Cytoplasmic Autoinhibition in HCN Channels is Regulated by the Transmembrane Region

Fig. 1. HCN4 TM-replacement in HCN2 inhibits activation but preserves cAMP-dependent V1/2 shift. a Schematics: composition of HCN2 and the chimeric Ch4-2 channel produced by HCN4 TM-replacement in HCN2. Rectangles represent transmembrane helices S1 through S6. HCN2 sequence is shown in thick curves and solid rectangles; HCN4 sequence is shown in thin curves and open rectangles. Traces: Representative inward currents of HCN2 and Ch4-2 (upper) and their corresponding R591E mutants lacking cAMP binding (lower). Recordings of HCN2 family channels are from Magee et al. (2015). Channel currents were elicited by step hyperpolarizations to various voltages followed by a tail epoch at − 120 mV for V1/2 determination. In each dataset, stars indicate traces for − 90 mV and − 130 mV. b Conductance–voltage relationships for HCN2 (open squares) compared to HCN2 R591E (filled squares) and for Ch4-2 (open circles) compared to Ch4-2 R591E (filled circles), using tail current data from panel a traces after leak-subtraction and normalization to maximal amplitude determined from Boltzmann equation fit (curves, see “Methods” section). HCN2 and Ch4-2 bind endogenous cAMP, whereas the corresponding R591E…

Fig. 2. HCN4 TM-replacement in truncated HCN2 derivatives does not negatively shift V1/2. a Schematics: composition of ΔCNB channels truncated after the C-linker. Traces: representative inward currents of HCN2 ΔCNB (from Magee et al. 2015) and Ch4-2 ΔCNB, formatted as in Fig. 1a. b Conductance–voltage relationships for HCN2 ΔCNB (point-up triangles) and Ch4-2 ΔCNB (point-down triangles) using data from a and constructed as in Fig. 1b. For comparison with fully autoinhibited channels, the relations for HCN2 R591E (long-dash curve) and Ch4-2 R591E (short-dash curve) are shown repeated from Fig. 1b. Boltzmann fit parameters of the representative curves are as follows: HCN2 ΔCNB, − 79.4 mV, s = 19.2 mV; Ch4-2 ΔCNB, − 83.2 mV, s = 15.4 mV. c Mean V1/2 values for full-length and truncated channels with the HCN2 TM region (squares) or the HCN4 TM region (circles). Error bars show SD, with > 6 recordings for each channel (see Online Resource 1). Selected pairwise comparisons are marked as either statistically significant (*p < 0.05) or not significant (ns). d Comparisons with ΔCNB and R591E channels. The autoinhibitory V1/2 shift of each channel (left panel) was calculated from that channel's…

Fig. 3. Effect of augmented autoinhibition on deactivation kinetics. a Mean t1/2 values with error bars showing SD, with ≥ 6 recordings for each channel. Datasets for HCN2 family channels are from Magee et al. (2015). Datasets are labelled with roman numerals for reference in b. Datasets are plotted for channels lacking cAMP potentiation: HCN2 ΔCNB, light solid squares, label i; HCN2 R591E, dark solid squares, label ii; Ch4-2 ΔCNB, light solid circles, label iii; Ch4-2 R591E, dark solid circles, label iv. b Correlation plot of deactivation t1/2 (+ 20 mV, from a) vs V1/2 (from Fig. 2c). Point symbols and roman numeral labels are as in a. The datapoints for ΔCNB and R591E channels are connected by a straight line representing autoinhibition. Pairwise comparisons were made using log t1/2 values and are marked as either statistically significant (*p < 0.05) or not significant (ns)

Fig. 4. Effect of augmented autoinhibition on deactivation kinetics is partially relieved by cAMP binding or truncation within CNB fold. a Representative deactivation currents at + 20 mV, each normalized to total decay amplitude. Roman numeral labels are as in Fig. 3a. Recordings for HCN2 family channels are from Magee et al. (2015). Within each channel family, the maximally autoinhibited channel (full-length R591E, ii and iv) exhibits the smallest t1/2 values. For these example traces, the t1/2 are as follows: HCN2 ΔCNB (label i), t1/2 = 133 ms; HCN2 R591E (label ii), t1/2 = 119 ms; HCN2 lig., t1/2 = 255 ms; Ch4-2 ΔCNB (label iii), t1/2 = 632 ms; Ch4-2 R591E (label iv), t1/2 = 245 ms; Ch4-2 lig., t1/2 = 340 ms. b Mean t1/2 values with error bars showing SD, with ≥ 6 recordings for each channel; for some channels the bars are drawn wider for clarity. For reference, datasets are repeated from Fig. 3b for corresponding ΔCNB and R591E channels (dashed lines, labels i through iv). In left panel, datasets are plotted for channels with cAMP potentiation: HCN2, open squares; Ch4-2 open circles. In right panel, datasets are plotted for truncated channels: HCN2 ΔαC, half-shaded squares; Ch4-2…

Fig. 5. Effect of augmented autoinhibition on activation kinetics. a Mean τw values with error bars showing SD, with ≥ 6 recordings for each channel; for some channels the bars are drawn wider for clarity. Selected datasets are labelled with roman numerals for reference in b and c. Datasets for HCN2 family channels are from Magee et al. (2015). Datasets are plotted as follows: HCN2 ΔCNB, light solid squares, label i; HCN2 R591E, dark solid squares, label ii; HCN2 ΔαC, half-shaded squares; Ch4-2 ΔCNB, light solid circles, label iii; Ch4-2 R591E, dark solid circles, label iv; Ch4-2 ΔαC, half-shaded circles. b Representative activation currents at − 130 mV, each normalized to total decay amplitude in the 3-s activation epoch. Roman numeral labels refer to channels in a. HCN2 R591E (labelled ii) exhibits Quick-Activation, whereas Ch4-2 R591E (labelled iv) exhibits autoinhibition. Recordings for HCN2 family channels are from Magee et al. (2015). For these example traces, the lag (d) and τw values are as follows: HCN2 ΔCNB (label i), d = 120 ms and τw = 380 ms; HCN2 R591E (label ii), d = 150 ms and τw = 190 ms; Ch4-2 ΔCNB (label iii), d = 60 ms and τw = 680 ms; Ch4-2 R591E (label iv), d = 410…

Fig. 6. Summary of effects of HCN4 TM-replacement on CNB fold-mediated regulatory mechanisms of HCN2. Depictions of TM and C-terminal regions for two adjacent subunits in the tetramer are schematic and not intended to portray structural details. Interdomain interactions are represented by corresponding pairs of symbols joined by bars, with solid-outline bars for stabilizing interactions and dashed-outline bars with starburst flare for destabilizing interactions. The four types of functional interdomain interaction represented are conventional autoinhibition (1, squares), augmented autoinhibition (2, circles), open-state trapping (3, triangles), and Quick-Activation (4, diamonds). Open-state trapping and Quick-Activation affect only the rate-limiting step(s) for deactivation and activation pathways, respectively