Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2019 Sep;51(9):1380-1388.
doi: 10.1038/s41588-019-0479-7. Epub 2019 Aug 19.

Transcriptionally active HERV-H retrotransposons demarcate topologically associating domains in human pluripotent stem cells

Yanxiao Zhang #  1 , Ting Li #  2 , Sebastian Preissl #  1   3 , Maria Luisa Amaral  4 , Jonathan D Grinstein  2 , Elie N Farah  2   5 , Eugin Destici  2 , Yunjiang Qiu  1   4 , Rong Hu  1 , Ah Young Lee  1 , Sora Chee  1 , Kaiyue Ma  1 , Zhen Ye  1 , Quan Zhu  1 , Hui Huang  1   5 , Rongxin Fang  1   4 , Leqian Yu  6   7 , Juan Carlos Izpisua Belmonte  8 , Jun Wu  6   7 , Sylvia M Evans  2   9 , Neil C Chi  10   11 , Bing Ren  12   13   14   15
Affiliations

Transcriptionally active HERV-H retrotransposons demarcate topologically associating domains in human pluripotent stem cells

Yanxiao Zhang et al. Nat Genet. 2019 Sep.

Abstract

Chromatin architecture has been implicated in cell type-specific gene regulatory programs, yet how chromatin remodels during development remains to be fully elucidated. Here, by interrogating chromatin reorganization during human pluripotent stem cell (hPSC) differentiation, we discover a role for the primate-specific endogenous retrotransposon human endogenous retrovirus subfamily H (HERV-H) in creating topologically associating domains (TADs) in hPSCs. Deleting these HERV-H elements eliminates their corresponding TAD boundaries and reduces the transcription of upstream genes, while de novo insertion of HERV-H elements can introduce new TAD boundaries. The ability of HERV-H to create TAD boundaries depends on high transcription, as transcriptional repression of HERV-H elements prevents the formation of boundaries. This ability is not limited to hPSCs, as these actively transcribed HERV-H elements and their corresponding TAD boundaries also appear in pluripotent stem cells from other hominids but not in more distantly related species lacking HERV-H elements. Overall, our results provide direct evidence for retrotransposons in actively shaping cell type- and species-specific chromatin architecture.

PubMed Disclaimer

Conflict of interest statement

Competing Interests Statement:

B.R. is a co-founder of Arima Genomics, Inc.

Figures

Figure 1.
Figure 1.. Reorganization of TADs during human cardiomyocyte differentiation.
(a) Hi-C contact matrices for each stage of cardiomyocyte differentiation at mega-base resolution. (b) A heatmap showing hierarchical clustering of dynamic chromatin compartments during cardiomyocyte differentiation. The pseudo-color reflects the PC1 values (compartment A/B) of compartment bins. Negative PC1 value stands for compartment B and positive for compartment A. Representative genes located in corresponding compartment bins are annotated to the right of the heatmap. (c) A heatmap showing the DI delta scores for the stage-specific TAD boundaries, ordered by the presence of TADs at six stages. vCM(−) stands for TAD boundaries lost in purified ventricular cardiomyocytes at D80, ESC(+) stands for hESC-specific TAD boundaries. (d) Scatter plot shows the fold enrichment and −log10(P values) for various repeat element classes at ESC(+) (N = 198) and vCM(−) (N = 329) TAD boundaries relative to the static TAD boundaries (N = 2,622). P values are from two-sided proportion test.
Figure 2.
Figure 2.. Transcriptionally active HERV-H forms human ESC-specific TAD boundaries.
(a) Aggregated RNA-seq expression profile (RPKM normalized) at ESC(+) TAD boundaries that overlap HERV-H element. (b) Scatterplot shows the expression levels (RPKM) across different HERV-H loci (ordered by expression levels from high to low in ESC) at D0, D2 and D5 stages of differentiation. (c) Heatmap of aggregated Hi-C contact matrix [log2(observed/expected)] within 200 kb of the top 50, 51-100 and 101-150 ranked HERV-Hs, at D0. (d) Heatmap of the aggregated Hi-C matrix [log2(observed/expected)] within 200 kb of the top 50 HERV-Hs, at D0, D2 and D5. (e,f) Representative Hi-C interaction matrices of two HERV-H loci located at ESC(+) TAD boundaries at D0, D2 and D5 (top) are shown as heatmaps along with genome browser tracks of POLR2A, SMC3, CTCF, H3K27ac ChIP-seq and RNA-seq data of the expanded genomic region containing the TAD boundary (arrow). (g) Aggregated genomic profiles of RNA-seq, POLR2A, SMC3 and CTCF ChIP-seq around top 50 HERV-Hs located on the ESC(+) TAD boundaries (red) and lower ranked HERV-Hs (grey).
Figure 3.
Figure 3.. Deletion of two HERV-H sequences leads to merging of TADs in hESCs.
(a) Hi-C interaction matrices of the wild-type (WT) and transgenic hESC lines (HERV-H1-KO and HERV-H2-KO) are shown, along with DI scores, expression levels (RPKM) and fold changes of gene expression at the HERV-H1 and HERV-H2 loci. The loss of TAD boundary in the transgenic cells is accompanied with decrease of RNA expression of genes 5’ terminus to the HERV-H sequences. (b) Boxplots show expression levels (RPKMs) of genes whose TSSs are located from −500 kb to the 5’ LTR (N = 49) and from 3’ LTR to +500 kb (N = 49) of boundary-associated HERV-Hs. P values are from two-sided paired t test on the log-transformed expression levels. The elements of the boxplot are: center line, median; box limit, upper and lower quartiles; whiskers, 1.5× interquartile range. (c) Line chart (mean ± standard error, N = 3 cardiomyocyte differentiations) shows percentage of TNNT2 positive cells during cardiomyocyte differentiation of WT and HERV-H1-KO hESC lines (two HERV-H1-KO clones analyzed). HERV-H1-KO hESCs display increased cardiomyocyte differentiation efficiency compared to WT hESCs.
Figure 4.
Figure 4.. Silencing of HERV-H sequences weakened the TAD boundaries in hESC.
(a) Design of the CRISPR-dCas9-KRAB system to silence HERV-H expression using sgRNAs targeting 5’ LTR7. (b) Gene expression values of HERV-H1 and HERV-H2 in WT and CRISPRi-targeted hESCs. (c) Hi-C interaction matrices of the CRISPRi targeted hESC lines (sgHERV-H1 and sgHERV-H2) are shown, along with DI scores, and log2(fold-change) of gene expression in engineered cell lines over control at the HERV-H1 and HERV-H2 loci.
Figure 5.
Figure 5.. HERV-H insertion creates de novo TAD boundaries.
(a) Design of the piggybac vector to “transpose” HERV-H to random genomic locations in the hESC (HERV-H2-KO line). (b) Hi-C contact matrices of the parental cell line (HERV-H2-KO) and the cell line with HERV-H insertions (HERV-H-ins.clone1) are shown, along with DI scores at the locus of one HERV-H insertion. (c) Line plots showing the cross-over contact scores of both alleles in parental cell line (HERV-H2-KO) and HERV-H-ins.clone1 at the same locus as in (b). Only the allele (B) harboring HERV-H insertion shows a decrease of cross-over contact score (increase of insulation). (d) Heatmap shows the z-transformed cross-over contact scores for all predicted knock-in (KI) alleles and unaffected alleles in the HERV-H inserted cells.
Figure 6.
Figure 6.. HERV-H introduces new TAD boundaries during primate evolution.
(a) A simplified tree of primate evolution with the copies of HERV-H annotated (left) and Hi-C interaction matrices of human ESCs, bonobo iPSCs, chimpanzee iPSCs, marmoset iPSCs and mouse ESCs are shown, along with DI scores at the syntenic regions to human HERV-H1 locus, HERV-H2 locus, and all top 50 transcribing HERV-Hs. The chimpanzee and bonobo syntenic regions are denoted with a star as they also contain HERV-H sequence. (b) Bar chart shows the percentage of HERV-H transcripts over all transcripts in the PSCs from each indicated species.
See this image and copyright information in PMC

References

    1. Dekker J & Mirny L The 3D Genome as Moderator of Chromosomal Communication. Cell 164, 1110–1121, doi:10.1016/j.cell.2016.02.007 (2016). - DOI - PMC - PubMed
    1. Yu M & Ren B The Three-Dimensional Organization of Mammalian Genomes. Annu Rev Cell Dev Biol 33, 265–289, doi:10.1146/annurev-cellbio-100616-060531 (2017). - DOI - PMC - PubMed
    1. Meaburn KJ & Misteli T Cell biology: chromosome territories. Nature 445, 379–781, doi:10.1038/445379a (2007). - DOI - PubMed
    1. Lieberman-Aiden E et al. Comprehensive mapping of long-range interactions reveals folding principles of the human genome. Science 326, 289–293, doi:10.1126/science.1181369 (2009). - DOI - PMC - PubMed
    1. Dixon JR et al. Topological domains in mammalian genomes identified by analysis of chromatin interactions. Nature 485, 376–380, doi:10.1038/nature11082 (2012). - DOI - PMC - PubMed

Methods-only references

    1. Lian X et al. Directed cardiomyocyte differentiation from human pluripotent stem cells by modulating Wnt/beta-catenin signaling under fully defined conditions. Nature protocols 8, 162–175, doi:10.1038/nprot.2012.150 (2013). - DOI - PMC - PubMed
    1. Hashem SI et al. Brief Report: Oxidative Stress Mediates Cardiomyocyte Apoptosis in a Human Model of Danon Disease and Heart Failure. Stem cells 33, 2343–2350, doi:10.1002/stem.2015 (2015). - DOI - PMC - PubMed
    1. Shen Y et al. A map of the cis-regulatory sequences in the mouse genome. Nature 488, 116–120, doi:10.1038/nature11243 (2012). - DOI - PMC - PubMed
    1. Gorkin D et al. Systematic mapping of chromatin state landscapes during mouse development. bioRxiv (2017).
    1. Jolma A et al. DNA-binding specificities of human transcription factors. Cell 152, 327–339, doi:10.1016/j.cell.2012.12.009 (2013). - DOI - PubMed

Publication types