UC Davis

Voltage-gated sodium (NaV) channels are transmembrane proteins that allow sodium cations to cross the cell membrane in response to changes in membrane potential, playing a crucial role in cellular electrical signaling. Many key physiological processes, such as muscle contraction and nerve impulse transmission, depend on the proper functioning of these channels. Consequently, NaV channels are involved in numerous disease processes and are important pharmaceutical targets for treating conditions like cardiac arrhythmia and pain. Despite their pharmacological significance, NaV channels present key challenges as molecular targets. There are nine human NaV channel subtypes, each with distinct physiological roles, and achieving high subtype selectivity—modulating the function of one particular subtype without affecting the others—has proven very difficult due to the high degree of sequence and structural conservation within the NaV family. As a result, most available NaV channel modulators are non-selective, which limits their therapeutic usefulness due to potential off-target side effects. One notable example involves NaV1.7 and NaV1.8, which have been identified as key players in pain signaling and are considered promising targets for developing novel pain therapeutics with reduced side effects. In recent years, academia and industry have devoted immense efforts to developing selective NaV1.7 and NaV1.8 modulators for non-addictive pain management. This is because the current use of opioids for pain treatment poses significant risks due to the potential for addiction, which has contributed to a severe public health crisis in the USA. Despite the ongoing efforts, our ability to develop potent and selective modulators for these key pharmaceutical targets remains limited.This dissertation focuses on the use of computational modeling and deep learning tools to rationally design novel, potent, and selective modulators of NaV1.7 and NaV1.8. First, it explores peptide toxins from animal venoms, which are naturally occurring modulators of voltage-gated ion channels, some of which inherently exhibit a high degree of subtype selectivity, making them an ideal inspiration for developing novel selective modulators. This work examines available structural and functional data on peptide toxin–ion channel complexes and develops a computational modeling pipeline for accurately predicting peptide toxin–ion channel interactions using protein-protein docking. Second, the dissertation discusses the current state of the art in de novo protein design targeting voltage-gated ion channels, highlighting the availability of abundant structural data with the emergence of deep learning methods for protein modeling and design. This work reviews the development of these methods and discusses design strategies for voltage-gated ion channel targeting and modulation, providing a theoretical framework for their application. Finally, utilizing the previously discussed framework, this dissertation describes the computational design of nanobodies and mini-proteins (a.k.a. binders) targeting an inactivated state of NaV1.7 and NaV1.8 utilizing deep learning methods. By stabilizing an inactivated state of these channels, these novel molecular tools are expected to inhibit pain signal transmission from peripheral sensory neurons to the central nervous system, providing a basis for the development of the next generation of biologic-based selective therapeutics for pain management. This work details the computational design pipeline, in silico results, and preliminary experimental data characterizing the designed proteins and nanobodies. Overall, this dissertation provides a foundation for integrating the structural biology of peptide toxins and NaV channels, computational modeling and design, and deep learning to drive major and innovative advances in NaV channel pharmacology.