Thermodynamic and structural determinants of calcium-independent interactions of calmodulin

Michael Dennis Feldkamp, University of Iowa


Calmodulin (CaM) is an essential protein found in all eukaryotes ranging from vertebrates to unicellular organisms such as Paramecia. CaM is a calcium sensor protein composed of two domains (N and C) responsible for the regulation of numerous calcium-mediated signaling pathways. Four calcium ions bind to CaM, changing its conformation and determining how it recognizes and regulates its cellular targets. Since the discovery of CaM, most studies have focused on the role of its calcium-saturated form. However, an increasing number of target proteins have been discovered that preferentially bind apo (calcium-depleted) CaM. My study focused on understanding how apo CaM recognizes drugs and protein sequences, and how those interactions differ from those of calcium-saturated CaM. I have used spectroscopic methods to explore CaM binding the drug Trifluoperazine (TFP) and the IQ-motif of the type 2 Voltage-Dependent Sodium Channel (Nav1.2IQp). These studies have shown that both TFP and Nav1.2IQp preferentially bind to the "semi-open" conformation of apo CaM. TFP was shown to be an unusual allosteric effector of calcium binding to CaM. Using 15N-HSQC NMR spectroscopy, I determined the stoichiometry of TFP binding to apo Cam to be 2:1 and to (Ca2+)4-CaM to be 4:1 TFP:CaM. That difference in stoichiometry determined whether TFP decreased or increased the affinity of CaM for calcium. Analysis of residue-specific chemical shift differences indicated that TFP binding to apo and (Ca2+)4-CaM perturbed the C-domain more than the N-domain, prompting high-resolution structural studies of the isolated C-domain of CaM. Crystallographic studies of TFP bound to a calcium-saturated C-domain fragment of CaM (CaM76-148) revealed that CaM adopted an "open" tertiary conformation. The unit cell contained two protein and 4 drug molecules. The orientation of TFP revealed that its trifluoromethyl group was found in two alternative positions (one in each protein in the unit cell), and that Met 144 acted as a gatekeeper to select the orientation of TFP.

In contrast to TFP binding to the "open" conformation of calcium-saturated CaM76-148, my NMR studies showed that TFP bound the "semi-open" conformation of apo CaM76-148. TFP interacted with CaM residues near the perimeter of the hydrophobic pocket, but did not contact residues that are solvent-accessible only in the "open" form. Allosteric effects due to TFP binding were observed in the calcium-binding loops of apo CaM76-148. These properties suggest that TFP may antagonize interactions between apo CaM and target proteins such as ion channels that preferentially bind apo CaM. Nav1.2, is responsible for the passage of Na+ ion across cellular membranes. Apo binding of CaM to Nav1.2 poises it for action upon calcium release in the cell. My NMR studies of CaM binding to the Nav1.2 IQ-motif sequence (Nav1.2IQp) showed that the C-domain of apo CaM was necessary and sufficient for binding. My high-resolution structure of the isolated C-domain of CaM bound to Nav1.2IQp revealed that the domain adopted a "semi-open" conformation. At the interface between the IQ-motif and CaM, the highly conserved I and two Y residues of Nav1.2IQp interacted with hydrophobic residues of CaM, while the invariant Q residue interacted with residues in the loop between helices F and G of CaM. This is the first CaM-IQ complex to be determined by NMR; the only other available structure of apo CaM bound to an IQ-motif was determined crystallographically. To accomplish its regulatory roles in response to cellular Ca2+ fluxes, CaM has evolved multiple binding interfaces that are allosterically linked to its Ca2+-ligation state. My studies of CaM binding to TFP and NaV1.2 demonstrate the versatility of CaM functioning as a regulatory protein comprised of domains having separable functions.