Programmable Biochemical Logic Computation Through Repurposing Natural bZIP Protein Interaction Networks

Unengineered bZIP interaction clusters from Nematostella vectensis, Ciona intestinalis, and Homo sapiens compute up to 15 of 16 two-input Boolean functions, including XOR and XNOR.

• Why it matters: Designing or tuning protein-protein affinities limits practical biochemical logic implementation, so it remains unclear whether evolved interaction networks can be repurposed without changing binding strengths.
• What they did: The authors repurposed experimentally characterized bZIP interactomes while keeping measured affinities fixed, then used mass-action equilibrium modeling with genetic optimization to tune accessory monomer concentrations and output dimers.
• The result: Optimized concentration regimes preserve logic-state separation under substantial variability and enable up to 13 functions in a 17-gate three-input benchmark, with a scaling framework supporting cell-free translation.

AI summary of the abstract · may contain errors · not medical advice · read the paper.