While examining the function of AIR-1 in spindle positioning (Kotak et al., 2016), as noted also previously (Noatynska et al., 2010; Schumacher et al., 1998), we observed that the majority of zygotes in which AIR-1 had been depleted by RNAi exhibited a remarkable polarity phenotype, with two distinct PAR-2 domains, one at each pole (see below). We set out to investigate this phenotype in depth and thus utilize it as a handle to decipher fundamental principles of symmetry breaking and A-P polarization in C. elegans.

We conducted time-lapse microscopy of embryos simultaneously expressing endogenously tagged fluorescent proteins marking NMY-2 to monitor the cortical acto-myosin network, SAS-7 to track centrosomes and PAR-2 to assay posterior polarization. As anticipated, by the time of pronuclear meeting, all control embryos established a single PAR-2 domain on the side of the centrosomes and the male pronucleus (Figure 1A and C, Video 1). In stark contrast,~76% air-1(RNAi) embryos formed two PAR-2 domains, one at each pole (Figure 1B and C, Video 2, referred to as ‘bipolar’). Another ~19% air-1(RNAi) embryos established a single PAR-2 domain, usually on the side of the maternal pronucleus (Figure 1C, referred to as ‘anterior’) or else on that of the male pronucleus (Figure 1C, referred to as ‘posterior’). Moreover,~5% air-1(RNAi) embryos harbored no plasma membrane PAR-2 (Figure 1C, referred to as ‘none’). Analogous distributions were found for endogenous PAR-2 by immuno labeling wild type embryos subjected to RNAi-mediated depletion of AIR-1 (Figure 1D; Figure 1—figure supplement 1A). Note that we never detected an anterior PAR-2 domain either early or later in the cell cycle in the endogenously tagged GFP::PAR-2 strain used in this study, in contrast to the situation in strains overexpressing GFP::PAR-2 (Figure 1—figure supplement 2). We noted also that progressive dilutions of air-1(RNAi) resulted in weaker spindle assembly defects and shifted the polarity phenotype away from harboring two poles (Figure 1—figure supplement 1B,C). Therefore, bipolarity is the most severe AIR-1 depletion phenotype obtained using RNAi, likely corresponding to a strong if not complete loss of function in the embryo. Note that null alleles of air-1 result in sterility (Furuta et al., 2002), precluding their analysis in the embryo. Overall, we conclude that AIR-1 ensures the establishment of a unique posterior PAR-2 region in the C. elegans zygote.

Figure 1 with 2 supplements see all

Download asset Open asset

Video 1

Download asset

Video 2

Download asset

We set out to analyze further features of PAR-2 domain formation in air-1(RNAi) embryos. We found that the fluorescence intensity of each of the two PAR-2 domains was weaker than that of the single PAR-2 domain in control embryos (Figure 1E,F; Figure 1—figure supplement 1F). Likewise, each of the two PAR-2 domains was smaller (Figure 1—figure supplement 1E). However, the summed fluorescence intensities and sizes of both PAR-2 domains did not differ significantly from that of the single PAR-2 domain in control embryos (Figure 1—figure supplement 1E,F). To investigate whether air-1(RNAi) embryos underwent bona fide bipolarization, and not merely aberrant distribution of PAR-2, we examined the distribution of other PAR proteins. Both PAR-6 and PKC-3 localized to the anterior pole in all control embryos but were restricted to the middle region in ~80% air-1(RNAi) embryos, as anticipated from the bipolar distribution of PAR-2 (Figure 1G–J). Furthermore, we found that the cytoplasmic polarity mediator protein MEX-5, which was enriched in the anterior of control embryos in response to proper A-P polarity (Schubert et al., 2000), was present primarily in the middle region of ~80% air-1(RNAi) embryos (Figure 1K,L). These findings indicate that the machinery that ensures polarization operates normally in air-1(RNAi) embryos, but that its spatial regulation is impaired.

To determine whether AIR-1 kinase activity is required for correct polarization, we analyzed embryos expressing GFP fusion proteins to either a wild type or a kinase inactive version of RNAi-resistant AIR-1 (Toya et al., 2011). We found that whereas wild type GFP::AIR-1 rescued posterior PAR-2 polarity upon depletion of endogenous AIR-1 in ~63% of embryos, the kinase inactive version did not, with all embryos being bipolar (Figure 1M,N). The fact that all embryos were bipolar in this case, compared to ~76% in plain air-1(RNAi) embryos, likely reflects the partial functional impairment that is also observed in the wild type construct (see Figure 1N), which may result from the presence of mCherry or else from altered expression levels. Overall, we conclude that AIR-1 kinase activity is required to ensure a unique symmetry breaking event in C. elegans zygotes. Whereas the identification of relevant AIR-1 substrates will be of interest for future investigations, hereafter we utilize this remarkable bipolar phenotype to decipher fundamental principles underlying symmetry breaking at the onset of A-P axis specification in C. elegans.

Since polarity establishment normally relies in a partially redundant manner on local RHO-1 inactivation mediated by cortical flows and on the PAR-2-dependent pathway (Motegi et al., 2011), we investigated these mechanisms in air-1(RNAi) embryos to assess whether their alteration could be responsible for imparting bipolarity. Examining cortical flows, we found that control embryos exhibited the expected anterior-directed movements of NMY-2 emanating from the centrosomes-bearing region, followed by the appearance of a posterior PAR-2 domain (Figure 2A). By contrast, embryos depleted of AIR-1 exhibited cortical flows stemming from both poles and directed toward the embryo center, followed by PAR-2 domain establishment at both poles (Figure 2B). Quantification of cortical flows using particle image velocimetry (PIV) of RFP::NMY-2 revealed that the average peak velocities were ~6.8 µm/min in control embryos, but ~3.1 µm/min for posterior- and ~2.9 µm/min for anterior-directed flows, respectively, in air-1(RNAi) embryos (Figure 2B,D). We conclude that AIR-1 is somehow required for robust cortical flows.

Figure 2 with 1 supplement see all

Download asset Open asset

We set out to test whether the decrease in cortical flow velocity in air-1(RNAi) embryos contributes to the bipolar phenotype. We reasoned that if this were the case, then increasing flow velocities in air-1(RNAi) embryos should rescue the phenotype. To this end, we set out to co-deplete AIR-1 and the RHO-1 GTPase activating protein (GAP) RGA-3 (Schonegg et al., 2007). On its own, RGA-3 depletion led to stronger cortical flows and to an enlarged posterior PAR-2 domain (Figure 2—figure supplement 1A,E) (Schonegg et al., 2007). We found that the average peak velocities of the cortical flows in doubly depleted air-1(RNAi) rga-3(RNAi) embryos were close to those of control embryos (Figure 2—figure supplement 1B). Despite this, the fraction of air-1(RNAi) rga-3(RNAi) embryos exhibiting a bipolar phenotype was not altered (Figure 2E,F). Therefore, while cortical flows are diminished upon AIR-1 depletion, this alone does not explain the bipolar phenotype. To investigate whether cortical acto-myosin flows present in air-1(RNAi) embryos are required for bipolarity, we set out to utilize nmy-2(ne3409) embryos. This temperature-sensitive mutant allele of NMY-2 exhibits severely reduced cortical flows (Liu et al., 2010), to an extent similar to those observed on either side of air-1(RNAi) embryos (Figure 2—figure supplement 1C). We found that nmy-2(ne3409) embryos did not exhibit bipolarity and instead harbored a single posterior PAR-2 domain, which formed later and was smaller than in the control (Figure 2G,H; Figure 2—figure supplement 1E), in line with prior work (Liu et al., 2010). Depletion of AIR-1 in nmy-2(ne3409) embryos almost completely abolished cortical flows (Figure 2—figure supplement 1C), but small bipolar PAR-2 domains formed nevertheless in most embryos, although with a substantial delay (Figure 2G,H; Figure 2—figure supplement 1F). Together, these findings indicate that cortical acto-myosin flows present in air-1(RNAi) embryos are not essential for bipolarity, but contribute to the robustness and extent of the two PAR-2 domains.

The above observations prompted us to test whether the bipolar phenotype relies on the PAR-2-dependent pathway (Motegi et al., 2011). To this end, we used a strain that expresses an RNAi-resistant GFP-tagged version of PAR-2, either in the wild type (WT) form or in mutant forms deficient for PAR-2 binding to microtubules (PAR-2 R163A or PAR-2 R183-5A) and thus lacking the second polarity pathway. As reported previously (Motegi et al., 2011), whereas all three transgenes rescued depletion of endogenous PAR-2 to almost 100%, diminishing cortical flows by co-depletion of ECT-2 abolished loading of the two PAR-2 mutants onto the plasma membrane (Figure 2I–K; Figure 2—figure supplement 1G–J). Strikingly, we found the same to be true in these mutants when AIR-1 was co-depleted with endogenous PAR-2 (Figure 2I–K; Figure 2—figure supplement 1G–J). Moreover, whereas cortical flows were normal when PAR-2 alone was depleted (Munro et al., 2004) (Figure 2—figure supplement 1K,M,N), we found that they were absent when AIR-1 was depleted concomitantly (Figure 2—figure supplement 1L,M,O), reflecting the feedback loop of membrane-bound PAR-2 onto cortical flows (Gross et al., 2019). Taken together, these results demonstrate that the bipolarity of air-1(RNAi) embryos relies on the PAR-2-dependent pathway.

As PAR-2 domains developed invariably at the poles of air-1(RNAi) embryos, we reasoned that curvature might be a contributing cue for establishing PAR-2 domains. To test this hypothesis, we used PDMS to microfabricate equilateral triangular chambers ~ 40 µm in side length and ~25 µm in depth. We reasoned that PAR-2 domains should form preferentially in the corners of such triangular geometries upon AIR-1 depletion if curvature is indeed a contributing factor. In this set of experiments, embryos were squeezed into the triangular chambers prior to symmetry breaking and filming. Despite being deformed, control embryos established a single PAR-2 domain in the vicinity of centrioles, which in ~50% of cases were along one side of the triangle and in the remaining ~50% in one of the corners (Figure 3A,B, Video 3). Approximately 20% of air-1(RNAi) embryos formed two PAR-2 domains, with the remaining ~80% forming one domain (Figure 3A,B, Videos 4 and 5). Remarkably, with one exception, PAR-2 domains in air-1(RNAi) embryos formed within a corner of the triangle (Figure 3C), usually in the corner where the embryo exhibited the highest degree of curvature before symmetry breaking (Figure 3D, Figure 3—figure supplement 1A). Moreover, as shown in Figure 3E, PAR-2 domains in air-1(RNAi) embryos formed independently of the position of centrosomes or polar bodies, in contrast to control embryos, in which PAR-2 domain formation coincided simply with centrosome position. Together, these experiments indicate that embryos depleted of AIR-1 recruit PAR-2 preferentially to curved regions.

Figure 3 with 1 supplement see all

Download asset Open asset

Video 3

Download asset

Video 4

Download asset

Video 5

Download asset

We investigated whether the preference of PAR-2 for curved regions stems from an accumulation of microtubules in these locations, which could conceivably enhance PAR-2 recruitment. However, immunofluorescence analysis did not reveal an accumulation of microtubules at the poles of air-1(RNAi) embryos (Figure 3—figure supplement 1B,C). Likewise, analysis of embryos expressing YFP::TBB-2 and placed vertically into cylindrical PDMS wells for end-on time lapse microscopy confirmed that there was no difference in microtubules distribution at poles when comparing control and air-1(RNAi) embryos (Figure 3—figure supplement 1D,E). Nevertheless, we tested whether microtubules might mediate PAR-2 recruitment in air-1(RNAi) embryos. We depleted microtubules shortly after fertilization using nocodazole, which resulted in a failure of female pronuclear migration (Figure 3—figure supplement 1F, arrow). As previously reported (Hayashi et al., 2012; Sonneville and Gönczy, 2004; Tsai and Ahringer, 2007), removing microtubules from control embryos did not prevent posterior PAR-2 domain establishment (Figure 3—figure supplement 1F,G). Importantly, nocodazole-treated air-1(RNAi) embryos recruited PAR-2 to both poles (Figure 3—figure supplement 1F,G). We found an analogous result in embryos lacking microtubules following RNAi-mediated depletion of the α-tubulin TBA-2 (Figure 3—figure supplement 1H). Overall, we conclude that the preference of PAR-2 for the polar regions of air-1(RNAi) embryos is not due to microtubule enrichment in those areas.

We developed a physical description of curvature-dependent symmetry breaking and PAR-dependent axis specification to better understand how the C. elegans zygote undergoes bipolarization when AIR-1 is lacking (Figure 4A, Appendix). Our 1D dynamical system is based on known interactions between anterior and posterior PAR proteins, which were also considered in a previous theoretical analysis (Goehring et al., 2011). In our system, the concentrations $A$ of membrane-bound anterior (PAR-6) and $P$ of posterior (PAR-2) proteins follow the dynamic equations

Figure 4 with 1 supplement see all

Download asset Open asset

where $D_A$ and $D_P$ denote the respective diffusion coefficients of the two species, $v$ the velocity of the cortical flow, $k_{on}$ and $k_{off}$ the respective attachment and detachment rates, and $A_{cyto}$ and $P_{cyto}$ their cytoplasmic concentrations. The terms $k_{AP}PA$ and $k_{PA}{A}^{2}P$ account for the mutual antagonism between anterior and posterior PAR proteins. In previous work, symmetry breaking was achieved by imposing a cortical flow from the posterior to the anterior side, which was assumed to be induced by centrosomes via an unspecified mechanism (Goehring et al., 2011). Here, we use instead active gel theory to include a physical mechanism that causes cortical flows (Kruse et al., 2004). In this approach, cortical flows are caused by gradients in the mechanical stress that is generated in the actin cortex notably through non-muscle myosin activity (Callan-Jones et al., 2016; Callan-Jones and Voituriez, 2013; Gross et al., 2019; Hannezo et al., 2015), following the force-balance relation

In this equation, $\eta$ denotes the viscosity of the cortex and $\gamma$ its friction against the membrane. ${\Pi }_{active}$ accounts for the active stress, leading to contraction using molecular motors, and ${\Pi }_{passive}$ for the passive stress originating from osmotic pressure. It has been shown that dissociation rates of cortical NMY-2 are regulated by PAR proteins, with lower dissociation rates occurring in the PAR-6-containing domain than in the PAR-2-containing domain (Gross et al., 2019). To account for this feedback loop, we assume that the active stress ${\Pi }_{active}$ effectively decreases with the density of membrane-bound PAR-2 $P$ as $\beta P$ via antagonism with PAR-6. Furthermore, we account for the preferential binding of PAR-2 to curved regions without assigning a specific molecular mechanism. This feature is incorporated by increasing the rate $k_{on,P}$ of PAR-2 binding at the poles by an amount $\alpha$ multiplied by the excess curvature compared to the central parts of the cortex (see Appendix for a detailed description of the physical model).

We first analyzed our dynamical system in the wild type situation. In this case, we capture the impact of centrosomes through an initial condition with reduced cortical active stress in their vicinity. The corresponding phase diagram exhibits a large zone of a single PAR-2 domain (Figure 4B, yellow area). Using moderate values of $\alpha$ and $\beta$ in this parameter space (Figure 4B, red arrows), the numerical simulation yielded a single PAR-2 domain (Figure 4C, Figure 4—figure supplement 1A, Video 6). Below a threshold of α, no polar PAR-2 domain was formed (Figure 4B, brown area), whereas for a very large curvature dependence of PAR-2 binding α and a very strong reduction of the cortical active stress β, a bipolar phenotype was obtained (Figure 4B, pale blue area). These two minor regions of the phase diagram seem outside of the parameter space in the wild type, since such embryos never exhibit bipolarity or lack polarity. Interestingly, this dynamical system also predicted that the PAR-2 domain relaxes to the nearby pole even if the initial active stress reduction occurred laterally (Figure 4—figure supplement 1B). This has been observed experimentally following lateral sperm entry in control embryos (Figure 4—figure supplement 1C, D) (Goldstein and Hird, 1996) or upon displacement of the PAR-2 domain from the pole using light-induced cytoplasmic streaming (Mittasch et al., 2018). This relaxation of the PAR-2 domain was also observed in air-1(RNAi) embryos, although in a less robust manner than in the wild type, likely due to impaired cortical flows (Figure 4—figure supplement 1D-F). Such congruence with experimental observations that were not included when developing our description provides further confidence that this dynamical framework provides a reliable representation of symmetry breaking and A-P polarization in C. elegans.

Video 6

Download asset

Next, we analyzed this dynamical system in a situation where cortical active stress is not reduced in a given location by a localized centrosome-derived cue, as may be the case in air-1(RNAi) embryos. Strikingly, the corresponding phase diagram exhibited a massive increase in the region of bipolar PAR-2 domains and a significant region of coexisting polar and bipolar states (Figure 4D, pale blue and green areas). Using the same values for $\alpha$ and $\beta$ as in the wild type, the numerical simulation yielded two PAR-2 domains (Figure 4E; Figure 4—figure supplement 1G, Video 7). Furthermore, we solved the equations in a situation corresponding to wild type embryos in a triangular geometry, which yielded a phase diagram exhibiting a single PAR-2 domain, except for extreme values of $\alpha$ and $\beta$ (Figure 4F, yellow area). A corresponding numerical simulation is shown in Figure 4G. By contrast, the phase diagram reflecting the situation in air-1(RNAi) embryos placed in a triangular geometry exhibited a region of a single PAR-2 domain that was larger than that of regular air-1(RNAi) embryos (Figure 4H, yellow area, compare to Figure 4D), in line with the increased proportion of air-1(RNAi) embryos with just one such domain observed experimentally, and an abrupt transition to regions with two and then three PAR-2 domains (Figure 4H, pale and dark blue areas). The corresponding numerical solution with the same values for $\alpha$ and $\beta$ as above yielded a single PAR-2 domain (Figure 4I). Overall, we conclude that the physical description of symmetry breaking and PAR-dependent axis specification developed here can explain bipolarity of embryos in which cortical active stress is not reduced by a localized centrosome-derived cue.

Video 7

Download asset

We aimed to understand when and where in the embryo AIR-1 acts to ensure proper symmetry breaking and polarization. To this end, we first analyzed the distribution of endogenously GFP-tagged AIR-1 (Sallee et al., 2018). Intriguingly, we found that, before symmetry breaking, GFP::AIR-1 localized in the vicinity of the cell cortex (hereafter referred to as ‘cortical’), as well as in the cytoplasm, and was barely detectable at centrosomes (Figures 5A and 00:00). Thereafter, the cortical and cytoplasmic pools decreased, which was paralleled by an increase of GFP::AIR-1 at maturing centrosomes (Figures 5A and 08:00-15:00). We found the same spatio-temporal distribution in wild-type embryos stained for endogenous AIR-1 (Figure 5B). Since the distribution of cortical AIR-1 is reminiscent of that of microtubules (O'Rourke et al., 2010), we assessed whether GFP::AIR-1 co-localized with mCherry::TBA-2, and found this to be the case indeed (Figure 5C). Likewise, endogenous AIR-1 co-localized with cortical microtubules before symmetry breaking in wild-type embryos (Figure 5H; Figure 5—figure supplement 1A). Importantly, we found that cortical AIR-1 was lost upon nocodazole-mediated depletion of microtubules (Figure 5E,I), whereas the centrosomal pool of AIR-1 was unaffected (Figure 5—figure supplement 1B). In contrast, depletion of AIR-1 did not alter the localization of cortical microtubules, but led to a slight increase in the corresponding signal intensity (Figure 5D, Figure 5—figure supplement 1A; Figure 3—figure supplement 1A,B). Overall, given that the absence of microtubules does not prevent A-P polarity (Cowan and Hyman, 2004; Sonneville and Gönczy, 2004; Tsai and Ahringer, 2007), these experiments lead us to propose that cortical AIR-1 does not modulate polarity establishment and that AIR-1 functions instead elsewhere, presumably in the cytoplasm, to prevent aberrant symmetry breaking early in the cell cycle.

Figure 5 with 1 supplement see all

Download asset Open asset

Interestingly, we found that SPD-2 or SPD-5, which are both required to recruit AIR-1 to centrosomes (Hamill et al., 2002; Kemp et al., 2004), are also needed for the presence of AIR-1 at the cell cortex, as well as for limiting the intensity of the cortical microtubule network signal (Figure 5F,G; Figure 5—figure supplement 1A). To test whether AIR-1 kinase activity is needed for cortical localization early in the cell cycle, we again analyzed transgenic embryos expressing GFP fusion proteins to either a wild type or to a kinase inactive version of RNAi-resistant AIR-1 (Toya et al., 2011), depleting endogenous AIR-1 using RNAi. Although wild-type GFP::AIR-1 was both cortical and centrosomal, as anticipated, the kinase inactive version did not localize to the cortex and barely to the centrosomes (Figure 5J,K). Overall, we conclude that microtubules, SPD-2, SPD-5 and AIR-1 kinase activity are needed for the presence of AIR-1 at the cell cortex early in the cell cycle.

To explore whether the cortical pool of AIR-1 might be sufficient to modulate A-P polarity, we utilized a strain expressing the GFP binding protein (GBP) fused to mCherry and to RNAi-resistant AIR-1 as a mean to localize AIR-1 at will to GFP-positive cellular compartments (Rothbauer et al., 2008). The GBP::mCherry::AIR-1 fusion protein rescued depletion of endogenous AIR-1 in ~67% of embryos, demonstrating that it possesses substantial activity (Figure 6B,C). As anticipated, co-depleting endogenous AIR-1 and SPD-2 in the presence of GBP::mCherry::AIR-1 resulted primarily in bipolar embryos as scored by mCherry::PAR-2 (Figure 6B,C). To target AIR-1 to the vicinity of the cell cortex, we crossed worms expressing GBP::mCherry::AIR-1 with animals expressing GFP::PH, which binds to phosphatidylinositol 4,5-bisphosphate (PI(4,5)P₂) at the plasma membrane (Figure 6A) (Audhya et al., 2005). This yielded a 2.5 fold increase of AIR-1 in the vicinity of the plasma membrane compared to cortical GBP::mCherry::AIR-1 alone (Figure 6—figure supplement 1A,B). Expression of mCherry::PAR-2 served as a readout for polarization in these experiments, whereas simultaneous depletion of SPD-2 ensured that no AIR-1 was targeted to centrosomes (Figure 6—figure supplement 1A). Importantly, we found that recruiting GBP::mCherry::AIR-1 to the vicinity of the plasma membrane prevented PAR-2 domain formation in ~72% of embryos, demonstrating that an excess of cortical AIR-1 can prevent symmetry breaking (Figure 6B,C).

Figure 6 with 1 supplement see all

Download asset Open asset

To test whether AIR-1 can also exhibit inhibitory function on symmetry breaking at endogenous protein levels, we made use of the temperature-sensitive mutant allele mat-1(ax161), which results in embryos being arrested in metaphase of meiosis I (Shakes et al., 2003). Such arrested embryos developed a small anterior PAR-2 domain when shifted to 24°C (Figure 6D,E; Figure 6—figure supplement 1C) (Wallenfang and Seydoux, 2000). Since AIR-1 distribution is cortical and cytoplasmic, but not centrosomal, during meiosis I (see Figures 5A), we assessed polarity in mat-1(ax161) air-1(RNAi) embryos to further assess the role of non-centrosomal AIR-1. When shifted to 24°C for 2 hr,~19% of mat-1(ax161) embryos developed a PAR-2 domain on the anterior side (Figure 6D,E; Figure 6—figure supplement 1C), a percentage that climbed to ~67% after 12 hr at 24°C (6D,E; Figure 6—figure supplement 1C). Strikingly, we found that ~39% mat-1(ax161) air-1(RNAi) embryos developed an anterior PAR-2 domain after 2 hr, and ~79% after 12 hr (Figure 6D,E; Figure 6—figure supplement 1C). Moreover, PAR-2 domains were more robust in mat-1(ax161) air-1(RNAi) embryos compared to the mat-1(ax161) mutant alone (Figure 6—figure supplement 1C). Next, we tested whether these findings could be captured by the dynamical system developed above. To this end, we translated the observed meiotic defects leading to anterior PAR-2 recruitment by imposing a weak anterior trigger (see Appendix). The resulting phase diagram representing mat-1(ax161) air-1(RNAi) embryos captured the experimentally observed shift from bipolarity towards a single PAR-2 domain (Figure 6F, yellow area, compare to Figure 4D). Taken together, these findings establish that endogenous levels of cortical or cytoplasmic AIR-1 can also negatively regulate polarity establishment upon compromised meiotic progression.

As a step towards investigating whether the centrosomal pool of AIR-1 contributes to proper symmetry breaking, we tested whether depleting other centrosomal components also gives rise to bipolar embryos. Importantly, we observed the same bipolar phenotype in mutant alleles of spd-2 and spd-5 (Figure 7A,B), as indicated previously (Hamill et al., 2002; O'Connell et al., 2000). By contrast, we found that depleting TBG-1, which is recruited to centrosomes in an AIR-1-dependent manner and is required for MTOC activity, does not impact polarity (Figure 7A,B) (Hannak et al., 2001).

Figure 7 with 1 supplement see all

Download asset Open asset

Centrosomes are thought to be essential for symmetry breaking and PAR-2 recruitment based on the outcome of laser ablation experiments (Cowan and Hyman, 2004). Why, then, do air-1(RNAi) embryos establish two PAR-2 domains instead of none, as would have been expected in the absence of functional centrosomes? We noticed that, in contrast to control embryos, the position of centrioles in air-1(RNAi) embryos does not always match the location or PAR-2 domain initiation (Figure 7A-Figure 7—figure supplement 1). This mirrors the findings in the triangular chambers and suggests that bipolarization occurs in a centrosome-independent manner. To thoroughly address this possibility, we first assayed embryos derived from such-1(t1668) mutant sperm, a subset of which do not harbor DNA or centrioles owing to meiotic defects during spermatogenesis (Figure 7C) (Bezler and Gönczy, 2010). Embryos fertilized by such sperm are recognizable by the absence of a male pronucleus, of microtubule asters and of GFP::SAS-7 signal (Figure 7D,E). To further test whether these embryos lack centrioles and not merely GFP::SAS-7, we labeled such-1(t1668) mutant embryos expressing GFP::PAR-2 and the centriolar marker GFP::SAS-4 with antibodies against GFP, as well as against AIR-1 as a proxy for centrosomes. As shown in Figure 7—figure supplement 1B, this experiment confirmed the absence of centrioles and centrosomes in embryos derived from such-1(t1668) mutant sperm. Strikingly, we found that such embryos usually establish a bipolar or an anterior PAR-2 domain (Figure 7D,E), indicating that centrosomes are dispensable for PAR-2 plasma membrane loading.

In an attempt to resolve the apparent discrepancy with the previous study (Cowan and Hyman, 2004), we conducted the same type of laser ablation experiment, but now using GFP::SAS-7 to monitor centrioles instead of GFP::SPD-2 as in the previous work. The bright GFP::SAS-7 signal enabled us to ablate centrosomes already during meiosis II. Importantly, we found that six out of nine embryos established bipolar GFP::PAR-2 domains upon centrosome ablation, with the remaining three developing an anterior PAR-2 domain (Figure 7F–H, Video 8). In contrast, embryos that underwent incomplete centrosome ablation and served as controls established a regular posterior PAR-2 domain (Figure 7—figure supplement 1C, Video 9). Although the above experiments further establish that centrosomes are dispensable for symmetry breaking, we considered whether sperm entry might impart a signal in the presumptive embryo posterior. To address this possibility, we took advantage of the fact that sperm occasionally enters the oocyte next to its nucleus (Goldstein and Hird, 1996). Importantly, we found that most air-1(RNAi) embryos were bipolar even in those cases where both male and female pronuclei were on the same side of the embryo (Figure 7I,J). Together, these experiments lead to the unequivocal conclusion that C. elegans embryos undergo spontaneous symmetry breaking in an unregulated manner in the absence of centrioles and centrosomes.

Video 8

Download asset

Video 9

Download asset

We wanted to test whether AIR-1 is not only necessary but also sufficient to ensure a single symmetry breaking event. To address this question, we again made use of the strain expressing the relocalizable GBP::mCherry::AIR-1. Strikingly, we found that localizing GBP::mCherry::AIR-1 to centrioles by crossing GFP::SAS-7 in this strain rescued posterior polarity of mCherry::PAR-2 in the majority of embryos (Figure 7K,L), although the PAR-2 domain was smaller compared to the control condition (Figure 7D). Importantly, centrioles in this experiment do not possess a PCM since SPD-2 was depleted as well; as a result, such embryos do not form a spindle (Figure 7—figure supplement 1E). We conclude that AIR-1 localized on centrioles is sufficient to impart substantial posterior polarity.

We discovered that AIR-1 localizes to the cell cortex prior to symmetry breaking, overlapping substantially with the cortical array of microtubules present at this early stage of development. We found also that SPD-2 and SPD-5 are needed for cortical AIR-1 distribution, in addition to their known requirement for centrosomal AIR-1 recruitment. In Xenopus egg extracts, interaction with the SPD-2 homologue CEP192 is required for autophosphorylation and thereby activation of Aurora A (Joukov et al., 2014), raising the possibility that SPD-2 is likewise needed for activating AIR-1 in C. elegans. This possibility is compatible with the observation that preventing AIR-1 activity using a kinase inactive mutant or a small molecule inhibitor results in cortical depletion of AIR-1 (Kotak et al., 2016). The dual requirement of SPD-2 and SPD-5 for correct AIR-1 distribution and/or activity both early and late in the cell cycle provides an explanation for why spd-2(RNAi) and spd-5(RNAi) embryos exhibit the same polarity defects as air-1(RNAi) embryos.

We found in addition that targeting excess AIR-1 to the vicinity of the plasma membrane is generally sufficient to prevent PAR-2 recruitment throughout the embryo, in a manner independent of SPD-2 and SPD-5. Therefore, cortical AIR-1 can act as a negative regulator of polarity establishment in some circumstances. However, when centrosomes are present, cortical AIR-1 is not needed to prevent polarity establishment, as evidenced by the fact that microtubule removal, which results in the loss of cortical AIR-1, does not impact A-P polarity. Furthermore, we found that endogenous cortical or cytoplasmic AIR-1 can limit the expansion of an anterior PAR-2 domain upon delayed meiotic cell cycle progression, and thus could be of importance under environmental stress conditions that cause such delays (Davis-Roca et al., 2017).

As the zygote progresses through the cell cycle, AIR-1 rapidly becomes enriched at centrosomes, in line with the fact that AIR-1 is very mobile (Kress et al., 2013). We found that targeting GBP::mCh::AIR-1 to centrioles in embryos depleted of SPD-2, and thereby of microtubule nucleation activity, is sufficient to recruit PAR-2 to the neighboring plasma membrane. Therefore, later in the cell cycle, centrosomal AIR-1 can act as a positive regulator of polarization.

What links the two facets of AIR-1 function in symmetry breaking? Our findings suggest two possible scenarios. In both of them, AIR-1 acts initially throughout the embryo as a negative regulator of symmetry breaking, perhaps by phosphorylating a cortical substrate protein. Thereafter, in a first scenario, AIR-1 could undergo substrate switching upon relocalizing to centrosomes, and phosphorylate another substrate in that location, which would be critical for symmetry breaking, possibly through diffusion from the centrosomes to the neighboring cortex. In a second scenario, AIR-1 could have just one cortical substrate of relevance for polarity throughout the first cell cycle, in which case centrosomes would serve to locally deplete AIR-1, thereby ensuring uniqueness of symmetry breaking.

There are countless Aurora A kinase substrates from yeast to man (Koch et al., 2011; Sardon et al., 2010; Tien et al., 2004; Willems et al., 2018), and it will be interesting to identify the one(s) that serve to polarize the C. elegans zygote. Moreover, it will be valuable to explore whether the dual function uncovered in C. elegans extends to homologous proteins, including the human Aurora A oncogene.

Centrosomes are sufficient to dictate symmetry breaking and thus polarity along the A-P axis of wild-type C. elegans zygotes. Indeed, the position of centrosomes within the embryo determines the location of symmetry breaking and of PAR-2 recruitment to the plasma membrane (Bienkowska and Cowan, 2012; Cowan and Hyman, 2004; Cuenca et al., 2003; Goldstein and Hird, 1996). Furthermore, centrosomes have been proposed to be necessary for A-P polarity, since their ablation by a laser microbeam has been reported to result in the absence of PAR-2 throughout the embryo (Cowan and Hyman, 2004). In contrast, using a centriolar marker bright enough to conduct physical centrosome ablation experiments at early time points, we found here that embryos usually recruit PAR-2 in a bipolar fashion in the absence of centrioles, as they do in air-1(RNAi) embryos. The timing of centrosome ablation might be critical for the outcome of such experiments and potentially might explain the discrepancy between our results and the earlier work. Since symmetry breaking as evidenced by cortical flows occurs before the posterior PAR-2 domain becomes detectable (see Figure 4—figure supplement 1C), it is plausible that centrosome ablation in the earlier study occurred when symmetry breaking had been initiated. Perhaps such initiation would no longer allow for spontaneous symmetry breaking and thus result in the complete absence of cortical PAR-2. Regardless, together with the genetic centrosome ablation experiments conducted with such-1 mutant sperm, our findings from the physical centrosome ablations firmly establish that polarization, albeit aberrant, occurs without centrosomes in C. elegans.

The mechanism that governs spontaneous polarization of the zygote in the absence of AIR-1 or of centrosomes is largely independent of cortical flows. Indeed, two PAR-2 domains can form in nmy-2(ne3409) air-1(RNAi) embryos, where cortical flows are essentially absent or below the detection limit. However, PAR-2 domains in such embryos are significantly smaller and develop later than in air-1(RNAi) embryos. Our findings also establish that the mechanism leading to spontaneous polarization in the absence of AIR-1 or of centrosomes functions independently of microtubules. Given this, it is intriguing that GFP::PAR-2 R163A or R183-5A, two mutants with impaired microtubule binding, fail to recruit PAR-2 onto the plasma membrane in air-1(RNAi) embryos. A potential explanation for this conundrum is that these mutants are impaired in more functions than just microtubule binding, for instance in PAR-2 oligomerization (Arata et al., 2016). In conclusion, our findings indicate that polarized membrane domains can be initiated largely independently of cortical flows through the instability of the PAR system, but that cortical flows help sharpen and accelerate polarization of the embryo.

We analyzed the dependence between PAR proteins and cortical flows in polarizing the zygote by coupling a reaction-diffusion-advection system to active gel theory. In previous work, cortical flows were considered to be induced by centrosomes, which cannot act as a trigger in air-1(RNAi) embryos, nor following their genetic or physical ablation. Instead, we assumed that PAR-2 indirectly reduces cortical contractility via antagonism with PAR-6 (Gross et al., 2019). Thereby, PAR-2 distribution can result in gradients of contractility and thus of cortical flows despite the absence of centrosomes.

Based on the findings with air-1(RNAi) embryos, including when placed into triangular PDMS chambers, we introduced a preferential attachment rate α of PAR-2 in the polar regions of the embryo in the dynamical system. Although we demonstrated experimentally that PAR-2 domains form preferentially in such regions, the underlying mechanism remains to be clarified further. Perhaps the preference stems from distinct lipid distributions in curved regions that could influence the attachment rate of PAR-2, which can bind phospholipids (Motegi et al., 2011). Moreover, purely geometric effects can cause preferential attachment at the poles (Rangamani et al., 2013) and have been suggested to play a role for PAR protein localization in C. elegans (Gessele et al., 2018). However, whereas geometry has been shown to be important for some systems of protein self-organization (Wettmann and Kruse, 2018), how exactly it contributes to PAR-2 membrane recruitment remains to be deciphered.

Regardless, the physical description we developed successfully captures the phenotypes of wild-type embryos, of embryos depleted of AIR-1, as well as of embryos placed in triangular chambers. It will be of interest to investigate embryos placed in differently shaped PDMS chambers, as this might help uncover the mechanism enabling curvature preference and further develop the physical model. Moreover, we realize that a potential limitation of our current physical analysis stems from having been performed in 1D. A future analysis using a 2D geometry will likely uncover further aspects of the self-polarization mechanism, including by explicitly accounting for the impact of centrosomes and geometry, as well as by providing a full description of cortical dynamics.

Spontaneous polarization mechanisms such as the one revealed here for C. elegans are encountered in other systems. For instance, a self-enhancing positive feedback loop of Cdc42-GTP contributes to polarization of the emerging bud in S. cerevisiae (Bendezú et al., 2015). Another example occurs in HeLa tissue culture cells forced to undergo monopolar division, which exhibit spontaneous symmetry breaking manifested by a concentration of actin, microtubules and cleavage furrow components on one side of the cell (Hu et al., 2008). Spontaneous symmetry breaking can also be recapitulated in cell-free assays, including in actin cortices formed at the periphery of Xenopus egg extracts encapsulated in an oil emulsion (Abu Shah and Keren, 2014). Likewise, beads coated with Listeria monocytogenes ActA, which catalyzes actin polymerization, spontaneously break symmetry and display directional motion due to small asymmetries of the polymerized actin around the bead (van Oudenaarden and Theriot, 1999). Our work echoes such findings from other domains of life and demonstrates that spontaneous polarization can also occur during the very first step of development in a metazoan organism. Such a centrosome-independent symmetry breaking mechanism could be of particular importance in systems devoid of centrioles such as parthenogenetic species. In the wild-type C. elegans zygote, AIR-1 funnels the inherent self-organizing properties of the cell onto a single site, next to paternally contributed centrioles, thus ensuring spatial coupling between paternal and maternal contributions at the onset of development.

Nematodes were maintained at 24°C using standard protocols (Brenner, 1974). Worms carrying temperature-sensitive mutations were maintained at the permissive temperature (16°C) and shifted to the restrictive temperature (24°C) for 24 hr prior to imaging. See Supplementary file 1 for the list of worm strains used in this study. New transgenic worm strains were generated as described just hereafter:

The GBP::mCherry::AIR-1 construct was obtained by inserting the mCherry coding sequence at the 3’-end of the GFP-binding protein (GBP) coding sequence, obtained from the pAOD-VHHGFP4 vector (a kind gift from Aurelien Olichon), followed by codon optimized RNAi-resistant air-1 cDNA (a kind gift from Asako Sugimoto). Flexible linkers encoding thirteen amino acids (GAGAGAGAGAFSV) were inserted between GBP and mCherry cDNAs, as well as between mCherry and AIR-1 cDNAs. The corresponding integrated transgenic line was generated by microparticle bombardment of unc-119(ed3) mutant worms (Praitis et al., 2001).

CRISPR/Cas9-mediated genome editing was used as described previously (Dickinson et al., 2015) to insert tagRFP-T in frame at the 5' of the sas-7 locus. To target Cas9 to the sas-7 locus, the 5'gcttaaaatcaactcaccg(TGG)3' −19 bp targeting sequence was inserted into the Cas9-sgRNA construct (pDD162) (Addgene). Homologous repair template for insertion of the self-excising selection cassette (SEC) was generated by modifying the pDD284 (Addgene) vector using Gibson assembly. Left and right homology arms of 655 bp and 652 bp were PCR-amplified, before insertion into pDD284 (Addgene) opened using ClaI, SpeI and PsiI to remove the Flag tag from the original vector. A silent mutation in the PAM motif (5'TGG3' -->5'TAG') was introduced to prevent repair template cutting and a 5'CACTCCACTGGAACCTCTAGA3'−21 bp linker was added between the tagRFP and sas-7 coding sequences. An injection mix containing 50 ng/μl targeting vector, 50 ng/μl homologous repair template, 10 ng/μl pGH8 (Prab-3::mcherry::unc-54 3'UTR) (Addgene), 5 ng/μl pCFJ104 (Addgene) (Pmyo-3::mcherry::unc-54 3'UTR) and 2.5 ng/μl pCFJ90 (Pmyo-2::mcherry::unc-54 3'UTR) (Addgene) was micro-injected into the gonads of wild type (N2) worms. Selection of insertion events and removal of the SEC cassette was performed as described (Dickinson et al., 2015).

A similar strategy was used to tag endogenous SPD-2 with RFP. The 5'tgttcattacagagattcat(TGG)3'−21 bp targeting sequence was cloned into the pDD162 (Addgene) vector and pDD286 (Addgene) was modified using Gibson assembly and Q5-directed mutagenesis to generate the homologous repair template. PCR amplified 639bp-left and 667bp-right homology arms were inserted into pDD286 (Addgene) digested with ClaI and NgoMIV; six silent mutations were introduced in the targeted sequence and the PAM site 5'tgctcgttgcacaggttcat(GGG)'3. As described above, an appropriate injection mix was micro-injected in the gonads of wild type N2 worms; screening was performed as described (Dickinson et al., 2015).

RNAi-mediated depletion was performed essentially as described (Kamath et al., 2001), using bacterial feeding strains from the Ahringer (Kamath and Ahringer, 2003) or Vidal libraries (Rual et al., 2004) (the latter a gift from Jean-François Rual and Marc Vidal). The bacterial feeding strain for depleting specifically endogenous AIR-1 without affecting air-1 RNAi resistant transgene (tjls173 and tjls188) expression was a kind gift from Asako Sugimoto (Toya et al., 2011). The bacterial feeding strains used to deplete specifically endogenous PAR-2 in strains expressing RNAi-resistant par-2 transgenes (axIs1933, axIs1936, axIs1934) was a kind gift from Fumio Motegi and Geraldine Seydoux (Motegi et al., 2011).

RNAi for air-1 (Vidal), rga-3 (Vidal), spd-2 (Vidal), spd-5 (Vidal) or par-2 (Ahringer) was performed by feeding L3-L4 animals with bacteria expressing the corresponding dsRNA at 24°C for 16–26 hr. In Figure 1—figure supplement 1B, air-1(RNAi) was diluted 1:5 or 1:10 with bacteria expressing an empty vector (L4440). RNAi for tbg-1 (Ahringer) and tbg-1 +air-1 was performed by feeding L2-L3 animals with bacteria expressing dsRNAs at 20°C for 48 hr and imaging their embryos. RNAi for tba-2 (Ahringer) and tba-2 +air-1 was performed by feeding L3 animals with bacteria expressing dsRNA at 24°C for 30–36 hr and imaging their embryos. Strains carrying temperature sensitive mutations were shifted to the restrictive temperature (24°C) for 24 hr and RNAi performed as described above.

Gravid hermaphrodites were dissected in osmotically balanced blastomere culture medium (Shelton and Bowerman, 1996) and the extracted embryos mounted on a 2% agarose pad. Dual fluorescence and DIC time lapse microscopy was performed at room temperature with a 60x CFI Plan Apochromat Lambda (NA 1.4) objective on a Nikon Eclipse Ti-U Inverted Microscope connected to an Andor Zyla 4.2 sCMOS camera, or with a 63x Plan-Apochromat (NA 1.4) objective on a Zeiss ObserverD.1 inverted microscope connected to the same type of camera. One frame was captured every 10 s, and a z-stack was acquired at every time point, covering 20 µm, with a distance of 0.7 µm between focal planes. Time lapse images in Figure 5A,C,J,K; Figure 3—figure supplement 1E; Figure 5—figure supplement 1B were acquired at 23°C using an inverted Olympus IX 81 microscope equipped with a Yokogawa spinning disk CSU - W1 with a 63x (NA 1.42 U PLAN S APO) objective and a 16-bit PCO Edge sCMOS camera. Images were obtained using a 488-nm solid-state laser at 60% laser power, with an exposure time of 400 ms. Embryos were imaged every 5 s, and a z-stack was acquired at every time point, covering 2 µm of the embryo cortex, with a distance of 0.5 µm between focal planes (i.e. four cortical planes).

Worms were dissected in 10 µg/ml Nocodazole (Sigma Aldrich, M1404), diluted in 3.3% DMSO and 96.7% osmotically balanced blastomere culture medium (Shelton and Bowerman, 1996); the extracted embryos were mounted on a 2% agarose pad. The drug entered the embryos due to the permeability of the eggshell during meiosis, and was effectivee in this manner, as evidenced notably by the lack of pronuclear migration. Epifluorescence imaging was performed as described above.

We conducted laser ablation on a previously described set-up (Mayer et al., 2010). Ablation was performed by applying 10 ultraviolet pulses at 1 kHz at equidistant sites on the circumference of a circle 0.5 μm in diameter. The centrosome was ablated at the end of meiosis II, before the centrioles moved towards the cortex. After ablation, the zygote was imaged every 10 s until nuclear envelope breakdown. Embryos were monitored for the re-appearance of the centrosomal signal, which occurred when the centrosome was only bleached and not destroyed, as well as for cell rupture or for membrane damage, which resulted in cytoplasmic leakage; all such embryos were excluded from the analysis.

Images acquired at the spinning disk were processed by z-projecting the four cortical planes using maximum intensity (ImageJ software). Images acquired with the epifluoresence microscope (Zeiss ObserverD.1or Nikon Eclipse Ti-U) were processed as follows using ImageJ: images underwent background subtraction using a ‘rolling ball’ algorithm of 10 pixels, followed by a maximum intensity z-projection. Grey levels were set identically for all images within each experimental series.

For flow velocity analysis, heat maps were obtained using Particle Image Velocimetry (PIV) with a freely available PIVlab MATLAB algorithm (pivlab.blogspot.de). Using PIVlab, we performed a four-step multi pass with a final interrogation area of 8 pixels with a step of 4 pixels. 2D velocity fields were obtained by averaging the x-component of velocity along the y-axis for each value in a single frame. All values were averaged over several embryos and represented in a heat map using MATLAB. Prior to averaging all values across multiple embryos, embryos were aligned temporally using the best fit of the pronuclei growth curves (De Simone et al., 2016). Pronuclear diameter was measured from the DIC images using ImageJ. For Figure 7—figure supplement 1A, centrosomes were tracked in three dimensions using the ImageJ plugin TrackMate (imagej.net/TrackMate).

Images were processed as described above using ImageJ following straightening of the embryo’s outline for each time point using the ImageJ plugin ‘Straighten’ (imagej.nih.gov/ij/plugins/straighten.html). The intensity of each pixel along the embryo’s straightened circumference was measured with ImageJ, and the resulting values displayed using Matlab in a heat map representing the GFP::PAR-2 intensity.

To calculate the extent of the PAR-2 domain, the circumference of the cell was smoothened in Matlab for each time point using a 23-point moving average to reduce acquisition noise. For each time point, a noise floor estimation was obtained by detecting the max and min values following a smoothing by applying a moving average. The noise floor value was then approximated using the minimum values. The value of the average background noise was used as a threshold to determine the points constituting the domain boundary. The same threshold was applied to all analyzed image sequences. In order to reduce noise-induced threshold crossing, the data values were squared.

SUCH-1 depletion causes, in addition to the paternal sperm phenotype, a maternally contributed delay in mitosis. We used such-1(t1168) males to fertilize fem-1(hc17) females, thus ensuring that all embryos resulted from such-1(t1168) mutant sperm and that the maternal contribution was normal. Gravid fem-1(hc-17) hermaphrodites expressing RFP::NMY-2; GFP::PAR-2; GFP::SAS-7 were shifted to the restrictive temperature (24°C) and L4 progeny were then mated with such-1(t1668) males at 20°C for 24 hr. Gravid adults were dissected and embryos imaged as described above. Prior to and after imaging, each embryo was thoroughly screened for the absence of centrosomes as evaluated by the lack of focal GFP::SAS-7 signal.

To target GBP::mCherry::AIR-1 to centrosomes, L4 larvae expressing mCherry::PAR-2 and GBP::mCherry::AIR-1 were mated with GFP::SAS-7 males for 24 hr at 20°C on spd-2(RNAi). Gravid adults were dissected and embryos imaged as described above. Only embryos that were fertilized, as evaluated by the presence of GFP::SAS-7, were imaged.

Embryos were permeabilized by freeze-cracking, followed by fixation in methanol at −20°C for 5 min and by incubation for 1 hr at room temperature with the following primary antibodies: mouse anti-α-tubulin, 1/200, (DM1A, Sigma), rabbit anti-PAR-2, 1/200 (Pichler et al., 2000), rabbit anti-AIR-1, 1/1000 (Hannak et al., 2001), chicken anti-GFP, 1/200 (Abcam). Secondary antibodies were Alexa-Fluor-647-coupled anti-mouse-IgG, Alexa-Fluor-488-coupled anti-chicken-IgY and Alexa-Fluor-568-coupled anti-rabbit-IgG, all used at 1:500 (Jacskon laboratories). Slides were counterstained with 1 mg/ml Hoechst 33258 (Sigma) to reveal DNA.

Micro-well structures were produced using standard soft-lithography methods. For triangular PDMS chambers, a custom lithography mask of a silicon waver coated with AZ 15nXT (MicroChemicals GmbH, Ulm, Germany) was illuminated with UV to generate a casting mould of triangles with a side length of ~40 µm. Micro-well structures were generated by pouring poly(dimethylsiloxane) (PDMS; Sylgard 184, VWR) into these casting moulds. Finally, triangular PDMS structures were cured for 1.5 hr at 75°C. Before each experiment, PDMS chambers were treated with oxygen plasma for 90 s to reduce hydrophobicity. Worms were dissected as described above, and embryos placed into the triangular chamber with an eyelash tool prior to symmetry breaking. Epifluoresence imaging was performed as described above.

PDMS cylinders with a diameter of 35 µm utilized in Figure 3—figure supplement 1D were generated in an analogous manner, and embryos also placed into the cylindrical chamber with an eyelash tool prior to symmetry breaking. Embryos entered the chamber with the anterior or posterior pole first, allowing imaging of one of the other pole using spinning disk confocal microscopy as described above.

Statistical significance for all shown quantification, except for Figure 3E; Figure 6—figure supplement 1B,C; Figure 7—figure supplement 1D, was tested using the non-parametric two-sided Fisher’s exact test, comparing a single category against the other pooled categories (e.g. posterior vs non-posterior) for two conditions (e.g. control versus air-1(RNAi)). The corresponding p-values for all experiments can be found in Supplementary file 2. A p-value<0.05 was considered as statistically significant. Statistical significance in Figure 3E; Figure 6—figure supplement 1B,C and Figure 7—figure supplement 1D was tested using the non-parametric Mann-Whitney-U test. A p-value<0.05 was considered as statistically significant.

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

We are grateful to Carrie Cowan and Sachin Kotak for sharing unpublished observations and for fruitful discussions. For their gift of worm strain to Bruce Bowerman, Fumio Motegi, Asako Sugimoto, Jessica Feldman, Anthony Hyman and Geraldine Seydouxs, as well as the Caenorhabditis Genetics Center (CGC), which is funded by NIH Office of Research Infrastructure Programs (P40 OD010440). We thank Alexandra Bezler, Alessandro De Simone and Radek Jankele for comments on the manuscript. We thank the Microstructure Facility (CMCB, TU-Dresden), in part funded by the State of Saxony and the European Fund for Regional Development — EFRE, for the production of the triangular micro-well chambers. This work was supported by post-doctoral fellowships from EMBO to KeKl (ALTF 81–2017) and to MP (ALTF 1426–2016), the Fondation Bettencourt.Schueller prize to NL, as well as grants from the Swiss National Science Foundation to PiGö (31003A_155942) and to KaKr, (205321_175996), and from the European Research Council (281903 and 742712) to SWG.

© 2019, Klinkert et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.