HDAC1/2-mediated repression of Wnt receptor expression orients asymmetric division polarity in Caenorhabditis elegans

ABSTRACT

Asymmetric cell division generates distinct daughter cells essential for tissue development, yet the mechanisms orienting division polarity within tissues remain incompletely understood. Here, we uncover a role for chromatin-mediated transcriptional repression in controlling polarity orientation during asymmetric division of Caenorhabditis elegans epidermal stem cells, known as the seam cells. Tissue-specific loss of the class I histone deacetylase hda-1, homologous to mammalian HDAC1/2, causes polarity reversals and reduces molecular asymmetry between daughter cells. Using Targeted DamID, we identify the Wnt receptors lin-17 (frizzled homologue) and cam-1 (Ror homologue) as key targets. Single-molecule fluorescence in situ hybridisation reveals opposing expression gradients along the body axis, with cam-1 enriched anteriorly and lin-17 posteriorly. In hda-1 mutants, both receptors are upregulated and differences between seam cells are flattened. Overexpression of either receptor alone is sufficient to reproduce the polarity reversals, while co-overexpression produces additive effects. The polarity phenotype is independent of the canonical NuRD and SIN3 complexes, suggesting that HDA-1 acts through an alternative mechanism. These findings link histone deacetylase activity to Wnt receptor expression and suggest that graded receptor expression may provide cues that orient asymmetric division polarity.

Summary: Targeted DamID and single-molecule fluorescence in situ hybridisation reveal that a histone deacetylase influences stem cell orientation by keeping Wnt signalling receptor expression in check.

INTRODUCTION

The development of multicellular organisms relies on the generation of diverse cell types organised into functional tissues. Stem cells play a crucial role in this process by undergoing asymmetric divisions, which generate daughter cells capable of differentiation as well as symmetric divisions that expand the stem cell population (Slack, 2018). In addition to generating daughter cells with distinct fates, asymmetric divisions must be precisely oriented within tissues. Proper orientation of polarity ensures that daughter cells are positioned appropriately and exposed to the correct extrinsic cues, as the directional specification of daughter cell fates determines their spatial organisation within tissues and establishes tissue architecture (Gómez-López et al., 2014; Morrison and Kimble, 2006; Williams and Fuchs, 2013). Defects in division orientation and polarity can disrupt tissue homeostasis and contribute to disease, including cancer, where loss of polarity and dysregulation of asymmetric division are frequently observed (Bajaj et al., 2020; Yamashita et al., 2010). Understanding the genes and pathways that regulate these fundamental processes is therefore crucial for elucidating mechanisms underlying both normal development and disease states.

Caenorhabditis elegans epidermal seam cells provide an excellent system for investigating stem cell behaviour and asymmetric division (Joshi et al., 2010). The seam cells are arranged in a linear pattern along the body axis (H0-H2, V1-V6 and T) and undergo both symmetric and asymmetric divisions throughout postembryonic development (Fig. 1A). During asymmetric division, posterior daughters typically retain seam cell fate and express genes such as those encoding the GATA factors ELT-1, EGL-18 and ELT-6, which promote seam cell fate maintenance (Gorrepati et al., 2013; Koh and Rothman, 2001; Smith et al., 2005). Seam cells contribute most of the epidermal nuclei during post-embryonic development and hence produce the protective cuticle (Chisholm and Hsiao, 2012). Additionally, specific seam cell lineages (H2, V5 and T) generate neuronal precursors (Sulston and Horvitz, 1977). During the early L2 stage, a symmetric division expands the seam cell population from 10 to 16 cells, with both daughters retaining stem-like properties. At the end of larval development, these 16 cells undergo terminal differentiation and fuse to form a syncytium that produces the alae, which are specialised ridges in the cuticle (Altun and Hall, 2002). The highly reproducible nature of seam cell division patterns and their fixed anterior-posterior positioning make them an excellent system for detailed analysis of developmental defects at single-cell resolution (Katsanos et al., 2017).

Fig. 1.

Analysis of hda-1 expression at the mRNA and protein level in seam cells. (A) Schematic showing seam cell divisions from L1 to L4. Seam cells are labelled in green, hypodermal cells are labelled in orange and neuroblasts are labelled in purple. (B) Seam cell number counts upon RNAi targeting members of the HDAC family in C. elegans. Only downregulation of hda-1 leads to a significant increase in the terminal seam cell number (*P<0.05 with a Krustal-Wallis test followed by Dunn's post-hoc test, 65≤n≤96 animals per condition). (C) Representative hda-1 smFISH image at late L1 stage. (D) Quantification of hda-1 mRNA spots in all seam cell lineages at late L1 (21≤n≤45 cells per condition). (E,F) Representative hda-1 smFISH images following the L2 symmetric division (E) and the L2 asymmetric division (F). (G) Quantification of hda-1 mRNA spots in V1-V4 seam cells before (p) and after the L2 symmetric division (pa and pp), or the L2 asymmetric division (paa, pap, ppa and ppp); n=166 (p), n=51 (pa and pp), n=90 (paa, pap, ppa and ppp) animals per condition. (H) Representative image of HDA-1::GFP at late L1. (I,J) Representative HDA-1::GFP images after the L2 symmetric (I) and L2 asymmetric (J) divisions. (K) Quantification of HDA-1::GFP fluorescence intensity following the L2 symmetric and L2 asymmetric divisions in V1-V4 lineages. Following symmetric division, anterior cells (pa) show similar expression of HDA-1 compared to posterior cells (pp) (n=71 cells per condition). After L2 asymmetric division, posterior daughter cells (pap and ppp) show higher HDA-1 expression compared to anterior cells (paa and ppa) (****P<0.0001 with a paired two-tailed t-test, n=61 cells per condition). Seam cell nuclei in C, E and F are labelled using SCM::GFP. In H, I and K, the membrane of the seam cells is labelled using the apical junction marker ajm-1p::ajm-1::mCherry. Scale bars: 5 μm in C,E,F,H-J. Data in B,D,G and K are mean±95% confidence intervals. In G,K, we calculated within-pair differences, tested these differences for normality and then used paired two-tailed t-tests or Wilcoxon signed-rank tests, as appropriate (see also Table S4). ****P<0.0001 and *P<0.05.

Correct seam cell patterning depends on a divergent version of the canonical Wnt signalling pathway called the Wnt/β-catenin asymmetry pathway (WβA), which ensures that, following asymmetric division, posterior daughters activate Wnt target genes and retain seam cell fate, while anterior daughters repress Wnt targets and differentiate (Jackson and Eisenmann, 2012; Sawa and Korswagen, 2013). Multiple Wnt ligands, including LIN-44, CWN-1, CWN-2, EGL-20 and MOM-2, are expressed in distinct body regions and act redundantly to control the orientation of seam cell polarity along the anterior-posterior axis (Harterink et al., 2011; Pani and Goldstein, 2018; Sawa et al., 2026; Yamamoto et al., 2011). Wnt receptors, including the Frizzled receptors LIN-17 and MOM-5, and the Ror receptor CAM-1 are essential for establishing cellular polarity (Forrester et al., 1999; Goldstein et al., 2006; Sawa et al., 1996, 2026; Yamamoto et al., 2011). In dividing seam cells, Wnt signalling promotes posterior cortical localisation of Wnt receptors and Dishevelled proteins, while negative regulators, such as PRY-1 (Axin homologue) and APR-1 (APC homologue), are enriched at the anterior cortex (Baldwin and Phillips, 2014; Mizumoto and Sawa, 2007a,b). This asymmetry generates daughter cells with distinct nuclear levels of POP-1 (TCF homologue) and SYS-1 (β-catenin homologue) (Mizumoto and Sawa, 2007b). Low POP-1 and high SYS-1 in posterior daughters activate Wnt target genes, whereas high POP-1 and low SYS-1 in anterior daughters lead to transcriptional repression (Bekas and Philips, 2022 preprint; Huang et al., 2007; Vora and Phillips, 2015).

While transcriptional regulation of seam cell development has been extensively studied, the role of chromatin-based mechanisms in this process remains largely unexplored. Components of the SWI chromatin remodelling complex have been implicated in asymmetric division of the T cell lineage (Cui and Han, 2007; Shibata et al., 2012), and the acetylated-histone-binding protein BET-1 contributes to fate maintenance in T cell progeny (Shibata et al., 2010). However, the broader impact of chromatin modifications on the expression of seam cell regulators remains unclear. Class I histone deacetylases (HDACs) regulate gene expression by removing acetyl groups from histone tails, promoting chromatin compaction and transcriptional repression (Seto and Yoshida, 2014; Shahbazian and Grunstein, 2007). These enzymes typically function within multiprotein repressor complexes such as SIN3, NuRD, CoREST, MiDAC, MIER, RERE and SMRT/NCoR, and play crucial roles in promoting cellular differentiation during animal development (Ahringer, 2000; Brunmeir et al., 2009; Derwish et al., 2017; Dovey et al., 2010b; Jaju Bhattad et al., 2020; Li et al., 2000; Plaster et al., 2007; Turnbull et al., 2020). The C. elegans class I histone deacetylase HDA-1, homologous to mammalian HDAC1 and HDAC2, has been implicated in the regulation of multiple developmental processes. During embryogenesis, HDA-1 acts together with UNC-37 (Groucho homologue) as a POP-1 co-repressor downstream of the Wnt/β-catenin asymmetry pathway to repress endoderm fate in the MS blastomere (Bekas and Philips, 2022 preprint; Calvo et al., 2001). HDA-1 also contributes to nervous system development by regulating cell migration and axon guidance (Zinovyeva et al., 2006) and, as part of the SIN3 complex, participates in the formation of male sensory rays (Choy et al., 2007). Furthermore, HDA-1 plays crucial roles in gonadogenesis and vulval development, including promoting the invasive behaviour of the anchor cell (Matus et al., 2015) and preventing ectopic vulval fate specification as part of the SynMuvB pathway (Dufourcq et al., 2002; Solari and Ahringer, 2000). Transcriptional profiling in the epidermis has shown that hda-1 is expressed in seam cells and that its downregulation affects terminal seam cell number (Katsanos et al., 2021). Despite these observations, the specific role of HDA-1 in regulating stem cell behaviour and asymmetric division has not been explored.

In this study, we identify a role for HDA-1 in regulating seam cell polarity orientation in C. elegans. We report that loss of hda-1 disrupts the orientation of asymmetric divisions and results in defects in molecular asymmetry between daughter cells. Using Targeted DamID, we find the Wnt receptors lin-17/Frizzled and cam-1/Ror as HDA-1 targets. Single-molecule fluorescence in situ hybridisation (smFISH) analysis reveals elevated expression of lin-17 and cam-1 in hda-1 mutants, reducing differences in expression observed in wild type across V1-V4 lineages. Furthermore, we show that overexpression of these Wnt receptors is sufficient to phenocopy the hda-1 mutant polarity defects. Together, our findings uncover a novel role for HDA-1 in ensuring proper orientation of asymmetric divisions, likely through regulation of Wnt receptor expression.

RESULTS

hda-1 expression is largely uniform in seam cells

We previously identified through transcriptional profiling hda-1 as a seam cell-expressed chromatin regulator. We reported that RNAi-mediated downregulation of hda-1 using a seam cell-driven hairpin construct caused seam cell hyperplasia (Katsanos et al., 2021). This phenotype is specific to hda-1 as RNAi targeting of other HDAC family members did not lead to statistically significant changes in seam cell number (Fig. 1B), so we decided to characterise in more detail the role of HDA-1 in seam cells.

To investigate the expression pattern of hda-1 in seam cells, we first quantified hda-1 transcripts using smFISH. hda-1 mRNA was detected in all seam cell lineages (Fig. 1C,E,F). At the late L1 stage, transcript abundance was highest in the mid-body V1-V5 seam cells, with lower levels in the H and T lineages at the head and tail (Fig. 1C,D). We next examined hda-1 expression during V1-V4 seam cell divisions. Following the L2 symmetric division, hda-1 transcript levels were similar between anterior and posterior daughters (Fig. 1E,G). After the L2 asymmetric division, hda-1 mRNA remained detectable in both daughters, with only mild, on average, enrichment in anterior cells (Fig. 1F,G); hence, hda-1 transcript is largely uniform between seam cell daughters post division.

To assess HDA-1 distribution at the protein level, we analysed an endogenous HDA-1::GFP reporter. HDA-1::GFP localised to the nuclei of all seam cells and hypodermal cells (Fig. 1H) with levels that were comparable between daughters after symmetric division (Fig. 1I,K). However, after asymmetric division, HDA-1::GFP showed moderate enrichment in posterior daughter cells (Fig. 1J,K). This pattern is not reflected at the mRNA level, so it suggests that HDA-1 may be subject to translational or post-translational regulation, as has been reported in other contexts (Poulin et al., 2005; Sengupta and Seto, 2004). Such regulation could affect its stability, activity or interactions with partners during asymmetric division.

hda-1 mutants display reversals in the polarity of asymmetric divisions

To investigate the function of HDA-1 in seam cell development, we generated a seam cell-specific hda-1 mutant using the Cre-lox recombination system. We confirmed that neither the introduction of loxP sites flanking the hda-1 locus nor the expression of Cre recombinase in the seam cells disrupted seam cell development (Fig. S1A). Furthermore, hda-1^loxP animals expressing Cre recombinase showed complete loss of hda-1 transcripts in seam cells by smFISH, while expression remained detectable in the surrounding epidermis, confirming the tissue specificity of the knockout (Fig. S1B).

Seam cell-specific hda-1 mutants exhibited clear patterning defects, with both extra cells and gaps in the seam cell line at the early adult stage, suggesting that loss of hda-1 disrupts the balance between seam cell self-renewal and differentiation (Fig. 2A). Quantification revealed an overall increase in mean seam cell number (Fig. 2B). To determine when seam cells are gained or lost in hda-1 mutant development, we scored seam cell number at late L2, L3 and L4 stages. We found that changes in seam cell number were restricted to the L4 stage (Fig. 2B).

Fig. 2.

A seam cell-specific hda-1 mutant displays gains and losses of seam cells. (A) Representative images of tissue-specific hda-1 mutant animals [simplified here as hda-1(-)] versus floxed animals not expressing the Cre recombinase (wild type) at the early adult stage. (B) Seam cell counts in tissue-specific hda-1 mutant animals at late L2, late L3 and late L4 stages. Late L2, *P<0.05; late L4, ****P<0.0001 with a Mann–Whitney test; n≥30 animals per condition. Data are mean±95% confidence intervals. (C) Representative images of hda-1 mutant defects at the early adult stage. Membrane breaks are represented with a blue bracket, mis-shapen membrane is indicated by an orange arrowhead, long connections are indicated by a blue arrowhead and membrane boundaries indicative of delayed cell fusion are indicated by a white arrowhead. The membrane of the seam cells is labelled using the apical junction marker ajm-1p::ajm-1::GFP. (D) Table showing the percentage of animals displaying at least one instance of each defect in hda-1 mutants (n=20 animals) compared to wild type (n=21 animals). In A and C, seam cell nuclei are labelled using SCM::GFP. Scale bars: 50 μm in A; 10 μm in C. See also Fig. S1.

At the late L4-early adult stage, seam cells normally fuse to form a syncytium, which can be visualised with the apical junction AJM-1::GFP marker as two perpendicular membrane lines with seam cell nuclei enclosed within (Sulston and Horvitz, 1977). In hda-1 mutants, however, we observed an array of defects, including membrane breaks, mis-shapen membrane, long connections between separated cells and delayed fusion with persistent boundaries between seam cells (Fig. 2C). These defects were highly penetrant, with 95% of hda-1 mutant animals displaying at least one type of abnormality (Fig. 2D). These strong phenotypes by the L4 stage were reminiscent of polarity defects and led us to explore whether earlier errors might also be present that do not directly impact seam cell number.

To investigate whether the stereotypic polarity of asymmetric divisions was affected in hda-1 mutants, we examined expression of the fusogen eff-1, which is normally restricted to differentiating anterior daughter cells following the L2 asymmetric division (Koneru et al., 2021). Because eff-1 expression is essentially all or none within each lineage, it provides an unambiguous, binary read-out of complete polarity reversal. In wild-type animals, eff-1 expression was exclusively observed in anterior daughter cells, as expected (Fig. 3A). In contrast, hda-1 mutants showed reversed eff-1 expression in some lineages, with posterior daughters expressing eff-1 while anterior daughters did not (Fig. 3A). Quantification revealed that 18% of hda-1 mutants displayed this reversal in the V2 lineage (Fig. 3B).

Fig. 3.

hda-1 mutants show polarity reversals upon asymmetric cell division. (A) Representative images of eff-1 smFISH following the L2 asymmetric division in wild-type and in tissue-specific hda-1 mutant animals [hda-1(-)]. Orange bracket indicates a division with reversed expression of eff-1. (B) Heatmap showing the frequency of cells expressing eff-1 following the L2 asymmetric division in wild type and hda-1 mutants (26≤n≤38 cells per condition). Expression frequencies significantly differed between wild type and hda-1 mutants in pap cells (Fisher's exact test). (C-H) Representative images (C,E,G) and quantification (D,F,G) of mRNA spots in wild-type (blue) and in hda-1 mutant animals (red) following the L2 asymmetric division for nhr-73 (C,D), elt-1 (E,F) and efl-3 (G,H). Orange brackets in C,E,G show a pair with reversed expression of the corresponding gene. Data were analysed using two-way mixed ANOVA followed by planned comparisons as described in the Materials and Methods. Corresponding anterior-posterior asymmetry distributions are shown in Fig. S1. 94≤n≤110 cells in D, 170≤n≤174 cells in F and 104≤n≤114 cells in G. In A,C,E and G, seam cell nuclei are labelled using SCM::GFP. Scale bars: 5 μm. Data in D, F and H show the mean±95% confidence intervals. ****P<0.0001, **P<0.01 (paired two-tailed t-tests for anterior-posterior comparisons; unpaired two-tailed Welch's t-tests for WT-hda-1(-) comparisons).

We next investigated whether molecular asymmetry between daughter cells was affected by assessing expression of genes that are differentially expressed between anterior and posterior daughter cells following asymmetric division. We examined nhr-73 and elt-1, which are distinctly enriched in posterior daughter cells (Miyabayashi et al., 1999; Smith et al., 2005), and efl-3, which is enriched in anterior daughter cells (Ferrando-Marco and Barkoulas, 2025) (Fig. 3C,E,G). Unlike eff-1, these genes are expressed in a more graded manner, so changes in their expression are better suited for assessing the degree of molecular asymmetry between daughter cells than for scoring discrete reversal events. In all cases, we found that molecular asymmetry between daughter cells was significantly reduced in hda-1 mutants (Fig. 3D-H; Fig. S1C-E). For nhr-73 and elt-1, we additionally scored, for each V1-V4 lineage, the percentage of pairs in which the normal posterior-enriched expression pattern was inverted (Fig. S1F). Reversal-like events were found across V1-V4 but were most enriched in the V2 lineage, consistent with the eff-1 expression data. Taken together, these findings suggest that loss of hda-1 leads to polarity defects and reduction in molecular asymmetry during asymmetric division, which precede and may contribute to the gains and losses of seam cells observed at later developmental stages.

Identification of putative HDA-1 targets in seam cells via TaDa

To identify downstream targets of HDA-1 in the seam cells, we performed Targeted DamID (TaDa) profiling. We expressed an hda-1::dam fusion at very low levels using a bicistronic TaDa cassette under the seam-cell promoter srf-3i, following established approaches to minimise Dam-associated toxicity (Katsanos and Barkoulas, 2022; Southall et al., 2013). Indeed, expression of the hda-1::dam fusion did not affect developmental speed or brood size (Fig. S2A,B). A nuclear GFP-Dam fusion expressed from the same cassette served as the control. Animals were collected at the late L3 stage, and genomic regions methylated by HDA-1::Dam were isolated and sequenced.

Aggregate genome-wide signal profiles across genes containing peaks between 6 kb upstream of the transcriptional start site (TSS) and 2 kb downstream of the transcription end site (TES) showed that HDA-1 binding is enriched within gene bodies (Fig. S2C). We identified 1886 statistically significant peaks (FDR<0.05) (Table S1). Hierarchical clustering of the localisation and score of statistically significant peaks within genes revealed clusters of signal enrichment upstream of the TSS, inside genes and downstream of the TES regions for HDA-1 TaDa (Fig. S2D). Peaks were assigned to possible genes when the centre of the peak fell between 6 kb upstream of the TSS and 2 kb downstream of the TES, leading to 3195 HDA-1 putative targets. The list of candidate HDA-1 target genes showed significant overlap with genes expressed in seam cells based on single-cell combinatorial indexing RNA-sequencing data (Cao et al., 2017) and with a lab-curated list of genes with known roles in seam cell development from the literature (Table S2). Gene ontology (GO) term analysis of putative HDA-1 targets in the seam cells revealed terms relating to development, Wnt signalling and transcription among others (Fig. S2E).

Loss of HDA-1 function disturbs the expression pattern of the Wnt receptors lin-17 and cam-1

To study how HDA-1 may exert its effect, we focused on Wnt signalling, which has been previously linked to seam cell polarity. We selected several candidates identified through TaDa for expression analysis by smFISH, including the Wnt receptors lin-17, cam-1 and mom-5, as well as the Dishevelled gene dsh-2 and the negative regulators apr-1 and pry-1 (Fig. S2F). These genes were prioritised based on their established, non-redundant roles in seam cell polarity in comparison to other Wnt pathway genes also identified through TaDa (e.g. mig-1, dsh-1 and vang-1; Table S2), which act redundantly or are dispensable in this context (Baldwin et al., 2016; Yamamoto et al., 2011).

Among all candidates examined, lin-17 and cam-1 showed the most striking and consistent expression changes in hda-1 mutants. As the mRNA expression patterns of these Wnt receptors had not been previously quantified in seam cells, we analysed cam-1 and lin-17 expression in wild-type animals. First, cam-1 expression was enriched in anterior seam cell lineages, with the highest expression in V1 and progressively lower levels in V2, V3 and V4 at late L1, and following the L2 symmetric and asymmetric divisions (Fig. S3A-D). Interestingly, following symmetric and asymmetric cell divisions, it was the anterior daughter cells that showed significantly higher cam-1 expression compared to their posterior counterparts (Fig. S3D). In hda-1 mutants, cam-1 transcript levels were significantly elevated in seam cells following both the L2 symmetric and asymmetric divisions (Fig. 4A-D). This upregulation was observed across all V lineages and in both anterior and posterior daughter cells, indicating that HDA-1 normally restricts cam-1 expression. As a consequence, hda-1 mutants showed significantly reduced anterior-posterior (A-P) asymmetry for cam-1 expression (Fig. 4E) and reduced V1-V4 differences compared to wild type (Fig. 4F). Analysis of cam-1 mRNA distribution more broadly across the lateral epidermis was consistent with graded expression along the anterior-posterior axis, with highest expression in the head area and lowest expression in the tail; this A-P gradient was not significantly altered in hda-1 mutants (Fig. 4G). This indicates that upregulation of cam-1 disrupts differences across V1-V4 lineages and asymmetric expression between daughter cells, while the broader A-P expression pattern across the lateral epidermis remains largely unaffected.

Fig. 4.

cam-1 expression is increased in hda-1 mutants. (A,B) Representative cam-1 smFISH images of wild-type and tissue-specific hda-1 mutant animals [hda-1(-)] following the L2 symmetric (A) and L2 asymmetric (B) divisions. Blue arrowheads indicate examples of cells with visually higher cam-1 mRNA spots. Seam cell nuclei are labelled using SCM::GFP. Scale bars: 5 μm. (C,D) Quantification of cam-1 mRNA spots in V1-V4 lineages in wild-type (blue) and hda-1 mutant (red) animals following the L2 symmetric (C) and L2 asymmetric (D) divisions (in C, 24≤n≤32; in D, 24≤n≤27 cells per condition). Wild-type and hda-1(-) values were compared within each cell type and lineage using a Mann–Whitney test or Welch's t-test following normality testing. (E) Normalised anterior-posterior (A-P) asymmetry of cam-1 mRNA following the L2 asymmetric division in wild-type (blue) and hda-1 mutant (red) animals. A-P asymmetry was calculated for each anterior/posterior sister-cell pair as a normalised difference (A-P)/(A+P+1), where values close to 0 indicate equal expression between sisters, positive values indicate higher expression in anterior cells and negative values indicate higher expression in posterior cells (202≤n≤206 cell pairs per condition). (F) Normalised V1-V4 lineage differences in cam-1 mRNA following the L2 asymmetric division in wild type (blue) and hda-1 mutants (red). V1-V4 differences were calculated for matched cell types (paa, pap, ppa and ppp) using a normalised difference (V1-V4)/(V1+V4), where values close to 0 indicate equalisation of expression between anterior (V1) and posterior (V4) lineages (n=92 cell pairs per condition). Normalised A-P asymmetry and V1-V4 difference scores were compared between wild-type and hda-1 mutants using a Mann–Whitney test. (G) Quantification of cam-1 smFISH spot density in the lateral epidermis across the anterior-posterior axis in wild-type and in hda-1(-) animals at late L1 stage; n=21 animals for wild type and n=17 animals for hda-1(-). Lines and ribbons show negative binomial fits ±95% confidence intervals. The genotype×AP-position interaction was not significant by likelihood-ratio test. Comparisons between wild-type and hda-1 mutant represent Wilcoxon rank-sum tests with Holm correction. Data in C-G show the mean±95% confidence intervals. ****P<0.0001, ***P<0.001, **P<0.01 and *P<0.05. See also Fig. S3.

Regarding lin-17, wild-type animals showed enrichment in posterior seam cell lineages, with V4 showing the highest expression and V1 the lowest among V1-V4 following the L2 symmetric and asymmetric divisions (Fig. S4A-D). In contrast to what was seen for cam-1 expression, posterior daughter cells showed significantly higher lin-17 expression compared to their respective anterior daughters after symmetric and asymmetric divisions (Fig. S4D). In hda-1 mutants, lin-17 expression was significantly increased following both the L2 symmetric and asymmetric divisions (Fig. 5A-D). This increase was observed across all V lineages analysed following the L2 symmetric division and mostly in V1 and V2 following the L2 asymmetric division. Analysis of A-P asymmetry revealed that lin-17 expression differences between anterior and posterior daughters remained largely unchanged in hda-1 mutants compared to wild-type animals (Fig. 5E). However, examination of V1-V4 differences showed that hda-1 mutants displayed significantly reduced differences across lineages compared to wild-type (Fig. 5F). Analysis of lin-17 mRNA across the lateral epidermis further confirmed enrichment in the posterior region of the body and a significant flattening of the gradient in hda-1 mutants (Fig. 5G). This indicates that HDA-1 loss disrupts the posterior-anterior gradient of lin-17 expression both within V1-V4 seam cell lineages and across the lateral epidermis, yet does not change daughter cell asymmetry.

Fig. 5.

lin-17 expression is increased in hda-1 mutants. (A,B) Representative lin-17 smFISH images of wild-type and tissue-specific hda-1 mutant animals [hda-1(-)] following the L2 symmetric (A) and L2 asymmetric (B) divisions. Blue arrowheads show examples of cells with visually higher lin-17 mRNA spots. Seam cell nuclei are labelled using SCM::GFP. Scale bars: 5 μm. (C,D) Quantification of lin-17 mRNA spots in V1-V4 lineages in wild-type (blue) and hda-1 mutant (red) animals following the L2 symmetric (C) and L2 asymmetric divisions (D) (in C, 30≤n≤32; in D, 17≤n≤21 cells per condition). Wild-type and hda-1(-) values were compared within each cell type and lineage using a Mann–Whitney test or Welch's t-test following normality testing. (E) Normalised anterior-posterior (A-P) asymmetry of lin-17 mRNA following the L2 asymmetric division in wild-type (blue) and hda-1 mutant (red) animals. A-P asymmetry was calculated as in Fig. 4 (150≤n≤162 cell pairs per condition). (F) Normalised V1-V4 lineage differences in lin-17 mRNA following the L2 asymmetric division in wild type (blue) and hda-1 mutants (red). V1-V4 differences were calculated as in Fig. 4 (60≤n≤72 cell pairs per condition). Normalised A-P asymmetry and V1-V4 difference scores were compared between wild type and hda-1 mutants using a Mann–Whitney test. (G) Quantification of lin-17 smFISH spot density in the lateral epidermis across the anterior-posterior axis for wild type and for hda-1(-) animals at late L1 stage; n=34 animals for wild type and n=18 animals for hda-1(-). Lines and ribbons show negative binomial fits ±95% confidence intervals. The genotype×AP-position interaction was significant by likelihood-ratio test (P<0.0001), consistent with a reduced lin-17 A-P gradient in hda-1(-) animals. Comparisons between wild type and hda-1 mutant represent Wilcoxon rank-sum tests with Holm correction. Data in C-G show the mean±95% confidence intervals. ****P<0.0001, ***P<0.001, **P<0.01 and *P<0.05. See also Fig. S4.

In contrast, other Wnt signalling pathway components explored showed weaker or no significant changes in transcript abundance in the hda-1 mutant background. mom-5 expression was elevated at the late L1 stage in hda-1 mutants but showed no significant changes following the L2 divisions (Fig. S5). dsh-2 displayed modest decrease in expression in hda-1 mutants following the L2 symmetric and asymmetric divisions (Fig. S6), while apr-1 and pry-1 expression levels showed both a modest decrease following the L2 symmetric division that was not consistent across lineages (Figs S7 and S8). Together, these findings indicate that HDA-1 influences expression of multiple Wnt pathway genes, with cam-1 and lin-17 showing the most robust and consistent changes in hda-1 mutants.

Overexpression of lin-17 and cam-1 in seam cells phenocopies the hda-1 mutant polarity defects

The upregulation of lin-17 and cam-1 expression in hda-1 mutants suggested that increased Wnt receptor expression may underlie the polarity defects observed in these animals. To test this hypothesis, we decided to overexpress cam-1 and lin-17 under the arf-5i element, which is contained within the SCM enhancer and is sufficient to drive uniform expression in seam cells (Ashley et al., 2021; Katsanos et al., 2021). Overexpression of lin-17 alone led to a significant increase in seam cell number at late L4, while overexpression of cam-1 alone did not significantly alter the average seam cell number but increased phenotypic variability (Fig. 6B). Co-overexpression of both lin-17 and cam-1 resulted in a pronounced increase in seam cell number, with animals simultaneously displaying gaps in the seam cell line (Fig. 6A,B). Phenotypic analysis across developmental stages revealed that, similar to hda-1 mutants, changes in seam cell number predominantly occurred at the L4 stage (Fig. 6B).

Fig. 6.

Overexpression of cam-1 and lin-17 is sufficient to cause polarity defects. (A) Representative images of control and animals overexpressing cam-1 and lin-17 under the arf-5i element (arf-5i::cam-1; arf-5i::lin-17). Blue bracket indicates a gap; white arrowheads indicate extra seam cells. (B) Seam cell counts in wild-type (control) and animals overexpressing cam-1 (orange), lin-17 (green) and both cam-1 and lin-17 (red) under the arf-5i element at the late L3 and late L4 stages. Data are mean±95% confidence intervals. *P<0.05, **P<0.01 and ****P<0.0001 with a Mann–Whitney test; 30≤n≤65 animals per condition. (C) Representative images of eff-1 smFISH following the L2 asymmetric division in wild-type and in control animals overexpressing cam-1, lin-17 and both cam-1 and lin-17. Orange brackets indicate divisions with reversed expression of eff-1. (D) Heatmap showing frequency of cells expressing eff-1 following the L2 asymmetric division in the conditions in C (14≤n≤37 cells per condition). Fisher's exact tests on pooled V1-V4 lineage data showed a significant increase in eff-1 polarity reversals in paa/pap pairs upon cam-1 overexpression, lin-17 overexpression and combined cam-1; lin-17 overexpression compared with control. In A and C, seam cell nuclei are labelled using SCM::GFP. Scale bars: 50 μm in A; 5 μm in C.

To assess whether overexpression of these Wnt receptors recapitulates the polarity defects observed in hda-1 mutants, we examined eff-1 expression following the L2 asymmetric division. Overexpression of either lin-17 or cam-1 individually resulted in polarity reversals closely resembling those of hda-1 mutants, with eff-1 expression detected in posterior daughters with no expression in anterior daughters (Fig. 6C). Such reversals were rarely observed in control animals (0-3% frequency). Notably, the V2 lineage, which was most prominently affected in hda-1 mutants, showed the highest frequency of polarity reversals upon Wnt receptor overexpression, with 21% of animals displaying anterior V2 reversals for cam-1 overexpression and 31% for lin-17 overexpression (Fig. 6D). Co-overexpression of both receptors further increased the frequency of V2 reversals to 63% and additionally induced polarity defects in other seam cell lineages, although V2 remained the most frequently affected (Fig. 6C,D). Taken together, these results show that increasing the expression of Wnt receptors in seam cells is sufficient to phenocopy the polarity defects observed in hda-1 mutants.

NuRD and SIN3 disruption does not fully phenocopy hda-1 polarity defects

HDA-1 is known to function as part of the NuRD and SIN3 chromatin-remodelling complexes, among others (Ahringer, 2000). To investigate whether these complexes mediate the HDA-1 role in polarity regulation, we first examined seam cell phenotypes in mutants of key complex components. Loss of the NuRD components dcp-66 and egl-27 resulted in an increase in the average seam cell number (Fig. 7A). By contrast, mutants of the NuRD component lin-53 and the SIN3 component sin-3 did not show significant differences compared to wild type (Fig. 7A). To determine a role in polarity, we assessed eff-1 expression following the L2 asymmetric division in these mutants. egl-27, dcp-66 and sin-3 mutants did not display polarity reversals (Fig. 7B,C), similar to hda-1 mutants (Fig. 3B). This suggests that, while these chromatin-remodelling complexes play a role in seam cell patterning, they may not mediate the polarity defects caused by loss of HDA-1.

Fig. 7.

NuRD and SIN3 disruption does not phenocopy the polarity defects of hda-1 mutants. (A) Seam cell counts in wild-type, and in lin-53(n833), egl-27(ok1670), dcp-66(gk370) and sin-3(syb2172) mutants at the late L4 stage. Data are mean±95% confidence intervals. ****P<0.0001 with a Mann–Whitney test; 34≤n≤76 animals per condition. (B) Representative images of eff-1 smFISH following the L2 asymmetric division in wild type and in egl-27(ok1670), dcp-66(gk370) and sin-3(syb2172) mutants. Seam cell nuclei are labelled using SCM::GFP. Scale bars: 5 μm. (C) Heatmap showing frequency of cells expressing eff-1 following the L2 asymmetric division in the conditions in B (9≤n≤28 cells per condition). No significant increase in eff-1 polarity reversals was observed in any of the mutants using a Fisher's exact test.

DISCUSSION

Histone deacetylases are well-established regulators of cell fate specification (Brunmeir et al., 2009; Dovey et al., 2010a; Jaju Bhattad et al., 2020; Lee et al., 2004). Previous work on hda-1 in C. elegans has shown roles in endoderm specification, development of the nervous system, sex-specific gene expression and vulval patterning (Calvo et al., 2001; Dufourcq et al., 2002; Grote and Conradt, 2006; Matus et al., 2015; Ranawade et al., 2013; Zinovyeva et al., 2006). We propose here that HDA-1 influences asymmetric cell division polarity in the C. elegans epidermal seam cells, at least in part, by repressing the expression of the Wnt receptors lin-17 and cam-1. In seam cells, Wnt receptors have been shown to be redundantly required for establishment of polarity, as simultaneous disruption of cam-1, lin-17 and mom-5 results in significant loss of polarity (Yamamoto et al., 2011). More recently, CAM-1 and LIN-17 have been implicated in polarity orientation during seam cell divisions, although how exactly they contribute to this process remains mechanistically unclear (Sawa et al., 2026). Our study suggests a link between chromatin-mediated transcriptional repression and Wnt-regulated polarity orientation in the context of asymmetric cell division.

Our findings demonstrate that Wnt receptors display graded expression patterns along the anterior-posterior axis. While cam-1 exhibits an anterior-to-posterior gradient, lin-17 shows an opposite posterior-to-anterior gradient, whereas mom-5 appears to be expressed more in the mid-body region. Gradients of Wnt receptors have been reported in other systems too, e.g. across the embryonic telencephalon in mice or in the developing wing epithelium in flies, and they are thought to influence graded competence of cells to Wnt ligands or shape the range of Wnt ligand action by limiting diffusion or protecting Wnt ligands from degradation (Cardigan et al., 1998; Chaudhary et al., 2019; Kim et al., 2001). It remains currently unknown how the cam-1 and lin-17 expression domains are established. HDA-1 shows a relatively uniform expression across seam cell lineages, so if HDA-1 contributes to shaping these gradients, it likely does so through interactions with transcription factors that provide positional information along the anterior-posterior axis. Hox genes, which specify positional identity along the body axis and regulate cell fate in a spatially restricted manner, are strong candidates for this role (Mallo and Alonso, 2013). In mammalian cells, Hox proteins have been shown to recruit HDACs through interactions with PBC co-factors to repress transcription (Li et al., 2026; Saleh et al., 2000). In C. elegans, several Hox genes show distinct expression domains along the anterior-posterior axis modulating seam cell development (Chisholm and Hsiao, 2012). Notably, Hox genes lin-39, mab-5 and ceh-13 are among the putative HDA-1 targets identified in our TaDa analysis. Therefore, HDA-1 may also influence Wnt receptor gradients indirectly through regulation of Hox gene expression. Interestingly, Wnt signalling itself can regulate Wnt receptor gradient establishment through feedback, as shown, for example, in the Drosophila wing where Wnt signalling represses frizzled 2 (Dfz2) expression (Cardigan et al., 1998; Sato et al., 1999). This is consistent with the expression patterns observed in daughter cells following L2 division, where cam-1 was found to be enriched in anterior daughters and lin-17 in posterior daughters. These patterns may reflect negative (cam-1) and positive (lin-17) feedback via Wnt signalling pathway activation (Ferrando-Marco et al., 2026).

In hda-1 mutants, both cam-1 and lin-17 receptors are upregulated, leading to reduced differences between V1-V4 lineages, and the lin-17 gradient being more broadly disrupted across the lateral epidermis. While these changes correlate with defects in polarity orientation, it remains unclear whether it is the disruption of the gradient or simply increased receptor abundance in seam cells that is linked to the polarity defect. Recently, overexpression of lin-17 in a cam-1;mom-5 mutant background was shown to result in random polarity orientation in V1-V4 seam cell lineages (Sawa et al., 2026), which was taken to suggest that LIN-17 does not itself encode polarity cues. In contrast, our observation that uniform lin-17 overexpression in an otherwise wild-type background leads to polarity reversals suggests that polarity defects can arise from changes in receptor abundance alone. It has been proposed that Wnt ligands can exert both instructive and permissive effects on polarity orientation, and that the permissive function of Wnt signalling requires additional unidentified polarity cues (Sawa et al., 2026; Yamamoto et al., 2011). Opposing anterior-posterior gradients of cam-1 and lin-17 could provide such intrinsic positional cues that allow seam cells to orient polarity in response to Wnt signals, with the relative levels of different receptors determining how individual cells interpret this signal (Fig. 8A). Differential receptor expression between anterior and posterior daughters following division may further refine polarity orientation at the level of individual lineages (Fig. 8A). Interestingly, both levels of spatial organisation are disrupted in hda-1 mutants: differences in lin-17/cam-1 receptor expression along the V1-V4 axis are reduced, and asymmetry between daughter cells is decreased in the case of cam-1 (Fig. 8B). Wnt receptors are thought to localise to the posterior cortex of the dividing seam cell, leading to asymmetric segregation of downstream regulators such that anterior daughters repress Wnt target genes while posterior daughters activate them (Goldstein et al., 2006; Mizumoto and Sawa, 2007b). An increase in receptor abundance could interfere with this process by destabilising receptor localisation and thereby impairing polarity orientation. However, we note that absolute receptor levels alone cannot fully account for the polarity defects observed in hda-1 mutants. If polarity reversals were solely caused by receptor levels exceeding a threshold that disrupts localisation, reversals would be expected to occur most frequently in lineages with the most extreme changes from baseline expression (V1 and V4). However, we found that polarity reversals occur preferentially in V2 across both hda-1 mutants and receptor overexpression conditions, suggesting that additional factors are likely to influence seam cell sensitivity to polarity defects.

Fig. 8.

Model for HDA-1-mediated regulation of Wnt receptor expression and polarity orientation in seam cells. (A) In wild-type animals, seam cell polarity is consistently oriented along the anterior-posterior axis. The Wnt receptors cam-1 (green) and lin-17 (blue) display complementary, graded expression across V1-V4 seam cell lineages, with cam-1 enriched anteriorly and lin-17 enriched posteriorly. In addition, within individual lineages, cam-1 and lin-17 show differential expression between anterior and posterior daughter cells following division. (B) In hda-1 mutants, asymmetric divisions show polarity reversals and expression of both cam-1 (green) and lin-17 (blue) is increased. Differences in receptor expression along the anterior-posterior axis and asymmetry between anterior and posterior daughter cells are reduced. Disruption of graded and asymmetric receptor expression may impair the interpretation of permissive Wnt signals, potentially contributing to defects in polarity orientation and polarity reversals.

It is notable that hda-1 mutants not only displayed polarity reversals at the L2 stage, but also exhibited both extra seam cells and loss of seam cells manifested at late L4. Wnt receptor overexpression was found to be sufficient to phenocopy both early reversals and late-stage symmetrisations, raising the possibility that polarity defects can manifest differently depending on the developmental stage. Symmetrisations of asymmetric divisions are often associated with impaired POP-1-mediated repression (Bekas and Philips, 2022 preprint; Ferrando-Marco and Barkoulas, 2025). In the embryonic MS blastomere, HDA-1 functions together with UNC-37/Groucho as a co-repressor of the TCF factor POP-1 (Calvo et al., 2001). Given that UNC-37 similarly acts as a POP-1 co-repressor in seam cells (Bekas and Philips, 2022 preprint), HDA-1/UNC-37 may also contribute to the repression of Wnt target genes in anterior differentiating daughters through regulation of POP-1. HDACs lack DNA-binding domains and require interactions with DNA-binding proteins or multiprotein complexes to be recruited to specific genomic locations. In C. elegans, HDA-1 typically functions as part of several distinct chromatin-remodelling complexes, including NuRD, SIN3, CoREST and the related MEP-1 containing MEC complex (Ahringer, 2000; Bender et al., 2007; Chen and Han, 2001; Choy et al., 2007; Passannante et al., 2010; Robert et al., 2023). While mutants of NuRD and SIN3 components show changes in terminal seam cell number, they do not display the same L2 polarity reversals. This suggests that, while HDA-1 may cooperate with POP-1 and function as part of the NuRD or SIN3 complex in certain aspects of seam cell gene regulation, its role in influencing polarity orientation likely depends on additional interacting partners that remain to be identified.

MATERIALS AND METHODS

C. elegans maintenance and genetics

All C. elegans strains in this study were maintained at 20°C on Nematode Growth Medium (NGM) plates seeded with E. coli OP50 (Brenner, 1974). A list of strains used in this study is provided in Table S3.

Molecular cloning and transient transgenesis

To overexpress cam-1 using the last intron of arf-5 (arf-5i), the cam-1^a fragment was amplified from N2 cDNA using the oligos cam-1CDS(pes-10) Fw and cam-1CDS(unc-54UTR) Rv. The amplicon was inserted by Gibson Assembly into pIR5 to create pMFM32. pMFM32 was injected into JR667 at 50 ng/μl with 5 ng/μl myo-2::dsRed as co-injection marker and 55 ng/μl BJ36 as carrier DNA. pMFM32 and pDK5 were injected into JR667 at 40 ng/μl each, with 5 ng/μl myo-2::dsRed as co-injection marker and 25 ng/μl BJ36 as carrier DNA.

The TaDa constructs for hda-1 were built using pDK7. NLS-GFP was amplified from pDK54 with NLS:GFPu-DamF and NLS:GFPu-DamR, and hda-1 was amplified from the fosmid WRM0627aC04 with (mCherry)_hda-1_Fw and (Dam)_hda-1_R. pDK7 was digested with XmaJI and NLS-GFP and hda-1 fragments were inserted upstream and in-frame with Dam to produce pMFM4 and pMFM6. The promoter srf-3i1::Δpes-10 was introduced using the Multisite Gateway Technology (Invitrogen). srf-3i1::Δpes-10 fragment was amplified from pDK82 using srf-3i_attB4 and delta-pes-10_attB1r_R. A donor vector was produced via a BP reaction, and was inserted in LR reactions with pMFM4 and pMFM6 to produce two different vectors: pMFM8 and pMFM10. A complete list of oligos and vectors used can be found in Table S3.

Stable transgenesis

Single-copy transgenic lines were generated by the MosSCI method using animals of the EG6699 strain with a Mos1 transposon insertion on chromosome II (ttTi5605 locus) showing the uncoordinated (Unc) phenotype (Frøkjær-Jensen et al., 2014). pMFM8 and pMFM10 vectors were injected at 50 ng/μl together with 50 ng/μl of pCFJ601 (Mos1 transposase), 10 ng/μl of pMA122 (heat-shock-inducible peel-1 toxin), and 10 ng/μl, 2.5 ng/μl and 5 ng/μl of the co-injection markers (pGH8, myo-2::dsRed and myo-3::mCherry). After injection, animals were kept at 25°C until they were starved. Heat shock was performed at 33°C for 3.5 h and the plates were allowed to recover for 3 h at room temperature before ‘reverse chunking’. The next day the top of the laws on each chunk was screened for normal roaming animals [unc-119(-) rescued animals] with the absence of co-injection markers. Homozygous lines were confirmed molecularly for the presence of single-copy transgenes using oligos NM3880 and NM3884.

CRISPR genome editing

LoxP sites flanking the hda-1 endogenous locus (394 bp upstream of the transcription start site and 435 bp downstream of the transcription termination site) were introduced by injecting JR667 animals with 0.25 ng/μl protein Cas-9, 2.25 μM tracrRNA, 2.38 μM of 5′_hda-1_crRNA and 3′_hda-1_crRNAs, 5.5 μM of 5′ and 3′ repair templates, and 5 ng/μl myo-2::dsRed. crRNAs targeted 414 bp upstream the TSS and 402 bp downstream the TES. 5′ and 3′ hda-1 repair templates presented 50 bp homology to either side of the Cas-9 cut site and included a loxP sequence, a ClaI cut site, and either an M13uni_43 (5′ template) or M13rev_49 (3′ template) sequence. Transgenic animals with red pharynx were genotyped using the oligos hda-1_5p_F_2, R and hda-1_3p_F, R for 5′ and 3′ modification, respectively. The PCR product was digested with ClaI restriction enzyme to confirm successful homologous repair.

RNAi feeding

RNAi by feeding was performed using E. coli HT115 expressing double-stranded RNA corresponding to the target gene or the control dsRNA plasmid (empty vector) as food source (Kamath et al., 2003). Bacteria cultures were grown overnight and 300 μl were seeded onto NGM plates containing 25 μg/ml ampicillin, 12.5 μg/ml tetracyclin and 1 mM IPTG. All RNAi treatments were performed during the postembryonic development by seeding eggs from the corresponding strain onto the RNAi plates. RNAi bacteria clones used in this study are from the commercially available Ahringer RNAi Library (Kamath and Ahringer, 2003) (Source Bioscience).

Analysis of brood size

To measure brood size, animals were synchronised by spot bleaching gravid adults. 48 h later (day 1), individual L4 stage hermaphrodites were transferred onto 35 mm Petri dishes seeded with 100 μl bacterial lawn. 10 individual hermaphrodites were transferred per condition. Animals were allowed to develop into adults and lay self-progeny for 24 h (I plates). On day 2, animals were transferred onto new 35 mm plates and allowed to lay eggs for another 24 h (II plates). On day 3, animals were transferred once more to III plates. The number of progeny was scored on subsequent days: plates I at day 3, plates II at day 4 and plates III at day 5. To facilitate scoring, a grid pattern was drawn one the plates using a fine marker (Kwah and Jaramillo-Lambert, 2023). Brood size was calculated by summing the number of progeny produced by each parental. Brood size bar plots represent the mean number of progeny per condition and error bars represent the standard deviation.

Analysis of developmental time

To measure the developmental time, animals were synchronised by spot bleaching gravid adults. Then, synchronised day 2 adults were let to lay progeny into new NGM plates for 3 h and then they were removed from the plates. 15 adults were used per plate and three plates representing three biological replicates per condition were set up in parallel. Plates were scored at 45 and 50 h after removal of the adult animals to find animals that had reached late L4 stage. To identify late L4 animals, the shape of the vulva was used as reference. Bar plots showing developmental time represent average percentage of animals that have reached L4 stage across the three replicates and error bars represent standard error of proportion. To calculate the standard error of proportion, average proportion across replicates (p) and total number of animals scored across replicates (n) was used as follows:

Phenotypic analysis and microscopy

For seam cell scoring, animals were anesthetised on 3% agarose pads containing 100 μM sodium azide (NaN₃), secured with coverslip and visualised under an AxioScope A1 (Zeiss) upright epifluorescence microscope with a LED light source with a RETIGA R6 camera controlled by the Ocular software. The seam cell number was scored on the lateral side closest to the objective.

Protein reporter and membrane marker images were acquired using a Leica SP8 inverted confocal microscope with a 63×/1.4 oil DIC objective controlled by LASX software. Animals were mounted in 5% agarose pads containing 100 μM sodium azide (NaN₃) and secured with a coverslip. To visualise L2 symmetric and asymmetric divisions, animals were observed at 26 h post-bleaching. HDA-1::GFP images were acquired in a single plane that provided the best focus for the seam cell nuclei. Z-stack slices of 0.4 μm were acquired for animals containing the AJM-1p::AJM-1::GFP at late L4 stage, 48 h post-bleaching.

Fluorescence intensity was quantified using FIJI macros developed by the Imperial College FILM facility. The data was normalised (rescaled) to [0,1] using the formula:

where X is the data, x is a datapoint and z is the corresponding rescaled datapoint. Picture editing was performed using straightening tool and stitching macro from FIJI (Preibisch et al., 2009).

Single molecule fluorescent in situ hybridisation

Animals were synchronised by bleaching and were fixed at the appropriate stage using 4% formaldehyde for 45 min and permeabilised with 70% ethanol. Samples were then incubated with a pool of 24-48 oligos labelled with Quasar 670 (Biosearch Technologies) for 16 h. Imaging was performed using the 100× objective of an epifluorescence Ti-eclipse microscope (Nikon) and DU-934 CCD-17291 camera (Andor Technology) operated via the Nikon NIS Elements Software. 23 z-stack slices of 0.8 μm were acquired for each animal. Spot quantification was performed using MATLAB (MathWorks), as previously described (Raj et al., 2008). For smFISH quantification in seam cells, regions of interest matching their eye-shaped morphology were manually drawn around each seam cell nucleus guided by the scm::GFP fluorescence and blinded to the smFISH signal. Only puncta within the ROI were included in the quantification.

For the epidermal quantification of cam-1 and lin-17, smFISH signal was analysed along the anterior-posterior axis using the epidermal seam cells as anatomical landmarks. Regions of interest (ROIs) were defined between adjacent seam cells (H0-H1, H1-H2, H2-V1, V1-V2, V2-V3, V3-V4, V4-V5 and V5-T). To minimize signal contribution from non-epidermal tissues in the head, mRNA quantification in the two most anterior ROIs (H0-H1 and H1-H2) was restricted to three z-planes corresponding to the epidermal layer, whereas five z-planes were analysed for all more posterior ROIs. ROIs were drawn manually in Fiji, and the area of each ROI was measured in pixel units and converted to physical units using the image calibration (7.6923 pixels=1 μm). ROI volumes were calculated as the product of ROI area (μm²), z-step size (0.8 μm) and the number of z-planes included for each ROI. For each region, mRNA density was calculated by dividing the number of detected mRNA molecules by the corresponding ROI volume and is reported as mRNAs per 1000 μm³. This volumetric normalisation accounts for differences in ROI size and z-depth across anterior and posterior regions. For nhr-73 and elt-1, reversal-like events (Fig. S1F) were defined as individual sister-cell pairs in which posterior expression was lower than anterior expression, scored separately for each V1-V4 lineage.

For presentation, z-slices containing spots where max-projected to give a single image, inverted and merged to the GFP channel using ImageJ (NIH). For the eff-1 smFISH experiment in dcp-66(-), the seam cell marker signal was adjusted locally in faint nuclei to facilitate lineage identification. A complete list of the smFISH probes used in this study is provided in Table S3.

DNA extraction and cDNA preparation

Animals from the N2 strain were collected at L2 stage and total RNA was extracted using the TRIzol reagent (Invitrogen) and isopropanol/ethanol precipitation. NanoDrop quantification and gel electrophoresis was used to assess the quantity and quality of RNA. 2 mg of RNA from all samples was subjected to DNase (Promega) treatment and cDNA was synthesized using Superscript IV (Invitrogen) with Oligo(dT) primers as per the manufacturer's instructions.

Extraction and amplification of methylated DNA

For TaDa experiments, animals were synchronised by bleaching and grown for three generations in E. coli dam−/− dcm−/− bacteria (New England Biolabs, C2925). Two biological replicates were grown in parallel for strains carrying the TaDa transgenes NLS-GFP::Dam and HDA-1::Dam. Animals were collected after 35 h post-bleaching, extensively washed five times in M9 buffer and stored at −20°C. Purification and amplification of methylated DNA was performed as previously described (Katsanos et al., 2021). Extracted gDNA was digested with DpnI and double stranded adaptors were ligated using T4 DNA ligase. Then, DNA fragments were digested with DpnII and amplified using MyTaq polymerase. Amplicons were purified using QIAquick PCR purification kit and adaptors were removed by AlwI digestion followed by another purification. Library preparation and Next Generation Sequencing on an Illumina HiSeq 4000 platform was performed by GENEWIZ.

NanoDam signal profiles, peak calling and gene assignment

FASTQ files representing paired-end reads for each sample and replicate were processed using the damidseq_pipeline (v.1.5.3) (Marshall and Brand, 2015). Reads were mapped to the C. elegans WBcel235 genome assembly using bowtie2 (v.2.4.5) and scores were calculated per GATC fragment of the genome. Pairwise genome-wide calculations of log2(Dam-fusion/Dam-control) ratio scores per GATC fragment of the genome were performed from bam files to produce bedGraph signal files that were arithmetically averaged into a single signal profile. Peak calling was performed using the perl script find_peaks (available at https://github.com/owenjm/find_peaks) with FDR<0.05. Statistically significant peaks were assigned to genes using UROPA (Kondili et al., 2017) with Caenorhabditis_elegans.WBcel235.106.gtf as the genome annotation file. For this analysis, the transcription start site (TSS) was defined as the annotated 5′ start coordinate of each transcript according to strand orientation. Peaks were assigned to genes when their centre coordinate was positioned up to 6 kb upstream of a gene start site or 2 kb downstream of the end site. Visualisation of signal tracks and other genomic features was performed using Integrative Genomics Viewer (IGV).

Aggregation plots and heatmaps of signal localization

Signal aggregation plots and peak localisation heatmaps were generated using the SeqPlots GUI application (Stempor and Ahringer, 2016). Signal and peaks were visualised across genes containing peaks up to 6 kb upstream of a gene start site or 2 kb downstream of the end site. Genes were plotted as ‘anchored’ features, where the TSS and the TES are fixed in the x-axis and the genic sequence is adjusted to a pseudo-length of 2 kb. In aggregation plots, signal is averaged in 10 bp bins and the y-axis represents deviations from the average signal across the plotted region (z-scores). In heatmaps, each coloured line represents the position of a statistically significant peak.

Gene set enrichment analysis

Gene sets identified in this study were assessed for enriched gene ontology (GO) terms using Gene Set Enrichment analysis (WormBase). P<0.005 was used for significance threshold and plots show −log₁₀(pvalue).

Statistical analysis

Graphic representation and statistical analysis were performed using R. Error bars used in all graphs represent 95% confidence intervals unless otherwise indicated. Statistical analyses were selected according to the experimental design and comparison structure of each dataset. Normality was assessed using Shapiro-Wilk tests. For independent-group comparisons, homogeneity of variance was assessed using Brown-Forsythe/Levene tests. For datasets with multiple experimental conditions compared with a shared control (Fig. 1B), data were analysed using one-way ANOVA followed by Dunnett's multiple-comparison test when assumptions of normality and equal variance were met. Otherwise, Kruskal–Wallis tests followed by Dunn's post-hoc comparisons were used. For pre-specified independent pairwise comparisons, including wild-type versus hda-1(-) comparisons within defined cell types, developmental stages or lineage groups, Welch's t-tests were used when both groups passed normality, whereas Mann–Whitney/Wilcoxon rank-sum tests were used when either group failed normality (Figs 2B, 4C-F, 5C-F, 6B, 7A; Figs S5D, S6D, S7D, S8D). For paired anterior/posterior comparisons, paired differences were calculated for each matched pair and tested for normality; paired two-tailed t-tests were used when paired differences were normally distributed, and Wilcoxon signed-rank tests were used otherwise (Fig. 1G,K). For datasets involving more than two matched or repeated measurements from the same animals, comparisons were analysed using Friedman tests followed by post-hoc Wilcoxon signed-rank tests with Holm correction (Figs S3D and S4D). For datasets testing whether anterior/posterior asymmetry depended on genotype, two-way mixed ANOVA was used with genotype as a between-pair factor and anterior/posterior position as a within-pair factor; the genotype×position interaction was used as the main test of altered asymmetry (Fig. 3D,F,H). For the lateral-epidermis gradient analyses in Figs 4G and 5G, smFISH spot counts were modelled as a function of genotype and AP position (centred prior to fitting) using negative binomial regression. To test whether the spatial gradient slope is dependent on genotype, for each gene, a model including a genotype×AP-position interaction was compared against the simpler null model without interaction using a likelihood-ratio test. Categorical heatmap data were analysed using Fisher's exact tests. Anterior-posterior asymmetry and V1-V4 differences in Fig. 4E and 4F and Fig. 5E and 5F were quantified using a normalised difference, (A−P)/(A+P+1), which is robust to zero counts and downweighs asymmetry estimates at very low expression levels. Asterisks in the figures indicate corresponding statistical significance as it follows: *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001.

Peer review history

The peer review history is available online at https://journals.biologists.com/dev/lookup/doi/10.1242/dev.205658.reviewer-comments.pdf