1.1 Rectification of K currents

The Kv channels, including HERG, are expressed in the CNS of vertebrates and invertebrates and may modulate the physiology of the brain (Fano et al., 2012). The Kv channels undergo diverse activation, inactivation and deactivation patterns. Based on these properties, they can be subdivided into the three groups, though there are also overlapping peculiarities. One of the first recognized types is a delayed rectifier, exhibiting a slow activation but not inactivation during steps (Figure 1a–d). Note that delayed activation can be more pronounced in some cell types. Rapid activation and fast inactivation of channels result in A-type currents (same as I_t0 or Kv4 and Kv1.4), which are comparable across the animal kingdom and organs (Figure 1e). Only HERG currents represent a unique type of channel gating, namely, a typical delayed activation and no inactivation during 2 s of mild depolarization and faster activation, albeit a decrease in amplitudes upon stronger steps within the physiological range of membrane potential (MP), the lower and upper limit of AP (Figure 1f).

The HERG gene was discovered in Drosophila, a truly insightful experimental model with genetic and physiological implications (Kodirov, 2022). Another distinguishing feature of the HERG channel is its deactivation—tail, which results in more facilitated currents than the main component (Figure 2b). Interestingly, currents via all Kv channels but HERG are of the clear-cut outwardly rectifying type (Figure 2a). HERG currents are also outwardly rectifying up to an MP of +20 mV, but thereafter, they decrease in amplitude, referred to as rectification (Figure 2c). Also, tails are progressively higher up to +30 mV and upon further depolarization saturate (Figure 2d). This substantiates the notion that currents during activation and deactivation are of the same nature and that the former is a driver while the latter is a follower and reflects closely the changes occurring during the former. Co-expression of KCNH2 and KCNQ1 α subunits increases outward currents and the extent of rectification (Figure 2c). The KCNQ1 α subunit acts as an auxiliary element since it does not alter the waveforms of HERG currents, including the saturation of tails (Figure 2d).

Thus, there is no rectification of A current in MNTB as observed for HERG (Trudeau et al., 1995), but the presumptive possibility is assumed to be caused by intracellular Mg²⁺ (Johnston et al., 2008) as in the case of Kir and NMDA receptors (Matsuda et al., 1987; Westbrook & Mayer, 1987). The currents via cloned Kv1.4 or Kv4.2 do not exhibit inward rectification, whether normalized as an ohmic or GHK I–V relationship when recorded in HEK293 cells or oocytes (Clay, 2000). Moreover, in MNTB neurons, there is not even a saturation of outward currents at terminal depolarization as established for I_Kr, I_Ks, Kv1.5 or Kv2.1 (Kodirov, 2020). Also, for delayed rectifier channels, an inward rectification is attributed, which is absent when no Na⁺ is present in the intracellular solution but emerges at 4 mM concentration (Johansson et al., 1996). The latter inward rectification occurs independently of the Mg²⁺ block and is present in the absence of this ion. However, the observed rectification is irrelevant since it starts after +50 mV and is recorded during excessive depolarization up to +120 mV.

The absence of A-channels in rat MNTB is also problematic when (Johnston et al., 2008), intriguingly, they are present in evolutionary unrelated animals, namely, in identified neurons of mollusks (Kiss et al., 2002; Neher, 1971; Nerbonne & Gurney, 1987). Besides, in Figure 2d, in traces from mice and rats, despite the presence of 3 mM TEA and 10 nM dendrotoxin-I, the delayed rectifier currents are activated at −17 mV from a prepulse of −107 mV, especially in the former (Johnston et al., 2008). The latter is in contrast to the waveforms of pharmacologically isolated A-type currents in Figure 1e, albeit with a prepulse of −97 mV. When cited in italics as above and elsewhere, the reader is prompted to learn from the corresponding figures in the referred studies.

The perikaryon of principal MNTB neurons is ellipsoid or spherical (globular or oval) in shape and relatively small at about 22 × 15 μm (Banks & Smith, 1992; Sommer et al., 1993). Therefore, perhaps electrodes with up to 12 MΩ tip resistance were used (Johnston et al., 2008), which are good only for single channel recordings, but not whole-cell configuration. The activation curve for the range of up to −7 mV is puzzling since the contribution of Kv4 channels to spike generation is expected after the overshoot starting at about +40 mV. Based on representative data points and perhaps a single case of rectification, the underlying currents might well be HERG-like. However, from recordings in Figure 1a,c,d, one will never observe a rectification, especially at +3 mV. The same concern applies as to saturation at −7 mV.

Perhaps still, ‘neurones of the medial nucleus of the trapezoid body (MNTB) are an attractive preparation for studying native K⁺ channels’ (Johnston et al., 2008). However, how should one understand that ‘these relatively compact neurones facilitate good voltage control’? Although neurons of stratum pyramidale of hippocampus are densely located in a lamina, adequate voltage control is not possible. However, it is not because of their perikaryon. But simply, neurons within a slice retain their axons and dendrites, which contributes to space-clamp limitations (MacDermott & Westbrook, 1986).

Even when there was a rectification, one could not define the half-maximal activation V_0.5 from a narrow range of −37 to −7 mV for an A-type or any Kv channel (Figure 3b). The legend of Figure 1b states that the V_0.5 values of activation and inactivation are −35.0 ± 1.3 and −76.9 ± 0.9 mV for a single neuron, respectively (Johnston et al., 2008). These values, however, appear identically for mean data (n = 9). The latter also includes the identical corresponding slope factors of 6.7 ± 0.3 and 7.0 ± 0.3 mV. The exclusion of activation data points at +3 and +13 mV (dashed box) during sigmoidal fit is not a doctrine of biophysics or ion channel physiology. The value of I_max is then artificially considered to be that of −7 mV. Those two values exhibit inward rectification here, but not in the original traces in Figure 5. A correction of experimental peak amplitudes for a single channel current non-linearity using a modified GHK formula should not unmask it. There is no GHK rectification for A-type channels, at least not for cloned Kv4.2 (Clay, 2000). In contrast to the ohmic I–V relationship, the GHK one exhibits a saturation of outward currents between +10 and +60 mV. Interestingly, the HH model of a single MNTB neuron or simulated A currents based on a mean of n = 5 has resembling waveforms, albeit the decreased amplitudes at −7 and +3 mV and increased ones at −37 and −27 mV, while unchanged at −17 mV for the latter (Figure 5b). In Figure 5a, perhaps the same traces as in Figure 4b are presented, albeit the excluded one at +13 mV (Figure 6).

From raw data as in Figure 1c (Johnston et al., 2008), one never obtains a rectification, though an A-type current activation via Kv4 channels was insightfully isolated from I_K by a double-test pulse protocol (T1 followed by T2 for each parallel step). Note that because of the short prepulse of −37 mV, there is no I_K activation at this voltage. In this case, for −37 mV at T2, the prepulse in turn is −117 mV.

Because the two current components have designated long prepulses, a holding potential of −77 mV plays no critical role in their activation. However, −77 mV is healthy for neurons, as it resembles the RMP. If a neuron requires up to −250 pA of current to reach the −77 mV level, it is of poor quality, as the RMP is severely depolarized.