Recurrent DNA break clusters drive replication-stress-induced copy number variants and genome diversification.

Abstract

Copy number variants (CNVs) are strongly implicated in neurological and psychiatric disorders and brain cancer, yet the process by which replication stress generates CNVs-and why some recur while others remain rare-remains poorly understood. Here, we show that recurrent DNA-break clusters (RDCs) act as common initiating lesions that drive both recurrent and non-recurrent CNVs. In murine neural progenitor cells subjected to chemically induced replication stress, bulk whole-genome sequencing identifies recurrent CNVs enriched at late-replicating RDCs within actively transcribed genes. Single-cell genome sequencing further uncovers frequent, non-recurrent CNVs associated with RDCs that arise during the transition from early to late DNA replication. These CNVs represent stable, heritable structural variants with breakpoints consistently enriched at RDCs. CRISPR/Cas9-mediated transcriptional suppression abolishes both RDC formation and CNV generation, establishing RDC-associated breaks as a shared upstream source. Mechanistically, CNV formation depends on DNA repair context: CNVs are Pol θ-dependent in NHEJ-deficient cells but arise independently of Pol θ in NHEJ-proficient cells. Together, these findings define RDCs as central drivers of replication-stress-induced genome diversification.