A synthetic peptide that prevents cAMP regulation in mammalian hyperpolarization-activated cyclic nucleotide-gated (HCN) channels

Abstract

Binding of TRIP8b to the cyclic nucleotide binding domain (CNBD) of mammalian hyperpolarization-activated cyclic nucleotide-gated (HCN) channels prevents their regulation by cAMP. Since TRIP8b is expressed exclusively in the brain, we envisage that it can be used for orthogonal control of HCN channels beyond the central nervous system. To this end, we have identified by rational design a 40-aa long peptide (TRIP8b_nano) that recapitulates affinity and gating effects of TRIP8b in HCN isoforms (hHCN1, mHCN2, rbHCN4) and in the cardiac current I_f in rabbit and mouse sinoatrial node cardiomyocytes. Guided by an NMR-derived structural model that identifies the key molecular interactions between TRIP8b_nano and the HCN CNBD, we further designed a cell-penetrating peptide (TAT-TRIP8b_nano) which successfully prevented β-adrenergic activation of mouse I_f leaving the stimulation of the L-type calcium current (I_CaL) unaffected. TRIP8b_nano represents a novel approach to selectively control HCN activation, which yields the promise of a more targeted pharmacology compared to pore blockers.

Keywords: HCN; cAMP; cardiac pacemaker; molecular biophysics; mouse; structural biology.

Figure 1.. Functional and structural characterization of TRIP8b_nano.

(A) Primary sequence of TRIP8b_nano. Amino acid numbering refers to full length mouse TRIP8b (1a4). (B) Binding of TRIP8b_core and TRIP8b_nano to purified His₆-MBP-CNBD measured by Isothermal titration calorimetry (ITC). Upper panel, heat changes (μcal/sec) during successive injections of 8 μL of the corresponding TRIP8b peptide (200 μM) into the chamber containing His₆-MBP-CNBD (20 μM). Lower panel, binding curve obtained from data displayed in the upper panel. The peaks were integrated, normalized to TRIP8b peptide concentration, and plotted against the molar ratio (TRIP8b peptide/His₆-MBP-CNBD). Solid line represents a nonlinear least-squares fit to a single-site binding model, yielding, in the present examples, a K_D = 1.2 ± 0.1 μM for TRIP8b_core and K_D = 1.4 ± 0.1 μM for TRIP8b_nano. (C) Evidence for TRIP8b_nano folding upon CNBD binding based on the superimposition of the [¹H, ¹⁵N] heteronuclear single quantum coherence (HSQC) NMR spectrum of CNBD-free TRIP8b_nano (black) and CNBD-bound TRIP8b_nano (red). The latter experiment was performed at the molar ratio ([CNBD]/[TRIP8b_nano])=3. The backbone amide (HN) signals of the residues of CNBD-bound TRIP8b_nano are labelled in red. (D) (Top) Ribbon representation of the 10 lowest energy conformers of TRIP8b_nano bound to CNBD used for in silico modelling of CNBD-TRIP8b_nano complex. The unfolded regions at the N- and C-termini of the construct (residues 235–237 and 270–275) are omitted for clarity. (Bottom) Chemical Shift Index (CSI, calculated using TALOS+) plotted as a function of the residue number of TRIP8b_nano bound to CNBD. Positive values represent helical propensity.

Figure 2.. NMR structure of CNBD bound to TRIP8b_nano.

(A) (Top) comparison of secondary structure elements of cAMP-free CNBD (Saponaro et al., 2014), cAMP-bound CNBD (Zagotta et al., 2003) and cAMP-free CNBD bound to TRIP8b_nano (this study). Secondary structure elements are indicated by arrows (β-strands) and cylinders (α-helices) and labeled. The loop between β₆ and β₇ constitutes the Phosphate Binding Cassette (PBC). The elements that fold upon binding of cAMP and TRIP8b_nano are shown in red. (Bottom) Chemical Shift Index (CSI, calculated using TALOS+) plotted as a function of the residue number of CNBD bound to TRIP8b_nano. Positive values represent helical propensity, while negative values represent strands. (B) Ribbon representation of the 10 lowest energy conformers of CNBD bound to TRIP8b_nano used for in silico modeling of CNBD - TRIP8b_nano complex. Secondary structure elements are coloured in light gray and labeled. Loop regions are colored in dark gray. The distal region of the C-helix (residues 657–662), which is unfolded in the free form of the CNBD (Saponaro et al., 2014) and folds upon TRIP8b_nano binding, is coloured in red. The unfolded regions at the N- and C-termini of the construct (residues 521–532 and 663–672 respectively) are omitted for clarity.

Figure 3.. Structural model of CNBD – TRIP8b_nano complex.

(A) Ribbon representation of the complex where CNBD is in gray and TRIP8b_nano is in orange. Helix N (N) and helix C (C) of TRIP8b_nano are labeled. C-helix of CNBD (αC) is labeled, while N-bundle loop is colored in yellow. Positively charged residues of C-helix CNBD (blue) and negatively charged residues of TRIP8b_nano (red) involved in salt bridges are shown as sticks and labeled. N₅₄₇ of the N-bundle loop (yellow) and D₂₅₂ of TRIP8b_nano (orange) are shown as sticks and labeled. (B) Close view of K₆₆₅ and K₆₆₆ of CNBD that interact, respectively, with E₂₆₅ and E₂₆₄ of TRIP8b_nano. C-helix (αC) of CNBD, and Helix C (C) of TRIP8b_nano are labeled. (C) Close view of R₆₅₀ of CNBD that is positioned between E₂₄₀ and E₂₄₁ of TRIP8b_nano. C-helix (αC) of CNBD, and Helix N (N) of TRIP8b_nano are labeled. (D) Close view of N₅₄₇ of N-bundle loop that forms a hydrogen bond (red dashed line) with D₂₅₂ of TRIP8b_nano. Helix E’ (αE') and C-helix (αC) of CNBD, and Helix N (N) of TRIP8b_nano are labeled.

Figure 3—figure supplement 1.. CNBD residues involved in TRIP8b_nano binding.

Combined chemical shift variations of NMR signals between CNBD unbound and TRIP8b_nano-bound state. Combined chemical shift variations are calculated from the experimental ¹H and ¹⁵N chemical shift changes (Δδ(¹H) and Δδ (¹⁵N), respectively) between the corresponding peaks in the two forms, through the following equation (Garrett et al., 1997): ${∆\partial }^{combined}=\sqrt{\frac{{\left(∆\partial \left(1_H\right)\right)}^2+\frac{1}{25}{\left(∆\partial \left(15N\right)\right)}^2}{2}}$. Residues experiencing intermediate exchange regime (whose NMR signal becomes broad beyond detection upon addition of TRIP8b_nano) are shown in grey. The horizontal dotted line indicates the average value plus one standard deviation. Residues above the line were set as ‘active’ in the docking calculation described in the text (see Materials and methods).

Figure 3—figure supplement 2.. Representative families of clusters obtained from the first docking calculation.

The clusters obtained can be grouped in two classes, shown here, on the basis of the position of helix N of TRIP8b_nano (N) relative to the CNBD. In the left representative structure, helix N is oriented in a way that allows E₂₄₁ to interact with R₆₅₀ of C-helix. In the right representative structure, E₂₄₀ - E₂₄₁ interact with N₅₄₇ of CNBD. In all clusters K₆₆₅ or K₆₆₆ of CNBD establish a contact with either E₂₆₄ or E₂₆₅ of TRIP8b_nano. The structural elements of CNBD are labeled: αE’, αA and αC. Helices N and C of TRIP8b_nano are labeled.

Figure 3—figure supplement 3.. Biochemical validation of CNBD – TRIP8b_nano complex.

Dissociation constant (K_D) of the interaction between the indicated CNBD and TRIP8b_nano peptides were measured by means of Isothermal Titration Calorimetry (ITC). CNBD WT - TRIP8b_nano WT (black filled circle)=1.4 ± 0.1 µM; CNBD K₆₆₅E - TRIP8b_nano WT (grey filled circle)=6 ± 0.3 µM; CNBD K₆₆₆E - TRIP8b_nano WT (grey filled circle)=9.9 ± 0.8 µM; CNBD WT - TRIP8b_nano E₆₄K (purple filled circle)=4.8 ± 0.2 µM; CNBD WT - TRIP8b_nano E₆₅K (purple filled circle)=6 ± 0.5 µM; CNBD R₆₅₀E - TRIP8b_nano WT (red filled circle)=9.6 ± 0.7 µM; CNBD WT - TRIP8b_nano E₄₀R (green filled circle)=5 ± 0.3 µM; CNBD WT - TRIP8b_nano E₄₁R (green filled circle)=3.8 ± 0.2 µM; CNBD N₅₄₇D - TRIP8b_nano WT (blue filled circle)=11 ± 0.9 µM; CNBD WT - TRIP8b_nano D₅₂N (orange filled circle)=3.4 ± 0.1 µM. Data are presented as mean ± SEM. Number of experiments (N) ≥ 3. K_D values of all tested combinations are statistically different from the K_D of CNBD WT for TRIP8b_nano WT (*p≤0.05; **p<0.01). Statistical analysis performed with ANOVA, followed by post-hoc Tukey test. Dashed black vertical line indicates the K_D of WT peptides.

Figure 3—figure supplement 4.. Different orientation of R₆₆₂ and K₆₆₅ in the cAMP-bound and TRIP8b_nano-bound conformation of the CNBD.

Residues R₆₆₂ and K₆₆₅ of CNBD C-helix (αC) which interact with cAMP (left, [Lolicato et al., 2011]) and TRIP8b_nano (right, this study) are represented as blue sticks and labeled. TRIP8b_nano residues E₂₅₀ and D₂₅₇ interacting with R₆₆₂ and residue E₂₆₅ interacting with K₆₆₅ are shown in red sticks and labeled.

Figure 3—figure supplement 5.. Mutation N₅₂₀D affects cAMP affinity in full-length HCN2 channel.

Note that in mouse HCN2 N₅₂₀D corresponds to N₅₄₇D in human HCN2 (used in our structural studies). (A) Representative whole-cell HCN2 currents recorded, at the indicated voltages (−70, –80, -90, –130 mV) in HEK 293 T cells transfected with WT and N₅₂₀D mutant, in the absence and presence of 5 µM cAMP in the pipette. (B) Mean activation curves obtained from the tail currents in (A) and other recordings in control solution (filled circle) and in the presence of cAMP (open circles), with HCN2 wt (black) and N₅₄₇D mutant. Data were fitted to a Boltzmann equation (solid and dotted lines) (see Materials and methods). (C) Half activation potential (V_1/2) values derived from Boltzmann equation fitting. HCN2 WT (black filled circle) V_1/2 = -95.9 ± 0.3 mV; HCN2 WT + 5 µM cAMP (black open circle) V_1/2 = 85.2 ± 0.4 mV; HCN2 N₅₂₀D (red filled circle) V_1/2 = -96.8 ± 0.2 mV; HCN2 N₅₂₀D + 5 µM cAMP (red open circle) V_1/2 = -91.2 ± 0.2 mV. Data are presented as mean ± SEM. Number of experiments (N) ≥ 6. Statistical analysis performed with two-way ANOVA, followed by post-hoc Tukey test (p<0.001) showed that wt and mutant channel are identical and that the cAMP-induced shifts are different.

Figure 3—figure supplement 6.. Structural characterization of N₅₄₇D CNBD protein.

Superimposition of the [¹H, ¹⁵N] HSQC spectra of CNBD wt (black) and CNBD N₅₄₇D mutant (red).

Figure 4.. TRIP8b_nano abolishes cAMP effect on HCN channel gating.

(A-C) Representative whole-cell HCN4, HCN2 and HCN1 currents recorded, at the indicated voltages, with control solution or with cAMP (1 μM for HCN4 and 5 μM for HCN2) or with cAMP + 10 µM TRIP8b_nano in the patch pipette (for HCN1, 10 µM TRIP8b_nano only was added). (D-F) Mean activation curves measured from HCN4, HCN2 and HCN1 in control solution (open circles), cAMP (filled circles), cAMP +TRIP8b_nano, or TRIP8b_nano only in the case of HCN1 (filled triangles). Solid and dashed lines indicate Boltzmann fitting to the data (see Materials and methods). (G) Half activation potential (V_1/2) values of HCN4 (blue), HCN2 (black) HCN1 (red) in control solution (open circle), cAMP (filled circle) and cAMP +TRIP8b_nano, or TRIP8b_nano only in the case of HCN1 (filled triangle). HCN4, control = −102.8 ± 0.3 mV; HCN4 +1 µM cAMP = −89.2 ± 0.6 mV; HCN4 +1 µM cAMP +10 µM TRIP8b_nano = −102.1 ± 0.6 mV, HCN4 +1 µM cAMP in the patch pipette +10 µM TRIP8b_nano in the bath solution = −91.7 ± 0.3 mV; HCN2, control = −93.7 ± 0.3 mV; HCN2 +5 µM cAMP = −83.5 ± 0.3 mV; HCN2 +5 µM cAMP +10 µM TRIP8b_nano = −94.5 ± 0.6 mV; HCN1, control = −72.8 ± 0.2 mV; HCN1 +10 µM TRIP8b_nano = −82 ± 0.5 mV. Data are presented as mean ± SEM. Number of cells (N) ≥ 11. There is no significant difference between the controls and the addition of TRIP8b_nano with (HCN4, HCN2) or without (HCN1) cAMP in the pipette. No significant difference was observed following the addition of TRIP8b_nano in the bath. Statistical analysis performed with two-way ANOVA, followed by post-hoc Tukey test (p<0.001).

Figure 4—figure supplement 1.. Comparison of half activation potentials (V_1/2) of HCN1 WT and HCN1 R₅₄₉E mutant (Chen et al., 2001).

(A) Currents were measured at the indicated voltages (- 50,–60, −70,–120 mV), from HEK 293 T cells with control solution or with 15 μM cAMP in the patch pipette. (B) Mean activation curves obtained from tail currents were fitted to a Boltzmann equation (see Materials and methods). (C) Half activation potential (V_1/2) determined from Boltzmann fitting. HCN1 WT (black filled circle)=−72.7 ± 0.2 mV; HCN1 WT + cAMP (black open circle)=−73.1 ± 0.3 mV; HCN1 R₅₄₉E (red filled circle)=−80.4 ± 0.2 mV; HCN1 R₅₄₉E + cAMP (red open circle) = −81 ± 0.3 mV. Note that in human HCN1, used here, R₅₄₉E corresponds to mouse HCN1 R₅₃₈E used in Chen et al. (2001). Data are presented as mean ± SEM. Number of experiments (N) ≥ 8. Wt values are significantly different from mutant. Statistical analysis performed with two-way ANOVA, followed by post-hoc Tukey test (p<0.001).

Figure 4—figure supplement 2.. Voltage-dependency of activation time constant (τ_on) of HCN4 channels in control solution (open circles), 1 µM cAMP (filled circles), 1 µM cAMP + 10 µM TRIP8b_nano (filled triangles).

Data are presented as mean ± SEM. Number of cells (N) = 5.

Figure 5.. Effects of TRIP8b_nano on voltage-dependent activation of I_f and spontaneous rate in rabbit sinoatrial node (SAN) myocytes.

(A) Representative whole-cell I_f currents recorded, at the indicated voltages, in the control solution and in the presence of 10 μM TRIP8b_nano, without (top) and with 1 μM cAMP in the pipette (bottom). (B) Mean I_f activation curves were measured using a two-step protocol (see Materials and methods) in control (filled circles) or in the presence of: 1 µM cAMP (open circles); 10 µM TRIP8b_nano (filled squares); 1 µM cAMP +10 µM TRIP8b_nano (open squares). Ligands were added in the patch pipette. Half activation potential (V_1/2) of I_f activation curves measured in control = −64.1 ± 0.4 mV or in the presence of: 1 µM cAMP = −59.9 ± 0.4 mV; 10 µM TRIP8b_nano = −67.7 ± 0.4 mV; 1 µM cAMP +10 µM TRIP8b_nano = −67.6 ± 0.7 mV. Data are presented as mean ± SEM. Number of cells (N) was ≥15. V_1/2 values are significantly different between each other’s whit the exception of V_1/2 obtained in the presence TRIP8b_nano and cAMP +TRIP8 _bnano. Statistical analysis performed with two-way ANOVA, followed by post-hoc Bonferroni test (*p<0.05) (C) (Left) Representative recordings of single SAN cell spontaneous activity in control and in the presence of 10 µM TRIP8b_nano. (Right) Mean spontaneous rate (Hz) recorded in control solution = 3.65 ± 0.29 Hz and in the presence of 10 µM TRIP8b_nano added to the pipette = 2.69 ± 0.27 Hz. Data are presented as mean ± SEM. Number of cells (N) was ≥7. Statistical analysis performed with t test (*p<0.05).

Figure 6.. Effect of TRIP8b_nano and TAT-TRIP8b_nano on I_f and I_Ca,L in mouse sinoatrial node (SAN) myocytes.

(A) Representative examples of I_f recordings at −95 mV in control conditions (top), in 10 µM TRIP8b_nano dialyzed cell (middle) and in cells perfused with 10 µM TAT-TRIP8b_nano (bottom), before (black trace) and after (red trace) application of ISO 100 nM. The voltage-clamp protocol used for recordings is shown above current traces. (B) Mean normalized I_f current density recorded at −95 mV in absence and in presence of 10 µM TRIP8b_nano in the patch pipette, before (open bars) and after (filled bars) 100 nM ISO perfusion. Data are presented as mean ± SEM. Number of cells (N) ≥ 6. Statistical analysis performed with two-way ANOVA test, followed by Sidak multiple comparisons test (**p<0.01). (C) Mean normalised I_f current density recorded at −95 mV in control solution or in the solution containing 10 µM TAT-TRIP8b_nano, in absence (open bars) and in the presence (filled bars) of 100 nM ISO. Data are presented as mean ± SEM. Number of cells (N) ≥ 8. Statistical analysis performed with two-way ANOVA test, followed by Sidak multiple comparisons test (****p<0.0001). (D) Representative examples of I_Ca,L recordings at −10 mV in control conditions (top), in 10 µM TRIP8b_nano dialyzed cell (middle) and in cells perfused with 10 µM TAT-TRIP8b_nano (bottom), before (black trace) and after (red trace) application of ISO 100 nM. The voltage-clamp protocol used for recordings is shown above current traces. (E) Mean normalized I_Ca,L current density recorded at −10 mV in absence and in presence of 10 µM TRIP8b_nano in the patch pipette, before (open bars) and after (filled bars) 100 nM ISO perfusion. Data are presented as mean ± SEM. Number of cells (N) ≥ 8. Statistical analysis performed with two-way ANOVA test, followed by Sidak multiple comparisons test (***p<0.001; ****p<0.0001). (F) Mean normalized I_Ca,L current density recorded at −10 mV in control solution or in the solution containing 10 µM TAT-TRIP8b_nano, in absence (open bars) and in the presence (filled bars) of 100 nM ISO. Data are presented as mean ± SEM. Number of cells (N) ≥ 7. Statistical analysis performed with two-way ANOVA test, followed by Sidak multiple comparison test (**p<0.01; ***p<0.001).

Figure 6—figure supplement 1.. Effect of TAT- (SCRAMBLED) TRIP8b_nano on I_f in mouse sinoatrial node (SAN) myocytes.

(A) Representative examples of I_f recordings (at −95 mV) in TAT-(SCRAMBLED)TRIP8b_nano (top) or in TAT-TRIP8b_nano (bottom) incubated cells. The voltage-clamp protocol used for recordings is shown above current traces. (B) Mean normalized I_f current density recorded at −95 mV in the presence of 10 µM TAT- (SCRAMBLED) TRIP8b_nano or TAT-TRIP8b_nano, before (open bars) and after (filled bars) 100 nM isoprenaline (ISO) perfusion. Data are presented as mean ± SEM. Number of cells (N) ≥ 11. Statistical analysis performed with two-way ANOVA test, followed by Sidak multiple comparisons test (****p<0.0001).