3D genome, on repeat: Higher-order folding principles of the heterochromatinized repetitive genome

Abstract

Nearly half of the human genome is comprised of diverse repetitive sequences ranging from satellite repeats to retrotransposable elements. Such sequences are susceptible to stepwise expansions, duplications, inversions, and recombination events which can compromise genome function. In this review, we discuss the higher-order folding mechanisms of compartmentalization and loop extrusion and how they shape, and are shaped by, heterochromatin. Using primarily mammalian model systems, we contrast mechanisms governing H3K9me3-mediated heterochromatinization of the repetitive genome and highlight emerging links between repetitive elements and chromatin folding.

Figure 1.. History of classic heterochromatin and repetitive genome discoveries

Key principles of heterochromatin acquisition, maintenance, and function since the initial discovery in 1928 by Emil Heitz (top). In parallel, insights into the abundance, regulation, and function of repetitive elements have also come to light, including Barbara McClintock’s 1950 foundational discovery of mobile genetic elements in maize and recent reports of the first gapless human genome (bottom). Crosstalk between heterochromatin and the repetitive genome has emerged as understanding has accumulated in both disciplines.

Figure 2.. Interplay between higher-order chromatin organization and heterochromatin

Mammalian genomes are folded into (A) A/B compartments, (B) chromatin-associated subnuclear bodies, (C) topologically associating domains (TADs), (D) subTADs, and (E) long-range looping interactions. (F) TADs/subTADs formed by cohesin-mediated loop extrusion manifest as corner-dot TADs/subTADs in ensemble Hi-C maps, whereas dot-less TADs/subTADs have weaker boundaries or are formed by cohesin-independent mechanisms. (G) Loop extrusion is refractory to B compartment and heterochromatin domain formation (based on Haarhuis et al. [2022]).

Figure 3.. Types and distribution of repetitive DNA elements across the human genome

(A) The human genome consists of ~54% repetitive sequences. Class I repeats comprise retrotransposable elements, including long interspersed nuclear elements (LINEs, grey), short interspersed nuclear elements (SINEs, dark orange), and long terminal repeat (LTR) endogenous retroviruses (ERVs, light orange). The human genome has primarily one abundant LINE, LINE-1, that is ~6 kb long and encodes its own requisite reverse transcription and transposition machinery. The most common class of SINE elements in the human genome are the ~280-bp-long Alu-SINEs which are non-autonomous and rely on LINE-1 encoded machinery to facilitate their transposition. ERV remnants are present in the human genome in fragments (HERVs) and in some cases can be actively transcribed. Tandem repeats are non-mobile DNA sequences in which the copies of one or more nucleotides are repeated in a head-to-tail manner. They are distributed throughout the genome and are most prevalent as satellite repeats at pericentromeric and centromeric chromatin regions. (B) The recent gapless telomere-to-telomere (T2T) human genome assembly consists of assemblies for all 22 human autosomes and chromosome X and comprises 3.05 gigabase pairs (Gbp) of nuclear DNA (CHM13v1) (adapted from Nurk et al. [2022]). Bar plot, recent T2T estimates of the percentage of the human genome made up of each repeat class. TSD, target site duplication; ORF, open reading frame; EN, endonuclease; RT, reverse transcriptase; A and B, split Pol III RNA promoter; LTR, long terminal repeats; Gag, group antigens; Prt, protease; Pol2, reverse transcriptase; Env, envelope protein.

Figure 4.. Distinct repeat elements require unique heterochromatinization mechanisms

Linear heterochromatin mechanisms play an integral role in repressing both the accessibility and expression of repetitive DNA elements. (A) At pericentric chromatin, H3K9me3 and DNA 5mC cooperate to silence satellite repeats. DNMT3B and the HMT SUV39H1 can be recruited to pericentromeric repeats and co-immunoprecipitation (co-IP) with HP1α. (B) The interplay between H3K9me3 and DNA 5mC can occur via two distinct models in a DNA-replication-dependent manner. In model A, UHRF1 is located to DNAP-replicating loci through an interaction with PCNA. At replicating loci, UHRF1 binds hemi-5mC DNA and H3K9me3 where it can recruit DNMT1 via a direct protein interaction for 5mC maintenance. In model B, UHRF1 ubiquitinates histone 3 residues which are recognized by DNMT1, leading to its local recruitment for 5mC maintenance. (C) Transcription of pericentromeric satellite repeats generates ncRNAs that can recruit the HMT SUV39H1 in cis for H3K9me3 deposition. (D) At telomeric short tandem repeats (STRs), TERRA ncRNA can recruit heterochromatin machinery to the TRF2 subunit of the shelterin complex. (E) ERVs are primarily targeted for repression by KRAB-containing zinc finger proteins (KRAB-ZFPs) which interact with the corepressor KRAB-associated protein 1 (KAP1) to facilitate recruitment of additional heterochromatin effector proteins. (F) Actively transcribed young LINE-1s are silenced by the HUSH complex, which recruits the H3K9me3 HMT SETDB1 to chromatin. Prematurely terminated transcripts can be degraded via the HUSH-associated NEXT complex. (G) SINEs can be targeted for H3K9me3-dependent repression by SUV39H1 as well as HP1α. SUV39H1, suppressor of variegation 3-9 homolog 1; DNMT3B, DNA methyltransferase 3B; HP1α, heterochromatin protein 1α; UHRF1, ubiquitin-like PHD and RING finger domain-containing protein 1; DNMT1, DNA methyltransferase 1; PCNA, proliferating cell nuclear antigen; DNAP, DNA polymerase; RNAPI, RNA polymerase I; ncRNA, non-coding RNA; TERRA, telomeric repeat-containing RNA; TRF2, telomeric repeat-binding factor 2; ORC1, origin recognition complex subunit 1; ERV, endogenous retroviruses; KRAB, Krüppel-associated box domain containing-zinc-finger proteins; C2H2, Cys-2 His-2 zinc finger; SETDB1, SET domain bifurcated 1; RNAPIII, RNA polymerase II; HUSH, human silencing hub; Transgene activation suppressor; MPP8, M-phase phosphoprotein 8; PPHLN, periphilin; NEXT, nuclear exosome targeting; RNAPIII, RNA polymerase III.

Figure 5.. STR expansion and HERV-H transcription influence TAD boundary integrity

(A) A CGG trinucleotide short tandem repeat (STR) tract residing in the 5′UTR of FMR1 is associated with fragile X syndrome. (B) A study by (Sun et al., 2018) reports that in healthy individuals, the CGG STR (i.e., 6–40 repeating monomer subunits) is present at a TAD boundary, where FMR1 is actively transcribed. (C) As the CGG STR expands to full mutation length of 200 triplets or more, CTCF is displaced and local TAD/subTAD boundaries within 1 Mb around FMR1 are disrupted. (D) A study by Zhang et al. (2019) shows that transcription of HERV-H by the RNA Polymerase II complex presents a physical barrier that interferes with cohesin-mediated loop extrusion, leading to the creation of a local CTCF-independent TAD boundary. (E) De novo HERV-H insertion is sufficient for the formation of a new TAD boundary. (F) Transcriptional repression of an endogenous HERV-H locus can dissolve a pre-existing TAD boundary.