Direct visualization of interaction between calmodulin and connexin45

Abstract

Calmodulin (CaM) is an intracellular Ca²⁺ transducer involved in numerous activities in a broad Ca²⁺ signaling network. Previous studies have suggested that the Ca²⁺/CaM complex may participate in gap junction regulation via interaction with putative CaM-binding motifs in connexins; however, evidence of direct interactions between CaM and connexins has remained elusive to date due to challenges related to the study of membrane proteins. Here, we report the first direct interaction of CaM with Cx45 (connexin45) of γ-family in living cells under physiological conditions by monitoring bioluminescence resonance energy transfer. The interaction between CaM and Cx45 in cells is strongly dependent on intracellular Ca²⁺ concentration and can be blocked by the CaM inhibitor, N-(6-aminohexyl)-5-chloro-1-naphthalenesulfonamide hydrochloride (W7). We further reveal a CaM-binding site at the cytosolic loop (residues 164-186) of Cx45 using a peptide model. The strong binding (K_d ∼ 5 nM) observed between CaM and Cx45 peptide, monitored by fluorescence-labeled CaM, is found to be Ca²⁺-dependent. Furthermore, high-resolution nuclear magnetic resonance spectroscopy reveals that CaM and Cx45 peptide binding leads to global chemical shift changes of ¹⁵N-labeled CaM, but does not alter the size of the structure. Observations involving both N- and C-domains of CaM to interact with the Cx45 peptide differ from the embraced interaction with Cx50 from another connexin family. Such interaction further increases Ca²⁺ sensitivity of CaM, especially at the N-terminal domain. Results of the present study suggest that both helicity and the interaction mode of the cytosolic loop are likely to contribute to CaM's modulation of connexins.

Keywords: calmodulin; connexin45; gap junctions; intracellular calcium; protein–protein interactions; regulation.

Figure 1. Cx45 membrane topology and the putative CaM-binding site

Cx45 has four transmembrane domains connected by two extracellular loops and one intracellular loop, which is longer than α- and β-class of connexins. The CaM-binding site is located in the second half of the intracellular loop. The predicted CaM-binding motif is conserved in Cx45 from different species.

Figure 2. CaM interacts with Cx45 in live cells in a Ca²⁺-sensitive manner

(A) Interaction between CaM and Cx45 was assessed by BRET in live HEK293 cells. HEK293 cells were cotransfected with a fixed amount of Cx45-Rluc and with an increasing DNA concentration of Venus–CaM (expression level was verified by fluorescence intensity measurement). BRET assays were performed 24 h after transfection. Cells were treated with 5 μM coelenterazine, followed by immediate BRET measurement. (B) Cells were treated with only BRET buffer, or buffer containing either 50 μM W7, or 5 mM Ca²⁺/10 μM ionomycin, or 50 μM BAPTA-AM for 30, 10, or 20 min, respectively, before BRET measurement. The net BRET ratios shown above are the mean ± S.E.

Figure 3. Ca²⁺-dependent specific interaction between CaM and Cx45p_164–186 characterized by NMR

(A) Overlay of HSQC spectra of holo-CaM (red) with the spectra of holo-CaM–Cx45p_164–186 complex (green). Inset, the MALDI-MS spectrum of the free form of CaM (left peak) and the CaM–Cx45p_164–186 complex (right peak) formed in the presence of Ca²⁺. (B) Overlay of HSQC spectrum of apo-CaM with the spectrum of apo-CaM–Cx45p_164–186 mixture. (C) The chemical shift change of A57 during titration of holo-CaM with Cx45p_164–186. The disappearance of the peak (free form) was accompanied by the appearance of the corresponding peak (bound form) at a shifted position.

Figure 4. Determination of the CaM–Cx45p_164–186-binding affinity by steady-state fluorescence studies

(A) The titration curve of D-CaM (0.25 μM) with Cx45p_164–186 in the presence of 5 mM Ca²⁺. The fluorescence spectra of D-CaM with (solid line) or without Cx45p_164–186 (dotted line) are seen in the inset. (B) The fluorescence spectra of D-CaM (0.25 μM) with 0 or 0.37 μM Cx45p_164–186 in the presence of 5 mM EGTA. All experiments were repeated in triplicate.

Figure 5. Ca²⁺-sensitive interaction between CaM and Cx45p_164–186 is confirmed by SPR analysis

The response unit was recorded with Cx45 peptide immobilized on a CM5 sensor chip under Ca²⁺-supplemented conditions using a sensitivity-enhanced Biacore T100 SPR system. The concentrations of CaM were 5–100 μM.

Figure 6. Revealing CaM–peptide interactions with CD spectroscopy

(A) Far UV spectra of Cx45p_164–186 were obtained in the presence of 0% (dotted line), 20% (thin solid line), and 60% (bold solid line) TFE (v/v). Inset: the helical content of Cx45p_164–186 was plotted against TFE concentration. (B) Far UV CD spectra of CaM in the presence of 5 mM EGTA (□) or CaCl₂ (○) and a 1 : 1 CaM/Cx45p_164–186 complex with 5 mM EGTA (▪) or CaCl₂ (●).

Figure 7. Chemical shift perturbation in CaM induced by Cx45p_164–186

The weight-average chemical shift changes (Δδ) were calculated and plotted as a function of amino acid residue number. The upper and lower values are chemical shift perturbations in CaM induced by Cx45p_164–186. The value of Δδ > 0.05 is considered as a significant change.

Figure 8. Cx45p_164–186 binding to CaM without significant hydrodynamic radius change

Hydrodynamics of the holo-CaM–connexin peptide complex were determined by pulsed-field gradient NMR. The holo-CaM (○) NMR signal decay curve did not change upon the addition of Cx45p_164–186 (●), which indicates that the hydrodynamic radius was not altered.

Figure 9. Binding of Cx45p_164–186 expands the Ca²⁺ sensitivity of both domains of CaM

Equilibrium Ca²⁺ titrations of CaM with (solid cycles) or without (open cycles) Cx45p_164–186 were evaluated by monitoring fluorescence emission. (A) The Phe fluorescence emission changes were utilized to reflect changes in the Ca²⁺-binding affinity of the N-domain of CaM. (B) Changes in CaM C-domain Tyr fluorescence emission upon the addition of Ca²⁺ in the presence or absence of Cx45p_164–186 were also monitored. Free Ca²⁺ concentration was determined using the Ca²⁺ indicator Oregon Green.