C-terminus–independent multimerization of Hv1 proton channels

Figures

• 

  Figure 1. Gating differences between WT and ΔC channels expressed in Xenopus oocytes.A, representative current families of WT and (B) ΔC channels. WT channels show a prominent delay. The dotted blue curve is the exponential fit to Equation 3 as indicated in the methods section. The second half of current activation used for the exponential fit is shown in purple for all traces. The panels to the right show that the delay in activation disappears in ΔC channels, since the exponential function extrapolates to the start of the voltage pulse. The voltage protocol used for currents in A and B is shown at the bottom of the figure. C, normalized conductance of both types of channels. The black curves are the fits of the data to Equation 2 with parameters: q = 1.79 eo, V0.5 = 23.9 mV for WT and for ΔC: q = 2.13 eo, V0.5 = 39.6 mV. Data points are the mean of n = 5 independent patches. Error bars are the SD The mean value of the V0.5 for WT is 24.0 mV and for ΔC is 40.7 mV, which are statistically different (t test, p = 3.90 x 10−5). Both expressed channels are N-terminal tagged with mVenus fluorescent protein.

• 

  Figure 2. Different activity levels in WT and ΔC human Hv1channels.A, image of a patch pipette superimposed on the fluorescence image of an inside-out patch containing WT channels tagged with the mVenus fluorescent protein. Next to the image is a family of proton currents recorded simultaneously. B, similar to (A) for ΔC channels. C, the tail current recorded at −80 mV from pulses at +80 mV and +100 mV for WT and ΔC, respectively, was plotted as a function of the fluorescence of the patch. Linear functions were least-squares fitted to the data. Correlation coefficients are r = 0.748 for WT and r = 0.888 for ΔC. The inset indicates the proton gradient used for these recordings.

• 

  Figure 3. Experimental detection of formation of dimeric channels in WT and ΔC channels by FRET.A, apparent FRET efficiency as a function of the donor-acceptor ratio Id/Ia measured from channels tagged with mCerulean and mCitrine at the N terminus. The blue squares and error bars are the mean and SD from measurements of 203 individual cells. The dark curve is the prediction of Equations 5 and 6, for a dimer with fluorophores separated by the indicated distance r, with a distance dispersion indicated by the ± value. The dotted curve is the prediction for a trimer and the thin solid curve the prediction for a tetramer, both with the same separation of 57 Å between fluorophores. B, WT dimers formed between subunits tagged in the N-terminus and C-terminus. C, dimers can be formed between ΔC subunits. Robust FRET is observed that is also consistent with a dimer. In all panes, the distance dispersion, A, predicts the FRET efficiency dispersion indicated by the shaded region. Below each data panel is a cartoon depicting the expected organization of the dimers and FPs in each condition. Data points and errors are the mean ± SD.

• 

  Figure 4. FRET in the N- and C-truncated hHv1 channel.A, functional expression of the ΔNΔC-hHv1 channels tagged with mCitrine or mCerulean in the N-terminus used to measure FRET. B, the G-V curves of these constructs compared to WT and ΔC channels (dotted and dashed curves as indicated in the figure). The mean value of the G-V curves of the ΔNΔC are indicated by colored circles. The error bars are the SD, n = 5 patches for each construct. C, apparent FRET efficiency is shown in the left panel. Symbols are the mean and the error bars represent SD; n = 98 independent cells. Dotted and dashed curves shown are the fits to dimer or trimer models with the main distance between FRET pairs indicated by r (right panel). The fit suggests that the most likely stoichiometry of this construct is as trimers in the membrane of HEK cells.

• 

  Figure 5. Native gel of WT and constructs lacking C and C-and-N terminal domains.A, non-denaturing gel blotted against mCitrine with an anti-GFP antibody. The ΔNΔC lane was exposed three times longer. The construct is indicated at the top of each lane. Blue arrows indicate the expected molecular weight of dimers for each construct. WT-hHv1 122 kDa, hHv1-ΔC 102 kDa, and hHv1-ΔNΔC, 80 kDa. Expected molecular weights for the monomers are: WT-hHv1 61 kDa, hHv1-ΔC 51 kDa, and hHv1-ΔNΔC, 40 kDa. The yellow arrow indicates the position of the monomer for hHv1-ΔC (50 kDa) and the purple arrow the monomer for hHv1-ΔNΔC. B, the 2D denaturing PAGE of the lanes in gel A. All high molecular weight bands in A collapse to the expected weight of the monomer for each construct. The second less intense higher molecular weight band observed is likely a glycosylated form of the channel.

• 

  Figure 6. Sigmoidicity of activation of AmHv1 and AmHv1-ΔC.A, mCitrine-WT currents (gray traces) are fitted with Equation 4 (dotted curves) with an n exponent that varies with voltage. Activation occurs with a delay and with a sigmoidal time course. B, currents of mCitrine-ΔC channels (gray traces) fitted to Equation 4 as in A. Note the near-exponential time course of activation. ΔpH = 1 for both experiment groups. C, voltage-dependence of conductance activation of WT (blue circles) and ΔC channels (yellow circles). Continuous curves are best fits to Equation 2 with parameters: WT, 1.97 eo, 37.1 mV; ΔC, 1.14 eo, 74.7 mV. D, summary of the values of the exponent n as a function of voltage derived from fits as in A and B. Blue circles are WT and yellow circles are ΔC channels. The continuous line is the mean value, and the shadowed areas are the SD. D, comparison of the activation kinetics as expressed by an exponential function raised to a power n. Dotted curves are predictions for different values of n. Continuous curves are the kinetics of an exponential raised to the average power n that was estimated from data at 60 mV, for both WT and ΔC channels.

• 

  Figure 7. Stoichiometry of AmHv1 channels estimated by FRET.A, WT channels tagged at the N-terminus with mCitrine or Cerulean fluorescent proteins were coexpressed in HEK293 cells. The assembled channels produced robust FRET signals that are consistent with a dimeric stoichiometry. An estimate of the distance r between the FRET pair is shown. B, the FRET signal from coexpressed ΔC-mCitrine and mCerulean subunits is also robust, and is consistent with these channels also assembling as dimers, as in the human ortholog, as indicated by the fit (continuous curve). Symbols are the mean apparent FRET and the error bars are the SD. Data from 155 cells for ΔC and 134 for WT. AmHv1, Acropora millepora human Hv1; HEK, human embryonic kidney.

• 

  Figure 8. Wild-type and ΔC human channels can theoretically form dimers.A, dimeric assemblies in WT and ΔC channels generated in HADDOCK. In the WT dimer, the main interaction surface is mediated by the S4-S4 and S1-S4 segments, as well as the C-terminal helix. Dimer interaction in ΔC channel is formed by S1-S4 and S4-S4 segments of different subunits. B, dimer assemblies of WT and ΔC channels generated using Alphafold2-multimer. In WT channels, the main interaction surface is formed by the S4 segments and the C-terminal helix. The interaction surface in ΔC channels is formed by the S4 segments of different subunits according to the AlphaFold prediction. In addition, some residues of the S1 segments contribute to the interaction surface with the S4 segment. The S1 and S4 segments of each monomeric subunit are depicted in pink and green, respectively. Interface residues are represented as spheres, color-coded as indicated. The segments S2, S3, and the N- and C-terminal regions are in gray. C, dimeric assemblies of WT and ΔC AmHv1 channels, respectively, generated in HADDOCK. In the WT channels, the primary interaction surface is formed between the S4 segments and the extended C-terminal…