Asymmetric segregation of NuRD during ACDs of C. elegans Q neuroblast.

(A) Schematic of the Q neuroblast lineages. QL or QR neuroblast each generates three neurons and two apoptotic cells (Q.aa/Q.pp, X). QL produces PQR, PVM, and SDQL. QR produces AQR, AVM, and SDQR.

(B) A model of the NuRD complex composition (Bracken et al., 2019; Lai and Wade, 2011).

(C) Protein domain structure of the GFP-tagged HDAC1/2 (HDA-1) or mNeonGreen-tagged RBBP4/7 (LIN-53). CAF1C: histone-binding protein RBBP4 or subunit C of CAF1 complex; WD: WD40 repeat tryptophan-aspartate domain. Scale bar: 100 amino acids.

(D) Representative images of endogenous HDA-1::GFP, LIN-53::mNeonGreen and overexpressed MYS-1::GFP during ACDs of QR.a. In each panel, the top row shows merged images, the middle row shows mCherry-tagged plasma membrane and histone, and the bottom row shows heatmaps of GFP/mNeonGreen fluorescence. The anterior of the cell is on the left. The GFP/mNeonGreen fluorescence intensity ratio between posterior and anterior chromatids or between QR.ap and QR.aa nuclei are shown in yellow at the lower right corner of heatmaps. Time is in minutes. Other frames are in Figure S3A, and the full movies are in Supplementary Movie S3-S5. Scale bar: 2 µm.

(E) Quantification of HDA-1::GFP, LIN-53::mNeonGreen and MYS-1::GFP fluorescence intensity ratio between QR.ap and QR.aa nuclei. Data are presented as mean ± SEM. N = 10 – 12.

(F) Quantification of HDA-1::GFP (magenta), LIN-53::mNeonGreen (orange) and MYS-1::GFP (gray) fluorescence intensity ratios between the posterior and anterior half of QR.a or between QR.ap and QR.aa. The anaphase onset is defined as the last frame without chromatids segregation. Data are mean ± SEM. N = 10 – 12.

(G) Quantification of HDA-1::GFP (left) and LIN-53::mNeonGreen (right) fluorescence intensity ratios between the large and small daughters of cells shown on the X-axis. Data are mean ± SEM. N = 7 – 14.

NuRD Asymmetry in C. elegans embryonic cell lineages.

(A) An x and y plane of 3D time-lapse imaging data at the ∼200 min after the first embryonic division (left) in the HDA-1::GFP knock-in animal. H2B::mCherry was expressed in the embryo ubiquitously. The bright green circle indicates the ABalaappa cell, whose division generates an apoptotic daughter cell. The cyan circle indicates the ABalaappp cell, whose division generates two surviving daughters. The orientation of the embryo is shown in the middle. The cell lineage of ABalaapp is shown on the right. Each vertical line represents a cell and each branch represents a cell division. X indicates an apoptotic cell. Scale bar: 5 μm.

(B) Representative images of dividing ABalaappp cell (top panel) and ABalaappa cell (bottom panel) from the lineage tracing data. GFP fluorescence are shown as heatmaps below merged images. Cyan or bright green lines mark the chromatids or nuclei of daughter cells according to the H2B::mCherry signal. The black and white arrowheads point to the anterior and posterior daughter cells, respectively. Scale bar: 5 µm.

(C) Tree visualization of HDA-1::GFP and LIN-53::mNeonGreen segregation between sister cells during embryonic development. Cells and their lineage relations are visualized as a tree structure, with vertical lines indicating cells and horizontal lines indicating cell divisions. Green vertical lines show cells with a significantly higher (>1.5-fold) nuclear HDA-1::GFP and LIN-53::mNeonGreen intensity than their sister cells. Red arrowheads point to apoptotic cells. Placement on the tree is according to the Sulston nomenclature. See also Supplementary Table 2.

(D) Quantifications of HDA-1::GFP and LIN-53::mNeonGreen fluorescence intensity ratios between nuclei of the live daughter cell and its apoptotic sister cell automatically identified and tracked in 3D time-lapse imaging of embryonic development (Supplementary Table 2). The lineage names of non-apoptotic sister cells of 17 cell pairs with apoptotic cells are shown on the X-axis. The apoptotic sister cell of MSpppaap (magenta circle) completes apoptosis approximately 400 minutes after birth. N = 3 - 4 in each measurement. Data are mean ± SEM. One-sample t-test compared to 1, *p< 0.05, **p< 0.01, ***p< 0.001, ns: not significant.

RNAi of hda-1 induces ectopic apoptosis and increases H3K27 acetylation of the egl-1 gene.

(A) Fluorescent images show Phsp::sAnxV::GFP from ced-1(e1735) and ced-1(e1735); ced-3(n2433) embryos between late gastrulation and bean stage, treated with control RNAi or hda-1 RNAi. Scale bars, 5 μm.

(B) Quantifications of cell corpse number in the embryos (N = 39 - 60) of indicated genotypes. Data are presented as mean ± SEM. Unpaired two-tailed t-test, compared with the control RNAi, ***p < 0.001.

(C) Scatter plot of the increased gene expression in both hda-1 RNAi and lin-53 RNAi animals. The egl-1 gene is marked in red. See also Supplementary Table 3.

(D) Quantitative real-time PCR (RT-PCR) measurement of egl-1 mRNA levels in the control, hda-1 or lin-53 RNAi embryos. Fold induction was calculated relative to levels in control RNAi embryos. Data represent the mean ± SEM of six biological replicates. Unpaired t-test, ***p < 0.001.

(E) Quantification of the number of cells expressing the Pegl-1::NLS::GFP reporter in embryos treated with control, hda-1 or lin-53 RNAi. Data are mean ± SEM and are compared by unpaired t-test, *p < 0.05, **p < 0.01, ***p < 0.001. N = 8 – 9.

(F) Normalized ChIP-Seq signal profiles of the H3K27 acetylation level in the control, hda-1 or lin-53 RNAi embryos at the egl-1 locus. The Y-axis shows the average sequencing coverage of bins per million reads of three biological replicates normalized to the input control sample. Ectopic H3K27ac enrichments at the egl-1 locus of hda-1 RNAi and lin-53 RNAi embryos are highlighted in magenta and cyan.

(G) ChIP-qPCR analyses using H3K27ac antibody or IgG (the negative control) at selected elements of egl-1 indicated in (F) by red lines. Results are shown as the percentage of input DNA. Data are mean ± SEM from three biological replicates. Unpaired t-test, *p < 0.05.

HAD-1 interacts with subunits of V-ATPase.

(A) The plot compares counts of proteins co-precipitated with HDA-1::GFP with those with the control ACT-4 (actin)::GFP. The PSM (Peptide-Spectrum Match) is the number of identified peptide spectra matched for the protein. Blue dots represent the subunits of V-ATPase. See also Supplementary Table 4.

(B) Schematic model of the V-ATPase complex. The known worm subunits are indicated. The mean PSM of the encoded protein from two co-IP and MS repeats is shown in blue after the gene name.

(C) Western blot (WB) showing co-immunoprecipitation (co-IP) of V-ATPase V1 domain A subunit (V1A) with the HDA-1 from worm lysates.

V-ATPase regulates NuRD asymmetric segregation and cell fates.

(A) Dynamics of the cytosolic pH indicated by super-ecliptic pHluorin during QR.a division in DMSO control or Baf A1-treated animals. In each panel, the top row shows mCherry-tagged plasma membrane and histone, and the bottom row shows heatmaps of super-ecliptic pHluorin signal. The anterior of the cell is on the left. Time 0 min is the onset of anaphase: scale bar, 2 µm. See also Supplementary Movie S8.

(B) The super-ecliptic pHluorin fluorescence intensity ratio between QR.ap and QR.aa. Data are mean ± SEM. N =11-15. Unpaired two-tailed t-test, **p < 0.01, ns: not significant.

(C) Images of HDA-1::GFP distribution during QR.a cell division in DMSO control or Baf A1-treated animals. In each panel, the top row shows mCherry-tagged plasma membrane and histone, and the bottom row shows heatmaps of HDA-1::GFP fluorescence. The GFP fluorescence intensity ratios between the posterior and anterior chromatids or between QR.ap and QR.aa nuclei are shown in green at the lower left corner of heatmaps. Anterior of the cell is left. Scale bar: 2 µm. See also Supplementary Movie S9.

(D) Quantification of HDA-1::GFP fluorescence intensity ratios between the large and small daughter cell nuclei and the cell size ratios between the large and small daughters. The names of mother cells are shown on the X-axis. Data are mean ± SEM. N = 9-12. Unpaired two-tailed t-test, **p < 0.01, ***p < 0.001, ns: not significant.

(E) Representative images showing fates of QR.aa after DMSO (top), Baf A1 (middle) and Baf A1 plus AID treatment (bottom). The Q cell plasma membrane and chromosome are labeled by mCherry. HDA-1::GFP is shown as inverted images. Yellow arrow head shows a short neurite-like outgrowth of QR.aa in Baf A1 treated larvae. Frequencies of QR.aa survival and HDA-1 maintenance are showed on the right. Time 0 min is the birth of QR.aa. Scale bar: 5 µm.

V-ATPase distribution during ACDs and a model.

(A) Dynamics of VHA-17::wrmScarlet during QR.a cell division in DMSO control or Baf A1-treated animals. VHA-17::wrmScarlet fluorescence is shown as heatmaps. Time 0 min is the onset of anaphase. Scale bar: 2 µm.

(B) Quantification of VHA-17::wrmScarlet (magenta) and HDA-1::GFP (blue) fluorescence intensity ratios between the posterior and anterior half of QR.a or between QR.ap and QR.aa. Data are mean ± SEM. N = 10 – 14. One-sample t-test compare to 1, ***p < 0.001, ns: not significant. See also Supplementary Movie S10.

(C) A proposed model. Asymmetric segregation of V-ATPase mediated the enrichment of its associated NuRD in the large daughter cell, where high level of NuRD deacetylates egl-1 and suppressed its expression.

Single-cell sequencing of the C. elegans embryonic cells.

(A) Schematic of the single-cell SPLiT-Seq workflow for C. elegans embryonic cells.

(B) Number of genes detected per cell is generated from exon, intron, and exon plus intron mapped reads. Each dot represents a cell, and each box represents the median and first and third quartiles per group.

(C) The distribution of reads into different mapping feature categories per cell. Each dot represents a cell, and each box represents the median and first and third quartiles.

(D) The left pie chart indicates the proportion of detected genes in SPLiT-Seq data of different cell types. Numbers within the pie chart indicate the number of genes. The right pie chart shows 806 genes out of 6487 genes that were not detected from the egl-1-positive cells with enrichment for the nucleus GO terms in the Gene Ontology Cellular Components (GOCC) analysis. See also Supplementary Table 1.

(E) Bar plot ranking of representative GOCC terms with a p-value lower than 10-6 of 806 genes localized to the nucleus. The “histone deacetylase complex” GOCC term includes hda-1, chd-3, lin-40, and egl-27, which encode NuRD subunits.

(F) Violin plots show counts per cell for egl-1/Pegl-1::gfp and NuRD component-encoding genes in apoptotic cells and surviving cells.

Asymmetric segregation of overexpressed NuRD during ACD of QR.a.

Fluorescence time-lapse images of overexpressed GFP-tagged NuRD subunits and mCherry-tagged plasma membrane and histone during asymmetric divisions of QR.a cells. In each panel, the top row shows merged images, the middle row shows mCherry-tagged plasma membrane and histone, and the bottom row shows inverted images of the GFP signal. Scale bar: 2 µm. See also in Supplementary Movie S1 and S2.

Asymmetric segregation of endogenous NuRD during ACDs of Q cells.

(A) Representative images of HDA-1::GFP (top), LIN-53::mNeonGreen (middle), and MYS-1::GFP (down) during QR.a division. The anterior of the cell is on the left. In each panel, the top row shows merged images, the middle row shows mCherry-tagged plasma membrane and histone, and the bottom row shows heatmaps of GFP/mNeonGreen fluorescence. Scale bar: 2 µm. See also Supplementary Movie S3-S5.

(B) Representative images show HDA-1::GFP (left) and LIN-53::mNeonGreen (right) in daughters of indicated Q cells. Inverted images of GFP GFP/mNeonGreen signal are on the right of merged images. Dotted purple lines show cell peripheries. The anterior of the cell is toward the left. Scale bar: 2 µm.

Quantifications of asymmetric NuRD segregation during ACD.

(A) Schematics of the fluorescence quantification method.

(B) Quantification of the fluorescence intensity ratio changes of HDA-1, LIN-53, and MYS-1 in dividing Q cells. Data are mean ± SEM. N = 10 – 12.

(C) The relative total fluorescence of HDA-1, LIN-53 and MYS-1 during ACDs in Figure 1F and S4B. The total fluorescence (NMF1*S1+ NMF2*S2) at each point was normalized to that at time point 0 min. Data are mean ± SEM. N = 10 – 12.

Symmetric NuRD segregation in pig-1 mutant.

(A) Representative images of HDA-1::GFP (left) and LIN-53::mNeonGreen (right) during ACDs of Q cells in pig-1 (gm344) mutant. In each panel, the top row shows merged images, the middle row shows mCherry-tagged plasma membrane and histone, and the bottom row shows inverted images of GFP/mNeonGreen fluorescence. Anterior of the cell is left. Scale bar: 2 µm. See also Supplementary Movie S6 and S7.

(B) Fluorescence intensity ratios of HDA-1::GFP (left) and LIN-53::mNeonGreen (right) between the surviving and apoptotic daughters of Q cells in WT or pig-1 (gm344) mutants. Data are presented as mean ± SEM. Data derived from WT and pig-1 mutants were compared by unpaired t-test, *p< 0.05, **p< 0.01, ***p< 0.001, N = 9– 19.

RNAi of hda-1 and lin-53 reduce fluorescence of HDA-1::GFP and LIN-53::mNeonGreen and enhance H3K27 acetylation level.

(A) Fluorescence images of oocytes (top and middle) and bright-field (bottom) in HDA-1::GFP (left) and LIN-53::mNeonGreen (right) KI animals raised on bacteria expressing control, hda-1 or lin-53 dsRNAs. Scale bar: 5 µm.

(B) Immunofluorescent images with the anti-H3K27ac antibody of control, hda-1 or lin-53 RNAi embryos around the same developmental stage. DAPI stained nuclei. Anterior of the cell is left. Scale bar: 5 µm.

(C) Fluorescence intensity quantification of H3K27ac, histone H3, and DAPI from control, hda-1 or lin-53 RNAi embryos. The fluorescence intensity is arbitrary units (A.U.). N = 16 - 17. Error bars indicate SEM. Unpaired two-tailed t-test, compared with the control RNAi, ***p < 0.001, ns: not significant.

HDA-1 regulates apoptotic cell fate through the canonical apoptosis pathway.

(A) Inverted images of Phsp::sAnxV::GFP from ced-1(e1735), ced-1(e1735); ced-4(n1162), ced-1(e1735); ced-9(n1950), or ced-1(e1735); egl-1(n1084n3082) embryos between late gastrulation stage and bean stage, treated with control or hda-1 RNAi. Scale bars: 5 μm.

(B) Bright-field and fluorescent images of the hda-1 or lin-53 RNAi embryos that expressed the Pegl-1::NLS::GFP reporter. The anterior is to the left, and the dorsal is up. Scale bar: 5 μm.

The composition of C. elegans NuRD and NuRD subunits identified by co-IP and mass spectrometry.

(A) Upper, a schematic representation of the NuRD complex (Bracken et al., 2019; Lai and Wade, 2011). Lower, C. elegans homologs of NuRD subunits and their function.

(B) Mass spectrometric analysis of proteins purified by anti-GFP agarose beads from HDA-1::GFP KI worms or transgenic worms expressing an actin protein ACT-4::GFP. The plot compares average counts of proteins from two biological replicates co-precipitated with HDA-1::GFP with protein counts co-precipitated with the control ACT-4::GFP. ACT-4::GFP affinity purification pulled down ACT-4 but none of the NuRD subunits (Y-axis), whereas HDA-1::GFP pulled down all the known NuRD components but not ACT-4 (X-axis). The number of PSM is the total number of identified peptide spectra matched for the protein. See also Supplementary Table 4.

V-ATPase co-localizes with endoplasmic reticulum.

(A) Representative double-labeling images of VHA-17 and the ER marker SP12 (left) or the late endosomal marker RAB-7 (right) at metaphase and anaphase in dividing QR.a cells. Scale bar: 2 µm.

(B) Quantification of VHA-17 with SP12 or RAB-7 colocalization using Manders overlap coefficient. N = 9 -13. Unpaired two-tailed t-test, ***p < 0.001.

C. elegans strains in this study.

Genomic targets for CRISPR

Plasmids and primers used in this study Captions for Supplementary Information.