Functional Parameters of Voltage-Activated Ca 2 Currents From Olfactory Interneurons in the Antennal Lobe of Periplaneta americana

Transcription

1 J Neurophysiol 99: , 28. First published November 4, 27; doi:.52/jn Functional Parameters of Voltage-Activated Ca 2 Currents From Olfactory Interneurons in the Antennal Lobe of Periplaneta americana Andreas Husch, Simon Hess, and Peter Kloppenburg Institute of Zoology and Physiology, University of Cologne, Cologne, Germany Submitted 27 June 27; accepted in final form 3 November 27 Husch A, Hess S, Kloppenburg P. Functional parameters of voltageactivated Ca 2 currents from olfactory interneurons in the antennal lobe of Periplaneta americana. J Neurophysiol 99: , 28. First published November 4, 27; doi:.52/jn Toward our goal to better understand the physiological parameters that mediate olfactory information processing on the cellular level, voltage-activated calcium currents (I Ca ) in olfactory interneurons of the antennal lobe from adult cockroaches were analyzed under two conditions: ) in acutely dissociated cells (in vitro) and 2) in an intact brain preparation (in situ). The study included an analysis of modulatory effects of potential inorganic and organic Ca 2 channel blockers. I Ca was isolated and identified using pharmacological, voltage, and ion substitution protocols. I Ca consisted of two components: transient and sustained. The decay of the transient component was largely Ca 2 dependent. In vitro, I Ca had an activation threshold of 5 mv with a maximal peak current at 7 mv and a half-maximal voltage (V.5act ) for tail-current activation of 8 mv. In situ these parameters were significantly shifted to more depolarized membrane potentials: I Ca activated at 4 mv with a maximal peak current at 8 mv and a V.5act for tail-current activation of mv. The sensitivity of I Ca to the divalent cations Cd 2, Co 2, and Ni 2 was dose dependent. The most effective blocker was Cd 2 with an IC 5 of 5 M followed by Ni 2 (IC M) and Co 2 (IC M). The organic channel blockers verapamil, diltiazem, and nifedipine also blocked I Ca in a dose-dependent way and had differential effects on the current waveform. blocked I Ca with an IC 5 of.5 4 M and diltiazem had an IC 5 of M. Nifedipine blocked I Ca by 33% at a concentration of 4 M. INTRODUCTION Address for reprint requests and other correspondence: P. Kloppenburg, University of Cologne, Institute of Zoology and Physiology, Weyertal 9, 593 Cologne, Germany ( peter.kloppenburg@uni-koeln.de). The first-order synaptic relay in olfactory systems of vertebrate and invertebrate animals have striking similarities of glomerular and neuronal organization, suggesting that olfactory information is processed through similar mechanisms in these evolutionarily remote animals (Hildebrand and Shepherd 997; Strausfeld and Hildebrand 999; Wilson and Mainen 26). One experimental system that has served very successfully as a model to understand olfactory information processing is the first-order olfactory relay or antennal lobe (AL) of insects (see Laurent 999; Wilson and Mainen 26). As an important step toward the long-term goal to better understand also the cellular mechanisms that mediate olfactory information processing we characterized the physiological and biophysical properties of voltage-activated Ca 2 currents in olfactory interneurons from the ALs of adult Periplaneta americana. Calcium plays a critical role in the control of a variety of neuronal processes such as synaptic release, membrane excitability, enzyme activation, and activity-dependent gene activation (Augustine et al. 23; Berridge 993). A main source of cytoplasmic Ca 2 that contributes significantly to the dynamics of intracellular Ca 2 signals is the Ca 2 influx through voltage-gated Ca 2 channels (VGCCs). Multiple types of voltagegated Ca 2 channels, characterized by different functional properties, are usually differentially distributed in functionally specialized subcellular compartments of the neuron and contribute to its whole cell Ca 2 current. Characterized by their physiological and biophysical phenotypes the following voltage-gated Ca 2 channel types can be presently distinguished in vertebrates: low-voltage activated (LVA) channels (T-type, Ca v 3. Ca v 3.3) and high-voltage activated (HVA) channels (L-, N-, P/Q-, R-type, Ca v. Ca v.4, Ca v 2. Ca v 2.3). A comparison of calcium channel gene sequences indicates that in certain structural aspects insect VGCCs might resemble the vertebrate VGCCs (Littleton and Ganetzky 2). Despite the similarities of Ca 2 channel sequences in invertebrates and vertebrates, invertebrate channels differ greatly in their pharmacological profile. For instance, one of the characteristics of L-type channels in vertebrates is their sensitivity to,4- dihydropyridines (e.g., nifedipine), whereas most invertebrate channels with homologous L-type like sequences lack this feature (for review see Jeziorski et al. 2; Wicher et al. 2). In the insect CNS, VGCCs can be separated electrophysiologically into LVA or mid-lva (M-LVA) and HVA calcium channels (Grolleau and Lapied 996; Wicher and Penzlin 997). It has been demonstrated that the LVA current in dorsal unpaired medican (DUM) neurons of P. americana could be further bisected into two components according to their sensitivity to Ni 2 ions: a transient (tlva) and a maintained LVA current (mlva; Grolleau and Lapied 996). In DUM neurons, LVA currents start to activate at 8 mv, M-LVA at 5 mv, and HVA currents at 4 mv. M-LVA and HVA currents were further characterized by their differential sensitivity to inorganic ions (Ni 2,Cd 2 ) and peptide toxins (conotoxins and agatoxins; Wicher and Penzlin 997). Differential sensitivity of HVA currents in embryonic brain neurons from P. americana to peptide toxins suggests two current components resembling the vertebrate P/Q- and R-type channels (Benquet et al. 999). Organic Ca 2 channel blockers that selectively block specific Ca 2 channel types in vertebrates also modify Ca 2 currents in insects. Earlier studies have demonstrated that Ca 2 currents in insect neurons are reduced by phenylalkylamines (PAAs; e.g., verapamil; Wicher and Penzlin 997), benzothia- The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked advertisement in accordance with 8 U.S.C. Section 734 solely to indicate this fact /8 $8. Copyright 28 The American Physiological Society

2 Ca 2 CURRENTS IN OLFACTORY INTERNEURONS 32 zepines (BZTs; e.g., diltiazem; David and Pitman 995),,4- dihydropyridines (DHPs; e.g., nifedipine; Schäfer et al. 994), and amiloride (Baines and Bate 998). Investigations in DUM neurons of P. americana showed that HVA currents are affected by verapamil and diltiazem, but not by nifedipine and amiloride (Wicher and Penzlin 997). However, nifedipine partially blocked barium currents in embryonic brain neurons of P. americana (Benquet et al. 999). In motoneurons of P. americana Ca 2 current components could be separated by their sensitivity to nifedipine (Mills and Pitman 997). Detailed analyses, however, of dose response relationships and effects of organic channel blockers on biophysical properties of Ca 2 currents in insect neurons are still lacking. The goal of this study was ) to characterize voltageactivated Ca 2 currents from olfactory interneurons of the antennal lobe in vitro and in situ and 2) to investigate the effects of some Ca 2 channel blockers that have been shown to be effective in vertebrates. The focus for the latter was on verapamil, diltiazem, and nifedipine, all of which selectively block vertebrate L-type channels and are members of different chemical classes. Such detailed knowledge on these readily available blockers could be helpful to block specific components of the calcium currents to better analyze their physiological function in neuronal information processing. METHODS Animals and materials P. americana were reared in crowded colonies at 27 C under a 3: h light/dark photoperiod regimen and reared on a diet of dry dog food, oatmeal, and water. All experiments were performed with adult animals of both sexes. Before dissection the animals were anesthetized by CO 2 or cooling (4 C) for several minutes. For cell culture they were then adhered in a plastic tube with adhesive tape and the heads were immobilized using dental modeling wax (S-U Modelierwachs, Schuler-Dental, Ulm, Germany) with a low solidification point (57 C). For in situ experiments the animals were placed in a custombuilt holder, and the head was immobilized with dental wax. The antennae were placed in small tubes on a plastic ring that was later used to transfer the preparation to the recording chamber. All chemicals, unless stated otherwise, were obtained from Aplichem (Darmstadt, Germany) or Sigma Aldrich (Taufkirchen, Germany) with a per-analysis purity grade. Cell culture To examine the electrophysiological properties of isolated antennal lobe neurons, cells were dissociated and cultured using modified protocols reported previously (Grolleau and Lapied 996; Hayashi and Hildebrand 99; Kirchhof and Mercer 997). The head capsule of anesthetized animals was opened and the antennal lobes were dissected with fine forceps. Typically, ALs from eight animals were pooled in sterile culture saline (kept on ice) containing (in mm): 85 NaCl, 4 KCl, 6 CaCl 2, 2 MgCl 2,35D-glucose, HEPES, and 5% fetal bovine serum (S-, c.c.pro, Neustadt, Germany), adjusted to ph 7.2 (with NaOH), which resulted in an osmolarity of 42 mosm. For dissociation the ALs were transferred for 2 min at 37 C into 5 l Hanks Ca 2 - and Mg 2 -free buffered salt solution (47, GIBCO, Invitrogen, Karlsruhe, Germany) containing (in mm): HEPES, 3 sucrose, 8 units ml collagenase (LS494, Worthington, Lakewood, NJ), and.7 units ml dispase (LS2, Worthington), adjusted to ph 7.2 (with NaOH) and to 45 mosm (with sucrose). Dissociation of neurons was aided by careful trituration with a fire-polished Pasteur pipette for 3 5 min. Enzyme treatment was terminated by cooling and centrifuging the cells twice through 6 ml of culture medium (4 C, 48 g, 5 min). The culture medium consisted of five parts Schneider s Drosophila medium (272, GIBCO) and four parts Minimum Essential Medium (2575, GIBCO) to which was added (in mm): HEPES, 5 glucose, fructose, 6 sucrose, and 5% fetal bovine serum, adjusted to ph 7.5 (with NaOH) and 43 mosm (with sucrose). After centrifugation, the cells were resuspended in a small volume of culture medium ( l per dish, eight ALs were plated in five to six dishes), and allowed to settle for 2 h to adhere to the surface of the culture dishes coated with concanavalin A (C-2, Sigma,.7 mg ml dissolved in H 2 O). The cultures were placed in an incubator at 26 C and then used for electrophysiological experiments on the same day. For recordings the cells were visualized with an inverted microscope (IX7, Olympus, Hamburg, Germany) using a 4 water objective (UAPO 4,.5 NA,.25 mm WD, Olympus) and phase-contrast optics. Intact brain preparation The intact brain preparation was based on an approach described by Kloppenburg et al. (999a,b), in which the central olfactory network was left intact. Shortly before the experiment, the head capsule of the anesthetized animal was opened by cutting a window between the two compound eyes and the bases of the antennae. The brain with antennal nerves and antennae attached was dissected from the head capsule and pinned with fine wire in a Sylgard-coated (Dow Corning, Midland, MI) recording chamber containing normal saline (see following text). To gain better access to the recording site and facilitate the penetration of pharmacological agents into the tissue, the brain was enzyme treated (papain, P4762, Sigma,.3 mg ml and L-cysteine, 39, Fluka/Sigma, mg ml dissolved in normal saline) for about 3 min at room temperature before the AL was desheathed using fine forceps. The AL neurons were visualized with a fixed-stage upright microscope (BX5WI, Olympus) using a 4 water-immersion objective (UMPLFL 4,.8 NA, 3.3 mm WD, Olympus) and IR-DIC optics (Dodt and Zieglgänsberger 994). Whole cell recordings Whole cell recordings were performed at 24 C following the methods described by Hamill et al. (98). Electrodes (tip resistance between 3 and 5 M ) were fashioned from borosilicate glass (.86 mm OD,.5 mm ID, GB5-8P, Science Products, Hofheim, Germany) with a temperature-controlled pipette puller (PIP5, HEKA Elektronik, Lambrecht, Germany), and filled with a solution containing (in mm): 9 CsCl, NaCl, CaCl 2, 2 MgCl 2, HEPES, and EGTA adjusted to ph 7.2 (with NaOH), resulting in an osmolarity of 45 mosm. During the experiments, if not stated otherwise, the cells were superfused constantly with saline solution containing (in mm): 85 NaCl, 4 KCl, 6 CaCl 2, 2 MgCl 2, HEPES, and 5 glucose. The solution was adjusted to ph 7.2 (with NaOH) and to 43 mosm (with glucose). To isolate the Ca 2 currents we used a combination of pharmacological blockers and ion substitution that has been shown to be effective in other insect preparations (Kloppenburg and Hörner 998; Kloppenburg et al. 999b; Schäfer et al. 994). Transient voltagegated sodium currents were blocked by tetrodotoxin (TTX, 7 to 4 M, T-55, Alomone, Jerusalem, Israel). 4-Aminopyridine (4- AP, 4 3 M, A7843, Sigma) was used to block transient K currents (I A ; nomenclature adapted from Connor and Stevens 97) and tetraethylammonium (TEA, 3 3 M, T2265, Sigma) blocked sustained K currents [I K(V) ] as well as Ca 2 -activated K currents [I K(Ca) ]. In addition the pipette solution did not contain potassium. Whole cell voltage-clamp recordings were made with an EPC9 patch-clamp amplifier (HEKA Elektronik) that was controlled by the

3 322 A. HUSCH, S. HESS, AND P. KLOPPENBURG program Pulse (version 8.63, HEKA Elektronik) running under Windows. The electrophysiological data were sampled at intervals of s ( khz), except the 5-ms tail current measurements were sampled at 2 khz. The recordings were low-pass filtered at 2 khz with a four-pole Bessel filter. Compensation of the offset potential and capacitance were performed using the automatic mode of the EPC9 amplifier. The liquid junction potential between intracellular and extracellular solution (see Neher 992) of 4.8 mv [calculated with Patcher s PowerTools plug-in from abteilungen/4/software/index.html for Igor Pro (WaveMetrics, Portland, OR)] was also compensated. To remove uncompensated leakage and capacitive currents, a p/6 protocol was used (see Armstrong and Bezanilla 974). Voltage errors due to series resistance (R S ) were minimized using the R S compensation of the EPC9. R S was compensated between 3 and 7% with a time constant ( ) of2 s. Stimulus protocols used for each set of experiments are provided in the RESULTS. Organic Ca 2 channel modulators The following organic Ca 2 channel modulators that modify L-type andt-type VGCCs in vertebrate preparations were used in this study: ( )-verapamil (V4629, Sigma), ( )-cis-diltiazem (D252, Sigma), nifedipine (N7634, Sigma), amiloride (A74, Sigma), and ( )-BAY K 8644 (B-35, Alomone). ( )- and ( )-cisdiltiazem were applied in concentrations ranging from 3 to 6 M. Nifedipine was dissolved in dimethyl sulfoxide (DMSO; D848, Sigma) and then added to the saline for final concentrations from 4 to 6 M. BAY K 8644 was tested at a concentration of 4 M. First, it was dissolved in DMSO and then stored in aliquots at 2 C. The aliquot was added to the saline shortly before the experiments were conducted. The DMSO concentration of both the nifedipine- and the BAY K 8644 containing salines was.5%. At this concentration it had no obvious effect on I Ca and was also added to the control, although it increased the osmolarity by about 5 mosm. Due to its photosensitivity (see product sheets from Sigma and Alomone, respectively), all experiments with nifedipine and BAY K 8644 were performed in the dark. Amiloride was dissolved in gently heated saline and used at a concentration of 3 M. Drug-containing as well as control saline were bath applied at a rate of 4 7 ml min. Data analysis The data from the dose response experiments were fit with a hill equation of the form I I max IC n H () 5 C where I is the peak amplitude of I Ca at a 5-mV test pulse from V h 6 mv in the presence of different concentrations of drugs ([C]), and I max is the peak amplitude of the control. IC 5 is the concentration where half of I Ca is blocked and n H is the Hill coefficient. Steady-state tail-current activation and steady-state inactivation data were fit using a first-order (n ) Boltzmann equation I (2) I max e V.5 V /s n where I max is the maximal current, V is the voltage of the test pulse, I is the current at voltage V, s is the slope factor, and V.5 is the voltage at which half-maximal activation occurs. To convert peak current to peak conductance we assumed an E Ca 6 mv (estimated with MaxChelator, cpatton/maxc.html; Patton et al. 24). For analysis of electrophysiological data we used the software Pulse (version 8.63, HEKA Elektronik), Igor Pro 4 (WaveMetrics, including the Patcher s PowerTools plug-ins), and Sigma Stat (Systat Software, Erkrath, Germany). All calculated values are expressed as means SD. Significance of differences between mean values were evaluated with Mann Whitney rank-sum tests or paired t-tests. Significance was accepted at P.5. RESULTS Whole cell voltage-clamp recordings were used to characterize voltage-dependent Ca 2 currents of olfactory interneurons in antennal lobe interneurons of P. americana under two conditions: in acutely dissociated cells (in vitro) and in an intact brain preparation (in situ). The study was expanded by an in vitro analysis of modulatory effects of potential inorganic and organic Ca 2 channel blockers. The recorded neurons were identified as antennal lobe interneurons by the position of their cell bodies within the antennal lobe. They were not unequivocally classified as local interneurons or projection (output) neurons. Besides the receptor neurons the main neuronal components of the AL are the local interneurons ( 5) and the projection neurons ( 5; numbers from Boeckh and Ernst 987). It can be expected that the different physiological and computational tasks of these neurons are reflected in the biophysical properties, which might cause a certain variability of the pooled data. If not stated otherwise, the membrane potential was clamped at 6 mv, which is in the range of the normal resting potential of these neurons in situ ( mv; n 28). Typically, depolarizing voltage steps evoked a combination of inward and outward currents in AL neurons. When voltage-gated Na,K, and H currents were reduced by ion substitution, pharmacological reagents, and appropriate voltage protocols, we recorded inward currents that had the characteristics of typical Ca 2 currents. Our recording conditions were designed to drastically minimize, or even abolish, outward currents through K channels and in most, but not all, recordings we did not observe outward currents. For recordings in which residual outward currents at strong depolarizations were detected we could not rule out that the K currents were not completely blocked and that there was some residual K in the neurons. However, we think it is more likely that the observed outward currents were carried by Cs through Ca 2 channels, which are not completely impermeable to Cs (Hess et al. 986). This scenario was supported by our observation that no residual outward currents could be detected when I Ca was blocked by Cd 2 (data not shown). To analyze I Ca, outward currents were minimized by substituting K in the pipette solution with Cs and adding 3 2 M TEA and 4 3 M 4-AP to the extracellular solution. The remaining inward current consisted of a transient very fast activating/inactivating component and a more slowly inactivating component (Fig. A), that did not completely inactivate even after long ( s) depolarizations (Fig. B). By using pharmacological blockers and ion substitution, the main charge carrier of both components could be separated and identified. The fast transient component was a TTX-sensitive sodium current (Fig. A), whereas the other components of the inward current were identified as voltage-activated Ca 2 conductances (see following text). To measure I Ca, the neurons were superfused with saline containing 7 M TTX, 4 3 M 4-AP, and 3 2 M TEA. In the pipette solution K was replaced with Cs.

4 Ca 2 CURRENTS IN OLFACTORY INTERNEURONS 323 A inward currents +TTX I Ca I Ca -5 mv na I Na B -5 mv ms FIG.. Whole cell recording in vitro of voltage-activated inward currents. A depolarizing voltage step (5 ms) from a holding potential of 6 to 5mV evoked a combination of inward and outward currents. Outward currents were blocked by substituting K in the pipette solution with Cs and adding 3 2 M tetraethylammonium (TEA) and 4 3 M 4-aminopyridine (4-AP) to the extracellular solution. The remaining inward current consisted of a transient very fast activating/inactivating component and a more slowly inactivating component. A: the fast transient component was a tetrodotoxin (TTX)- sensitive Na current, whereas the other components of the inward current were identified as voltage-activated Ca 2 conductances, I Ca (see Fig. 2). B: I Ca consisted of an inactivating and sustained component, which did not completely inactivate even after long ( s) depolarization. na Charge carrier Experiments with varying extracellular Ca 2 concentrations confirmed that Ca 2 was the charge carrier of the investigated inward current. Low extracellular Ca 2 concentrations acted quickly to reduce the inward current reversibly (n 5; Fig. 2A). This reduction of I Ca was concentration dependent. When EGTA containing Ca 2 -free extracellular solution was applied, the inward current was completely abolished (data not shown). When extracellular Ca 2 was substituted with Ba 2 (Fig. 2, B E), the maximal I Ca was enhanced by about 2% (Fig. 2, B and D), indicating that the channels are more permeable to Ba 2 than to Ca 2, as described for Ca 2 channels in other cell types (Hille 2). Simultaneously the current voltage (I/V) relation shifted by about mv (n 4) to more hyperpolarized membrane potentials (Fig. 2D), which might result from changes in the surface charge of the cell membrane (Fedulova et al. 985). The resulting voltage-dependent differences between I Ca and I Ba are demonstrated by steps to 2 and 5 mv in Fig. 2, C and C2, respectively. Furthermore, the amount of inactivation during a depolarizing voltage pulse was reduced, which became especially obvious during long-lasting depolarizations ( s; 5 mv; Fig. 2E). These results suggest that inactivation during a voltage step is, at least in part, a Ca 2 -dependent mechanism. I Ca in vitro The characteristics of I Ca are shown in Fig. 3. The I/V relationship of the peak currents was determined by increasing FIG. 2. Ca 2 is the charge carrier of the TTX-insensitive voltage-activated inward currents. All experiments are in vitro recordings. The cells were held at 6 mv. A: experiments with varying extracellular Ca 2 concentrations indicated that Ca 2 was the main charge carrier of the investigated inward current. A: inward currents elicited by 5-ms voltage steps to 5 mv were reduced reversibly by decreasing the extracellular Ca 2 concentration from 6 3 to 2 3 M(n 5). A2: time course of peak inward current during a decrease in extracellular calcium concentration, from 6 3 to 2 3 M. B: voltage-activated inward currents when Ca 2 was the main charge carrier (I Ca ) and when Ca 2 was substituted with Ba 2 (I Ba ). Each series represents current responses to increasing voltage steps between 6 and 5 mv in 5-mV increments. When extracellular Ca 2 was substituted with Ba 2 the peak amplitude was enhanced by about 2%. C: currents elicited by voltage steps to 2 mv (C) and 5 mv(c2) demonstrate the difference between I Ca and I Ba at different command potentials. D: the current voltage (I/V) relation of I Ba was shifted by about mv to more hyperpolarized membrane potentials compared with I Ca (n 4). E: I Ca and I Ba during a long-lasting depolarization ( s; 5 mv) demonstrating decreased inactivation for I Ba.

5 324 A. HUSCH, S. HESS, AND P. KLOPPENBURG FIG. 3. Calcium currents (I Ca ) in vitro. A C: example current traces for steady-state activation, steadystate activation of tail currents, and steady-state inactivation, respectively. A: steady-state activation. The holding potential was 6 mv and I Ca was evoked by 5-ms voltage steps from 6 to 4 mv in 5-mV increments. B: tail currents. The holding potential was 8 mv and tail currents were evoked by 5-ms voltage steps from 8 to 4 mv in 5-mV increments. C: steady-state inactivation. The holding potential was 6 mv. Test pulses to 5 mv (5 ms) were preceded by 5-ms pulses ranging from 95 to 5 mv in 5-mV increments. D: voltage dependence of peak I Ca from 65 neurons. D: data from single neurons. D2: the averaged data. The mean peak amplitude (I max )is.7.6 na. E: I/V relation of peak I Ca normalized to the maximal current of each cell, I max. E: data from single neurons. E2: the averaged data. The current is activated at command potentials more depolarized than 5 mv with a maximum at mv (n 65). F: current density/voltage relation. Current density was calculated from the ratio of I Ca and the cell s capacitance. F: data from single neurons. F2: the averaged data. The mean maximal current density was pa/pf. G: I/V relation of tail currents normalized to the maximal tail current of each cell. Curves are fits to a first-order Boltzmann relation (Eq. 2). The mean maximal tail current is.9.4 na (n 2). H: I/V relations for steady-state inactivation of neurons. Curves are fits to a first-order Boltzmann relation (Eq. 2). I: mean I/V relations of steady-state inactivation of peak I Ca (filled squares) and tail-current activation (filled circles). The curves are fits to first-order Boltzmann equations (Eq. 2) with the following parameters: Tail-current activation: V.5act mv; s act ; n 2. Steady-state inactivation: V.5inact mv; s inact 8.7 2; n. voltage steps (5 ms, 5 mv) between 6 and 4 mv from a holding potential of 6 mv (Fig. 3A). The voltage dependence of activation of I Ca was determined from tail currents that were evoked by 5-ms voltage steps from 8 mv holding potential to 4 mv in -mv increments (Fig. 3B). The I/V relations were fit to a first-order Boltzmann equation (Eq. 2; Fig. 3, G and I). Steady-state inactivation of I Ca was measured from a holding potential of 6 mv. Prepulses (5 ms) were delivered in 5-mV increments from 95 to 5 mv, followed by a 5-ms test pulse to 5 mv, and the peak currents were determined (Fig. 3C). The I/V relations were fit to a first-order Boltzmann equation (Eq. 2; Fig. 3, H and I).

6 Ca 2 CURRENTS IN OLFACTORY INTERNEURONS 325 In situ A steady state activation B tail currents C steady state inactivation 2 ms 5 pa 4 mv 4 mv 5 mv n=22-8 mv D I E F I [na] Ca D2 E2 F2 I [na] G H I current density [pa/pf] current density [pa/pf] mv I Ca /C FIG. 4. Calcium currents (I Ca ) in the intact brain preparation (in situ). For details see legend of Fig. 3. A C: current traces for steady-state activation, steadystate activation of tail currents, and steady-state inactivation, respectively. The voltage step for steady-state inactivation was 5 mv. D: voltage dependence of peak I Ca of 22 neurons. D: data from single neurons. D2: the averaged data. The current is activated at command potentials more depolarized than 4 mv with a maximum around mv (n 22). The mean peak amplitude (I max )is.2.4 na. E: current density/ voltage (I/V) relation. The mean maximal current density was pa/pf. F: I/V relation of peak I Ca normalized to the maximal current of each cell, I max.on average the mean maximal current (I max )of.2.4 na is reached at a membrane potential (E max )of8 7.8 mv. G: I/V relation of tail currents normalized to the maximal tail current of each cell. The mean maximal tail current is.8.4 na (n 3). H: I/V relations for steady-state inactivation of 8 neurons. I: mean I/V relations of steady-state inactivation of peak I Ca (filled squares) and tail-current activation (filled circles). The Boltzmann fits have the following parameters: Tailcurrent activation: V.5act.5 6 mv; s act 7.5.8; n 3. Steady-state inactivation: V.5inact mv; s inact 8.7.9; n n= n= During a depolarizing voltage step I Ca activated relatively quickly and decayed during a maintained voltage step (Fig. 3A). The current waveforms and I/V relations for activation were typical for I Ca, but varied between cells (Fig. 3, D F). In vitro I Ca started to activate with voltage steps more depolarized than 5 mv (Fig. 3D). The mean peak currents reached a maximum amplitude (I max )of.7.6 na (Fig. 3D2) at mv (n 65; Fig. 3E) and decreased during more positive test pulses as they approached the calcium equilibrium potential (Fig. 3D). Given a mean whole cell capacitance of

7 326 A. HUSCH, S. HESS, AND P. KLOPPENBURG A B -5 M Cadmium wash control 5 pa -3 M Nickel A2 I [na] B2 I [na] M Cadmium 2 4 t [min] -3 M Nickel pf (n 65), this corresponds to a mean current density of pa/pf (Fig. 3F). The activation and inactivation kinetics during a voltage step are voltage dependent (Fig. 3, A and B); the time to peak current and the time constant for the decay during a voltage pulse decreased when voltage steps of increasing amplitude were applied. These parameters, however, were not analyzed quantitatively. The tail currents that are independent of the changing driving force during the series of voltage pulses had a maximum amplitude of.9.4 na (n 2). This corresponds to a mean maximal conductance (G max ) of ns and a mean current density of pa/pf. The I/V relation of the tail currents was fit by a first-order Boltzmann equation with a mean voltage for half-maximal activation (V.5act )of mv (s ; n 2; Fig. 3, G and I). Steady-state inactivation started in vitro at prepulse potentials around 6 mv and increased with the amplitude of the depolarizing prepulse (Fig. 3, C, H, and I). The I/V relationship was fit with a first-order Boltzmann equation (Eq. 2) with a voltage for half-maximal inactivation (V.5inact )of mv (s 8.7 2; n ; Fig. 3, H and I). C D Block of I Ca [%] wash control 5 pa 2-3 M Cobalt wash control 5 pa Cadmium C2 I [na] Concentration [M] t [min] 2-3 M Cobalt t [min] Nickel Cobalt 8 2 I Ca in situ I Ca recorded in situ (Fig. 4) showed characteristics similar to those in vitro. However, the I/V relationships had a larger variability and the means were shifted significantly to more depolarized membrane potentials. The mean voltage for activation of the maximal peak current was shifted by 4.4 mv (P.). The voltage for half-maximal tail-current activation (V.5act ) was shifted by 7.3 mv (P.). In situ, I Ca started to activate at command potentials more depolarized than 4 mv (Fig. 4D). The mean peak currents reached a maximum amplitude (I max )of.2.4 na (n 22; Fig. 4D) at mv (Fig. 4E). Based on a mean whole cell capacitance of pf, this corresponds to a mean current density of pa/pf (Fig. 4F). The tail currents had a mean maximum of.8.4 na and a mean voltage for half-maximal activation (V.5act )of.5 6mV (s 7.5.8; n 3; Fig. 4, G and I). In situ steady-state inactivation started at prepulse potentials around 6 mv and had a mean voltage for half-maximal inactivation (V.5inact )of mv (s 8.7.9; n 8; Fig. 4, H and I). FIG. 5. The divalent inorganic ions Cd 2 (A), Ni 2 (B), and Co 2 (D) reduce I Ca in a concentration-dependent way. A: effect of 5 MCd 2 on I Ca. A: I Ca evoked by a 5-ms voltage step to 5 mv from a holding potential of 6 mv before (control), during, and after (wash) application of 5 M Cd 2. A2: time course of Cd 2 ( 5 M) induced reduction of peak I Ca. The black bar indicates the time of Cd 2 application. B: effect of 3 MNi 2 on I Ca. B: I Ca before (control), during, and after (wash) application of 3 M Ni 2. B2: time course of Ni 2 ( 3 M) induced reduction of peak I Ca. The black bar indicates the time of Ni 2 application. C: effect of 2 3 MCo 2 on I Ca. C: I Ca before (control), during, and after (wash) application of 2 3 MCo 2. C2: time course of Co 2 (2 3 M) induced reduction of peak I Ca. The black bar indicates the time of Co 2 application. D: dose response curves for the aforementioned blockers. Curves are fits by a Hill function (Eq. ) with the following parameters: cadmium: IC 5 5 M; n H.87; nickel: IC M; n H.; cobalt: IC M; n H.4; h 3 for all data points.

8 Ca 2 CURRENTS IN OLFACTORY INTERNEURONS 327 Inorganic ions and organic Ca 2 channel blockers A B -4 M -5 M C -3 M -5 mv M Block of I Ca [%] M A Concentration [M].5-4 M (IC 5 ) +5 mv C2 na τ [msec] n=4 (.5-4 M) In a series of in vitro experiments we analyzed the effect of several inorganic ions and organic substances on I Ca, which block or enhance voltage-gated calcium currents in vertebrate preparations. I Ca was blocked by the divalent cations Cd 2, Ni 2, and Co 2. We also found that verapamil, diltiazem, and nifedipine, which all belong to different chemical classes (phenylalkalamine, benzothiazepine, and,4-dihydropyridine, respectively) and are known to selectively block vertebrate L-type channels, differentially modify I Ca. Amiloride ( 3 M), a T-type channel blocker, and ( )-BAY K 8644 ( 4 M), a,4-dihydropyridine and L-type channel agonist, had no effect (data not shown). INORGANIC IONS. The divalent cations Cd 2,Co 2, and Ni 2 blocked I Ca in a dose-dependent way. Figure 5, A C shows single example experiments for each of the blockers and Fig. 5D shows the dose-inhibition curves for all three cations. I Ca was elicited by a 5-ms voltage pulse to 5 mv from a holding potential of 6 mv. The dose-inhibition curves were well fit with a Hill equation (Eq. ). The most effective blocker was Cd 2 with a half-inactivating concentration (IC 5 )of 5 M (Hill coefficient n H.87) followed by Ni 2 (IC M; n H.) and Co 2 (IC M; n H.4). ORGANIC CA 2 CHANNEL BLOCKERS., diltiazem, and nifedipine effected I Ca with different concentration dependencies. For verapamil and diltiazem we were able to measure dose-inhibition curves and determine the concentrations for full and half-maximal block of the peak current (IC 5 ). Nifedipine was soluble only in concentrations below its IC 5.Ifnot stated otherwise, any further experiments were conducted at the respective IC 5 or, in the case of nifedipine, at the maximal usable concentration ( 4 M). To compare the voltage dependence for steady-state activation and inactivation we compared D n= n=5 n=5 E n=9 FIG. 6. Effect of verapamil on I Ca. In all experiments except the doseinhibition curve verapamil was applied at its IC 5 (.5 4 M; see Fig. 6A). The cells were held at 6 mv. A: the reduction of I Ca by verapamil is dose dependent. A: the current traces are from different neurons treated with increasing verapamil concentrations ( 6 to 3 M). Each trace is normalized to its control (%). I Ca was evoked by 5-ms depolarizing voltage steps to 5 mv. A2: dose-inhibition curve of verapamil on peak I Ca (n is given for each data point). The curve is a fit to a Hill function (Eq. ) with IC M and n H.55. B: current traces of I Ca before and during application of.5 4 M verapamil. Each series represents currents evoked by 5-ms voltage steps to 5 mv in 5-mV increments. Inset: the inactivation of I Ca during a sustained voltage step was increased during verapamil application. The decay time constant (from a monoexponential fit) of I Ca, which was evoked by 5-ms voltage steps to 5 mv, changed significantly from ms (control) to 4.2 ms during application of verapamil (P.6; n 4). C: I/V relations of I Ca from 5 neurons before (squares) and during application of verapamil (triangles). C: the data that are normalized to the controls from single neurons. C2: the averaged data scaled to the controls. did not change the voltage dependence of peak-current activation. D and E: I/V relations for tail-current activation and steady-state inactivation of I Ca under control conditions (solid squares) and in verapamil (solid triangles). The open triangles represent the currents scaled to the control. For voltage protocols see legend of Fig. 3. Curves are fits to a first-order Boltzmann relation (Eq. 2) with the following parameters. D: tail-current activation was not significantly changed by verapamil (n 5): : V.5act mv; s act : V.5act mv; s act E: steady-state inactivation was significantly changed by verapamil (P.5; n 5): : V.5inact mv; s inact : V.5inact mv; s inact

9 328 A. HUSCH, S. HESS, AND P. KLOPPENBURG the I/V relation for tail-current activation and peak-current inactivation that were both well fit by a Boltzmann function. A 5-4 M M B C D M -5 mv -4 M M (IC 5 ) +5 mv -2 M, -6 M.5 A2 Block of I Ca [%] C2 E na τ [msec] Concentration [M] n=6 (2.5-4 M) n=6 n= n= n= VERAPAMIL. The phenylalkalamine verapamil was tested at concentrations ranging from 6 to 3 M. Example experiments demonstrating the verapamil effect are shown in Fig. 6A and the dose response curve is given in Fig. 6A2. An obvious effect of verapamil started at 4 M and a nearly complete block was achieved at 3 M (Fig. 6). The effects reached a steady state within 3 5 min. The dose-inhibition curve was well fit by a Hill equation (Eq. ) with a half-maximal block of I Ca at.5 4 M(n H.55; Fig. 6A2). increased the decay rate of I Ca during depolarizing voltage pulses in a dose-dependent manner with a maximum effect at concentrations between 4 and M. For example, at the IC 5 verapamil decreased the time constant for the decay (from a monoexponential fit) during a voltage pulse to 5 mv from to 4.2 ms(p.6; n 4; Fig. 6B and inset). The voltage dependence of the peak- and tail-current activation was not modified by verapamil (Fig. 6, C, C2, and D). However, the voltage dependence for steady-state inactivation of I Ca was changed (Fig. 6E): The mean voltage for halfmaximal inactivation (V.5inact ) was shifted to hyperpolarized membrane potentials from mv (s ) to mv (s 6.5.5; P.5; n 9). DILTIAZEM. The benzothiazepine diltiazem was tested at concentrations ranging from 6 to 3 M. Example experiments demonstrating the effect of diltiazem are shown in Fig. 7A and the dose response curve is given in Fig. 7A2. The effects reached a steady state within 3 5 min. A clear diltiazem effect started at 4 M. The dose-inhibition curve was well fit by a Hill equation (Eq. ) with a half-maximal block of I Ca at a concentration of M(n H.6; Fig. 7A2). The I Ca block was accompanied by a change in the waveform of I Ca (Fig. 7B and inset). During a depolarizing voltage pulse diltiazem increased the rate of inactivation. The mean time constant ( ) of the decay of I Ca (from a monoexponential fit) during a voltage pulse to 5 mv was significantly increased from to ms (P.3; n 6) when the IC 5 of diltiazem (3 5 M) was applied. This effect was FIG. 7. Effect of diltiazem on I Ca. In all experiments except the doseinhibition curve diltiazem was applied near its IC 5 (2.5 4 M; see A). The cells were held at 6 mv. For details see the legend of Fig. 6. A: the reduction of I Ca by diltiazem is dose dependent. was applied in the range of 6 to 2 M. A2: the curve is a fit to a Hill function (Eq. ) with IC M and n H.6. B: current traces of I Ca before and during application of diltiazem. Inset: the inactivation of I Ca during a sustained voltage step was decreased during diltiazem application. The decay time constant (from a monoexponential fit) of I Ca changed significantly from ms (control) to ms (P.3; n 6) during application of diltiazem. C: I/V relations of I Ca from 6 neurons before (squares) and during application of diltiazem (triangles). The diltiazem-insensitive current started to activate at more hyperpolarized potentials than the controls (for quantitative analysis see Fig. 6E). D and E: I/V relation for tail-current activation and steady-state inactivation of I Ca under control conditions (solid squares) and in diltiazem (solid triangles). The Boltzmann fits have the following parameters. D: tail-current activation was significantly changed by diltiazem (P.; n 8): : V.5act mv; s act : V.5act mv; s act E: steady-state inactivation was significantly changed by diltiazem (P.; n 7): : V.5inact mv; s inact : V.5inact mv; s inact 7..3.

10 Ca 2 CURRENTS IN OLFACTORY INTERNEURONS 329 even more prominent with higher concentrations of diltiazem (data not shown). The I Ca component that was not blocked by diltiazem at or above its IC 5 was activated at slightly, but significantly more hyperpolarized potentials (Fig. 7, C2 and D). Accordingly, the mean V.5act for steady-state activation of the tail currents A B -4 M Nifedipine C -5 M D M, -6 M -5 mv Nifedipine Nifedipine A2 Block of I Ca [%] mv C2 E na Nifedipine Nifedipine n=6 n= Concentration [M] Nifedipine n=5 n= (from a first-order Boltzmann fit) was shifted from mv (s 5.4.3) in the control to mv (s 5.5 ; P.; n 8) during diltiazem application (Fig. 7D). In addition, diltiazem shifted the voltage for half-maximal voltage for inactivation (V.5inact ; from a first-order Boltzmann fit) to more hyperpolarized membrane potentials from mv (s 9.2.6) in the control to mv (s 7..3; P.; n 7) during diltiazem application (Fig. 7E). NIFEDIPINE. The,4-dihydropyridine nifedipine was tested in concentrations from 6 to 4 M, in which range it was relatively easy to dissolve. Example experiments demonstrating the effect of nifedipine are shown in Fig. 8A and the mean dose response data are given in Fig. 8A2. An obvious effect of nifedipine on I Ca was detectable at 5 M (Fig. 8A). The maximal usable concentration of nifedipine ( 4 M) blocked about 33% of I Ca (Fig. 8A). The remaining experiments with nifedipine were carried out with a concentration of 4 M. At this concentration nifedipine reduced I Ca without significantly changing the waveform and the voltage dependence for activation and inactivation (Fig. 8, B E). NONSPECIFIC EFFECTS OF CA 2 ORGANIC BLOCKERS., diltiazem, and nifedipine had to be used in the millimolar range to yield a significant block. To test whether these substances cause effects on other channels we tested all blockers at their IC 5 on the voltage-activated whole cell sodium and potassium currents. The main components of the potassium current are I A and I K(V) in these neurons. To record the sodium and potassium currents K was used instead of Cs in the pipette solution. The normal extracellular solution was used without TTX, 4-AP, and TEA, whereas the change in osmolarity was compensated with NaCl. We did not perform a detailed biophysical analysis, but all three organic blockers clearly had substantial effects on the voltage-activated sodium and potassium currents (Fig. 9). All blockers drastically reduced the amplitude and/or waveform of the sodium and potassium currents. seemed to preferentially block the sustained component [I K(V) ] of the potassium current (Fig. 9C). However, the effect of verapamil on the isolated I K(V) (Fig. 9D) indicated that verapamil increased the inactivation of I K(V) and/or might act as an open channel blocker for I K(V). DISCUSSION This study is an initial step to obtain detailed information about physiological and biophysical properties of I Ca in inter- FIG. 8. Effect of nifedipine on I Ca. In all experiments except the doseinhibition curve nifedipine was applied at a concentration of 4 M. This was the maximal concentration in which nifedipine could be easily dissolved. The cells were held at 6 mv. For details see the legend of Fig. 6. A: the reduction of I Ca by nifedipine is dose dependent. Nifedipine was applied in the range of 6 to 4 M. B: current traces of I Ca before and during application of nifedipine. C: I/V relations of I Ca from 6 neurons before (squares) and during application of nifedipine (triangles). C and C2: nifedipine did not change the voltage dependence of peak-current activation (for quantitative analysis see Fig. 6E). D and E: I/V relation for tail-current activation and steady-state inactivation of I Ca under control conditions (solid squares) and in nifedipine (solid triangles). The Boltzmann fits have the following parameters. D: tailcurrent activation was not significantly changed by nifedipine (n 5): : V.5act 9 3 mv; s act Nifedipine: V.5act mv; s act E: steady-state inactivation was not significantly changed by nifedipine (n 6): : V.5inact mv; s inact Nifedipine: V.5inact mv; s inact 8.5.

11 33 A. HUSCH, S. HESS, AND P. KLOPPENBURG neurons from the adult insect olfactory system and to test pharmacological tools to manipulate I Ca. The insect AL serves as an important model for olfactory information processing on the network level (Hildebrand and Shepherd 997; Laurent 999; Strausfeld and Hildebrand 999; Wilson and Mainen A B C D I A I Na I A I Na I A I Na 5 na I K(V) I K(V) 2 ms I K(V) I K(V) M Nifedipine -4 M Nifedipine.5-4 M.5-4 M - mv 5 ms +4 mv FIG. 9. Effects of diltiazem, nifedipine, and verapamil on voltage-activated sodium and potassium currents. and verapamil were applied at their IC 5 (2.5 4 and.5 4 M, respectively) and nifedipine at 4 M, the maximal concentration that could be easily dissolved. The cells were held at 6 mv. After prepulses to mv (5 ms), the voltage was stepped from 8 to 4 mv in -mv increments. The main components of the potassium current are I A and I K(V) in these neurons (A C). We did not perform a detailed biophysical analysis, but diltiazem (A), nifedipine (B), and verapamil (C) clearly affected the voltage-activated sodium and potassium currents (n 3 for each blocker). All blockers drastically reduced the amplitude and/or waveform of the sodium and potassium currents. C and D: verapamil seemed to preferentially block the sustained component [I K(V) ] of the potassium current (C), but the effect of verapamil on the isolated I K(V) (D) indicated that verapamil increased the inactivation of I K(V) and/or might act as an open channel blocker for I K(V) (n 3). 26). Together with future research, this work aims to better understand the role of Ca 2 currents in mediating olfactory information processing on both the cellular and the subcellular levels. I Ca recorded in vitro from isolated somata of adult AL olfactory interneurons activated at membrane potentials above approximately 5 mv with a maxiumum current around 5 mv. I Ca consisted of both relatively fast activating/inactivating and noninactivating components. In situ the I/V relation for steady-state activation/inactivation was shifted to more depolarized membrane potentials. Interestingly, in a similar study of honeybee antennal motoneurons, differences between in vitro and in situ recordings were not observed (Kloppenburg et al. 999b). After taking into account the different experimental conditions between in vitro and in situ experiments, however, such a shift of parameters was not unexpected. One reason might be imperfect voltage control across the whole neuron in the intact brain preparations, in which the neuronal arborizations are still intact, and/or differences in voltage dependence of calcium channels that are localized in distal regions of the neurons. Previous studies in other insect and invertebrate preparations have demonstrated neuritic or axonal localizations of calcium channels (Duch and Levine 22; Haag and Borst 2; Kloppenburg et al. 2, 27). However, given the complex morphology of these neurons it can be assumed that a major part of the measured in situ currents originates from the cell body. Nevertheless, both the in vitro and in situ parameters of I Ca are well in the range of Ca 2 currents described in other insect preparations including Drosophila neurons (Byerly and Leung 988; Saito and Wu 99), honeybee Kenyon cells (Schäfer et al. 994), Manduca motor neurons (Hayashi and Levine 992), embryonic cockroach neurons (Benquet et al. 999), cockroach DUM neurons (Heidel and Pflüger 26), locust thoracic neurons (Laurent et al. 993; Pearson et al. 993), cricket giant interneurons (Kloppenburg and Hörner 998), and honeybee antennal motoneurons (Kloppenburg et al. 999b). The average current density of about 5 pa/pf in vitro was in the same range as that found in vitro for honeybee projection neurons (Grünewald 23) and cockroach DUM neurons (Heidel and Pflüger 26). In this study we presented averaged data from a large number of experiments to characterize the parameter space of I Ca in AL interneurons. It is the first stage to characterize I Ca of adult AL interneurons in detail and to get pharmacological tools to manipulate I Ca. The variability of the data, however, should not be ignored. It might be due to differential expression of Ca 2 channel types in different cell types. This hypothesis can be tested by recordings from neurons that are unequivocally identified by single-cell labeling. In vertebrates, Ca 2 currents are usually classified into low-voltage activated (LVA or T-type, with activation starting above approximately 7 mv) and high-voltage activated (HVA, activation starting above approximately 3 mv) classes. Subtypes such as L-, P/Q-, N-, and R-type are defined by biophysical and pharmacological properties (Ertel et al. 2; Hille 2; Triggle 26). In accordance with previous studies, we found that the pharmacological classification of vertebrate calcium currents is difficult to transfer to I Ca of insects (for review see King 27). The I Ca in AL interneurons exhibits some characteristics that are typical for some HVA channel types: Ba 2 is a better charge carrier than Ca 2,