KRAB zinc finger protein-mediated control of transposon-embedded regulatory sequences from human germline to early embryos

Transposable elements (TEs), also called jumping genes, are genetic elements capable of changing their position within the genome of their host. They make up large fractions of genomes, including 45% of human DNA content, according to current estimates.

Some of these elements are of retroviral origin and are known as endogenous retroviruses (ERVs). Long terminal repeats (LTRs) of ERVs, like other TEs, influence host gene expression in various ways. They can for example create new promoters, enhancers or imprinted regions upon integration. As ERVs derive from a germline infection in the common ancestor of a species, they are often restricted to a particular evolutionary clade.

While rewiring by TEs can harm a host genome, some of the features of TE invasions can increase host fitness and appear to be universally exploited by various organisms. For instance, TE insertions happen faster than protein-coding genes evolve. They can therefore be coopted for a species-specific evolutionary remodeling of transcriptional networks. Specifically, the LTR12/ERV9 TE family has colonized primate genomes over the past 30 million years and has been widely reported to contribute regulatory elements to the human genome.

In this work, we set out to establish the features that made this family successful, and to assess the prevalence of LTR12-derived regulatory elements in early development. We observe the hominoid-restricted LTR12C subfamily to be a driver of transcription in both gametogenesis and early embryogenesis. LTR12C loci have rewired genes involved in ciliogenesis and flagellogenesis that are critical for male fertility. Relatedly, they also present unusual behavior during the epigenetic reprogramming characteristic of these periods by resisting genome-wide demethylation in primordial germ cell (PGC) development.

To identify the determinants of the epigenetic status of LTR12C elements that made them attractive for cooption, we turn to KRAB zinc finger proteins (KZFPs). TEs appear to be prominent targets of this large family of epigenetic repressors.

Through functional studies, we determine ZNF676 is a regulator of LTR12C, with its targets resisting genome-wide demethylation. This resistance possibly pre-marks them for zygotic rather than maternal expression. We show that ZNF676 levels influence the expression of some cilium-related genes and are able to disrupt ciliogenesis.

We conclude this KZFP likely coevolved with its target ERVs, which carried its binding sites into the vicinity of multiple genes, including those important for male fertility. KZFPs can therefore become regulators of genes with a common function along with the transposable elements they target.

More generally, in our study of TEs expressed in human embryonic development, we find the expression of some TE families to be a hallmark of a particular development stage, an indicator of the current chromatin state. These markers provide a criterion that can be used to assess the similarity of embryonic development models to real embryos. Transcription of some marker TEs anticorrelates with that of their putative KZFP regulators, implying proper KZFP expression to be important to properly recapitulate embryonic development.

Our results provide insight into the TE- and KZFP-mediated rewiring of transcriptional networks related to the germline and early development, and suggest that the improper expression of a KZFP can have consequences ranging up to infertility.