During the last few years, we have developed structural models of the transmembrane and extracellular segments of Shaker, KvAP, hERG, NaChBac, and Ca2+ channels in resting, open, and numerous transition conformations. Molecular dynamic simulations of these channels embedded in a lipid bilayer were performed to evaluate and refine the models. The models were constrained by recently obtained experimental data;e.g., the crystal structure of the Kv1.2 channel, electron paramagnetic resonance (EPR) studies of KvAP channels, thermodynamic cyclic mutagenesis studies of the binding of BeKM1 toxin from scorpions to the hERG channel, and cysteine scanning mutagenesis (SCAM) studies of Ca2+ channel pores. We have demonstrated that the helical screw model for the voltage-dependent movement of the S4 voltage-sensor segment that we proposed first in 1986, is consistent with virtually all experimental results and energetic criteria, including analyses using molecular dynamic simulations. Recent experimental and computational studies from other groups have provided additional support for our models. The NaChBac channel is a prokaryotic Na+ channel that has similarities to K+, Ca2+, and Na+ channels. We were the first group to identify this sequence in the prokaryotic sequence data base. Since then, it has been expressed and its properties have been studied expensively. Efforts are underway to solve its crystal structure. Our NaChBac was develop using the crystal structure of the Kv1.2 channel as an initial template. The resulting NaChBac model has several unique features involving the ion selective region formed by the P segments, the activation gate formed by the S6 segment, and the interaction between the voltage-sensing (S1-S4) and pore-forming (S5-P-S6) domains. We are now using the NaChBac model as a stepping stone to model more complex eukaryotic Ca2+ and Na+ channels. So far we have modeled the transmembrane regions of human and fungal Ca2+ channels. We are collaborating with Steffen Herrings group to test experimentally these models. Specifically, we are using the models to analyze the molecular pharmacology of Ca2+ channel blockers (important in treating hypertension and heart disease in humans) and to better understand how mutations associated with genetic diseases alter the gating properties of Ca2+ channels. We have started new projects to develop structural and functional models of HCN and TREK channels. A crystal structure of the cyclic nucleotide-binding domain of the HCN channel is being used to model the cytoplasmic domain. HCN is a member of a channel family that includes the hERG channel, which we have modeled previously. TREK channels are mechanosensitive (stretch-activated) K+ channels. We are collaborating with Sergei Sukharev's group at the University model in developing and testing the models. Our previous collaborations with Sukharev's group on other mechanosensitive channels were very productive.
