Candidate Background: In graduate school at the University of Virginia, I built on my undergraduate spectroscopy education by using spectroscopic tools to investigate membrane protein flexibility. As a Postdoctoral Fellow at Vanderbilt, I transitioned to membrane protein structural biology involved in human disease, specifically KCNQ and KCNE family-associated channelopathies. As a Postdoctoral Fellow, I have been involved in several projects concerning the structural underpinnings of disease mechanisms, most recently proposing a mechanism for diminished apical chloride secretion through an estrogen-induced loss of KCNQ1- KCNE3 channel conduction. Research Strategy: The human voltage-gated sodium channel Nav1.5 (encoded by SCN5A) is implicated in several diseases of the heart including dilated cardiomyopathy, cardiac conduction disease, sick sinus syndrome, type 3 longQT syndrome, and Brugada syndrome. Several algorithms accurately predict SCN5A variants that are ultimately harmful (SIFT, PolyPhen-2, PredSNP, etc.). However, there is a significant gap in the negative predictive ability of these methods, i.e. the ability to accurately classify a variant as benign. The approach I am proposing is to tackle this problem on two fronts: 1) incorporating channel-specific, quantitative information-rich data into predictive model construction?the objective being to predict channel function, instead of disease- inducing propensity?and 2) including a set of point mutation variants enriched in WT/neutral phenotypes to improve discrimination power during model training and evaluation. This project aims to ultimately predict Nav1.5 channel phenotypes for all possible amino-acid changing single nucleotide polymorphisms (nsSNP) by balancing high-throughput computation and rigorous experimental validation with model systems: predicting the nearly 15,000 possible SCN5A missense nsSNPs is currently only feasible in silico, i.e. leveraging calculable channel-specific protein sequence and structure-based features. The availability of a high-throughput electrophysiology instrument allows for an unprecedented amassing of ion channel functional output from heterologously expressed Nav1.5; the evaluation of SCN5A variants impact on action potential in the more native like human induced pluripotent stem cell cardiomyocytes is possible in low-throughput. During the mentored (K99) phase of this award, I will generate (mis)trafficking and electrophysiology current output data from missense nsSNPs of SCN5A, focusing on the Voltage-Sensing Module (VSM) of domain IV (Aim 1) and train an SCN5A VSM IV-specific phenotype prediction model using trafficking and electrophysiology data from Aim 1 and the literature (Aim 2). As an independent investigator, I will determine structure and flexibility-induced changes from selected variants using a combination of Rosetta modeling and nuclear magnetic resonance (NMR) to refine the predictive model (Aim 3). Career Development and Training: My training proposal is ambitious covering several disciplines, some of which will be new to me. The skills I will acquire are developing computational predictive models of ion channel phenotypes, trafficking/expression quantitation through Fluorescence Activated Cell Sorting (FACS), CRISPR/Cas9 gene manipulation, and hiPSC cardiomyocyte production. Though there are many activities planned, I will be trained directly in the laboratories of prominent scientists in their respective fields: Charles Sanders, Jens Meiler, and Dan Roden.
Missense mutations in the sodium channel Nav1.5, encoded by SCN5A, are implicated in several diseases of the heart including dilated cardiomyopathy, cardiac conduction disease, sick sinus syndrome, type 3 longQT syndrome, and Brugada syndrome. While multiple algorithms are available to predict whether SCN5A variants are deleterious (SIFT, PolyPhen-2, PredSNP, etc.), there is a significant gap in the negative predictive ability of these methods, i.e. the number of neutral variants classified as disease-causing. Closing this gap would enable the consistent use of such methods in clinical diagnosis and thus enable a more prevalent use of genomic profiling in patients; closing this gap is the goal of this proposal.