Single‐cell dynamics of chromatin activity during cell lineage differentiation in <i>Caenorhabditis elegans</i> embryos

Article26 April 2021Open Access Source DataTransparent process Single-cell dynamics of chromatin activity during cell lineage differentiation in Caenorhabditis elegans embryos Zhiguang Zhao orcid.org/0000-0002-0084-3783 State Key Laboratory Molecular Developmental Biology, Institute Genetics and Chinese Academy Sciences, Beijing, China University ChinaThese authors contributed equally to this work Search for more papers by author Rong Fan Weina Xu orcid.org/0000-0002-0798-0000 Yahui Kou Yangyang Wang Xuehua Ma Zhuo Du Corresponding Author [email protected] orcid.org/0000-0002-6322-4656 Information Zhao1,2, Fan1,2, Xu1,2, Kou1,2, Wang1, Ma1 *,1,2 1State 2University *Corresponding author. Tel: +86 10 64801699; E-mail: Mol Syst Biol (2021)17:e10075https://doi.org/10.15252/msb.202010075 PDFDownload PDF article text main figures. Peer ReviewDownload a summary the editorial decision including letters, reviewer comments responses feedback. ToolsAdd favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Abstract Elucidating that orchestrate embryogenesis is fundamental question developmental biology. Here, we exploit position effects on expression as an indicator infer landscape every lineaged early embryogenesis. Systems-level analyses reveal distinguishes cellular states correlates with fate patterning embryos. As unfolds, diversifies lineage-dependent manner, switch-like changes accompanying anterior–posterior asymmetry characteristic landscapes being established different lineages. Upon tissue differentiation, from distinct lineages converges according types but retains stable memories history, contributing intra-tissue heterogeneity. However, cells organized left–right symmetric pattern are predetermined be analogous progenitors so pre-set equivalent states. Finally, genome-wide analysis identifies many regions exhibiting concordant mediate co-regulation functionally related genes differentiation. Collectively, our study reveals genomic at single-cell level. SYNOPSIS This investigates influence local environment reporter gene embryogenesis, based tracing live-cell imaging. The inferred lineage-resolved single C. Chromatin commitment anterior-posterior asymmetry. tissue-specific origins contribute heterogeneity within tissues. Predetermination occurs progenitor left-right symmetry establishment. Introduction state regulates diverse biological processes controlling potential. During vivo development, specific identical genotypes differentiate into through spatiotemporal sets (Yadav et al, 2018; Packer 2019). how dynamic cell-specific regulate has been fields epigenetics biology 2018). Accurate assessments essential understanding it functions development. A combination molecular approaches high-throughput sequencing widely used analyze biochemical biophysical properties chromatin, histone modifications, accessibility, spatial organization (Goodwin 2016). These epigenomic have significantly enhanced profiling mechanistic regulation (Nicetto In addition described above, another powerful strategy elucidating functional position-effect variegation (Schotta 2003; Elgin Reuter, 2013), which classic phenotype associated Drosophila eye color. white-eyed mutant caused epigenetic silencing normally active white genes, due misplacement heterochromatin (Timms findings environments exhibit strong positional modulating potential nearby genes. color provides straightforward readout chromatin; thus, genetic screens identify regulators activity. efforts identified critical constitute much current knowledge biology, protein 1 (HP1) (James Elgin, 1986), tri-methylation H3 lysine 9 (Rea 2000), deacetylase (Mottus 2000). Furthermore, also novel higher organisms, such human hub complex HUSH (Blewitt 2005; Ashe 2008; Tchasovnikarova 2015). identifying regulators, exploited elucidate across genome. these studies, status same integrated dozens thousands positions sensor (Gierman 2007; Akhtar 2013; Chen Frokjaer-Jensen genome provided insight several aspects throughout regulated, domain structural 2007), repressive role lamina (Akhtar germline (Frokjaer-Jensen compared existing sequencing-based approaches, assessment represents unique approach can measure measures ultimate levels, outcomes state. applying developing remains challenging. Previous large-scale primarily focused dissecting Thus, studies involved lines, measuring multicellular organisms tissue/organism systematic not previously performed. work, elegans, model organism studying Using live-imaging approach, quantified levels was than 100 genome, resulting data were (changes positions) corresponding all We revealed general patterns commitment, asymmetry, heterogeneity, bilateral Our systems-level Results Quantification collection transgenic strains (integrants) generated 2014), each containing copy GFP-expressing cassette driven ubiquitous promoter (eef-1A.1, translational elongation factor) (Fig 1A Dataset EV1). Imaging representative integrants position- cell-dependent variation GFP 1B), suggests Peef-1A.1::GFP highly responsive exhibits specificity. Figure 1. Construction Transgenic carrying Peef-1A.1::GFP::NLS (green boxes) NLS denotes nuclear localization signal. Representative micrographs showing mCherry (left) (right) when positions. Scale bar = µm. 3D time-lapse imaging maximum projections show (lineage marker) five stages. Schematic automated identification (circles), (arrows bifurcations indicating divisions), continuous quantification (green). cell. color-coded tree visualizes (green) cells. Vertical lines indicate traced over time, horizontal divisions. landscape. Left: example illustrating inference integrating four (black circle, ABplpappaa). Right: Circos plot landscape, radial line location gradient Images (top) (bottom) ABplpappaa integration sites chromosome I. plots nine embryonic Download figure PowerPoint To allow quantification, nucleus-localized, ubiquitously expressed transgene crossed above nuclei (Dataset performed record high resolution, images analyzed determine individual 1C) (Murray 2014). Cell reconstructed automatic via signal using StarryNite AceTree software (Bao 2006; Boyle Santella 2010; 2014; Katzman 2018), followed multiple rounds manual curations 1D). Simultaneously, intensity nucleus time point measured averaged certain 1D E). On average, 38 consecutive points (range 13–83) (Figs EV1A). total, 268 quantify 113 (at 0.88-Mb resolution) 722 (364 terminal cells), covers up 350-cell stage, specification completed (Appendix Fig S1 EV2). Click here expand figure. EV1. Bar graph number (mean ± SD) (n 268). intensities then represent Verification lines. Green boxes examined study, 18 verified PCR numbered sequences flanking provided. fluorescent attenuation depth. Due embryo orientations, located depth relative microscope objective. DNO VNO dorsal or ventral side near objective, respectively. Heatmap comparing between experimental replicates orientation. Cells far center Z plane grouped. Differences orientations factor. ordered lineage, grouped red green bars representing embryos, upper panel shows raw results, lower result adjusted factor 0.054 exhibited best performance. examples performance correction. Scatter strain before after compensation effect. Distribution average Pearson correlation coefficient consistency on/off replicates. Tree visualization fraction validated accuracy annotation had done EV1B) reliability results (Materials Methods). Additionally, optimized method compensate fluorescence sample EV1C–F) precisely instantaneous improvements allowed reliable comparison Cellular reproducible, median 0.83 binarized 0.90 EV1G). follows fixed generates set differentiated embryo, making entirely comparable resolution (Sulston 1983). although site assayed S1), invariant cell-by-cell lineage-equivalent 1F). activity, constructed distribution (termed landscape) lineage-traced 1F–H EV3). For integrant 2–8), only those 60% considered expressed, consensus Aggregating showed predominantly majority stage (with 364 cells) onward EV1H), most informative. Unless otherwise specified, Position reliably systematically relevance determining whether features its regulating EV2 EV4). EV2. A. Comparison autosomes 97) X 16). B. 66) arm 47) chromosomes. Correlation observed values predicted 19 modifications 113). D. E, F. Changes perturbing H3K9me3 (E) H4K16ac (F). experiment, top micrograph bottom (log2 fold Benjamini–Hochberg-adjusted P-values, Wilcoxon signed-rank test). Cell-by-cell comparisons wt perturbed G. (states 1–3, n 12) silent 10–13, 19) defined Ho al (2014). H. 1–5, 14) 17, 18, 20, 26) Evans (2016). accessible 39) non-accessible 74) chromatin. J. LAD 57) non-LAD 56) regions. K. intragenic 53) intergenic 60) L. endogenous 5-kb 500-kb intervals centered 89 5-kb; 500-kb). M. interval sites. Only transcriptomes assigned identity 38, left panel) two possible identities 267, right analyzed. Given generally very similar anatomy function, treated transcriptome “unique” if pair scRNA-seq (lower panel). Data information: All inter-group statistics Mann–Whitney U-test; correlation. Box plot: band median, box limits first third quartiles, length indicates IQR, whiskers either 1.5 times IQR minimum/maximum value falls IQR. Outliers shown. First, consistent known Similar previous finding 2016), central region EV2A B). Second, modification, key determinant Prediction (Ho 2014) correlated ones EV2C; R 0.82, P 8.27E-29). Twelve EV2D), writers (tri-methylation 9, H3K9me3) activating (acetylation H4 16, H4K16ac) induced expected EV2E F). confirm both indicative segregated various states/domains differential dataset, found domain/states EV2G H). Third, ATAC-seq stages (Janes rest EV2I). Moreover, attached tend heterochromatic low transcriptional (Reddy Kind 2013). reduced lamina-associated (LAD). information obtained determined LEM-2, inner membrane protein, mixed-stage (Ikegami 2010). Indeed, EV2J). Conversely, environments. Integrants EV2K). no level EV2L, left, 0.13, 0.24), (cis-elements genes) did expression. yeast kanR TEF do correlate (Chen modest significant size extended kb, yielding strongest right, 0.57, 3.76E-11). Since large likely governed cis-elements would partially normalized, again supported proxy RNA-sequencing (scRNA-seq) (Packer EV2M EV5), supporting scarcity data, unable assess context other properties. distribution, biochemical, support informative distinguishing Having reliable, next extent 2A). considerably position, taking variability account 2B). quantitative expression,