Co-optation of Tandem DNA Repeats for the Maintenance of Mesenchymal Identity

Tandem repeats (TRs) are generated by DNA replication errors and retain a high level of instability, which in principle would make them unsuitable for integration into gene regulatory networks. However, the appearance of DNA sequence motifs recognized by transcription factors may turn TRs into functional cis-regulatory elements, thus favoring their stabilization in genomes. Here, we show that, in human cells, the transcriptional repressor ZEB1, which promotes the maintenance of mesenchymal features largely by suppressing epithelial genes and microRNAs, occupies TRs harboring dozens of copies of its DNA-binding motif within genomic loci relevant for maintenance of epithelial identity. The deletion of one such TR caused quasi-mesenchymal cancer cells to reacquire epithelial features, partially recapitulating the effects of ZEB1 gene deletion. These data demonstrate that the high density of identical motifs in TRs can make them suitable platforms for recruitment of transcriptional repressors, thus promoting their exaptation into pre-existing cis-regulatory networks. Tandem repeats, iterated and unstable sequences generated by DNA replication errors, can be integrated into ancient gene regulatory networks controlling mesenchymal identity and stabilized in the human genome.