F-actin organization excludes E-cadherin from the division furrow to ensure cytokinesis fidelity
Debodyuti Mondal, Megha Rai, Anup Padmanabhan
TL;DR
This study shows how actin filaments prevent E-cadherin from entering the cell division zone, ensuring accurate cell division during embryonic development.
Contribution
The paper reveals a novel mechanism where actin filaments physically exclude E-cadherin from the division furrow to ensure cytokinesis fidelity.
Findings
CYK-1/formin-polymerized actin filaments restrict HMR-1/E-cadherin mobility into the cytokinetic furrow.
Disrupting cortical actin leads to HMR-1 accumulation in the furrow and increased cytokinesis failure.
Co-depletion of HMR-1 rescues cytokinesis failure caused by actin disruption.
Abstract
Localization of the cell adhesion receptor E-cadherin within the furrow is associated with reduced cytokinetic fidelity. Cortical organization of formin-polymerized linear actin filaments at the division zone physically associates with E-cadherin and restricts its lateral mobility into the cytokinetic furrow. Maintaining intercellular adhesion during proliferative cell division is essential for ensuring tissue integrity during embryonic development. How these two processes are coordinated in vivo remains poorly understood. During early Caenorhabditis elegans embryonic divisions, the cell adhesion receptor, HMR-1/E-cadherin, is absent from the cytokinetic furrow zone and the newly established cell–cell interface. We investigated mechanisms underlying this spatial exclusion of HMR-1 from the furrow zone. We find that crosslinking and compact alignment of CYK-1/formin-polymerized…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure S1
Figure 2
Figure S2
Figure 3
Figure S3
Figure 4
Figure S4
Figure 5
Figure S5
Figure 6
Figure S6- —Wellcome Trust DBT India Alliance (India Alliance)http://dx.doi.org/10.13039/501100009053
- —DST | Science and Engineering Research Board (SERB)http://dx.doi.org/10.13039/501100001843
- —NIH Office of Research Infrastructure Programshttp://dx.doi.org/10.13039/100016958
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsGenetics, Aging, and Longevity in Model Organisms · Cellular Mechanics and Interactions · Microtubule and mitosis dynamics
Introduction
Eukaryotic cytokinesis relies on the constriction of the actomyosin-based contractile ring to partition the segregating genetic material. Assembly of the cytokinetic apparatus in animal cells is initiated upon anaphase onset, after the activation of the small GTPase RhoA (Fededa & Gerlich, 2012; Sugioka, 2022). Active RhoA is limited to a narrow zone at the medial cortex through a combination of positive and inhibitory signals from astral microtubules and the mitotic spindle (Chircop, 2014; Basant & Glotzer, 2018; Sugioka, 2022). Localized RhoA activation triggers the recruitment and assembly of key components of the cytokinetic ring, including diaphanous formin, type II nonmuscle myosin, anillin, septins, and actomyosin-binding and regulatory proteins (Severson et al, 2002; Matsumura, 2005; Maddox et al, 2007; Liu et al, 2012; Basant & Glotzer, 2018; Sugioka, 2022). Formin promotes the nucleation and polymerization of unbranched actin filaments, a key structural component of the cytokinetic ring (Swan et al, 1998; Otomo et al, 2005; Kühn & Geyer, 2014). Activation of nonmuscle type II myosin is mediated via RhoA-associated kinase (ROCK), enabling myosin head domain to bind and crosslink actin filaments while translocating them through ATP hydrolysis (Osório et al, 2019; Sobral et al, 2021). In addition to core actomyosin components, several actin-binding proteins (ABDs) are recruited to the medial cortex. These include crosslinkers, bundling proteins, nucleation promotion factors, and filament severing proteins that regulate the constriction of circumferentially assembled contractile actomyosin ring (Fededa & Gerlich, 2012). Although the motor activity of type II myosin is the primary force generator driving cytokinesis in many cell types (Green et al, 2012; D’Avino et al, 2015; Osório et al, 2019), evidence for myosin-independent cell division in certain contexts has led to alternative models of furrow constriction (Lord et al, 2005; Ma et al, 2012; Mendes Pinto et al, 2012). During early embryonic divisions in Caenorhabditis elegans, actomyosin enrichment in the equatorial cortex establishes a contractility gradient with its maxima at the furrow zone (Mayer et al, 2010). This results in “ring-directed” cortical flows from the poles, proposed to enable embryonic asymmetry, as well as the assembly and constriction of the cytokinetic actomyosin ring (Reymann et al, 2016; Khaliullin et al, 2018; Hsu et al, 2023; Ng et al, 2023).
Although significant progress has been made in understanding cytokinesis through studies in isolated individual cells, cell division in metazoan systems presents additional complexity because of cell adhesion receptors on the surfaces of dividing cells. The best-understood transmembrane adhesion receptor, E-cadherin, is essential for embryonic development and epithelial tissue integrity in metazoans (Costa et al, 1998; Harris & Tepass, 2010; Bulgakova et al, 2012; Chihara & Nance, 2012; Ninomiya et al, 2012; Takeichi, 2014). E-cadherin–mediated adhesion requires receptors expressed on the cell surface to interact in-trans through their extracellular domains. Adaptor proteins—β-catenin and ⍺-catenin—are recruited to the cytoplasmic domain of E-cadherin. Cortical actin associates with the ⍺-catenin, establishing a physical connection between E-cadherin–mediated cell adhesion and the actin cortex (Bertocchi et al, 2017). Although E-cadherin is predominantly localized to cell junctions mediating intercellular adhesion, a significant population of E-cadherin is also found at nonjunctional surfaces (Wu et al, 2015; Padmanabhan et al, 2017). Both junctional and nonjunctional E-cadherin receptors recruit an assemblage of ∼150 proteins (“cadhesome”) to their cytoplasmic domains (Zaidel-Bar, 2013; Guo et al, 2014; Bertocchi et al, 2017). Notably, a significant fraction of these components overlap with the cytokinetic machinery, suggesting an intricate sharing of components between cell adhesion and cell division processes (Padmanabhan et al, 2015). In addition to carrying out intercellular adhesion, these receptors interact with cortical actomyosin, modulating cell division dynamics even in isolated cells devoid of cell–cell contacts (Padmanabhan et al, 2017). Accordingly, E-cadherin has been shown to regulate the plane and pace of cytokinetic furrow ingression in mammalian cells, Drosophila epithelia, and C. elegans embryos (den Elzen et al, 2009; Founounou et al, 2013; Guillot & Lecuit, 2013; Herszterg et al, 2013; Morais-De-Sá & Sunkel, 2013; Padmanabhan et al, 2017; Pinheiro et al, 2017; Derksen & van de Ven, 2020). Thus, spatiotemporal remodeling of surface-localized E-cadherin machinery is necessary to resolve any competition between cell adhesion and cell division complexes for a shared pool of components and to ensure accurate cell divisions during proliferation and developmental morphogenesis (Founounou et al, 2013; Guillot & Lecuit, 2013; Herszterg et al, 2013; Takeichi, 2014; Pinheiro et al, 2017). The assembly of the actomyosin ring during early embryonic divisions in C. elegans provides a powerful system to investigate the crosstalk between cell adhesion and cell division machinery. Previous studies showed that HMR-1, the sole E-cadherin ortholog in C. elegans, associates with actin filaments during cortical flows and inhibits RHO-1–mediated type II myosin/NMY-2–dependent contractility. Embryos depleted of HMR-1 divided faster, indicating that HMR-1 slows down cytokinetic furrow ingression (Padmanabhan et al, 2017). Intriguingly, the medial zone was largely devoid of HMR-1 during ring assembly and furrow ingression. Similar observations were made for DE-cadherin during epithelial divisions in Drosophila (Pinheiro et al, 2017). The absence of E-cadherin in the furrow zone during cytokinesis remains unexplained.
Our study reveals that during early embryonic divisions in C. elegans, orthogonally aligned actin filaments polymerized by formin/CYK-1 at the medial cortex and crosslinked by PLST-1 and NMY-2 corral the contractile furrow zone. Cortical flows transport surface-localized HMR-1 clusters toward the furrow, where they associate via the catenin HMP-1, with the compactly aligned cytokinetic actin architecture. This association impedes further movement of HMR-1 into the furrow. Disrupting this actin organization in the cytokinetic ring allows HMR-1 to localize into the cytokinetic zone, contributing to the slowdown of furrow closure and cytokinesis failure. These findings demonstrate a reciprocal inhibitory regulation between E-cadherin/HMR-1–mediated cell adhesion and actomyosin-based cell division machinery.
Results
Surface-localized HMR-1 is excluded from the division zone during cytokinesis
To characterize HMR-1 dynamics at the site of actomyosin ring assembly and furrow ingression, we imaged the surface plane during the first (P0 cell) and second (AB cell) cytokinetic divisions in embryos co-expressing GFP fused to the C terminus of HMR-1 (HMR-1::GFP) and mCherry fused to the plasma membrane marker, PLC1δ-PH (mCherry::PH), or type II nonmuscle myosin, NMY-2 (NMY-2::mCherry). Consistent with previous reports, HMR-1::GFP formed discrete clusters at nonjunctional surfaces (Figs 1A and B and S1A; yellow arrows and Video 1), and at intercellular boundaries (Fig S1A; cyan arrows, Fig 1C and D; time = 140 s; yellow arrow) (Padmanabhan et al, 2017). Despite their presence at noningressing regions, HMR-1 was largely excluded from the contractile division zone during cytokinetic ring assembly in P0, AB, and P1 cells (Figs 1A and B and S1A; yellow arrowhead in enlarged ROI [region of interest]). Notably, the double bilayer formed by the plasma membrane invagination also lacked detectable HMR-1 during cytokinesis (Figs 1C and D and S1B, yellow arrowheads) (Padmanabhan et al, 2017). However, faint HMR-1::GFP signals could be detected toward the conclusion of cytokinesis (Fig 1C; time = 100 s, and Fig 1D; time = 60 s; yellow arrowhead). HMR-1 localization intensified at the newly formed interface between daughter cells only after the membrane ingression was complete (Fig 1C and D; time = 140 s; yellow arrow, and Video 2).
Cytokinetic furrow zone is devoid of surface-localized HMR-1.(A, B) Cell surface view of furrow ingression in one-cell (P0)-stage embryos co-expressing HMR-1::GFP, and (A) mCherry::PLC1δ-PH and (B) NMY-2::mCherry. Yellow arrows indicate surface-localized nonjunctional clusters of HMR-1::GFP. HMR-1 is largely absent (yellow arrowhead) in the medial furrow zone, enlarged as the region of interest (ROI). Scale bar—10 μm (see Video 1). (C, D) Time series of equatorial plane of an embryo co-expressing NMY-2::mCherry, HMR-1::GFP, and Histone::mCherry while undergoing (C) P0 cell division (first division) and (D) AB cell division (second division). Whereas NMY-2 is enriched at the edge of the ingressing membrane (cyan arrows), this membrane region is devoid of HMR-1 (indicated by yellow arrowheads). After cytokinesis completion (140 s), HMR-1 is detected at the newly formed cell–cell boundary (yellow arrows) (see Video 2 and Video 3; control). (E) Enlarged ROI from the time series described in Fig 1D, showing HMR-1 dynamics at the boundary between AB and P1 cells during AB division. Yellow arrows denote junctional HMR-1 that progressively disappears from the site of cytokinetic ring constriction, resulting in the tricellular junction lacking HMR-1 (yellow asterisks). Cyan arrows indicate that NMY-2 is detected at these junctions. Bottom: plots corresponding to the time series show the line-scan intensities of HMR-1 and NMY-2 along the cell junction between AB and P1 cells, showing depletion of HMR-1 (marked as a yellow asterisk) during AB ring constriction (see Video 3; control). (F) Equatorial view of the furrow zone during cytokinesis in P0 (top row) and AB (bottom row) cells expressing HMR-1::GFP, PEZO-1::mScarlet, or SAX-7::GFP. In contrast to the absence of HMR-1 in furrows (indicated by arrowheads), PEZO-1 and SAX-7 are detected in ingressing furrows (yellow arrows).
(Related to this Figure).(A) Cell surface view of dividing AB cell and P1 cell in embryos co-expressing HMR-1::GFP and mCherry::PLC1δ-PH. HMR-1 clusters localize to nonjunctional cell surfaces (yellow arrow) and at cell–cell junctions (cyan arrows). The yellow arrowhead in the enlarged ROI shows the absence of HMR-1::GFP in the ingressing furrow zone. Scale bar—10 μm. (B) Medial time-lapse images of cytokinetic membrane ingression during AB cell and P1 cell division in embryos co-expressing NMY-2::mCherry, HMR-1::GFP, and Histone::mCherry. The yellow arrowhead shows the absence of HMR-1::GFP in the ingressing furrow. HMR-1::GFP localizes to the intercellular boundary upon completion of furrow ingression. (C) Time-lapse images of cytokinetic membrane ingression during the P0 and AB cell divisions in embryos expressing endogenously tagged HMR-1::GFP. Similar to cytokinesis in embryos expressing transgenic HMR-1::GFP, the P0 and AB cell divisions in endogenously tagged embryos did not localize HMR-1 during furrow ingression (yellow arrowhead), but only after the completion of cytokinesis (arrowhead). (D) Time-lapse images showing the surface plane during P0 furrow ingression in embryos co-expressing SAX-7::GFP and Lifeact::RFP. SAX-7 is detected at the sites of cell division and during furrow ingression; clusters of SAX-7 are seen moving in (yellow arrowheads) (see Video 4).
Video 1Surface-localized HMR-1 clusters are excluded from contractile furrow zone. The cortical plane of P0-stage embryos expressing HMR-1::GFP, the membrane marker mCherry::PH. Three representative clusters tracked (arrowheads—cyan/white/yellow) during furrow ingression show minimal furrow-directed movement. Frame rate—10 fps. Download video
Video 2HMR-1 localizes to the newly formed cell boundary only after cytokinesis completion. Equatorial ROI from the medial plane of P0 embryos expressing HMR-1::GFP, NMY-2::mCherry, and Histone::mCherry, from metaphase to cytokinesis completion (white arrowheads in the HMR-1 channel). Frame rate—3 fps. Download video
We next asked whether junctional HMR-1, like nonjunctional HMR-1 clusters, is excluded from the cytokinetic zone. To address this, we analyzed the AB-P1 cell boundary during AB cytokinesis. During actomyosin ring assembly, medially localized NMY-2 in the AB-P1 cell boundary (cyan arrow in Fig 1E) co-localized with junctional HMR-1 (Fig 1E; time = “t_0_,” cyan arrow). As the ring began to constrict, HMR-1 intensity progressively diminished at the ring assembly site along the AB-P1 boundary (Fig 1E; time = +60 s onward, yellow asterisks). In contrast, HMR-1 levels remained stable at junctional regions adjacent to the furrow ingression site (Fig 1E; time = +70 s, +80 s, and +140 s; yellow arrows). The tricellular junction formed after AB cell ring constriction was devoid of HMR-1 but enriched in NMY-2 (Fig 1E; time = +140 s, yellow asterisk and cyan arrow; Video 3, control). To rule out the potential artifacts from transgenic HMR-1 expression driven by the mex-5 promoter, we examined embryos expressing GFP-tagged HMR-1 from its endogenous locus (Marston et al, 2016). Similar to the transgenic reporter, endogenously tagged HMR-1 failed to localize to ingressing membranes during P0 and AB cell divisions and was absent from tricellular junction during AB cytokinesis (Fig S1C). To determine whether this exclusion from the division zone was specific to HMR-1 or represented a general feature of transmembrane proteins, we imaged embryos expressing mScarlet-tagged PEZO-1 (mScar::PEZO-1), a multipass calcium-gated mechanosensory protein, and GFP-tagged SAX-7 (SAX-7::eGFP), an L1-IgCAM cell adhesion receptor (Chen et al, 2001; Wang et al, 2005; Bai et al, 2020). In contrast to HMR-1, both PEZO-1 and SAX-7 localized to ingressing membranes throughout cytokinesis in P0 and AB cells (Figs 1F and S1D, yellow arrowhead; Video 4) (Chen et al, 2001). These observations demonstrate that HMR-1 is specifically excluded from the medial cortex and ingressing furrow during cytokinesis.
Video 3Depletion of junctional HMR-1 in the furrow zone is disrupted in cyk-1(RNAi) embryos. Equatorial ROI from the medial plane of control and cyk-1(RNAi) embryos undergoing AB division and expressing HMR-1::GFP, NMY-2::mCherry, and Histone::mCherry. HMR-1 is enriched at the junctions between AB and P1 cells. In control embryos, HMR-1 is depleted from the site of ring assembly (yellow asterisk). In mild cyk-1(RNAi) embryos, HMR-1 is accumulated in the ingressing membrane surface and the site of ring assembly at the AB-P1 cell interface (yellow arrows). Frame rate—3 fps. Download video
Video 4Surface-localized SAX-7 localizes to the medial cytokinetic zone during furrow ingression. The cortical plane of the medial cytokinetic region from P0-stage embryos expressing SAX-7::GFP and Lifeact::mCherry. Clusters of SAX-7::GFP are seen moving unhindered into the cytokinetic zone (cyan arrowhead) and into the furrow. Download video
Localization of HMR-1 in the division furrow is regulated by RHO-1 and NMY-2
Previous work showed that depletion of HMR-1 resulted in elevated levels of active RHO-1, the C. elegans ortholog of RhoA GTPase, and NMY-2 at nonjunctional surfaces in early embryos (Padmanabhan et al, 2017). To test the possibility of reciprocal inhibition—whereby RHO-1 or NMY-2 might restrict HMR-1 dynamics at the division zone, we quantified the levels of HMR-1::GFP and NMY-2::mCherry in ingressing cytokinetic membrane of embryos partially depleted for RHO-1 (Fig 2B and C) or NMY-2 (Fig 2D and E). Complete loss of RHO-1 activity severely disrupted germline development and prevented furrow ingression. However, a mild (partial) depletion of rho-1(RNAi) or nmy-2(RNAi) resulting in ∼50% reduction in NMY-2::mCherry intensity allowed slower cytokinetic ring constriction. Whereas ingressing membranes in control embryos displayed negligible HMR-1::GFP signal, partial depletion of RHO-1 or NMY-2 led to an ∼4-fold accumulation of HMR-1 (normalized to HMR-1::GFP levels in control) (yellow arrows in Fig 2B and D, HMR-1::GFP fold change in Fig 2C and E). A similar accumulation of HMR-1 was observed when temperature-sensitive nmy-2(ne3409) (hereafter nmy-2(ts)) embryos were shifted to the restrictive temperature (25°C) before anaphase onset (Fig S2A) (Liu et al, 2010). This allele has been proposed to impair NMY-2 motor activity at 25°C. We next examined whether perturbation of NMY-2 regulators affected HMR-1 localization. Partial depletion of myosin light chain/MLC-4 or RHO kinase/LET-502 resulted in delayed cytokinesis, confirming reduced NMY-2 activity (Fig 2H and I). However, in contrast to nmy-2(RNAi), neither let-502(RNAi) nor mlc-4(RNAi) embryos exhibited detectable HMR-1 accumulation in the furrow or ingressing membranes (Fig 2F; yellow arrowhead, Figs 2G and S2A and B). Similar results were obtained upon temperature upshift of let-502(sb106) embryos expressing HMR-1::GFP and mCherry::PH (Fig S2D and E). Together, these findings indicate that distinct perturbations of the actomyosin machinery have differential effects on HMR-1 localization.
*HMR-1 localization in the furrow is regulated by RHO-1 and NMY-2 levels.(A) Schematic showing the ROI (highlighted in yellow rectangle) during P0 division that is used to measure the fluorescence intensities of HMR-1::GFP (green) and NMY-2::mCherry (magenta) along the ingressing furrow. The pixel with maximum NMY-2 intensity is considered the edge of the furrow and therefore taken as the center of ROI. (B, D, F) Medial plane images of the cytokinetic furrow during P0 division in embryos co-expressing NMY-2::mCherry, HMR-1::GFP, and Histone::mCherry in control, (B) rho-1(RNAi), (D) nmy-2(RNAi), and (F) mlc-4(RNAi) embryos. Accompanying plots show the mean ± 95% CI of two-color line scans of NMY-2::mCherry (magenta) and HMR-1::GFP (green) intensities measured along the ROI in furrows. The yellow arrow in (B, D) indicates the accumulation of HMR-1::GFP in the furrows of rho-1(RNAi) and nmy-2(RNAi) embryos. (F) Depletion of MLC-4 does not lead to HMR-1 accumulation in the furrow (yellow arrowhead). (C, E, G) Plots showing normalized fold change in HMR-1::GFP and NMY:: mCherry fluorescence intensities in ingressing surface of (C) rho-1(RNAi) (n = 10), (E) nmy-2(RNAi) (n = 19), and (G) mlc-4(RNAi) (n = 8) embryos. Error bars denote the mean ± 95% CI of intensity normalized to the mean value of control embryos. Control embryos therefore show a mean fold increase of 1. (H) Plot showing the time taken for ring constriction from anaphase onset to completion of cytokinesis in control and mlc-4(RNAi) embryos analyzed in Fig 2F and G (n = 8, **P < 0.0001, two-tailed, unpaired, nonparametric, Mann–Whitney test). (I) Fold change in HMR-1::GFP intensity at the invaginating membrane is plotted against time taken for completion of P0 cytokinesis. Data points indicating specific RNAi conditions are grouped based on the extent of HMR-1 fold change in the furrow—minimal change (red); moderate increase (green); and maximal increase (blue). ROI with crowded data points in the red group is enlarged. Error bars are the mean ± 95% CI.
(Related to this Figure).(A) Diminished NMY-2 activity results in HMR-1 accumulation in the ingressing cytokinetic membranes. Medial confocal images of ingressing furrow during P0 cytokinesis from temperature-sensitive allele—nmy-2(ne3409)—co-expressing HMR-1::GFP and Histone::mCherry taken at 15°C (permissive temperature) and 25°C (restrictive temperature). (B) Medial image of P0 division in control and let-502(RNAi) embryos expressing NMY-2::mCherry, Histone::mCherry, and HMR-1::GFP. Accompanying plots show line-scan intensity profiles of the ingressing membrane. Error bars indicate the mean ± 95% CI. (C) Fold change in levels of HMR-1::GFP and NMY-2::mCherry in the cytokinetic membranes during P0 divisions in control and let-502(RNAi) embryos. Plots indicate the mean ± 95% CI. (D) Diminished LET-502 activity does not result in HMR-1 accumulation in ingressing cytokinetic membranes. Equatorial images of P0 cytokinesis of embryos from temperature-sensitive allele let-502(sb102) co-expressing HMR-1::GFP and mCherry::PLC1δ-PH taken at 15°C (permissive temperature, n = 8) and 25°C (restrictive temperature, n = 9). Accompanying plots indicate the line intensity profiles of HMR-1::GFP and plasma membrane. Error bars indicate the mean ± 95% CI. (E) Fold change in furrow levels of HMR-1::GFP and mCherry::PLC1δ-PH at 15°C and 25°C. Plots indicate the mean ± 95% CI. (F) Medial images of P0 division showing membrane ingression in tpxl-1(RNAi) embryos co-expressing NMY-2::mCherry, Histone::mCherry, and HMR-1::GFP. (G) Fold change in levels of HMR-1::GFP and NMY-2::mCherry in the cytokinetic membranes during P0 divisions in control and tpxl-1(RNAi) embryos. Plots indicate the mean ± 95% CI. (H) Candidate RNAi screen to identify factors regulating HMR-1 exclusion at the cytokinetic zone. Medial images of P0 division showing membrane ingression in embryos co-expressing NMY-2::mCherry, Histone::mCherry, and HMR-1::GFP after RNAi-mediated depletion of indicated target genes. (I) Surface plane images during polarization and cytokinesis stage in control and cdc-42(RNAi) embryos co-expressing HMR-1::GFP and Lifeact::RFP. Cyan arrows in the enlarged ROI indicate the localization of HMR-1 being restricted to the edge of F-actin–enriched zone at the furrow. Scale bar—10 μm (see Video 6). (J) Equatorial images of furrow ingression during P0 divisions in embryos co-expressing NMY-2::mCherry, HMR-1::GFP, and Histone::mCherry in control and cdc-42(RNAi). Accompanying plots indicate the mean ± 95% CI of NMY-2::mCherry and HMR-1::GFP line-scan intensities in the ingressing furrows of control (n = 10), cdc-42(RNAi) (n = 6), par-6(RNAi) (n = 9), and par-5(RNAi) (n = 11). (K) Plots show normalized fold change in levels of HMR-1::GFP and NMY:: mCherry in ingressing furrows of control, cdc-42(RNAi), par-6(RNAi), and par-5(RNAi). Error bars show the mean ± 95% CI.
Delayed cytokinesis does not lead to HMR-1 accumulation in the division zone
LET-502 and MLC-4, like RHO-1 and NMY-2, regulate both actomyosin ring assembly and constriction kinetics. let-502(RNAi) and mlc-4(RNAi) embryos exhibited delayed cytokinesis, requiring 466 ± 84 s and 342 ± 121 s, respectively, from anaphase onset to ring constriction (Fig 2H and I). Despite this extended duration, HMR-1 did not accumulate in the ingressing membrane, arguing against the idea that additional time allows HMR-1 diffusion into the furrow. To confirm this, we analyzed tpxl-1(RNAi) zygotes, in which spindle shortening delays the anaphase onset and cytokinetic ring assembly (Özlü et al, 2005; Lewellyn et al, 2010). Although tpxl-1(RNAi) embryos took significantly longer to complete cytokinesis than controls (377 ± 54 s versus 227 ± 21 s), HMR-1::GFP levels at the invaginating membrane were indistinguishable from controls (Figs 2I and S2F and G). Thus, delayed cytokinesis per se is insufficient to drive HMR-1 accumulation in the ingressing membrane. Moreover, depletion of several components known to regulate cytokinesis either directly or indirectly—including PAR-5, CDC-42, ARX-3, AFD-1—did not result in HMR-1 accumulation at the division site (Figs S2H and 2I - highlighted in red zone) (Pittman & Skop, 2012; Wang et al, 2013; Gao et al, 2017; Chan et al, 2019; Onwubiko et al, 2023).
Exclusion of HMR-1 from the cytokinetic furrow zone is independent of vesicle trafficking, PAR-dependent polarity complex, and protein degradation
Intracellular vesicle trafficking has been implicated in the redistribution of adhesion molecules during cytokinesis (Lock & Stow, 2005; Ai & Skop, 2009; Gillard et al, 2015; Sakaguchi et al, 2015; Kakar-Bhanot et al, 2019). We therefore asked whether HMR-1 might be actively removed from the division zone via endocytosis or vesicular transport. However, RNAi-mediated depletion of components involved in clathrin-mediated endocytosis (chc-1, unc-11), caveolin-dependent pathways (cav-1), or vesicle trafficking (dhc-1, rab-11.1) did not result in detectable accumulation of HMR-1::GFP at the division site (Fig S2H).
After polarization of the zygote, HMR-1 clusters are enriched at the anterior cortex and largely absent on the posterior surface (Fig 1A and B) (Padmanabhan et al, 2017). Because CDC-42/PAR-dependent polarity and cortical actomyosin contractility regulate asymmetric localization of many cell surface proteins in the C. elegans zygote (Guo & Kemphues, 1995; Hung & Kemphues, 1999; Cheeks et al, 2004; Motegi & Sugimoto, 2006; Jordan et al, 2016; Wang et al, 2017), we considered whether polarity cues might influence HMR-1 exclusion during cytokinesis. Disruption of anterior–posterior (A-P) polarity via depletion of CDC-42, PAR-5, or PAR-6 abolished anterior enrichment of HMR-1, resulting in a dispersed distribution of HMR-1::GFP clusters during polarization and cytokinesis (Fig S2I, yellow arrows). The surface plane of cdc-42(RNAi) embryos showed few HMR-1 clusters near the furrow zone (Fig S2I; cyan arrows in ROI). Despite this redistribution, HMR-1 did not accumulate in ingressing furrows when examined in the equatorial plane (Fig S2J and K and Video 5). Thus, polarity-dependent sequestration does not account for HMR-1 exclusion from the division zone.
Video 5CYK-1 depletion results in accumulation of HMR-1 in the furrow zone and cytokinesis failure. Equatorial ROI from the medial plane of cyk-1(RNAi) P0 embryos expressing HMR-1::GFP, NMY-2::mCherry, and Histone::mCherry, from metaphase onward. HMR-1 accumulates in the ingressing furrow (yellow arrows). Fully ingressed furrow immediately retracts (cyan arrows) resulting in cytokinesis failure. Frame rate—8 fps. Download video
In mammalian cells, ubiquitination-mediated trafficking and degradation regulate E-cadherin surface levels (Fujita et al, 2002; Hartsock & Nelson, 2012; Niño et al, 2019). To test whether similar mechanisms operate for HMR-1, we depleted the sole C. elegans ortholog of ubiquitin-activating enzyme, UBA-1, as well as E2 ubiquitin–conjugating enzymes—UBC-1, UBC-9, and UEV-1 (Kramer et al, 2010; Papaevgeniou & Chondrogianni, 2014). Disruption of the ubiquitin–proteasome pathway did not result in increased HMR-1 accumulation in the invaginating cytokinetic membrane (Fig S2H). Together, these results indicate that HMR-1 exclusion from the cytokinetic furrow is independent of vesicle trafficking, polarity cues, and ubiquitin-mediated degradation.
Formin/CYK-1 depletion leads to HMR-1 accumulation in the medial furrow and constricting cytokinetic ring
Cortical F-actin in early C. elegans embryos is composed predominantly of unbranched filaments polymerized by the diaphanous-related formin CYK-1, whereas branched actin nucleated by the ARP2/3 complex appears as dense punctate structures (Swan et al, 1998; Severson et al, 2002; Sawa et al, 2003; Roh-Johnson & Goldstein, 2009). Because HMR-1 accumulated in ingressing membranes after RHO-1 depletion, we asked whether CYK-1–mediated F-actin organization downstream of RHO-1 might also regulate HMR-1 dynamics at the division zone. To test this, we selectively disrupted linear or branched actin networks by partial depletion of their respective nucleators in embryos co-expressing HMR-1::GFP and NMY-2::mCherry. In contrast to control, cyk-1(RNAi) embryos exhibited a pronounced (∼4-fold) accumulation of HMR-1 at the ingressing cytokinetic membrane (Fig 3A and B; yellow arrow, and Video 6). A similar phenotype was observed in temperature-sensitive cyk-1(or596) (hereafter cyk-1(ts)) embryos shifted to the restrictive temperature before anaphase onset (Fig S3A) (Davies et al, 2014). HMR-1 accumulation was also evident along the invaginating membrane during AB cell division (Fig 3C, yellow arrow).
*Accumulation of HMR-1 in the cytokinetic region is dependent on formin/CYK-1 activity.(A) Equatorial plane images of the medial cytokinetic zone during P0 division in embryos co-expressing NMY-2::mCherry, HMR-1::GFP, and Histone::mCherry in control, cyk-1(RNAi), and arx-3(RNAi) embryos. Accompanying plots show the mean ± 95% CI of two-color line scans of NMY-2::mCherry (magenta) and HMR-1::GFP (green) fluorescence intensities measured along ROI in the furrow. The yellow arrow denotes the HMR-1 accumulation in the ingressing furrows of cyk-1(RNAi) embryos (see Video 5; cyk-1(RNAi)). (B) CYK-1 depletion shows ∼4-fold increase in the HMR-1 levels in the constricting ring and the ingressing membrane. Plots showing fold change in HMR-1::GFP and NMY:: mCherry fluorescence intensities in cyk-1(RNAi), and arx-3(RNAi) embryos. Error bars denote the mean ± 95% CI of intensity normalized to the mean value of control embryos. Control embryos therefore show a mean fold change of 1. (C) Equatorial images of the cytokinetic membrane during AB division in embryos co-expressing NMY-2::mCherry, HMR-1::GFP, and Histone::mCherry in control and cyk-1(RNAi) embryos. Accompanying plots show the mean ± 95% CI of two-color line scans of NMY-2::mCherry (magenta) and HMR-1::GFP (green) fluorescence intensities measured along the ingressing membrane. The yellow arrow denotes the HMR-1 accumulation in the membrane partitioning ABa and ABp cells in cyk-1(RNAi) embryos. (D) Time-lapse images of enlarged ROI from Fig 3C, depicting the AB-P1 cell boundary during AB cell division. Depletion of CYK-1 results in HMR-1 accumulation (yellow arrow) at the tricellular junction formed at the AB-P1 cell boundary, whereas control embryos show a progressive clearance of junctional HMR-1 at the site of ring constriction (depicted by a yellow asterisk). (E) Equatorial image of P0 furrow ingression in control and pfn-1(RNAi) embryos co-expressing NMY-2::mCherry, HMR-1::GFP, and Histone::mCherry. The plot denotes the mean ± 95% CI of NMY-2::mCherry and HMR-1::GFP line-scan intensities of the ingressing furrow. The yellow arrow shows the furrow localization of HMR-1 in pfn-1(RNAi) embryos. (F) Plots show normalized fold change in furrow intensities of HMR-1::GFP and NMY:: mCherry in control (n = 15) and pfn-1(RNAi) (n = 12) embryos. Error bars are the mean ± 95% CI (*P < 0.05, **P < 0.005, **P < 0.0001, two-tailed, unpaired, nonparametric test, Mann–Whitney test).
Video 6CDC-42–mediated embryonic polarity does not regulate HMR-1 dynamics at the medial cytokinetic zone. Surface plane of cdc-42(RNAi) embryos co-expressing HMR-1::GFP and Lifeact::mCherry shows exclusion of HMR-1 clusters from the medial cytokinetic zone (white rectangular ROI). The cyan arrowhead indicates anterior clusters stalling at the edge of the ROI. The yellow arrowhead indicates events where HMR-1 clusters that are trapped within the medial region are gradually excluded out the ROI. Download video
(Related to this Figure).(A) Equatorial image of furrow ingression during the P0 cytokinesis of cyk-1 (ts) embryo co-expressing HMR-1::GFP and NMY-2::mCherry. Acquired at the restrictive temperature (25°C), HMR-1::GFP is localized in the furrow upon loss of CYK-1 activity (yellow arrow). (B) Medial plane of AB cell of nmy-2(RNAi) embryo undergoing cytokinesis. HMR-1 is not depleted from the site of ring constriction (yellow arrows) (tricellular junction).
We next examined whether CYK-1 and NMY-2 also regulate the exclusion of junctional HMR-1 during cytokinesis. Time-lapse imaging of the AB-P1 boundary during AB cell division revealed that in control embryos, HMR-1 intensity at the furrow zone progressively declined as the cytokinetic ring constricted (Fig 3D; control—yellow asterisk; and Video 3). In contrast, cyk-1(RNAi) and nmy-2(RNAi) embryos displayed sustained accumulation of HMR-1 at the site of furrow ingression throughout AB cell cytokinesis (Fig 3D; cyk-1(RNAi)—yellow arrow; and Fig S3B). Notably, depletion of factors promoting branched actin assembly—including ARP2/3 components, wip-1, wve-1, and rac-1—had little effect on the HMR-1 levels in the ingressing membrane (Figs 3A-arx-3(RNAi), Fig S2H). Similarly, depletion of the scaffolding proteins anillin (ANI-1) or septin (UNC-59), or the F-actin–severing protein cofilin (UNC-60), did not alter HMR-1 exclusion from the division zone (Fig S2H). Because elongation of CYK-1–polymerized actin filaments requires profilin-bound G-actin, we examined the effect of PFN-1 depletion (Severson et al, 2002). Consistent with CYK-1 depletion, partial knockdown of PFN-1 resulted in ∼5-fold enrichment of HMR-1 within the cytokinetic zone during ring constriction (Fig 3E, yellow arrow, Figs 3F and 4G). Collectively, these findings identify CYK-1–dependent unbranched F-actin assembly as a key determinant of HMR-1 exclusion from the cytokinetic furrow.
*CYK-1–polymerized unbranched actin filaments at the furrow zone restrict HMR-1 mobility during cytokinesis.(A) Intensity-coded temporal tracks of HMR-1::GFP clusters at the cell cortex during cytokinesis in control (upper row panel) and cyk-1(RNAi) (lower row panel) embryos. “ay” and “by” denote rotational flow of HMR-1::GFP clusters that are tracked perpendicular (y-axis, blue arrow) to the embryonic axis. Furrow-directed cortical flow (“ax” and “bx”) carries HMR-1 parallel (x-axis) to the embryonic axis. Control embryo tracks show movement of HMR-1 clusters paused at the cytokinetic furrow region (blue arrowhead), whereas cyk-1(RNAi) embryos show the movement and accumulation of these puncta at the furrow zone (red arrow). (B, C) Furrow-directed cortical movement of HMR-1 clusters is enhanced in cyk-1(RNAi) embryos. Normalized frequency histogram plot showing velocities in y and x directions, (B) ∣VY∣ and (C) ∣VX∣, of HMR-1::GFP clusters during ring-directed cortical flow in control (blue, n = 15 embryos) and cyk-1(RNAi) (red, n = 18 embryos). ((B) P = ns (0.137), (C)****P < 0.0001; nonparametric, Mann–Whitney U statistic test) (see Video 8). (D) Cortical view of control and cyk-1(RNAi), pfn-1(RNAi), and arx-3(RNAi) embryos expressing HMR-1::GFP and Lifeact::RFP undergoing cytokinesis. Parallelly arranged actin filaments are compacted at the medial cortex in control and arx-3(RNAi) embryos (bottom merge panel, enlarged blue ROI). HMR-1 clusters in control remain outside yellow margin, the boundary marking compact actin organization. cyk-1(RNAi) and pfn-1(RNAi) embryos show reduced linear actin filaments in medial cortex (middle row, cyan arrowheads) and increased branched actin structures (middle row, cyan arrows). arx-3(RNAi) embryos lack branched actin structures and show linear actin filaments. The yellow arrow (top row) highlights accumulation of HMR-1 clusters in the ingressing furrow zone in cyk-1(RNAi) and pfn-1(RNAi) embryos. Scale bar—10 μm (see Video 8 and Video 10). (E) Time series of medial cortex in control and cyk-1(RNAi) embryos from the ROIs in Fig 4C. The yellow arrowhead in control embryo tracks a HMR-1 cluster moving along the edge of compact actin-enriched zone, but not into the furrow. The cyan arrowhead in the cyk-1(RNAi) embryo indicates the movement of HMR-1 cluster into the furrow zone because of disruption of compacted actin in the medial cortex (see Video 7 and Video 9). (F) Schematic representation of the parallel and compact organization of unbranched actin filaments in the furrow zone (indicated as black ROI) during cytokinesis. The enlarged red ROI shows the angle (θ) of orientation between actin filaments and the embryonic axis. For clarity, ARP2/3-polymerized branched F-actin structures are not depicted in this schematic. (G) Normalized polar histogram plots showing average orientation of medially assembled actin filaments in control (blue, n = 10) and cyk-1(RNAi) (red, n = 10) embryos. Actin filaments in the cytokinetic furrow zone (black ROI in Fig 4E) are considered for the orientation measurement. (H) Quantification of HMR-1 clusters localized to the cytokinetic furrow zone in control (n = 38), cyk-1(RNAi) (n = 17), arx-3(RNAi) (n = 18), cyk-1;arx-3(RNAi) (n = 7), pfn-1(RNAi) (=20), nmy-2(RNAi) (n = 16), hmp-1(RNAi) (n = 11), and plst-1(RNAi) (n = 6) embryos. Error bars are the mean ± 95% CI (*P < 0.05, **P < 0.005, ****P < 0.0001, two-tailed, unpaired, nonparametric, Mann–Whitney test). (I) Graph shows linear fit of the Loge (mean square displacement, <r2>) versus Loge (track duration of clusters, T) of HMR-1 clusters during furrow invagination (∼80 to 200 s) in control (blue, n = 15), cyk-1(RNAi) (red, n = 18), and arx-3(RNAi) (gray, n = 7) embryos. Plot shows the mean ± s.d. (J, K) Truncated violin plots showing (I) diffusion exponent “α” and (J) diffusion coefficient “D” of HMR-1::GFP clusters estimated from Fig 4H for control, cyk-1(RNAi), and arx-3(RNAi) embryos. (I) “α > 1” in control and arx-3(RNAi) embryos suggests super-diffusion, whereas “α∼1” in cyk-1(RNAi) embryos indicates random (Brownian) diffusion. (J) The value of “D” was significantly higher in cyk-1(RNAi), compared with control or arx-3(RNAi) embryos, suggesting the role of CYK-1–polymerized unbranched actin filaments in hindering the mobility of HMR-1 clusters. The violin plots show median (solid bar) along with first and second quartiles (dashed bar). arx-3(RNAi) resulted in fewer HMR-1 clusters in the medial surface. Error bars are the mean ± 95% CI (*P < 0.05, **P < 0.005, ***P < 0.0001, two-tailed, unpaired, nonparametric test, Mann–Whitney test). (L) cyk-1(RNAi) embryos demonstrate reduced resistance to HMR-1 diffusion in the medial zone. Each datapoint represents the mean diffusion coefficient for a given surface region, and each pair of points corresponds to the anterior and medial furrow zone from the same embryo. Compared to the control, cyk-1(RNAi) embryos display higher mean diffusion coefficient values. Additionally, difference in the diffusion coefficients between anterior and medial cortical surfaces is absent in cyk-1(RNAi) embryos. (P < 0.05, Wilcoxon matched-pairs signed-rank test).
CYK-1–polymerized F-actin restricts the cortical mobility of HMR-1 clusters at the cytokinetic furrow zone
During cytokinesis, cortical flows directed along the embryonic axis and toward the furrow contribute to both assembly and constriction of the cytokinetic actomyosin ring (Khaliullin et al, 2018; Hsu et al, 2023; Alonso-Matilla et al, 2024). Surface-localized HMR-1 clusters physically associate with cortical F-actin and are known to retard cortical flows (Padmanabhan et al, 2017). We therefore asked whether CYK-1–polymerized actin filaments at the medial cortex limit the lateral mobility of HMR-1 clusters and prevent their entry into the invaginating membrane. To address this, we tracked individual HMR-1 clusters in embryos co-expressing HMR-1::GFP and Lifeact::RFP. In control embryos, rotational cortical flows induced by embryo compression transported HMR-1 clusters orthogonal to the embryonic A-P axis (i.e., y-axis, Fig 4A (a_y_), blue arrow) (Padmanabhan et al, 2017; Singh et al, 2019). After this initial displacement, HMR-1::GFP clusters paused and redirected toward the cytokinetic ring (i.e., x-axis, Fig 4A (a_x_); blue arrowhead, and Video 7, Video 8, and Video 9), with median |V_Y_^Contol^| and |V_X_^Contol^| values to be 0.089 and 0.044 μm/s, respectively (Fig 4B and C). Strikingly, in control embryos, this ring-directed movement stalled at the outer boundary of the medial ingression zone enriched in parallel F-actin filaments (Fig 4D and E; control, and Video 7). This boundary corresponds to the region in which CYK-1–polymerized unbranched actin filaments are oriented orthogonal to the embryonic axis and parallel to the ingressing furrow (schematic shown in Fig 4F and orientation quantified in Figs 4G and S4A; control) (Reymann et al, 2016; Leite et al, 2020; Li et al, 2021).
Video 7Surface mobility of HMR-1 clusters is stalled at the boundary of the cytokinetic furrow zone. The cortical plane of control embryos expressing HMR-1::GFP and Lifeact RFP undergoing cytokinesis. Laterally moving HMR-1 clusters (yellow and cyan arrowheads) are excluded from the ingressing furrow zone (indicated by white rectangular ROI). Download video
Video 8Increased furrow-directed lateral mobility of HMR-1 is observed in cyk-1(RNAi) embryos. The cortical plane of control and cyk-1(RNAi) embryos expressing HMR-1::GFP and Lifeact RFP undergoing cytokinesis. Depletion of CYK-1 results in increased HMR-1 mobility toward the ingressing furrow. Frame rate—8 fps. Download video
Video 9CYK-1–polymerized unbranched actin filament organization in cytokinetic ring impedes furrow-directed movement of HMR-1 clusters. ROI from the cortical plane of a dividing P0 embryo expressing HMR-1::GFP and Lifeact::RFP in control and *cyk-*1(RNAi) embryos shows an accumulation of HMR-1 in the furrow zone upon disruption of cytokinetic actin architecture in cyk-1(RNAi) embryos. Frame rate—4 fps. Download video
*(Related to this Figure).(A) Bar plot quantification showing most of the cortical filaments during cytokinesis in control embryos (blue) are oriented perpendicular to the embryonic axis 90° < θ < 60°, whereas cyk-1(RNAi) embryos (red) show randomly oriented filaments 60° < θ < 30°. (B) Time-lapse cortical plane images of embryos expressing Lifeact::RFP and HMR-1::GFP depleted of both CYK-1 and ARX-3. F-actin structures are significantly reduced. The yellow arrowhead indicates the cytokinetic furrow. The white arrow denotes medial accumulation of HMR-1 during furrow ingression (see Video 11). (A, B, C, D) Medial plane images of two-cell-stage embryos expressing (A) F-actin and HMP-2, or (B) F-actin and HMP-1. The localization of HMP-2 and HMP-1 to nonjunctional surfaces (yellow arrowhead) and cell junctions (yellow arrow) is HMR-1–dependent. (A) HMP-2 and (B) HMP-1 fail to localize to cell surface in hmr-1(RNAi) embryos. Scale bar—10 μm. (E) Representative medial plane images showing ingressing cytokinetic membrane during P0 division in control (n = 18) and hmp-1(RNAi) (n = 13) embryos co-expressing NMY-2::mCherry, HMR-1::GFP, and Histone::mCherry. Graphs adjoining the image show the mean ± 95% CI of NMY-2::mCherry and HMR-1::GFP intensities. The yellow arrow indicates accumulation of HMR-1::GFP in the ingressing furrow upon HMP-1 depletion. (F) Plot showing normalized fold change fluorescence intensities of HMR-1::GFP and NMY-2::mCherry in the ingressing furrows of control (n = 17) and hmp-1(RNAi) (n = 14) embryos. Error bars shown are the mean ± 95% CI. (**P ≤ 0.01, Mann–Whitney test). (G) Time taken for cytokinesis in control, cyk-1;rol-6(RNAi), cyk-1;hmp-1;rol-6(RNAi), and cyk-1; hmp-1;hmr-1(RNAi) embryos. The plot shows a simultaneous depletion of CYK-1 and HMP-1 exacerbates the delay in ring constriction exhibited by embryos depleted of CYK-1 alone. Additional removal of HMR-1 in embryos lacking CYK-1 and HMP-1 rescues the delay (mean ± 95% CI, *P ≤ 0.05, ****P ≤ 0.0001, two-tailed, unpaired, nonparametric test, Mann–Whitney test). (H) Bar graph shows the percentage of successful (blue) and failed cytokinesis (red) in cyk-1;rol-6(RNAi), cyk-1;hmp-1;rol-6(RNAi), and cyk-1; hmp-1;hmr-1(RNAi) embryos. Co-depleting HMR-1 rescues cytokinesis failure in cyk-1;hmp-1;rol-6(RNAi) embryos. Statistical significance was determined using two-tailed Fisher’s exact test (P > 0.05, *P ≤ 0.05, *P ≤ 0.01). (I) Plot showing normalized fold change fluorescence intensities of HMR-1::GFP and NMY-2::mCherry in the ingressing furrows of control (n = 17), cyk-1(RNAi) (n = 17), and cyk-1(RNAi);hmp-1(RNAi) (n = 14) embryos. Error bars shown are the mean ± 95% CI (P ≤ 0.0001, Mann–Whitney test). (J) AFD-1 depletion has no effect on HMR-1 localization during furrow ingression. Equatorial image of P0 furrow ingression in control and afd-1(RNAi) embryos co-expressing NMY-2::mCherry, HMR-1::GFP, and Histone::mCherry. (K) Plot shows the mean ± 95% CI of normalized fold change in furrow intensities of NMY-2::mCherry and HMR-1::GFP in afd-1(RNAi) (n = 11) embryos.
In contrast, in cyk-1(RNAi) embryos, the x-directed movement of HMR clusters was significantly enhanced (median |V_X_^cyk−1^| - 0.062 μm/s), whereas the vertical component of the velocity |V_Y_| remained comparable to that in HMR cluster control embryos (median |V_X_^cyk−1^| - 0.085 μm/s) (Fig 4B and C and Video 8 and Video 9; cyk-1(RNAi)). Under these conditions, HMR-1 clusters no longer stalled at the furrow boundary but instead continued unimpeded into the center of the division zone (Fig 4E, cyan arrowheads, and Video 8 and Video 9; cyk-1(RNAi)). Consistent with this behavior, cyk-1(RNAi) embryos showed a marked reduction in the density and alignment of unbranched actin filaments perpendicular to the embryonic A-P axis (78.1% in control versus 28.6% in cyk-1(RNAi) of total filaments in the furrow zone) (Fig 4D, F-actin; red ROI—cyan arrowheads; and quantified in Figs 4G and S4A). CYK-1 depletion also led to an increase in ARP2/3-dependent punctate F-actin structures, at the cortex (Fig 4D, F-actin panel; cyan arrows), as previously reported (Burke et al, 2014; Suarez et al, 2015; Chan et al, 2019).
Consistent with unrestricted mobility, the number of HMR-1 clusters detected within the division zone increased significantly in cyk-1(RNAi) embryos (12.2 ± 3.8 clusters) compared with controls (Fig 4H). In contrast, depletion of ARP2/3 (arx-3(RNAi)) resulted in a reduced anterior polarity domain, and HMR-1 clusters remained excluded from the furrow, similar to control embryos (Fig 4D; arx-3(RNAi), Fig 4H, and Video 10). Embryos co-depleted of CYK-1 and ARP2/3 were devoid of most of the cortical F-actin structures. In arx-2(RNAi); cyk-1(RNAi) embryos, HMR-1 clusters initially localized to the anterior cortex (like in arx-2(RNAi) embryos). Although absent from the cytokinetic furrow zone during initial stages of cytokinesis, HMR-1 clusters rapidly translocated along the surface and accumulated in the equatorial division zone (like in cyk-1(RNAi) embryos) (Fig S4B; Video 11). Together, these observations suggest that CYK-1–polymerized unbranched actin filaments form a physical barrier that restricts HMR-1 mobility during ring-directed cortical flow.
Video 10HMR-1 clusters localized to the anterior end in arx-3(RNAi) embryos are excluded from the cytokinetic zone. The cortical plane of arx-3(RNAi) embryos co-expressing HMR-1::GFP and Lifeact::mCherry showing enhanced actin filaments and HMR-1 clusters excluded from the cytokinetic zone. Frame rate—10 fps. Download video
Video 11Co-depleting CYK-1– and ARP2/3–polymerized cortical actin rapidly translocates to medial zone during cytokinesis. HMR-1 clusters in cyk-1(RNAi);arx-3(RNAi) embryo are initially limited to the anterior end, and are excluded from the cytokinetic furrow during early stages of cytokinesis. In the absence of cortical actin, thereafter, HMR-1 clusters rapidly are redistributed along the cell surface and accumulate in the equatorial division zone. Download video
To quantitatively assess how CYK-1–dependent actin organization influences HMR-1 mobility during cytokinetic ring constriction, we analyzed the trajectories of individual HMR-1::GFP clusters by calculating their mean square displacement (MSD) during ring-directed cortical flow in control, cyk-1(RNAi), and arx-3(RNAi) embryos (Fig 4I). Assuming two-dimensional surface mobility, we estimated the mean diffusion exponent (α) and diffusion coefficient (D) for each condition. In control embryos, HMR-1 clusters exhibited anomalous, super-diffusive behavior, with a mean α value of ∼1.76 ± 0.263 (median “α” = 1.84), consistent with directed motion coupled to cortical flows (Fig 4J). The mean diffusion coefficient (D) value for HMR-1::GFP clusters in control embryos was 0.01 ± 0.004 μm^2^ s^−1^, comparable to previously reported values for E-cadherin dynamics in other experimental systems (Fig 4K) (Kusumi et al, 1993; Adams et al, 1998; Biswas et al, 2015; Fujiwara et al, 2016). Notably, within individual embryos undergoing cytokinesis, the diffusion coefficient of HMR-1 clusters was consistently lower in the medial furrow zone (0.010 μm^2^s^−1^) than at the anterior cortex (0.015 μm^2^s^−1^) (Fig 4L), supporting the idea that actin architecture in the furrow imposes constraints on HMR-1 mobility. In contrast, HMR-1 clusters in cyk-1(RNAi) embryos displayed near-Brownian motion with “α” values ∼1.028 ± 0.161 (median “α” = 1.017), indicating loss of directional constraint. Importantly, the difference in diffusion coefficients between the medial (0.030 μm^2^s^−1^) and anterior (0.027 μm^2^s^−1^) cortex observed in control embryos was abolished upon CYK-1 depletion (Fig 4L). Correspondingly, the mean diffusion coefficient in cyk-1(RNAi) embryos (0.030 ± 0.012 μm^2^ s^−1^) was ∼3-fold higher than in control or arx-3(RNAi) embryos (D∼ 0.01 ± 0.005 μm^2^ s^−1^) (Fig 4K). In arx-3(RNAi) embryos, which contain increased density of CYK-1–polymerized unbranched actin filaments but lack branched actin networks, HMR-1 clusters exhibited a higher diffusion exponent (α∼1.73 ± 0.21). We hypothesize that enhanced association of HMR-1 clusters with this denser actin network in arx-3(RNAi) embryos increases directional persistence (higher “α”) while limiting overall diffusion (lower “D”). These data indicate that CYK-1–polymerized unbranched actin filaments impose spatially localized constraints on HMR-1 mobility, thereby preventing its diffusion into the furrow zone during cytokinesis-associated ring-directed cortical flow.
HMP-1–mediated linkage with cortical actin restricts HMR-1 localization at the division furrow
Physical coupling between E-cadherin and cortical F-actin regulates cadherin clustering and surface mobility (Sako et al, 1998; Wu et al, 2015; Rao & Zaidel-Bar, 2016; Chandran et al, 2021). In C. elegans, the β- and α-catenin homologs HMP-2 and HMP-1 link the cytoplasmic domain of HMR-1 to actin and are recruited to the embryonic cell surfaces in an HMR-1–dependent manner (Fig S4C and D). Disruption of either HMP-2 or HMP-1 decouples HMR-1 from F-actin, and compromises junctional integrity during epithelial morphogenesis (Costa et al, 1998; Loveless & Hardin, 2012; Choi et al, 2015; Kang et al, 2017). Given the parallel alignment of actin filaments at the furrow in control embryos (Fig 4C and F) and the increased mobility of HMR-1 clusters upon disruption of this architecture in cyk-1(RNAi) embryos (Fig 4B and F), we tested whether the catenin-mediated linkage between HMR-1 and F-actin contributes to HMR-1 exclusion during cytokinesis. To this end, we quantified HMR-1 levels in the furrow region of hmp-1(RNAi) embryos. Compared with controls, depletion of HMP-1 resulted in a ∼3-fold increase in HMR-1 levels at the ingressing cytokinetic membrane (yellow arrow in Fig S4E and F). Co-depletion of HMP-1 and CYK-1 further exacerbated this phenotype, leading to an eightfold increase in HMR-1 accumulation, prolonged actomyosin ring constriction (∼400 s), and a marked increase in cytokinesis failure (∼50%) (Fig S4G–I). Strikingly, additional depletion of HMR-1—triple hmr-1(RNAi);hmp-1(RNAi);cyk-1(RNAi) embryos—nearly completely rescued both the delayed constriction and cytokinesis failure phenotypes (Fig S4G and H). These results demonstrate that HMP-1–mediated coupling of HMR-1 to cortical F-actin contributes to impede HMR-1 mobility and restrict its entry into the cytokinetic furrow zone. AFD-1, the C. elegans homolog of afadin, genetically and physically interacts with the cadherin–catenin complex and functions synergistically with HMP-1 during epidermal morphogenesis (Lynch et al, 2012; Indra et al, 2014; Gong et al, 2024 Preprint). Because AFD-1 is also expressed in early embryos (Serre et al, 2023; Slabodnick et al, 2023; Hall et al, 2024), we tested whether it, similar to HMP-1, contributes to exclusion of HMR-1 during cytokinesis. In contrast to hmp-1(RNAi) embryos, depletion of AFD-1 did not result in detectable HMR-1 accumulation in the invaginating membrane (Fig S4J and K), indicating HMP-1 plays a dominant role in mediating actin-dependent restriction of HMR-1 mobility during cytokinesis.
NMY-2– and PLST-1–mediated crosslinking and compaction of actin filaments restrict HMR-1 mobility during cytokinesis
During cytokinetic ring assembly, CYK-1–polymerized unbranched actin filaments are aligned and compacted through the combined actin-binding and crosslinking activities of NMY-2 and the actin-bundling protein plastin/PLST-1 (Leite et al, 2020; Henson et al, 2024). We therefore asked whether actin filament crosslinking and compaction at the medial cortex contribute to restricting the lateral mobility of HMR-1 clusters. To address this, we quantified the width (d) of the medial cortical region enriched in parallel actin filaments during cytokinesis. In control embryos, this actin-enriched zone was relatively narrow, with a mean width of 10.92 ± 2.17 μm (Fig 5A). In contrast, nmy-2(RNAi) embryos displayed highly disorganized actin architecture at the equatorial cortex and a significantly broadened actin-enriched zone (∼12.81 ± 2.18 μm; Fig 5A and C). The extent of disorganization varied with RNAi penetrance, with stronger NMY-2 depletion resulting in more severe disruption of filament alignment and compaction (Fig 5B and Video 12). Consistent with this loss of actin compaction, nmy-2(RNAi) embryos exhibited a dramatic increase in HMR-1 accumulation within the furrow region, with an average of ∼20.9 ± 9.1 HMR-1 clusters detected in the medial division zone—∼10-fold higher than in control embryos (Figs 4G and 5D).
*Compact organization of actin filaments during cytokinetic ring assembly prevents HMR-1 clusters from localizing at the furrow zone.(A, B) Images showing surface plane of (A) nmy-2(RNAi) and (B) plst-1(RNAi) embryos co-expressing HMR-1::GFP and F-actin (Lifeact::RFP). Variability in RNAi penetrance results in variation in severity of disruption in cytokinetic actin organization. The radial plot indicates the orientation of actin filaments in the furrow zone. (A, B) Embryos with lower actin crosslinking during actomyosin ring assembly in (A) nmy-2(RNAi) and (B) plst-1(RNAi) embryos allow HMR-1 access the cytokinetic furrow zone (yellow arrows) (see Video 12). (C) Plot showing broadening of the width of the actin-enriched region during cytokinesis in nmy-2(RNAi) and plst-1(RNAi) embryos. The accompanying schematic shows the actin-enriched region in the furrow zone whose width “d” is measured (black ROI). The width “d” of the Gaussian is measured at 10% of the increase in intensity across the length of the embryonic axis (red ROI) (*P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ***P ≤ 0.0001, Mann–Whitney test). (D) Plot showing correlation between the width of the actin-enriched zone (“d”) during furrow formation and the number of medially localized HMR-1 clusters in control (black), nmy-2(RNAi) (green), and plst-1(RNAi) (yellow) embryos. Increase in “d” upon disruption of actin crosslinking results in an increased HMR-1 localization. Error bars show the mean ± 95% CI. (E) Medial plane images of ingressing membrane during P0 division in embryos co-expressing NMY-2::mCherry, HMR-1::GFP, and Histone::mCherry in control and plst-1(RNAi). Plots show the mean ± 95% CI of NMY-2::mCherry and HMR-1::GFP intensity in control (n = 12) and plst-1(RNAi) (n = 13) embryos. The yellow arrow shows accumulation of HMR-1 in the constricting ring upon PLST-1 depletion. (F) Plot showing fold change in HMR-1::GFP and NMY:: mCherry fluorescence intensities in cytokinetic membrane of control (n = 12) and plst-1(RNAi) (n = 13) embryos. The fold change of HMR-1::GFP and NMY::mCherry in plst-1(RNAi) is normalized to corresponding mean intensities in the control. Error bars denote the mean ± 95% CI.
Video 12Depletion of NMY-2 results in accumulation of HMR-1 clusters in the contractile furrow zone. The cortical plane of nmy-2(RNAi) embryos co-expressing HMR-1::GFP and Lifeact::mCherry showing clusters of HMR-1 accumulating in the medial furrow zone during furrow ingression. Download video
PLST-1 has been shown to tightly crosslink linear and branched actin filaments and, together with the β_H_-spectrin (SMA-1), plays an essential role in maintaining cortical actomyosin connectivity during polarization and cytokinesis (Ding et al, 2017; Sobral et al, 2021; Silva et al, 2023). Embryos depleted of PLST-1 displayed defects in actin structures and delayed furrow ingression, consistent with impaired crosslinking of the cytokinetic actin network (Fig 5B). Similar to nmy-2(RNAi) embryos, PLST-1 depletion resulted in an expanded medial F-actin–enriched region (∼14.44 ± 3.61 μm; Fig 5B and C) and pronounced accumulation of HMR-1::GFP within the ingressing furrow zone (Figs 4H and 5E and F). Notably, the degree of actin zone broadening correlated positively with the number of HMR-1 clusters detected in the furrow (Fig 5D). Consistent with this idea, depletion of either α or β_H_-spectrins, via spc-1(RNAi) or sma-1(RNAi), respectively, resulted in a modest but reproducible increase in HMR-1 levels (Fig S5A and B). Together, these observations indicate that multiple actin crosslinkers cooperate to compact CYK-1–polymerized actin filaments at the medial cortex, thereby restricting HMR-1 mobility. Disruption of this compact actin architecture creates permissive gaps that allow HMR-1 clusters to diffuse into the ingressing furrow.
(Related to this Figure).(A) Equatorial image of P0 furrow ingression in control (n = 10), sma-1(RNAi) (n = 11), and spc-1(RNAi) (n = 14) embryos expressing NMY-2::mCherry, HMR-1::GFP, and Histone::mCherry. The accompanying graph indicates the mean ± 95% CI of line-scan intensities of NMY-2::mCherry and HMR-1::GFP along the ingressing furrow. (B) Plot shows normalized fold change in furrow intensities of HMR-1::GFP and NMY:: mCherry in control (n = 10), sma-1(RNAi) (n = 14), and spc-1(RNAi) (n = 12) embryos. Error bars denote the mean ± 95% CI.
Accumulation of HMR-1 in the medial division zone compromises cytokinetic fidelity
Loss of CYK-1, NMY-2, or their upstream regulator RHO-1 resulted in the accumulation of HMR-1 along invaginating membrane surfaces and at the edges of constricting cytokinetic rings. This aberrant localization of HMR-1 correlated with a delayed or failed furrow ingression depending on the severity of depletion (Fig 6A, cyk-1;rol-6(RNAi)—48 h; Fig S6A–C) (Swan et al, 1998; Ding et al, 2017). Because anterior localization of HMR-1 has previously been shown to slow cytokinetic furrow ingression in C. elegans zygote (Padmanabhan et al, 2017), we asked whether aberrant localization of HMR-1 within the actomyosin ring might contribute to impaired cytokinesis. Notably, the extent of HMR-1 accumulation at the furrow zone correlated inversely with successful ring constriction (Fig 6B). We first co-depleted HMR-1 together with key regulators of actin organization—RHO-1, CYK-1, NMY-2, PLST-1, or ARP2/3—and quantified the kinetics of cytokinetic ring closure during P0 divisions. In control embryos, furrow ingression was completed 223 ± 12 s after anaphase onset (Fig 6F). In contrast, only a subset of embryos depleted of actomyosin regulators successfully completed cytokinesis: 10 out of 31 cyk-1;rol-6(RNAi), 18 out of 24 nmy-2;rol-6(RNAi), and 8 out of 10 rho-1;rol-6(RNAi) embryos, requiring substantially longer times for furrow completion (397 ± 88 s, 526 ± 80 s, and 504 ± 41 s, respectively; Figs 6D and F and 2I). Strikingly, co-depletion of HMR-1 with CYK-1 restored cytokinesis in most of the embryos: 24 out of 26 cyk-1;hmr-1(RNAi) embryos (as well as 19 out of 19 nmy-2;hmr-1(RNAi) and 7 out of 7 rho-1;hmr-1(RNAi)) successfully completed furrow ingression (Fig 6A, C, and D). Although ring constriction in cyk-1;hmr-1(RNAi) embryos remained slow (428 ± 43 s), this represented a dramatic rescue of the cytokinesis failure observed in cyk-1;rol-6(RNAi) embryos. Similarly, plst-1;hmr-1(RNAi) embryos, compared with plst-1(RNAi) embryos, exhibited accelerated ring assembly but required similar times for furrow initiation and ring constriction (Fig S6D). In contrast, depletion of HMR-1 did not accelerate cytokinesis in embryos depleted of ARP2/3 activity (Fig S6E). Together, these results indicate that aberrant accumulation of HMR-1 within the furrow exacerbates cytokinetic defects caused by disrupted actin organization, ultimately leading to cytokinesis failure.
*HMR-1 and UNC-59-ANI-1 pathways regulate cytokinetic furrow ingression via distinct mechanisms in cyk-1(RNAi) embryos.(A) Time series of equatorial plane showing ingression during P0 division in control, cyk-1;rol-6(RNAi), and cyk-1;hmr-1(RNAi) embryos co-expressing NMY-2::mCherry, HMR-1::GFP, and Histone::mCherry. RNAi was carried out for 24, 36, and 48 h as indicated to demonstrate dose-dependent effect of loss of CYK-1 activity. (B) Plot quantifying HMR-1::GFP intensity at the medial surface during furrow initiation (i.e., post-anaphase onset until the start of back-to-back membrane invagination) in control (black curve; n = 6 embryos) and cyk-1(RNAi) embryos. HMR-1::GFP levels are quantified for cyk-1(RNAi) embryos that failed to initiate ingression (red curve; 48- to 56-h RNAi; n = 8 embryos), showed partial ingression (orange curve; 36- to 48-h RNAi; n = 5 embryos), and carried out successful ingression (blue curve; 24- to 36-h RNAi; n = 5 embryos). (C) Kinetics of cytokinetic furrow closure post-anaphase onset in control (n = 23), cyk-1;rol-6(RNAi) (n = 18), and cyk-1;hmr-1(RNAi) (n = 25) embryos. (D) Bar graph shows the percentage of successful (blue) and failed cytokinesis (red) after co-depletions with rol-6(RNAi) or hmr-1(RNAi). Co-depleting HMR-1 rescues cytokinesis failure in cyk-1(RNAi), nmy-2(RNAI), or rho-1(RNAi) embryos. Statistical significance was determined using two-tailed Fisher’s exact test (*P ≤ 0.05, ****P ≤ 0.0001). (E) Time lapse of equatorial plane showing actomyosin ring constriction in cyk-1(ts);rol-6(RNAi), cyk-1(ts);unc-59(RNAi), and cyk-1(ts);unc-59;hmr-1(RNAi) embryos co-expressing NMY-2::mCherry and HMR-1::GFP. Yellow arrows indicate HMR-1 accumulation in the medial surface and invaginating membrane during cytokinesis. (F) Time taken for cytokinesis in control, cyk-1;rol-6(RNAi), cyk-1;unc-59;rol-6(RNAi), and cyk-1;unc-59;hmr-1(RNAi) embryos. Plot shows HMR-1 co-depletion rescues the delay in the furrow ingression time in cyk-1;rol-6(RNAi) and cyk-1;unc-59;rol-6(RNAi) embryos (mean ± 95% CI, P > 0.05, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ***P ≤ 0.0001, two-tailed, unpaired, nonparametric test, Mann–Whitney test). (G) Model of the proposed mechanism for actin architecture in the actomyosin ring impeding movement of HMR-1 into the cytokinetic furrow zone.
*(Related to this Figure).(A) Medial plane images of cyk-1(RNAi) embryos expressing HMR-1::GFP and NMY-2::mCherry with partially constricted cytokinetic rings. HMR-1 was enriched at the edge of all rings we observed in cyk-1(RNAi) embryos. All rings indicated here retracted thereafter resulting in cytokinesis failure. (B) Upshifting cyk-1(ts) embryos co-expressing HMR-1::GFP and NMY-2::mCherry to semi-restrictive 20°C results in HMR-1 accumulation at the edge of constricting ring resulting in increased instances of furrow retraction. HMR-1 depletion rescues these embryos from cytokinesis failure. (C) Co-depleting HMR-1 rescues cytokinetic failures in nmy-2(RNAi) embryos. Kinetics of cytokinetic furrow closure post-anaphase onset is in control (n = 11), nmy-2;rol-6(RNAi) (n = 18), and nmy-2;hmr-1(RNAi) (n = 23) embryos. (D, E) Plot denotes the time taken for various cytokinesis stages in (D) plst-1(tm4255), plst-1(tm4255);hmr-1(RNAi), and (E) arx-2(RNAi) and arx-2(RNAi);hmr-1(RNAi) embryos (mean ± 95% CI, *P ≤ 0.05, ***P ≤ 0.0001, two-tailed, unpaired, nonparametric test, Mann–Whitney test). (F) Time lapse of equatorial plane showing actomyosin ring constriction in control, cyk-1;rol-6(RNAi), cyk-1;unc-59(RNAi), and cyk-1;unc-59;hmr-1(RNAi) embryos co-expressing NMY-2::mCherry, HMR-1::GFP, and Histone::mCherry. (G) Schematic explaining that HMR-1 regulates cytokinesis independent of actomyosin contractility and anillin-mediated pathways.
HMR-1 exclusion operates independently of the septin/anillin-mediated pathway
Recent work has shown that anillin/ANI-1-dependent mechanisms can compensate for the loss of CYK-1–polymerized unbranched actin when septin/UNC-59 function is reduced, thereby restoring cytokinetic ring constriction (Jordan et al, 2016; Lebedev et al, 2023). To determine whether exclusion of HMR-1 contributes to this compensatory pathway, we analyzed unc-59;rol-6(RNAi) in embryos depleted of CYK-1 in cyk-1(ts) allele or simultaneous RNAi. Consistent with previous reports, co-depletion of UNC-59 and CYK-1 accelerated furrow closure (322 ± 38 s) compared with embryos depleted of CYK-1 alone (509 ± 144 s) (Figs 6E and F and S6F) (Lebedev et al, 2023). Despite this faster kinetics, HMR-1 remained detectable within the equatorial division zone of embryos lacking CYK-1 and UNC-59 (Figs 6E and S6F), indicating that the rescue of cytokinesis in these embryos occurs independently of HMR-1. We confirmed this by quantifying ring closure in embryos depleted of CYK-1, UNC-59, and HMR-1, which exhibited even faster ingression (291 ± 33 s) compared with embryos lacking CYK-1 and UNC-59 but not HMR-1 (Fig 6E and F). Efficient depletion of HMR-1 was verified by loss of HMR-1::GFP signal at the cortex (Fig 6E, bottom panel).
Remarkably, cytokinesis in embryos lacking CYK-1 and co-depleted of UNC-59 and HMR-1 occurred with kinetics comparable to control embryos (271 ± 25 s; Figs 6D and E and S6F), demonstrating near-complete suppression of cytokinesis delay caused by CYK-1 depletion. These findings indicate that HMR-1 and UNC-59-ANI-1 pathways regulate furrow ingression through distinct and parallel mechanisms. Thus, although the contractile actomyosin ring assembly (mediated by RHO-1, ANI-1, CYK-1, NMY-2, PLST-1, etc.) promotes furrow ingression, medial localization of HMR-1 inhibits cytokinesis (Fig S6G).
Discussion
HMR-1 is restricted from localizing to the medial furrow zone
Cortical actin architecture significantly influences the clustering and mobility of both junctional and nonjunctional E-cadherin molecules at the cell surface (Sako et al, 1998; Wu et al, 2015; Li et al, 2021). In C. elegans embryos, HMR-1 assembles into clusters that are transported along the plasma membrane by cortical actomyosin flows (Padmanabhan et al, 2017). Actin flows directed toward the cytokinetic furrow have been observed in various cell types, including C. elegans embryos (Cao & Wang, 1990; Reymann et al, 2016; Pinheiro et al, 2017; Khaliullin et al, 2018; Hsu et al, 2023). Because HMR-1 localization on the cell surface mirrors cortical flow direction, this would suggest that actin flow during cytokinesis should facilitate the localization of HMR-1 in the furrow zone and at the newly formed cell–cell interface formed by plasma membrane invagination. Contrary to this expectation, HMR-1 is absent from the cytokinetic furrow and the invaginating membrane during ring constriction. Instead, HMR-1 localizes to the junction between newly divided cells only after cytokinesis is complete. This behavior contrasts with that of SAX-7, the L1-IgCAM cell adhesion receptor, which localizes to the cell division site and the invaginating membrane. In adult tissue, SAX-7 associates with F-actin via DYS-1/dystrophin and UNC-44/ankyrin in adult stages (Zhou et al, 2008; Zhou & Chen, 2011). The distinct behavior of HMR-1 and SAX-7 in early embryos may therefore reflect differences in adaptor-mediated coupling to F-actin. Consistent with this idea, depletion of HMP-1, the cytoplasmic adaptor connecting HMR-1 to F-actin, results in accumulation of HMR-1 at the division zone (Fig S4F).
The absence of HMR-1 in the contractile zone during ring assembly and ingression suggests an active mechanism regulating HMR-1 dynamics. We investigated four possibilities: (a) endocytosis and trafficking of HMR-1 away from the medial cortex, (b) selective degradation of HMR-1 at the furrow zone, (c) anterior localization of HMR-1 after polarity establishment, and (d) physical restriction on the lateral mobility of HMR-1 during the furrow-directed cortical flow. A targeted RNAi screen ruled out the first three possibilities. In contrast, depletion of factors required for assembly and organization of F-actin in the cytokinetic actomyosin ring resulted in a robust (fourfold to sixfold) accumulation of HMR-1 in the division zone. Because a strong depletion of essential genes leads to embryonic lethality, these experiments were performed using mild RNAi. As a caveat, we cannot fully exclude the possibility that incomplete knockdown contributes to HMR-1 exclusion in some conditions.
CYK-1–polymerized unbranched F-actin restricts HMR-1 mobility during cytokinesis
Through physical association, cortical actin has been shown to drive the movement of E-cadherin, as well as act as a fence, limiting the lateral mobility and size of HMR-1/E-cadherin clusters (Sako et al, 1998; Fujiwara et al, 2002; Wu et al, 2015; Padmanabhan et al, 2017). Our biophysical analysis supports such a barrier function at the cytokinetic furrow region. During cytokinesis-associated cortical flows, nonjunctional clusters demonstrated anomalous diffusion within the division zone, suggesting association with the cortical actin network. Disruption of CYK-1–polymerized unbranched actin filaments abolished this behavior, causing these clusters to display Brownian diffusion. On the contrary, ARP2/3 disruption resulted in increased diffusion exponent values of HMR-1 clusters. These observations suggest that association with cortical actin underlies the anomalous obstructive diffusion of HMR-1 clusters. Consistently, HMR-1 clusters displayed lower diffusion coefficients in control and arx-3(RNAi) embryos, which retain CYK-1–polymerized actin filaments, and higher diffusion coefficients in cyk-1(RNAi) embryos, where unbranched actin “corrals” are disrupted. Although our measurements characterize the diffusion parameters of nonjunctional HMR-1, comparable values for E-cadherin engaged in cell adhesion have been previously reported in reconstituted and in vivo systems (Adams et al, 1998; Sako et al, 1998; Biswas et al, 2015).
F-actin crosslinking and myosin activity contribute to medial exclusion of HMR-1
Depletion of NMY-2 also resulted in aberrant HMR-1 accumulation at the cytokinetic ring. Temperature upshift of nmy-2(ne3409) embryos caused HMR-1 to localize within the ingressing membranes, suggesting a role for the motor activity of NMY-2. However, depletion of myosin regulators LET-502 or MLC-4, essential for assembling contractility-competent myosin filaments, did not affect HMR-1 exclusion. This discrepancy may reflect differences in the extent of NMY-2 impairment or engagement of compensatory mechanisms.
Equatorial accumulation of HMR-1 in plst-1(RNAi) embryos further highlights the importance of actin filament crosslinking in restricting HMR-1 mobility. In addition to driving F-actin contractility, NMY-2 organizes medial actin filaments into a dense, compact, parallel alignment during actomyosin ring assembly (Leite et al, 2020). We propose that disruptions in actin organization, as observed in plst-1(RNAi) or nmy-2(RNAi) embryos, create gaps in the cortical actin network that permit diffusion of HMR-1 clusters into the furrow. Given the interdependence of actin organization and myosin contractility, disentangling their individual contributions to HMR-1 exclusion will require further investigation.
We also ruled out the possibility that cytokinetic delays alone allow sufficient time for HMR-1 accumulation in the ingressing furrow. Embryos depleted of LET-502 and TPXL-1 require approximately twice as long to complete cytokinesis post-anaphase onset, yet do not accumulate HMR-1 in the division zone. These embryos assemble the cytokinetic actin ring with largely intact filament organization, and HMR-1 remains excluded. In contrast, embryos with disrupted actin crosslinking (plst-1(RNAi)) or impaired HMR-1–F-actin coupling (hmp-1(RNAi)) or mild pfn-1(RNAi) display pronounced (4–6 fold) accumulation of HMR-1 in the ingressing membrane (Fig 6A, highlighted in green zone). Notably, these embryos divide faster than tpxl-1(RNAi) or let-502(RNAi) embryos, reinforcing the conclusion that compact actin architecture—rather than cytokinesis duration—restricts HMR-1 entry into the medial cortex.
HMR-1 localized to the anterior cortex and the medial cytokinetic zone modulates cytokinesis
Loss of CDC-42 disrupts embryonic polarity and results in uniform surface distribution of HMR-1 clusters. Nevertheless, cdc-42(RNAi) embryos exclude HMR-1 from the ingressing membrane during cytokinesis, similar to control embryos. This observation argues against polarity-dependent sequestration as a mechanism for medial exclusion and supports a dominant role for actin organization at the division site. Importantly, co-depleting HMR-1 rescued cytokinesis failure in cyk-1(RNAi) embryos. The extent of HMR-1 accumulation in the medial division zone correlates with the severity of cytokinesis failures (Fig 6C). Given that HMR-1 clusters localize to the division zone and anterior surface of cyk-1(RNAi) embryos (Fig 4C), the specific contribution of these two populations in regulating cytokinesis remains unclear. HMR-1 can associate with cortical actin in both regions. We therefore propose that both medial and anterior HMR-1 pools could contribute to the observed slowdown of cytokinesis in cyk-1(RNAi) embryos (Fig 6G).
This behavior parallels the exclusion of DE-cadherin from the ingressing membrane during Drosophila epithelial divisions, a process mediated by anillin- and septin-dependent actomyosin flows (Guillot & Lecuit, 2013; Pinheiro et al, 2017). Unlike Drosophila epithelia, however, early C. elegans lack polarized adherens junctions and instead display nascent HMR-1 clusters at intercellular boundaries. Whether actomyosin flow-mediated mechanisms contribute to junctional HMR-1 remodeling in C. elegans remains an open question.
Reciprocal antagonism between cell division and cell adhesion
What might be the physiological relevance of restricting HMR-1 from entering the cytokinetic furrow zone? Early C. elegans embryos undergo rapid cell cycles with minimal G1 and G2 phases, alternating primarily between S and M phases (Brantley & Di Talia, 2021). Accurate and timely positioning, assembly, and constriction of the cytokinetic actomyosin ring are thus essential for division fidelity (Carvalho et al, 2009). Disruption of actin architecture compromises myosin contractility and slows down furrow ingression. Our findings suggest that HMR-1 localization within the invaginating membrane independently contributes to delayed ring constriction.
We speculate two possibly overlapping mechanisms could underlie the inhibitory effect of HMR-1 accumulation at the division site. First, HMR-1 receptors on opposing ingressing membranes could engage “en face” trans-adhesive interactions, increasing mechanical resistance to inward-directed contractile forces generated by the cytokinetic ring. Second, numerous cytoplasmic components—including RHO-1, myosin II, ECT-2, anillin, septins, citron kinase, and formins—are shared between cadherin adhesome and the cytokinetic machinery (Ratheesh et al, 2012; Guo et al, 2014; Padmanabhan et al, 2015; Jones et al, 2019). Accumulation of the cadherin receptors at the division site may therefore sequester these components away from the cytokinetic apparatus, impairing efficient ring constriction.
Our previous work showed that cortical actin–HMR-1 association impedes actomyosin flows and inhibits furrow ingression (Padmanabhan et al, 2017). Here, we demonstrate that the compact architecture of cortical actin filaments within the cytokinetic ring forms a mechanical barrier that restricts HMR-1 mobility at the furrow. Thus, the actomyosin ring not only drives membrane constriction but also excludes negative regulators of cytokinesis from the division zone. Together, our results support the hypothesis of reciprocal regulation and provide a mechanistic framework for the mutual antagonism between E-cadherin–mediated cell adhesion and actomyosin-based cell division.
Materials and Methods
C. elegans culturing and maintenance
All C. elegans strains were grown at 20°C on nematode growth media (NGM) seeded with OP50 E. coli. Temperature-sensitive strains were maintained at 15°C. C. elegans strains used in this study are listed in Table S1.
Table S1. List of. C. elegans strains used in this study.
RNA interference
Table S2 details the plasmids and primers used for RNA interference in this study. RNAi via the feeding method was employed for the candidate screening described in Fig S2H. In vitro*–*synthesized dsRNAs were microinjected for all other experiments. HTT5 E. coli cells expressing target candidate-specific dsRNA (rab11.1, chc-1, unc-11, cav-1, zen-4, cyk-4, dhc-1, jac-1, rac-1, cgef-1, nop-1, pat-3, pph-6, ani-1, unc-60, vab-10, vav-1, wip-1, wve-1, mel-11) in L4440 were grown in LB containing 100 μM ampicillin for 12–14 h at 37°C. This was subcultured to 0.5OD and induced with 1 mM IPTG for 3 h. The bacterial cultures were spotted on NGM plates, which were supplemented with 100 μM ampicillin and 1 mM IPTG. Subsequently, synchronized L4-staged larvae were transferred to appropriate RNAi plates. Hermaphrodites were dissected after 36–58 h of feeding on the RNAi plate, and appropriate cell division–staged embryos were collected, mounted on 3% agarose pads, and imaged. All the feeding constructs were obtained from the Ahringer library (Fraser et al, 2000; Kamath et al, 2003).
Table S2. List of oligonucleotides used for dsRNA synthesis.
dsRNA synthesis was carried out using the MEGAscript T7 kit (Thermo Fisher Scientific). Amplified PCR products derived from RNAi plasmids (Ahringer Library) or N2 genomic DNA were used as templates. Details of primers used are listed in Table S1. dsRNAs were injected into the L4-stage hermaphrodites at the concentration of ∼1 μg/μl using Eppendorf FemtoJet mounted on an IX53 Olympus microscope. Phenotypic confirmation of RNAi-mediated knockdown was carried out as follows. Images of embryos that displayed loss-of-function phenotypes were considered for analysis. Embryos, subsequent to imaging on agarose pads, were hydrated, sealed, and stored at 20°C until hatching. Loss-of-function phenotypes were microscopically examined and confirmed the next day. Phenotypic details of RNAi experiments are listed in Table S3.
Table S3. List of phenotypes observed after dsRNA microinjection.
To carry out simultaneous knockdown of two (double knockdown) or three (triple knockdown) mRNA transcripts, equimolar amounts of dsRNAs were mixed and injected. rol-6 dsRNA was used as a dilution control.
Live imaging and image acquisition
After 24–48 h postinjection, hermaphrodites were dissected and appropriate stage embryos were collected, mounted on 3% agarose pads, and imaged.
All imaging was carried out on Olympus IX83 Inverted Microscope equipped with CSU-W1 spinning-disk confocal head (Yokogawa Corporation) using a 100× UPlanSApo 1.40 objective. Embryo samples were excited at 488 (for GFP) and 561 nm (for RFP/mCherry) using the Coherent OBIS laser system. Images were acquired using the PrimeBSI sCMOS camera (Photometrics). Fig 3D was acquired using the Evolve delta EMCCD camera (Photometrics). Image acquisition was controlled by Olympus cellSens Dimension software. All images acquired using PrimeBSI were acquired with the camera set at 1,024 × 1,024 pixels.
Imaging was primarily carried out at 20–22°C. let-502(ts) and cyk-1(ts) were imaged at 25°C and 20°C, respectively. In this case, few late L4 worms are transferred to an OP50-seeded NGM plate at the restrictive temperature. Worms were then imaged after 16–20 h (let-502(ts)) or 5 min (cyk-1(ts)) of incubation at appropriate restrictive temperatures. During quantification of HMR-1 and NMY-2 intensities at the ingressing membrane, ROIs of the medial plane were taken from the penultimate frame in a time lapse (acquired every 10 s) before cytokinetic ring closure.
Image analysis
Image analysis and quantifications were carried out using Fiji (Schindelin et al, 2012), Python, and Microsoft Excel; Jupyter Notebook (Python) and Prism 6 (GraphPad) software packages were used for graph plotting and statistical analysis. All embryo images presented in the figures were rotated and cropped to orient anterior to the left. These images were subjected to Gaussian filter—0.75 pixel. And brightness and contrast were set to same values in control and RNAi conditions.
Intensity line-scan plots and co-localization analysis
To estimate the intensity profile of NMY-2::mCherry and HMR-1::GFP in the ingressing membrane, a 5-pixel wide and 100-pixel-long ROI was drawn along the ingressing furrow. Fluorescence intensity of both channels was measured along the ROI. For background fluorescence calculation, a similar ROI was placed in the cytoplasmic region in embryonic and away from the ingressing furrow, and these intensity values were subtracted from furrow values. Mean ± 95% CI intensities were plotted.
Estimating fold change in intensity of HMR-1 and NMY-2
Fold change in intensities of HMR-1 and NMY-2 was estimated to compare the effect of specific RNAi on HMR-1 localization in the furrow compared with the mock control. A specific ROI is drawn along the ingressing furrow in both channels. For background, ROI was placed in the embryonic surface away from the ingressing furrow, and intensities from both channels were subtracted. This protocol was followed for the control and all the RNAi; the intensity in RNAi is normalized with the control intensity.
Measurement of medial actin filament orientation
To analyze the orientation of actin filaments, embryos expressing Lifeact::RFP were imaged and a rectangular ROI was selected covering the cortical plane of the division zone. For each ROI, the filament orientation information was extracted using “OrientationJ (OJ) Distribution (Rezakhaniha et al, 2012)” plugin in Fiji. This workflow returns coherency-weighted histogram of orientations. Coherency indicates orientation of local image features. Parameters optimized for the workflow using OJ Distribution plugin were as follows: orientation—degree (−90° to 90°); gradient—cubic spline; local window σ = 5 pixels (scale 1 pixel = 0.13 μm); minimum energy—1%; minimum coherency—5%. A custom-made Python code was used to carry out the following analysis. First, the absolute value of the filament orientation histogram (0°–90°) was extracted from OJ Distribution data. Using these data, a normalized radial histogram of filament orientation was made for each embryo during actomyosin ring assembly. Thereafter, to combine orientation data for all the embryos into a single radial histogram, average of each bin (bin width = 15°) radial histogram was calculated. These data were used to plot Fig 5A and B. To calculate the average across multiple embryos, mean radial histogram with each bin containing the average of that particular bin for all embryos was obtained. This was normalized (to the average total filaments across all embryos) and plotted to obtain a normalized mean radial histogram, which is shown in Fig 4F. This histogram shows the average distribution of orientation of actin filaments across different bins.
Speed and MSD measurement of HMR-1 clusters
TrackMate plugin (Tinevez et al, 2017) in Fiji was used to track HMR-1::GFP clusters in a 28 × 16 μm ROI covering the medial zone. LoG (Laplacian of Gaussian) detector was used to select the HMR-1::GFP clusters in the ROI, using the diameter of blob∼0.9 μm and threshold value 350–450 to 65,535 (16-bit images). LAP tracker was used to track each cluster’s trajectory. 4 μm was taken as the linking distance. Image acquisition was carried out every 5 s. The penalties used are Std intensity ch 1:1.0, median intensity ch 1: 1.0, Visibility: 1.0, radius: 1.0, signal/noise ratio: 1. No gap closing distance was used. TrackMate returned tracking data in three .csv files, namely, Spot, Edge, and Track statistics.
Instantaneous speeds in x and y direction were measured for clusters in control and cyk-1(RNAi) embryos using the TrackMate data obtained above. Instantaneous cluster velocities were computed from tracked Spot and Edge statistics file by measuring absolute x- and y-displacements between consecutive frames and dividing by the frame interval (5 s) to obtain |V_X_| and |V_Y_|, respectively. Edge times were aligned to the onset of furrow initiation. Velocity measurements were calculated independently for each embryo and pooled across embryos for each RNAi (cyk-1 RNAi or l4440). For frequency distribution analysis, |V_X_| and |V_Y_|, within the furrow ingression window (∼47–122 s), were combined across embryos and plotted as probability density–normalized histograms (Fig 4B).
To calculate the mean square displacement (MSD), x- and y-coordinates of HMR-1::GFP clusters (extracted from Spot statistics) were used.
To find the diffusion exponent “α” (in Fig 4I) and diffusion coefficient “D” values (in Fig 4J) of HMR-1::GFP clusters in the furrow ingression region, clusters were tracked from ∼40 s until ∼end of cytokinesis in each embryos. The values of slope and y-intercept of the best fit of MSD versus track duration for each embryo was used to find the diffusion exponent, “α” (Fig 4I), and diffusion coefficient “D” (Fig 4J).
Measuring the width of actin-enriched region
To measure the width/spread of actin-enriched region during the assembly of contractile actomyosin ring formation in early embryos (1-cell stage), actin intensity was measured across a 70-pixel-wide ROI placed along the major axis of embryo. These intensity values along the length of embryo were fitted using Gaussian function. We considered an increase in the intensity of actin at the furrow zone. Therefore, the width (d) of the fitted Gaussian at 10% actin intensity was used as a parameter to measure the spread of this actin-enriched region. The peak intensity (y_max_) and the minimum intensity (y_min_) of the fitted Gaussian plot are used to find “d” (d = [(y_max_ − y_min_)*0.1] + y_min_).
Quantification of HMR-1 clusters in the cytokinetic furrow zone
To count the number of HMR-1::GFP clusters at the cytokinetic furrow zone (Fig 4G), cortical plane images were analyzed in FIJI. Images are subtracted for background, filtered (mean filtering), and subjected to intensity thresholding. These images were then subjected to watershed segmentation, and the number of HMR-1::GFP clusters was estimated using “Analyse particles.”
Measurement of HMR-1 fluorescence at the medial cortex
Medial plane images of control or cyk-1(RNAi) embryos co-expressing NMY-2::mCherry, HMR-1::GFP, and Histone::mCherry were acquired every 10 s post-anaphase onset. Ring assembly time was taken as time from anaphase onset until a back-to-back membrane fold was detected. The cortical region between segregating genetic material was used as the furrow zone. HMR-1 intensity was measured over a 10-pixel-wide manually drawn line over the surface. The average of two edges was considered before background subtraction and plotting.
Measurement of kinetics of the furrow closure
Medial plane images of embryos co-expressing NMY-2::mCherry, HMR-1::GFP and Histone::mCherry subjected to appropriate RNAi were acquired every 10 s post-anaphase onset, and the distance between the ingressing furrows was measured. The distance between the two furrows is normalized to the starting width of the embryo.
Supplementary Material
Reviewer comments
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Adams CL, Chen YT, Smith SJ, Nelson WJ (1998) Mechanisms of epithelial cell-cell adhesion and cell compaction revealed by high-resolution tracking of E-cadherin-green fluorescent protein. J Cell Biol 142: 1105–1119. 10.1083/jcb.142.4.11059722621 PMC 2132880 · doi ↗ · pubmed ↗
- 2Ai E, Skop AR (2009) Endosomal recycling regulation during cytokinesis. Commun Integr Biol 2: 444–447. 10.4161/cib.2.5.893119907714 PMC 2775247 · doi ↗ · pubmed ↗
- 3Alonso-Matilla R, Lam AR, Miettinen TP (2024) Cell-intrinsic mechanical regulation of plasma membrane accumulation at the cytokinetic furrow. Proc Natl Acad Sci U S A 121: e 2320769121. 10.1073/pnas.232076912138990949 PMC 11260091 · doi ↗ · pubmed ↗
- 4Bai X, Bouffard J, Lord A, Brugman K, Sternberg PW, Cram EJ, Golden A (2020) Caenorhabditis elegans PIEZO channel coordinates multiple reproductive tissues to govern ovulation. Elife 9: e 53603. 10.7554/e Life.5360332490809 PMC 7340503 · doi ↗ · pubmed ↗
- 5Basant A, Glotzer M (2018) Spatiotemporal regulation of Rho A during cytokinesis. Curr Biol 28: R 570–R 580. 10.1016/j.cub.2018.03.04529738735 PMC 6508076 · doi ↗ · pubmed ↗
- 6Bertocchi C, Wang Y, Ravasio A, Hara Y, Wu Y, Sailov T, Baird MA, Davidson MW, Zaidel-Bar R, Toyama Y, (2017) Nanoscale architecture of cadherin-based cell adhesions. Nat Cell Biol 19: 28–37. 10.1038/ncb 345627992406 PMC 5421576 · doi ↗ · pubmed ↗
- 7Biswas KH, Hartman KL, Yu C-H, Harrison OJ, Song H, Smith AW, Huang WYC, Lin W-C, Guo Z, Padmanabhan A, (2015) E-cadherin junction formation involves an active kinetic nucleation process. Proc Natl Acad Sci U S A 112: 10932–10937. 10.1073/pnas.151377511226290581 PMC 4568248 · doi ↗ · pubmed ↗
- 8Brantley SE, Di Talia S (2021) Cell cycle control during early embryogenesis. Development 148: dev 193128. 10.1242/dev.19312834164654 PMC 8255047 · doi ↗ · pubmed ↗