de novo genes originate in regions of ancestrally closed chromatin

Abstract The evolution of de novo genes from ancestrally non-transcribed sequences has been reported in a wide range of taxa. However, the DNA sequence changes responsible for the origin of de novo genes are poorly understood. In this study, we investigated whether de novo genes evolve by exploiting ancestral regulatory sequences, or are associated with the evolution of new regulatory regions. We focused on the youngest stratum of de novo genes that have polymorphic expression in the male accessory gland of Drosophila melanogaster and show no evidence of transcription in other Drosophila species. ATAC-seq in five D. melanogaster genotypes showed that open chromatin correlates with the expression state of de novo genes, and is more prevalent in the transcribed than in the non-transcribed alleles. In the outgroup species D. simulans and D. yakuba, the sequences homologous to most D. melanogaster de novo genes have closed chromatin, suggesting that open chromatin around de novo genes is a derived state. We used 29 D. melanogaster genotypes to map cis-regulatory variants associated with de novo gene expression in accessory glands. Despite low power, we were able to identify at least five de novo genes for which local sequence variants are associated with transcription. Many of these variants are located in and around the regions of differentially accessible chromatin that correlate with genotype-specific transcription. We conclude that most de novo genes are associated with the origin of new regulatory sequences, which typically requires multiple DNA sequence changes.