Structural mechanism for the regulation of HCN ion channels by the accessory protein TRIP8b

Abstract

Hyperpolarization-activated cyclic nucleotide-gated (HCN) ion channels underlie the cationic Ih current present in many neurons. The direct binding of cyclic AMP to HCN channels increases the rate and extent of channel opening and results in a depolarizing shift in the voltage dependence of activation. TRIP8b is an accessory protein that regulates the cell surface expression and dendritic localization of HCN channels and reduces the cyclic nucleotide dependence of these channels. Here, we use electron paramagnetic resonance (EPR) to show that TRIP8b binds to the apo state of the cyclic nucleotide binding domain (CNBD) of HCN2 channels without changing the overall domain structure. With EPR and nuclear magnetic resonance, we locate TRIP8b relative to the HCN channel and identify the binding interface on the CNBD. These data provide a structural framework for understanding how TRIP8b regulates the cyclic nucleotide dependence of HCN channels.

Figure 1. TRIP8b_core binds to and regulated HCN2 channels

A) Cartoon showing the interactions between HCN channels and full length TRIP8b. The intracellular C-linker and the CNBD of HCN2 are labeled. Only two of the four subunits are shown. The variable, core, and TPR domains of TRIP8b are highlighted (see also Figure S1). B) Representative current traces elicited by hyperpolarizing voltage steps from inside-out patches of oocytes expressing HCN channels with no ligand (left), in the presence of 1 μM cAMP (center), and in the presence of both 1 μM cAMP and 10 μM TRIP8b_core (right). (c) Normalized conductance-voltage relationships of HCN alone (cyan), in the presence of 1μM cAMP (red), and in the presence of both 1 μM cAMP and 10 μM TRIP8b_core (black). Data are represented as mean ± SEM.

Figure 2. TRIP8b_core and TRIP8b(1a-4) alter the separation of spin labels attached to HCN2-CNBDxt at V537C and R635C

A) Structure of the HCN2 C-terminal fragment (HCN2-CNBDxt) used in this study (based on accession number 3ETQ) consisting of the cyclic nucleotide binding domain (cyan) and a fraction of the C-linker (gray). B) HCN2-CNBDxt with predicted spin label rotamers highlighted in color. The colored spheres indicated the predicted positions of the midpoints of the N-O bond in MTSL. C) DEER time traces of HCN2-CNBDxt V537C,R635C in the absence (cyan) and presence of TRIP8b(1a-4) (black). D) DEER time traces of HCN2-CNBDxt V537C,R635C in the absence (cyan) and presence of TRIP8b_core (black). E) Distance distributions of HCN2-CNBDxt V537C,R635C calculated from the time traces in (A). F) Distance distributions of HCN2-CNBDxt V537C,R635C calculated from the time traces in (d) (see also Figure S2).

Figure 3. TRIP8b_core binding only alters spin label position at distal C-helix

DEER distance distributions of HCN2-CNBDxt residues in the presence of TRIP8b_core (black traces), cAMP (red traces) and in the apo state (cyan traces). DEER distributions for A) HCN2-CNBDxt S563C,R635C, B) HCN2-CNBDxt V537C,A624C, C) HCN2-CNBDxt S563C,A624C, and D) HCN2-CNBDxt V537C,K570C. The HCN2-CNBDxt structure above each distribution highlights the predicted spin label rotamers (see also Figure S3).

Figure 4. TRIP8b_core binding decreased the mobility of a spin label at the distal end of the C-helix

CW EPR spectra for spin labels attached to indicated residues on HCN in the absence (cyan) and presence of TRIP8b_core (black). Significant line broadening of the R635C spectrum is seen upon addition of TRIP8b_core.

Figure 5. Inter-protein DEER measurements between HCN and TRIP8b_core

A) Structure of HCN2-CNBDxt with the positions of predicted spin label rotamers highlighted in color. B) Distance distributions obtained from DEER measurements between the indicated residues of singly labeled 150 μM HCN and 40 μM TRIP8b_core. No oscillations were observed in the time trace of A261C of TRIP8b_core and A624C of HCN2 indicating the absence of binding (see also Figure S4).

Figure 6. Probability density distributions for the locus of TRIB8b_core residues 248 and 261 relative to HCN2-CNBDxt

Two views of a structural model of the apo state of HCN2-CNBDxt with calculated trilateration isosurfaces for A248C of TRIP8b_core (yellow) and A261C of TRIP8b_core (green), contoured at a probability density of 1.1×10^-4 Å^-3. Residues 624 and 635 on the C-helix are highlighted in magenta (see also Figure S5).

Figure 7. NMR residues on the CNBD perturbed by TRIP8b binding

A) ¹H, ¹⁵N HSQC spectrum of 100 μM labeled HCN4-CNBDxt either unbound (black) or in the presence of 25 μM TRIP8b_core (red). Selected residues from the C-terminus (Ile720, top right), the B4-B5 loop (Lys645, Glu649, middle right), and both regions (Lys648, Ile714) are highlighted. Unlabeled resonances in the highlighted portions of the spectra were not significantly affected by the addition of TRIP8b_core. B) Change in peak intensities for residues in HCN4-CNBDxt upon addition of TRIP8b, measured as a ratio of the intensities of peaks in the bound state over their corresponding intensities in the free state. Unassigned residues are omitted. Peaks with z-scores less than −1 are highlighted in red (see also Figure S6).

Figure 8. NMR and DEER localization of a TRIB8b binding site along the C-helix of the CNBD

A) Two views of a surface representation of the CNBD of HCN2. Residues that were affected by 25 μM TRIP8b binding as measured by NMR are shown in red. Blue residues were assigned but unaffected in the NMR experiments. White residues were not assigned. B) Cartoon summary that shows a tetrameric CNBD (from accession number 1Q5O) with the residues affected by TRIP8b binding colored in red. TRIP8b_core 248C (yellow) and 261C (green) trilateration isosurfaces are shown on the tetramer of the HCN2 CNBD to demonstrate that the localized positions do not clash with the neighboring subunits. The TPR domain of TRIP8b bound to the front subunit is colored in green and shown binding to the distal c-terminus of the HCN channel.