C terminus L-type Ca2+ channel calmodulin-binding domains are 'auto-agonist' ligands in rabbit ventricular myocytes

Abstract

L-type Ca2+ channel C terminus calmodulin (CaM)-binding domains are molecular determinants for Ca(2+)-CaM-dependent increases in L-type Ca2+ current (ICa), and a CaM-binding IQ domain mimetic peptide (IQmp) increases L-type Ca2+ channel current by promoting a gating mode with prolonged openings (mode 2), suggesting the intriguing possibility that CaM-binding domains are 'auto-agonist' signalling molecules. In order to test the breadth of this concept, we studied the effect of a second C terminus CaM-binding domain (CB) mp (CBmp), in conjunction with IQmp, on single L-type Ca2+ channel currents in excised cell membrane patches from rabbit ventricular myocytes. Here we show that both CBmp and IQmp are agonist ligands that non-additively increase L-type Ca2+ channel opening probability (Po) by inducing mode 2 gating. CBmp and IQmp agonist effects were lost under conditions favouring calcification of CaM (Ca(2+)-CaM, 150 nM free Ca2+ and 10-20 microM CaM), but persisted in the presence of CaM (0-20 microM) under conditions adverse to Ca(2+)-CaM (20 mM BAPTA), indicating that CaM-binding domains increase L-type Ca2+ channel Po by a low Ca(2+)-CaM activity mechanism. Increasing Ca(2+)-CaM in the bath (cytosol) reduced the efficacy of CBmp and IQmp signals with Ba2+ as charge carrier, suggesting that CaM binding motifs target a site outside of the pore region. We measured the combined effects of CBmp and Ca(2+)-CaM-dependent protein kinase II (CaMKII) on L-type Ca2+ channels by using an engineered Ca(2+)-CaM-independent form of CaMKII that remains active under low Ca(2+)-CaM conditions, permissive for CBmp signalling. CBmp and CaMKII increased L-type Ca2+ channel Po in a non-additive manner, suggesting that low and high Ca(2+)-CaM-dependent L-type Ca2+ channel facilitation pathways converge upon a common signalling mechanism.

Figure 1. L-type Ca²⁺ channel C terminus CaM-binding domains increase L-type Ca²⁺ channel opening probability (P_o)

A, schematic representation of the L-type Ca²⁺ channel with the amino acid sequences and relative positions of the IQ and CB CaM-binding motifs. Numerals I– IV indicate homolgous domains in α_1C. B, representative L-type Ca²⁺ channel current recordings under control conditions and after addition of CB or IQ mimetic peptides (mp) to the bath (cytoplasmic) solution. The scale bars indicate 20 ms (horizontal) and 2 pA (vertical). C, mean L-type Ca²⁺ channel P_o under basal conditions (control) or after stepwise increases in CBmp or IQmp (indicated by horizontal bars, with concentrations on the abscissa). The rightmost bar represents L-type Ca²⁺ channel P_o after 10 μm CBmp and 10 μm IQmp added to the bath solution. * P < 0.001 versus control; † P < 0.05 versus CBmp. Differences between 10 and 20 μm CBmp and IQmp were not significant.

Figure 2. The CaM-binding domain mimetic peptide (CBmp) induces mode 2 gating

A and B, representative L-type Ca²⁺ channel recordings (top), ensemble-averaged currents (200–300 tracings, middle), and opening probability (P_o) diary plots (bottom). C and D, histograms plot logarithmically binned L-type Ca²⁺ channel open times (abscissa) against the number of events (ordinate). Open-time distributions were best fitted with the sum of two exponentials, and the long (τ_l) and short (τ_s) time constants for these fits are shown. The percentage of total openings described by τ_l and τ_s are shown in brackets. E and F, quantification of gating modes was performed by plotting P_o (top) against the maximum open time (T_omax, bottom) for each non-blank sweep. A vertical line is placed at the minimum of two Gaussian distributions between long and short L-type Ca²⁺ channel openings, while the horizontal line marks P_o = 2 %. Numerals indicate gating modes 0–2 (Yue et al. 1990). Representative gating mode analysis is shown for E, the inactive control peptide, and F, CBmp (10 μm). Data from panels A, C and E are taken from the same cell membrane excised patch treated with inactive control peptide (10 μm) while panels B, D and F are results from a different excised cell membrane patch treated with CBmp (10 μm).

Figure 3. Elevated Ca²⁺–CaM prevents CBmp actions at L-type Ca²⁺ channels

A-C, L-type Ca²⁺ channel open time histograms with two exponential fits, and long (τ_l) and short (τ_s) time constants describing opening times are shown, as in Fig. 2C and D. While CBmp (10 μm) with CaM (2 μm) increased the percentage of long L-type Ca²⁺ channel openings (B), this effect was prevented by increased CaM (20 μm, C). Data in panels A and B are the same as panels C and D in Fig. 2), and are shown for comparison with panel C. Summary data show the percentage of long L-type Ca²⁺ channel openings (D) and τ_l and τ_s (E) with control peptide + CaM (2 μm) (Control, open bar), CBmp + CaM (2 μm) (hatched bar), and CBmp + CaM (20 μm) (black bar), where the number of L-type Ca²⁺ channels studied is indicated in parentheses. F, summary data show the distribution of gating modes from excised cell membrane patches (n = 3 for each group) in response to control + CaM (2 μm) (open bars), CBmp + CaM (2 μm) (hatched bars) and CBmp + CaM (20 μm) (black bars). The numbers of sweeps (ordinate) with L-type Ca²⁺ channel openings representative of each gating mode are shown. * P = 0.01 for mode 2 sweeps after CBmp + CaM (2 μm) versus control + CaM (2 μm) and CBmp + CaM (20 μm).

Figure 4. CBmp increases L-type Ca²⁺ channel P_o only under low Ca²⁺–CaM activity conditions

A, mean L-type Ca²⁺ channel P_o in the presence of a control peptide (Control) + CaM, CBmp + CaM, or CBmp + AC3-I + CaM. CBmp, Control, and AC3-I were all used at 10 μm, and CaM concentrations are indicated (abscissa) in B. The bath free (cytosolic) Ca²⁺ = 150 nm before CaM addition in A, while the Ca²⁺ was buffered with BAPTA (20 mm) in B. The number of excised cell membrane patches studied in each group is indicated by numerals (abscissa). * P < 0.001 versus control (A) and * P = 0.02 versus control (B).

Figure 5. CBmp and CaMKII increase L-type Ca²⁺ channel P_o by distinct low- and high-Ca²⁺–CaM activity mechanisms

Mean L-type Ca²⁺ channel P_o increases significantly from control (0 μm CaM) after CBmp (10 μm, P = 0.008), but this increase is reversed in step-wise fashion by increasing concentrations of CaM up to 20 μm, in the continued presence of CBmp. This trend reverses at CaM = 40 μm (but does not reach significance versus control) in the presence of CBmp. Mean L-type Ca²⁺ channel P_o significantly increases (P = 0.002) with increasing CaM (from 10–20 μm) in the absence of CBmp, and this increase is occluded by the CaMKII inhibitory peptide AC3-I (10 μm), indicating endogenous CaMKII is localized near the L-type Ca²⁺ channel in excised cell membrane patches (Dzhura et al. 2002). Labelled horizontal bars indicate the presence of CBmp and AC3-I. CaM concentrations and the number of excised cell membrane patches studied are indicated (abscissa). * P < 0.05 compared to the 0 μm CaM-0 μm CBmp condition, and ** P < 0.05 compared to the 2 μm CaM-0 μm CBmp condition.

Figure 6. A model for increasing L-type Ca²⁺ channel P_o in low- and high-Ca²⁺–CaM conditions

Under basal conditions (centre panel) CaMKII is largely inactive (dark folded, membrane-associated bar) and CaM is bound to the L-type Ca²⁺ channel C terminus by a low-Ca²⁺ activity mechanism (Pitt et al. 2001) that does not permit Ca²⁺–CaM-dependent binding with IQ or CB motifs, or significant interaction of these motifs with a facilitation site (fs). Increased Ca²⁺–CaM (left panel) results in Ca²⁺–CaM binding to CaMKII, activation and autophosphorylation of endogenous CaMKII (dark, extended bar), which binds the L-type Ca²⁺ channel C terminus (Hudmon & Schulman, 2002). CaMKII binding to the L-type Ca²⁺ channel C terminus disrupts CaM binding to the C terminus CaM-binding domains (Hudmon et al. 2002) and improves the interaction of IQ and CB with the facilitation site (fs). CaMKII activation and C terminus binding results in increased L-type Ca²⁺ channel P_o (Dzhura et al. 2000; Wu et al. 2001a; Dzhura et al. 2002). During low-Ca²⁺–CaM conditions (right panel), C terminus CaM-binding domain mimetic peptides (CBmp and IQmp) are ‘enhanced’ auto-agonist ligands, by virtue of reduced steric constraint compared to the endogenous counterparts, for interacting with the facilitation site (fs) to induce mode 2 gating. CBmp (Figs 3 and 4) and IQmp (Wu et al. 2001a) cannot interact with the facilitation site while bound to Ca²⁺–CaM, and so are ineffective during increased Ca²⁺–CaM activity conditions.