The transcription of genes and engineered circuits can deeply vary when inserted into different genomic loci. This unpredictable performance, termed context sensitivity, complicates strain development. Although the causes and mechanisms of context sensitivity are emerging, it is poorly known how to engineer circuits and synthetic pathways isolated from it. Using tools of synthetic biology for designing and inserting various reporter cassettes in the Escherichia coli genome and RT-qPCR for directly measuring gene transcription, we first surveyed the genomic landscape for context sensitivity at 214 positions in cells grown in glucose or glycerol. The results show deep variations in cassette transcription with respect to position (up to 160-fold) and growth condition (up to a 30-fold). We then demonstrated that this position-dependent transcription variability is strongly reduced when the reporter cassette is insulated in an artificial protein-bound DNA loop. Finally, we measured the transcription of two loop-insulated genes at different genomic positions. The results show that transcription strongly depends on the relative orientation of the genes, promoter strength, and positive supercoiling. We present a model suggesting that DNA looping is an important cause of context sensitivity and can be used for better controlling the transcription of engineered circuits.

The expression of genes and engineered circuits can deeply vary when inserted into different genomic loci. This unpredictable performance of expression, termed context sensitivity, complicates strain development. Although the causes and mechanisms of context sensitivity are emerging, it is poorly known how to engineer circuits and synthetic pathways isolated from it. Using tools of synthetic biology for designing and inserting various reporter cassettes in the Escherichia coli genome and RT-qPCR for directly measuring gene expression, we first surveyed the genomic landscape for context sensitivity at 209 positions in cells grown in glucose or glycerol. Results showed deep variations in cassette expression with respect to position (up to 150-fold) and growth condition (up to a 1400-fold). We then demonstrated that this position-dependent expression variability is strongly reduced when the reporter cassette is insulated in an artificial protein-bound DNA loop. Finally, we measured expression of two loop-insulated genes at different genomic positions. Results showed that expression strongly depends on the relative orientation of the genes, promoter strength and positive supercoiling. We present a model suggesting that DNA looping is an important cause of context sensitivity and can be used for better controlling expression of engineered circuits.