| Tunable control of protein production inside cells is one of the main challenges in metabolic engineering and pharmaceutical industries. There are many cellular processes that are responsible for altering protein expression levels, including transcription initiation and termination, translation initiation and elongation, and co-translational protein folding. Successful manipulation of each of these processes requires the understanding of their mechanism, and how macromolecular machineries such as RNA polymerases and ribosomes interact with DNAs and mRNAs. In particular, the ribosome's interactions with mRNA govern its translation rate and the effects of post-transcriptional regulation. It has been shown that translation initiation is a rate limiting step in protein production, in which ribosome must efficiently make the first contact with mRNA. Although translation initiation is one of the well-studied processes inside cells, there are still many unanswered questions that puzzled researchers for several decades. For example, what is the role of 5' untranslated regions on protein synthesis rate? How the N-terminal coding section controls ribosome binding to mRNA? How the regulatory components such as small RNAs and ligand-responsive riboswitches alter the ribosome binding rate? Can RNA folding kinetics compete with ribosome binding rate to alter the binding equilibrium between mRNA and ribosome?;In this dissertation, we aim to answer these questions by taking a "learn-by-design" approach, in which we first propose a plausible mechanism, and then we conduct specific experiments to test our hypothesis while eliminating other confounding interactions, followed by developing mechanistic models that can explain the experimental observations. Our ultimate goal is to develop simplified yet accurate models that can predict the rate of translation initiation and riboswitch regulation directly from DNA sequence, using kinetics and thermodynamics first principles. These biophysics-based models can be further improved by identifying new interactions, characterizing their effects, and incorporating them into the existing models.;Overall, through our biophysical modeling approach, we developed an accurate and comprehensive model of bacterial translation initiation that predicts the rate of translation under both equilibrium and non-equilibrium conditions. In addition, we also developed the first sequence-structure-function relationship for translation-regulating riboswitches that predicts the function of riboswitches from the energetic interactions between mRNA, ligand, and ribosome. Importantly, the incorporation of this biophysics-based model into a web-interface called the Riboswitch Calculator, allows the scientific community to design new generation of highperformance biosensors for diverse applications. |
