It’s clear that humans and mice have taken very different evolutionary paths, but new research suggests that transcription from long terminal repeat retrotransposons (LTRs), which are endogenous retroviral elements that can induce DNA methylation in oocytes, may have acted as guiding lights along the forks in the road.
Besides the obvious differences in size, furriness, and ability to use Google, humans differ from one of our favorite model organisms in terms of which genes are genomically imprinted in the germline. Imprinted genes are typically silenced through DNA methylation, on either the paternal or maternal allele, but the evolutionary switch that causes a gene to become imprinted in one species but not the other has been a roadblock for researchers.
A new study from the lab of Louis Lefebvre (University of British Columbia, Canada) built on their previous egg-ceptional collaborative work with the Lorincz lab which found that LTR-initiated transcription units (LITs) are involved in the deposition of de novo DNA methylation. The talented team navigated through published datasets from mice and humans to find that:
- There are 21 regions of the mouse genome that are methylated in oocytes but not sperm, and 125 similarly imprinted regions in humans (igDMRs), according to whole genome bisulfite sequencing (WGBS) data
- 4 igDMRs in mice and 17 in humans are embedded
within or downstream of LITs when overlapped with RNA-seq data from oocytes
- The human-specific LITs belong to families of retroviruses that are found in primate lineages, while the mouse-specific LITs come from rodent endogenous retroviruses (ERVs)
- 16/17 human-specific loci retain their differential methylation in placental cells, 5 of them belong to ERVs common to humans and chimpanzees, and 7 are common to all primates
- This imprinting is retained through the blastocyst but is lost in adult tissues, and the LIT-associated igDMRs are only found in maternally imprinted genes
- LTR insertions that are common to all primates
are usually associated with imprinted LIT-associated genes, as determined by
DNA methylation and gene expression data from macaque oocytes
- Some regions, like the SORD gene, are imprinted in macaques and humans but different LTRs direct methylation traffic in each species
- At some genes, including MCCC1, even though humans and macaques both have an LTR-insertion, it only drives imprinting in humans because it has been mutated in macaques
- LTRs that are conserved between macaques and chimpanzees are imprinted in the placenta of both species when examined through targeted bisulfite sequencing
Next, the researchers threw the evolutionary car in reverse, and went back to rodents to explore whether the presence of specific LTRs is necessary to cause imprinting at a nearby gene. They used CRISPR-Cas9 to delete two LTRs in embryonic stem cells, generated knockout mice, and found that:
- Oocytes without the LTR upstream of Slc38a4 do not gain DNA methylation at the Slc38a4 igDMR postnatally, and the resulting placental cells lack imprinting at this locus, as measured by targeted bisulfite sequencing and allele-specific expression analysis in the progeny
- When the LTR upstream of the Impact gene is deleted, oocytes fail to acquire methylation and it is absent in the maternally inherited allele of offspring.
Lefebvre shares, “We show that DNA methylation at the canonical promoter of species-specific imprinted genes can be established in oocytes via transcription from nearby endogenous retroviral promoters. Our work proposes a new evolutionary model for the switch from biallelic to imprinted expression in related species.”
Find your way over to the original article in Nature Communications, December 2019.