Computer-aided biochemical programming of synthetic microreactors as diagnostic devices (vol 14, e7845, 2018)

Biological systems have evolved efficient sensing and decision-making mechanisms to maximize fitness in changing molecular environments. Synthetic biologists have exploited these capabilities to engineer control on information and energy processing in living cells. While engineered organisms pose important technological and ethical challenges, de novo assembly of non-living biomolecular devices could offer promising avenues toward various real-world applications. However, assembling biochemical parts into functional information processing systems has remained challenging due to extensive multidimensional parameter spaces that must be sampled comprehensively in order to identify robust, specification compliant molecular implementations. We introduce a systematic methodology based on automated computational design and microfluidics enabling the programming of synthetic cell-like microreactors embedding biochemical logic circuits, or protosensors, to perform accurate biosensing and biocomputing operations invitro according to temporal logic specifications. We show that proof-of-concept protosensors integrating diagnostic algorithms detect specific patterns of biomarkers in human clinical samples. Protosensors may enable novel approaches to medicine and represent a step toward autonomous micromachines capable of precise interfacing of human physiology or other complex biological environments, ecosystems, or industrial bioprocesses. For the first time we show that the implementation of Boolean logic circuits with reactive biochemical species can be automated to satisfy user defined temporal logic specifications and their behavior optimized for robustness. We demonstrate the programming, synthesis and operability of three different instances of synthetic biochemical logic circuits encapsulated within synthetic phospholipid bilayers. Using these methodologies, we generate proof-of-concept diagnostic microreactors, or protosensors, biochemically programmed to detect specific patterns of biomarkers and classify pathological states insitu. We demonstrate their capabilities for the diagnosis of acute diabetes complications in human clinical samples.