Journal of Neurophysiology

Spatial variation in membrane excitability modulated by 4-AP-sensitive K+ channels in the axons of the crayfish neuromuscular junction

Jen-Wei Lin

Abstract

Current-clamp recordings were made from the primary (1°) and secondary (2°) branching points (BPs) of axons at the crayfish neuromuscular junction. Action potential (AP) firing initiated by current injected at the 2° BP showed strong adaptation or high-frequency firing at threshold current, whereas AP firing frequency at the 1° BP exhibited a gradual rise with increasing current amplitude. The voltage threshold for AP (VTH) was higher at the 2° BP than the 1° BP. 4-Aminopyridine (4-AP) at 200 μM increased AP amplitude and duration at both BPs but reduced threshold current at the 2° BP more than at the 1° BP. This blocker lowered VTH at both BPs, but the difference between the BPs remained. Firing patterns evoked at the 2° BP became similar to those evoked at the 1° BP in 4-AP. Thus 4-AP-sensitive channels may be more concentrated in the distal axon and control AP initiation and firing patterns there. Orthodromic APs between the two BPs were also compared. There was no difference in AP amplitude between the two BPs, but AP half-width recorded at the 2° BP was longer than that at the 1° BP. AP duration at both BPs increased gradually, by ∼17%, during a 100-Hz, 500-ms train (in-train rise). Normalized AP half-widths revealed a smaller fractional in-train rise at the 2° BP. Thus, although distal APs were broader, AP duration there was under more stringent control than that of the proximal axon. 4-AP increased AP amplitude and duration of the entire orthodromic train and reduced the magnitude of the in-train rise in AP half-width at both BPs. However, this blocker did not uncover a clear difference between the two BPs. Thus 4-AP-sensitive channels concentrated in distal axon may be essential in preventing unintended firing and modulating AP waveform without interfering with orthodromic AP propagation.

  • localized distribution

axons, initial segments, soma, and dendrites are the four well-defined compartments of neurons. Membrane excitability and synaptic interactions have been studied extensively in soma, dendrites, and the initial segments due to their accessibility to patch electrodes. By comparison, electrophysiological studies in distal axons have been less extensive, although distal axons are also capable of signal processing (Debanne 2004).

Recent patch-clamp and imaging studies of presynaptic terminals have demonstrated an important and diverse role for K+ channels in regulating action potential (AP) initiation, waveform, and firing patterns (Johnston et al. 2010; Rudy and McBain 2001). Detailed analyses of K+ channels at the calyx of Held (Dodson et al. 2003; Ishikawa et al. 2003), hippocampal mossy fiber terminals (Alle et al. 2011; Geiger and Jonas 2000), and cerebellar basket cell terminals (Robertson and Southan 1999) have established a close link between K+ channel kinetics and specific aspects of axonal and terminal firing. However, given the diversity of neuronal types in the mammalian central nervous system (CNS), it is conceivable that the function of individual classes of K+ channels may differ among neurons. For example, some Kv1 channel blockers did not alter AP waveform or transmitter release when applied alone at the calyx of Held (Dodson et al. 2003; Ishikawa et al. 2003), whereas Kv1 blockers enhanced transmitter release at the neuromuscular junction (Anderson and Harvey 1988), at inhibitory synapses in the cerebral cortex and cerebellum (Cunningham and Jones 2001; Southan and Robertson 1998; Tan and Llano 1999), and at synapses between cortical pyramidal cells (Kole et al. 2007). Thus, although the kinetics of specific K+ channel types often dovetail with their expected function in regulating AP firing, their role in axons of different neurons may vary depending on the type and density of other voltage-gated channels present. The full range of functions that can be ascribed to each class of K+ channel remains to be further explored in axons of different neurons.

The studies reviewed above were performed by directly patching synaptic terminals or by indirect inference from Ca2+ imaging or transmitter release. As a result, the spatial distribution of the K+ channels involved and their functional roles in axon branches proximal to the terminals remain unclear. The majority of axons in the mammalian brain branch repeatedly before terminating at synapses. Thus it is unknown whether different K+ channel types are distributed uniformly along gradually tapering and branching axons, or are they localized exclusively at release sites? How far “upstream” of the axon should these channels be distributed to shape AP waveform in terminals effectively? Although immunocytochemical studies have demonstrated the presence of K+ channels in thin axons connecting terminal varicosities (Alonso-Espinaco et al. 2008; Puente et al. 2010), the relative impact of distributed vs. localized K+ channel distribution on the AP waveform of axons and terminals has not been studied in detail.

Axons in the opener neuromuscular junction of the crayfish first walking leg represent a unique model system for the study of local variations in membrane excitability of branching axons. These axons branch repeatedly and terminate as strings of varicosities (Cooper et al. 1996; Florey and Cahill 1982), similar in morphology to axons in the mammalian CNS. The axons are large enough for sharp electrode penetration at multiple locations, and membrane activities at synaptic terminals can be monitored with voltage indicators (Lin 2008). This report uses two-electrode current-clamp (TECC) from proximal and distal regions of axons to examine variation in the spatial distribution of 4-aminopyridine (4-AP)-sensitive K+ channels. The effect of this blocker on orthodromically conducting AP trains is also investigated.

METHODS

Crayfish (Procambarus clarkii) were purchased from Atchafalaya Biological Supplies (Raceland, LA). The small animals, 4–6 cm, head to tail, were maintained in tap water at room temperature (22°C), fed three times per week, and water was replaced twice per week. All experiments were performed at 22°C. The first walking leg was removed by autotomy and fixed dactylopodite-side down to a 15-mm petri dish with cyanoacrylate. The opener axon-muscle preparation was then dissected in saline. To ensure complete drug access to axons, the upper half of the shell of the carpopodite was also removed such that the entire length of the axons was directly exposed to perfusing saline. Only the inhibitory axon was used in this study.

Physiological saline contained (in mM): 195 NaCl, 5.4 KCl, 13.5 CaCl2, 2.6 MgCl2, and 10 HEPES, titrated to pH 7.4 with NaOH. The saline was circulated by a peristaltic pump at the rate of 1 ml/min. Glucose (1 mg/ml) and gentamicin (1 μg/ml) were added to the saline. 4-AP was added to the saline directly from a 1 M stock solution, prepared by dissolving it with HEPES free acid at 1:1 molar ratio and storing at −20°C. After adding the blocker, a new steady-state was typically achieved within 20 min. All data presented in this report were collected 30–45 min after the blocker was added. All chemicals were purchased from Sigma.

Two microelectrode amplifiers (Warner IE-210) were used to perform intracellular recordings. Voltage signals were filtered at 2 kHz and digitized at 50 μs. Data were collected and analyzed with IGOR (WaveMetrics, Lake Oswego, OR). Microelectrodes were filled with 500 mM K-methanesulfonate and 5 mM K-HEPES and had a resistance of 50–100 MΩ. The pipette solution was prepared by titrating KOH of known quantity with methanesulfonic acid to pH ∼7.5. K-HEPES (1 M, pH 7.6) and water were then added to adjust the final concentrations of K-methanesulfonate to 500 mM and K-HEPES to 5 mM, pH 7.4.

Axon penetration was performed under a ×60 water immersion lens on a fixed-stage microscope (Axioskop). The typical resting membrane potential (Vm) was −70 to −80 mV, and preparations with a Vm less than −65 mV were not used. Additional criteria used to monitor the viability of axons included: 1) stability of input resistance (Rin) and resting Vm; 2) smooth consistency of axoplasm as inspected visually; and 3) functional synaptic transmission tested at the end of some experiments.

Orthodromic stimulation of AP was triggered by 0.3-ms pulses delivered by a bipolar electrode located ∼3 mm proximal to the intracellular electrodes. The change in AP conduction velocity was estimated from the time delay between the orthodromic stimulus artifact and the maximum of the time derivative of the first AP of an orthodromic train.

Measurement of fluorescence transients using JPW 1114.

Details of dye injection and fluorescence signal detection have been described previously (Lin 2008). JPW 1114 was injected by a pipette containing JPW 1114 (0.5 mg/ml), K-methanesulfonate (330 mM), K-HEPES (2 mM), and K-GABA (250 mM). An overnight protocol was followed where 3-nA, 300-ms steps were delivered with alternating polarity on a 30-s cycle. Intracellular and fluorescence recordings were carried out the next day.

Fluorescence transients were measured with a Hamamatsu photodiode (S5973-01) coupled to a single-channel headstage GeneClamp 500. A 150-W Xenon lamp was powered by an Opti Quip 1600 power supply with a 770 lamphouse. In some experiments, a high-power LED (528 nm; Diamond DRAGON; Osram) was used as the light source. The filter set used for JPW 1114 included an excitor 525/45 band-pass filter, a dichroic 560-nm long pass, and an emitter 575-nm long pass. Xenon lamp illumination was gated by a shutter (Uniblitz; Vincent Associates) mounted to the legs of a vibration isolation table to uncouple it from the lamphouse and microscope and to avoid mechanical artifacts picked up by the photodiode. The area of illumination was restricted by an iris diaphragm customized to allow an illumination diameter of approximately 20–50 μm with a ×60 objective. Normalization of fluorescence transients was calculated as ΔF/F = [F(t) − Frest]/Frest × 100, where Frest represents the fluorescence intensity of stained varicosities in the absence of activity. Background fluorescence in unstained regions was not subtracted.

RESULTS

AP firing patterns are different at the 1 and 2° BPs.

A typical opener axon is shown in Fig. 1D where the image was constructed from a confocal stack of an inhibitor filled with Calcium Orange. The primary (1°), secondary (2°), and tertiary (3°) branching points (BPs) are labeled. Two microelectrodes were typically placed at the 1 and 2° BPs simultaneously. Sometimes, the 3° BP was thick enough for electrode penetration. (BPs were preferred because they provide mechanical “anchoring” for electrode penetration.) The two electrodes were separated by 700–1,500 μm depending on the branching pattern of the axon and the size of the animal. Using this distant TECC, excitability at the two BPs was probed by 1-s current steps. Figure 1A3 shows two traces recorded at the 2° BP and evoked by subthreshold and threshold current (ITH) steps delivered to the 1° BP. [The lowest current amplitude capable of evoking APs is designated ITH.] The axon fired two APs at ITH and could fire continuously by large current steps (Fig. 1A2).

Fig. 1.

Two-electrode current-clamp (TECC) recordings from primary (1°) and secondary (2°) branching points (BPs). A3: voltage traces evoked by subthreshold (black) and threshold (ITH; gray) current steps injected at the 1° BP and recorded at the 2° BP (inset). Current amplitudes were 9 and 10 nA, respectively. A2: high-frequency firing evoked by a 25-nA current step. A1: raster plot of action potential (AP) firing. Dots on each row represent APs evoked by a specific current intensity (y-axis; see B1 for vertical scale). The raster plot is aligned to the same time scale as the traces in A2 and A3. B3: APs evoked by current injected at the 2° BP (inset). ITH evoked high-frequency and continuous firing (gray trace). Subthreshold and ITH were 9 and 10 nA, respectively. B2: AP train triggered by a 15-nA current step. B1: raster plot of AP evoked by current steps injected at the 2° BP. C: current dependence of AP firing (AP#-I), displayed as the number of APs fired during 1-s steps (AP #/sec) plotted against current amplitude. AP firing initiated at the 1° BP exhibits a smooth rise (circles), whereas the firing initiated at the 2° BP shows a step rise (circles). D: confocal stack of a typical axon labeled with Calcium Orange. The 1, 2, and 3° BPs are labeled. This preparation was not used for the traces shown in A and B. Traces in A2, A3, B2, and B3 share the same vertical and time scales.

For current steps injected at the 2° BP of the same axon, ITH initiated a high-frequency and continuous train (Fig. 1B3). Raising current amplitude further increased firing frequency with minor adaptation (Fig. 1B2). Since firing did not always last during the entire current step (see Fig. 4), a raster plot was used to illustrate the progression of AP firing with increasing current amplitude (Fig. 1, A1 and B1). Here, current amplitude is displayed on the y-axis, whereas the dots representing individual APs are plotted on the same time scale as that in the traces below. AP firing evoked at the 2° BP started abruptly at a high frequency (Fig. 1B1), whereas firing frequency increased gradually when initiated from the 1° BP (Fig. 1A1). The current dependence of AP firing has been summarized by plotting the number of APs during the 1-s current step against the magnitude of injected current (AP#-I; Fig. 1C). ITH values at the two locations of this preparation were identical, but the AP#-I plot resulting from current injection at the 2° BP exhibits a discontinuity, whereas the plot derived from the 1° BP has a smooth rising phase. Thus current steps injected at the two BPs initiated distinct firing patterns.

Since distant TECC does not report Vm at the current injection site, it remains unclear whether the distinct firing patterns evoked at the two BPs are indeed generated locally. Two separate approaches were used to resolve this uncertainty.

First, a “hybrid” simultaneous recording from the two BPs was implemented by combining photometric measurement of voltage indicator JPW 1114 with distant TECC. Figure 2 shows fluorescence transients and microelectrode recordings obtained simultaneously. In each inset, the gray microelectrode indicates the recording electrode that gives rise to the AP trace displayed in gray (APe), whereas the black circle represents the imaging site that gives rise to the fluorescence transient trace in black (APf). APs were initiated by current injection at the 1 and 2° BP and by orthodromic stimulation. The three stimulation protocols were delivered in order in a cyclic manner for each imaging location. When APf was scaled to match APe evoked by orthodromic stimulation (Fig. 2, A1 and B1), the same scaling factor also resulted in a near-perfect fit between APf and APe when the two recordings were obtained from the same location (Fig. 2, A3 and B2). When APf was measured at the current injection site and APe recorded from the other BP, the APf exhibited a higher threshold and earlier onset than the APe (Fig. 2, A2 and B3), suggesting that APs were initiated near the current injection site and propagated to the voltage recording electrode. This was a consistent observation in seven preparations.

Fig. 2.

A hybrid simultaneous recording achieved by combining distant TECC with photometric recording of voltage transients. A1–A3: APf recorded at the 1° BP (black) as AP was evoked by orthodromic stimulation (A1), ITH at the 1° BP (A2), and ITH at the 2° BP (A3). The gray microelectrode in each inset identifies the location from which APe (gray) was obtained. APf measured at the 1° BP overlapped perfectly with APe recorded there (A1 and A3) but exhibited an earlier onset when APe was recorded at the 2° BP and AP initiated at the 1° BP (A2). APf traces in A1–A3 have the same vertical scale and were obtained from the same recording session. B1–B3: APf measured at the 2° BP overlapped perfectly with APe recorded there (B1 and B2) but exhibited an earlier onset when APe was recorded at the 1° BP and AP initiated at the 2° BP (B3). APf traces in B1–B3 have the same vertical scale and were obtained from the same recording session. Time calibration applies to all panels. All traces in this graph were averages of 20 trials. Jittering in AP onset was corrected by aligning the peaks of APe traces before fluorescence transients were averaged; see Lin (2008) for details. APe, AP recorded with microelectrode; APf, AP measured from fluorescence transients.

Because of relatively low signal-to-noise ratio of the voltage transients, the APf traces shown in Fig. 2 were averages from 20 trials. The need to average precluded the analysis of firing patterns, which typically requires inspection of single traces.

To seek further support for the hypothesis that the different firing patterns shown in Fig. 1 indeed reflect localized difference in membrane excitability between the two BPs, local TECC was performed. For this configuration, two electrodes are placed within 50 μm of each other so that firing patterns at the current injection site can be directly measured. At the 1° BP, AP firing evoked by ITH and higher current amplitudes (Fig. 3, A1 and A2) gave rise to a smoothing rising AP#-I plot (Fig. 3C, gray dots) similar to that obtained with distant TECC (Fig. 1C). In a different preparation, local TECC performed at the 2° BP showed an abrupt onset of high-frequency firing at ITH (Fig. 3B2) and a discontinuous AP#-I plot (Fig. 3C, black dots). These features are similar to those evoked at the 2° BP and recorded with distant TECC (Fig. 1, B and C).

Fig. 3.

Local TECC recordings from the 1 and 2° BPs. A1 and A2: membrane responses recorded and evoked at the 1° BP. ITH (17 nA) evoked a single AP (A2), and larger current (24 nA) evoked continuous firing (A1). The subthreshold response shown in A2 was evoked by a 16-nA step. B1 and B2: membrane responses recorded and evoked at the 2° BP. ITH (13 nA) evoked high-frequency and continuous firing (B2), and a larger current step (20 nA) further increased firing frequency (B1). C: AP#-I plots measured from the 2 BPs shown in A and B. D: phase plots of APs shown in A2 (gray) and B2 (black). Inset shows the beginning of the upswing of the phase plots, with APs at the 2° BP (black) showing a more depolarized threshold. The horizontal line in the inset identifies AP threshold at 5 V/s. Vm, membrane potential.

AP firing threshold (VTH) at the two locations was examined by phase plots (Fig. 3D). Phase plots from both BPs exhibited a smooth upswing (inset), indicative of local initiation of AP (Bean 2007; Yu et al. 2008). When VTH was set as the level of Vm crossing 5 V/s (Fig. 3D, inset, horizontal line), VTH of the first AP at the 1 and 2° BP had a threshold of −41 and −36 mV, respectively. Averaged VTH recorded at the 1° BP was −47.0 ± 3.7 mV, which is significantly more hyperpolarized than −37.1 ± 3.8 mV as measured at the 2° BP (n = 5; P = 0.001). Thus the phase plots demonstrated that distinct firing patterns were indeed initiated locally and that VTH at the 2° BPs was significantly more depolarized than that of the 1° BP.

Results in Fig. 3 support the assumption that different firing patterns measured by distant TECC at 1 and 2° BPs reflect a difference in membrane excitability between these locations. Because of the technical difficulties involved with using local TECC, most pharmacological experiments presented in this report were obtained using distant TECC.

Effects of 4-AP on membrane excitability at the 1 and 2° BPs.

Since K+ channels have been shown to play a prominent role in controlling AP initiation and firing pattern, the possible presence of spatial variation in K+ channels was investigated with K+ channel blocker 4-AP (200 μM). Although this blocker at the concentrations used here may not be entirely selective to specific K+ channel isoforms, the results did prove informative.

Figure 4 shows the impact of 4-AP on AP firing recorded with distant TECC. In this preparation, ITH delivered to the 1° BP initiated a single AP (Fig. 4A2). A further increase in current amplitude resulted in continuous firing (Fig. 4A1, black dots). The current dependence of firing (Fig. 4A4, black circles) was similar to that obtained from the 1° BP in the experiments shown in Figs. 1 and 3. At the 2° BP, ITH initiated a burst of APs at the beginning of the current step (Fig. 4B2). Larger current steps increased and then decreased AP number, and firing continued to show strong adaptation (Fig. 4B1, black dots).

Fig. 4.

Effect of 200 μM 4-aminopyridine (4-AP) on the axonal firing pattern recorded with distant TECC. A2: ITH (7 nA) at the 1° BP evoked a single AP (dotted trace). The subthreshold response (thin trace) was evoked by a 6-nA step. A3: membrane responses evoked by subthreshold (4 nA) and ITH (5 nA) steps in 4-AP. The inset details the increase in AP amplitude and duration resulting from 4-AP. Asterisks in A2 and A3 identify the APs used for comparison in the inset. A1: raster plots of AP firing evoked in control saline (black dots) and in 4-AP (gray dots). A4: AP#-I plot constructed before and after 4-AP. The inset shows that membrane responses evoked by −5-nA steps were not altered in 4-AP (gray). The scale bars, also applicable to the inset in B4, represent 2 mV and 40 ms. B2 and B3: responses evoked by current step delivered to the 2° BP before (B2) and after (B3) 4-AP. This axon fired in high-frequency bursts at the beginning of ITH in control saline but was able to fire singly in 4-AP. The inset in B3 compares APs identified by asterisks in B2 and B3. B1: raster plot shows that the phasic nature of AP firing in control saline (block dots) was converted to 1 that fires continuously in 4-AP (gray dots). B4: AP#-I plots constructed from firing evoked at the 2° BP. The leftward shift in the AP#-I curve is larger than that shown for the 1° BP (A4). The inset shows that membrane responses evoked by −5-nA steps were not altered in 4-AP (gray). A1–A3 and B1–B3 share the same time scale. The insets in A3 and B3 share the same time scale. A2, A3, B2, B3, and the insets share the same vertical scale. Ctl, control.

Addition of 4-AP increased AP amplitude and duration at both BPs (Fig. 4, A3, B3, and insets). For APs initiated at the 1° BP, 4-AP caused a reduction in ITH and a leftward shift in the AP#-I plot (Fig. 4A4, gray circles). For APs initiated at the 2° BP, AP fired continuously with minor adaptation in 4-AP (Fig. 4B1, gray dots). The AP#-I plot after addition of 4-AP shows a smooth rise and a large leftward shift (Fig. 4B4, gray circles). The large reduction in ITH was not due to the blocking of 4-AP-sensitive channels opened at rest, since membrane responses evoked by −5-nA steps were not changed by 4-AP (Fig. 4, A4 and B4, insets). Thus 4-AP, in addition to increasing AP amplitude and duration at both BPs, reduced ITH at the 2° BP more than that at the 1° BP. Furthermore, this blocker qualitatively altered firing patterns at the 2° BP.

The effect of 4-AP was also investigated with local TECC (Fig. 5). The enhancing effect of 4-AP on AP amplitude and duration (Fig. 5, A1, A2, B1, and B2) was similar to that shown with distant TECC in Fig. 4. [The only exception was that 4-AP raised VTH at the 2° BP under distant TECC (Fig. 4B3, inset), whereas the blocker lowered VTH under local TECC (Fig. 5B2). See discussion for further comments.] 4-AP converted a discontinuous AP#-I plot to one with continuous rise at the 2° BP (Fig. 5B3), whereas this blocker created a small left shift and resulted in no qualitative change in the firing patterns obtained from the 1° BP (Fig. 5A3).

Fig. 5.

Effect of 4-AP at 200 μM on the axonal firing pattern examined with local TECC. A1: AP firing evoked by ITH before (black) and after (gray) 4-AP at the 1° BP. A2: increase in AP amplitude and duration in 4-AP, shown in detail by aligning the rising phase of the 1st AP from the traces shown in A1. A3: AP#-I plots derived from the recording series shown in A1. A4: phase plots of the 2 traces shown in A1. 4-AP increased the maximal rate of rise in the APs (gray dotted trace). The inset expands the initial upswing of the phase plot, demonstrating a slight leftward shift in the voltage threshold for AP (VTH). B1: AP firing evoked by ITH before (black) and after (gray) 4-AP at the 2° BP. B2: increase in AP amplitude and duration shown in detail by aligning the rising phase of the 1st AP from the traces shown in B1. B3: AP#-I plots derived the recording session shown in B1. B4: phase plots of the 2 traces shown in B1. 4-AP increased the maximal rate of rise in the APs (gray dotted trace). The inset expands the initial upswing of the phase plot, demonstrating a large leftward shift in VTH. The horizontal line indicates the 5 V/s level. The x-axis of the 2 insets are aligned to illustrate that AP thresholds at the 2° BP (bottom) are higher than those at the 1° BP (top).

Phase plots obtained before (black) and after (gray) 4-AP were compared at the two BPs (Fig. 5, A4 and B4). 4-AP accelerated the rate of rise at both BPs. In addition, 4-AP shifted VTH in a hyperpolarizing direction at both BPs (inset between Fig. 5, A4 and B4). Although the shift in VTH resulted from 4-AP appeared to be smaller at the 1° BP than at the 2° BP in this figure, the magnitude of the shift between the two BPs was not significantly different when averaged (n = 5). At the 1° BP, VTH was −47.0 ± 3.7 mV in control saline and −52.9 ± 5.4 mV in 4-AP (n = 5; P = 0.015, paired t-test). At the 2° BP, VTH was −37.1 ± 3.8 and −44.4 ± 4.3 mV in control saline and 4-AP, respectively (n = 5; P = 0.0013, paired t-test). The VTH at the two BPs was significantly different in control saline, and the difference remained significant in 4-AP. Thus, although 4-AP homogenized the firing pattern at the two BPs, the remaining difference in VTH between the two locations suggests that other voltage-gated channels may also have nonuniform spatial distribution.

Firing patterns initiated at the 2° BP shown in Figs. 4 and 5 are different, representing the range of variation among animals. The source of this variability is unknown, but the impact of 4-AP was consistent, namely that it converted the AP#-I from discontinuous or strongly adaptive shapes to one with a continuous rising phase and minimal adaptation.

Table 1 summarizes the averaged results (n = 11) of the effects of 200 μM 4-AP on AP amplitude, duration, and ITH obtained with distant TECC. AP amplitude and duration were measured from the first AP initiated by ITH (following the convention used in Figs. 1 and 4 where 1° BP and 2° BP designate parameters measured from APs evoked by current injection at the 1 and 2° BPs, respectively). AP amplitude was measured from the baseline before ITH onset to the peak of the first AP. AP duration was measured at 75% of peak amplitude, rather than at 50%, to avoid an overestimate of this parameter when the AP was riding on a large depolarization. Statistical analysis showed that 4-AP significantly increased AP amplitude and duration, and reduced ITH, at both BPs. 4-AP did not alter Rin or the time constant measured at resting Vm, suggesting an absence of 4-AP-sensitive K+ channel activity at resting Vm (see insets in Fig. 4, A4 and B4, for example).

View this table:
Table 1.

Summary on 4-AP effects on AP waveform evoked by distant TECC

Since the main focus of this report is spatial variation in membrane excitability, the relative impact of 4-AP at the two BPs was also compared. The most significantly changed parameter was ITH. The averaged reduction in ITH due to 4-AP was 3.9 ± 0.7 nA at the 1° BP, significantly lower than the 5.6 ± 0.7 nA decrease measured the 2° BP. The larger reduction in ITH at the 2° BP is even more significant considering the higher local Rin there. [Rin measured by local TECC was 1.8 ± 0.5 and 3.0 ± 0.3 MΩ at the 1 and 2° BP, respectively (n = 5).] There was no difference in the extent to which 4-AP increased AP amplitude at the two BPs. The increase in AP duration evoked at the 2° BP, 0.20 ± 0.020 ms, was significantly larger than that evoked at the 1° BP, 0.13 ± 0.02 ms.

In summary, the averaged results suggest that 4-AP-sensitive channels are likely to be more concentrated in the distal axon such that blocking these channels results in greater changes in AP initiation and waveform at the 2° BP. The pronounced effect of 4-AP on ITH suggests that this blocker targets channels active in subthreshold range. This pharmacological and kinetic profile is consistent with Kv1 channels (Johnston et al. 2010). However, α-dendrotoxin at 1 μM did not have any effect on crayfish axon (n = 5, data not shown).

In addition to the statistical comparison of parameters related to AP waveform, AP#-I plots compiled from 11 to 13 preparations were used to evaluate firing patterns. In Fig. 6, A and B, AP#-I plots obtained from different preparations were aligned by setting ITH in control saline to 1 nA. A sigmoid fit to data points from all preparations was used to generate an averaged trend for each condition: AP#Ifit=Max1+e(11/2I)/rate where Max (number of APs) represents the maximal number of APs fired by 1-s current steps, I1/2 (nanoamperes) represents the current level that activates 50% of maximal firing, and rate (nanoamperes) defines the rate of rise of the curve as a function of injected current.

Fig. 6.

Summary of the impact of 4-AP on firing patterns. A and B: effects of 4-AP on AP#-I plots initiated at the 1° (A) and 2° (B) BPs. AP#-I plots constructed in control saline are in open circles, whereas those obtained in 4-AP are in open triangles. Data from different preparations were aligned by setting ITH in control saline to 1 nA. A sigmoid function was used to fit data points from all preparations. The sigmoid fit in control saline and 4-AP are displayed in continuous and dotted gray lines, respectively. See text for comments on connected data points in B. Iinj, current injected.

In control saline, the sigmoid fit to data recorded at the 2° BP exhibited a steeper rise than that obtained from the 1° BP (Fig. 6, A and B, solid gray lines). The steep rise at the 2° BP can be partially attributed to the high local Rin, but a discontinuous firing pattern could also contribute to the steep rise. In fact, as suggested in Figs. 1 and 4, the firing pattern was variable at the 2° BP in control saline. Of the 13 preparations plotted in Fig. 6B, 4 exhibited discontinuous firing at ITH (connected gray circles), 5 showed phasic firing with strong adaptation (connected black circles), whereas the remaining 4 showed a continuous rise. In contrast, AP#-I plots constructed from current injection at the 1° BP consistently showed a smooth rise in control saline (Fig. 6A, solid gray line).

Introduction of 4-AP left-shifted ITH and significantly reduced the slope of the AP#-I plot at the 2° BP: the maximal slope of the sigmoid fit was 12.4 APs/nA in control saline and 5.7 APs/nA in 4-AP (Fig. 6B). At the 1° BP, 4-AP also created a leftward shift in the sigmoid fit, but the reduction in the slope was small, from 6.1 APs/nA in control saline to 4.2 APs/nA in 4-AP (Fig. 6A). [Since the left shift and the reduced slope in 4-AP were observed in every preparation, changes in these parameters were statistically significant in paired t-tests (data not shown).] In addition, 4-AP raised the maximal level of the sigmoid fit at both BPs. Thus 4-AP-sensitive channels play a more dominant role in shaping the current dependence of AP firing at the 2° BP compared with the 1° BP.

Impact of 4-AP on orthodromic AP.

Under physiological conditions, axons conduct APs initiated from the soma. To understand better the function of 4-AP-sensitive K+ channels and the significance of nonuniform spatial distribution of these channels under physiological conditions, the effect of this blocker on an orthodromic train, 100 Hz for 500 ms, was compared between BPs. In control saline, the peaks of the 1st 5 APs exhibited a slight rise and remained constant thereafter at both BPs (Fig. 7, A1 and B1). (Recordings obtained from the 1 and 2° BPs are displayed in A and B columns, respectively.) Since Vm between APs did not return to rest, peak level of AP was determined by the sum of AP height and interspike potential. To eliminate any contribution from interspike potential, AP amplitude was measured from the “foot” to the peak of the AP (double arrows in Fig. 7A, inset). The trend in AP amplitude showed an initial dip followed by a small rise (Fig. 7, A4 and B4, black dots). The initial dip occurred because the 1st AP took off from resting level, whereas the 2nd AP and those following it took off from the afterdepolarization of the preceding AP. AP half-width during the 100-Hz train increased rapidly during the 1st 7–10 APs (in-train rise) and was followed by a slower rise (Fig. 7, A5 and B5, black dots, and insets in A1 and B1). The in-train rise in amplitude and half-width followed a similar time course.

Fig. 7.

Effect of 4-AP on an orthodromic train. A1 and B1: 51 orthodromic APs, recorded simultaneously from the 1 and 2° BPs, were fired at 100 Hz in control saline. Insets in A1 and B1 compare the 1st and last APs to highlight the increase in AP amplitude and duration by the end of a high-frequency train. The superimposed traces also show that the last AP taking off from a slightly depolarized potential. The double-headed arrow in A1 indicates how AP amplitude was measured. A2, B2, and insets: high-frequency train recorded in 4-AP. The insets show that the peak and duration of the last AP in a train were similar to those of the 1st. A3 and B3: detailed comparison of orthodromic AP waveform resulting from 4-AP. A4, A5, B4, and B5: summary plots of AP amplitude (amp; A4 and B4) and duration (dur.; A5 and B5) for all APs of the high-frequency train. The parameters measured in control saline are shown in black dots, and those measured in 4-AP are in gray. A1–A3 and B1–B3 share the same vertical scale. Traces in A1 and A2 insets, B1 and B2 insets, and A3 and B3 share the same time scale. The time scale in B2 is shared by A1, A2, and B1.

The amplitude of orthodromic APs at the 1° BP was larger than that of those at the 2° BP (Fig. 7, A4 and B4, black dots). The half-widths of orthodromic APs recorded at the 2° BP were longer than those at the 1° BP (Fig. 7, A5 and B5, black dots).

Adding 4-AP resulted in an increase in AP amplitude for the entire AP train at both BPs (Fig. 7, A2, A3, B2, and B3). The initial rise in AP amplitude seen in control saline was no longer present (Fig. 7, A4 and B4, gray dots, and insets in A2 and B2). Instead, following the first AP, there was a slight and continuous decline in AP amplitude during the entire train (Fig. 7, A4 and B4, gray dots). The half-width of all APs increased in 4-AP and the in-train rise in this parameter decreased (Fig. 7, A5 and B5, gray dots). Thus, in addition to controlling the amplitude and duration of individual orthodromic APs, 4-AP-sensitive K+ channels play a role in modulating AP amplitude and duration during a high-frequency train. However, 4-AP did not change existing differences in AP amplitude and duration between the two BPs during orthodromic trains.

In control saline, the example in Fig. 7 suggested a smaller AP amplitude at the 2° BP than at the 1° BP. However, this difference was not statistically significant (data not shown). On the other hand, the larger AP half-width measured at the 2° BP was a consistent finding (Fig. 8; n = 7). Averaged results showed that this parameter was significantly larger at the 2° BP than at the 1° BP (Fig. 8A, bottom left bracket). To compare the relative magnitude of the rise in half-width during the high-frequency train, the half-width was normalized by that of the first AP in the train. The normalized plots show that this parameter rose by 18 and 16% for the 1 and 2° BPs, respectively, by the end of the train (Fig. 8B). Although the difference in the percentage of in-train rise between the two BPs was small, it was statistically significant (Fig. 8B, bottom right bracket). Thus, whereas distal APs exhibited a longer half-width, this parameter remained slightly more stable than that at the 1° BP during a high-frequency train.

Fig. 8.

Summary of the effects of 4-AP on the half-widths (HW) of orthodromic APs. A: half-widths of all APs in the 100-Hz train measured from the 1° (○) and 2° (□) BPs. Measurements obtained from control and 4-AP saline are plotted in black and gray symbols, respectively. Error bars represent standard errors of means (n = 7). The half-widths measured at the 2° BP were significantly larger than those at the 1° BP (left bottom bracket, P < 0.03). 4-AP significantly lengthened half-widths at both BPs (right brackets, P < 0.001). The difference in AP half-width between the 2 BPs was still significant in 4-AP except for the 1st 5 data points (left top bracket). B: normalized (norm.) AP half-width shows that, in control saline, the relative in-train rise in half-width at the 1° BP was significantly larger than that at the 2° BP (bottom right bracket P < 0.026). The fractional increase in AP half-width resulting from 4-AP was not statistically different between the 2 BP (ns; top right bracket). *Statistically significant.

4-AP significantly increased AP amplitude for the entire train at both BPs. Since this result is similar to the effect of 4-AP on APs evoked by TECC, statistical results will not be further detailed. The half-widths of APs in the orthodromic train were also significantly increased by 4-AP (Fig. 8A, gray symbols and right brackets), but the degree of in-train rise was reduced. The difference in AP half-width between the two BPs remained significant in 4-AP (Fig. 8A, top left brackets). Thus, unlike its impact on the firing patterns and ITH evoked with current steps, the impact of 4-AP on orthodromic AP did not exhibit clear spatial variation.

Finally, 4-AP-sensitive, Kv1-like channels have been shown to play an important role in controlling Vm between APs (Brew and Forsythe 1995; Gittelman and Tempel 2006; Klug and Trussell 2006). There was no difference in interspike potential between the two BPs in control saline, and 4-AP caused a small (<2 mV) depolarization of this parameter at both BPs (data not shown). Despite the large increase in AP amplitude and duration in 4-AP, the conduction velocity of orthodromic AP was either unchanged or slightly slowed in the blocker, by 3.4 ± 3.1% (n = 10).

DISCUSSION

In this report, TECC was used to investigate local variation in membrane excitability of the axons of the crayfish opener neuromuscular junction. The main finding was that the distal axon exhibited firing patterns distinct from those of the proximal axon. AP initiation in the distal axon required a higher ITH, with respect to local Rin, and had a higher VTH than APs initiated in the proximal axon. These differences were mostly homogenized by 4-AP at 200 μM, suggesting that the distal axon has a higher density of 4-AP-sensitive K+ channels than the proximal axon. The difference between 1 and 2° BPs was then compared when high-frequency trains were fired orthodromically. In control saline, AP duration recorded at the 2° BP was significantly longer than that at the 1° BP. There was a gradual increase in AP half-width during the high-frequency train (in-train rise). The normalized in-train rise of AP half-width was smaller at the 2° BP than at the 1° BP, suggesting a tighter in-train control of AP duration there. Although 4-AP increased the AP amplitude and duration of the entire orthodromic train, this blocker did not uncover significant differences between the two BPs. These results suggest that, although the higher density of 4-AP-sensitive channels in the distal axon may have an important role in preventing unintended firing and can modulate AP waveform and firing pattern, these channels do not interfere with orthodromic AP conduction.

Crayfish axons as a model system for the study of axon excitability.

Historically, the crayfish opener preparation has been used mainly for the study of short- and long-term synaptic plasticity (Atwood and Karunanithi 2002; Atwood and Wojtowicz 1986; Bittner 1989; Dudel 1989; Parnas and Parnas 1994; Zucker and Regehr 2002). Muscle fibers controlling the opening of the claw are innervated by one excitor and one inhibitor. These axons run in parallel, branch repeatedly, and synapse onto muscle fibers by strings of varicosities (Florey and Cahill 1982). Both axons and terminals are covered in glia cells (Govind et al. 1995). Conduction of AP in these axons is considered continuous rather than saltatory (Lin 2008). Thus the morphology and physiology of opener axons are comparable with that of unmyelinated axons in other nervous systems. The accessibility of crayfish axons to sharp electrodes and imaging provide a unique opportunity for detailed analysis of axonal excitability.

Potassium channels in opener axons have been examined previously, with a predominant focus on using Ca2+-activated K+ channels as an indicator of Ca2+ dynamics around release sites (Blundon et al. 1995; Sivaramakrishnan et al. 1991). It has been shown that 4-AP effectively increases presynaptic AP amplitude, duration, and release. In contrast, tetraethylammonium (TEA) alone, even at 20 mM, only slightly broadens the duration of an isolated AP (Sivaramakrishnan et al. 1991). However, these studies focused on excitor axons that, due to muscle contraction, precluded the use of the high-frequency or prolonged stimulation necessary for more detailed analysis of membrane excitability. Experiments reported here represent a logical extension of the previous studies and were inspired by advances in the study of localized channel distribution in dendrites, soma, and initial segments in mammalian preparations.

There is currently no molecular characterization of K+ channel subtypes in crayfish. Thus results presented in this report cannot be easily mapped onto the K+ channel families found in mammals. The main effects of 4-AP at 200 μM, a large reduction in ITH and VTH and a qualitative changed in firing patterns at the 2° BP, are consistent with the functions of Kv1 channels reported in mammalian preparations (Johnston et al. 2010). However, α-dendrotoxin did not have any effect on crayfish axon. Nevertheless, since access to glia-encased axons could potentially account for the absence of a α-dendrotoxin effect, the possibility that 4-AP targets Kv1-like channels has not been completely ruled out. Alternative techniques such as RNA probes or immunocytochemistry will be needed to determine the molecular identities of K+ channels in the crayfish axons.

Subthreshold membrane excitability is not uniform in a branching axon.

One of the main conclusions in this report is that the distal axon exhibits stronger subthreshold K+ channel activity than the proximal axon. Although local Rin at the 2° BP was nearly twice that at the 1° BP, ITH was the same at both BPs in control saline. This finding suggests that a relatively large ITH was needed at the 2° BP to counter the strong subthreshold K+ conductance. 4-AP preferentially blocked the distal K+ conductance and resulted in a larger reduction in ITH there than at the 1° BP. It should be noted that 4-AP did not completely equalize VTH at the two BPs: this parameter was still more depolarized at the 2° BP than at the 1° BP in 4-AP. Nonuniform distribution of 4-AP-insensitive K+ channels or Na+ conductance may account for the difference that remained.

The higher level of subthreshold K+ channel activity at the 2° BP can explain an apparent anomaly shown in Fig. 4. Whereas 4-AP lowered VTH at both BPs for local TECC, distant TECC often revealed a higher VTH for AP initiated at the 2° BP in 4-AP (Fig. 4B3, inset). This anomaly was observed in four out of six preparations studied with distant TECC. A possible explanation is that subthreshold depolarization preceding the first AP fired by ITH at the 2° BP would activate significant 4-AP-sensitive K+ conductance locally and severely reduce the length constant in antidromic direction. The antidromic attenuation of subthreshold signals would be reduced in 4-AP and give rise to VTH that appears to be higher than VTH recorded in control saline. A lower density of 4-AP-sensitive channels at the 1° BP would result in a small or no change in length constant in orthodromic direction by 4-AP. This prediction is consistent with the finding that the hyperpolarizing shift of VTH in 4-AP was observed in both distant and local TECC for AP initiated at the 1° BP.

Distinct firing patterns at the 2° BP.

The firing pattern at the 2° BP is variable, and the source of this variability is unclear. Two main patterns evoked by ITH were: 1) brief bursting with strong adaptation (Fig. 4); and 2) high-frequency and nonadapting firing (Figs. 1 and 3).

Principle neurons in the medial nucleus of the trapezoid body and their presynaptic terminals have been shown to fire one or two APs in response to ITH and suprathreshold current steps (Brew and Forsythe 1995; Dodson et al. 2003; Ishikawa et al. 2003; Nakamura and Takahashi 2007). Dendrotoxin converted this phasic firing pattern to one that fired continuously. Similar behavior has also been observed in somatic current-clamp studies in cortical and sensory neurons (Guan et al. 2007; Hsiao et al. 2009). The effects of specific Kv1 blockers resembled that of 4-AP on the 2° BP of crayfish axons shown in this report. Although the brief burst shown here may involve both Na+ and K+ channels, the effect of 4-AP suggests that 4-AP-sensitive K+ channels play a dominant role in shaping the phasic firing pattern shown here.

The abrupt onset of high-frequency firing, with minimal adaptation, at ITH has also been observed in the mammalian CNS and classified as type 2 dynamics (Tateno et al. 2004). The underlying mechanism of type 2 firing was assumed to involve both inward and outward current. Although type 2 firing has not been specifically associated with Kv1 channels, the striking effect of 4-AP, in converting the discontinuous AP#-I curve to a smooth rising one, suggests the importance of 4-AP-sensitive K+, low-voltage-activated channels on type 2 firing patterns. It should be noted that although this report has emphasized the impact of 4-AP-sensitive K+ channels in distal axons, the increase in AP amplitude and duration at the 1° BP was also large. It remains unclear whether these effects at the 1° BP were due to the presence of 4-AP-sensitive channels there or due to antidromic spread of signals from distal branches and terminals.

Functionally, the presence of strong subthreshold K+ channel activity could be relevant to several situations. For example, injuries may occur at crustacean claws. The presence of K+ channels that are active in a subthreshold range would be a valuable mechanism for preventing unwanted firing and transmitter release. In addition, ligand-gated channels and G protein-coupled receptors at presynaptic terminals are known to modulate resting Vm (Beaumont and Zucker 2000; Fischer and Parnas 1996; MacDermott et al. 1999; Schramm and Dudel 1997). In this context, the 4-AP-sensitive channels in the distal axon could be essential for limiting this modulation to a subthreshold range.

Effect of 4-AP on a high-frequency orthodromic train.

One notable feature of this report is that although membrane responses evoked by current steps have provided clear evidence for local variation in membrane excitability, this spatial variation did not translate into an obvious distinction between orthodromic trains recorded at the proximal and distal axons. This apparent dilemma perhaps makes sense in light of the fact that a large charging current from the proximal axon rapidly initiates AP from a resting level. Thus the 4-AP-sensitive K+ channels could shape AP waveform but are unlikely to interfere with forward AP conduction. Since 4-AP at 200 μM is known to block both Kv1 and Kv3 channels (Johnston et al. 2010; Rudy and McBain 2001), the increase in AP amplitude and waveform reported here could be due to the block of both channel types. In preliminary studies, TEA at 1 mM, which has been shown to block Kv3 channels (Erisir et al. 1999), had minimal impact on AP waveform (data not shown). Therefore, 4-AP-sensitive, Kv1-like channels remain the most likely candidate for controlling the AP waveform observed in this report.

Modulation of AP duration during a high-frequency train has been reported in both soma and synaptic terminals. In hippocampal mossy fiber varicosities, both Kv1 and Kv3 channels have been shown to underlie an in-train rise in AP duration (Alle et al. 2011; Geiger and Jonas 2000). In the present study, there was a ∼20% increase in AP duration at the end of a 100-Hz, 500-ms train (Fig. 8). An in-train rise of such magnitude should have a significant impact on the magnitude of transmitter release and synaptic delay (Boudkkazi et al. 2011). The importance of 4-AP-sensitive channels in this modulation is indicated by the flattened in-train rise of AP half-width observed in 4-AP (Fig. 8). Thus it is possible that some of the 4-AP-sensitive channels play a role in controlling AP duration during a high-frequency train. Since 4-AP at 200 μM is known to block both Kv1- and Kv3-type K+ channels (Johnston et al. 2010), the identity of the channel underlying the in-train rise remains to be clarified.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

J.-W.L. conception and design of research; J.-W.L. performed experiments; J.-W.L. analyzed data; J.-W.L. interpreted results of experiments; J.-W.L. prepared figures; J.-W.L. drafted manuscript; J.-W.L. edited and revised manuscript; J.-W.L. approved final version of manuscript.

ACKNOWLEDGMENTS

I thank Nicky Schweitzer and Bernardo Rudy for editing and commenting on this manuscript, respectively.

REFERENCES

View Abstract