Human-specific retrotransposons encode the regulatory logic of two lineages in a bipotent neural crest and central nervous system precursor

Abstract

Transposable elements are sources of cis-regulatory innovation, and the youngest families are particularly interesting because only recent insertions can account for species-specific regulatory divergence. SVAs (subfamilies -E, F), LTR5Hs and L1HS are human-specific retroelements. They are polymorphic, and in some cases still mobile. Regulatory activity in these families has been surveyed mainly in pluripotent cells and differentiated tissues, leaving the early embryonic precursors in which lineage decisions are made largely unexplored. We profiled chromatin accessibility and gene expression in human iPSC-derived bipotent neuroectodermal aggregates harvested at a stage that precedes lineage commitment and retains competence for both the neural crest and the central nervous system. Transcriptional profiling confirmed this dual competence: neural plate border, neural crest and forebrain programs were concurrently active, and definitive neural crest specifiers not yet induced. Of ~4,300 annotated human-specific transposons, 289 were reproducibly accessible in the aggregates. Relative to inaccessible elements, these were closer to genes, occupied gene-dense neighborhoods, and had more highly expressed neighboring transcripts. They were enriched for transcription factor binding motifs drawn from both neural crest and neuronal programs, mirroring the dual competence of the cells. Co-option was strongly family-biased in both extent and content, with SVA-F and LTR5Hs over-represented among accessible elements and each family carrying a distinct motif repertoire. Interestingly, protein interaction analysis of the genes near accessible elements recovered a module built around TRIM28 and multiple KRAB zinc-finger proteins, the machinery that silences transposons through H3K9me3 suggesting a potential feedback loop. These data identify human-specific elements as candidate regulatory substrates at a bipotent developmental stage.