Page 34 - Cardiac Electrophysiology | A Modeling and Imaging Approach
P. 34

P. 34
        P . 34
               When paced at slow rate (CL=4000ms) with I  block, Pcell develops EADs. Interestingly, the
                                                                 Kr
        simulations (see reference ) show that the ionic mechanism of EAD formation is different in Pcell
                                     93
        from that in Vcell. As shown earlier, recovery and reactivation of I   CaL  underlies EAD formation in
        Vcell. In Pcell EAD generation is mostly attributable to reactivation of I   NaL2  (type 2 component
        of I NaL ).



           2.8. Structure – Function Relationship of an Ion Channel During Gating


                                    Reductionist Approach at the Molecular Scale


               Markov models of ion channels represent kinetic states of the channel and reproduce
        the kinetics of channel opening and closing. However, the structural molecular conformations
        represented by the Markov model states are not identified. A Markov scheme is not a unique
        model of the channel function and different schemes that fit the same experimental data can

        be interpreted differently in terms of underlying gating mechanisms. To overcome this
        limitation, and relate explicitly molecular structure and channel function, a multiscale
        computational approach that combines molecular dynamics and continuum electrostatics was

        introduced and applied to link KCNQ1 and I  movement during gating to the ionic current they
                                                        Ks
        generate and to the cardiac action potential.  A molecular model of KCNQ1 is constructed by
                                                         101
        aligning the KCNQ1 sequence with K 1.2 and using the known crystal structure of K 1.2 (in the open
                                                 v                                                 v
        state)  as a template (Figure 2.20A). This homology-based model is then refined by minimizing
              102
        an objective function that accounts for this alignment, as well as van der Waals forces, dihedral

        angles, bond lengths and torsions, and Coulombic interactions, using the computer software
        Modeller . As shown in Figure 2.20A, there is good agreement between KCNQ1 and K 1.2 with
                  103
                                                                                                       v
        respect to position of charged residues. Further refinement of the model is done with molecular
        dynamics simulations (using the NAMD simulation package)  to include lipid and water mole-
                                                                           104
        cules surrounding the channel protein (Figure 2.20B, C).


               The voltage-sensing region of each of the four KCNQ1 channel subunits includes the
        transmembrane segments S1-S4 (Figure 2.20B-D). S4 contains a series of positively charged

        residues (R1-R6) that interact with negatively charged residues on S2 and S3 (see Figure 2.20F,
        top). During channel gating, S4 moves up and down through the membrane approximately 12
        Angstrom   105,106  (up towards the extracellular space during activation and down toward the

        intracellular space during channel closing) and rotates about its axis (a simplifying assumption
        is made that S4 remains rigid and that other helices do not move). The simulated movement
        (translation and rotation) of S4 is shown in Figure 2.20D. During this movement, positive charges
        on S4 interact with different negative charges on S2 and S3, generating an electrostatic energy at
        each protein conformation that can be computed using the Poisson-Boltzmann equation.               107,108,109

        As an approximation, other forces (van der Waals force, bending and twisting of residues into
        favorable and unfavorable conformations) are not considered.
   29   30   31   32   33   34   35   36   37   38   39