A robust expression system reveals distinct gating mechanisms and calmodulin regulation of Na V 1.9 channels

Fig. 1.. Characterization of mNaV1.9 currents in ND7/23 cells. (A) Trace and (B) activation voltage (G-V) and channel availability (I-V) curve comparison between mNaV1.9 (blue) and hNaV1.5 (gray) highlight typical features of NaV1.9 gating including large persistent currents, hyperpolarized activation voltage, depolarized voltage dependency of inactivation, and slow kinetics of activation as well as inactivation. hNaV1.9 currents were too small for accurate analysis (A). mNaV1.9 currents were recorded in ND7/23 cells and hNaV1.5 currents in HEK293T cells. mNaV1.9 channel activation over time [5 min (dark blue) versus 7 min (teal)] and in the presence of 200 μM GTP-γ-S (yellow), illustrating (C) current density augmentation and (D) channel opening at more negative voltages. (E) Time to peak is not significantly different under measured conditions. ns, not significant. (F) GTP-γ-S accelerates inactivation kinetics (τ) and (G) reduces persistent currents. Because a time frame of 100 ms was insufficient to calculate the time constant (τ) of inactivation, recordings at the 7-min time point were used where the test pulse duration was 500 ms. *P < 0.05 and ***P < 0.001.

Fig. 2.. Comparison of mNaV1.9 gating properties in ND7/23 cells to DRG neurons. Endogenous mNaV1.9 currents in DRG neurons (red) show markedly slower kinetics compared to those observed in ND7/23 cells (blue). This is illustrated by (A) the NaV1.9-current trace recorded from DRG on two different timescales (left: first 100-ms pulse duration; right: 500-ms pulse duration), (B) peak currents being reached up to four times slower, and (D) slowed inactivation as indicated by a larger persistent current component at 100 ms but not 500 ms. mDRG, mouse DRG. In addition, (C) the channel availability curve of currents recorded from DRG was depolarized by ~20 mV compared to mNaV1.9 currents measured in ND7/23 cells. Although the offset of the channel availability curve from heterologous mNaV1.9 seems to suggest that persistent currents were larger in ND7/23 cells compared to DRG, this was not reflected in absolute measurements of persistent current at peak voltage (D). Most likely, this discrepancy is due to the smaller size of heterologous currents and the presence of other endogenous currents in ND7/23 cells, especially at higher voltages (fig. S2). DRG data were obtained from two biological…

Fig. 3.. Gating characterization of hNaV1.9-based chimeras in ND7/23 cells. (A) Left: Only transient transfection of mNaV1.9 (dark blue) and the hNaV1.9-1.5Ctail (teal) construct generated measurable currents. hNaV1.9-1.5DI currents (yellow) were observed in 3 of 52 cells. Right: Half-maximal activation voltage of both of these NaV1.9-based chimeras revealed sizable shifts. (B) Comparison of traces at peak voltage illustrates the resemblance of hNaV1.9-1.5DI to hNaV1.5 (gray) and of hNaV1.9-1.5Ctail to mNaV1.9. However, kinetics of (C) activation as well as (D) inactivation of the hNaV1.9-1.5Ctail construct are faster, and (E) persistent current fraction is lower. (F) Both activation voltage and channel availability curves of hNaV1.9-1.5Ctail are hyperpolarized compared to mNaV1.9. *P < 0.05 and **P < 0.01.

Fig. 4.. Current density and voltage dependency of activation of hNaV1.5-based C-tail chimeras. (A) Current density comparison of all hNaV1.5-1.9 chimeras (HEK293T), hNaV1.5 (HEK293T), hNaV1.9-1.5Ctail (ND7/23), and mNaV1.9 (ND7/23). (B) Conductance (G)-voltage (V) relationships of all constructs. (C) Individual half-maximal activation voltage and (D) slope factors calculated separately for each cell.

Fig. 5.. Voltage dependency of inactivation of hNaV1.5-based C-tail chimeras. (A) Half-maximal inactivation voltage and (B) slope factors of channel availability curves calculated per cell for all hNaV1.5-1.9 chimeras (HEK293T), hNaV1.5 (HEK293T), hNaV1.9-1.5Ctail (ND7/23), and mNaV1.9 (ND7/23). (C) Overview of mean channel availability curves for all constructs.

Fig. 6.. Comparison of activation and inactivation kinetics of all tested chimeras. (A) The time constant of inactivation (τ) and (C) time to peak in the function of pulse voltage for all constructs illustrate differences between NaV1.9- and hNaV1.5-based constructs. (B) Detailed comparison of inactivation time constants at voltages at which peak current was reached (−40 to −10 mV; see Materials and Methods) between hNaV1.5-based constructs. Although the increase in hNaV1.5-1.9IQ compared to hNaV1.5 did not reach statistical significance after post hoc corrections (P = 0.11), the difference was significant compared to NaV1.5-1.9preIQ-EF (yellow). (D) Similar analysis of activation parameters also indicated slower kinetics for hNaV1.5-1.9preIQ, while hNaV1.5-1.9preIQ-EF tended to be faster than other constructs. The difference between hNaV1.5-1.9preIQ-EF and hNaV1.5 was, however, not significant (P = 0.19). (E) A small but significant increase in persistent current was found for the hNaV1.5-1.9IQ construct compared to some of the other constructs, although the difference fell just below the significance criterion for the comparison with hNaV1.5 (P = 0.078). All P values can be found in…

Fig. 7.. Investigation of CaM binding to the hNaV1.9 C-tail using ITC. Titration of (A) apoCaM (Ca2+-free) into the hNaV1.9 IQ domain (dark blue), (B) Ca2+/CaM into the hNaV1.9 IQ domain, (C) apoCaM into the hNaV1.9 C-tail, and (D) Ca2+/CaM into the NaV1.9 C-tail containing IQ, pre-IQ (teal), and EF-hand (yellow) indicates low affinities that are unlikely to occur under physiological conditions. In titration (D), the heat is low and the binding isotherm is featureless, resulting in a poor fit that again indicates very low affinity. Fitting parameters are shown in table S5.

Fig. 8.. Characterization of the p.L811P (human) and p.L799P (mouse) mutations. Mutations were introduced in mNaV1.9 and hNaV1.9-1.5Ctail and transfected into ND7/23 cells. Both mutants demonstrated (A) a similar shift in the voltage dependency of activation and (B) a distinctive flattening of the channel availability curve. (C) Current densities were augmented in both variants, although the difference did not reach statistical significance in hNaV1.9-1.5Ctailp.L811P. Kinetics of (D) activation and (E) fast inactivation were slower for both orthologs. The fraction of persistent current in (E) was measured after 100 ms. As apparent from current traces in (F), a 500-ms depolarizing pulse was insufficiently long to achieve full inactivation. An overview of all parameters can be found in tables S14 and S15. ***P < 0.001 and ****P < 0.0001.

Fig. 9.. Overview of targeted regions within the C-tail of hNaV1.5 and hNaV1.9. (A) Sequence alignment of hNaV1.5 and hNaV1.9 C-tail regions, starting behind the last segment of domain IV (DIV-S6, gray) and the regions for CaM binding: the EF-hand domain (yellow), pre-IQ (teal), and IQ domain (dark blue). Identical (dark gray) and conserved (light gray) residues are highlighted. (B) AlphaFold model of hNaV1.5, showing a close-up around the proximal C terminus. In this model, the linker between DIII and DIV (red) makes several interactions with the EF-hand domain, including a salt bridge that includes K1504 in the DIII-DIV linker. The K1504E/K1505E double mutant was shown to decrease inactivation and to cause a left shift in the steady-state inactivation curve (43), suggesting that this interaction is involved in inactivation gating. (C) Superposition of the model from (B) with a crystal structure of the hNaV1.5 C-tail that does not include the DIII-DIV linker [PDB (Protein Data Bank) ID: 6MUD; (39)]. Here, the pre-IQ domain is at a different location that would clash with the DIII-DIV linker, suggesting that it has to move to allow the interaction of the DIII-DIV linker with the EF-hand…