Summary of the main results on protease based signaling

• Three TEVp homologues (PPVp, SbMVp and SuMMVp) were tested and proved to be fully orthogonal.
• Upon overexpression none of the tested proteases was toxic to mammalian cells, demonstrating that they do not interfere with essential cellular processes.
• Four new active split site-specific proteases were designed in addition to the previously published TEVp.
• We demonstrated higher cleavage activity of the novel split proteases against their respective substrates in comparison to the already existing split TEVp.
• The activity of the new split proteases was regulated by different external stimuli, such as chemical ligands and light.
• This extends the set of functional split proteases from one to five, including TEVpE, PPVp, SuMMVp and SbMVp and provides the building blocks for the design of complex logic operations based on post-translational protein modifications instead of the widely used transcriptional regulation.

Due to a delay inherent to the genetically encoded circuits based on transcription/translation, we decided to engineer a system where all components of the signaling cascade would already be synthetized and present in the cell in their inactive state. Having protein components ready to be activated by post-translational modifications would make the response of our system to activation by external stimuli faster.

Fast signal processing in cells is mainly performed by the phosphorylation cascades. These depend on complex interaction between substrates and protein kinases or phosphatases, resulting in rapid and reversible phosphorylation. However, specificity for kinases may be difficult to engineer in order to maintain the orthogonality of complex signaling pathways. On the other hand, the principles of protease target recognition are well understood. There are in fact several protease-initiated signaling pathways, such as the caspase-induced apoptosis, Notch signaling or the coagulation cascade that all respond fast. Protease-based signaling may therefore represent a versatile system for modular construction of complex logic functions.

To control designed signaling circuits based on proteolysis we needed to have available several specific orthogonal proteases which don’t interfere with native mammalian signaling cascades or with basic cellular processes. Furthermore, to prevent interference between different steps in the signaling pathways the designed proteases should be orthogonal to each other. To allow for external control, these proteases should be expressed in the cell in an inactive state and should switch to the proteolytically active state after the induction with a selected input signal.

Therefore, in order to design protease-based signaling pathways we had to design and test a set of orthogonal split proteases, inducible by a wide range of different input signals.