Abundance-Biased Codon Diversification prevents recombination in AAV production and ensures robust in vivo expression of functional FRET sensors

The delivery of genetically encoded fluorescent sensors via adeno-associated viral vectors (AAVs) enables the quantification of biological analytes with high spatiotemporal resolution in living animals. In this study, we expose an unreported problem of the approach, in which the presence of repeated subsequences in the sensor’s DNA sequence triggers recombination during AAV production. In the case of Förster Resonance Energy Transfer (FRET) sensors, recombination leads to a mixture of fluorescent products, severely compromising in vivo functionality. To counter this phenomenon, we introduced Abundance-Biased Codon Diversification (ABCD), a computational method that prevents recombination without sacrificing codon optimization for a target organism. We demonstrate that ABCD greatly facilitates in vivo studies by restoring the functionality of FRET sensors and advanced inducible expression systems delivered via AAV vectors. We thereby offer a robust solution to a previously overlooked challenge for quantitative imaging and genetic manipulation in living animals using AAV-mediated strategies.