ABSTRACT
In the Caenorhabditis elegans zygote, astral microtubules generate forces, pushing against and pulling from the cell periphery. They are essential to position the mitotic spindle and in turn the cytokinesis furrow, ensuring the proper distribution of fate determinants to the daughter cells. By measuring the dynamics of astral microtubules, we revealed the presence of two populations, residing at the cortex during 0.4 s and 1.8 s, proposed to reflect the pulling and pushing events, respectively. This is a unique opportunity to unravel the time and space variations of these both spindle-positioning forces, to study their regulation under physiological conditions. By an investigation at the microscopic level, we first confirmed that the asymmetry in pulling forces was encoded by an anteroposterior imbalance in dynein-engaging rate, and that this asymmetry exists from early metaphase on and accounts for the final spindle position. More importantly, we obtained direct proof of the temporal control of pulling forces through the force-generator processivity increase during anaphase. Lastly, we discovered an anti-correlation between the long-lived population density and the stability of the spindle position during metaphase, which strongly suggests that the pushing forces contribute to maintaining the spindle at the cell centre.
INTRODUCTION
During asymmetric divisions, the position of the mitotic spindle is precisely regulated to correctly partition cell fate determinants, which is crucial to ensure faithful divisions during developmental processes (Gönczy, 2008; Neumüller and Knoblich, 2009; Morin and Bellaïche, 2011; McNally, 2013). In the one-cell embryo of the nematode Caenorhabditis elegans, the mitotic spindle is first oriented along the polarity axis and positioned at the cell centre. Then, the spindle is precisely maintained at that position for a few minutes during metaphase. Finally, it is displaced out of the cell centre prior to division (Gönczy, 2008; McNally, 2013). So far, cell-scale investigations revealed some forces at the core of this precise choreography, in particular, the pulling on astral microtubules from the cell cortex accounting for the posterior displacement of the spindle starting from mid metaphase. The force generators are composed of the dynein/dynactin complex, the LIN-5NuMA protein, and the G-protein regulators GPR-1/2LGN, and are anchored at the membrane through Gα subunits (Gotta and Ahringer, 2001; Colombo et al., 2003; Srinivasan et al., 2003; Couwenbergs et al., 2007; Nguyen-Ngoc et al., 2007). These force generators contribute by molecular motor action and by coupling with the microtubule depolymerisation (Schmidt et al., 2005; Kozlowski et al., 2007; Nguyen-Ngoc et al., 2007; O’Rourke et al., 2010; Laan et al., 2012a). Balancing this cortical pulling, the centring force which maintains the spindle at the cell centre during metaphase is still under debate (Reinsch and Gonczy, 1998; Wühr et al., 2009; Zhu et al., 2010; Kimura and Kimura, 2011a; McNally, 2013; Wu et al., 2017), with three possible forces implicated: the cortical pulling forces (Grill and Hyman, 2005; Kimura and Onami, 2005); the cytoplasmic pulling forces (Kimura and Kimura, 2011a; Shinar et al., 2011; Barbosa et al., 2017); and the cortical pushing forces (Pecreaux et al., 2006a; Garzon-Coral et al., 2016). So far, these studies were all based on cell-scale measures.
How these two mechanical forces, previously studied separately and spatially or temporally averaged, are regulated and coordinated in space and throughout mitosis? The cortical pulling forces are asymmetric, because of a higher number of active force generators – trimeric complexes engaged in pulling events with astral microtubules – at the posterior-most side of the embryo (Gotta et al., 2003; Grill et al., 2003; Pecreaux et al., 2006a; Nguyen-Ngoc et al., 2007), in response to polarity cues (Grill et al., 2001; Colombo et al., 2003; Tsou et al., 2003a; Park and Rose, 2008; Bouvrais et al., 2018). On top of that, the physical ground of the progressive increase in the pulling forces along the course of the division, was inferred from cell scale measurements (Labbe et al., 2004; Pecreaux et al., 2006a; Campbell et al., 2009; Bouvrais et al., 2018), and a molecular mechanism is still missing. Finally, little is known on the spatiotemporal regulation of the centring forces.
Astral microtubules are involved in generating all these forces. These semi-flexible filaments emanate from the spindle poles. They are dynamic, switching alternatively from growing to shrinking and back, at the catastrophe and rescue rates, respectively (Mitchison and Kirschner, 1984). At the cortex, astral microtubules can be in three different states: shrinking and pulling from there through minus-end directed dynein molecules anchored at the cortex, which coordinate their motor action with microtubule shrinkage (Gonczy et al., 1999; Dujardin and Vallee, 2002; Grishchuk et al., 2005; Gusnowski and Srayko, 2011; Laan et al., 2012a; Rodriguez-Garcia et al., 2018); pushing and growing against the cortex (Faivre-Moskalenko and Dogterom, 2002; Dogterom et al., 2005; Howard, 2006), helped by stabilizing associated proteins like CLASP (Espiritu et al., 2012); or stalled, with microtubules being clamped at the cortex, neither growing or shrinking, possibly through dynein tethering (Hendricks et al., 2012; Laan et al., 2012a; Laan et al., 2012b) or stabilizing proteins (Labbé et al., 2003; Sugioka et al., 2018). These different states are revealed by the microtubule dynamics at the cortex, especially their residence times that reflect their catastrophe rates. Furthermore, in vitro and computational studies have shown that dynein tethering delayed microtubule catastrophe (Hendricks et al., 2012; Laan et al., 2012a), while larger pushing forces would advance it (Janson et al., 2003). Consequently, at cell-scale, different residence times are expected for microtubules involved either in pulling forces during anaphase (Pecreaux et al., 2006a) or in pushing forces during metaphase (Pécréaux et al., 2016) in C. elegans embryo, revealing the cortical microtubule-mediated force generating events.
Previous studies focussing on the residence times of astral microtubules at the C. elegans embryo cortex revealed some spatial modulations. In particular, microtubules are suggested to be more dynamic (lower lifetime) at the posterior cortex compared to anterior cortex, although the reported residence times are strikingly different between distinct studies (Labbé et al., 2003; Kozlowski et al., 2007; Gerhold et al., 2018; Sugioka and Bowerman, 2018; Sugioka et al., 2018). Furthermore, these studies provide reduced indications about how the contact durations evolve throughout mitosis. This limitation is made more acute by the short duration of these contacts (a few frames), in low signal-to-noise ratio images. Recent developments in microscopy and image processing tools call for revisiting this problem (Chenouard et al., 2014; Meiniel et al., 2018).
Beyond imaging improvements, the statistical analysis of the microtubule-end tracks at the cortex could also be significantly refined in contrast to the classic fit with a mono-exponential distribution (Kozlowski et al., 2007; Sugioka et al., 2018). In particular, and to distinguish several co-existing dynamical behaviours, we here fitted the experimental distribution of the contact durations with a finite-mixture-of-exponential models and then used an objective criterion to choose which model best fits the data. Such an approach, although delicate, benefits of developments in applied mathematics (Grinvald and Steinberg, 1974; James and Ware, 1985; Vieland and Hodge, 1998; Jae Myung et al., 2000; Turton et al., 2003). The microtubule residence times could last only a few tenths of a second, corresponding to a few frames, calling for specific analysis as performed in photon counting experiments. This field has designed appropriate fitting strategies that offer a firm starting point to analyse microtubule dynamics (Maus et al., 2001; Turton et al., 2003; Nishimura and Tamura, 2005; Laurence and Chromy, 2010).
In the present paper, we aimed to study the spatiotemporal regulation of the spindle-positioning forces during the first division of the C. elegans embryo. To do so, the microtubule dynamics at the cortex were instrumental. We designed the DiLiPop assay for “Distinct Lifetime subPopulation assay” to faithfully disentangle several microtubule populations, distinct by their dynamics at the cortex and found two of them, which we could associate with different microtubule functions. Equipped with this assay, we could for the first time resolve in time and space, and at the microscopic level, how the forces positioning the spindle are regulated during mitosis. In particular, first, we measured a force-generator processivity increase during anaphase revealing how pulling forces are regulated throughout mitosis; and, second, we identified which mechanism enables to maintain the spindle at cell centre during metaphase. Importantly, we also suggest how these forces coordinate in space and time.
RESULTS
The Distinct Lifetime (sub)Population assay reveals two populations of microtubules at the cortex
To investigate the regulation of the forces exerted on the mitotic spindle during the first mitosis of the C. elegans zygote, we set to measure the dynamics of astral microtubules at the cortex. The microtubules were entirely fluorescently labelled using YFP::α-tubulin (Supp. Text §1.1.1), since we aimed to visualise the microtubules in all their force generating modalities, revealed by growing, shrinking or stalled states. We performed spinning disk microscopy at the cortical plane, at 10 frames per second similarly to (Bouvrais et al., 2018) (Supp. Text §1.1.2). The microtubules contacting the cortex end-on, they appeared as spots during metaphase and anaphase (Movie S1) (Gusnowski and Srayko, 2011). The high frame rate, needed to resolve microtubule residing briefly at the cortex, led to images with a low signal-to-noise ratio (Figure 1A, top). We thus denoised them using the Kalman filter (Figure 1A, middle) (Kalman, 1960) before tracking the microtubule contacts using u-track algorithm (Figure 1A, bottom, Table S1, Supp. Text §1.1.3) (Jaqaman et al., 2008). This image-processing pipeline is further named KUT.
We computed the duration distributions of the tracks for each embryo separately (to avoid averaging artefacts) (Figure 1B). When all the microtubules have the same catastrophe rate, this distribution follows an exponential decay (Kozlowski et al., 2007; Floyd et al., 2010). However, we also envisaged that multiple force-generating mechanisms involving microtubules are superimposed, corresponding to distinct catastrophe rates. Therefore, we fitted the distribution with finite-mixture-of-exponential models, in particular double- and triple-exponential models (Supp. Text §1.2.1). Double-exponential better fitted the duration-distribution suggesting that we observed at least two populations of microtubules at the cortex of C. elegans embryo, distinct by their residence time (Figure 1C).
To faithfully distinguish the number of microtubule-contact populations present at the cortex of the C. elegans embryo and to finely characterise them, we implemented an advanced statistical analysis of the contact duration distribution (Figure 1D). This offers improvements with respect to published approaches on four aspects. (a) When fitting data points, whose standard errors are Gaussian-distributed, maximising the model likelihood often corresponds to least-square fitting. However, we fitted a histogram with few counts in some bins. Thus, the error distribution followed a Poisson law. Therefore, we designed the corresponding objective function (Figure 1D, grey) (Supp. Text §1.2.2) (Laurence and Chromy, 2010). (b) To distinguish between multiple microtubule-population dynamics within each embryo, and embryos with single dynamics but different from each other, we avoided averaging the data from multiple embryos but instead fitted the same model on each embryo by global fitting, i.e. maximizing the global likelihood being the product of the embryo-wise likelihoods (Figure 1D, orange) (Supp. Text §1.2.3). (c) We performed an objective selection of the best mixture-of-exponential model using the Bayesian Inference Criterion (BIC) (Schwarz, 1978). The number of exponentials in the best model suggests, without any hypothesis, the number of microtubule dynamical behaviours within an embryo (Figures 1D, blue) (Supp. Text §1.2.4). (d) To ensure proper comparison of the results between conditions, we computed the confidence intervals on the fitted parameters using two different approaches, bootstrapping and likelihood ratio (Bolker, 2008; Agresti, 2013) (Figures 1D, purple) (Supp. Text §1.2.5). Applying this approach to untreated embryos of C. elegans, we found two dynamical behaviours for the microtubule contacts at the cortex (Figure 1EF).
We reasoned that supporting models featuring more parameters needs more data points to be selected by methods like the Bayesian Inference Criterion. We thus set to explore whether a third population with a very-short or very-long lifetime might exist. A very-short-lived population may correspond to astral microtubules growing, touching and immediately leaving because they are neither captured by a motor (dynein) nor stabilised by a MAP (Janson et al., 2003). However, the triple-exponential model fitting rather suggests a longer lifetime for the third population (Table S2). Furthermore, microtubules performing touch-and-go at the cortex are not likely to contribute to cortical force generation there. The triple-exponential model suggested the possible presence of a 4 s lifetime third population, however with a proportion below 4%. Biologically, some microtubules could be stabilised at the cortex (Sugioka et al., 2018) or stalled there (Hendricks et al., 2012; Laan et al., 2012a; Laan et al., 2012b). We asked whether the ∼20 000 microtubule contacts per embryo in the experimental data are sufficient to support the detection of a putative third population and addressed this question in silico (Supp. Text §2.2.2). We fabricated embryo data simulating contacts, referred as “fabricated embryos”, with three dynamical behaviours of lifetimes 0.4 s, 1.5 s and 4 s, and population proportions set to 55%, 40% and 5%, respectively. To mimic the experimental investigations of a given condition, here untreated embryos, we generated 10 “simulated samples”, composed of 25 fabricated embryos, which we analysed using the DiLiPop statistical assay. A sample size of 20 000 tracks or more per embryo was necessary for the triple-exponential model to be selected as the best one in most of the fabricated samples (Supplemental Figure S1D). Larger sample sizes enabled more accurate model-parameter estimates. Our experimental data thus contained enough tracks to allow identifying a putative third population. We reckoned that another limitation precluding the detection of a third population could be its small proportion and set to determine in silico the minimum requirement. We fabricated embryos with 20 000 tracks for each and three populations of lifetimes 0.4 s, 1.5 s and 4 s, the proportion of the short-lived population was kept to 55% and the third population one ranged from 2.5% to 10%. We found that third-population proportions at least equal to 5% were necessary to support the triple-exponential model as preferred one (Supplemental Figure S1E). We concluded that in untreated embryos, there is few (less than 5%) or no very-long-lived population of astral microtubules, contributing little or no to cortical force generation. Finally, we asked whether the dual-exponential dynamics of the microtubule contacts correspond to two well-defined astral microtubule populations or may reflect a broad variety of dynamical behaviours. This latter case corresponded to the stretched-exponential model and was excluded by the model selection criterion (Table S2) (Lee et al., 2001; Siegel et al., 2001), but we asked once again whether the amount of experimental data may have allowed identifying such a case. By in silico investigation, using fabricated embryos with contacts displaying a stretched exponential behaviour of lifetime 0.1 s and power 2.2, we found that 500 tracks were sufficient (Supp. Text §2.2.1), although a larger number increased the accuracy of recovered parameters (Supplemental Figure S1E). Because we had much more tracks in experimental data, we concluded that the two dynamical behaviours measured in vivo really corresponded to two populations of microtubule-contact tracks.
We next asked whether the dynamical behaviours of the tracks reflect the presence of two populations of microtubules and investigated possible artefacts caused firstly by the microtubule labelling. Indeed, it was recently proposed that tubulin paralogs composition of the microtubules impacted their dynamics (Honda et al., 2017) or the spindle positioning (Wright and Hunter, 2003) in the C. elegans one-cell embryo. Thus, we asked whether YFP::α-tubulin fluorescent labelling could create two populations. We performed a similar DiLiPop analysis in GFP::β-tubulin-labelled embryos and measured two dynamical behaviours again at the cortex, with characteristic lifetimes very similar to the ones obtained with α-tubulin labelling (Figure 2A). We concluded that the two dynamically-distinct microtubule populations do not depend on a specific entire-microtubule labelling.
To complete artefact exclusion, we challenged the image-processing pipeline in two ways. Firstly, and in contrast with simulations of tracks durations above, we fabricated images containing particles with a single dynamical behaviour (Figure 2E, 2F black, Table S3). By DiLiPop analysis of in silico images, we recovered the mono-exponential as the best model, with the correct parameters (Figure 2F, red). In contrast, when we performed a similar simulation with two dynamical behaviours, we recovered double-exponential as best model and accurate parameters (Figure 2G, blue). Overall, this suggests that the KUT image-processing pipeline used here performed without causing artefacts. Secondly, and to gain certainty, we repeated the analysis on in vivo data using a different image-processing pipeline composed of algorithms based on different hypotheses. This pipeline, named NAM, encompasses the ND-SAFIR denoising (Boulanger et al., 2010) (Figure 2D2), the ATLAS spot-detecting (Basset et al., 2015), and the MHT linking (Multiple Hypothesis Tracker) (Reid, 1979) (Figure 2D3). Corresponding settings are listed in Table S4 (Supp. Text §2.1.1). Applied to untreated C. elegans embryos, the NAM pipeline combined to DiLiPop statistical analysis produced results similar to the KUT pipeline (Figure 2C). Therefore, the two dynamical behaviours suggested by the analysis are not artefactual.
Finally, to test whether this double dynamical behaviour of microtubule cortical contacts is specific to C. elegans, we investigated the microtubule dynamics at the cortex in Caenorhabditis briggsae β-tubulin-labelled embryos, a nematode cousin of C. elegans, and we also observed two populations distinguished by their different dynamical behaviours (Figure 2B). We concluded that these two populations are not a peculiarity of C. elegans embryo. They may instead originate from conserved cortical force generation mechanisms between these two species, which share already spindle positioning systems (Riche et al., 2013; Bouvrais et al., 2018).
Overall, by viewing entirely-labelled microtubules, we measured two populations of astral microtubules at the cortex. There, they can generate pulling and pushing forces and in particular, pushing microtubules, in growing state, are likely to reside longer, while pulling ones coordinate dynein action, lasting about 0.5 s, with depolymerisation (Rodriguez-Garcia et al., 2018). Therefore, we formulated our two-population hypothesis, stating that long-lived microtubules could be the pushing ones and those short-lived the pulling ones (Figure 3E).
The short- and long-lived microtubules contribute to pulling from and pushing against the cortex
We set to test whether the short-lived population reflected the pulling-force-generating events. If so, its cortical density is expected to be polarised, especially during anaphase, to cause the spindle posterior-displacement. We refined the DiLiPop assay to measure the evolution of both microtubule populations along the mitosis and to map them in the embryo. The region and period used to perform such a block-analysis were selected to preserve sample size large enough to guarantee proper detection of the two populations (Supp. Text §1.3, Supplemental Figure S2), while the procedure to represent each population in DiLiPop-maps was similar to our previous one used to map all microtubules without dynamic classification (Bouvrais et al., 2018). Applied to untreated C. elegans embryos (N = 25), this analysis recovered the high-density ridgelines of microtubule-contacts for both populations (Figure 3A), which corresponded to the centrosomal positions (Bouvrais et al., 2018). Importantly, from 100 s before anaphase on, the short-lived contact-density was larger in the 70-100% region (classically measured along the anteroposterior (AP) axis) than in the anterior cortex (Figure 3A1), while the long-lived population was only slightly polarized at late anaphase, as expected because of the posterior displacement of the spindle (Figure 3A2) (Bouvrais et al., 2018). The polarisation of the short-lived but almost not the long-lived population suggests that the former reflects pulling-force events.
To strengthen our two-population hypothesis, we genetically decreased or increased cortical pulling forces and observed how the spatial distribution of the short-lived-microtubule density is affected (Rodriguez-Garcia et al., 2018). To do so, we respectively depleted GPR-1/2LGN, the well-established force generator regulator (Colombo et al., 2003; Grill et al., 2003; Pecreaux et al., 2006a), and CNSK-1 involved in repressing anterior forces (Panbianco et al., 2008). Firstly, we treated gpr-2(ok1179) mutant embryos by gpr-1/2(RNAi) to ensure a more pronounced depletion and computed the DiLiPop-map. We observed a strong reduction in the short-lived-microtubule density in the posterior region in comparison with the control (Figure 3B1-2, Supplemental Figure S4B1) leading to an almost symmetric anteroposterior distribution, while the anterior region was mostly not affected. The slight asymmetry of the long-lived-microtubule density in late anaphase also tended to vanish (Supplemental Figure S4A2,B2), as expected since the spindle position was more centred (Supplemental Figure S5C1-2). To check on the penetrance of the GPR-1/2 depletion, we monitored the cortical pulling forces, using spindle pole oscillations and displacements, in N = 8 embryos treated in the same way and imaged at the spindle plane. We observed no oscillation and a strong reduction of the posterior centrosome displacement as reported previously (Supplemental Figure S5C1-2) (Colombo et al., 2003; Pecreaux et al., 2006a). Similarly to the above analysis, we tested in silico whether the number of microtubule-contacts in each region and period was sufficient to permit detecting two populations (Supplemental Text §2.2.3) and found that with Ne = 8 fabricated embryos and sample size of 3000 tracks, mimicking the experiments with protein depletion, lifetime difference T2-T1>0.4 s is sufficient if the proportion P1 is between 30% and 70%, else T2-T1>1 s is required (Supplemental Figure S3B-G). We concluded that the short-lived-population distribution strongly depends on GPR-1/2 supporting that these microtubules contribute to pulling force generating.
Secondly, we performed the converse experiment, enriching the force generators anteriorly through a cnsk-1(RNAi) treatment (Panbianco et al., 2008). Computing the DiLiPop map, we observed an increase in the short-lived microtubule-contact densities anteriorly, compared to controls (Figure 3B3, Supplemental Figure S4B1), as expected. We also observed a slight decrease in the short-lived microtubule-contact densities at the posterior crescent attributed to the anterior displacement of the spindle (Supplemental Figure S4B1, S5C1,3) (Bouvrais et al., 2018; Rodriguez-Garcia et al., 2018). The long-lived microtubule density overall increased possibly because an altered actin-myosin network is indirectly affecting microtubule dynamics (Severson and Bowerman, 2003; Redemann et al., 2010) or because of a more direct CSNK-1 dependent change in microtubule dynamics (Fievet et al., 2013) (see also micrographs in (Panbianco et al., 2008)) (Supplemental Figure S4A3,B2). To check on the penetrance of our treatment, we imaged N = 9 similarly cnsk-1(RNAi)-treated embryos at the spindle plane and observed mildly increased oscillations at the posterior pole, with a peak-to-peak amplitude of 8.10 ± 1.99 µm compared to 6.43 ± 1.89 µm (p = 0.079), and more substantially increased ones anteriorly, 6.64 ± 0.76 µm compared to 2.66 ± 0.84 µm in control embryos (p = 7.5×10−9), confirming the significant increase in pulling forces mostly at anterior (Panbianco et al., 2008). Overall, the short-lived-microtubule density correlates with the cortical force intensity and the number of force generators (Grill et al., 2003; Panbianco et al., 2008; Rodriguez-Garcia et al., 2018), supporting our two-population hypothesis.
Our hypothesis also suggests that the long-lived microtubules correspond to cortical-pushing events. To support this view, we investigated the two cortical-contact populations upon genetically altering the dynamics of the microtubules and then studied microtubule dynamics using an EB-labelling.
Firstly, we impaired microtubules growing by depleting the microtubule polymerase ZYG-9XMAP-215 by RNAi (Figure 3C1) and performed a DiLiPop analysis during metaphase considering all regions. We found two populations but interestingly, only the long-lived microtubules had their lifetime reduced while the short-lived one was unaltered (Figure 3C2-3, green). This was consistent with the microtubule polymerase activity of ZYG-9 (Brouhard et al., 2008) and supported our hypothesis that the long-lived population accounts for growing microtubules. We checked on the penetrance by repeating the experiment and imaging at the spindle plane. We observed, during the centring establishment, an incomplete anterior displacement of the spindle, which reached 0.63 ± 0.85 µm upon zyg-9(RNAi) (N = 10) compared to – 0.7 ± 0.5 µm in control embryos (N = 13, p = 6.5×10−4). This phenotype was expected and confirmed the ZYG-9 penetrance (Grill et al., 2001). To confirm that the observed alteration of the long-lived microtubule dynamics corresponds specifically to a perturbation of the growth rate and not to generic altered dynamics, we depleted the microtubule depolymerase KLP-7MCAK by RNAi (Figure 3C1). The DiLiPop analysis revealed no significant change in the lifetime of the long-lived population (Figure 3A2, blue). We also found that the short-lived population displayed a slightly increased residence time (Figure 3A3, blue) consistent with the increased pulling forces previously reported (Gigant et al., 2017). To check on the penetrance, we repeated the experiment imaging at the spindle plane and observed, during anaphase, a faster elongation of the spindle (0.166 ± 0.056 µm/s (N = 12) compared to 0.116 ± 0.025 µm/s for the control embryos, p = 0.012) in agreement with (Bellanger et al., 2007; Maton et al., 2015), which confirmed KLP-7 penetrance. Both KLP-7 and ZYG-9 depletions suggest that the short-lived population may reveal the microtubules in a shrinking state while the long-lived population contains growing microtubules.
Secondly, and to strengthen the link between long-lived population and growing state, we analysed, with the DiLiPop assay, microtubules specifically in growing state, using an EBP-2::GFP strain. We found a single microtubule-contact population with an intermediate lifetime compared to the two population lifetimes found when labelling all microtubules using YFP::α-tubulin (Figure 3D1). Finding a single population is consistent with our hypothesis that the two dynamical behaviours of the microtubule-contacts correspond to distinct states. The intermediate lifetime could be attributed either to a direct effect of EB protein over-expression, increasing microtubule dynamics in other organisms (Duellberg et al., 2016), or to a reduction/disappearance of the GTP cap, which concentrates EB proteins (Bieling et al., 2007; Zanic et al., 2009), earlier than the switch to catastrophe, making the contact disappears prematurely (Figure 3D2) (Kozlowski et al., 2007). Indeed, some proteins like CLASP can stabilise the microtubule (Espiritu et al., 2012), while dynein can even clamp the microtubule ends preventing depolymerisation (Laan et al., 2012a). Overall, these observations suggest that the long-lived microtubules reflect microtubules growing against the cortex, which would generate a pushing force.
Altogether, these experiments focusing mostly on the spatial distribution of the microtubules support the two-population hypothesis. On the one hand, the short-lived microtubules are likely involved in pulling-force generating and their lifetime is consistent both with dynein residence time at the cortex (Rodriguez-Garcia et al., 2018) and estimates of force-generator runtimes obtained by spindle-pole oscillation modelling (Pecreaux et al., 2006a). On the other hand, the long-lived microtubules include those growing against the cortex, thought to generate the pushing forces maintaining the spindle in cell-centre during metaphase. Indeed, their residence time agrees with the estimation coming from centring stability measurements (Pécréaux et al., 2016) (Figure 3E).
The polarised short-lived-population density, reflecting dynein on-rate, accounts for the pulling force imbalance and the final spindle position
Our recent study of dynein dynamics at the embryo cortex suggested that force imbalance is caused by a higher dynein-binding rate of the cortical force generators at the posterior cortex compared to anterior one (Rodriguez-Garcia et al., 2018). Equipped with the DiLiPop assay, we aimed to go beyond and investigate how the pulling forces are regulated along time since the short-lived microtubules reveal them. We performed the analysis using 60 s time-windows and found a higher density of short-lived posteriorly from the beginning of our recording corresponding to prometaphase (Figure 4A), while the lifetime remained symmetric during metaphase and mostly during anaphase (Figure 4B). These results support that polarity is reflected by a force generator on-rate imbalance (Rodriguez-Garcia et al., 2018). Furthermore, this density ratio got more significantly pronounced during anaphase, when most of the posterior displacement of the spindle intervenes, consistent with the increase in the cortical pulling imbalance along metaphase and anaphase previously proposed at the cellular level in (Labbé et al., 2004). Interestingly, the long-lived population, despite it is not involved in pulling, became also posteriorly enriched, but only at late anaphase (Figure 4C), while its lifetime stayed close to the symmetry all the time (Figure 4D). We attributed this to the posterior displacement of the spindle (Bouvrais et al., 2018). Such a “geometrical” effect, reported to reinforce the posterior force generator engagement upon displacement of the centrosome (Bouvrais et al., 2018), would account at least for a part of the density increase in the short-lived population with mitotic progression.
We wondered to which extent it causes the asymmetric short-lived microtubule density. To address this question, we performed a time-resolved DiLiPop analysis of CSNK-1 depleted embryos. This protein acts upstream polarity mechanism mostly limiting the pulling forces generated anteriorly (Panbianco et al., 2008). We found that the long-lived population became asymmetrically distributed during late anaphase (Supplemental Figure S5B1,3), consistent with the almost normal posterior displacement of the posterior centrosome (Supplemental Figure S5C1,3), but the distribution of the short-lived population was symmetric (Supplemental Figure S5A1,3). This excludes that the asymmetrical short-lived-microtubule density could be caused only by a geometrical effect.
We set to go beyond our previous study relating the rate of dynein microtubule-binding/engaging in pulling and the force imbalance (Rodriguez-Garcia et al., 2018), and asked whether some additional mechanism could contribute to reflect polarity into force imbalance. In particular, it was proposed that microtubule dynamics to the force imbalance and the posterior displacement of the spindle (Kozlowski et al., 2007). Interestingly, EFA-6 negatively regulates the force generator, dynein, and microtubule dynamics (O’Rourke et al., 2007; O’Rourke et al., 2010). DiLiPop analysis of efa-6(RNAi) showed a modest increase in the short-lived-microtubule density leading to a delay in the appearance of an anteroposterior asymmetry and a strong increase in the long-lived-population density (Supplemental Figure S5A1,4;B1,4). Repeating the experiment in the same conditions at the spindle plane, we observed as expected a reduction in anaphase oscillations, the peak-to-peak oscillation of the posterior centrosome being equal to 2.58 ± 2.71 μm (N = 12) compared to 6.43 ± 1.89 μm for control embryos (N = 9, p = 0.001) but an exaggerated posterior displacement (Supplemental Figure S5C1,4) (O’Rourke et al., 2010). This is consistent with an increase in centring forces that also dampen the oscillations, especially on the transverse axis because the embryo oblong shape decreases these forces along the anteroposterior axis (Howard, 2006; Pecreaux et al., 2006a; Garzon-Coral et al., 2016; Pécréaux et al., 2016). In contrast, the pulling forces only mildly increase on the posterior during anaphase. This, in addition to the long-lived population being symmetric, suggests that the long-lived population opposed to the cortical pulling forces, supporting the hypothesis that only the short-lived population reflects the polarity.
To gain certainty, we analysed, through the DiLiPop assay, the two-population phenotype after depleting proteins controlling the density of the pulling force generators, as reflected by short-lived population, and upstream polarity proteins. Firstly, we depleted GPR-1/2LGN, LIN-5NuMA and Gα proteins, which are components of the trimeric force generator and its cortical anchorage module. We found that the density imbalance was abolished until mid-anaphase, and was strongly reduced thereafter upon gpr-2(ok1179);gpr-1/2(RNAi), lin-5(RNAi) and goa-1;gpa-16(RNAi) (Supplemental Figure S5A1-2, Supplemental Figure S6A,). As expected, the long-lived population showed no asymmetry under these treatments (Supplemental Figure S5B1-2, Supplemental Figure S6B), because the spindle posterior-displacement was reduced (Supplemental Figure S5C1-2, Supplemental Figure S6C). To confirm the efficiencies of these depletions, we imaged embryos in the same conditions at the spindle plane and noticed no spindle pole oscillation and a reduced spindle posterior-displacement and elongation (Supplemental Figure S5C1-2, Supplemental Figure S6C). Importantly, par-3(RNAi) and even more strongly par-2(RNAi) resulted in a reduction in the density imbalance of the short-lived-population (Supplemental Figure S7A). We confirmed the PAR-protein depletion efficiencies by imaging similarly treated embryos at the spindle plane and observed divisions closed to the symmetry for both depletions, with an increased anterior displacement after depleting PAR-3 and a reduced posterior displacement after depleting PAR-2 (Supplemental Figure S7B), as expected (Grill et al., 2001). We concluded that the PAR proteins related polarity is likely and only reflected by the short-lived-population density, necessary and sufficient to translate polarity into force imbalance.
To strengthen this hypothesis, we asked whether the final position of the spindle, measured by imaging at the spindle plane and tracking the centrosomes, correlated with the asymmetric density of the short-lived population, obtained from DiLiPop analysis after imaging at the cortex plane, in control and genetically perturbed embryos. We plotted the final spindle position along the AP axis (Sx (Pécréaux et al., 2016)) versus the anterior (0-45% of AP axis) to posterior (70-100% of AP axis) density ratio of the two populations during anaphase. We obtained a more pronounced correlation using the short-lived-microtubule density-ratio compared to the long-lived-population one (Figure 4EF). Interestingly, the spindle centred position, Sx = 50%, corresponded – through a linear fit estimate of density ratio – to a ratio equal to 1.035 for the short-lived population (Figure 4E, dotted line), suggesting quasi-equal pulling from anterior and posterior cortices. We concluded that the pulling force imbalance is recapitulated by the asymmetric density ratio of the short-lived population. In turn, this density corresponds to the binding rate of force generator dynein to the microtubule or to the initiation of a molecular motor run to exert a pulling force (globally termed on-rate (Rodriguez-Garcia et al., 2018)). By combining cell-scale and single-component measurements, we can overall conclude that polarity is translated into a force imbalance solely by this asymmetric on-rate.
The short-lived-microtubule lifetime reflects the mitotic progression control through the force generator processivity
A temporal control of the cortical pulling forces was previously reported (McCarthy Campbell et al., 2009) and, based on cell-scale measurements during anaphase, we suggested that it acted through the force generator processivity, reflected into the short-lived-microtubule lifetime (Pecreaux et al., 2006a; Bouvrais et al., 2018). We sought a direct proof of such a regulation and studied the temporal evolution of the two-population residence times using a time resolution window of 30 s, which still ensured accurate estimates. We found a steep increase in the short-lived-microtubule lifetime during the early anaphase, continued by a shallower one in late anaphase, in contrast to a constant lifetime during metaphase (Figure 5A). This observation, consistent with an increase in force generator processivity during anaphase, brings for the first time a direct observation of this at the single-microtubule scale. Furthermore, it raised the question of whether this corresponds to a force control mean or to a bare consequence of some other process. Interestingly, we observed that the long-lived-microtubule lifetime undergoes the same time evolution but with a reduced amplitude variation, (Figure 5B). This suggests that some distinct processes might regulate each microtubule population. The short-lived-population lifetime would increase beyond the consequence of the bare increased microtubule-nucleation rate (Srayko et al., 2005) or of the posterior displacement, which makes more microtubules available at the cortex (Bouvrais et al., 2018).
The polymerising microtubules, viewed in the long-lived population, contribute to maintaining the spindle in the cell-centre
We recently proposed that microtubule pushing against the cortex contributes to maintaining the spindle in the cell centre during metaphase (Garzon-Coral et al., 2016; Pécréaux et al., 2016). However, other mechanisms were proposed, involving pulling from either the cortex or the cytoplasm (Grill and Hyman, 2005; Kimura and Kimura, 2011a; Kimura and Kimura, 2011b; Shinar et al., 2011; McNally, 2013; Barbosa et al., 2017). We set to investigate how the long-lived population of microtubules varied upon various genetic perturbations of microtubule dynamics, to discriminate between those hypotheses. We used the stability of the metaphasic spindle in the cell centre as a readout of the centring (Pécréaux et al., 2016): this is recapitulated into the diffusion coefficient DSy of the spindle along the transverse axis. The larger this value, the poorer the stability. This value was computed from imaging at the spindle plane. We found a strong anti-correlation between the diffusion coefficient and the density of long-lived microtubules (Figure 6A) but not with the short-lived-population density (Supplemental Figure S8A). These observations suggest that microtubule pushing rather than pulling contributes to maintaining the spindle in the cell centre. Furthermore, because the microtubule density at the cortex impacted the centring, these results are not consistent with the cytoplasmic pulling hypothesis. We also found an anti-correlation between the centring stability and the long-lived-population lifetime (Figure 6B) as expected, since longer pushing improves the centring (Howard, 2006; Pécréaux et al., 2016). Interestingly, the short-lived-microtubule lifetime anti-correlated also with the centring stability (Supplemental Figure S8B), in contrast to the density, consistent with the stochastic attachment and detachment of the cortical pulling force generators creating noise that perturbs the centring: higher dynamics of force generators leading to poorer stability (Pécréaux et al., 2016). Through direct observation of pushing force events, we conclude that the spindle is likely maintained at cell centre during metaphase by microtubule pushing against the cortex, while the pulling events add some biological noise.
DISCUSSION
Through an advanced and careful analysis of microtubule-contact dynamics at the cortex of C. elegans embryos, based on high frame-rate imaging, we could for the first time resolve two microtubule populations, distinct by their residence time. Measured as 0.4 s and 1.8 s, these lifetimes are rather short with respect to previously published values, which range between 1 s and 15 s (Labbé et al., 2003; Kozlowski et al., 2007; Gerhold et al., 2018; Sugioka and Bowerman, 2018). These differences could be attributed to different microtubule labellings, growing microtubule plus-end or whole microtubule, mitotic phase considered, or various sample-mounting techniques. Noticeably, approaches with higher frame rate, consistent with microtubule growth and shrink rates, provide smaller residence times, close to the values found here. Furthermore, and key to DiLiPop out-performing existing methods, the imaging and tracking ensure the capture of a much larger number of microtubule contacts, ensuring strong representative statistics. Finally, the DiLiPop advanced and tailor-made statistical analysis enables to disentangle mixed dynamics in a set of microtubule-contact durations.
Towards a multi-scale approach by combining the DiLiPop assay with cell-scale investigations
One decisive trait of our approach is the combining of measurements at cell scale and single-microtubule scale (microscopic). In particular, it enables us to understand the spatial and temporal regulation of the cortical pulling forces as well as support our hypothesis of spindle maintenance in the cell centre by microtubule pushing. Firstly, and because short- and long-lived microtubules reflect pulling and pushing force generating events, respectively, our assay enables for the first time to delineate the concomitant time and space evolution of these forces at the microscopic level, paving the way to understand how it accounts for the complex choreography of the mitotic spindle. In particular, we confirmed that the asymmetry in pulling forces – visible here through the density of the short-lived population – was encoded as an imbalance in dynein binding rate rather than in dynein detachment rate, as proposed in (Rodriguez-Garcia et al., 2018). Secondly, and superseding our study focusing on dynein, we suggest that no other mechanism downstream of polarity proteins PAR-2/-3 and CSNK-1 contributes significantly to the pulling force imbalance. The time evolution of the short-lived-microtubule densities suggests that their asymmetric distribution, thus the force imbalance, is present from early prometaphase, as previously inferred from cell-scale measurements (Labbé et al., 2004). Thirdly, the pulling force amplitude – related to the force-generator persistence pulling (processivity) based on modelling cell-scale centrosome kinematics (Pecreaux et al., 2006a) – viewed here through the cortical residence time of the short-lived microtubule, stays in contrast at constant value until anaphase onset, when it starts increasing with time, while the spindle migrates posteriorly and elongates due to the cortical pulling forces. Fourthly, observing that the spindle stability in the cell-centre during metaphase correlates with the density of pushing microtubules at the cortex, reflected by the long-lived population, we suggest that these pushing microtubules contribute to the centring-maintenance mechanism.
The short-lived population might not account for only pulling events
Measuring the displacement of centrosome fragments after laser disintegration suggests that twice more force generators are active at the posterior cortex (Grill et al., 2003) while the anterior-to-posterior density-ratio of the short-lived microtubules is above 0.5. It suggests that some of these short-lived contacts might represent microtubules not involved in pulling events. Interestingly, analysing the dynein dynamics at the cortex led to a similar ratio (Rodriguez-Garcia et al., 2018). It could correspond to stalled microtubule-ends/dyneins. Indeed, in vitro and in vivo studies showed that anchored dynein could serve as microtubule plus-end tether (Dujardin and Vallee, 2002; Hendricks et al., 2012; Laan et al., 2012b; Perlson et al., 2013; Yogev et al., 2017). Consistently, the number of short-lived microtubules contacting the cortex is a bit large compared to the expected values of 10-100 per cortex half (Grill et al., 2003; Redemann et al., 2010). Indeed, we measured, during late anaphase, about 40 short-lived microtubules contacting the visible cortex each second, extrapolating to about 120 per half cortex. A fraction of these could be stalled dynein tethering the microtubule ends. This can be the signature of a mechanism regulating dynein run initiation from a state where it captures the microtubule (Laan et al., 2012a; Jha et al., 2017).
Microtubules pushing against the cortex likely perform the centring maintenance
The spindle was proposed to be maintained in the cell centre by three families of mechanisms: pushing forces generated by microtubule growth against the cortex, similarly to what was found in yeast e.g.; the pulling on astral microtubules from the cortex, carefully regulated in space; and forces generated within the cytoplasm by force generators attached on small organelles (Reinsch and Gonczy, 1998; Grill and Hyman, 2005; Kimura and Onami, 2005; Wühr et al., 2009; Zhu et al., 2010; Kimura and Kimura, 2011b; Kimura and Kimura, 2011a; Shinar et al., 2011; Laan et al., 2012a; McNally, 2013; Garzon-Coral et al., 2016; Pécréaux et al., 2016; Barbosa et al., 2017; Wu et al., 2017). Because the dynamics of microtubules at the cortex correlate with the centring stability, our finding is poorly consistent with cytoplasmic pulling. Furthermore, during metaphase, the density of long-lived – but not the short-lived – microtubules anti-correlates with the centring stability, consistently with centring by pushing. Interestingly, the short-lived-population density was low during that period, indicating that it likely does not contribute to centring. As previously suggested, pulling forces revealed by short-lived population contribute to de-centring and are down regulated during metaphase to let the spindle maintain in cell centre (Dogterom et al., 2005; Grill and Hyman, 2005; Zhu et al., 2010; Pécréaux et al., 2016). Noticeably, we found no polarisation of the long-lived population, beyond the density imbalance due to the spindle displacement in late anaphase. Out of the scope of this work, our multi-scale approach could also be used to decipher centring establishment prior to mitosis (Ahringer, 2003) and the orientation of the pronuclei-centrosome complex (PCC), for which the mechanisms are still debated (Tsou et al., 2003b; Labbé et al., 2004; Kimura and Onami, 2005, 2007; Park and Rose, 2008; Gusnowski and Srayko, 2011; Kimura and Kimura, 2011a; Kimura and Kimura, 2011b; Barbosa et al., 2017).
The short-lived-microtubule density reflects positional regulation while its lifetime the mitotic progression
We recently proposed that the cortical pulling forces are regulated both by the position of the centrosome, which in turn controls the amount of microtubules that can be captured by the cortical force generators on the one hand, and on the other hand, by the processivity of the force generators increasing along time (Pecreaux et al., 2006a; Bouvrais et al., 2018). Our assay here offers the unique opportunity to observe separately these two, as the former controls the density of pulling microtubules and the later their lifetime. Indeed, the time evolutions of the density of short-lived and long-lived microtubules are quite alike and thus suggest a global control by the position. The cell cycle control of the number of nucleated microtubules, known to increase at anaphase (Srayko et al., 2005), will further reinforce this density increase at that moment. Noticeably, we found both long-lived and short-lived microtubules at the cortex along the whole mitosis, including when the short-lived ones are limiting the number of engaged force generators since force-generator to microtubule attachment corresponds to a pseudo-chemical reaction as modelled in (Bouvrais et al., 2018). This is because the force generators or their anchor sites are in limited number and thus sparse at the cortex (Grill et al., 2003; Grill et al., 2005; Redemann et al., 2010; Rodriguez-Garcia et al., 2018). Consequently, not all microtubules can find a force generator to pull on (Bouvrais et al., 2018). In contrast, the lifetime evolutions along metaphase and anaphase are different for short- and long-lived microtubules, suggesting that a specific mechanism regulates the short-lived-microtubule lifetime as previously suggested based on modelling at cell scale (Pecreaux et al., 2006a; Bouvrais et al., 2018).
In contrast to previous studies measuring the resulting forces on the centrosome along time (Grill et al., 2003; Labbé et al., 2004), the DiLiPop enables to separate the contributions of pulling and pushing over time, in a physiological context. While time evolutions of the densities of the both populations, reflecting both forces, are similar excluding a different mechanism of regulation, the short-lived-microtubule lifetime undergoes a more pronounced increase in anaphase compared to long-lived one. It suggests that this controls the transition between metaphase – with dominant pushing forces maintaining the spindle in the cell centre – and anaphase when pulling dominates to displace the spindle posteriorly. While the pulling increases at the latter stage, the pushing, thus the centring remains mostly constant or even slightly increases. In contrast, the pulling force plateau (constant residence time) during metaphase ensures the slow posterior displacement but lets the centring forces dominate (Ahringer, 2003; Garzon-Coral et al., 2016; Pécréaux et al., 2016). This regulation through intensifying the pulling/displacement forces contrasts with recent finding in sea urchin zygote where a reduction of the centring forces accounts for the decentration after the maintenance in cell-centre (Sallé et al., 2018).
Which molecular mechanisms could regulate these forces?
Three forces and regulations are involved in controlling the spindle positioning: a spatial asymmetry of cortical pulling, reflecting polarity; a time increase of the same, corresponding to mitotic progression; and a constant centring. Firstly, the limited amount of force generator regulators, GPR-1/2, or anchor sites e.g. Gα proteins GOA-1/GPA-16, more abundant on the posterior side, is the most likely to account for the asymmetric density (Grill et al., 2003; Pecreaux et al., 2006a; Wu and Rose, 2007; Park and Rose, 2008; Riche et al., 2013; Rodriguez-Garcia et al., 2018). Restricting the pulling force-generators to a posterior-most crescent through LET-99 reinforces this asymmetry, and consistently we found a density of short-lived microtubules in the 45-70% similar to the anterior side (Figure 4A) (Krueger et al., 2010; Bouvrais et al., 2018). Recently, the APR-1/APC complex was suggested to decrease the cortical forces anteriorly (Sugioka et al., 2018). Since we do not see an increase in growing microtubules anteriorly, the balance between stalled (for a short while) and shrinking microtubules, all reflected by the short-lived population, could be different between anterior and posterior, with a larger proportion of stalled microtubules anteriorly. Finally, the denser actin-myosin cortex at the anterior cortex could play a similar regulatory role (Goulding et al., 2007; Colin et al., 2018). Secondly, the pulling forces are regulated along the mitosis as discussed above. The first and obvious possible player is dynactin, known to be essential to cortical pulling forces and able to make dynein more processive (Skop, 1998; Gonczy et al., 1999; King and Schroer, 2000; Rodriguez-Garcia et al., 2018). Beyond, some other players were involved in other systems in vitro or in vivo, for example: the end-binding (EB) proteins putatively involved in initiating dynein pulling once a microtubule is captured (Jha et al., 2017), and whose homolog EBP-2 contributes, although modestly, to cortical forces (Rodriguez-Garcia et al., 2018); LIS-1 essential for spindle positioning in nematode and reported to be also implicated in dynein regulation beyond its classic inhibitory role (Cockell et al., 2004; Baumbach et al., 2017; DeSantis et al., 2017); BICD-1BICD2 involved in nuclear migration in nematode hypodermis (Fridolfsson et al., 2010) but no strong early embryonic phenotype was reported, despite indication of its role in other organisms (Swan et al., 1999; Splinter et al., 2012; Jha et al., 2017; Urnavicius et al., 2018). Thirdly and finally, efficient pushing by microtubules against the cortex requires some microtubule-associated proteins to prevent the switch to catastrophe (Janson et al., 2003). Viewing of microtubule buckling in nematode suggests that microtubules are able to efficient pushing (Kozlowski et al., 2007) although the involved proteins are still not clearly known. One possibility is CLS-2CLASP reported to affect spindle positioning (Espiritu et al., 2012). Indeed, the inability to separate the measurements of pulling and pushing forces limited the ability to delineate the involved players and this work may help to work around this issue.
Beyond these findings, this work is also a rare example of combining investigations at two scales to decipher some key mechanisms regulating mitosis, especially the forces responsible for positioning the mitotic spindle (Figure 7). It offers the unparalleled ability to view the pulling and pushing force-generating events directly, through the short-lived and long-lived microtubule contacts, and to explore the regulation of these forces in time and space. We expect that this novel approach will find applications beyond that work and that the DiLiPop tool will be used in another context, in which an experimental distribution has to be fitted, for instance in photon count experiments.
MATERIALS AND METHODS
Culturing C. elegans
C. elegans nematodes were cultured as described in (Brenner, 1974), and dissected to obtain embryos. The strains were maintained at 25°C and imaged at 23°C. The strains were handled on nematode medium plates and fed with OP50 bacteria.
Strains of C. elegans and C. briggsae used
C. elegans TH65 YFP::TBA-2 (α-tubulin) strain (Srayko et al., 2005) having a fluorescent labelling of the whole microtubule (MT) was used for the DiLiPop assay as well as C. elegans AZ-240 GFP::TBB (β-tubulin) strain (Praitis et al., 2001) and C. briggsae ANA020 GFP::TBB (β-tubulin) strain. TH27 TBG-1::GFP (γ-tubulin) strain (Oegema et al., 2001) displaying a centrosomal fluorescent labelling was the standard for the “centrosome-tracking” assay. TH66 EBP-2::GFP strain (Srayko et al., 2005) that displays a labelling of microtubule plus-ends was used for comparison of its effects on microtubule dynamics. The JEP-18 gpr-2(ok1179) strain was used to target GPR-1/2 protein though a mutation.
Gene inactivation through protein depletion by RNAi feeding
RNA interference (RNAi) experiments were performed by feeding using the Ahringer-Source BioScience library (Kamath and Ahringer, 2003), except for GOA-1;GPA-16 depletion, whose clone was kindly given by Prof P. Gönczy. The feedings were performed at 25°C for various durations according to the experimental goals. The treatment lasted 24h for lin-5, goa-1;gpa-16, efa-6 and klp-7 genes. When we aimed for stronger phenotypes (e.g. symmetric divisions), we used duration of 48h (cls-2, par-2, par-3 and gpr-1/2). The duration was reduced to 4h and 6-10h when targeting zyg-9 and cnsk-1, respectively. The control embryos for the RNAi experiments were fed with bacteria carrying the empty plasmid L4440. We did not notice any phenotype suggesting that the meiosis was impaired during these various treatments.
Preparation of the embryos for imaging
Embryos were dissected in M9 buffer and mounted on a pad (2% w/v agarose, 0.6% w/v NaCl, 4% w/v sucrose) between a slide and a coverslip. Depending on the assay (landing or centrosome-tracking ones), embryos were observed using different microscopic setups. To confirm the absence of phototoxicity and photodamage, we checked for normal rates of subsequent divisions (D.L. Riddle, 1997; Tinevez, 2012). Fluorescent lines were imaged at 23°C.
Imaging of microtubule contacts at the cortex
We imaged C. elegans one-cell embryos at the cortex plane in contact with the glass slide, viewing from the nuclear envelope breakdown (NEBD) until late anaphase. We used a Leica DMi8 spinning disk microscope with Adaptive Focus Control (AFC) and a HCX Plan Apo 100x/1.4 NA oil objective. Illumination was performed using a laser with emission wavelength of 488 nm and we used GFP/FITC 4 nm band pass excitation filter and a Quad Dichroic emission filter. To account for the fast microtubule dynamics at the cortex, images were acquired at an exposure time of 100 ms (10 Hz) using an ultra-sensitive Roper Evolve EMCCD camera that was controlled by the Combox. During the experiments, the embryos were kept at 23°C. To image embryos at the cortex, we typically moved the focus to 12 to 15 µm below the spindle plane. Images were then stored using Omero software (Li et al., 2016).
Spindle pole imaging
Embryos were observed at the midplane using a Zeiss Axio Imager upright microscope modified for long-term time-lapse. First, an extra anti-heat filter was added to the mercury lamp light path. Secondly, to decrease the bleaching and obtain optimal excitation, we used an enhanced transmission 12 nm band pass excitation filter centred on 485 nm (AHF analysentechnik). We used a 100x/1.45 NA Oil plan-Apo objective. Images were acquired with an Andor iXon3 EMCCD 512×512 camera at 33 frames per second and using their Solis software. Images were then stored using Omero software (Li et al., 2016).
Centrosome-tracking assay
The tracking of labelled centrosomes and analysis of trajectories were performed by a custom tracking software (Pecreaux et al., 2006a) and developed using Matlab (The MathWorks). Tracking of −20°C methanol-fixed γ-tubulin labelled embryos indicated accuracy to 10 nm. Embryo orientations and centres were obtained by cross-correlation of embryo background cytoplasmic fluorescence with artificial binary images mimicking embryos, or by contour detection of the cytoplasmic membrane using background fluorescence of cytoplasmic TBG-1::GFP with the help of an active contour algorithm (Pecreaux et al., 2006b). The results were averaged over all of the replicas for each condition.
Simulation of microscopy images
To validate the image processing pipeline, we fabricated fluorescence images of known dynamics, which mimic our cortical images using the algorithm developed by (Costantino et al., 2005) that we adapted to our needs. In further details, we simulated stochastic trajectories of particles (Figure 2E1) that displayed a limited random motion: xi,j(t+1) = xi,j(t) + ξ (2Dt)1/2, where xi,j(t) represent the coordinates in two dimensions at time t, ξ is a random number and D is the diffusion coefficient. The duration of the tracks (i.e. their length) was sampled from an exponential distribution. The intensity was set similar to experimental ones and encoded by the quantum yield parameter (Qyield). We then plotted the instantaneous positions and applied a Gaussian filter to mimic the effect of the point-spread function (PSF) in fluorescence microscopy. Then, we mimicked the background noise by adding at each pixel a sampling of a Gaussian distribution normalized to ε, with formula reading Anoisy = A + εM and corresponding to a SNR max(A)/ε (Figure 2E2). This simulation provided a realistic scenario to test the image-processing pipeline. Details of the parameters used for simulation can be found in Table S3.
Statistics
For classic statistical analyses (excluding model parameter comparison), averaged values of two conditions were compared using the two-tailed Student’s t-test with correction for unequal variance except where otherwise stated. For the sake of simplicity, we recorded confidence levels using stars (*, p ≤ 0.01; **, p ≤ 0.001; ***, p ≤ 0.0001; ****, p ≤ 0.00001) and n.s. (non-significant, p > 0.01; sometimes omitted to save room). We abbreviated standard deviation by SD, standard error by s.e., and standard error of the mean by s.e.m.
Data and image processing
All data analysis was developed using Matlab (The MathWorks).
ACKNOWLEDGMENTS
The bacterial clone of GPA-16;GOA-1 was a kind gift from Prof P. Gönczy. We thank Dr. Gregoire Michaux for the feeding clone library and technical support. We also thank Drs. Anne Pacquelet, Giulia Bertolin, Xavier Pinson, Grégoire Michaux, Sébastien Huet and Marc Tramier for discussions about the project. Some strains were provided by the Caenorhabditis Genetics Center (CGC), which is funded by National Institutes of Health Office of Research Infrastructure Programs (P40 OD010440; University of Minnesota). J.P. was supported by a Centre National de la Recherche Scientifique (CNRS) ATIP starting grant and La Ligue nationale contre le cancer. We also acknowledge Plan Cancer grant BIO2013-02, COST EU action BM1408 (GENiE) and La Ligue contre le cancer (comités d’Ille-et-Vilaine et du Maine-et-Loire). Microscopy imaging was performed at the Microscopy Rennes Imaging Center, UMS 3480 CNRS/US 18 INSERM/University of Rennes 1. Spinning disk microscope was co-funded by the CNRS, Rennes Métropole and Région Bretagne (AniDyn-MTgrant). DF’s postdoctoral fellowship was funded by Région Bretagne (pRISM grant). HB’s postdoctoral fellowship was funded by the European Molecular Biology Organization. TP was supported by the France-BioImaging infrastructure (ANR-10-INBS-04).
Footnotes
Email: helene.bouvrais{at}univ-rennes1.fr