From Natural Discovery to AI-Guided Design: A Curated Collection of Compact Enhancers for Crop Engineering

Area of Science:

• Plant Biotechnology
• Genomics
• Molecular Biology

Background:

• Precise transgene-free gene upregulation is difficult in crop biotechnology.
• Conventional enhancers face size constraints for CRISPR knock-ins and regulatory issues.

Purpose of the Study:

• To create a cross-species resource of compact transcriptional enhancers for crop biotechnology.
• To identify and validate short transcriptional enhancers (STEs) for precise gene upregulation.
• To develop an AI framework for designing synthetic enhancers.

Main Methods:

• Utilized STEM-seq, a high-throughput screening platform, to evaluate 81,475 genomic elements across maize, wheat, tomato, and soybean.
• Developed BaseSearch, an AI-driven design framework, to computationally generate synthetic STE candidates.
• Validated enhancer activity and performance through experimental screening.

Main Results:

• Identified 6904 natural short transcriptional enhancers (STEs; 60-80 bp) with varying activation efficiencies.
• Wheat-derived enhancers showed potent activation (up to 46.3-fold).
• AI-designed synthetic STEs achieved a 9.1% success rate, with ten ultra-potent enhancers outperforming natural ones by 2.27-fold (64.5-fold activation).

Conclusions:

• The developed STEs are compact, aligning with regulatory preferences for endogenous sequences, facilitating policy-compatible precision breeding.
• This integrated platform offers a substantial collection of validated enhancers for engineering crop agronomic traits.
• Advances fundamental understanding of plant cis-regulation and provides applicable tools for crop improvement.