A multiscale approach for bridging the gap between potency, efficacy, and safety of small molecules directed at membrane proteins

Figures

• 

  Figure 1. Structures of three activators known to target the transmembrane domain of the calcium pump.

• 

  Figure 2. Bilayer-crossing free energy profiles of SERCA2a activators. Profiles are represented by the potential of the mean force (PMF) of SERCA2a activators CDN1163, CP-153526, and Ro 41-0960 along the normal z-axis of a POPC membrane; the distance along the z-axis was used here as a reaction coordinate (ξ). The chemical structures in the upper right indicate the location of titratable groups and their corresponding pKa values (red circles); based on these pKa values, Bilayer-crossing free energy profiles were calculated for either neutral or ionized states of the small molecules. The gray area illustrates only one leaflet of the lipid bilayer.

• 

  Figure 3. Single-molecule atomistic simulations of ligand–membrane interactions. The plots represent the position of a single molecule of (A) CDN1163, (B,C) CP-154526, and (D,E) Ro 41-0960 bound to a solvated POPC membrane. We show the time-dependent position of the molecule along the z-axis of the lipid–water system (left), the relative density (δ) profile of the molecules in the lipid bilayer (center), and the orientation of the molecules in the lipid bilayer extracted at the end of the trajectories (right). In all plots, the shaded area represents the location of the lipid bilayer. The small molecules are shown as ball and stick representation; lipid headgroups and tails are shown as light gray spheres and sticks, respectively; water molecules are shown as white and blue sticks.

• 

  Figure 4. Ensemble-based atomistic simulations of ligand–membrane interactions of SERCA2a activators. The left panels show the time-resolved behavior of SERCA2a activators in a lipid–water environment and the relative density profiles of SERCA2a activators in a lipid–water system obtained from the ensemble-based MD simulations. The figure shows the position of the molecules relative to the z-axis of the system (left) and the relative density (δ) profile (right) calculated for (A) CDN1163, (B,C) CP-154526, and (D,E) Ro 41-0960. The gray horizontal lines (left) and the shaded area (right) represent the boundaries of the lipid bilayer. We show time-dependent configurations of each system containing a solvated membrane and ten copies of each SERCA2a actiator. The simulations capture spontaneous insertion of small molecules into the membrane as well as ligand–ligand interactions during the 300 ns of simulation time. CDN1163 (blue), CP-154526 (green), and Ro 41-0960 (red) are shown as spheres; lipids and water molecules are shown in gray and blue, respectively.

• 

  Figure 5. Expression of major Ca2+ handling proteins in human iPSC-derived cardiomyocytes by western blot analysis. (A) Western blot analysis of human cardiomyocytes showed that these cells express all major proteins involved in Ca2+ transport and regulation of the cardiac calcium pump SERCA2a. We analyzed the expression of the following proteins: RyR, ryanodine receptor; SERCA2a, the cardiac isoform of sarcoplasmic reticulum Ca2+-ATPase; Cav1.2, voltage-dependent L-type Ca2+ channel, α-1C subunit; NCX1, Na+–Ca2+ exchanger; PLN, phospholamban. (B) Detection of phosphorylated phospholamban (PLN) at positions Ser16 (PLN-pSer16) and Thr17 (PLN-pThr17) in response to pharmacological β2-adrenergic stimulation with isoproterenol (ISO, 1 µM); untreated cells were included in the analysis (basal, negative control). In all cases, we used a pig heart membrane preparation as a control (‘CSR’ lane) for the expression of Ca2+ handling proteins tested in this study.

• 

  Figure 6. Functional characterization of SERCA2a activators in human iPSC-derived cardiomyocytes. Concentration-dependent effects of SERCA2a activators CP-154526, CDN1163, and Ro 41–0960 on (A) the Ca2+ transient amplitude, (B) the Ca2+ transient duration at 50% recovery (CaTD50), and (C) beat rate in normal human iPSC-derived cardiomyocytes. We compared these effects to those induced by isoproterenol (ISO), which has both positive inotropic and chronotropic effects on the cardiomyocyte. Results are expressed as percent change from baseline values in Hank’s balanced salt solution and reported as mean ± SEM (n = 6). Groups were compared using a one-way analysis of variance (ANOVA) with the Dunnett test.

• 

  Figure 7. Conduction velocity and APD90 after treatment of human iPSC-derived cardiomyocytes with the SERCA activator CDN1163. (A) Representative impulse propagation during the spontaneous activity of untreated monolayer syncytium formed of human iPSC-derived cardiac cells. Impulse propagation is represented by the activation maps of action potential propagation at different activation times, with the arrows showing the direction of the propagation wave across the monolayer syncytium. (B) Representative trace of the spontaneous action potential (AP) obtained by optical mapping of a syncytium formed by human iPSC-derived cardiomyocytes. The inset represents single-pixel signals of spontaneous optical AP recordings. (C) Conduction velocity (CV) calculated from each activation map in the presence and absence (untreated control) of CDN1163. We found no significant changes in CV between the control and treatment groups. (D) Quantification of APD90 from optical mapping readings shown in panel (B). We found that treatment of human iPSC-derived cardiomyocytes with CDN1163 does not have significant effects on the APD90 compared to the untreated control. Data are reported as mean ± SEM (n = 6).…