| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
1Saxon Academy of Sciences, Department of Neurohormones, Jena; 2Institute of Physiology, Department of Neuroendocrinology, Philipps University Marburg, Marburg; and 3Max-Planck Institute of Neurobiology, Department of Systems and Computational Neurobiology, Martinsried, Germany
Submitted 30 August 2005; accepted in final form 18 September 2005
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONAPPENDIXGRANTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONAPPENDIXGRANTSREFERENCES |
|---|
; Van der Horst 2003). As found recently, they also enhance locomotor activity in response to starvation (Lee and Park 2004). In cockroach (Periplaneta americana), there are two AKH peptides (Scarborough et al. 1984; Witten et al. 1984), one of which, AKH I (pQVNFSPNWamide, also called Neurohormone D) (Baumann et al. 1984; Baumann and Penelin 1984) accelerates spiking and alters the shape of action potentials in abdominal dorsal unpaired median (DUM) neurons. These DUM neurons are of general interest inasmuch as they release the biogenic amine octopamine. The octopaminergic system of insects, comparable to the noradrenergic system of mammals, plays an important role in arousal and affects general activity levels (e.g., Roeder 2005).
Previous investigations revealed that AKH I modulates a whole set of ion currents in the DUM neurons (Wicher et al. 2001). AKH I leads to potentiation of a P/Q-type Ca2+ current due to channel phosphorylation via protein kinase A (PKA) (Wicher 2001b); reduction of the fast Na+ current, again via PKA (Wicher 2001a); potentiation of a BK-type Ca2+–activated K+ (KCa) current (Derst et al. 2003; Wicher et al. 1994); potentiation of a Ca2+ background current via Gq-mediated stimulation of phospholipase C (PLC) (Wicher et al. 2004). This latter current contributes to pacemaker depolarization and spontaneous firing of the abdominal DUM neurons. Other currents investigated, such as voltage-dependent K+ currents, were not affected by the peptide (Wicher et al. 1994).
The present study analyzes how the complex modulation of several currents leads to the increase in spike frequency and to the altered shape of the action potential. To simulate spiking of DUM neurons, we developed a one-compartment model that includes a set of 10 ion currents described in terms of the Hodgkin-Huxley formalism. Our studies indicate that accelerated spiking is solely due to the change in the Ca2+ background current, whereas the increase in fast afterhyperpolarization (fAHP) of action potentials (Wicher et al. 1994) requires the change in all other modulated currents.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONAPPENDIXGRANTSREFERENCES |
|---|
Isolation of dorsal unpaired (DUM) neurons from the fifth abdominal ganglion of the cockroach P. americana was performed as described previously (Wicher et al. 1994). Briefly, the ganglia were excised, desheathed, and incubated for 10 min at room temperature in saline (for composition see bath solution used for spike recordings) containing 0.5 mg/ml trypsin (type II, Sigma, Deisenhofen, Germany) and 0.5 mg/ml collagenase (type I, Sigma). After washing off the enzyme, the ganglia were stored in saline for 
1 h. Prior to the measurements, the somata of DUM neurons were separated using thin metal needles.
Electrophysiology
Ion currents in HEK293 cells and in DUM neurons were measured at room temperature using whole cell patch-clamp with appropriate compensation of series resistance and of capacitive and leakage currents. Pipettes having resistances of 2–4 M
(HEK cells) or 0.5–0.8 M (DUM cells) were pulled from borosilicate capillaries. Current measurements and data acquisition were performed using an EPC9 patch-clamp amplifier controlled by PULSE software (HEKA Elektronik).
HEK CELLS. For experiments with Periplaneta Slowpoke (pSlo) channels, the bath solution contained (in mM) 135 NaCl, 5 KCl, 1 MgCl2, 1 CaCl2, 0.33 NaH2PO4, 2 Na-pyruvate, 10 glucose, and 10 HEPES. The pipette solution contained 140 KCl, 4 NaCl, 2 Mg-ATP, and 10 HEPES. The free Ca2+ concentration was adjusted to 150 µM by adding CaCl2 and measuring [Ca2+] with calcium-sensitive electrodes (KWIK tips; WPI, Berlin). The pH of bath and pipette solution were adjusted to 7.4 and 7.3, respectively.
DUM NEURONS.
Spiking of neurons was recorded under current-clamp conditions without current injection. The bath solution contained (in mM) 190 NaCl, 5 KCl, 5 CaCl2, 2 MgCl2, and 10 HEPES (pH = 7.4), and patch pipettes (resistance >1.5 M) were filled with solution composed of (in mM) 190 K-gluconate, 5 NaCl, 2 Mg-ATP, 1 CaCl2, 3 EGTA, and 10 HEPES (pH = 7.25). Between recordings (duration: 1 s) the cells were held at –70 mV under voltage clamp.
Na+ currents were separated as described (Wicher 2001a) with pipette solution composed of (in mM) 5 NaCl, 100 choline methylsulfate, 30 TEA-Br, 3 CsCl, 60 CsOH, 2 Mg-ATP, 1 CaCl2, 5 EGTA, and 10 HEPES and bath solution containing (in mM) 60 Na isethionate, 90 choline methylsulfate, 40 TEA-Br, 7 MgCl2, 1 CdCl2, and 10 HEPES. For Ca2+ current measurements (Wicher and Penzlin 1997), the pipette solution contained (in mM) 100 choline methylsulfate, 30 TEA-Br, 8 CsCl, 60 CsOH, 2 Mg-ATP, 1 CaCl2, 5 EGTA, and 10 HEPES, and the bath solution (in mM) had 5 CaCl2, 190 choline methylsulfate, and 10 HEPES and 5 µM tetrodotoxin.
K+ currents in DUM neurons were measured with pipette solution containing (in mM) 180 K-gluconate, 10 NaCl, 2 Mg-ATP, 1 CaCl2, 5 mannitol, 3 EGTA, and 10 HEPES. The bath solution to measure KCa currents contained (in mM) 190 NaCl, 5 KCl, 5 CaCl2, 2 MgCl2, and 10 HEPES and 0.5 µM tetrodotoxin (TTX). KCa currents were separated from other currents by applying the BK-channel blocker iberiotoxin (100 nM). The purely voltage gated delayed rectifier (DR) and the A-type current were measured with the above bath solution containing in addition 2 CdCl2. To isolate the DR, the A-type current was depressed by 3 mM 4-aminopyridine. The bath solution to measure KNa currents contained (in mM) 190 NaCl, 5 KCl, 1 CaCl2, 2 MgCl2, 2 CdCl2, and 10 HEPES. KNa currents, together with Na currents, were separated by comparing recordings performed in the absence and presence of the Na+ channel blocker TTX (0.5 µM). The pH of bath solutions was adjusted to 7.4 and that of pipette solutions to 7.25. Liquid junction potentials were compensated.
Application or washout of blocking agents was performed by transferring the cell (attached to the pipette tip) within a glass tube into the various solutions. A complete and fast solution change within a few milliseconds was achieved by sucking a small amount of solution into the tube.
Data analysis
Results are given as means ± SE (n = number of cells). Fitting procedures were performed with the software Prism2 (Graph Pad, San Diego, CA) and with IGOR (WaveMetrics, Lake Oswego, OR).
Chemicals
Salts, tetraethylammonium, tetrodotoxin, and verapamil were purchased from Sigma (Deisenhofen, Germany). 
-Agatoxin IVA and -conotoxin GVIA were obtained from Alomone Labs (Jerusalem, Israel) and iberiotoxin from Calbiochem (Schwalbach, Germany).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONAPPENDIXGRANTSREFERENCES |
|---|
Periplaneta DUM neurons have previously been shown to express pSlo, the 
subunit of large conductance KCa (BK) channels (Derst et al. 2003). While pSlo channels heterologously expressed in HEK293 cells give rise to a purely noninactivating KCa current, the KCa current in DUM neurons is composed of a transient (KCa,t) and a sustained (KCa,s) component (Derst et al. 2003; Grolleau and Lapied 1995; Wicher et al. 1994). pSlo channels are sensitive to the BK channel blocker iberiotoxin (IbTx; IC50 = 45 nM). The KCa channels in DUM neurons seem to be more sensitive to IbTx because both KCa current components are blocked by 100 nM IbTx (Derst et al. 2003). This is illustrated in Fig. 1A, which shows sample traces of K+ currents, measured prior and after application of IbTx and the KCa current obtained by subtraction of these currents. There is an initial rapid KCa current rise (the time to peak ranges from 7 ms at –20 mV to 3 ms at 20 mV) and a fast decay (decay to 50% takes 
2 ms) of the first part of the KCa current which is dominated by KCa,t. It then reaches a sustained level, i.e., KCa,s, some 5 ms after stepping the voltage. Activation of both components of the KCa current starts around –50 mV, i.e., at about the same voltage as that of the start of the Ca2+ current (Fig. 1B, top). The time-to-peak of the KCa current is comparable to that of the voltage-gated Ca2+ current (Fig. 1D).
|
1 min later, indicating that KCa,s requires a higher [Ca2+]i than KCa,t to be activated by depolarization (Derst et al. 2003). The fact that KCa,t inactivated in spite of a constant [Ca2+]i indicates that inactivation is an intrinsic channel property perhaps produced by accessory 
subunits. It cannot be excluded that in our present measurements there is some overlap of currents, i.e., that there is a small contribution of KCa,s to the KCa peak current at least at higher depolarizations. Because, however, such error is expected to be small and will not have consequences for the results obtained in the following text, we will refer the KCa peak current as KCa,t.
The Ca2+ current activating positive to –50 mV is composed of three subtypes (Wicher and Penzlin 1997): P/Q-type current sensitive to -agatoxin IVA (-AgaTx), N-type current sensitive to -conotoxin GVIA (-CgTx), and L-type current sensitive to verapamil (Fig. 1B, bottom). The effect of -AgaTx (50 nM) and verapamil (10 µM) on Ca2+ current is illustrated in Fig. 1C; the concentrations used have been previously shown to block the respective currents within 1 min (Wicher and Penzlin 1997).
We applied the Ca2+ channel blockers to learn to what extent L-, N-, and P/Q-type Ca2+ currents might supply the Ca2+ for activation of the KCa current. The three Ca2+ currents cause a Ca2+ influx of different size and activation kinetics (Fig. 1, B, bottom, and C), and their suppression affected the KCa currents differently: -AgaTx (50 nM) led to reduction of both the transient and the sustained component of KCa current (Fig. 2, A and C), -CgTx (1 µM) had no effect on KCa current (n = 5, not shown), whereas verapamil (10 µM) had a clearly weaker effect on the transient component than -AgaTx and suppressed the sustained KCa current to a similar degree (Fig. 2, B and D). The lack of effect of N-type channel block might indicate a spatial separation of N-type Ca2+ channels and KCa channels so that Ca2+ entering through N-type channels hardly diffuses to KCa channels. On the other hand, the weak inhibition of the transient KCa current component by the block of the L-type channel may be largely due to the rather slow activation kinetics of L-type channel. As illustrated in Fig. 2B, the peak of L-type current is attained after the peak of the verapamil-sensitive KCa current. The rise in intracellular [Ca2+] due to the L-type current might thus overlap with the intrinsic inactivation of the channel conducting KCa,t. By contrast, the P/Q-type current activates more rapidly and reaches the peak earlier than the KCa current sensitive to -AgaTx (Fig. 2A). Thus the fast activating P/Q-type current seems to supply most of the Ca2+ required for the activation of the transient KCa current component. On the other hand, the sustained KCa current may rely on Ca2+ entering via both types of Ca2+ channels because it is reduced by about the same extent by blocking P/Q- or L-type channels (Fig. 2, C and D).
|
The Periplaneta adipokinetic hormone AKH I, a peptide released from the corpora cardiaca, is known to enhance KCa currents (Wicher et al. 1994). AKH I also potentiates the P/Q-type Ca2+ but not the N- and L-type Ca2+ currents in DUM neurons (Wicher 2001b). Figure 3 demonstrates the AKH-induced effect on KCa currents: the total outward current produced by a voltage step to 0 mV before and after application of 10 nM AKH I, and the difference current representing the peptide-induced KCa current are shown in Fig. 3A, 1 and 2, respectively. AKH I thus upregulates both the transient and the sustained KCa component. This may be plausibly attributed to the upregulation of the P/Q-type current as this current was seen in the preceding text to provide Ca2+ for the activation of both KCa components. The threshold concentration for the AKH I-induced potentiation of the P/Q-type Ca2+ current was 1 pM (Wicher 2001b). At this concentration, there was already some increase in the KCa current (Fig. 3A3), which, however, was mainly restricted to the transient component (Fig. 3B1). Increasing [AKH I] from 1 pM to 10 nM induced progressively larger KCa currents (Fig. 3C). Compared with the KCa current under control conditions (isolated with IbTx), the current produced by 10 nM AKH I differed in that the transient component activated at more negative potentials. The nonlinearity in the I-V relation for this component around –10 mV (Fig. 3B2) is especially remarkable. It seems to reflect the increase in P/Q-type current by AKH I that is most pronounced at voltages ranging from –20 to 0 mV (Wicher 2001b). The possibility that AKH I affects purely voltage-gated K+ currents such as the delayed rectifier (IK DR) and the A-type current (IK A) could be ruled out because in the presence of 1 mM Cd2+, which completely blocks the Ca2+ currents (Wicher and Penzlin 1997), AKH I failed to affect K+ currents (n = 5 cells, not shown). Taken together, these results are compatible with the assumption that the upregulation of the KCa current is caused by the increase in P/Q-type current.
|
). Conversely, a reduction of KCa current was seen to prolong action potentials and to reduce the hyperpolarization (Derst et al. 2003). Activation of PKA does not affect IKCa
Some Slo channels are known to be modulated by phosphorylation via protein kinases such as PKA (Schubert and Nelson 2001). AKH I in fact activates PKA, which is a necessary step in the upregulation of the P/Q-type Ca2+ current (Wicher 2001b). In principle the regulation of the KCa currents by AKH I might be dual, i.e., be partly due to KCa channel phosphorylation. To evaluate whether PKA can affect the pSlo channel, we coexpressed pSlo together with the Periplaneta AKH receptor (pAKHR) in HEK293 cells. Stimulation of heterologously expressed pAKHR with AKH I was previously shown to increase [cAMP] via activating Gs proteins in HEK293 cells (Wicher, unpublished observation). Application of AKH I (10 nM), after loading the cell with a Ca2+-rich pipette solution, had, however, no effect on pSlo currents activated by depolarization (Fig. 4A, left). There was neither a change in the size of currents nor in the kinetics. In another series of experiments, we induced a rise in [cAMP] by bath application of the membrane-permeable cAMP analogue 8-bromo-cAMP. Again, there was no change in the properties of pSlo currents (Fig. 4A, right). To exclude the possibility that PKA might be fully activated under control conditions, we tested whether inhibition of PKA would affect pSlo currents. Application of 10 µM KT5720, which was previously shown to completely abolish the AKH I effect on P/Q-type Ca2+ current (Wicher 2001b), did not change pSlo currents. These results, which are summarized in Fig. 4B, do not support the possibility that PKA regulates pSlo. Recently direct binding of a PKA and a Src tyrosin kinase to domains in the C-terminus of the Drosophila Slo channel (dSlo) and phosphorylation of the dSlo channel protein by both kinases has been demonstrated (Wang et al. 1999). However, no difference in peak current amplitude nor in voltage dependence of dSlo gating was observed after coexpression of dSlo with either protein kinase (Wang et al. 1999). On the other hand, the free catalytic subunit of PKA binds to dSlo and leads to downregulation of channel activity (Zhou et al. 2002). This modulation, however, does not involve phosphorylation of the only consensus PKA-substrate site in the C-terminal domain of dSlo.
|
DUM neurons express Na+-dependent K+ channels (Wicher et al. 1994; Grolleau and Lapied 1994). Therefore blocking the Na+ current with tetrodotoxin (TTX) leads to disappearance of a transient outward current component (Fig. 5A), and the total TTX-sensitive current is the sum of the Na+ current and the Na+-dependent K+ (KNa) current (Fig. 5B). Due to the lack of a specific blocker, it is impossible to separate the KNa current from the Na+ current (Grolleau and Lapied 1994). Although the TTX-sensitive outward current is contaminated with the Na+ current, we will refer to it as the KNa current. It activates very rapidly, its peak following that of the Na+ current within <1 ms (Fig. 5C). Because KNa channels are not voltage-gated, the I-V relation of the KNa peak current mirrors the I-V relation of the Na+ current (Fig. 5D).
|
AKH I accelerates the inactivation of the Na+ current in DUM neurons, thereby reducing both peak size and duration and thus also the net Na+ influx (Wicher 2001a). Because activation of the KNa current requires a high intracellular Na+ concentration (Dryer 1994), any reduction of Na+ influx should entail a reduction of the KNa current. An example of a current suppressed by AKH I (i.e., the difference of the current measured under control conditions and in presence of AKH I) is shown in Fig. 5E. The relation between the AKH-I-sensitive Na+ current and the AKH-I-sensitive KNa current (Fig. 5F) resembles that between the TTX-sensitive Na+ current and KNa current under control conditions (Fig. 5D). Thus the reduction of KNa peak current was proportional to the reduction of Na+ peak current, and the downregulation of KNa current by AKH I, therefore is probably solely due to the reduction of the Na+ influx. A test of a possible effect of PKA on the KNa channel (Slo2 or slack) (Yuan et al. 2003) could not be performed because this channel has not yet been cloned in Periplaneta. Furthermore, an experimental analysis of the role of the peptidergic modulation of Na+ current and KNa current on action potential shape is impaired by the lack of tools to separate the currents: Li+ permeates Na+ channels but fails to activate KNa channels in some preparations. By contrast, in DUM neurons Li+ also activates the KNa current, and it can thus not be used as tool to separate the Na+ current from KNa currents (Grolleau and Lapied 1994).
Modeling the effect of peptidergic counter-regulation of Ca2+- and Na+-dependent K+ currents on the action potential shape
We performed a modeling study of the differential modulation of the K+ currents with the aim of answering two main questions: Does modulation of these currents affect the pacemaker depolarization and does superposition of the modulatory effects explain quantitatively the observed increase in fast AHP (fAHP) of action potentials?
To simulate the endogeneous spiking of DUM neurons, we included 10 currents in a one-compartment model of a DUM neuron: one Na+ current, four Ca2+ currents (the background current, a low-voltage activated current, and 2 high-voltage activated currents, i.e., P/Q-type and non-P/Q-type current), and five K+ currents (the delayed rectifier, KDR, the A-type current, KA, the KNa current and the KCa currents, KCa,t and KCa,s). The currents were described in terms of the Hodgkin-Huxley formalism (cf. APPENDIX, Fig. A1 and Table 1). Figure 6 compares recorded spikes from a DUM neuron with the spike pattern generated by the model, using the parameters given in Table 1. The model reproduced the experimentally observed characteristics of the action potential such as threshold, overshoot, and AHP as well as the resting firing frequency.
|
|
|
). The dependence of fAHP on IKNa was similar to that on IKCa,t but considerably steeper (Fig. 7B).
|
The main conclusions about the role of the KNa current and the transient KCa current in the regulation of spiking are there is a critical size of KNa current that is required for stable spiking (Fig. 7, A and B) and the AHP is determined by up- and downregulation of the KCa current by the P/Q-type Ca2+ current and by downregulation of the KNa current by the Na+ current.
Application of 10 nM AKH I produces faster spiking of DUM neurons and affects the action potential shape. The spike frequency increase ranges from 15 to 34%, the mean is 25.2 ± 4.0% (n = 7, Fig. 8A). The hyperpolarization increases by 2–6 mV, on average by 4.0 ± 1.0 mV (n = 7, Fig. 8B). The question, then, is whether our model reproduces these changes and whether this requires that in fact all observed AKH I-induced changes in ion currents are implemented, i.e., upregulation of P/Q-type Ca2+ current and KCa currents, downregulation of Na+ current, and KNa current as well as upregulation of Ca2+ background current. In terms of our model (cf. Table 1), this means the following changes: G (Ca P/Q) from 209 to 287 nS, 
h Slope up (Na) from +11 to +13 mV, h Slope dn (Na) from –22 to –18 mV, and G (Ca, back) from 0.023 to 0.07 nS. These changes produced indeed an accelerated spiking. In the example shown in Fig. 8C, the increase in spike frequency amounts to 25%; this is in accordance with the mean AKH I effect. Furthermore, the hyperpolarization is increased by 3 mV (Fig. 8D). The model also predicts a slight reduction in overshoot (by 3 mV, Fig. 8C), which was, however, not observed with DUM neurons (Fig. 8A).
|
We finally simulated a situation where AKH I would only lead to upregulation of the P/Q-type Ca2+ current and both KCa currents (Fig. 9A). This caused an increase in hyperpolarization (by 5 mV), which is in line with the measured AKH I effect (4 mV), whereas the spike frequency was hardly affected (reduction by 2%). Downregulation of Na+ current and KNa current accelerated spiking (by 7%) and slightly reduced the hyperpolarization (by 2 mV; Fig. 9B). Another effect was a slight reduction of the overshoot (by 3 mV). When the modulations of P/Q-type Ca2+ current and KCa currents were combined with those of Na+ and KNa currents, the change of the hyperpolarization was in accord with that observed experimentally (increase by 3 mV; Fig. 9C). The spike frequency, however, remained virtually unaffected (increase by 2%). When in addition the upregulation of Ca2+ background current was implemented, the result was an increase in spike frequency by 25% with no further effects on the action potentials i.e., the observed AKH I effects were now correctly reproduced as already mentioned in the preceding text (Fig. 9D).
|
), the acceleration of spiking by AKH I solely relies on the upregulation of the Ca2+ background current. |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONAPPENDIXGRANTSREFERENCES |
|---|
; Saraswati et al. 2004). In Drosophila, it has recently been shown that, on starvation, AKH-release produces enhanced locomotor activity (Lee and Park 2004). Conceivably, the physiological relevance of the AKH I-induced modulation of spiking lies in the enhanced release of octopamine both to local targets (e.g., respiratory muscles) and to the hemolymph thereby contributing to a rise of systemic octopamine level.
Neuropeptides, in general, affect not a single but several ionic currents in the excitable cells they act on, e.g., neurons and skeletal muscles of invertebrates (Brezina et al. 2005; Nusbaum et al. 2001; Wicher et al. 2001) or neurons of vertebrates (e.g., Pena and Ramirez 2004). In the present study on cockroach DUM neurons, we have completed knowledge of the neuro-modulatory actions on ion currents of the hormone AKH I, and we have tested, in a model of this neuron, how the interplay of these effects leads to the observed changes in firing and time course of action potentials.
To understand the role of different ion currents in the discharge patterns of neurons, various computational models have been developed for well studied examples from invertebrates and vertebrates (e.g., Dale 1995; Golowasch et al. 1992). This approach has also been used in analyzing the modulation of currents by chemical messengers (e.g., Harris-Warrick et al. 1995). Neuronal models designed to simulate spontaneous activity either implement KCa currents (Buchholtz et al. 1992), KNa currents (Dale 1995), or KNa and slowly activating KCa currents (Dale and Kuenzi 1997). In our model, the KNa current and the KCa,t current, two rapidly activating currents contributing to action potential repolarization, act in parallel and are both subject to modulation. In contrast to various other models, ours lacks the hyperpolarization-activated cation current, Ih. This type of current, which has not been found in cockroach DUM neurons (Grolleau and Lapied 2000) but in Drosophila CNS (Marx et al. 1999) and in bee olfactory receptor neurons (Gisselmann et al. 2003), is present in some but not all spontaneously active insect neurons (Wicher et al. 2001). Generally, for the neurons to spontaneously spike in the tonic mode, only some persistent inward current plus a low-threshold Ca2+ current seem to be required. In the case of cockroach DUMs, a Ca2+ background current, a low-voltage-activated Ca2+ current and a persistent Na+ current contribute to pacemaking (Grolleau and Lapied 2000).
How AKH I leads to changes of Ca2+- and Na+-dependent K+ currents
Because AKH I produces changes in Ca2+ and Na+ currents, the KCa and the KNa current must also change. For IKCa, we experimentally ruled out the possibility of an additional direct modulation of the respective channels. For IKNa, this is also rather unlikely because our model calculations returned realistic results without implementing any direct modulation of this current. The observed slight discrepancy with respect to the amplitude of the action potential (i.e., reduction in overshoot by 3 mV) indicates that in the real neurons the effect of Na+ current reduction on overshoot is compensated for by the reduction of KNa current. There might be a more sophisticated relationship between Na+ and KNa current than included in our model.
Modulation of the KCa current
Although three types of voltage-gated Ca2+ channels are expressed in DUM neurons (Fig. 1B), only P/Q-type channels appear to provide Ca2+ for activation of the transient KCa current component. The sustained KCa component can additionally be activated by Ca2+ influx through L-type channels (Fig. 2). Coupling of BK channels to L- and Q-type channels has been reported for rat chromaffin cells (Prakriya and Lingle 1999). In mouse sympathetic neurons, however, BK channels are linked to N-type channels, whereas L- and P-type channels activate Ca2+-activated Cl– currents (Martinez-Pinna et al. 2000). In hippocampal neurons, N-type channels activate BK channels, L-type channels activate small conductance (SK) KCa channels and P/Q-type channels fail to activate either type of channel (Marrion and Tavalin 1998).
Blocking the BK currents in DUM neurons under current-clamp conditions demonstrated that only the rapidly activating, transient component KCa,t affects the shape of the action potential. KCa,t controls its duration (for further K+ currents involved in this process, cf. Wicher et al. 2001). In addition KCa,t determines the size of fAHP (Derst et al. 2003). These roles of the BK current are comparable to those in various mammalian neurons (Faber and Sah 2002). The sustained KCa component, according to the predictions of our model, controls the interspike interval similar to the purely voltage-dependent K currents KA and KDR although, of course, in Ca2+-dependent fashion (not shown).
Modulation of the KNa current
KNa channels require a relatively high intracellular Na+ concentration (50 mM) to become activated. In some preparations, prolonged discharges or long-lasting depolarizations are required to activate IK,Na (Dryer 1994). Co-localization of Na+ and KNa channels may, however, allow a sufficient rise in [Na+] to activate the KNa current already during a single action potential (Koh et al. 1994). In hippocampal CA1 neurons, for example, the KNa current is responsible for the fAHP after a single action potential (Liu and Stan Leung 2004).
In bursting neocortical neurons, the KNa current is the main cause of the postexcitatory hyperpolarization (Franceschetti et al. 2003). In these cells, the KCa current, too, makes an albeit small contribution to the hyperpolarization. In DUM neurons, the fAHP is about equally dependent on KNa and KCa current, the former activating slightly more rapidly than the latter (which accords to the somewhat different activation kinetics of the Na+ and the Ca2+ currents). The modulations of the P/Q-type Ca2+ current and the Na+ current by AKH I change the initial proportion of KCa to KNa current in favor of the KCa current.
Why counterregulation of KCa and KNa currents?
Would one considers acceleration of spiking as the main effect of AKH I on the DUM neurons, then it remains enigmatic why a whole combination of currents (P/Q-type Ca2+ current, Na+ current, KCa and KNa current) is modulated by the peptide because this does not produce a change in spike frequency. In this light, it would be sufficient to upregulate the Ca2+ background current to get the observed acceleration of spiking.
Spike frequency, however, is not the sole parameter involved in controlling secretion. It is known from vertebrate neurons that presynaptic voltage-gated Ca2+ currents (N and P/Q types) play a central role in triggering transmitter release (Reid et al. 2003). Although we have no evidence that the upregulation of P/Q-type current in the soma of DUM neurons observed with AKH I also occurs in the release zones of its terminals, this would be a possible means to enhance octopamine release. That a modulator may produce the same changes in the soma and the terminals of a neuron is shown by the example of neurons in rat superior cervical ganglion where application of norepinephrine reduces both neurotransmission and somatic Ca2+ currents (Stephens and Mochida 2005). Another important principle involved in the control of transmitter release is the regulation of the BK channel-mediated fAHP. On the one hand, the fAHP can limit transmitter release (Raffaelli et al. 2004). On the other hand, it is necessary for an effective repolarization of action potentials, thus supporting the terminal's capacity to produce full-size action potentials on repetitive firing (e.g., Sausbier et al. 2004). In the case of the DUM neurons, it seems possible that the downregulation of Na+ current and KNa current is necessary to limit the effect of Ca2+ and KCa current modulation on hyperpolarization.
Conclusion
We have demonstrated a considerable complexity behind the "simple" increase in a neuron's firing rate induced by a hormone. Such complexity may be of functional relevance in the terminals of the neuron. Alternatively—or additionally—it could enable the neuron to respond differently to a (more or less) constant synaptic input. Such kind of mechanism was, for example, found for dopamine modulation in lobster stomatogastric neurons (Harris-Warrick et al. 1995). It might also be involved in the adaptation of insect neurons to different behavioral situations as observed, for example, in locust thoracic DUM neurons - ranging from patterned activity to complete inhibitory silencing (Pflüger 1999).
|
APPENDIX |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONAPPENDIXGRANTSREFERENCES |
|---|
We simulated a one-compartment model with each current being described by a Hodgkin-Huxley formalism (Yamada et al. 1998). Beside the leak current, there are 10 currents implemented: One Na+ current (INa), four Ca2+ currents, and 5 K+ currents. The Ca2+ currents are the voltage-independent background current ICa,back, the low-voltage-activated (LVS) current ICa LVA, and the two high-voltage-activated currents ICa P/Q (P/Q-type current) and ICa nP/Q (non-P/Q-type current = total HVA current –P/Q-type current). The K+ currents are the voltage-activated delayed rectifier IKDR and the A-type current IKA, the voltage-independent Na+-activated current IKNa, and the Ca2+-activated current composed of the transient IKCa,t and the sustained IKCa,s. In general, each current I(t) is modeled according to the following equation [G(t) is the conductance, Veq the equilibrium potential, Vm(t) the membrane potential]
 | (A1) |
 | (A2) |
 | (A3) |
 | (A4) |
 | (A5) |
 | (A6) |
 | (A7) |
 | (A8) |
The three ion-activated currents, i.e., IKNa and both the transient and the sustained IKCa, deviated from the preceding description in the following aspects.
First, to model IKNa, the actual Na+-inward current carried by INa was calculated and multiplied with an empirical determined scaling factor called sensitivity factor. This approximation appears to be justified at least for the KNa peak current, which was found to be proportional to the Na current (Fig. 5D). Moreover, using this approximation, we were able to simulate a KNa current with amplitude and kinetic properties similar to the current obtained in the cells. The Na+-inward current times the sensitivity factor was used in Eqs. A5 and A6 instead of the membrane voltage to calculate m and h, respectively. Because the time constants of the KNa current are virtually voltage independent (Fig. A1, E2 and E3), they were fixed, i.e., Eqs. A7 and A8 did not apply. However, the current had a particular delay of 2 ms with respect to the Na-inward current. The sensitivity factor and the delay were adjusted to reproduce time course and size of the modeled current according to the measured current at different voltages (for size, cf. Fig. 5D).
Second, to model IKCa,t and IKCa,s, we had to make some simplifying assumptions. We first considered these currents purely voltage dependent. To separate the currents, we modeled KCa,t with a m2h kinetics and subtracted it from the total KCa current. Under these conditions, we determined the voltage dependence of parameters as given in Table 1 and Fig. A1, F1–F3. These parameters, of course, reflect a "mean" Ca2+ supply through voltage-gated Ca2+ channels. To link the Ca2+ to the KCa current, we determined the effect of increase and decrease of the Ca2+ current (using 10 and 2 mM Ca2+ in the bath solution, respectively) on the KCa current. With this information, we calculated a sensitivity factor linking the amplitudes of KCa,t and KCa,s to the Ca2+ current amplitude (Table 1). For modeling the KCa currents, we made a further simplifying assumption in that we calculated these currents from the actual Ca2+ current. According to the results shown in Fig. 2, we coupled the KCa currents to the P/Q-type Ca2+ current. Although this neglects the fact that also the L-type Ca2+ current provides Ca2+ for the activation of the KCa,s component (Fig. 2, B and D), it is more important for the purpose of this study to investigate the effect of P/Q-type current modulation on KCa currents and spiking. To reproduce the kinetics and voltage dependence of the KCa currents measured in neurons by the modeled currents, we had to introduce a delay (1.5 ms) between Ca2+ current and KCa currents.
The background Ca2+ current, ICa,back, was modeled with a constant, voltage- and ion-independent conductance (cf. Wicher et al. 2004).
The total number of parameters including leak conductance, leak potential and membrane capacitance amounted to 156. The simulation could be run in either the current- or voltage-clamp mode at a variable time-resolution ranging from 1 to 50 µs. With a maximum number of 20,000 time steps, this corresponded to a total simulation time ranging from 20 ms to 1 s. All parameters could be saved in a parameter file in ASCII format, which allowed for easy editing.
In voltage-clamp, the total current, i.e., the sum of all voltage- and ion-activated currents, was calculated as leak-subtracted and without capacitive current
 | (A9) |
 | (A10) |
 |
 |
 |
|
GRANTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONAPPENDIXGRANTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: D. Wicher, Saxon Academy of Sciences, Dept. Neurohormones, Erbertstr. 1, 07743 Jena, Germany (E-mail: b6widi{at}pan.zoo.uni-jena.de)
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONAPPENDIXGRANTSREFERENCES |
|---|
Brezina V, Horn CC, Weiss KR, and Orekhova IV. Modeling neuromuscular modulation in Aplysia. III. Interaction of central motor commands and peripheral modulatory state for optimal behavior. J Neurophysiol 93: 1523–1556, 2005.
Buchholtz F, Golowasch J, Epstein IR, and Marder E. Mathematical model of an identified stomatogastric ganglion neuron. J Neurophysiol 67: 332–340, 1992.
Dale N. Experimentally derived model for the locomotor pattern generator in the Xenopus embryo. J Physiol 489: 489–510, 1995.
Dale N and Kuenzi FM. Ion channels and the control of swimming in the Xenopus embryo. Prog Neurobiol 53: 729–756, 1997.[CrossRef][ISI][Medline]
Derst C, Messutat S, Walther C, Eckert M, Heinemann SH, and Wicher D. The large conductance Ca2+-activated potassium channel (pSlo) of the cockroach Periplaneta americana: structure, localisation in neurones and electrophysiology. Eur J Neurosci 17: 1197–1212, 2003.[CrossRef][ISI][Medline]
