and size very early in late telophase/earlyG1, which might be particularly important in cells like these which have a very short cell cycle (45 min to 1.30 h on average). Semi-closed mitosis may also aid asymmetric nuclear division itself. In addition, this mode of division may allow nascent nuclei to compete for common factors prior to cell division, facilitating the break in symmetry required to distinguish the neural stem cell from the more differentiated GMC.

Finally, while many of the consequences of this mechanism of semi-closed mitosis we describe here remain to be explored, our analysis suggests that asymmetric nuclear division might contribute, alongside cortical polarity, to the formation of sibling cells with different identity/fate.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Mouse anti-LamDm0 (1:100)	DHSB	Cat# ADL67.10; RRID:AB_528336
LamDm0 Ser25 (1:100)	DHSB	Cat# ADL84.12; RRID:AB_528338
Rabbit anti-dMoesin (1:1000)	François Payre lab	N/A
Chicken anti-GFP (1:1000)	Abcam	Cat# ab13970; RRID:AB_300798
Rabbit anti-H3K4me2 (1:100)	Abcam	Cat# ab7766; RRID:AB_2560996
Mouse anti-Prospero (1:100)	DHSB	Cat# MR1A; RRID:AB_528440
Vectashield with DAPI	Vector Laboratories	Cat# H1200; RRID:AB_2336790
Biological Samples
Larval brain tissues from Drosophila melanogaster strains	This study	N/A
Chemicals, Peptides, and Recombinant Proteins
Schneider’s Insect Medium	Sigma-Aldrich	Cat# S0146
Insulin	Sigma-Aldrich	Cat# I6634
L-Glutamine	Sigma-Aldrich	Cat# G7513
L-Glutathione reduced	Sigma-Aldrich	Cat# G6529
20-Hydroxyecdysone	Sigma-Aldrich	Cat# H5142
FBS	Sigma-Aldrich	Cat# F6178
Chan and Gehring’s medium	35	N/A
Papain	Sigma-Aldrich	Cat# P4762
Collagenase Type I	Sigma-Aldrich	Cat# C2674
Concanavalin A	Sigma-Aldrich	Cat# L7647
Cytochalasin D	Sigma-Aldrich	Cat# C8273
Latrunculin A	Sigma-Aldrich	Cat# L5163
Binuclein 2	Sigma-Aldrich	Cat# B1186
MitoTracker Deep Red	Invitrogen	Cat# M22426
Experimental Models: Organisms/Strains
D.melanogaster: y, w; KlaroidCB04483	Bloomington Drosophila Stock Center	BDSC: 51525 FBal0211641
D.melanogaster: w; UAS BiP::sfGFP::HDEL	Bloomington Drosophila Stock Center	BDSC: 64749 FBst0064749
D.melanogaster: UAS Klarsicht::GFP	Elhanany-Tamir et al.36	N/A
D.melanogaster: UAS RpS13::GFP	Rugjee et al.37	FBal0284510
D.melanogaster: gNup58::EGFP	Radermacher et al.38	FBtp0093711
D.melanogaster: w; EGFP::Nup107	Bloomington Drosophila Stock Center	BDSC: 35514 FBst0035514
D.melanogaster: yw; UAS LamDm0::GFP	Bloomington Drosophila Stock Center; Buszczak et al.39	BDSC: 7376 FBst0007376
D.melanogaster: w; WorGal4, UAS Cherry::Jupiter / Cyo; UAS CD8::GFP / TM6B,Tb	This study	N/A
D.melanogaster: UAS EGFP::NLS(tra)	Bloomington Drosophila Stock Center	BDSC: 65402 FBst0065402
D.melanogaster: UAS Pebble::GFP	Somers et al. 34	N/A
D.melanogaster: w; WorGal4, KlaroidCB04483, UAS Cherry::Jupiter / Cyo; UAS Dicer / TM6B,Tb	This study	N/A
D.melanogaster: y,sc,v,sev; +; UAS dsRNA LamDm0	Bloomington Drosophila Stock Center	BDSC: 36617 FBst0036617
D.melanogaster: y, sc, v, sev; UAS dsRNA LamDm0; +	Bloomington Drosophila Stock Center	BDSC: 57501 FBst0057501
D.melanogaster: UAS Cherry::Jupiter	Chabu and Doe40	FBal0220463
D.melanogaster: w;WorGal4, His2A::mRFP / Cyo; MKRS / TM6B,Tb	Roubinet et al.15	N/A
Software and Algorithms
ImageJ	Schneider et al.41	https://imagej.nih.gov/ij/index.html; RRID:SCR_003070
Imaris	Bitplane	https://imaris.oxinst.com/packages; RRID:SCR_007370
Volocity	PerkinElmer	http://quorumtechnologies.com/index.php/component/content/category/31-volocity-software; RRID:SCR_002668
NIS-Element	Nikon	https://www.microscope.healthcare.nikon.com/products/software; RRID:SCR_014329
Slidebook	3i	https://www.intelligent-imaging.com/slidebook; RRID:SCR_014300
Prism	GraphPad	https://www.graphpad.com/; RRID:SCR_002798
Photoshop	Adobe	https://www.adobe.com/uk/products/photoshop.html; RRID:SCR_014199
Illustrator	Adobe	https://www.adobe.com/uk/products/illustrator.html; RRID:SCR_010279

Resource availability

Lead Contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Chantal Roubinet (c.roubinet@ucl.ac.uk).

Materials Availability

All unique/stable reagents generated in this study are available from the Lead Contact without restriction.

Data and Code Availability

• •All data reported in this paper will be shared by the lead contact upon request.
• •This paper does not report original code.
• •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Experimental model and subject details

Several lines of the fruit fly Drosophila melanogaster were used in this study. Flies and larvae were raised in vials containing standard cornmeal-agar medium supplemented with baker’s yeast and incubated at 25°C. Third-instar larvae of 3-days old (expressing live imaging markers) or 5-days old (expressing RNAi) were dissected to extract the brains, then live imaging was performed directly on brains or after brain dissociation (see “Method details”). Experiments utilized both female and male larvae, and data from both sexes was pooled.

Flies used in this study include: Klaroid^CB04483 (BDSC-51525),³⁹ UAS BiP::sfGFP::HDEL (BDSC-64749),⁴² UAS Klarsicht::GFP,³⁶ UAS RpS13::GFP,³⁷ gNup58::EGFP,³⁸ EGFP::Nup107 (BDSC-35514),⁹ UAS LamDm0::GFP (BDSC-7376), UAS 5EGFP::NLS (BDSC-65402), UAS dsRNA LamDm0 (BDSC-36617 and BDSC-57501), WorGal4, UAS Cherry::Jupiter; UAS CD8::GFP (this study), WorGal4, Klaroid^CB04483, UAS Cherry::Jupiter; UAS Dicer (this study), UAS Galphai (BDSC-44600), w; UAS GBP::PonLD; UAS GFP::Patronin,⁴³ UAS AuroraB RNAi (VDRC-35107),⁴⁴ UAS AuroraB RNAi (BDSC-35299),⁴⁴ w; WorGal4, UAS His2A::mRFP,¹⁵ UAS Tumbleweed RNAi (BDSC-28982 and VDRC-106850), UAS Pavarotti RNAi (BDSC-35649), w; WorGal4, His2A::mRFP, Sqh::GFP.¹⁵ Transgenes were expressed using the neuroblast-specific driver worGal4.⁴⁵ Detailed allele information is listed in the Key resources table.

Method details

Live imaging sample preparation

For experiments performed on intact brains, imaging medium (Schneider’s insect medium mixed with 10% FBS (Sigma), 2% PenStrepNeo (Sigma), 0.02 mg/mL insulin (Sigma), 20mM L-glutamine (Sigma), 0.04 mg/mL L-glutathione (Sigma) and 5 μg/mL 20-hydroxyecdysone (Sigma)) was warmed up to room temperature before use. Seventy-two hours after egg laying, larvae were dissected in imaging medium then brains were transferred onto a gas-permeable membrane (YSI Life Sciences #13-298-83) fitted on a metallic slide. Brains were oriented with the brain lobes facing the coverslip, and excess media was removed until the brain lobes were in contact with the coverslip. The sample was sealed with Vaseline.

Primary neuroblast cultures

For experiments done on primary neuroblast culture, seventy-two hour larvae were dissected in Chang & Gerhing solution (3.2 g/L NaCl, 3 g/L KCL, 0.69 g/L CaCl2-2H2O, 3.7 g/L MgSO4-7H2O, 1.79 g/L tricine buffer pH 7, 3.6 g/L glucose, 17.1 g/L sucrose, 1 g/L BSA)³⁵ at room temperature. Brains were then incubated in Chang & Gerhing solution supplemented in collagenase from Clostridium histolyticum (Sigma) and papain from papaya latex (Sigma) at a final concentration of 1 mg/mL each, during 45 min at 29°C. Brains were washed with imaging medium (see above) then dissociated in imaging medium by pipetting 20–25 times. The cell culture was then imaged using Ibidi chambers (15μ-slide angiogenesis).

Antibodies and immunostaining

The following primary antibodies were used for this study: mouse anti-LamDm0 Ser25 (DHSB; #ADL84.12; 1:100), mouse anti-LamDm0 (DHSB; #ADL67.10; 1:100), chicken anti-GFP (Abcam; #ab13970; 1:1000), rabbit anti-dMoesin (gift from Francois Payre; 1:1000), rabbit anti-H3K4me2 (Abcam; #ab7766; 1;100) and rabbit anti-Miranda (gift from Emmanuel Caussinus; 1:50). Secondary antibodies were from Invitrogen and Phalloidin-FITC (Sigma; #P5282; 1:100) was used to stain F-Actin.

For immunostaining, seventy-two hour larvae were dissected in Schneider’s insect medium (Sigma-Aldrich S0146) and brains were fixed for 20 min in 4% paraformaldehyde in PEM (100 mM PIPES pH 6.9, 1 mM EGTA and 1 mM MgSO4). After fixing, the brains were washed with PBSBT (1 × PBS (pH7,4), 0.1% Triton X-100 and 1% BSA) then blocked with PBSBT for 1 h. Primary antibody dilution was prepared in PBSBT and brains were incubated for 48 hours at 4°C. Brains were washed with PBSBT four times for 30 min each, then incubated with secondary antibodies diluted in PBSBT at 4°C overnight. The next day, brains were washed with PBST (1x PBS, 0.1% Triton X-100) four times for 20 min each and kept in Vectashield antifade mounting medium with DAPI (Vector laboratories; H1200), at 4°C.

Small molecule inhibitors and chemical treatments

Following dissection and prior montage for live imaging, brains were incubated in imaging medium (see above) containing the following chemical reagents, as indicated on the Figures: Latrunculin A (Sigma; #L5163; final concentration of 10 μM), Cytochalasin D (Sigma; #C8273; final concentration of 5 μg/mL or 10 μg/mL) or Binuclein 2 (Sigma; #B1186; final concentration of 10 μM or 20 μM). Cerulenin (Sigma; #C2389; 10μM) was added during live imaging. After brain dissociation, the following final concentrations were used on primary neuroblast culture: Latrunculin A: 1 μM; Cytochalasin D: 0.5 μg/mL.

For control experiments, dissected brains were incubated in DMSO at the same final concentrations than the ones used to dissolve the reagents.

In order to induce binucleated cells, primary neuroblast culture was platted on dish coated with Concanavalin A (Sigma # L7647).

Observation of mitochondria distribution during mitosis was done by incubating primary neuroblast culture with MitoTracker Deep Red (Invitrogen # M22426; final concentration of 0.1nM).

Electron microscopy

Cells were cultured on gridded coverslip-bottomed dishes (MatTek, P35G-1.5–14-CGRD) to facilitate correlation between light and electron microscopy, and processed for the latter following a protocol adapted from Deerinck et al. (NCMIR methods for 3D EM: A new protocol for preparation of biological specimens for serial block face scanning electron microscopy). Briefly, coverslips were fixed in 2% PFA / 2.5% Gluteraldehyde solution (EM grade, TAAB) in 0.1M sodium cacodylate buffer for 30 min at room temperature, washed in 0.1M sodium cacodylate buffer and post-fixed in 1% OsO₄/1.5% potassium ferricyanide for 1 h at 4°C. Samples were then treated with 1% thiocarbohydrazide (TCH) for 20 min at room temperature, 2% OsO₄ for 30 min at room temperature, 1% uranyl acetate (UA) overnight at 4°C and lead aspartate for 30 min at 60°C, with washing in dH₂O between each step. This was followed by dehydration of the samples by graded ethanol incubations in 70%, 90% and 100% ethanol, and incubations for 1 hour in a 1:1 mix of propylene oxide: epon resin (TAAB), 1 hour in fresh 100% epon resin, and a change to a final 1 hour epon resin incubation all at room temperature before mounting the coverslip by inversion onto a pre-polymerized epon resin stub and polymerization overnight at 60°C. Coverslips were removed by plunging into liquid nitrogen. Cells of interest were relocated on the blockface using the transferred alphanumeric grid with reference to light microscopy images, and the resin block trimmed accordingly. Ultrathin sections of 70 nm were cut using a diamond ultra 45-degree knife (Diatome) on a Leica UC7 ultramicrotome, and collected on formvar coated copper 2 × 1mm slot grids. EM images were acquired on a transmission electron microscope (T12 Tecnai Spirit biotwin, FEI).

Confocal microscopy

Fixed samples were imaged using an inverted Leica SP5 confocal microscope. For representative images, a 60 × /1.40 N.A oil immersion objective was used. Live cell imaging was performed on a UltraView Vox spinning disc confocal microscope (Perkin Elmer Nikon TiE; Yokogawa CSU-X1 spinning disc scan head) with 60 × /1.40 N.A oil objective and equipped with a Hamamatsu C9100-13 EMCCD camera, or a 3I spinning disc confocal microscope (Zeiss AxioObserver Z1; Yokogawa CSU-W1 spinning disc scan head) with 63 × /1.40 N.A objective and equipped with a photometrics prime 95B scientific CMOS camera. Both spinning disc microscopes are equipped with a temperature-controlled environment chamber set at 26C for the experiments. Airyscan pictures were taken on fixed and stained neuroblasts using an inverted LSM880 Multiphoton or on live samples using a LSM900 point scanning laser confocal with Airyscan 2 on an inverted Axio Observer Z1 microscope stand.

Quantification and statistical analysis

Image processing and calculations

Images were processed using Imaris × 64 7.5.2 and ImageJ softwares.

Measurements of intensity was obtained by using the oblique slicer in Imaris oriented along the cell division axis and going through the mid-plan of both daughter nuclei (neuroblast and GMC, respectively). The corresponding images were exported as TIFF files and opened with ImageJ to measure the mean intensity or integrated density, as mentioned on graphs. To measure the ratio cytoplasmic / nuclear NLS::5GFP or Rps13::GFP, background correction was performed by measuring the green channel intensity in the media. For the cell width and nuclear width values over time, a grid was used to measure the cell and nuclear width along each of the line of the grid, starting from the apical cortex toward the basal one for each time point. Cell and nuclear furrowing values were obtained by measuring over time both cell and nuclear width along the red line, which represents the cleavage furrow plan. To do so, images were exported as TIFFs then combined in a stack using ImageJ. The red line was drawn at anaphase when the cleavage furrow positioning is set and easy to determine. Measurements of both compartment widths were then done along this line, prior and after anaphase onset, to quantify the cleavage furrow ingression and the deformation of the nuclear compartment at the cleavage furrow region. Measurements of nuclear or cell area ratio were performed by using oblique slicer in Imaris as previously described, then images were exported as TIFFs and opened with ImageJ to measure manually the area or perimeter (as mentioned on the figures) by drawing the border of each nucleus and cell. For the graphs showing nuclear growth, TIFF files were opened with ImageJ then the nuclear diameter was manually measured for each time point. Microtubule density was based on the measurement of Cherry::Jupiter intensity, as reported in Derivery et al.⁴³ For the experiments with nanobody (GBP-PonLD), the nuclear size was measured using the Patronin channel, since Patronin is also present in the cytoplasm and excluded from the nuclei. ER partitioning between the daughter neuroblast and daughter GMC was measured using two different ER probes, measuring the sum intensity into the entire volume of the nuclei. H3K4me2 measurements were done by quantifying the sum intensity of this Histone mark in the entire volume of the two nuclei for each binucleated cell, plotting the ratio on the graph. All figures were assembled using Adobe Illustrator CS6 software.

Statistical analysis

For each experiment, the data was collected from at least 3 independent experiments. For each independent experiment, at least 8 larvae were dissected. For all the analysis, “n” refers to the number of cells analyzed and is mentioned on each graph as well as in figure legends. Statistical significance was determined with Student’s t test using GraphPad Prism software. In all Figures the Prism convention is used: n.s. (p > 0.05), ^∗(P ≤ 0.05), ^∗∗(P ≤ 0.01), ^∗∗∗(P ≤ 0.001) and ^∗∗∗∗(P ≤ 0.0001). In all graphs showing mean, the error bars correspond to standard deviation.

Published: July 22, 2021