INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Around stage 3 of the 18 stages of metamorphic adult development of the sphinx moth, Manduca sexta, the arrival of antennal sensory receptor axons in the antennal lobes (ALs) of the brain triggers the formation of subunits of synaptic neuropil, called glomeruli (reviewed by Oland et al. 1998; Rössler et al. 2000). Local AL interneurons, projection (output) neurons, and centrifugal neurons that project to the ALs from elsewhere in the brain undergo extensive growth and restructuring as glomeruli develop in a lateral-to-medial wave across the AL neuropil (reviewed by Hildebrand et al. 1997; Oland and Tolbert 1996). Recordings from AL neurons in vitro and in situ in semi-intact brain preparations reveal dramatic changes in the electrical properties and response characteristics of the cells during this period (see accompanying paper). Action potentials elicited from Manduca AL neurons at the onset of metamorphosis are small in amplitude and long in duration, but as metamorphosis proceeds they become larger in amplitude and shorter in duration, and cell-type-specific differences in the response characteristics of the cells begin to emerge. Changes in whole cell current profiles that coincide with developmental changes in spike waveform have been described, but specific currents contributing to these changes have yet to be identified.

This study describes voltage-gated and Ca²⁺-dependent currents in developing Manduca AL neurons and compares their properties early (stage 5) and late (stage 14) in metamorphosis. Two distinct cell types are examined in vitro: Rick Rack (RR) neurons and Proximal Branching (PB) neurons, identified elsewhere as putative local AL interneurons and projection (output) neurons, respectively (Oland and Hayashi 1993; Oland et al. 1992). Pharmacological manipulations, together with voltage inactivation and subtraction protocols, have enabled us to identify six distinct current types in these cells. Changes in the density of Na⁺, Ca²⁺, and K⁺ currents during adult metamorphosis are described, and currents that are likely to contribute to developmental changes in action potential waveform in Manduca AL neurons have been identified. The results provide an important starting point for investigations into the development of electrical excitability in Manduca AL neurons, and its role in the development of primary olfactory centers in the brain of the moth.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Manduca sexta (Lepidoptera: Sphingidae) were reared on an artificial diet (modified from Bell and Joachim 1976) and maintained at 25°C under a long-day photoperiod regimen (17 h light/7 h dark). Larvae hatch from eggs and pass through five larval instars before undergoing metamorphosis from larva to pupa to adult. Metamorphic adult development in the moth can be divided into 18 stages, each of which lasts approximately 1 day and is accompanied by well-defined changes in pupal morphology (Sanes and Hildebrand 1976a,b; Tolbert et al. 1983). AL neurons from pupae at stages 2-16 were examined in this study, but the study focuses on a comparison of currents recorded in cells at pupal stages 5 and 14. In vitro analysis is used to examine two morphologically distinct cell types: RR neurons, the out growing neurites of which have a striking zig-zag appearance in vitro (Mercer et al. 1996a; Oland and Hayashi 1993), and PB neurons, which possess distinctive clusters of arbors that arise from the primary neurite close to the cell body (Hayashi and Hildebrand 1990; Mercer et al. 1996a; Oland and Hayashi 1993). PB neurons were derived from the median group of AL neurons, a cluster of cells that contains projection (output) neurons, but that lacks local neurons (Homberg et al. 1989). RR neurons were derived from either, the lateral cell cluster, or from dissociations of whole ALs. To avoid the possibility that small differences in culture medium or developmental stage may affect the results of comparisons between RR neurons and PB neurons, only cells examined in parallel from the same sets of animals were used to compare current densities in the two cell types.

Changes in the whole cell current profiles of these neurons during adult metamorphosis have been described elsewhere (see accompanying paper). In RR neurons, these changes are represented by four distinct types of whole cell current profile (types 1-4), whereas for PB neurons, three types of current profile have been described. Approximately 15% of the stage 5 RR neurons examined in this study exhibited RR-type 1 whole cell current profiles, but the majority exhibited profiles of types 2 and 3. All stage 14 RR neurons exhibited RR-type 4 current profiles. In the case of PB neurons, stage 5 cells displayed PB-type 1 or type 2 profiles, whereas only cells with PB-type 3 profiles were included in the analysis of stage 14 PB neurons. Descriptions of whole cell current profiles are provided in the accompanying paper.

Cell culture

Cells were dispersed and cultured according to methods described by Hayashi and Hildebrand (1990). Briefly, brains were removed from cold-anesthetized pupae and placed into sterile culture saline. ALs were transferred into Hanks' Ca²⁺- and Mg²⁺-free buffered salt solution containing 0.5 mg/ml collagenase and 2 mg/ml dispase for 2 min at 37°C to dissociate the tissue, which was then dispersed by trituration with a fire-polished Pasteur pipette. Enzyme treatment was terminated by centrifugation, first through culture saline and then through culture medium (for recipes, see accompanying paper). Dissociated cells were allowed to settle and adhere to the surface of culture dishes coated with Concanavalin A (200 µg/ml) and laminin (2 µg/ml). Cultures were placed in a humidified incubator at 26°C and maintained for 5-6 days. Only cells without contacts to other neurons were selected for recording.

Electrophysiological recording

Whole cell voltage-clamp recording techniques (Hamill et al. 1981) were used to examine voltage-activated currents in neurons from pupae at stages 2-16 of the 18 stages of metamorphic adult development. Electrodes with resistances of 1-2 M were made from borosilicate glass (VWR Scientific, West Chester, CA) and filled with a solution containing (in mM) 150 K-aspartate, 5 NaCl, 2 MgCl₂, 1 CaCl₂, 11 EGTA, and 10 HEPES, adjusted to pH 7 and 330 mOsm. Cells were viewed under an Olympus IMT-2 inverted microscope equipped with Hoffman modulation contrast optics. Prior to recording, culture medium was replaced with insect saline containing (in mM) 100 NaCl, 4 KCl, 6 CaCl₂, 5 D-glucose, and 10 HEPES (pH 7), adjusted to 360 mOsm with mannitol to facilitate the formation of high resistance (gigaohm) seals. Saline levels in the perfusion chamber were maintained as low as possible to reduce electrode capacitance transients and improve compensation. Junction potentials were nullified prior to seal formation. To obtain whole cell recordings, light suction and brief high-voltage pulses were used to rupture the cell membrane beneath the recording electrode. Recordings were made using an AxoPatch 1B amplifier (Axon Instruments), and data were acquired and analyzed using pClamp6 software (Axon Instruments, version 6.02) run on an i30486 microcomputer. Membrane currents were sampled at intervals of 100 µs and were filtered at 2 kHz using a low-pass 4-pole Bessel filter. Cell capacitance was determined from the capacitative current elicited by a 10-mV depolarizing voltage step following establishment of the whole cell clamp. The whole cell capacitative charge (calculated as the integral of the whole cell capacitative current transient) was divided by the applied voltage to obtain cell capacitance. Membrane capacitance measurements of 86 ± 25 (SE) pF and 63 ± 27 pF were recorded for RR neurons (n = 133) and PB neurons (n = 89), respectively. No stage-related differences in cell capacitance were observed. Linear leakage currents and residual capacitance artifacts were subtracted using a P/4 protocol included in the data acquisition software (Axon Instruments). All recordings were performed at room temperature (20-22°C).

Series resistance compensation was applied, where possible 80%. However, large series resistance errors associated with the measurement of Na⁺ currents in mature neurons made it difficult to routinely apply adequate compensation. The largest currents, which were recorded in stage 14 RR neurons, approached 7.5 nA in some cells. Electrode resistance was low (1-2 M, see above), and series resistance was estimated to be <4 M, suggesting residual errors (with 70% compensation) of 9 mV in these cells. For this reason, Na⁺ currents were examined in detail only in immature neurons at pupal stage 5, where residual errors were estimated to be 5 mV or less. For all isolated currents other than I_Na, residual errors were estimated to be <5 mV. Series resistance errors will result in underestimates of developmental increases in current amplitudes and activation rates. No correction has been made to plots showing current-voltage (I-V) relations, or to estimates of time to peak current.

Current isolation

Components of the whole cell current profile were isolated using routine pharmacological techniques and subtraction protocols described previously for Manduca neurons (Duch and Levine 2000; Grünewald and Levine 1998; Hayashi and Hildebrand 1990; Hayashi and Levine 1992; Kloppenburg et al. 1999a; Mercer et al. 1996b; Zufall et al. 1991). Na⁺ currents were blocked with tetrodotoxin (TTX, 10⁸ M) and Ca²⁺ currents with 5 × 10⁴ M CdCl₂. K⁺ currents, collectively, were blocked by replacing KAsp in the electrode with CsCl. A-type K⁺ currents were blocked with 5 × 10³ M 4-aminopyridine (4-AP), and 3 × 10² M tetraethylammonium chloride (TEA) was used to block non-A-type K⁺ currents. Ca²⁺-activated currents were blocked indirectly, by blocking Ca²⁺ currents with Cd²⁺. In Na⁺-free solutions, NaCl was replaced with Tris-Cl. To normalize for differences in cell size, peak current amplitudes (in pA) were divided by cell capacitance (in pF). Measurements of activation threshold, time-to-peak current, and maximum current density were recorded for each cell. The significance of differences in the properties of neurons at developmental stages 5 and 14 was tested by two-tailed Student's t-tests. For comparison of multiple group means, one-way ANOVA was performed using Minitab 8.2 software (State College, PA, 1991); post hoc comparisons were performed using Tukey tests. A significance level of 0.05 was accepted for all tests. Data are presented as means ± SE.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Inward currents

NA⁺ CURRENTS. To isolate Na⁺ currents in Manduca AL neurons, outward currents were blocked by substituting K⁺ in the pipette solution with Cs⁺ and adding 3 × 10² M TEA and 5 × 10³ M 4-AP to the extracellular solution. Ca²⁺ currents were blocked with 5 × 10⁴ M CdCl₂. Large amplitude Na⁺ currents present in mature neurons made it difficult to obtain adequate series resistance compensation. For this reason, Na⁺ currents were examined in detail only at pupal stage 5, which falls within a narrow time window in which the percentage of RR neurons exhibiting rapidly activating, transient Na⁺ current (I_Na) increases significantly, but current amplitudes are still relatively small (Fig. 1, A1A4). Cells in which no sodium current could be detected (n = 7) were discarded. In 2 of the remaining 11 stage 5 RR neurons, no fast, transient Na⁺ currents could be detected, but a small amplitude sustained inward current was observed (Fig. 1A1). The characteristics of this current have yet to be examined in detail and are not considered further in this report. In all remaining cells (including PB neurons), I_Na activated at voltages between 50 and 45 mV. Stage 5 RR neurons used to examine more closely the activation kinetics of this current were grouped initially into those showing high (I_Na max) and low (I_Na min) density Na⁺ currents. From these groupings, it was apparent that comparisons would be improved by introducing a third category, I_Na mid, in which I_Na densities were intermediate between the two extremes (Fig. 1, A2A4). Normalizing for cell size did not alter the distribution of recordings between these three categories. The large range of Na⁺ current densities found in stage 5 RR neurons is illustrated in Fig. 1B1. If peak current amplitudes at each voltage step are expressed as fractions of the maximum I_Na amplitude observed in each cell (I/I_max, Fig. 1B2), the normalized I-V plots indicate that despite rapid changes in I_Na density around pupal stage 5, the voltage dependence of I_Na remains unchanged. In the three groups of stage 5 RR neurons shown in Fig. 1B, mean times to reach peak current recorded during a depolarizing voltage step from 70 to 10 mV were 1.9 ± 0.5 ms (I_{Na min}, n = 3), 1.2 ± 0.2 ms (I_{Na mid}, n = 4) and 1.5 ± 0.2 ms (I_{Na max}, n = 2). Although I_Na appeared to activate more slowly in cells exhibiting low density currents (Fig. 1A2), differences between the 3 groups in mean time to peak current were not statistically significant (df 8, F = 1.71; P = 0.26).

View larger version (37K): [in this window] [in a new window]   Fig. 1. AC: Na+ currents in stage 5 Rick Rack (RR) neurons. Cells in which no inward current could be identified (n = 7) were discarded. A: in 2 of 11 stage 5 RR neurons, no fast-activating, transient Na+ currents (INa) could be detected, but a small amplitude, sustained inward current was apparent (A1). Those cells in which INa could be identified were divided into 3 groups (see text); cells displaying low (INa min, A2), medium (INa mid, A3), and high (INa max, A4) density Na+ currents. B: properties of Na+ currents. B1: voltage dependence of INa. Peak current density plotted as a function of membrane potential in cells exhibiting low (INa min, , n = 3), medium (INa mid, , n = 4), and high (INa max, , n = 2) density Na+ currents. Each point represents the mean peak current density (±SE). B2: the same data as in B1 expressed as a percentage of the maximum peak current density in each cell. Despite significant differences in current density, the voltage dependence of INa in these cells was similar. Series resistance errors associated with measurements of the large amplitude Na+ currents in stage 14 RR neurons [and Proximal Branching (PB) neurons] made it difficult to apply adequate compensation. For this reason, properties of this current were not examined in these cells. C1: inactivation protocol. To examine INa inactivation, cells (n = 8) were held at a resting potential of 70 mV and stepped to a potential of 15 mV following a conditioning prepulse, as shown. C2: Na+ currents elicited after each prepulse were normalized to the Na+ current evoked when no conditioning prepulse was applied. This value is expressed here as a function of the prepulse potential. As no significant differences in steady-state inactivation between cells exhibiting different INa densities were identified, the data were combined and fitted with a Boltzmann distribution to determine V1/2 (50.34 mV). D: comparisons of mean peak current densities (±SE) in RR neurons and PB neurons early (stage 5, white bars) and late (stage 14, gray bars) in metamorphosis. The significance of differences between groups is indicated by letters above the bars in each graph. Groups that share a letter are not significantly different; groups that do not share a letter differ significantly (P < 0.05). The density of Na+ currents in stage 5 RR neurons was significantly lower than in RR neurons at pupal stage 14. A similar trend in PB neurons was not statistically significant.

CA²⁺ CURRENTS. Prior to the appearance of fast, transient Na⁺ currents, small amplitude action potentials could be elicited from Manduca AL neurons. Preliminary evidence suggests that these are Ca²⁺ spikes (see accompanying paper). To examine Ca²⁺ current densities in the cells, Na⁺ currents were abolished with TTX (10⁸ M), and K⁺ currents were blocked by replacing KAsp in the electrode with CsCl, and adding 4-AP (5 × 10³ M) and TEA (3 × 10² M) to the external medium. I_Ca was identified at all stages of development examined in this study and, as shown elsewhere (Hayashi and Levine 1992; Mercer et al. 1995), I_Ca could be abolished by the addition of 5 × 10⁴ M CdCl₂ to the external medium. Recordings from RR neurons revealed marked differences in the amplitude of Ca²⁺ currents early (Fig. 2A1) and late (Fig. 2A2) in metamorphosis.

View larger version (42K): [in this window] [in a new window]   Fig. 2. Ca2+ currents recorded from RR neurons. A: recordings from RR neurons early (stage 5, A1) and late (stage 14, A2) in metamorphosis. B: properties of Ca2+ currents. B1: voltage dependence of Ica in stage 5 (, n = 8) and stage 14 (, n = 10) RR neurons. Peak density is plotted as a function of membrane potential. Each point represents the mean (±SE) peak current density. B2: the same data as in B1 expressed as a percentage of the maximum peak current amplitude in each cell suggests a similar voltage dependence. C: comparisons of peak current densities (±SE) in RR neurons and PB neurons early (stage 5, white bars) and late (stage 14, gray bars) in metamorphosis. The significance of differences between groups is indicated by letters above the bars in each graph. Groups that share a letter are not significantly different; groups that do not share a letter differ significantly (P < 0.05). Ca2+ current densities in stage 14 RR neurons were significantly higher than those recorded at stage 5. In contrast, there was no significant difference in ICa densities of PB neurons at stages 5 and 14. Ca2+ current densities in neurons examined at stage 14 were significantly higher in RR neurons than in PB neurons. D1: inactivation protocol. To examine ICa inactivation, cells were held at a resting potential of 70 mV and stepped to a potential of 10 mV following a conditioning prepulse, as shown. D2: Ca2+ current inactivation in stage 5 RR neurons. Data were fitted with a Boltzmann distribution to determine V1/2 (42.55 mV).

Contribution of Ca²⁺ currents to spike waveform

RR neurons were used to examine in detail the contribution of Ca²⁺ currents to spike waveform. Brief (20-ms) injections of depolarizing current (0.2-2 nA) were applied to elicit single action potentials from cells early (stage 3, Fig. 3A1 and stage 5, Fig. 3A2) and late (stage 10, Fig. 3A3 and stage 14, Fig. 3A4) in metamorphic adult development. Cells represented in Fig. 3, A1A4, exhibited whole cell current profiles of types 1-4, respectively (see accompanying paper). Figure 3, B1B4, shows the impact on spike waveform of blocking Ca²⁺ currents with Cd²⁺. Early in metamorphosis, depolarizing current pulses failed to elicit action potentials in some RR neurons (Fig. 3A1). The prominent sag to more hyperpolarized potentials that occurred in these cells during the depolarizing current pulse (arrow, Fig. 3A1) was partially blocked by Cd²⁺ (Fig. 3B1). Small-amplitude spikes characteristic of RR neurons early in adult metamorphosis (Fig. 3A2) were also blocked by Cd²⁺ (Fig. 3B2). However, at later stages of development, action potentials remained intact in Cd²⁺ but showed significant changes in spike waveform (Fig. 3, B3 and B4). Cd²⁺ reduced spike amplitude, increased spike duration, and reduced spike afterhyperpolarization (Fig. 3B3). Cd²⁺-induced reduction of spike amplitude was significantly greater in RR neurons with type 3 current profiles (31.7 ± 4.2% reduction, e.g., Fig. 3, A3 and B3) than in cells exhibiting type 4 profiles (9.5 ± 0.7% reduction, e.g., Fig. 3, A4 and B4; T = 8.87, P = 0.012), suggesting that Ca²⁺ currents contribute progressively less to spike amplitude as development proceeds. Ca²⁺ currents contribute to spike waveform in PB neurons also (Fig. 4A), but the effects of Cd²⁺ in these cells were not examined in detail in this study. That K⁺ currents in general contribute significantly to repolarization of the action potential in PB neurons and RR neurons is indicated by the effects of blocking outward currents with 4-AP and TEA (e.g., Fig. 4B).

View larger version (25K): [in this window] [in a new window]   Fig. 3. Responses of RR neurons to depolarizing current pulses show effects on spike waveform of blocking Ca2+ currents with Cd2+. A1A4: recordings in current-clamp mode from RR neurons that exhibited whole cell current profiles of types 1-4, respectively (see accompanying paper), illustrate characteristic changes in spike waveform associated with metamorphic adult development. B1B4: recordings from the same cells as in A1A4, respectively, 2-3 min after the addition of 5 × 104 M Cd2+ to the medium bathing the cells. Cell membrane potential was held at the level recorded prior to Cd2+ treatment. A1 and B1: recordings from a stage 3 RR neuron. Prior to the appearance of spike activity (A1), Cd2+ blocks a large component of the sag (arrow) apparent in responses elicited by depolarizing current pulses (B1), indicating that Ca2+-activated outward currents contribute to the sag. A2 and B2: recordings from a stage 5 RR neuron. Ca2+ spikes apparent in RR neurons prior to the appearance of Na+-dependent action potentials (arrow, A2) were blocked by Cd2+ (B2), but not affected by TTX (see accompanying paper). A3 and B3: recordings from a stage 8 RR neuron. During development, the action potential waveform became larger in amplitude, shorter in duration, and increasingly Na+ dependent. Blocking Ca2+ currents with Cd2+ increased spike duration and reduced the size of the afterhyperpolarization (B3). Cd2+ also reduced spike amplitude. A4 and B4: recordings from a stage 14 RR neuron. The results suggest that Ca2+ currents contribute significantly to spike waveform late in metamorphosis. However, the magnitude of the effects on spike amplitude of blocking Ca2+ currents with Cd2+ was greater in cells midway through metamorphosis (e.g., see B3) than in stage 14 RR neurons (e.g., see B4, see text).

View larger version (13K): [in this window] [in a new window]   Fig. 4. Responses of PB neurons to brief (20-ms) depolarizing current pulses (approximately 0.5 nA) show effects on spike waveform of (A) blocking Ca2+ currents with Cd2+, and (B) blocking K+ currents with tetraethylammonium (TEA) and 4-aminopyridine (4-AP). A: recording in current-clamp mode from a PB neuron that exhibited a PB-type 2 whole cell current profile (see accompanying paper). Responses obtained before (Control) and after (+Cd2+) the addition of 5 × 104 M Cd2+ are superimposed to show spike broadening induced by Cd2+. B: effects on spike waveform of blocking outward currents with 4-AP (5 × 103 M) and TEA (3 × 102 M). In the presence of these specific K+ channel blockers, repolarization of the membrane (arrow) is slowed dramatically.

Outward currents

Previous studies have shown that Na⁺ currents, and Na⁺-dependent action potentials, in Manduca AL neurons can be blocked with TTX, or by perfusing cells in Na⁺-free medium (Mercer et al. 1996b; see also accompanying paper). Interestingly, these treatments were found in the present study to have a significant impact also on a rapidly activating, transient outward current that was particularly prominent in RR neurons (Fig. 5A1). As shown in Fig. 5, TTX not only abolished I_Na, but also reduced the amplitude of this outward transient (Fig. 5A2). The effects of TTX on I_Na and on the fast outward transient were at least partially reversible (Fig. 5A3) and could be mimicked by perfusing cells with Na⁺-free saline (Fig. 5B). Neither of these treatments blocked Ca²⁺-independent K⁺ currents in Manduca AL neurons (I_A and I_KV) (Hayashi et al. 1992; Kloppenburg et al. 1999a; Mercer et al. 1996b; see following text). However, the complete abolition of total outward current by the specific K⁺ channel blockers, 4-AP and TEA, together with internal CsCl (see METHODS), suggests that the TTX-sensitive outward current identified in this study is carried mostly by K⁺ ions. In RR neurons and PB neurons, the current is also reduced significantly by the Ca²⁺ channel blocker, Cd²⁺ (see following text).

View larger version (20K): [in this window] [in a new window]   Fig. 5. A and B: Na+-dependent outward current. A: a series of depolarizing voltage steps from 70 to +50 mV shows the whole cell current profile of a stage 14 PB neuron exhibiting a PB-type 3 current profile (see accompanying paper). No Cd2+ was present in the medium bathing these cells. A1: prior to applying 108 M TTX to the cell (Control), a fast, transient outward component was apparent in the outward current profile (arrow). A2: in the presence of 108 M TTX this outward component was blocked (TTX). A3: the effects of TTX were partly reversed when cells were washed in TTX-free saline (Wash). B1: single voltage steps from 70 to +40 mV are used to illustrate the effects of perfusing cells in Na+-free saline. Recordings from a stage 14 RR neuron in normal saline (gray arrow) and in Na+-free saline (arrow, Na+-free) are superimposed to show that in Na+-free saline both, the fast transient inward current, INa, and a fast, transient outward current disappear. B2: subtraction of the current profile recorded in Na+-free saline from the current profile recorded in normal saline reveals the 2 currents that are abolished in Na+-free saline.

CA²⁺-DEPENDENT K⁺ CURRENTS. Changes in spike waveform resulting from addition of Cd²⁺ to the external medium (Figs. 3 and 4) suggest that Ca²⁺-dependent outward currents contribute to membrane repolarization in Manduca AL neurons. To examine Ca²⁺-dependent components of the outward current profile, whole cell current profiles recorded in the presence of Cd²⁺ (Fig. 6, A2 and B2) were subtracted from those recorded prior to Cd²⁺ treatment (Fig. 6, A1 and B1, respectively). Subtraction currents reveal the time course and amplitude of outward currents blocked by Cd²⁺ (Fig. 6, A3 and B3). Cells examined in this way were analyzed under two different sets of conditions: in the first group, Na⁺ currents were blocked with TTX (10⁸ M) prior to the application of Cd²⁺ (Fig. 6A), whereas in the second group, Na⁺ currents were left intact (see Fig. 6B).

View larger version (47K): [in this window] [in a new window]   Fig. 6. Ca2+-dependent K+ currents (IKCa). Whole cell current profiles of stage 14 RR neurons recorded in 108 M TTX (A) and in normal saline (B). A1 and B1: recordings made prior to Cd2+ treatment. A2 and B2: effects of applying 5 × 104 M Cd2+. A3 and B3: subtraction of current profiles recorded in the presence of Cd2+ (A2 and B2) from those recorded prior to Cd2+ treatment (A1 and B1, respectively) reveal the amplitude and time course of currents blocked by Cd2+. Cd2+-sensitive currents in cells not exposed to TTX (B3) displayed both, a fast transient outward component (IKCa transient, arrow) and a sustained outward component (IKCa sustained). In cells treated with TTX (A3), the transient component was greatly reduced in amplitude. C: peak current density plotted as a function of membrane potential. Na+ currents in these cells were blocked with TTX. D: comparison of IKCa profiles recorded (in the absence of TTX) from a stage 2 RR neuron (D1), a stage 14 RR neuron (D2), and a stage 5 PB neuron (D3). E and F: comparisons of mean current densities (±SE) in RR neurons and PB neurons early (stage 5, white bars) and late (stage 14, gray bars) in metamorphosis. Voltage steps from 70 to +40 mV were used to examine the density of IKCa transient (E) and IKCa sustained (F). The significance of differences between groups is indicated by letters above the bars in each graph. Groups that share a letter are not significantly different; groups that do not share a letter differ significantly (P < 0.05). E: the density of IKCa transient was calculated from Cd2+ subtraction currents recorded in TTX-free saline (e.g., B3). A significant difference between the density of IKCa transient in RR neurons at stages 5 and 14 was not apparent in PB neurons. At stage 14 there was no significant difference between RR neurons and PB neurons in the density of IKCa transient. F: the density of IKCa sustained was calculated from Cd2+ subtraction currents recorded in the presence of TTX (e.g., A3). To minimize contamination of the sustained current with IKCa transient, measurements of IKCa sustained were taken from the end of the 100-ms depolarizing voltage step. IKCa sustained density in RR neurons was significantly higher at stage 5 than at 14. This was not the case for PB neurons. Differences in IKCa sustained density between RR neurons and PB neurons were not significant.

CA²⁺-INDEPENDENT K⁺ CURRENTS. Ca²⁺-independent K⁺ currents are likely to contribute also to membrane repolarization in Manduca AL neurons (Kloppenburg et al. 1999a; Mercer et al. 1996b). In cells used to examine developmental changes in the density of Ca²⁺-independent K⁺ currents, TTX (10⁸ M) was applied to block Na⁺ currents and Ca²⁺ currents were blocked with 5 × 10⁴ M CdCl₂. Replacement of Ca²⁺ ions with Ba²⁺, and removal of Ca²⁺ from the solution used to fill the recording electrodes have been used previously to confirm the Ca²⁺ independence of the K⁺ currents examined below (Mercer et al. 1996b). Two distinct types of current contribute to the Ca²⁺-independent K⁺ currents in Manduca AL neurons: rapidly activating, transient A-type current (I_A) of which two forms have been described, I_{A fast} and I_{A slow}, and a sustained component (I_KV) that inactivates slowly, if at all, during a 200-ms depolarizing voltage step (Hayashi et al. 1992; Kloppenburg et al. 1999a; Mercer et al. 1996b).

TRANSIENT K⁺ CURRENT. For measurement of I_A, 3 × 10² M TEA was added to the saline bathing the cells to block the sustained current, I_KV. I_{A fast} was identified in all (n = 5) of the stage 5 RR neurons examined (Fig. 7A1), but in only three of eight RR neurons at developmental stage 14. I_{A slow} alone was detected in the remaining stage 14 RR neurons (Fig. 7A2). I_{A fast} differed from I_{A slow} in the following ways: first, peak current density was consistently higher for I_{A fast} than for I_{A slow} (Fig. 7B); second, I-V plots used to examine the voltage dependence of I_A (Fig. 7B) indicate that the threshold for activation of I_{A fast} occurs at a more hyperpolarized potential (around 45 mV) than for I_{A slow} (around 25 mV); and third, between 0 and +70 mV, time-to-peak was significantly faster for I_{A fast} than for I_{A slow}. With a depolarizing voltage step from 70 to +20 mV, mean time to peak current was 8.0 ± 1.3 ms for I_{A fast} and 15.9 ± 1.7 ms for I_{A slow} (T = 3.75, P = 0.005). As the two A-type currents are difficult to separate, measurements comparing A-type current densities in RR neurons and PB neurons (Fig. 7C) did not discriminate between these two subtypes. ANOVA revealed significant differences between the four groups of cells tested (df 3, F = 6.81, P = 0.005). While the mean density of I_A in stage 5 RR neurons appeared lower than in stage 14 RR neurons, differences between these two groups were not statistically significant. In contrast, the mean density of I_A in stage 14 PB neurons was significantly higher than in stage 5 PB neurons (Fig. 7C). This is consistent with the observation that, in marked contrast to RR neurons, I_{A fast} remained as prevalent in PB neurons at stage 14 as it was at stage 5. 

View larger version (20K): [in this window] [in a new window]   Fig. 7. Ca2+-independent K+ current, IA. Previous analyses have revealed 2 distinct forms of A-type current in Manduca AL neurons, IA fast and IA slow (Mercer et al. 1996b). In the present study, IA fast (A1) was prominent in all stage 5 RR neurons (n = 5), whereas IA slow only (A2) was observed in 6 of 8 stage 14 RR neurons. B1: voltage dependence of IA fast () and IA slow (). Peak current density is plotted as a function of membrane potential. Activation of IA fast occurs around 45 mV. In contrast, IA slow activation occurs at a voltage closer to 25 mV. C: comparisons of mean current densities (±SE) in RR neurons and PB neurons early (stage 5, white bars) and late (stage 14, gray bars) in metamorphosis. Voltage steps from 70 to +40 mV were used. The significance of differences between groups is indicated by letters above the bars in each graph. Groups that share a letter are not significantly different; groups that do not share a letter differ significantly (P < 0.05). In PB neurons, A-type current densities at stage 14 were significantly higher than at stage 5, but this was not the case for RR neurons. Interestingly, at stage 14, IA densities were significantly higher in PB neurons than in RR neurons.

SUSTAINED K⁺ CURRENT. To measure I_KV (Fig. 8A), A-type K⁺ currents were blocked with 5 × 10³ M 4-AP, or alternatively, measurements were taken from a holding potential of 20 mV, where I_A has been shown to be almost completely inactivated (Kloppenburg et al. 1999a). I-V plots of data taken from RR neurons indicate that there was no difference in the voltage dependence of I_KV between cells at stages 5 and 14 of metamorphic adult development (Fig. 8B). I_KV reached peak current levels in 44.5 ± 11 ms (stage 5 RR neurons) and 37.2 ± 8.6 ms (stage 14 RR neurons) after the onset of depolarizing voltage steps from 70 to +20 mV (T = 0.52; P = 0.64). I_KV densities in RR neurons and PB neurons were similar (Fig. 8C), and neither cell type exhibited significant differences in I_KV density between pupal stages 5 and 14 (df 3, F = 0.44, P = 0.73).

View larger version (24K): [in this window] [in a new window]   Fig. 8. A Ca2+-independent K+ current, IKV. Recording from a stage 5 RR neuron. B: voltage dependence of IKV in RR neurons at stage 5 () and stage 14 (). Peak current density is plotted as a function of membrane potential. Activation of IKV occurs around 15 mV. C: comparisons of mean current densities (±SE) in RR neurons and PB neurons early (stage 5, white bars) and late (stage 14, gray bars) in metamorphosis using voltage steps from 70 to +40 mV. The significance of differences between groups is indicated by letters above the bars in each graph. Groups that share a letter are not significantly different; groups that do not share a letter differ significantly (P < 0.05). There was no significant difference between the density of IKV in cells from stage 14 pupae and stage 5 pupae, either in PB neurons or in RR neurons.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The principal aim of this study was to identify ionic currents that contribute to changes in the electrical excitability of Manduca AL neurons during metamorphic adult development of the moth (see accompanying paper). The results show significant changes in the density of inward and outward ionic currents in Manduca AL neurons that correlate temporally with developmental changes in spike waveform and cell excitability. It is unlikely that we have isolated all voltage-gated currents present in these neurons, but the study provides important first steps toward understanding the functional development and membrane biophysics of cells in central olfactory pathways of the moth.

Action potentials in Manduca AL neurons are initially Ca²⁺ dependent

Early in metamorphosis, action potentials in Manduca AL neurons are Ca²⁺ dependent, but they become increasingly Na⁺ dependent during the course of adult metamorphosis. Developmental changes in the ionic dependence of action potentials have been described in many neuronal systems (e.g., Baccaglini and Spitzer 1977; Blair 1983; Mori-Okamoto et al. 1983; Willard 1980). While events that trigger such changes have yet to be fully elucidated, there is increasing evidence that early forms of electrical excitability have a developmental function (reviewed by Spitzer 1991, 1994). Manduca AL neurons exhibit spontaneous activity in vitro. This activity is particularly prevalent prior to and during the transition from Ca²⁺- to Na⁺-dependent spikes (see accompanying paper), and recent studies show that the cells exhibit distinctive plateau properties during this period (Mercer et al. 2000). The shift to Na⁺-dependent action potentials is not accompanied by reductions in Ca²⁺ current density. On the contrary, in RR neurons the density of Ca²⁺ currents was significantly higher at stage 14 than at pupal stage 5. In mature neurons it is likely that Ca²⁺ spikes are masked by the large K⁺ currents present in these cells.

Ca²⁺ spikes have been identified also in the Manduca motor neuron, MN5, and may play a role in the postembryonic dendritic remodelling of this cell (Duch and Levine 2000). Duch and Levine (2000) show that the transient occurrence of Ca²⁺ spikes correlates temporally with cessation of the formation of higher-order branches in MN5. In embryonic spinal neurons it has been shown also that the frequency of growth cone Ca²⁺ transients is inversely proportional to the rate of axon outgrowth (Gomez and Spitzer 1999; Lautermilch and Spitzer 2000). The role of Ca²⁺ spikes in Manduca AL neurons has yet to be determined, but it is interesting to note in this regard that the neuromodulator 5-HT not only promotes neurite outgrowth of immature RR neurons in vitro (Mercer et al. 1996a), but also significantly reduces Ca²⁺ currents in these cells (Mercer et al. 1995).

Cell-type specific differences in the appearance of I_Na

The inward and outward currents identified in Manduca AL neurons (Hayashi and Hildebrand 1990; Kloppenburg et al. 1999a; Mercer et al. 1995, 1996b; present investigation) are similar to those described in other insect neurons (e.g., Baines and Bate 1998; Baker and Salkoff 1990; Byerly and Leung 1988; Grünewald and Levine 1998; Hayashi and Levine 1992; Kloppenburg et al. 1999b; Laurent 1991; O'Dowd and Aldrich 1988; Saito and Wu 1991; Schäfer et al. 1994; Solc and Aldrich 1988). Voltage-gated currents appear in developing central neurons in a predictable sequence, the order of which varies between different types of neurons (Baines and Bate 1998; O'Dowd et al. 1988; Ribera and Spitzer 1990; Spitzer 1994). In Manduca AL neurons, A-type currents can be detected prior to the appearance of fast transient Na⁺ currents in the cells, whereas in central neurons of the Drosophila embryo the opposite is reported (Baines and Bate 1998). Interestingly, blockade of synaptic transmission in Drosophila embryonic motor neurons results in significant increases in voltage-gated Na⁺ (and K⁺) currents in the cells (Baines et al. 2001). That the expression of voltage-gated currents can be influenced by local cues is suggested also by results showing the effects of the steroid hormone 20-hydroxyecdysone (20-HE) on Ca²⁺ currents in Manduca motor neurons. Grünewald and Levine (1998) have shown that pupal leg motor neurons exposed in vitro to 20-HE exhibit significantly larger Ca²⁺ current levels than untreated cells. Effects of 20-HE on the density of currents in developing Manduca antennal-lobe neurons have yet to be determined.

I_Na could not be detected in all stage 5 RR neurons. This is consistent with the finding (see accompanying paper) that a lower percentage of RR neurons exhibit Na⁺-dependent action potentials at early stages of metamorphosis than PB neurons. The earlier development of voltage-dependent ionic currents in PB neurons may explain why differences in the density of ionic currents (present investigation), and changes in spike waveform and cell excitability (accompanying paper), between cells at pupal stages 5 and 14 were less pronounced in PB neurons than in RR neurons. Developmental changes in the morphology of projection (output) neurons, such as PB neurons, also occur slightly ahead of changes in local AL interneurons. Dendritic processes of projection neurons invade developing glomeruli before the processes of multiglomerular local AL interneurons (Malun et al. 1994; Oland et al. 1990). Increases in Na⁺ current density during adult metamorphosis are consistent with observed increases in spike amplitude during this period (see accompanying paper). However, the absence of differences in the voltage dependence or activation kinetics of Na⁺ currents in RR neurons exhibiting different Na⁺ current densities, suggests that these parameters do not contribute significantly to changes in spike waveform in these cells at the stages of metamorphosis examined in this study.

Development of outward currents

The dramatic effects on action potential duration of reducing outward currents with K⁺ channel blockers indicate that K⁺ currents contribute significantly to action potential repolarization in Manduca AL neurons. In addition to shaping action potential waveforms, K⁺ currents play an important role in setting resting membrane potentials, and in modulating the frequency of neuronal firing (reviewed by Salkoff et al. 1992). Increases in the density of K⁺ currents in AL neurons during adult metamorphosis are consistent with K⁺ currents contributing not only to decreases in action potential duration, but also to changes in cell excitability during this period. There was no evidence in these neurons of developmental shifts in the I-V curves of outward currents, suggesting that action potential waveforms between pupal stages 5 and 14 are not affected by this parameter. Moreover, no significant changes in times to peak current were identified.

Complex changes in K⁺ current profiles

Changes in K⁺ current densities that coincide with adult metamorphosis are complex and cell-type specific. In RR neurons, the percentage of cells exhibiting rapidly activating, transient A-type current (I_{A fast}) fell as development progressed, but the density of Ca²⁺-dependent K⁺ currents in these cells increased. In contrast, analysis of K⁺ current densities in PB neurons at pupal stages 5 and 14 revealed significant increases in the density of I_A. The persistence of I_{A fast} in PB neurons correlates well with the identification of large amplitude currents of this type in projection neurons of the adult moth (Kloppenburg et al. 1999a). K⁺ currents in stage 14 PB neurons were similar in amplitude to those seen in adult projection neurons, suggesting that by pupal stage 14 the K⁺ current profile of projection neurons may be close to being fully mature. The contribution that each of the K⁺ currents identified in this study makes to developmental changes in cell excitability has yet to be determined. However, a lack of change in peak I_KV current densities between pupal stages 5 and 14, suggests that in RR neurons and PB neurons, I_KV does not contribute significantly to changes in spike waveform during this period. This is in marked contrast to Xenopus spinal neurons, where delayed rectifier K⁺ current plays a major role in the developmental shortening of the action potential (reviewed by Spitzer 1994; Spitzer and Ribera 1998). However, effects on spike waveform of blocking Ca²⁺ currents in AL neurons with Cd²⁺, and large increases in the density of Ca²⁺-dependent currents apparent in RR neurons, suggest that Ca²⁺-dependent K⁺ currents contribute significantly to repolarization of the action potential in mature AL neurons of the moth. Both sustained and transient Ca²⁺-dependent K⁺ currents have been described previously in insect neurons (Schäfer et al. 1994; Thomas 1984; Torkkeli and French 1995; Wegener et al. 1992; Zufall et al. 1991), and evidence suggests that Ca²⁺-dependent K⁺ currents contribute to repolarization of the action potential both in vertebrate and invertebrate neurons (Adams et al. 1982; Gao and Ziskind-Conhaim 1998; MacDermott and Weight 1982; Storm 1987; Walsh and Byrne 1989). In Manduca AL neurons, removal of Na⁺ ions from the external medium, or blocking Na⁺ influx with TTX, reduces the transient component, while leaving IK_{Ca sustained} intact.

Na⁺-dependent outward current in Manduca AL neurons

K⁺ channels activated by Na⁺ have been reported in many types of neuronal membranes (Bischoff et al. 1998; Dryer 1991, 1994; Egan et al. 1992a,b). Transient, Na⁺-activated K⁺ current has been identified also in dorsal unpaired medium (DUM) neurons of the cockroach, Periplaneta americana (Grolleau and Lapied 1994), and a large sustained Na⁺- and voltage-dependent K⁺ current has been reported in spinal neurons of the frog embryo (Dale 1993). The functions of Na⁺-activated K⁺ channels are not well understood. K_Na channels in vertebrate neurons appear not to contribute to action potential waveform (Safronov and Vogel 1996), but their activation in rat motoneurons during repetitive firing of action potentials leads to pronounced slow hyperpolarization of the cells. Transient Na⁺-activated K⁺ current identified in insect DUM neurons can be elicited in Ca²⁺-free medium or in solutions containing CoCl₂ (Grolleau and Lapied 1994), but in Manduca AL neurons the Na⁺-dependent outward current is blocked by Cd²⁺ (see earlier), suggesting that IK_Na and the transient component of IK_Ca in these cells may be one and the same current. A better understanding of the macroscopic K⁺ currents identified in Manduca AL neurons awaits analysis at a single-channel level.

The diversity of ionic currents in Manduca AL neurons, and the possibility of cell-type-specific differences in their developmental regulation, is revealed for the first time in this study. The functional significance of developmental changes in electrical excitability is the subject of intense interest and investigation. Manduca AL neurons provide an experimentally tractable system to explore the impact of such changes on the growth, electrical properties, and connectivity of developing central neurons.