Regulation of endogenous retroelements

Retrotransposons can be divided into those that contain long terminal repeats (LTRs) and those that lack LTRs. Retrotransposons that contain LTRs are believed to be remnants of retroviruses, are therefore called endogenous retroviruses, and constitute about 9.45% of the human genome. No endogenous retroviruses are known to be currently transpositionally active in humans, with the most recent transpositions estimated to have occurred about 0.67 million years ago.

Retrotransposons that lack LTRs constitute about 34% of the human genome and mostly have uncertain origins: Long interspersed nuclear elements (LINEs, about 18,99% of the genome) encode two proteins, an RNA-binding protein (ORF1p) and a reverse transcriptase/endonuclease (ORF2p), that confer autonomous transposition activity; short interspersed nuclear elements (SINEs, about 14.66% of the genome) do not encode proteins and require proteins produced in trans by LINEs for transposition. Of the 3 families of LINEs, only the LINE1 family is known to be currently active. Of the SINEs, only Alu elements, which evolved from the 7SL RNA of the signal recognition particle, are known to be currently active.

Surprisingly, although retroelements can cause deleterious mutations due to insertions and recombination, their genomic elements and expression are required for embryogenesis. The envelope proteins ERVW-1 (Syncytin-1) of the HERV-W endogenous retrovirus and ERVFRD-1 (Syncytin-2) of the HERV-FRD endogenous retrovirus act to fuse cells in the trophoblast (reviewed in Gholami Barzoki et al. 2023). TEs contain binding sites for pluripotency factors (for example OCT4, SOX2, and NANOG) as well as cell lineage-specific factors, In humans, a hominid-specific TE family, LTR5Hs, become transcriptionally active during specification of primordial germ cells (PGCs) and serve to bind PGC transcription factors. In mice, the retroviral MERVL element is transcribed in cleavage-stage embryos and during this time MERVL regulatory DNA elements activate expression of more than one hundred genes involved in zygotic genome activation. In mouse embryos, transcripts from LINE1 retroelements are required to exit the 2-cell stage by serving as scaffolds for Nucleolin (NCL) and TRIM28 (KAP1) to repress Dux expression and activate rRNA expression. LINE1 elements are also active in expressing cell type-specific transcripts in the developing and adult human brain. By binding transcription activators, retroelements such as LTR5 and LTR7 can also act as cis-regulatory networks in somatic cells (reviewed in Sundaram and Wysocka 2020, Low et al. 2021, Fueyo et al. 2022).

Retroelements are generally silenced transcriptionally by DNA methylation and histone modifications (reviewed in Geis and Goff 2020, Carotti et al. 2023, Stamidis and Zylicz 2023) or post-transcriptionally by RNA interference (reviewed in Geis and Goff 2020), however repressive chromatin marks are lost and retroelements are transcribed at two points during mammalian development: in early germ cells prior to meiosis and in zygotes immediately after fertilization (reviewed in Low et al. 2021). In germ cells, small RNAs of 24-31 nucleotides known as PIWI-interacting RNAs (piRNAs) are generated from transcripts of retroelements and, when bound to PIWI proteins, guide post-transcriptional decay of retroelement transcripts and re-impose chromatin modifications that repress transcription of retroelements.

Endogenous retroelements are also transcriptionally silenced by zinc finger-containing proteins, KRAB-ZFPs, that bind specific DNA sequences (reviewed in Yang et al. 2017). The human genome contains 423 KRAB-ZFP genes that encode 742 KRAB-ZFP proteins that appear to evolve in response to invasion by retroelements (reviewed in Huntley et al. 2006, Lupo et al. 2013). Specific KRAB-ZFPs bind specific families of retroelements and recruit the TRIM28 scaffold protein (also known as KAP1), which assembles a complex containing the histone methylase SETDB1 and the NuRD repressor complex to silence transcription of the retroelements (reviewed in Almeida et al. 2022).

The Human Silencing Hub (HUSH) complex directly binds RNA of LINE1 elements and trimethylates lysine-9 of histone H3 (H3K9) of nucleosomes assembled on the LINE1 that produced the RNA (reviewed in Seczynska and Lehner 2023). Trimethyl H3K9 (H3K9me3) is a repressive mark thus the result is de novo formation of heterochromatin at LINE1 elements. The HUSH complex also binds H3K9me3 and trimethylates H3K9 of adjacent nucleosomes, resulting in propagation of heterochromatin (reviewed in Seczynska and Lehner 2023).

In mice, N6 methylation of adenosine residues in RNAs can cause destabilization of the RNA and transcriptional silencing of the locus that produced the RNA. The effect of N6-methyladenosine on expression of human retroelements is less clear. Xiong et al. (2021) observed that N6-methyladenosine in transcripts of young LINE1 elements increased RNA expression while N6-methyladenosine in transcripts of ancient LINE1 elements decreased RNA expression. The mechanism responsible for the difference may involve the promotion of translation of young LINE1 elements by N6-methyladenosine residues located in the 5' untranslated region.