Niu L et al. (2021), Three-dimensional folding dynamics of the Xenop...

XB-ART-58165

Nat Genet 2021 Jul 01;537:1075-1087. doi: 10.1038/s41588-021-00878-z.

Show Gene links Show Anatomy links

Three-dimensional folding dynamics of the Xenopus tropicalis genome.

Niu L , Shen W , Shi Z , Tan Y , He N , Wan J , Sun J , Zhang Y , Huang Y , Wang W , Fang C , Li J , Zheng P , Cheung E , Chen Y , Li L , Hou C .

???displayArticle.abstract???
Animal interphase chromosomes are organized into topologically associating domains (TADs). How TADs are formed is not fully understood. Here, we combined high-throughput chromosome conformation capture and gene silencing to obtain insights into TAD dynamics in Xenopus tropicalis embryos. First, TAD establishment in X. tropicalis is similar to that in mice and flies and does not depend on zygotic genome transcriptional activation. This process is followed by further refinements in active and repressive chromatin compartments and the appearance of loops and stripes. Second, within TADs, higher self-interaction frequencies at one end of the boundary are associated with higher DNA occupancy of the architectural proteins CTCF and Rad21. Third, the chromatin remodeling factor ISWI is required for de novo TAD formation. Finally, TAD structures are variable in different tissues. Our work shows that X. tropicalis is a powerful model for chromosome architecture analysis and suggests that chromatin remodeling plays an essential role in de novo TAD establishment.

???displayArticle.pubmedLink??? 34099928
???displayArticle.pmcLink??? PMC8270788
???displayArticle.link??? Nat Genet
???displayArticle.grants??? [+]

Species referenced: Xenopus tropicalis
Genes referenced: crebbp ctcf ctrl psmd6 rad21 smarca5

???attribute.lit??? ???displayArticles.show???

	Fig. 1: De novo assembly of the reference genome of X. tropicalis by using Hi-C and single-molecule sequencing. a, Heatmap of chromosome 5 as an example to show assembly errors using the v.9.1 reference genome of X. tropicalis. b, Procedure of de novo assembly of the reference genome of X. tropicalis. c, Comparison between the v.9.1, de novo assembled and v.10.0 chromosome 5. The red lines show sequences with the orientation reversed. d, Heatmap of chromosome 5 to show that assembly errors are mostly corrected in the new version of the reference genome. e, Ideograms of X. tropicalis new reference pseudomolecules. The top track shows the positions of gaps (dark blue). Contigs longer than 1 Mb are shown in black and contigs shorter than 1 Mb are shown in light gray. The Hi-C datasets for genome assembly were generated from s9 embryos.
	Fig. 2: De novo TAD establishment during embryogenesis of X. tropicalis. a, Schematic representation of the ten developmental stages examined by in situ Hi-C. b, Chromatin interaction frequency mapped at a 5-kb resolution. c, Clusters of TAD borders appear for ordinary domains (light blue), loop domains (red) and non-domains (black) at each specific developmental stage. ‘Non-domain border’ refers to a genomic region not identified as a TAD border at a specific embryonic developmental stage, which switches to a TAD border at other development stages. d, Heatmaps of aggregated TADs for the eight developmental stages. The interaction frequency of aggregated TADs from s10 to s23 was normalized against s9. Three phases of change in the TAD structure are shown below, with the developmental stages also shown (TAD number at s9, s10, s11, s12, s13, s15, s17 and s23: 2,471, 2,805, 3,599, 4,036, 4,160, 3,164, 3,609 and 3,199).
	Fig. 3: Orientation-biased CTCF and Rad21 enrichment at TAD borders of higher DI values. a, DI for three clusters of TADs identified in embryos at s13: 555, 3,191 and 414 TADs in clusters 1, 2 and 3, respectively. b, CTCF enrichment is biased to borders with higher DI values. c, Rad21 enrichment is biased to borders with higher DI values and drops toward the border on the other side of the TAD. a–c, Data are presented as the mean ± s.e.m. d, DI for TADs of cluster 1, five subclusters of clusters 2 and cluster 3. According to the ranking of DI strength, TADs in cluster 2 were divided into five subclusters. Data are presented as box plots within violin plots. The minima, maxima, center and bounds of each box plot refers to quartile 1–1.5 interquartile range, quartile 3 + 1.5 interquartile range, median, first and third quartiles of the data. e, Examples of TADs for the three different clusters. f, Aggregated heatmaps for three clusters of TADs. g, Histone modifications and p300 enrichment patterns across the borders of three clusters of TADs. Data are presented as the mean ± s.e.m.
	Fig. 4: De novo TAD establishment is independent of transcription. a, Western blot of RPB1 in WT embryos at the four developmental stages. b, Western blot of proteins in embryos with RPB1 knocked down by morpholinos and in embryos that were rescued. Note that the CTCF and Rad21 protein levels at delayed s10 were similar to s9 WT. Morpholino control (Ctrl); no morpholino (−); rpb1 morpholino (+); rpb1 rescue. See also Supplementary Table 4 for the rpb1 coding sequence for the rescue experiment. c, Schematic representation of the embryogenesis process arrested by RPB1 knockdown and transcription inhibition by α-amanitin. d, Example of a region showing the RPB1 knockdown effects on TAD structure at s9. e,f, Aggregated and normalized heatmaps for s9 and delayed s9 (e), and s10, s13 and delayed s10 (f) for the RPB1 knockdown experiment. g, Example of a region showing the RPB1 knockdown and rescue effects on TAD structure at s11. h, Western blot of proteins in embryos inhibited with α-amanitin. Note that the CTCF and Rad21 protein levels at s9 sustained were similar to s9 WT. Water as control (Ctrl); no α-amanitin (−); the amount of α-amanitin injected was 2 ng per embryo. All western blot experiments in this figure were repeated at least twice unless otherwise stated. i, Example of a region showing the effects of α-amanitin inhibition on TAD structure. j, Aggregated and normalized TAD analysis for embryos of WT s11 and α-amanitin-inhibited s9 sustained.
	Fig. 5: Requirement of CTCF and Rad21 for TAD establishment in X. tropicalis embryos. a, Western blot of CTCF and Rad21 knockdown by morpholinos in embryos at s13. WT, morpholino control (Ctrl), CTCF morpholino (ctcf), Rad21 morpholino (rad21), ctcf rescue (rsc1), rad21 rescue (rsc2), CTCF and Rad21 morpholinos (c+r) and double rescue of ctcf and rad21 (rsc3) are shown. See Supplementary Table 4 for the gene coding sequences for rescue. The western blot experiments in this figure were repeated at least twice with similar results. b, Example region showing the knockdown and rescue effect on TAD structures. y axis: normalized ChIP–seq read count. c, Arrowhead corner score distribution for WT, morpholino Ctrl and knockdowns with CTCF morpholino, Rad21 morpholino, combined CTCF and Rad21 morpholinos and rescued. All experiments were carried out on s13 embryos. Higher corner score values on the x axis indicate a greater likelihood of being at the corner of a domain; the height of the curve indicates the density of corner score values within a specific range. d, Heatmaps of aggregated TADs in WT and Ctrl embryos. e, Heatmaps of aggregated TADs in CTCF and Rad21 knockdown embryos. f, Heatmaps of aggregated TADs in CTCF and Rad21 expression–rescued embryos. In d–f, The black arrows point to interacting borders. The interaction frequency of aggregated TADs was normalized against WT s13.
	Fig. 6: Chromatin remodeling is required for TAD establishment and embryo development. a, Schematic representation of the embryogenesis process arrested by the knockdown of SNF2H, the ATPase component of the ISWI complex. b, Western blot of proteins in embryos with SNF2H knocked down by morpholinos and in embryos that were rescued. WT, morpholino control (Ctrl), SNF2H morpholino (snf2h knockdown), SNF2H rescue (snf2h rescue). The western blot experiments in this figure were repeated at least twice with similar results. c, The domain arrowhead corner score distribution of snf2h knockdown embryos. Morpholino control (Ctrl). d, Example region to show snf2h knockdown and rescue effect on TAD establishment. e, Heatmaps of aggregated TADs normalized against s11.
	Fig. 7: Continuous compartmentalization during embryogenesis. a, Heatmaps of chromosome 2 plotted at a 50-kb resolution at multiple developmental stages. b, Hi-C matrices (observed/expected) for chromosome 2 at a 50-kb resolution at multiple developmental stages. c, Hi-C matrices (observed/expected) of an example region between 0 and 50 Mb in chromosome 2. d, Chromatin switched between A and B compartments during embryo development. e, PCA of compartment scores derived from adjusted Cscore values of Hi-C matrices at multiple developmental stages.
	Fig. 8: Tissue-specific genome architecture in mature brain, liver and sperm cells. a, Heatmaps of chromosome 2 plotted at a 50-kb resolution for brain and liver cells. b, Hi-C matrices (oberved/expected) and Cscore for chromosome 2 at 50-kb resolution for brain and liver cells. c, Heatmaps of an example region in chromosome 2 showing the gain and loss of TAD structures in the brain, liver and sperm cells compared to s13. The single asterisk next to Liver indicates repeated Hi-C on liver cells with K562 as the spike-in control. d, Aggregated TADs of brain, liver and s13 embryo cells and normalization of aggregated TADs against s13 embryos. e, DI cluster for TADs of brain and liver cells. In brain cells, we identified 874, 3,361 and 912 TADs in clusters 1, 2 and 3, respectively. In liver cells, we identified 11, 2,469 and 5 TADs in clusters 1, 2 and 3, respectively. f, Western blot of CTCF, Rad21, SNF2H, RPB1 and histone H3 as a control in brain and liver cells. The western blot experiments in this figure were repeated at least twice with similar results. g, Heatmap of chromosome 2 plotted at a 50-kb resolution for mature sperm cells.
	Extended Data Fig. 1. Heatmap of contact frequency for each chromosome.Assembly errors are shown using the v9.1 reference genome of X. tropicalis for heatmap plotting. Embryos from stage 9 were used for the Hi-C library preparation for Extended Data Figs. 1 and 2.
	Extended Data Fig. 2. Comparison of v9.1, Niu et al. assembled, and v10.0 genome for each chromosome.Red lines show sequences with orientations reversed.
	Extended Data Fig. 3. Analysis for TADs identified at different developmental stages.a, Number of TADs identified at different developmental stages. b, TAD size distribution at different developmental stages. c, Percentage of genome folded into TADs at different developmental stages. d, Density of TSS across TAD borders at s9 and s13. e, Gene expression level across TAD borders at s9 and s13. Data in d and e are represented as mean±SEM. f, Loop and ordinary domains identified at different developmental stages. g & h, PCA analysis of insulation scores for Hi-Cs on embryos at different developmental stages.
	Extended Data Fig. 4. TAD analysis at different developmental stages.a, CTCF and Rad21 binding across loop domain and ordinary domain borders. b, TSS density and gene expression level across loop domain and ordinary domain borders. c, Histone modifications and p300 ChIP-seq signals across loops and ordinary domain borders. All data in this figure are represented as mean±SEM.

	Extended Data Fig. 6. Effects of rpb1 knock-down on TAD establishment.a, Effect of rpb1 knock-down on embryo development. Knock-down of rpb1 was repeated for at least two times with similar results. b, Arrowhead corner score distribution. c, TAD size distribution. d, Percentage of genome folds in TADs.

	Extended Data Fig. 8. Compartment score for the assignment of compartments A and B.Chromosome 2 is shown as example.
	Extended Data Fig. 9. Chromatin switches between compartments A and B.Compartment switches are shown for each chromosome through multiple embryo developmental stages. Red and blue colors show chromatin in compartment A and B, respectively. Grey, yellow and purple lines show no switch, B to A, and A to B switch, respectively.
	Extended Data Fig. 10. TAD structure in terminally differentiated brain and liver tissues.a, Comparison of arrowhead corner score distribution for s13 embryos, brain, and liver tissues. b, Number of TADs and size distribution. c, The percentage of genome folded into TADs in brain and liver tissues. d-f, Example regions to show TADs structure in biological replicate Hi-Cs for brain, liver, and spike-in K562 cells. For each Hi-C replicate, 8 million paired reads from the genome-wide interactions were randomly selected and used for the heatmap plotting. * indicates Hi-C on liver tissue with human K562 as spike-in control.

References [+] :

, American Psychiatric Association issues two policies on AIDS. 1988, Pubmed