The Rac1 homolog CED-10 is a component of the MES-1/SRC-1 pathway for asymmetric division of the Caenorhabditis elegans EMS blastomere

Abstract

Asymmetric cell division is essential for the creation of cell types with different identities and functions. The endomesodermal precursor cell (EMS) of the 4-cell Caenorhabditis elegans embryo undergoes an asymmetric division in response to partially redundant signaling pathways. One pathway involves a Wnt signal from the neighboring P₂ cell, while the other pathway is defined by the receptor-like MES-1 transmembrane protein localized at the EMS-P₂ cell contact and the cytoplasmic kinase SRC-1. In response to these signals, the EMS nuclear–centrosome complex rotates, so that the spindle forms on the anterior–posterior axis; after division, the daughter cell contacting P₂ becomes the endodermal precursor cell. Here, we identify the Rac1 homolog CED-10 as a new component of the MES-1/SRC-1 pathway. Loss of CED-10 affects both spindle positioning and endoderm specification in the EMS cell. SRC-1 dependent phosphorylation at the EMS-P₂ contact is reduced. However, the asymmetric division of the P₂ cell, which is also MES-1 and SRC-1 dependent, appears normal in ced-10 mutants. These and other results suggest that CED-10 acts upstream of, or at the level of, SRC-1 activity in the EMS cell. In addition, we find that the branched actin regulator ARX-2 is enriched at the EMS-P₂ cell contact site, in a CED-10-dependent manner. Loss of ARX-2 results in EMS spindle orientation defects, suggesting that CED-10 acts through branched actin to promote spindle orientation in the EMS cell.

Introduction

Asymmetric cell division is the process by which dividing cells give rise to daughter cells with different identities and fates. This type of division occurs in all multicellular organisms and is necessary for cell type diversification during both embryonic development and adult tissue homeostasis (Morin and Bellaiche 2011; Inaba and Yamashita 2012; Venkei and Yamashita 2018). A key aspect of asymmetric division is alignment of the mitotic spindle along a specific axis (D’avino et al. 2005; McNally 2013; Bergstralh and St Johnston 2014; di Pietro et al. 2016; Kotak 2019). In many metazoan cell divisions requiring an oriented spindle, the microtubule pulling forces necessary to move the spindle are generated by the asymmetric localization of a conserved cortical complex. This force-generating complex contains the minus-end–directed motor dynein, its partner dynactin, and the adaptor protein NuMA (LIN-5 in C. elegans, mud in Drosophila). The complex can be recruited to the membrane or cortex by various adaptor and anchor proteins, and several different polarized cues have been identified that instruct spindle orientation (Morin and Bellaiche 2011; McNally 2013; di Pietro et al. 2016). For example, in many types of metazoan asymmetric divisions, the evolutionary conserved PAR polarity proteins establish the polarity axis and influence the localization or activity of NuMA and the force-generating complex. Wnt signaling pathways also act via NuMA and dynein to align mitotic spindles with tissue polarity in several systems. However, for many cell types, the detailed mechanisms by which spindle orientation is regulated remain to be elucidated (Gillies and Cabernard 2011; di Pietro et al. 2016; Kotak 2019).

The early Caenorhabditis elegans embryo serves as an excellent model for studying the molecular mechanisms of asymmetric division in different developmental contexts, as the embryo exhibits an invariant division pattern that includes multiple asymmetric divisions regulated by different cues (Rose and Gonczy 2014; Griffin 2015; Pacquelet 2017). After fertilization, the zygote (or P₀ cell) establishes molecular asymmetries that define the anterior and posterior embryonic poles. Following pronuclear meeting, the spindle aligns with the anterior–posterior axis and thus division creates an anterior somatic cell, AB, and a posterior germ cell precursor, P₁. The AB daughter divides symmetrically to give rise to ABa and ABp at the anterior and dorsal aspects of the embryo, respectively. The P₁ cell divides asymmetrically to give rise to the EMS and P₂ cells at the ventral and posterior aspects of the embryo, respectively. EMS and P₂ both divide asymmetrically again (Fig. 1a) (Rose and Gonczy 2014; Griffin 2015; Pacquelet 2017).

The asymmetric division of the P₀ cell is instructed by PAR polarity. In these cells, PAR proteins antagonize one another's cortical localization to form mutually exclusive anterior and posterior domains (Rose and Gonczy 2014; Griffin 2015; Pacquelet 2017). In response to the formation of the PAR domains, a cytoplasmic gradient of cell fate determinants forms along the anterior–posterior axis. Downstream of PAR polarity cues, the force-generating complex described previously localizes asymmetrically on the cortex and exerts microtubule pulling forces that cause the nuclear–centrosome complex to rotate, such that the spindle forms along the axis of PAR asymmetry (Rose and Gonczy 2014; Pacquelet 2017; Kotak 2019). The resulting mitotic division gives rise to daughters that have inherited different quantities of PAR proteins and cytoplasmic fate determinants and therefore adopt different fates. In the asymmetric divisions of the posterior P₁ daughter, mutually exclusive cortical PAR domains form and the nuclear–centrosome complex rotates onto the axis of PAR polarity. The mechanisms regulating nuclear rotation and unequal division of P₁ are thought to be similar to those in the 1-cell, P₀ (Rose and Gonczy 2014; Pacquelet 2017; Kotak 2019).

In contrast, the asymmetry of the EMS division is generated by 2 partially redundant signaling pathways that require contact with EMS's neighboring cell, P₂: a well-described Wnt pathway and a less understood MES-1/SRC-1 pathway involving the transmembrane tyrosine kinase-like receptor MES-1 and the cytoplasmic tyrosine kinase SRC-1 (Fig. 1a; reviewed in Sawa 2012; Rose and Gonczy 2014; Maduro 2017). In response to these pathways, the mitotic spindle orients on the anterior–posterior axis (Fig. 1a). The anterior daughter of this division, MS, becomes a mesoderm progenitor, while the posterior daughter, E, born adjacent to P₂, becomes an endoderm progenitor.

Double-mutant analysis has been the standard method for assigning genes to one of the partially redundant EMS asymmetric division pathways. Homozygous maternal loss of function for MES-1, SRC-1, or any of several Wnt pathway members causes a low rate of late or failed spindle orientation in the EMS cell. However, when a mes-1 or src-1 mutation is combined with mutation of a Wnt pathway component, the rate of failed orientation increases to 80–100%. In contrast, Wnt component double mutants and mes-1;src-1 double mutants have the same low frequency of orientation defects as single mutants (Rocheleau et al. 1997; Thorpe et al. 1997; Schlesinger et al. 1999; Bei et al. 2002; Walston et al. 2004). The MES-1 protein is localized to the EMS-P₂ cell contact, apparently in both cells, as this localization requires cell–cell contact specifically between these cells (Berkowitz and Strome 2000; Bei et al. 2002). Further, MES-1 is required for activation of SRC-1 at the EMS-P₂ contact, based on the staining pattern of an antibody that detects SRC-1-dependent tyrosine phosphorylation (Bei et al. 2002). Thus, MES-1 and SRC-1 are part of a pathway that is functionally redundant with Wnt signaling in promoting EMS spindle positioning. The 2 pathways may converge on the conserved force-generating complex described earlier. Loss of the dynein heavy chain, dynactin, or LIN-5 results in stronger spindle orientation defects than single mutants in either the MES-1/SRC-1 or Wnt pathways, suggesting that LIN-5 and dynein may act in both pathways (Zhang et al. 2007; Liro and Rose 2016).

The induction of the E cell to form endoderm also depends on the MES-1 and Wnt pathways. Endoderm fate specification requires the Wnt ligand (MOM-2 in C. elegans, expressed in the P₂ cell), the transmembrane Frizzled receptor (MOM-5), and other conserved Wnt pathway components (Rocheleau et al. 1997; Thorpe et al. 1997; Schlesinger et al. 1999). Activation of this pathway ultimately results in nuclear export of the transcription factor POP-1 in the E cell, which allows endoderm-specific gene expression in this cell (Sawa 2012; Maduro 2017). While mutations in either MES-1 or SRC-1 do not result in endoderm defects, mutation of either in combination with loss of function mutations in Wnt pathway components results in increased numbers of embryos without endoderm; loss of SRC-1 was also shown to enhance the POP-1 nuclear export defect of Wnt pathway mutants (Sugioka and Sawa 2010; Sugioka et al. 2011; Sumiyoshi et al. 2011).

A previous study found that CED-10, 1 of 3 C. elegans Rac proteins, is required for spindle orientation in the ABar blastomere of the 8-cell embryo, which is a Wnt-dependent pathway (Cabello et al. 2010). Rac proteins, a conserved subclass of the Rho GTPase family, are involved in cytoskeletal regulation across diverse developmental contexts and many organisms (Boreaux et al. 2007; Bustelo et al. 2007; Hall 2012; Duquette and Lamache-Vane 2014). The same C. elegans study reported that ced-10 mutant embryos exhibited abnormal EMS spindle positioning in the 4-cell embryo (Cabello et al. 2010). However, this phenotype was not characterized, nor was the relationship between CED-10 and the signaling pathways known to promote EMS spindle positioning. Therefore, we set out to define the genetic and molecular role CED-10 plays in this asymmetric division. Here, we demonstrate that CED-10 works in parallel with the Wnt pathway in the EMS cell, where it acts upstream or at the level of SRC-1 for both spindle orientation and endoderm specification.

Materials and methods

Strains

Nematodes were maintained on Modified Youngren's Only Bacto-Peptone agar plates, seeded with Escherichia coliOP50 using standard procedures (Brenner 1974; Church et al. 1995; Stiernagle 2006). Temperature-sensitive strains were maintained at 16°C, and others were maintained at ambient temperature (22–24°C). The full genotypes of the strains used in this study are found in Supplementary Table 1.

RNA interference

For all target genes, RNAi was performed by feeding as described (Ahringer 2006), with the modification of adding Isopropyl β-D-1-thiogalactopyranoside (IPTG) to the bacterial culture, to a final concentration of 1 mM, just prior to seeding the culture on the feeding plates. Plates were then kept in the dark at 4°C for no more than 5 days prior to use. The following RNAi feeding clones used were obtained from the Ahringer library (Kamath et al. 2003): arx-2 (V-7M13), mom-2 (V-6A13), and mes-1(X-5L23); the L4440 empty vector was used as control for all RNAi experiments. For all conditions, worms at the late L3 through L4 stages were placed on plates seeded with RNAi bacteria and grown at 20°C unless otherwise noted, with imaging carried out at room temperature. For arx-2, worms were grown for 48–52 hr RNAi was judged to be effective if many small cortical extrusions were present in early embryos and there were unhatched siblings on the RNAi plate (Patel et al. 2008; Roh-Johnson and Goldstein 2009); however, when individual worms were examined, lethality of their embryos ranged from 100 to 0% and thus RNAi depletion was variable and incomplete. Attempts to produce stronger RNAi by incubating worms at 16°C for 72–110 h still resulted in variable lethality, and the spindle positioning defects were similar to those reported in Fig. 5. For mom-2 RNAi in experiments to score spindle orientation, worms were grown for 36–48 h, and RNAi was considered effective if mes-1(bn7); GFP::tubulin; mom-2RNAi embryos filmed in parallel exhibited 80–100% abnormal spindle orientations; mes-1 RNAi was carried out for 24–20 h, and considered effective if mom-5(zu193); GFP tubulin; mes-1(RNAi) embryos treated in parallel exhibited 80–100% abnormal spindle orientations as in (Liro and Rose 2016). For mom-2 and mes-1 RNAi used to test for endoderm specification, worms were grown at 16 for 48–72 h, with RNAi effectiveness determined by enhancement of mes-1(bn7) or mom-2(or42).

The ced-10 RNAi clone (IV-1J04) obtained from our copy of the Arhinger library was found to contain no insert. An RNAi feeding construct was generated by subcloning the ced-10 cDNA region of pPR37 (constructed by Peter Reddien, courtesy of Erik Lundquist) into the NcoI and SacI sites of the L4440 vector. The plasmid was transformed into HT115 and sequence was confirmed. Defects in P₁ and EMS spindle rotation in mom-5(zu193); GFP::tubulin; ced-10(RNAi) embryos confirmed that CED-10 knockdown with this plasmid was sufficient to induce an enhanced spindle phenotype, after 48–52 h feeding at 20°C. Many small cortical extrusions were also present in these embryos, as reported for ced-10(t1875) (Price et al. 2022). In parallel experiments, knockdown efficacy was assessed by quantifying the depletion of fluorescent protein in GFP::CED-10; ced-10(RNAi) embryos after 48–52 h at 20°C, which resulted in a reduction of the GFP signal to background N2 levels as described in the text. Nonetheless, older embryos still exhibited CED-10::GFP signal, and most ced-10(RNAi) embryos hatched; these data indicate that while RNAi treatment causes a strong loss of function in the early embryo, it does not produce a complete loss. The following conditions were then used for RNAi of CED-10 for different experiments. For immunostaining of MES-1, 28–30 h of feeding at room temperature was conducted, with the appearance of cortical blebs as a readout of RNAi efficacy. For SRC-1::GFP and endoderm/gut granule assays, feeding was carried out at 48–72 h at 16°C to allow for longer production of embryos. In addition to scoring for cortical blebs, enhancement of mom-2 spindle orientations defects was scored first as above, to ensure for RNAi effectiveness.

Live brightfield and epifluorescence microscopy

Embryos were removed from gravid hermaphrodites in water or 1× egg buffer and mounted on 2% agarose pads under coverslips. Typically, 2 worms at a time were dissected, and several 2- and 4-cell embryos mounted; thus on average, the number of individual mothers sampled was about half the number of embryos presented in the data. For experiments scoring spindle positioning and PAR-2 domain formation, embryos were observed on an Olympus BX60 microscope equipped with PlanApo N 60X, 1.42 NA oil immersion objective lens, a CoolLED light source, a Hamamatsu Orca 12-bit digital camera, and MicroManager software (Strange et al. 2007; Hardin 2011).

To score P₁, EMS, and P₂ spindle positioning, images were acquired every 10 s with an exposure of 10 ms in bright field illumination. Focus was manually adjusted to follow centrosomes during each division. In strains expressing GFP::tubulin, acquisition was switched to 488 nm light at approximately nuclear envelope breakdown (NEB) of the ABa and ABp cells to more precisely follow the position of centrosomes and spindles in the EMS cell in mutants. The same conditions were used to visualize GFP::tubulin (10 ms, 488 nm) and PAR-2::mCherry (200 ms, 560 nm) localization, with the addition of fluorescence images taken at the start of the P₁ cell cycle, at NEB/metaphase, and at P₁ cytokinesis to score the P₁ and P₂PAR-2 domains. To visualize GFP::CED-10, single-plane images were acquired with an exposure of 100 ms of 488 nm light every 30 s.

For the mom-2(or42) single- and double-mutant experiments and embryos filmed as part of the endoderm specification assay (see below), embryos were prepared as above and imaged on an Olympus BX53 microscope equipped with PlanApo N 60X, 1.42 NA oil immersion lens, a Hamamatsu Orca Fusion BT camera, a SpectraX light engine, and motorized turret, all run by Olympus Cellsens software. Differential interference contrast (DIC) images were acquired every 10 s.

Scoring asymmetric division

Nuclear rotation and spindle orientation were scored by following centrosome movements in brightfield and/or via tubulin fluorescence, with the angle of the nuclear–centrosome complex defined based on a line through the 2 centrosomes relative to a line drawn on the anterior–posterior axis of the embryo. For the P₁ cell, rotation/orientation was scored as “normal” if the nuclear–centrosome complex rotated to within 30° of the anterior–posterior axis (=0°) prior to NEB, late if the spindle oriented after NEB but prior to cytokinesis onset, and failed if orientation onto the anterior–posterior axis did not occur. With either normal or late rotation, one centrosome was also close to the AB-P₁ cell contact after rotation. For the EMS cell, in most cells the centrosomes initially migrate onto the left–right axis or an oblique axis; in these cases, nuclear rotation aligns the centrosomes onto the anterior–posterior axis prior to NEB of the EMS cell (Bei et al. 2002; Zhang et al. 2007). In some cases, centrosomes migrate directly onto the anterior–posterior axis before NEB, and no rotation is needed for alignment. Previous reports did not identify any differences in centrosome migration between MES-1 and Wnt pathways mutants (Liro and Rose 2016). Because we could not follow the precise path of centrosome migration in all movies, especially in backgrounds without GFP::tubulin, we thus scored EMS spindle orientation as “normal” if the centrosomes were visible and aligned within 30° of the anterior–posterior axis prior to NEB. EMS spindle orientation was scored as “late” if alignment on the anterior–posterior axis occurred via spindle movements after NEB; in both cases, the posterior centrosome was closer to the EMS-P₂ contact at the end of an oriented EMS division. A “failed” EMS orientation was defined as a spindle that was oriented more than 45° off the anterior–posterior axis in any direction at the end of division; the daughter cells in this case were typically in different focal planes after division. For the P₂ cell, control GFP::tubulin embryos were filmed to determine the time at which close apposition of one centrosome to the EMS-P₂ cell contact occurred, which can occur through direct centrosome migration or a slight rotation of the nucleus. The nuclear–centrosome complex is thus oriented ∼45° relative to the anterior–posterior axis. This “centrosome cueing” occurred in all control embryos, and the time ranged from 60 s before EMS NEB to 40 s after NEB in this dataset. Centrosome position was scored as normal in mutant strains if centrosome cueing occurred within this time range. In control embryos, the spindle forms at this same position but then becomes more dorsal-ventrally oriented as division of EMS and P₂ proceeds.

Polarity was examined using mCh::PAR-2; GFP::tubulin expressing strains. The P₁ cell inherits PAR-2 around the entire periphery; as the cell cycle proceeds, the PAR-2 signal is greatly reduced on the anterior side of the cell including the AB-P₁ cell contact and thus PAR-2 is present in a posterior domain (Rose and Gonczy 2014; Koch and Rose 2023); embryos were scored as normal if a PAR-2 posterior domain was present by metaphase. Embryos in which PAR-2 was inherited abnormally from the first division (e.g. laterally in both AB and P₁) were excluded from analysis, as the P₁ cell would not be fated normally; this was observed in 1/11 control embryos. For the P₂ cell, a “normal” PAR-2 domain was restricted to the ventral half to three-quarters of the cell; that is, PAR-2 disappeared from at least the dorsal half of the ABp contact and the dorsal-most aspect of the outer posterior membrane of P₂, and this occurred before P₂ NEB (Arata et al. 2010). “Abnormal” P₂PAR-2 domains observed in both controls and ced-10 embryos included uniform presence of mCh::PAR-2 on the entire surface of P₂ and the absence of mCh::PAR-2 from the posterior cell surface only. P₂ nuclear–centrosome orientation in PAR-2 strains was scored as above, but we noted more variability in the time of P₂ centrosome cueing in this strain, with some control embryos cueing as late as EMS NEB + 160 s; ced-10 embryos were scored as normal if they exhibited cueing within this time range.

To determine if endoderm specification occurred in embryos, the presence of intestinal cells (an E cell derivative) were scored in embryos mounted on agar pads as outlined for imaging of early divisions above. After filming some embryos to confirm spindle positioning phenotypes (e.g. in RNAi conditions), these and sibling embryos on agar pads were incubated in a moist chamber at 16°C for 18–24 h, until control embryos had hatched. Polarization optics were then used to identify the presence of birefringent gut granules, which are a marker of intestinal differentiation (Laufer et al. 1980).

Immunostaining

Embryos from gravid hermaphrodite were dissected in phosphate buffered saline (PBS) solution on slides coated with poly-L-lysine, and the embryos were then permabilized using the freeze-crack method (Duerr 2006) except that liquid nitrogen was used. To stain SRC-1 dependent phosphotyrosine as in (Bei et al. 2002), slides were immediately fixed in prechilled methanol at −20°C for 5 min, followed by 5 min in ice-cold acetone, air-dried for 5–10 min at room temperature, and then rehydrated in PBS for 5 min. Slides were incubated in 50 μL of blocking solution [5% w/v bovine serum albumin (BSA) in PBS + 0.1% v/v Tween] in a moist chamber for 2 h at room temperature. Blocking solution was removed, and samples were incubated in 50 μL of primary antibody solution (Santa Cruz Biotech mouse monoclonal PY99 diluted 1:200 in PBS + 1% w/v BSA + 0.1% v/v Tween) overnight in a moist chamber at 4°C. Slides were washed 3× in PBS + 0.1% v/v Tween and then incubated in 50 μL secondary antibody solution (Alexa594-congugated goat monoclonal anti-mouse diluted 1:200 in PBS + 1% w/v BSA + 0.1% v/v Tween) in a moist chamber for 2 h at room temperature. After washing 3× in PBS + 0.1% v/v Tween, slides were incubated in 4′,6-Diamidino-2-phenylindole dihydrochloride (DAPI)/PBS, washed once in PBS; excess PBS was wicked away and samples were mounted with 20 μL of Vectashield under a 20 × 40 mm coverslip sealed with nail polish. To stain mNG::3xFLAG::MES-1, the same protocol was used, omitting the acetone fixation and air-drying steps, and using anti-FLAG Mab primary (Invitrogen), diluted 1:500. Images of samples were acquired in the mid-focal plane of the EMS-P₂ contact with an exposure of 200 ms of 488 nm light and 10 ms of 360 nm light on the Olympus BX60 microscope described above.

Live confocal microscopy

Embryos were removed from gravid hermaphrodites and mounted as above. SRC-1::GFP embryos were observed using the spinning disc module of an Intelligent Imaging Innovations (3i) Marianas SDC Real-Time 3D Confocal-TIRF microscope fit with a Yokogawa spinning disc head and EMCCD camera. Single mid-focal plane images were acquired using a 60× 1.4 numerical aperture oil immersion objective, 100% 488 nm laser power, an intensification 0, and a 500-ms exposure every 30 s. Acquisition was controlled by Slidebook 6 software. ARX-2::mNG imaging was performed on a Nikon Ti2 microscope equipped with a Yokogawa CSU-X1 confocal scanner unit and Hamamatsu ORCA-Flash4.0 LT3 Digital sCMOS camera. Single mid-focal plane images were acquired using a Nikon Plan Apo 60× 1.4 NA oil objective, with 100 ms exposure at 25% 488 nm laser power followed by 200 ms brightfield exposure, for every 10 s. The microscope was controlled by NIS Elements 5.42.

Analysis of cortical fluorescence intensity

All fluorescence quantification was carried out in ImageJ/FIJI using the raw image files. For determination of whole-cell intensities, the “Freehand line” tool was used to draw a line precisely around the entire 4-cell embryo, followed by the “Measure” function to obtain the mean fluorescence intensity within that area. For determining the intensities of cell contacts, the “Segmented Line” tool was used to trace a 2-pixel-wide line along the entire EMS-P₂ contact and other contacts, followed by the “Measure” function to obtain the mean intensity of the line. The same line traces were placed ∼10 pixels from the cell–cell contacts, avoiding the nucleus or centrosome, to sample mean cytoplasmic fluorescence intensity within the adjacent cells, and a line of similar length was used to measure background fluorescence on the slide away from the embryo. Measurements were assembled in Excel, and the background value was subtracted from all cortical and cytoplasmic measurements. Background-subtracted cell–cell contacts, or “cortical” intensities, were then divided by the mean background-subtracted cytoplasmic intensities of the 2 cells on either side of it, to produce a relative cortical intensity. The enrichment index was calculated as the ratio of the relative cortical intensity of the EMS-P₂ cell contact to that of the EMS-ABp contact in each embryo.

For mNG::MES-1, all 4 cell embryos were used; we did not observe any obvious differences in the relative cortical intensities at different stages of the cell cycle (as staged by DAPI staining). The EMS-P₂ and EMS-ABp cell contacts were measured; previous reports did not observe any signal at contacts other than EMS-P₂ (Berkowitz and Strome 2000). For Y99, only embryos from early prophase through metaphase were used, based our finding that control interphase embryos exhibited a weaker signal. The EMS-P₂ and EMS-ABp cell contacts were measured to calculate the enrichment index as previously done (Liro et al. 2018); in addition, we measured the ABp-P₂ cell contact since MES-1/SRC-1 signaling is used for both the EMS and P₂ divisions. Because CED-10, SRC-1 and ARX-2 have not been analyzed at the 4-cell stage, we measured all cell–cell contacts. For CED-10::GFP and SRC-1::GFP, qualitative analysis did not suggest strong asymmetry at the EMS-P₂ contact at any time in the cell cycle. Thus, measurements were made at NEB of the AB cells (visible by loss of exclusion of fluorescent signal from the nucleus) for CED-10::GFP or 2 min before EMS NEB for SRC-1::GFP (due to the nature of the signal, EMS NEB was more visible). These timepoints are within the time range where Y99 is asymmetric as noted above and when there is strong LIN-5 asymmetry at the EMS-P₂ contact (Heppert et al. 2018 our unpublished data).

Statistical analysis and figure compilation

Final statistics and graph creation were carried out using GraphPad Prism software version 10.4.1. Specific tests used are listed in figure legends, and specific P-values are provided in supplementary tables. Images were cropped and rotated in ImageJ. DIC, GFP tubulin, and PAR-2 images used in the figures, where levels of the fluorescence were not being quantified and were adjusted to best illustrate the phenotype being shown. For all fluorescence images in Figs. 3–5, raw images were used or were scaled the same. Images and graphs were imported into PowerPoint for arrangement into Figures.

Results

The Rac1 homolog CED-10 is a member of the MES-1/SRC-1 pathway for EMS spindle positioning

Previous work showed that embryos from mothers homozygous for a null allele of ced-10(t1875), hereafter referred to as ced-10 mutant embryos, exhibited defects in the division of the ABar cell of the 8-cell embryo; this same study reported that CED-10 is required for spindle positioning in EMS, but no data were shown for this or earlier divisions (Cabello et al. 2010). In a separate study, we found that ced-10 mutant embryos exhibit normal nuclear and spindle positioning movements during the division of the 1-cell embryo (Price et al. 2022). However, normal EMS division also requires the successful completion of the P₁ division to produce properly fated P₂ and EMS cells (Fig. 1a). Thus, here we first examined the P₁ division in ced-10 mutant embryos before proceeding to characterize the role of CED-10 in the EMS division. To facilitate scoring spindle positioning, we generated a strain with ced-10(t1875) in a GFP::tubulin background and compared embryos from this strain to control GFP::tubulin embryos.

In control embryos, the P₁ nuclear–centrosome complex rotated onto the anterior–posterior axis prior to NEB in all embryos (Fig. 1b). In comparison, 20% of ced-10; GFP::tubulin embryos had a late or failed P₁ spindle rotation (Fig. 1b; n = 25, 4 late, one failed; see also Supplementary Table 2). PAR polarity is required for spindle orientation and for the identity of P₁'s daughter cells, and so we next observed the localization of endogenously tagged mCherry::PAR-2 (mCh::PAR-2) in ced-10 and control strains that were also expressing GFP::tubulin. In the control strain, one 2-cell embryo had lateral PAR-2 domains in both cells (indicative of abnormal 1-cell division) and so was excluded from further analysis. Of the remaining control embryos, all established a normal PAR-2 domain in which PAR-2 was absent or present at only very low levels on the anterior-most membrane of the cell; these embryos exhibited timely P₁ nuclear rotation (n = 11, Fig. 1c). Late and failed P₁ nuclear rotations were again seen for ced-10 embryos in this mCh::PAR-2 background. Nonetheless, all embryos formed normally oriented PAR-2 domains, and the P₂ cell inherited PAR-2 around its entirety in embryos with normal or late rotation (n = 12, Fig. 1c). A small proportion of ced-10 embryos also inherited some PAR-2 in the EMS cell (1/12), but this was also seen in the control embryos (2/11). In ced-10 embryos in which the P₁ spindle oriented with normal or late timing, the EMS blastomere was larger than P₂ and divided before P₂, as in controls. Together these results suggest that CED-10 may have a subtle role in P₁ nuclear rotation, but in the vast majority of embryos, overall polarity appears normal and the P₁ cell divides asymmetrically.

We next examined the EMS division in ced-10; GFP::tubulin embryos compared with GFP::tubulin control embryos. In most control embryos, centrosomes migrate onto the left–right axis or an axis oblique to the anterior–posterior axis; subsequently, the nuclear–centrosome complex rotates onto the anterior–posterior axis. However, in some embryos, the centrosomes migrate directly onto the anterior–posterior axis (Bei et al. 2002; Zhang et al. 2007; Liro and Rose 2016). In all cases in controls, the centrosomes are aligned on the anterior–posterior axis prior to NEB (Fig. 2, a and b). Thus, in mutants, we scored orientation as normal if such alignment occurred prior to NEB and late if the spindle aligned after NEB. If the spindle set up and remained on any axis oblique to the anterior–posterior axis, orientation was scored as failed (also see ‘Methods’). We observed that 36% of ced-10 embryos had a late or failed EMS spindle orientation (Fig. 2, a and b). For the subset of embryos where P₁ and EMS rotation could both be scored, EMS defects were observed in the embryos with normal and late P₁ rotation at similar frequency (Fig. 2c; the single ced-10 embryo with a failed P₁ orientation was excluded). These results indicate that CED-10 plays a role in EMS nuclear rotation.

CED-10 was reported to act in a Wnt-dependent asymmetric division to orient the spindle of the ABar cell at the 8-cell stage (Cabello et al. 2010). We therefore scored spindle positioning defects in ced-10 mutants depleted of MES-1, expecting that if CED-10 contributes to Wnt signaling for EMS spindle positioning, ced-10; mes-1(RNAi) embryos would show a higher rate of complete failures of EMS orientation. However, the proportion of abnormal orientations in mes-1(RNAi); ced-10 embryos (46%) was not enhanced significantly compared to the ced-10 embryos (36%; Fig. 2b). Instead, the combination of RNAi depletion of MOM-2 (Wnt) and the ced-10 mutant background increased the overall proportion of abnormal EMS spindle positions to 74% and increased the rate of failed EMS spindle orientation events to 52%, compared with 9% for ced-10 alone (Fig. 2, a and b). As with ced-10 embryos, some ced-10; mom-2 embryos had abnormal P₁ spindle orientation (31.6%), but the failed orientation rate was low (4/19 late, 2/19 failed); defects in EMS spindle rotation were observed at a high rate in embryos with both normal and late P₁ rotation (Fig. 2c). These results suggest that CED-10 plays a role in the MES-1/SRC-1 pathway for EMS spindle positioning, in parallel to Wnt signaling.

CED-10 is not required for spindle positioning in the P₂ cell

The MES-1 protein is localized exclusively to the EMS-P₂ cell contact in wild-type embryos (Berkowitz and Strome 2000). The MES-1/SRC-1 signaling pathway, in addition to its importance in EMS spindle positioning and endoderm fate specification, is required for orienting PAR polarity and spindle positioning in the P₂ cell (Fig. 1a) (Berkowitz and Strome 2000; Bei et al. 2002). At birth, P₂ inherits the posterior PARs, such as PAR-2, uniformly around the cortex. Before division, new PAR domains form, but in a different orientation to the previous P lineage divisions (Fig. 3) (Arata et al. 2010; Rose and Gonczy 2014). Specifically, the “posterior” PARs disappear from the dorsal side of the cell and most of the ABp-P₂ cell contact, such that they are present in a ventral domain that includes the EMS-P₂ cell contact. The “anterior” PARs localize to the reciprocal dorsal side of the cell. The P₂ spindle orients along this new axis of PAR asymmetry. In mes-1(RNAi) or mutant embryos, reciprocal PAR polarity domains form in the P₂ cell but are misoriented in ∼70% of embryos; specifically, PAR-2 disappears from the EMS-P₂ and the ABp-P₂ cell contact in these embryos, resulting in a posterior PAR-2 domain. The P₂ spindle is oriented along this anterior–posterior axis in ∼50% of mes-1 null mutant embryos, the proportion being lower because spindle and polarity are uncoupled in some embryos; many divisions are also equal in terms of daughter cell size (Arata et al. 2010).

If loss of CED-10 disrupts MES-1 function, then asymmetric division of the P₂ cell should be affected. We first examined the P₂ division from the same GFP::tubulin movies used above. Because the focal plane of these movies was optimized for scoring the EMS division, the final orientation or equalness in the P₂ division was not always visible. Thus, we scored the initial orientation of the P₂ nuclear–centrosome complex. The P₂ nuclear–centrosome complex aligns with the PAR polarity axis through either centrosome migration or partial nuclear rotation where one centrosome becomes closely associated with EMS-P₂ cell contact (Berkowitz and Strome 2000; Arata et al. 2010); the nuclear–centrosome complex and the subsequent spindle are thus oriented initially at an ∼45° angle relative to the anterior–posterior axis (Fig. 3a). For simplicity, we refer to the close association of the centrosome with the cell contact as “centrosome cueing.” In control GFP::tubulin embryos, centrosome cueing occurred in all embryos, usually before EMS NEB (11/13; 2 embryos cued by EMS NEB + 40 s). Centrosome cueing also occurred in 100% of ced-10(t1875) embryos, within the same time range as controls (n = 9) (Fig. 3a). In comparison, in the majority of mes-1(bn7) embryos, the P₂ nuclear–centrosome complex did not cue to the cell contact (13/14); these embryos exhibited centrosomes oriented on the dorsal ventral axis (Fig. 3a) or on the anterior–posterior axis, as previously reported (Arata et al. 2010).

To further test for a mes-1 like polarity phenotype, we examined P₂ polarity in embryos expressing mCh::PAR-2 and GFP::tubulin. In the majority of control embryos and ced-10(t1875) mutant embryos, PAR-2 signal was lost from the ABp-P₂ cell contact and a PAR-2 domain was formed on the ventral side of the cell as expected (Fig. 3b). In one case for each, PAR-2 remained uniform, and in one case for each, PAR-2 disappeared from the posterior of the P₂ cell first. The latter is an opposite pattern to the defects noted above for mes-1 embryos. Thus, ced-10 does not affect the orientation of the PAR-2 domain in the P₂ cell. In this dataset, we again observed that in 100% of ced-10 embryos (n = 11), the P₂ nuclear–centrosome complex cued to the EMS-P₂ cell contact, within the same time range as controls cued (n = 10). Together these data indicate that polarity and orientation of the P₂ division are not affected by loss of CED-10.

We also examined the localization of MES-1. Although endogenously tagged mNeonGreen::3xFLAG::MES-1 (mNG::MES-1) is not visible in live embryos until the 8-cell stage (Heppert et al. 2018), anti-FLAG antibody staining showed the expected MES-1 localization at the EMS-P₂ cell contact (Fig. 3c). MES-1 was also localized at the EMS-P₂ contact in ced-10(RNAi) embryos and the intensity at this contact was enriched compared with that of the adjacent EMS-ABp contact (Fig. 3c; Supplementary Fig. 1). However, staining intensities of the entire embryo were lower in ced-10 embryos compared to controls (Fig. 3c); thus CED-10 may be required for overall levels of MES-1. Nonetheless, from these data it is clear that CED-10 is not required for MES-1 localization to the EMS-P₂ contact per se. In addition, the lower levels of MES-1 in ced-10 embryos do not appear to reduce the ability of MES-1 to direct P₂ polarity and spindle orientation.

CED-10 is required for normal levels of SRC-1-dependent phosphotyrosine staining and endoderm specification

We next set out to determine whether CED-10 affects SRC-1 activity or localization. Staining of 4-cell C. elegans embryos with the Y99 antibody, which recognizes phosphotyrosine, results in an enriched signal at the EMS-P₂ contact compared with other contacts, which is SRC-1 dependent (Bei et al. 2002) (Fig. 4a). MES-1 appears to activate SRC-1 at the EMS-P₂ contact, because mes-1 mutants show a loss of Y99 signal enrichment but still exhibit cell contact staining; note that even in wild type, there is variability and some embryos have only a slight enrichment at the EMS-P₂ contact, as previously reported (Bei et al. 2002) (Fig. 4a). Quantification of Y99 antibody staining showed that while many ced-10 embryos showed enrichment at the EMS-P₂ cell contact compared with EMS-ABp, the enrichment on average was lower than in wild type; the overall levels of Y99 in ced-10 embryos were not different from controls (Fig. 4a; Supplementary Fig. 1 and Table 4). Thus, loss of CED-10 affects the levels of SRC-1-dependent phosphorylation. The phosphotyrosine accumulation is presumed to be in both the EMS and P₂ cell, because SRC is required autonomously for the asymmetric division of both cells (Arata et al. 2010). However, due to the limitations of light microscopy, we cannot distinguish whether the reduction in Y99 signal in ced-10 embryos is specifically in the EMS cell, the P₂ cell, or both.

Whether the SRC-1 protein itself is enriched at that EMS-P₂ contact relative to other contacts in the 4-cell embryo has not been tested. We therefore examined the localization of SRC-1 using a strain with SRC-1 endogenously tagged with GFP (Zhu et al. 2020). SRC-1::GFP was cortically localized to all cell contacts at the 4-cell stage. There was a slight enrichment at the EMS-P₂ contact compared to the EMS-ABp contact (Fig. 4b). The cortical enrichment of SRC-1::GFP at the EMS-P₂ contact was similar in ced-10(RNAi) embryos (Fig. 4b). Further, the whole-cell intensities of SRC-1::GFP were not significantly reduced in ced-10 (RNAi) embryos (Fig. 4b) nor were the cortical intensities of SRC-1 GFP at cell–cell contacts (Supplementary Fig. 1c).

In addition to EMS spindle positioning, MES-1 and SRC-1 are required for endoderm fate specification and thus the formation of intestinal cells derived from the E daughter cell. Although loss of MES-1 or SRC-1 alone does not affect the formation of intestinal cells, loss of either enhances the variable intestine-minus phenotypes of Wnt pathway mutants (Bei et al. 2002). Since our results suggested that CED-10 affects SRC-1 activity at the EMS-P₂ contact, we therefore asked whether loss of CED-10 affects the formation of intestinal tissue in late-stage embryos. As in prior studies (Rocheleau et al. 1997; Thorpe et al. 1997; Bei et al. 2002), we found that mom-2(RNAi) embryos did not hatch but did form intestinal tissue as scored by the presence of gut granules, while the majority of mom-2(RNAi); mes-1(bn7) embryos did not form intestinal tissue (Fig. 4c). All ced-10(t1875) embryos had intestinal tissue, but a large proportion of ced-10(t1875); mom-2(RNAi) embryos did not exhibit gut granules (Fig. 4c). The percentage of ced-10; mom-2(RNAi) embryos exhibiting EMS spindle orientation defects in this dataset (without GFP::tubulin in the strain background) was similar (Fig. 4d), but slightly lower than for the experiments in Fig. 2. These data indicate that CED-10 is required for proper specification of endoderm tissue in parallel with Wnt signaling. Together with the results of the anti-phosphotyrosine staining, the data suggest that CED-10 acts upstream of both EMS spindle positioning and endoderm fate specification by promoting enhanced SRC-1 activity or the accumulation of SRC targets at the EMS-P₂ contact.

CED-10 and ARX-2 are localized at cell contacts at the 4-cell stage

To gain further insight into the function of CED-10 in MES-1/SRC-1 signaling, we characterized the localization of GFP::CED-10 in 4-cell embryos, using an integrated GFP::CED-10 transgene under transcriptional control by the ced-10 promoter (Ziel et al. 2009). This transgene was previously shown to rescue the defects of ced-10 (n1993), which is a viable allele of ced-10 (Lundquist et al. 2001). We crossed this transgene into the ced-10 null background and found that GFP::CED-10; ced-10(t1875) embryos appeared less round than ced-10 null embryos, and embryonic lethality was sufficiently rescued to maintain the strain homozygous for ced-10. In early embryos, GFP::CED-10 was present on all cell–cell contacts including at the 4-cell stage (Fig. 5a). The cell contact signal and the cytoplasmic signal were both reduced to background levels by RNAi of ced-10 (Supplementary Fig. 2). Although GFP::CED-10; ced-10(t1875) embryos exhibited many small cortical protrusions throughout the first few mitotic divisions, similar to ced-10(t1875) embryos (Price et al. 2022), GFP::CED-10; ced-10(t1875) embryos did not exhibit EMS division orientation defects. In addition, GFP::CED-10; ced-10(t1875); mom-2(RNAi) embryos did not show the higher rate of defective spindle orientations observed for ced-10(t1875); mom-2(RNAi) embryos examined in parallel (Supplementary Fig. 2a). These results indicate that GFP::CED-10 is functional in the early embryo for EMS spindle positioning. Quantification of GFP::CED-10 at the EMS-P₂ contact revealed a very slight enrichment at the EMS-P₂ cell contact compared with the EMS-ABp contact (Fig. 5c). The GFP::CED-10 localization pattern is consistent with a cortex-localized mechanism of action for CED-10 in EMS.

Rac proteins can signal through the WAVE complex or WASp proteins, which both activate the Arp2/3 complex to nucleate branched actin (Shakir et al. 2008; Saenz-Narciso et al. 2016). Therefore, we hypothesized that CED-10 contributes to EMS asymmetric division by promoting branched actin at the EMS-P₂ contact. Although prior studies have shown that branched actin reporters localize to the cell surface at the 1-cell stage and to cell–cell contacts during gastrulation and morphogenesis (Sawa et al. 2003; Roh-Johnson and Goldstein 2009; Chan et al. 2018), the localization of branched actin at the 4-cell stage has not been reported. We visualized ARX-2, the Arp2 member of the Arp2/3 branched actin nucleating complex, using an endogenously tagged ARX-2::mNG reporter (P. Zhang and B. Goldstein, personal communication). Consistent with our hypothesis, ARX-2::mNG was present at all cell–cell contacts of the 4-cell embryo. Further ARX-2::mNG was enriched at the EMS-P₂ contact relative to the EMS-ABp contact (Fig. 5b). The cortical enrichment of ARX-2::mNG at the EMS-P₂ contact was decreased in ced-10(t1875) embryos (Fig. 5b), and the relative cortical intensity of ARX-2::mNG at all cell contacts was lower in ced-10 embryos compared with the control (Supplementary Fig. 2). However, overall embryo intensities were not different in ced-10 embryos compared with controls (Fig. 5b). These results indicate that CED-10 is required to promote the normal cell contact accumulation of ARX-2 at the 4-cell stage and its higher enrichment at EMS-P₂ cell contact.

ARX-2 is required for EMS spindle orientation

To test whether ARX-2 and thus branched actin contributes to EMS spindle positioning, we first depleted ARX-2 by RNAi in GFP::tubulin;mom-5(zu193)(Frizzled null) embryos, one of the standard Wnt mutant backgrounds used in prior studies of EMS spindle positioning (Cabello et al. 2010; Liro and Rose 2016). Unexpectedly, we found that mom-5(zu193); arx-2(RNAi) embryos displayed a high rate of P₁ spindle rotation defects, with 56% completely failed rotations, compared with 8% for mom-5(zu193) and 7% for arx-2(RNAi) GFP::tubulin embryos (Supplementary Fig. 3a). EMS spindle positioning was then scored in only the embryos with a normal or late P₁ orientation, which would allow for normal fating of the EMS and P₂ blastomeres. In this subset, depletion of ARX-2 in a mom-5(zu193) background appeared to enhance the rate of defective spindle rotations observed in mom-5 alone (Supplementary Fig. 3a); the difference was not quite significant, potentially due to the small sample size, making it difficult to conclude whether ARX-2 is required for EMS spindle positioning. Failures of P₁ rotation also trended higher in ced-10(RNAi); mom-5 embryos compared to mom-5 or ced-10(RNAi) alone; however, we had not observed an increase in abnormal P₁ orientations in ced-10(t1875); mom-2(RNAi) embryos (Supplementary Fig. 3, b and c). Thus, we repeated RNAi of ARX-2 in embryos from the strong loss of function Wnt, mutant, mom-2(or42) (Thorpe et al. 1997). In these experiments, we did not observe a high level of P₁ spindle rotation defects for mom-2, mom-2; arx-2(RNAi) or mom-2; ced-10(RNAi) embryos (Supplementary Fig. 3). These results suggest the P₁ failure phenotype in the double-mutant combinations is due to an interaction between the Frizzled ortholog MOM-5 and the branched actin cytoskeleton.

Examination of mom-2(or42); arx-2(RNAi) embryos at the 4-cell stage revealed EMS spindle positioning defects that were enhanced relative to mom-2 mutant embryos (Fig. 5c). On the other hand, ced-10(t1875); arx-2(RNAi) embryos exhibited a low level of EMS late rotations similar to ced-10 single mutants (Supplementary Fig. 3d). These data show that ARX-2 is required for spindle positioning in the EMS cell in a pathway parallel to the Wnt pathway, and the results are consistent with CED-10 acting through ARX-2 to promote spindle orientation. The penetrance of EMS defects for arx-2(RNAi) embryos was lower than that observed for ced-10 mutants or ced-10(RNAi) both singly and in combination with mom-2 (Fig. 2; Supplementary Fig. 3). In addition, no enhancement of the endoderm defects of mom-2 mutants was seen in mom-2; arx-2(RNAi) embryos. The mom-2 mutant already has a high level of failure to produce gut cells, but nonetheless the mom-2 intestine-minus phenotype was enhanced by mes-1 RNAi and ced-10 RNAi (Fig. 5c; Supplementary Fig. 3e). These data raise the possibility that CED-10 acts through additional effectors to promote endoderm specification.

Discussion

In this study, we demonstrate that the C. elegans Rac protein CED-10 plays an important role in the asymmetric division of the EMS cell of the early embryo, which is regulated by partially redundant Wnt and MES-1/SRC-1 signaling pathways. Our data are consistent with CED-10 being a novel member of the MES-1/SRC-1 pathway, acting upstream or at the level of SRC-1. First, even though there were lower levels of MES-1 in ced-10 embryos, MES-1 was still present at the EMS-P₂ contact site. Further, ced-10 embryos did not exhibit the defects in PAR-2 polarity and spindle orientation that are known to result from loss of MES-1. Thus, although we cannot rule out a redundant role in the P₂ division, CED-10 does not appear to be essential for MES-1 activity. At the same time, ced-10 embryos did show reduced enrichment of SRC-dependent phosphotyrosine signal at the EMS-P₂ contact. The antibody used for this immunostaining experiment recognizes phosphorylated tyrosine and is a readout for both SRC-1 autophosphorylation and phosphorylation of its targets (Bei et al. 2002). A GFP-tagged version of SRC-1 was present at all cell–cell contacts and showed only a slight enrichment at the EMS-P₂ contact compared to the EMS-ABp contact. In addition, SRC-1 localization at cell contacts was not reduced by ced-10(RNAi). These data argue against a primary role in the asymmetric recruitment of SRC-1 itself to the EMS-P₂ cell contact. Instead, we propose that CED-10 promotes the activity of SRC-1 or the recruitment or retention of SRC-1 targets at the EMS-P₂ contact. A detailed analysis of this role will require future work uncovering direct targets of SRC-1 in EMS, since only one target in the early embryo has thus far been proposed (Sumiyoshi et al. 2011).

Our studies also reveal that ARX-2 and thus branched actin are required for the asymmetric division of the EMS cell. The defects in spindle positioning in arx-2(RNAi) embryos were enhanced by loss of MOM-2, similar to what was observed for ced-10 embryos. In addition, ARX-2::GFP was present at higher levels at all cell contacts in controls compared with ced-10 embryos. Thus, CED-10 appears to act through the generation of branched actin to promote spindle positioning in the EMS cell. Interestingly, drug inhibition experiments on 4-cell embryos showed that actin is required for endoderm specification (Goldstein 1995). However, we did not uncover a role for ARX-2 in endoderm specification in this study. It is possible that endoderm specification is less sensitive to the loss of arx-2 by RNAi than is spindle positioning. Although our RNAi conditions resulted in a strong loss of function phenotype based on the appearance of cortical blebs, lethality was variable indicating that depletion of ARX-2 was incomplete (see “Methods”). In addition, the frequency of defects in arx-2(RNAi) and mom-2; arx-2(RNAi) embryos was lower than that observed in ced-10 embryos, and this may also be due to incomplete RNAi. Alternatively, Rac is known to have several downstream effectors and to exhibit crosstalk with other actomyosin regulators (Bustelo et al. 2007; Saenz-Narciso et al. 2016; Nguyen et al. 2018), and thus it is possible that CED-10 acts through other factors during the EMS division, especially for endoderm specification. For example, nonmuscle myosin (NMY-2) is a component of the MES-1/SRC-1 pathway. Using a temperature-sensitive mutant, it was shown that NMY-2 activity is needed for the normal enrichment of Y99 staining at the EMS-P₂ contact and for endoderm specification, but no defects in spindle orientation were identified (Liu et al. 2010).

The finding that both Rac and ARX-2 regulate spindle positioning in the EMS cell indicates that the actin cytoskeleton is important for spindle positioning in the EMS cell. The role of actomyosin flows in establishing polarity in the 1-cell embryo is well characterized, and loss of ARX-2 or its regulators has been shown to have mild effects on PAR polarity at this stage (Xiong et al. 2011; Shivas and Skop 2012). Actomyosin flows have also been implicated in polarity and spindle positioning at the 2-cell stage, and actomyosin flows have been visualized late in the EMS cell (Singh and Pohl 2014; Caroti et al. 2021; Koch and Rose 2023; Ng et al. 2023). However, it seems unlikely that CED-10 and ARX-2 promote spindle orientation in the EMS cell through actomyosin cortical flows that polarize the PAR proteins. In the EMS cell, the spindle aligns on the anterior–posterior axis, while the PAR proteins form inner/outer domains in this cell (Rose and Gonczy 2014). In addition, the cortical flows identified occur late during the EMS division and are dependent on the Wnt pathway (Caroti et al. 2021), while CED-10 and ARX-2 act in parallel to Wnt signaling to promote spindle orientation. We favor a model in which branched actin promotes the recruitment or activity of force-generating motors to the cortical cytoskeleton. Both LIN-5 (NuMA and Mud in vertebrates and Drosophila, respectively) and dynein play a role in spindle orientation in the EMS cell (Zhang et al. 2007; Liro and Rose 2016). In other systems, actin regulators such as Afadin and ERM proteins can influence the localization of NuMA. In addition, in cultured cells, myosin-10 and the ARX-2 ortholog Arp2 orient the spindle in response to specific substrate adhesion patterns in a pathway that appears to act in parallel to dynein (di Pietro et al. 2016; Kotak 2019). It will be interesting to determine if any of these other cytoskeletal components play a role in the EMS division of C. elegans. It will also be important to use high resolution microscopy or other methods to determine the precise localization of LIN-5 and ARX-2 in the EMS cell. It is still not clear if the activity or the localization of LIN-5 is regulated by the MES-1 and Wnt pathways in the EMS cell; due to the limitations of light microscopy, one cannot determine whether proteins enriched at the EMS-P₂ cell contact are actually asymmetrically localized at the EMS cell cortex in the intact embryo (Liro and Rose 2016; Heppert et al. 2018). Thus, while ARX-2 could promote localized branched actin networks important for spindle orientation, it could also act more generally at the cell cortex.

Our experiments also uncovered an apparent genetic interaction between branched actin and the Frizzled ortholog MOM-5 at the 2-cell stage. In the absence of MOM-5, branched actin depletion with RNAi against arx-2 or ced-10 caused a high rate of failed P₁ spindle rotations. Previous work implicated the Frizzled/MOM-5 protein and the adenomatous polyposis coli homolog APR-1 in the regulation of microtubule dynamics during mitotic spindle positioning of the 1-cell C. elegans embryo. This mechanism is dependent on PAR polarity but independent of the MOM-2 ligand (Sugioka et al. 2011, 2017). As the P₁ cell divides asymmetrically in a PAR-dependent manner, it is possible that Frizzled and APR-1 participate in a similar mechanism of regulating microtubule dynamics for P₁ spindle rotation as well. CED-10 and ARX-2 would act in a parallel pathway, based on the genetic enhancement data. Such a parallel pathway could be for either spindle positioning or PAR polarity itself. Actin flows in the AB cell have been implicated in promoting midbody positioning and P₁ spindle orientation (Singh and Pohl 2014), and it has recently been shown that PAR polarity in the P₁ cell is established through redundant mechanisms, at least one of which involves the actomyosin skeleton (Koch and Rose 2023; Ng et al. 2023).

Prior work characterized a role for Rac1/CED-10 in spindle positioning of the ABar cell of the 8-cell C. elegans embryos (Cabello et al. 2010). Wnt signaling instructs the mitotic spindle of ABar to set up at a perpendicular angle to the orientation of the 3 other AB cell spindles. This ABar spindle orientation pathway, like the Wnt-dependent pathway for EMS spindle positioning, is transcription independent (Schlesinger et al. 1999; Walston et al. 2004). In the ABar cell, Wnt signaling appears to be the predominant mechanism because mutations in the Wnt components cause failure of spindle orientation with high penetrance. Further, overexpression of CED-10 resulted in slight rescue of Wnt mutant spindle orientation defects, and genetic analysis showed that CED-10 acts downstream of Wnt signaling for cell-corpse clearance in the later embryo and distal tip cell migration in the larvae (Cabello et al. 2010). In contrast, our analyses show that CED-10 is a new component of the MES-1/SRC-1 pathway for both endoderm specification and spindle positioning in the EMS cell, rather than part of the Wnt pathway. SRC-1 has also been shown to contribute to ABar spindle positioning, but its precise role is unknown, and MES-1 does not appear to be present in the AB cells. Together these observations and our findings raise the possibility that SRC-1, CED-10 and ARX-2 work together in the ABar cell to promote spindle position, but downstream of Wnt signaling.

The data presented here on the role of CED-10 in the EMS division are very similar to those reported previously for the PIG-1 kinase. PIG-1 is a PAR-1 related kinase involved in the asymmetric division of several cell types in C. elegans (Cordes et al. 2006; Liro et al. 2018). PIG-1 has partially redundant roles with PAR-1 in the 1-cell embryo, but in the EMS division, genetic analyses placed PIG-1 in the MES-1/SRC-1 pathway while PAR-1 appears to act in the Wnt pathway. Loss of PIG-1 resulted in defects in both endoderm specification and spindle positioning, as seen here for loss of CED-10 (Liro et al. 2018). Further studies are required to determine the mechanistic links among SRC-1, CED-10, ARX-2, and PIG-1 during this asymmetric division in C. elegans, and it will be interesting to learn if this mechanism is utilized by other cells in C. elegans or in other organisms.

Data availability

Data necessary for confirming the conclusions of this study are present within the manuscript text or figures and supplementary material; this research did not generate large datasets, but raw data and images are available upon request. Strains are available upon request from the corresponding author or the CGC.

Supplemental material available at GENETICS online.

Acknowledgments

We thank members of the Rose and McNally labs for helpful discussions and Drs Daniel A. Starr, Bruce W. Draper, and Jennifer Heppert for experimental advice. We are grateful to the labs of Drs Bob Goldstein, Erik Lundquist, Frank McNally, Karen Oegema, Guangshuo Ou, and Martha Soto for strains and plasmids. We thank the MCB Light Microscopy Imaging Facility, which is a UC-Davis Campus Core Research Facility, for the use of these microscopes.

Funding

Many strains were obtained from the Caenorhabditis Genetics Center, which is funded by the National Institutes of Health Office of Research Infrastructure Programs (P40 OD010440) for strains. The 3i Marianas spinning disk confocal and the Nikon laser scanning confocal used in this study were purchased using a National Institutes of Health Shared Instrumentation grant (1S10RR024543-01). This research was funded by awards from the National Institutes of Health grant (R01GM68744), the National Institute of Food and Agriculture (CA-D*-MCB-6239-H), and the UC Davis Office of Research Bridge Funding Program to LSR. Additional support for HL and MJL was provided by the UC Davis BMCDB Graduate Group and a National Institutes of Health training grant to HJL (T32 GM 007377).

Literature cited

Author notes