Decoding tissue-specific enhancers in plants using massively parallel assays and deep learning

Enhancers control gene expression, orchestrating plant development and responses to stimuli. However, the regulatory codes of enhancers that confer tissue-specific expression in plants remain largely unexplored. Using massively parallel reporter assays (MPRAs) in tomato tissues, we tested the enhancer activity of 11,180 promoter fragments derived from fruit-specific genes. We discovered 2,436 active fruit enhancer sequences, a subset of which showed differential activity between fruit and leaves, suggesting that they can drive fruit-specific gene expression in tomato. We dissected the sequence determinants of fruit enhancers using deep learning. Guided by the regulatory rules learned from our MPRA dataset, we designed synthetic enhancers and experimentally validated their ability to specifically target tomato fruit. Our study provides a comprehensive landscape of functional enhancers in tomato fruit, facilitating the de novo design of synthetic enhancers for tissue-specific gene expression in plants.