The Antisense RNA Calculator: Rational Design of Synthetic Regulatory RNAs for Controlling Protein Expression

Tunable, rational control of chromosomally encoded protein expression levels would significantly advance our ability to engineer cellular metabolism and physiology for biotechnological applications. Bacterial, non-coding, antisense RNAs have the ability to regulate gene expression by manipulating mRNA translation and stability. Natural regulatory RNAs are ubiquitously found in bacteria, controlling their metabolic responses to environmental stimuli. Synthetic antisense RNAs have the potential to rationally control the translation rate of selected mRNA targets, either increasing or decreasing a targeted protein expression level. However, the field currently lacks a biophysical model that predicts specific antisense RNAs to regulate targeted mRNAs according to a desired specification.

We have developed a statistical thermodynamic model to quantitatively predict the effect of an antisense RNA on a mRNA's translation initiation rate and its resulting protein expression level.  By combining this model with an optimization technique, we create a new automated design method, called the Antisense RNA Calculator, that generates synthetic antisense RNAs according to a desired specification. Antisense RNAs may be designed to increase or decrease a protein's expression level by a selected factor. These antisense RNAs act in trans and thus the target protein may be chromosomally encoded or on a plasmid.

The model relates changes to translation initiation rate to the antisense RNA concentration and the calculated Gibbs free energies of the rate-limiting steps involved in antisense RNA regulation, including ribosome-mRNA binding, antisense RNA-mRNA binding, and mRNA-antisense RNA-ribosome binding. The model predicts the fold-change in protein expression level for any antisense RNA sequence, mRNA sequence, and antisense RNA concentration. The optimization method may design antisense RNAs to target natural mRNAs or co-design a synthetic mRNA for optimal regulation.

We experimentally test the model predictions by rationally designing many antisense RNAs and measuring their effect on the steady-state protein expression level, using flow cytometry and spectrophotometry. We use a synthetic measurement plasmid that contains two fluorescent proteins (mRFP1 and GFPmut3b) and the synthetic antisense RNA. The intracellular concentration of antisense RNA is controlled by a constitutive promoter, selected from a toolbox of promoters with characterized transcription rates.

The Antisense RNA Calculator has broad applications in metabolic engineering and medical therapeutics. By combining our model with a multi-objective optimization algorithm, we may design nucleic acids drugs to selectively kill specific bacterial species within a bacterial community while not harming other species. These nucleic acids, which are chemically modified to increase their bioactivity, regulate genes of the selected species and shut down their metabolism to kill them.