Abstract
Metazoan cells accurately attach to, congress and segregate chromosomes during mitosis Additionally, hybrid cells derived through fertilization or somatic cell fusion also employ mechanisms to recognize and separate chromosomes of different origin. The underlying mechanisms are mostly unknown but could prevent aneuploidy and tumor formation. Here, we acutely induce fusion between Drosophila neural stem cells (neuroblasts) and differentiating ganglion mother cells (GMCs) in vivo to define how epigenetically distinct chromatin is recognized and segregated. We find that Nb-GMC hybrid cells align both endogenous (neuroblast-origin) and ectopic (GMC-origin) chromosomes at the metaphase plate through centrosome derived dual spindles. Mixing of endogenous and ectopic chromatin is prevented through an asymmetric, microtubule-dependent chromatin capture mechanism during interphase and physical boundaries imposed by nuclear envelopes. Although hybrid cells fail to accurately segregate ectopic chromatin, manifested in lagging chromosomes and chromosome bridges, transplanted brain tissue containing hybrid cells neither reduce the lifespan nor form visible tumors in host flies. We conclude that fly neural stem cells utilize asymmetric centrosome activity in interphase to capture and physically separate epigenetically distinct chromatin in a microtubule-dependent manner. We propose a novel chromosome recognition and separation mechanism that could also inform biased chromatid segregation observed in flies and vertebrates.
Dividing cells distribute replicated chromosomes equally between the two forming sibling cells through microtubule-dependent attachment and segregation mechanisms (McIntosh, 2016; Maiato et al., 2017). Chromosomes are connected with microtubules via kinetochore proteins, which are localized on centromeric DNA (Thomas et al., 2017; Yu et al., 2019). The occurrence of biased segregation of sister chromatids or homologous chromosomes during mitosis and meiosis respectively, as well as the spatial separation of the maternal and paternal genomes during early vertebrate development, implies the presence of accurate chromatin recognition mechanisms (Akera et al., 2017; Ranjan et al., 2019; Sunchu and Cabernard, 2020; Reichmann et al., 2018). Chromosome separation has also been observed in fused human cells (Heasley et al., 2017; Rieder et al., 1997). The molecular nature of these separation mechanisms is unclear but could entail asymmetries in centromere binding proteins, kinetochore size or kinetochore composition (Arco et al., 2018; Ranjan et al., 2019; Akera et al., 2019; Drpic et al., 2018).
Over a century ago, cell – cell fusion has also been proposed to initiate tumor formation (Aichel, 1911), possibly through chromosome missegregation and aneuploidy (Boveri, 2008; Molina et al., 2020). Here, we ask how hybrid cells accurately recognize, separate and segregate epigenetically distinct chromosomes. To this end, we acutely fused Drosophila neural stem cells (neuroblasts (NBs), hereafter) with differentiating ganglion mother cells (GMCs) in the intact larval fly brain to create hybrid cells, containing both neuroblast and GMC chromosomes. Unperturbed Drosophila neuroblasts divide asymmetrically, self-renewing the neural stem cell while forming a differentiating GMC. Neuroblasts are twice the size of GMCs and neurons and express the transcription factor Deadpan (Dpn+), whereas GMCs and neurons can be identified based on Prospero (Pros+) expression (Gallaud et al., 2017) (Fig. S1A). We used a 532nm pulsed laser to induce a small lesion at the Nb – GMC interface between a dividing neuroblast and GMC, causing the GMC chromatin to enter the neuroblast cytoplasm. Targeted mitotic neuroblasts often retain the GMC chromosomes, creating a large apical hybrid cell containing a Dpn+ and Pros+ nucleus (Nb – GMC hybrid). Most Nb – GMC hybrid cells normally localize the contractile ring marker non-muscle Myosin to the cleavage furrow and complete cytokinesis (Fig. S1B-E). Acute cell fusion can also result in the expulsion of the GMC chromatin, forming GMC – GMC hybrids (see Supplementary Fig. 3c in (Roubinet et al., 2017)) but the focus here is on Nb – GMC hybrids (hybrid cells, hereafter) only.
To better characterize the dynamics of neuroblast (endogenous) and GMC (ectopic) chromosomes during mitosis, we induced cell fusion at different cell cycle stages in wild type neuroblasts, expressing the canonical chromosome marker His2A::GFP. Hybrid cells derived from Nb-GMC fusions early in the cell cycle could (1) align only the neuroblast chromosomes at the metaphase plate, (2) congress a mix of neuroblast and GMC chromosomes or (3) separate and align the two chromosome pools at the metaphase plate (Fig. 1A). We found that the endogenous and ectopic chromatin was separated and distinguishable when fusions were induced in early mitosis. Both the ectopic and endogenous chromatin aligned at the metaphase plate (Fig. 1B; video 1&2). Nb-GMC fusions can be induced at all cell cycle stages but GMC chromosomes align at the metaphase plate more accurately in hybrid cells derived from interphase or early prophase fusions (Fig. 1C, D). We conclude that hybrid cells derived from fusions between interphase Nbs and non-mitotic GMCs accurately align ectopic and endogenous chromatin at the metaphase plate.
Nb-GMC hybrid cells independently align Nb and GMC chromatin at the metaphase plate
We next asked whether GMC chromatin congresses independently of neuroblast chromatin. To this end, we measured the time between nuclear envelope breakdown (NEB) and chromosome alignment at the metaphase plate for Nb and ectopic GMC chromatin in hybrid cells derived from interphase and early prophase fusions (Fig. 1E). In most hybrid cells, ectopic and endogenous chromatin was distinguishable based on differences in location and intensity (see Fig. 1B; video 1&2). In untargeted control neuroblasts (unfused), Nb chromosomes aligned at the metaphase plate within 6.5 minutes after NEB (SD = 1.55; n = 6), which is insignificantly faster than the neuroblast chromosomes of hybrid cells (t = 6.9 mins; SD = 2.84; n = 11). GMC chromosomes aligned within 7.63 mins (SD = 3.69; n = 11), statistically not significantly different from unperturbed wild type chromosomes (Fig. 1E). In most Nb-GMC hybrids, the endogenous neuroblast and the ectopic GMC chromosomes aligned at the metaphase plate with no significant time difference. However, in a few cases, ectopic chromatin aligned before or after the neuroblast chromatin (Fig. 1F). These results suggest that the neuroblast and GMC chromatin can move independently to the metaphase plate in hybrid cells.
Ectopic spindles distinguish between Nb and GMC chromatin
We next investigated the mechanisms underlying independent Nb/GMC chromosome alignment. Ectopic chromosomes could be aligned together with the endogenous chromosomes via a single bipolar spindle. Alternatively, hybrid cells could form multiple bipolar spindles, which attach to either the neuroblast’s, GMC’s, or chromosomes from both cell types (Fig. 2Aa). Live cell imaging showed that hybrid cells derived from interphase Nb-GMC fusions contained double spindles in almost all cases, whereas the vast majority of cell fusions induced in metaphase only formed one mitotic spindle (Fig. 2B, C & video 3). Most hybrid cells contained two mitotic spindles, which predominantly formed at the same time (Fig. 2D, E). We quantified spindle alignment and positioning dynamics by measuring the angle and distance between the two spindles during metaphase (Fig. 2F and methods). The two spindles were initially misaligned and separated but decreased their inter-spindle distance and angle during metaphase (Fig. 2G-I). We conclude that Nb-GMC hybrid cells align neuroblast and GMC chromosomes separately at the metaphase plate through the formation of independent mitotic spindles.
Ectopic spindles are nucleated from GMC centrosomes
Mitotic spindles can be nucleated through the centrosome-dependent, chromatin or microtubule pathway but when centrosomes are present, the centrosome pathway prevails (Prosser and Pelletier, 2017). To elucidate the mechanisms underlying ectopic spindle formation, we induced Nb-GMC fusions in interphase wild type neuroblasts expressing a live centriole (Asterless; Asl::GFP) and spindle marker (cherry::Jupiter), and assayed centrosome dynamics and spindle formation throughout mitosis. Normal wild type neuroblasts usually contain two Asl::GFP positive centrioles in mitosis, forming a single bipolar spindle. However, in Nb-GMC hybrids, we predominantly found four Asl::GFP positive centrioles, two of which seemed to be introduced from the GMC (Fig. 2J,K). The GMC centrosomes nucleated an ectopic bipolar spindle that subsequently aligned with the main neuroblast spindle. Although multipolar spindle formation can be prevented through centrosome clustering (Quintyne et al., 2005), we observed that GMC centrosomes approached neuroblast centrosomes but still formed two parallel bipolar spindles. We conclude that ectopic spindles are formed through the centrosome pathway, using centrioles originating from GMCs.
Microtubule-dependent, asymmetric chromatin-centrosome attachments retain chromosomes close to the apical neuroblast cortex during interphase
Our data suggest that Nb and GMC chromatin are being separated through an endogenous, Nb-derived and an ectopic, GMC-derived mitotic spindle. We next investigated how these spindles distinguish between Nb and GMC chromosomes. During mitosis, microtubules emanate from centrosomes and attach to sister chromatids through kinetochore proteins, localizing to the centromeric region (Fukagawa and Earnshaw, 2014). Drosophila male germline stem cells contain asymmetric levels of the centromere-specific H3 variant (Centromere identifier (Cid) in flies (Henikoff et al., 2000)) (Ranjan et al., 2019), prompting us to investigate whether Cid intensity differs between endogenous and ectopic chromatin in hybrid cells. We induced Nb-GMC fusions of wild type cells expressing Cid::EGFP (Ranjan et al., 2019) and measured Cid intensity on both GMC and Nb chromatin. These measurements did not reveal a significant intensity difference between Nb and GMC Cid (Fig. S2A). However, we noticed that endogenous Cid::EGFP was localized in very close proximity to the apical centrosome in unperturbed interphase and prophase wild type neuroblasts (Fig. S2B & video 4). Cid::EGFP remained associated with chromatin throughout the neuroblast cell cycle, excluding the possibility that early Cid clusters are not connected with chromatin (Fig. S2F & video 5).
Interphase wild type neuroblasts contain only one active apical microtubule organizing center (MTOC), retaining the daughter-centriole containing centrosome close to the apical neuroblast cortex. The mother-centriole-containing centrosome is inactive in interphase but matures from prophase onward, positioning itself on the basal cell cortex (Januschke et al., 2013, 2011; Januschke and Gonzalez, 2010; Gallaud et al., 2020). We measured the distance of individual Cid::EGFP clusters to the apical and basal centrosome in unperturbed wild type neuroblasts (Fig. S 2C) and found that interphase Cid was always in close proximity to the apical centrosome. After nuclear envelope breakdown (NEB), Cid moved progressively towards the metaphase plate (Fig. S 2D). Once the basal centrosome appeared (0 mins), Cid was still closer to the apical than the basal centrosome and this distance asymmetry was still intact 6 minutes after the appearance of the basal centrosome (Fig. S2E).
The proximity of Cid::EGFP clusters to the active interphase MTOC suggests a microtubule-dependent chromosome attachment mechanism. Indeed, wild type neuroblasts treated with the microtubule-depolymerizing drug colcemid showed a strong correlation between apical MTOC activity and Cid localization; as MTs depolymerize after colcemid addition, Cid progressively moved away from the apical cortex towards the cell center (Fig. S2G, H & video 6). To test whether MTOCs are connected to Cid clusters during interphase, we removed the centriolar protein Centrobin (Cnb; CNTROB in humans). Neuroblasts lacking Cnb fail to maintain an active apical interphase MTOC but regain normal MTOC activity during mitosis (Januschke et al., 2013). Neuroblasts expressing cnb RNAi lost apical Cid localization after the apical centrosome downregulated its MTOC activity. However, maturing centrosomes reconnected with Cid in prophase (Fig. S2I, J & video 7). Cid’s proximity to the apical MTOC (‘apical’ refers to the centrosome destined to move to the apical cortex) in cnb RNAi expressing neuroblasts was much more varied compared to wild type (Fig. S2L). At 6 minutes after centrosome maturation onset, Cid – apical MTOC distance was comparable to wild type, as were Cid – basal centrosome distance relationships (Fig. S2K-M). We conclude that Cid-containing chromatin is already attached to the apical centrosome prior to entry into mitosis.
Asymmetric centrosome-chromatin attachments contribute to the separation of endogenous and ectopic chromosomes
Based on these observations, we hypothesized that the separation between endogenous and ectopic chromosomes could be due to a pre-attachment mechanism, preventing mixing of Nb and GMC chromosomes. To test this hypothesis, we first analyzed Cid localization in relation to the endogenous and ectopic centrosomes in wild type hybrid cells. Similar to unperturbed wild type neuroblasts, endogenous Cid is also localized in close proximity to the endogenous apical centrosome in wild type hybrid cells (Fig. 3A-C, G, H & video 8; ‘apical’ refers to the centrosome destined to segregate into the large apical sibling cell) and appeared further away from ectopic Cid clusters compared to ectopic centrosomes (Fig. 3A-C, I & video 8; 0 mins refers to the appearance of the endogenous basal centrosome; ‘0’ refers to the appearance of the ectopic centrosomes).
We next attempted to randomize Cid – centrosome distance relationships by inducing acute fusions in cnb RNAi expressing neuroblasts, since loss of interphase MTOC activity released endogenous Cid from the apical centrosome (see above; Fig. S2G-M). In contrast to wild type hybrid cells, endogenous Cid is roughly equidistant to the endogenous and ectopic centrosomes in cnb RNAi expressing hybrid cells at 0 mins and ‘0’ mins respectively (Fig. 3D-F, J, K & video 9). However, ectopic Cid was still closer to the ectopic centrosomes in cnb RNAi expressing hybrid cells compared to the endogenous apical centrosome (Fig. 3D, F, L & video 9). In both wild type and cnb RNAi expressing hybrid cells, ectopic and basal centrosomes were about equidistant to endogenous Cid, but ectopic centrosomes were closer to ectopic Cid (Fig. S3A-D). Taken together, we conclude that in hybrid cells the apical MTOC forms an asymmetric attachment to endogenous Cid-containing chromosomes prior to entry into mitosis.
Nuclear envelopes separate ectopic and endogenous chromatin in hybrid cells
We next asked whether early MTOC – Cid attachments are sufficient to prevent endogenous and ectopic chromosome separation, tracking endogenous and ectopic Cid::EGFP after induced cell fusion. In wild type hybrid cells, endogenous and ectopic CID clusters started to congress at the metaphase plate and became difficult to clearly separate 8.8 mins (SD = 5.63; n = 5; Fig. 3M) after NEB. However, 50% of cnb RNAi hybrid cells showed endogenous and ectopic Cid mixing prior to NEB (Fig. 3D, N), although the time difference was not significantly different to wild type hybrid cells (Fig. 3M).
Neuroblasts undergo semi-closed mitosis, mostly retaining a matrix composed of nuclear envelope proteins around the mitotic spindle (Katsani et al., 2008). We imaged hybrid cells with the nuclear envelope marker Klaroid, using the protein-trap line koi::EGFP (Buszczak et al., 2007). We confirmed that unperturbed wild type neuroblasts contain a nuclear envelope matrix surrounding the mitotic spindle during mitosis (Fig. S3E). Similarly, wild type hybrid cells contain two nuclear envelopes during mitosis (Fig. 3O and Fig. S3F). Taken together, these data suggest that asymmetric MTOC-Cid attachments in interphase and nuclear envelopes establish and maintain the physical separation between endogenous and ectopic chromosomes in hybrid cells. Loss of biased interphase MTOC activity removes asymmetric MTOC-Cid attachments and allows for cross-connections between endogenous centrosomes and ectopic Cid (Fig. 3P).
Hybrid cells segregate endogenous and ectopic chromosomes independently
We next investigated whether both bipolar spindles are functional in Nb-GMC hybrid cells. Erroneous or incomplete microtubule-kinetochore attachments trigger the spindle assembly checkpoint (SAC), preventing or delaying anaphase entry (Musacchio, 2015). Since the kinetochore-derived ‘wait anaphase’ signal is diffusible (Heasley et al., 2017), ectopic spindles should thus also obey the SAC in Nb-GMC hybrid cells. We tested whether hybrid cells contain functional microtubule-kinetochore attachments by measuring the time between finished chromosome alignment at the metaphase plate and chromosome separation in Nb-GMC hybrids expressing His2A::GFP and cherry::Jupiter (Fig. 4A, B). Unperturbed control neuroblasts usually initiate anaphase onset within 2.63 minutes (SD=1.84; n=12) after chromosomes are aligned at the metaphase plate. In hybrid cells derived from interphase fusions, endogenous and ectopic chromatin entered anaphase 4.14 mins (SD= 2.23; n=14) and 5.12 minutes (SD= 1.85; n=13) after metaphase alignment. Only ectopic chromatin for prophase induced hybrid cells showed a significantly delayed anaphase onset (Average: 10.25 mins; SD= 4.12; n=6) (Fig. 4C). Ectopic chromatin never separated before endogenous chromosomes but entered anaphase with a few minutes delay (Fig. 4D). We conclude that in Nb-GMC hybrids, endogenous and ectopic chromosomes establish correct MT-kinetochore attachments, thereby fulfilling the spindle assembly checkpoint necessary to enter anaphase. However, given the delays in ectopic chromosome separation, we further conclude that ectopic spindles can initiate chromatid separation independently from the endogenous neuroblast spindle.
Hybrid cells are insufficient to induce tumors in wild type host flies
Finally, we assessed the accuracy of chromosome segregation in wild type hybrid cells. Using the canonical chromosome marker His2A::GFP we detected chromosome missegregation - ranging from lagging chromosomes to chromosome bridges - in all wild type hybrid cells (Fig. 4E,F & video 10). Chromosome segregation defects can result in aneuploidy and micronuclei formation (Molina et al., 2020). We found a small percentage of hybrid cells containing micronuclei, but more frequently discovered heterokaryons (hybrid cells containing two nuclei of different origin). In most cases, hybrid cells fused both nuclei into one, forming synkaryons (Fig. 4G).
Aneuploidy has been proposed to be a hallmark of cancer (Molina et al., 2020) but appears to be context dependent (Ben-David and Amon, 2020). To test whether neuroblast – GMC derived hybrid cells are sufficient to induce tumor formation, we grafted His2A::GFP expressing larval fly brains after successful induction of cell fusion into the abdomen of wild type hosts (Rossi and Gonzalez, 2015) and monitored the host flies for tumor formation and life span changes. As previously reported (Caussinus and Gonzalez, 2005), brat RNAi expressing brains formed visible tumors in host flies by day 30 and caused a reduction in lifespan of the host. However, larval brains without attempted fusions (wild type transplants), attempted but unsuccessful fusions (controlling for the effect of laser ablation), or induced fusion, showed neither tumor growth by day 30 (or after), nor a reduction in lifespan (Fig. 5A,B). We conclude that chromosome segregation is defective in Nb-GMC hybrid cells but insufficient to form visible tumors in otherwise normal and unperturbed larval fly brains.
Discussion
Cell – cell fusion can occur under normal physiological conditions and has been implicated in malignancy (Platt and Cascalho, 2019) but how hybrid cells recognize and separate endogenous and ectopic chromosomes during mitosis is unclear. Here we have shown in vivo that Nb-GMC derived hybrid cells separate endogenous chromosomes from the introduced ectopic GMC chromosomes until metaphase and potentially during anaphase. Chromosome separation is achieved through a dual-spindle mechanism; both endogenous and ectopic chromosomes nucleate two independent spindles, most likely through the canonical centrosome pathway. These dual spindles co-align during metaphase, thereby congressing the two chromosome clusters at the metaphase plate. Endogenous and ectopic chromosomes independently segregate during anaphase, manifested in delayed segregation onset of ectopic chromatin. Although chromosome missegregation is frequent in hybrid cells, potentially leading to aneuploidy, we failed to detect malignant tumor formation when hybrid cell – containing larval brains were grafted into wild type host flies.
Taken together, we propose that endogenous and ectopic chromosome separation is achieved through an early microtubule-dependent chromosome capture mechanism that retains endogenous chromosomes in close proximity to the apical neuroblast cortex during interphase. In addition to this geometrical separation, the nuclear envelope might impose a physical boundary, preventing random chromosome mixing prior to chromosome congression (Fig. 5C). In pre-mitotic neuroblasts, chromatin can be connected with centrosomes through the Linc complex (Lee and Burke, 2018), potentially implicating the SUN domain protein Klaroid (Kracklauer et al., 2007) and the KASH-domain protein Klarsicht (Razafsky and Hodzic, 2009; Lee and Burke, 2018) in asymmetric chromatin clustering and the prevention of chromatin mixing during interphase. The chromatin separation mechanisms described here could be applicable to chromosome separation occurring in the first cleavage after fertilization in insects, arthropods and vertebrates (Reichmann et al., 2018; Kawamura, 2001). Similarly, biased chromatid segregation has been observed in stem cells (Yadlapalli and Yamashita, 2013; Ranjan et al., 2019) and meiosis (Akera et al., 2017). Since centromeres have also been found to be confined to specific nuclear locations in many organisms (Weierich et al., 2003; Muller et al., 2019), it will be interesting to see whether microtubule-dependent chromatin attachment provides an alternative mechanism for biased sister chromatid segregation or other important cellular functions.
Supplemental Table 1: Statistical information
Complete statistical information for the data shown in the corresponding Figures.
Methods
Fly Strains
Transgenes and fluorescent markers: worGal4, UAS-mCherry::Jupiter (Cabernard and Doe, 2009); worGal4, UAS-mCherry::Jupiter, Sqh::GFp (Cabernard et al., 2010); His2A::GFp (Bloomington stock center); UAS-mCherry::CAAX, UAS-iLID::CAAX:;mCherry (A. Monnard & C. Cabernard; unpublised); Cid::EGFP (Ranjan et al., 2019); pUbq-Asl::GFp (Blachon et al., 2008); worgal4, UAS-mCherry::Jupiter, Asl::GFp (this work); pros::EGFp (endogenously tagged with CRISPR; this work); koi::GFP (CB04483) (Buszczak et al., 2007), cnbGD11735 RNAi line (v28651) (Dietzl et al., 2007).
Generation of pros::EGFP with CRISPR
Target specific sequences with high efficiency were chosen using the CRISPR Optimal Target Finder (http://tools.flycrispr.molbio.wisc.edu/targetFinder/), the DRSC CRISPR finder (http://www.flyrnai.org/crispr/), and the Efficiency Predictor (http://www.flyrnai.org/evaluateCrispr/) web tools. Sense and antisense primers for these chosen sites were then cloned into pU6-BbsI-ChiRNA (Gratz et al., 2013) between BbsI sites. To generate the replacement donor template, EGFP and 1 kb homology arms flanking the insertion site were cloned into pHD-DsRed-attP (Addgene plasmid #51019) using Infusion technology (Takara/Clontech). Injections were performed in house. Successful events were detected by DsRed-positive screening in the F1 generation. Constitutively active Cre (BDSC#851) was then crossed in to remove the DsRed marker. Positive events were then balanced, genotyped, and sequenced.
Live cell imaging acute cell-cell fusion
Imaging medium consists of Schneider’s insect medium (Sigma-Aldrich S0146) mixed with 10% BGS (HyClone). Third instar larvae were dissected in imaging medium and the brains were transferred into a μ-slide Angiogenesis or μ-slide 8 well (Ibidi). Live samples were imaged with an Intelligent Imaging Innovations (3i) spinning disc confocal system, consisting of a Yokogawa CSU-W1 spinning disc unit and two Prime 95B Scientific CMOS cameras. A 60x/1.4NA oil immersion objective mounted on a Nikon Eclipse Ti microscope was used for imaging. Live imaging voxels are 0.22 × 0.22 × 0.75-1μm (60x/1.4NA spinning disc).
Neuroblast-GMC fusions were induced using a 3i Ablate! ablation system, consisting of a 532nm pulsed laser. We used a pulse width of 7 ns, targeting the membrane interface between the neuroblast and the adjacent GMC.
Colcemid treatment
Dissected brains were incubated with Colcemid (Sigma) in live imaging medium at a final concentration of of 25 μgmL−1.
Transplantation experiments
Brain lobes containing the hybrid cells expressing His2A::GFp and worGal4, UAS-mCherry::Jupiter were transplanted into 3 to 4 day old, well fed adult w1118 female host flies as described previously (Rossi and Gonzalez, 2015). Custom made needles were prepared from Narishige GD-1 glass capillaries using a Narishige, needle puller. Injection needles were shaped with forceps to have a smooth, 45° opening. Transplanted flies were transferred into fresh vials each day for the first three days, followed by biweekly flipping. The tumor growth was monitored and recorded with a Leica MZ FLIII fluorescence stereomicroscope.
Image processing and measurements
Live cell images were processed using imaris x64 8.3.1 and image J.
For angle and distance measurements, the coordinates for the two spindle poles were determined in Imaris. From these coordinates, angles and distances between spindles were derived based on the calculations outlined below.
Angle between spindles:
Dot product: n · e = (X1*X2)+(Y1*Y2)+(Z1*Z2)
Magnitude of vectors: Where n corresponds to the spindle vector: n(x1,y1,z1) = (N1-N1’, N2-N2’, N3-N3’) and e to the ectopic spindle vector: e(x2,y2,z2) = (E1-E1’, E2-E2’, E3-E3’) N1, N2, N3 and N1’, N2’ and N3’ are coordinates of the two poles of the Nb spindle. Similarly, E1, E2, and E3 and E1’, E2’ and E3’ are coordinates of the ectopic spindle poles.
Distance between spindle vectors
The midpoints of the two spindle vectors are calculated from coordinates of the poles on either side of the respective spindle. This is followed by calculating the distance between these midpoints.
Midpoint of the Nb spindle vector =
Midpoint of the GMC spindle vector =
Distance between these two points =
Centrosome - Cid distance
The centrosome (CS) - Cid distance was calculated using Cid and CS coordinates.
CS –Cid distance:
Where x1,y1,z1 correspond to CS and x2,y2,z2 to Cid coordinates, respectively.
Plotted values correspond to averaged values of all CS –Cid punctae distances and the corresponding standard deviations.
0 and 6 mins corresponds to the appearance of the basal centrosome and 6 minutes thereafter. ‘0’ mins and ‘6’ mins corresponds to the appearance of the ectopic centrosome appearance and 6 minutes thereafter.
Statistical analysis
Statistical analysis was performed using Graphpad prism 8. Statistical significance was determined using paired or unpaired t-test and one-way ANOVA. Significance was indicated as following: *; p<0.05, **;p<0.01, ***;p<0.001,****p<0.0001, ns; not significant. Exact p values and complete statistical information can be found in Extended data table 1.
Author contributions
This study was conceived by B.S., N.L. and C.C. B.S and N.L performed all the experiments with significant help from C.S. B. S, N.L., C.S., and C.C analyzed the data. B.S and C.C. wrote the manuscript.
Competing interest declaration
The authors declare no competing financial interests.
Additional information
This paper contains three Fig. Ss, a supplementary table and 10 videos to support the main conclusions.
Video legends
Video 1: Wild type neuroblast division; related to Fig. 1B
Wild type control (unfused) neuroblast expressing the microtubule binding protein Cherry::Jupiter (white) and the canonical Histone marker His2A::GFP (cyan). Time scale is h:mm:ss and the scale bar is 5 μm
Video 2: Wild type hybrid cell; related to Fig. 1B
Wild type hybrid cell derived from a neuroblast –GMC fusion in vivo, expressing the canonical Histone marker His2A::GFP (white in single channel; cyan in merge) and the microtubule binding protein Cherry::Jupiter (white). The blue and orange arrows mark endogenous and ectopic chromatin, respectively. Time scale is h:mm:ss and the scale bar is 3 μm.
Video 3: Wild type hybrid cell; related to Fig. 2B
Wild type hybrid cell derived from a neuroblast –GMC fusion in vivo, expressing the canonical Histone marker His2A::GFP (white in single channel; cyan in merge) and the microtubule binding protein Cherry::Jupiter (white). The blue and orange arrows mark endogenous and ectopic chromatin, respectively. Time scale is h:mm:ss and the scale bar is 2 μm.
Video 4: Wild type neuroblast division; related to Fig. S2B
Wild type control (unfused) neuroblast, expressing the microtubule binding protein Cherry::Jupiter (white) and the centromere-specific H3 variant Cid::EGFP (Cyan). Purple and yellow arrows point to the apical and basal centrosome, respectively. The blue arrow refers to moving Cid clusters. Time scale is h:mm:ss and the scale bar is 1 μm.
Video 5: Wild type neuroblast division; related to Fig. S2F
Wild type control (unfused) neuroblast, expressing the membrane marker mCherry::CAAX (white), the canonical Histone marker His2A::GFP (white) and Cid::EGFP (Cyan). The green arrow points to Cid clusters. Time scale is h:mm:ss and the scale bar is 5 μm.
Video 6: Wild type neuroblast exposed to Colcemid; related to Fig. S2G
Wild type control (unfused) neuroblast exposed to the microtubule depolymerizing drug Colcemid, expressing the membrane marker mCherry::CAAX (white), the microtubule binding protein Cherry::Jupiter (white) and Cid::EGFP (Cyan). The yellow arrow points to the apical centrosome, the blue arrow to Cid clusters. Time scale is h:mm:ss and the scale bar is 1 μm.
Video 7: Cnb RNAi expressing neuroblast; related to Fig. S2I
Cnb RNAi expressing (unfused) neuroblast, co-expressing the microtubule binding protein Cherry::Jupiter (white) and Cid::EGFP (Cyan). The orange and blue arrow points to the apical and basal centrosome, respectively. The green arrow highlights Cid clusters. Time scale is h:mm:ss and the scale bar is 1 μm.
Video 8: wild type hybrid cell; related to Fig. S3A
Wild type hybrid cells expressing the microtubule binding protein Cherry::Jupiter (white) and Cid::EGFP (white in single channel; Cyan in merge). The blue and orange arrow highlights endogenous and ectopic Cid clusters, respectively. Time scale is h:mm:ss and the scale bar is 1 μm.
Video 9: Cnb RNAi expressing hybrid cell; related to Fig. S3D
Cnb RNAi expressing hybrid cell, expressing the microtubule binding protein Cherry::Jupiter (white) and Cid::EGFP (white in single channel; Cyan in merge). The blue and orange arrow highlights endogenous and ectopic Cid clusters, respectively. The yellow arrow indicates mixing of endogenous and ectopic Cid clusters. Time scale is h:mm:ss and the scale bar is 1 μm.
Video 10: wild type hybrid cell; related to Fig. 4E
Wild type hybrid cell, expressing the canonical Histone marker His2A::GFP (white). The blue and orange arrow highlights endogenous and ectopic chromosomes, respectively. The magenta arrowhead highlights the fate of missegregated chromosomes. This hybrid cell forms a heterokaryon. Time scale is h:mm:ss and the scale bar is 1 μm.
Acknowledgements
We thank Xin Chen for fly stocks, David Salvador Garcia for generating the Pros::EGFP transgenic line, Sue Biggins and members of the Cabernard laboratory for helpful discussions and comments. This work was supported by the National Institutes of Health (1R01GM126029-03) and a Research Scholar grant from the American Cancer Society (130285-RSG-16253-01-CSM). Stocks obtained from the Bloomington Drosophila Stock Center (NIH P40OD018537) and from the Vienna Drosophila Resource Center (VDRC).