Optimising CNT-FET Biosensor Design: Predictive Modelling of Biomolecular Electrostatic Gating and its Application to Beta-Lactamase Detection

Carbon nanotube field effect transistor (CNT-FET) setups hold great promise for constructing next generation miniaturised biosensors whereby a biomolecular event gates conductance. The main issue is predicting how proteins, with their innate mosaic and distinctive electrostatic surfaces, interact with and thus modulate conductance of the CNT-FET. To overcome this barrier, we used advanced sampling molecular dynamics combined with non-canonical amino acid chemistry, to model the protein electrostatic potential imparted on SWCNTs. Here, we focused our efforts using β-lactamase binding protein (BLIP2) as the receptor due to its potential as a biosensor for the most common antibiotic degrading enzymes, the β-lactamases (BLs). Modelling was confirmed experimentally by attaching BLIP2 at single designed residues positions directly to SWCNTs using genetically encoded phenyl azide photochemistry. Our devices were able to successfully detect the two different BLs, TEM-1 and KPC-2, with each BL generating distinct conductance profiles due to differences in their unique surface electrostatic profiles presented close to the SWCNT surface. The changes in conductance closely matched the predicted electrostatic profile sampled by the SWCNTs on BL binding. Thus, our modelling approach combined with new and straight-forward residue-specific receptor attachment techniques, provides a general approach for more effective and optimal CNT-FET biosensor construction.