Back

Tissue-specific restriction of TE-derived regulatory elements safeguards cell-type identity

Milovanovic, D.; Duc, J.; Matsushima, W.; Hamelin, R.; Planet, E.; Offner, S.; Rosspopoff, O.; Trono, D.

2025-05-15 genomics
10.1101/2025.05.13.653700 bioRxiv
Show abstract

Transposable elements (TEs) have extensively reshaped the cis-regulatory landscape of mammalian genomes, yet the mechanisms that govern their context-specific activity remain incompletely understood. KRAB zinc finger proteins (KZFPs), a large family of transcription factors specialized in TE recognition, are known repressors of TE-derived regulatory activity through TRIM28-mediated H3K9me3 deposition. Here, we expand this paradigm by uncovering noncanonical relationships between TEs and KZFPs. By generating a comprehensive epigenomic map of KZFP-bound TEs, we find that the regulatory activity of ancient mammalian L2/MIR elements is broadly delineated by KZFP binding patterns, despite low H3K9me3 enrichment. We further dissect this relationship by investigating the function of ZNF436, a non-canonical KZFP, highly expressed during in vivo human fetal heart development. Using loss-of-function approaches, we show that ZNF436 preserves cardiomyocyte function by promoting cardiac gene expression while restricting the activation of alternative lineage programs. Mechanistically, ZNF436 recruits specialized SWI/SNF chromatin remodeling complexes to limit the accessibility of L2/MIR-derived enhancers, many of which are active in non-cardiac tissues. These findings reveal a noncanonical, TRIM28-independent role for KZFPs in shaping cell-type-specific regulatory landscapes and emphasize the importance of repressing alternative regulatory programs alongside activating lineage-specific ones to safeguard cell identity.

Matching journals

The top 7 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.