(a) A space-filled model of the CaM-CaMBD2-b complex (3SJQ), depicting the location of PHU (yellow) between the interface of CaM (salmon) and CaMBD2-b (marine) at the CaM N-lobe. Only half of the 2×2 complex is shown for clarity. (b) Difference Fourier electron density map constructed using mFo–DFc coefficients calculated in PHENIX prior to modeling PHU and side chain rearrangements that occur in the binding site. The map is contoured at 2.5σ and displayed within a 6 Å sphere of PHU. The map is overlaid with the final coordinates for PHU and its surrounding amino acid residues in the CaM-CaMBD2-a complex. Marine are amino acid residues from CaMBD2-a and salmon are amino acid residues from CaM. Coordinates of the same amino acid residues (gray) from the original CaM-CaMBD2-a complex (without PHU, 1G4Y) are provided for comparison. (c) A space-filled model of PHU at the CaM-CaMBD interface. Marine represents CaMBD and salmon represents CaM. (d) Comparison of the chemical structure of PHU with that of 1-EBIO and NS309. Both 1-EBIO and NS309 are known SK/IK channel modulators.

(a) Raw current traces from an inside-out patch with SK2-a channels expressed. The SK2 channels are activated by 200 nM Ca²⁺, with subsequent potentiation by PHU at the concentrations indicated (all in the presence of 200 nM Ca²⁺). A voltage ramp, from −100 mV to +100 mV, was applied. (b) Dose-response curves for PHU (n = 4) and 1-EBIO (n = 5) for their potentiation of the SK2 channel activities (all in the presence of 200 nM Ca²⁺). The current amplitudes were measured at −90 mV and normalized to the maximal current obtained either with PHU or 1-EBIO. (c) EC50s of potentiation of the SK2 channel activities by PHU (n = 4), 1-EBIO (n = 5), DCEBIO (n = 8), CyPPA (n = 6) and NS309 (n = 3). All modulators were in a 200 nM Ca²⁺ solution. Note that the y-axis is in the log scale. (d) SK2-a (n = 3) and SK2-b (n = 6) have the same responses to application of NS309. Because SK2-b is less sensitive to Ca²⁺ for its activation, 500 nM Ca²⁺ was used instead of 200 nM Ca²⁺ for SK2-a channels. All data are mean ± s.e.m.

(a) Difference Fourier electron density map constructed using mFo–DFc coefficients calculated in PHENIX prior to modeling 1-EBIO and side chain rearrangements that occur in the binding site. The map is contoured at 2.2s and displayed within a 6 Å sphere of 1-EBIO. The map is overlaid with the final coordinates for 1-EBIO and its surrounding amino acid residues in the CaM-CaMBD2-a complex. Marine are amino acid residues from CaMBD2-a and salmon are amino acid residues from CaM. Coordinates of the same amino acid residues (gray) from the original CaM-CaMBD2-a complex (without 1-EBIO, 1G4Y) are provided for comparison. (b) Comparison of key amino acid residues which form the binding pocket for PHU (orange for CaMBD residues and purple for CaM residues) and that for 1-EBIO (marine for CaMBD residues and salmon for CaM residues). (c) Formation of the 1-EBIO binding pocket at the CaM-CaMBD interface. 1-EBIO disrupts the interaction between M71 of CaM (salmon) and A477/L480 of CaMBD (marine). (d) Overlay of the coordinates for 1-EBIO (yellow) and PHU (Cyan) obtained from their respective protein crystal structures.

(a) Conformations of PHU and 1-EBIO determined from the crystal structures. They serve as reference conformations. (b) The predicted conformation of DCEBIO in the binding pocket (left), which is superimposed with the conformations of PHU or 1-EBIO (right). (c) The predicted conformation of NS309 in the binding pocket (left), which is superimposed with the conformations of PHU or 1-EBIO (right). (d) The predicted conformation of CyPPA in the binding pocket (left), which is superimposed with the conformations of PHU or 1-EBIO (right). (e) Correlation of the predicted binding free energies of the five ligands using the S2 scoring function and the EC50s of these compounds in potentiation of SK2 channels. The EC50 for NS309 deviates from the prediction.

(a) Sequence alignment of the fragments (equivalent of E469 to Q487 of SK2-a) of CaMBD at the CaM N-lobe from SK1, SK2, SK3 and IK channels. A477, L480 and V481 are implicated in formation of the binding pocket for the channel modulators from the structure data. Used in the alignment are the human SK1, SK3 and IK channel sequences and the rat SK2 channel sequence. (b) A space-filled model of the A477V/L480M mutation (green) with 1-EBIO in the binding pocket. Changes are made based on the structure data using Pymol. Note that T479 points away from 1-EBIO. (c) Dose-response curves for potentiation by 1-EBIO of the SK2 channel activities from WT (n = 5), T479S (n = 4) and A477V/L480M (n = 4). 200 nM Ca²⁺ was present in all 1-EBIO solutions. (d) A477V/L480M mutation significantly increases the potency of 1-EBIO (10-fold). All data are mean ± s.e.m.

(a) A space-filled model of the A477L mutation (green) with 1-EBIO in the binding pocket. The Ala residue at position 477 is changed to Leu based on the structure data using Pymol. Introduction of such a bulky amino acid residue may interfere with the interaction of 1-EBIO and its binding pocket. (b) A space-filled model of the A477I mutation (green) with NS309 in the binding pocket. NS309 is docked to the binding pocket using MD modeling from the structure of the 1-EBIO-bound CaM-CaMBD2-a complex. The A477I change is then made using Pymol. (c) Responses of the A477I mutant, from an inside-out patch, to sequential applications of solutions containing (1) 0 Ca²⁺, (2) 200 nM Ca²⁺, (3) 3 μM NS309 with 200 nM Ca²⁺, and (4) 0 Ca²⁺ (as indicated). The current amplitudes were measured at −90 mV. (d) Dose-response curves for potentiation by NS309 of the SK2 channel activities from WT (n = 3), A477V/L480M (n = 4) and A477I (n = 3). 200 nM Ca²⁺ was present in all NS309 solutions. On average, the A477V/L480M mutation shifts the NS309 dose-response curve to the left by 3.4-fold, while the A477I mutation results in a 5-fold shift of the NS309 dose-response curve to the right. All data are mean ± s.e.m.

(a) A space-filled model of CyPPA docked into the binding pocket. Molecular docking and MD simulation were performed to dock CyPPA into the PHU binding pocket based on the structure of PHU in the CaM-CaMBD2-a complex. (b) A space-filled model of CyPPA docked into the binding pocket with the A477V/L480M mutation (green). Molecular docking and MD simulations were performed to introduce the A477V/L480M mutation and then docking of CyPPA. Note that the A477V/L480M mutation is predicted to force CyPPA to adopt a different conformation in the pocket. (c) The A477V/L480M mutation on the effect of CyPPA. Responses of the A477V/L480M mutant, from an inside-out patch, to sequential application of solutions containing (1) 0 Ca²⁺, (2) 200 nM Ca²⁺, (3) 100 μM CyPPA with 200 nM Ca²⁺, (4) 10 μM NS309 with 200 nM Ca²⁺ and (5) 0 Ca²⁺. The current amplitudes were measured at −90 mV. Note that CyPPA, at 100 μM, barely potentiates the mutant channel activity. (d) Responses of SK2 channels, WT (n = 6), V418T (n = 3) and A477V/L480M (n = 5), to increasing concentrations of CyPPA. 200 nM Ca²⁺ was present in all CyPPA solutions. Note that there is little response from the A477V/L480M mutant at the CyPPA concentration range tested. All data are mean ± s.e.m.