Design principles for controlling gene expression
Joao C. Guimaraes · Portuguese National Funding Agency for Science, Research and Technology (RCAAP Project by FCT) · 2014
Control of gene expression underlies the majority of cellular processes and, hence, it is of utmost importance to understand how living organisms tailor protein levels precisely at all times. In addition to comprehending how natural systems tune endogenous expression levels, it has recently become critical to develop genetic tools enabling reliable control of gene expression within synthetic circuits for biotechnology purposes. To this end, synthetic biologists seek parts (DNA segments) with diverse functional properties that once assembled together yield predictable behavior. Nonetheless, the design cycle of synthetic genetic circuits remains heavily dependent on multiple rounds of trial-and-error and manual tinkering. One of the main hurdles faced by synthetic biology is the unpredictable behavior resulting from the reuse of genetic elements whose activities vary across changing contexts. Methods are lacking for researchers to affordably coordinate the quantification and analysis of part performance in different environments, as needed to identify, evaluate and improve problematic part types. We demonstrate how the combination of careful experimental designs and appropriate statistical frameworks can be used for quantifying the performance of genetic elements as they are reused in varying contexts. This methodology revealed design flaws of current gene expression platforms leading to unpredictable behavior. It further motivated the engineering of enhanced genetic elements that can reliably express sequence distinct genes across a 1,000-fold observed dynamic range and within twofold relative target expression windows with ~93% reliability. Other than engineering efforts, a better understanding of how natural systems precisely control gene expression is equally important. However, living organisms optimized by evolution are inherently complex and, commonly, difficult to understand. In this case, systems must be analyzed using integrative approaches that consider the multiple factors potentially affecting the observed phenotype. To facilitate in silico analyses of these multi-factorial behaviors, we have developed an extendable software framework, D-Tailor, affording the automated inference of multiple relevant biological signals from plain genomic sequences. The software also implements a design module that allows researchers to generate artificial sequences exploring a wide range of parameters of interest so as to create more robust datasets to support the hypothesis being tested. We further demonstrated the validity of the above-mentioned integrative approach by evaluating more than 100 sequence features impacting translation efficiency across the E. coli genome, and also by exploring the determinants of specificity and functionality of the RNA-IN/OUT antisense RNA regulation system. In summary, the work presented here shows how computational analysis frameworks can be efficiently combined with experimental approaches to get new insights into the design principles of natural and engineered genetic elements controlling gene expression. Such approaches will be essential for the engineering of more robust artificial systems and, ultimately, lead to the full understanding and modeling of natural biological systems.