Developmental changes in expression of ion currents accompany maturation of locomotor pattern in frog tadpoles

Abstract

1. The K⁺ currents of spinal neurons acutely dissociated from Xenopus larvae were studied and compared with those of neurons dissociated from Xenopus embryos.

2. The density of total outward current in the larval and embryonic neurons remained the same from stage 37/38 to stage 42.

3. Almost all neurons at stage 42 expressed a fast activating Ca²⁺-dependent K⁺ current (I_KCa) that was largely absent from embryonic neurons. Whereas I_KCa became larger and more prevalent during development, the delayed rectifier K⁺ currents were down-regulated.

4. About 53% of I_KCa was selectively blocked by iberiotoxin which had no effect on the delayed rectifier K⁺ currents or the K⁺ currents of embryonic neurons.

5. The firing properties of neurons isolated from embryos were unchanged by iberiotoxin. However, the toxin greatly increased the frequency of firing in larval neurons.

6. Iberiotoxin extended the duration of ventral root bursts during fictive swimming in larvae at stages 41 and 42 but had no effect at stage 40. The progressive expression of I_KCa thus contributed to burst termination.

7. We have found that changes in expression of outward current closely correlate with the maturation of the motor pattern during development. At a time when the motor pattern has a need for a burst-terminating mechanism, the larval neurons express a channel with properties appropriate for such a role.

In a variety of animals the capacity to generate both locomotor and respiratory patterns arises early in embryonic development (Sillar, Wedderburn & Simmers, 1991; Ho & O'Donovan, 1993; Sillar, 1994; Casasnovas & Meyrand, 1995; Paton & Richter, 1995; Ho, 1997). As development proceeds, these motor patterns gradually mature. Although changes to network circuitry (Antal, Berki, Horvath & O'Donovan, 1994; Paton, Ramirez & Richter, 1994; Berki, O'Donovan & Antal, 1995) and the intrinsic properties of the neurons (Sillar, Simmers & Wedderburn, 1992) accompany this maturation, causal links between changes at cellular and network levels and maturation of motor patterns have been harder to establish.

In Xenopus, the embryonic motor pattern is characterized by neurons firing only a single spike per cycle. This changes to the larval pattern (stage 40 onwards) in which neurons fire bursts of spikes (Sillar et al. 1991). Although descending serotonergic reticulospinal fibres have been proposed as a trigger for this change (Sillar, Woolston & Wedderburn, 1995), the mechanisms underlying the change in the motor pattern remain unknown. Nevertheless the observation that neurons individually become more excitable (Sillar et al. 1992) suggests that changes in the expression of ion channels may be important. Previous studies have extensively characterized the ionic currents of embryonic spinal neurons at stage 37/38 (Dale, 1993; Dale, 1995a;Wall & Dale, 1995). We have extended this analysis to the K⁺ currents of larval neurons to see whether new channels are expressed and whether embryonic channels may be down-regulated.

We have found that a process of ion current substitution accompanies the development of the larval motor pattern. While the total amount of depolarization-activated K⁺ current remains the same in embryonic and larval neurons, the proportion that is Ca²⁺-independent drops with development, and a new fast-activating Ca²⁺-dependent K⁺ current, I_KCa, appears in the larval neurons. By using a specific blocker of I_KCa, iberiotoxin, we have shown that this current plays a role both in limiting the repetitive firing of the larval neurons and in the termination of bursts during the larval swimming pattern.

METHODS

All embryos and larvae were staged according to the external criteria of Nieuwkoop & Faber (1956).

Preparation of neurons

Acutely isolated neurons from the spinal cords of Xenopus embryos and larvae were obtained using methods previously described (Dale, 1991; Dale, 1995a). In brief, embryos were anaesthetized by immersion in 3-aminobenzoic acid ethyl ester (MS-222, 0.5 mg ml⁻¹; Sigma), and their spinal cords carefully dissected free. A combination of brief enzymatic treatment (30 s to 1 min in 8 mg ml⁻¹ pronase E, Sigma) and mechanical dissociation reliably gave a good yield of healthy neurons. These were plated onto poly-D-lysine-coated dishes. Neurons derived from all developmental stages were left for about 1 h before recording commenced.

Patch clamp recordings from isolated neurons

Whole cell patch clamp recordings were made from neurons acutely isolated from the Xenopus embryo and larval spinal cord (Dale, 1991) that had multipolar and commissural neuron-like morphologies and would thus have constituted a mixture of motoneurons and excitatory and inhibitory premotor interneurons (Dale, 1991). All of these are involved in motor pattern generation.

Electrodes were fabricated using a Sutter Instruments P97 puller from glass obtained from World Precision Instruments (TW 150F). A List LM_PC amplifier together with a DT2831 interface (Data Translation) were used to record and digitize (at 40 kHz) the voltage and current records. The recordings had access resistances ranging from 4-12 MΩ. Between 70 and 85 % of this access resistance was compensated for electronically.

The external medium contained (mM): 115 NaCl, 2.4 NaHCO₃, 3 KCl, 10 CaCl₂, 1 MgCl₂, 10 Hepes adjusted to pH 7.4 at room temperature (18-22°C); while the pipette solution contained (mM): 100 KCH₃SO₃, 5 KCl, 5 MgCl₂, 20 Hepes, 5 ATP and 2 BAPTA, adjusted to pH 7.4 at room temperature. To block the Ca²⁺-dependent currents, the Ca²⁺ in the recording medium was replaced by an equimolar amount of Mg²⁺. Drugs were applied via a multibarrelled microperfusion system. Iberiotoxin was obtained from Research Biochemicals Inc. Currents were measured from a holding potential of -50 mV.

Extracellular recordings

Xenopus embryos and larvae were prepared for extracellular ventral root recordings in accordance with the UK Animals (Scientific Procedures) Act, 1986, using previously described methods (Sillar et al. 1991; Dale, 1995b). Drug access was facilitated by bilateral removal of the rostral myotomes and loosening the dorsal attachment of the remaining myotomes. The spinal cord was transected at the first or second post-otic myotome. The preparation was placed in a small bath (0.5 ml volume) and continuously superperfused with saline containing (mM): 115 NaCl, 3 KCl, 1 MgCl₂, 2 CaCl₂, 2.4 NaHCO₃, 10 Hepes, adjusted to pH 7.4 at room temperature. Drugs were applied by superfusion of the bath. Electrical stimuli to the skin were used to trigger swimming with an interval of 3 min between the beginning and end of consecutive episodes.

Analysis of burst duration

By means of a Data Translation DT31EZ, swimming episodes were digitized and stored on a computer. The ventral root activity was then rectified and integrated with a 2 ms period. Changes in burst duration were assessed in two ways. The first was a threshold crossing method in which a burst was defined as the time from the first crossing above threshold to the last crossing below threshold. The second method involved averaging the rectified and integrated ventral root bursts. Using the start of each burst as a reference point, the computer averaged a stretch of activity on either side of the burst. Both methods gave similar results, although the threshold crossing method tended to underestimate the duration of the bursts.

Data are given as means ±s.e.m. where n indicates the number of neurons.

RESULTS

Larval neurons express a new I_KCa

The firing properties of Xenopus neurons change with development: larval neurons are capable of more repetitive firing than their embryonic counterparts (Sillar et al. 1992). We therefore examined the K⁺ currents in acutely isolated spinal neurons of Xenopus embryos and larvae. To casual inspection, the currents in control appear very similar (Fig. 1A and B). To account for the fact that larval neurons are smaller than their embryonic counterparts, we normalized current density to cell capacitance. The density of the total outward currents was the same in embryonic and larval neurons (Fig. 1C, open bars).

Figure 1. The delayed rectifier K⁺ currents are down-regulated and I_KCa is newly expressed during development.

A, the voltage-gated currents, elicited from a holding potential of -50 mV by 10 mV steps from -40 to +50 mV, of a neuron acutely dissociated from a stage 37/38 embryo recorded in (from the top) control saline, zero Ca²⁺ saline, wash and the control minus zero Ca²⁺ difference currents. There is no Ca²⁺-sensitive current. B, voltage-gated currents (evoked by same voltage protocol as in A for a neuron acutely dissociated from a stage 42 larva (traces equivalent to those in A). There is a large Ca²⁺-sensitive current in this neuron which has both a transient and a sustained component. C, the total and Ca²⁺-independent outward current density (normalized to cell capacitance) shown for neurons from stages 37/38 and from stages 41 and 42 combined. Note that although the total current density remains the same, the density of the Ca²⁺-independent currents declines (**P < 0.01 stages 37/38 vs. 41/42, Student's t test). D, the development of I_KCa. Both the proportion of neurons that possessed the current and the density of the current increased with developmental stage (**P < 0.01 stages 39/40 vs. 37/38, and stages 41/42 vs. 39/40, Student's t test for current density and a χ² contingency test for percentage of neurons)

However, when the Ca²⁺ dependence of outward currents was examined, a dramatic difference between larval and embryonic neurons was seen. Whereas very few embryonic neurons possessed I_KCa, virtually all neurons at stage 42 exhibited this current (compare Fig. 1A and B). Thus at stage 42, the delayed rectifier K⁺ currents have been down-regulated (Fig. 1C, filled bars), and a new current, I_KCa, has been expressed which constitutes 29 ± 2.5 % (no. of neurons, n= 25) of the total outward current. By looking at a developmental sequence from stage 37/38 to stage 42, we found that both the density of I_KCa in individual neurons and the proportion of neurons expressing this current increased (Fig. 1D). The I_KCa is kinetically complex having both transient and sustained components and may be similar to the currents previously described in cultured Xenopus neurons (Blair & Dionne, 1985; Ribera & Spitzer, 1987; O'Dowd, Ribera & Spitzer, 1988).

To check that the I_KCa was also present at more physiological levels of Ca²⁺, we compared the Ca²⁺ sensitivity of outward currents at 2 and 10 mM external Ca²⁺. The amount of I_KCa remaining at the lower level of 2 mM was 64 ± 6 % (n= 4) of that present at 10 mM Ca²⁺. This means that I_KCa is able to contribute to neural function under physiological conditions.

The new I_KCa is partially blocked by iberiotoxin

To test the functional roles of the newly expressed I_KCa, we used iberiotoxin, a specific blocker of the BK Ca²⁺-dependent channels (Galvez, Gimenez-Gallego, Reuben, Katz, Kaczowrovski & Garcia, 1990). We found that iberiotoxin, at doses from 10 to 100 nM, partially and selectively blocked I_KCa in larval neurons in a non-voltage-dependent manner (Fig. 2A, B and D). Consistent with their being generated by a large conductance channel, the iberiotoxin difference currents were very noisy (Fig. 2B). We therefore suggest that the I_KCa is an example of a BK current. By comparing the effects of iberiotoxin with those of zero Ca²⁺ in the same cell (Fig. 2B), we found that the iberiotoxin-sensitive current comprised 53 ± 10 % (n= 9) of the total Ca²⁺-dependent K⁺ current.

Figure 2. The new Ca²⁺-dependent K⁺ current is partially blocked by iberiotoxin.

A, the effect of 100 nM iberiotoxin (IbTx) on the outward currents evoked in a stage 42 neuron by 10 mV steps from +10 to +80 mV from a holding potential of -50 mV. The total current is reduced at all potentials (i) and the percentage block (relative to the current in control) is constant at all potentials (ii) suggesting that the block is non-voltage-dependent (n= 6). B, a comparison of the total I_KCa with its IbTx-sensitive and -insensitive components (control minus zero Ca²⁺, control minus 100 nM IbTx and IbTx minus zero Ca²⁺ difference currents, respectively). These were evoked in the same neuron, from a holding potential of -50 mV, by 10 mV steps from -40 to +40 mV. The currents are characterized by large noise, suggesting that they are carried by relatively few high conductance channels. The slowly activating component of the IbTx-insensitive current is almost certainly the SK current that is present in these neurons. C, the iberiotoxin-sensitive current shows the same development of expression as the I_KCa. Iberiotoxin (100 nM) had no significant effect on the embryonic stage 37/38 neurons, but significantly blocked neurons from stages 41 and 42 (P < 0.01, Student's t test). D, the dose-response for iberiotoxin plotted together with the best-fitting Hill equation (line). The IC₅₀ was 1.1 ± 0.6 nM (n= 4).

To check that iberiotoxin had no effect on the delayed rectifier and other non-Ca²⁺-dependent K⁺ currents, we blocked the Ca²⁺-dependent currents by either removing external Ca²⁺ or applying Cd²⁺. Under these conditions, iberiotoxin had no effect on the outward currents (mean change, 1.0 ± 0.9 %, n= 6). As might be expected from the development of the I_KCa, we found that iberiotoxin had very little effect on the outward currents of the embryonic neurons (Fig. 2C).

The role of I_KCa in control of repetitive firing and motor pattern generation

We next examined the effects of iberiotoxin on the repetitive firing characteristics of acutely isolated embryonic and larval neurons. In embryonic neurons, iberiotoxin had no effect on the number of spikes evoked by a series of increasing current pulses, the threshold current, or the frequency of firing (Fig. 3A, C and D, respectively). However, in larval neurons, iberiotoxin increased the number of spikes evoked by current pulses, reduced the current threshold, and increased the frequency of firing (Fig. 3B, C and D, respectively). Clearly the newly expressed I_KCa plays an important role in controlling the excitability of the larval neurons.

Figure 3. I_KCa controls excitability of larval but not embryonic neurons.

A, current clamp recording from an acutely isolated embryonic neuron (stage 37/38) showing response to an ascending series of current pulses. Iberiotoxin (IbTx; 100 nM) had no effect on the firing in response to current injection. B, similar recordings from a neuron isolated from a stage 42 larva showed that blockade of the I_KCa by 100 nM iberiotoxin both reduced the threshold and increased the number of spikes evoked by current injection. C, summary data for the effects of 100 nM iberiotoxin on the mean number of spikes evoked during the sequence of current pulses shown in A and B (left) and the threshold current for firing (right). D, 100 nM iberiotoxin had no effect on the frequency of firing in embryonic neurons (n= 5), but enabled larval neurons to fire at a significantly higher rate (n= 6). Frequency of firing was calculated as the inverse of the first interspike interval.

We also examined whether iberiotoxin changed spike width. At stage 42 the time from the peak of the action potential to half-repolarization for the first spike in a train was 1.2 ± 0.2 ms in control and 1.3 ± 0.2 ms in iberiotoxin (n= 9, P < 0.01). Although this small change is statistically significant it is probably of no functional importance. At stage 37/38 there was no significant change in spike width (1.3 ± 0.2 ms and 1.3 ± 0.2 ms for control and iberiotoxin, respectively, n= 6).

To test the role of the I_KCa in motor pattern generation, we applied iberiotoxin to the intact larva while monitoring ventral root activity. At stage 40, we found that iberiotoxin had no significant effect on burst duration or cycle period of the ventral root activity (Table 1, Fig. 4A). However, at both stage 41 and stage 42, despite blocking only around half of the total I_KCa, iberiotoxin significantly increased burst duration by 24-33 % (Table 1, Fig. 4B and C). As iberiotoxin had no significant effect on cycle period at stages 41 and 42, we conclude that the I_KCa contributes to burst termination during swimming and that block of this current not only increases absolute burst duration but also the proportion of the motor cycle that each burst lasts (Table 1).

Table 1.

The effect of 50 nM iberiotoxin on the mean duration of ventral root bursts during swimming at different stages of larval development expressed both as an absolute value and as a proportion of cycle period

	Burst duration (ms)				Burst duration (% cycle)
Stage	Control	Iberiotoxin	n	P	Control	Iberiotoxin	n	P
40	8.8 ± 1.2	8.5 ± 1.3	5	—	15.5 ± 2.1	14.1 ± 2.4	5	—
41	13.2 ± 1.5	17.5 ± 2.0	6	< 0.05	18.0 ± 2.2	23.4 ± 2.3	6	< 0.01
42	13.3 ± 0.7	16.6 ± 0.8	5	< 0.01	26.3 ± 1.8	29.8 ± 1.0	5	< 0.05

Burst duration was measured by the threshold crossing method. Each value is shown as mean ±s.e.m., Student's paired t test was used to assess changes in burst duration.

Figure 4. I_KCa contributes to termination of ventral root bursts during swimming in larvae.

A, in a stage 40 larva, iberiotoxin (IbTx, 50 nM) had no effect on the ventral root activity (shown as rectified and integrated traces). The averaged ventral bursts in control and iberiotoxin (right) almost exactly overlapped (fine line shows averaged ventral root burst in iberiotoxin). In a stage 41 (B) and stage 42 (C) larva, 50 nM iberiotoxin increases the burst duration during swimming which can be more easily seen in the averaged ventral root bursts (shown to the right). All averaged ventral root bursts were obtained from two swimming episodes in control and iberiotoxin.

DISCUSSION

Our results show that as the motor pattern develops from the embryonic to the larval pattern, the expression of ion currents changes. The delayed rectifier currents are down-regulated and are partially replaced by the I_KCa. In earlier work, a similar transition in the Ca²⁺ sensitivity of K⁺ currents in cultured Xenopus neurons was interpreted as an alteration of channel structure to endow pre-existing K⁺ channels with Ca²⁺ sensitivity (Blair & Dionne, 1985). However, in the light of more recent knowledge about the genes that encode Ca²⁺-dependent K⁺ channels (Pallanck & Ganetzky, 1994), we interpret the acquisition of Ca²⁺- and iberiotoxin sensitivity as resulting from a switch in gene expression and synthesis of new BK channel proteins. Given the massive conductance of BK channels, the newly expressed channels may not be very numerous: for a whole cell conductance of 8 nS only around forty to eighty channels would be needed.

Since the fast and slow components of the delayed rectifier limit the rate of firing of the embryonic neurons and contribute to spike threshold (Dale, 1995a, b), down-regulation of these currents may increase the excitability of the larval neurons and contribute to the developmental increase in burst duration during swimming. Nevertheless other alterations to the network circuitry that also increase burst duration independently of ion channel expression are likely.

One effect of the switch in the expression of currents is to endow a significant proportion of the total outward current (around 30 %) with a dependence on the levels of intracellular Ca²⁺. This will accumulate in a cyclical fashion during rhythmic motor activity, owing to voltage activation of Ca²⁺ channels and synaptic activation of ligand-gated channels, most notably the NMDA receptor. Thus I_KCa is likely to exert a stronger repolarizing influence at the end of the motor cycle compared with the beginning (before significant accumulation of Ca²⁺). When a motor pattern involves alternating burst activity between mutually antagonistic motor half-centres, mechanisms must be present to terminate the bursts in each half-centre so that alternation ensues rather than unceasing, high frequency firing in one dominant half-centre. Ca²⁺-dependent K⁺ currents are thought to contribute to such burst termination in the lamprey (El Manira, Tegnér & Grillner, 1994). Our results show very clearly that as the motor pattern matures in Xenopus and a need for burst terminating mechanisms arises (at stages 41 and 42), the delayed rectifier K⁺ currents are partially substituted by Ca²⁺-dependent K⁺ currents that specifically meet this new requirement. The changes in ionic currents, and by implication, ion channel expression are thus closely matched to the requirements of the motor pattern.

Interestingly, Xenopus embryos express a Ca²⁺-dependent K⁺ current (probably an SK channel; Wall & Dale, 1995) that activates some three orders of magnitude more slowly than I_KCa and is sensitive to apamin. This SK current is present also in the larva and is therefore not subject to the same developmental regulation as I_KCa (Q.-Q. Sun & N. Dale, unpublished observations). In the embryo the SK current activates too slowly to play a role in the cycle-by-cycle regulation of the motor pattern. Instead it slowly accumulates over many cycles and contributes to the termination of swimming episodes (Wall & Dale, 1995). The role of this current in the larva has not been examined; however, since its kinetics of activation remain slow compared with the duration of a single locomotor cycle, a role similar to that demonstrated in the embryo is plausible.

The identification of a current whose expression is so closely matched to the maturation of the motor pattern provides a new experimental approach to identifying the developmental cues that underlie this maturation. Any factors that change expression of this current in single neurons will be key candidates for those underlying the triggering of a wider developmental programme that results in maturation of the motor pattern. Interestingly, cultured Xenopus neurons (isolated from the neural plate and allowed to develop in vitro), express Ca²⁺-sensitive K⁺ currents within a few hours of dissociation (Blair & Dionne, 1985; O'Dowd et al. 1988). This is probably 2-3 days in advance of when I_KCa is expressed in vivo. If the currents expressed in the cultured neurons are the same as those in vivo, this implies the existence of extrinsic mechanisms that delay expression of the new channels until an appropriate stage of larval development.