Drosophila Mushroom Body Kenyon Cells Generate Spontaneous Calcium Transients Mediated by PLTX-Sensitive Calcium Channels

Jiang, Shaojuan Amy, Jorge M. Campusano, Hailing Su, and Diane K. O’Dowd. Drosophila mushroom body Kenyon cells generate spontaneous calcium transients mediated by PLTX-sensitive calcium channels. J Neurophysiol 94: 491–500, 2005. First published March 16, 2005; doi:10.1152/jn.00096.2005. Spontaneous calcium oscillations in mushroom bodies of late stage pupal and adult Drosophila brains have been implicated in memory consolidation during olfactory associative learning. This study explores the cellular mechanisms regulating calcium dynamics in Kenyon cells, principal neurons in mushroom bodies. Fura-2 imaging shows that Kenyon cells cultured from late stage Drosophila pupae generate spontaneous calcium transients in a cell autonomous fashion, at a frequency similar to calcium oscillations in vivo (10–20/h). The expression of calcium transients is up regulated during pupal development. Although the ability to generate transients is a property intrinsic to Kenyon cells, transients can be modulated by bath application of nicotine and GABA. Calcium transients are blocked, and baseline calcium levels reduced, by removal of external calcium, addition of cobalt, or addition of Plectreurys toxin (PLTX), an insect-specific calcium channel antagonist. Transients do not require calcium release from intracellular stores. Whole cell recordings reveal that the majority of voltage-gated calcium channels in Kenyon cells are PLTX-sensitive. Together these data show that influx of calcium through PLTXsensitive voltage-gated calcium channels mediates spontaneous calcium transients and regulates basal calcium levels in cultured Kenyon cells. The data also suggest that these calcium transients represent cellular events underlying calcium oscillations in the intact mushroom bodies. However, spontaneous calcium transients are not unique to Kenyon cells as they are present in approximately 60% of all cultured central brain neurons. This suggests the calcium transients play a more general role in maturation or function of adult brain neurons.


I N T R O D U C T I O N
Transient increases in intracellular calcium regulate a wide range of cellular processes in developing and mature neurons, from rapid enhancement of neurotransmitter release to longlasting changes in gene expression (Carey and Matsumoto 1999;Emptage et al. 2001;Spitzer et al. 2004;Syed et al. 2004;Yuste et al. 1992). In Drosophila, GAL4 driving expression of the calcium-sensitive luminescent protein, apoaequorin, reveals slow rhythmic oscillations in intracellular calcium levels localized in mushroom bodies, a region of the insect brain required for olfactory associative learning (Rosay et al. 2001). An increase in oscillation amplitude in the mushroom bodies of amnesiac memory mutants suggests these slow wave calcium oscillations may contribute to memory consolidation in the fly.
Calcium oscillations are not only present in intact flies, but also occur in isolated brains, indicating these spontaneous events are independent of sensory input (Rosay et al. 2001). Blockade of a variety of voltage-gated ion channels and neurotransmitter receptors modulate the oscillation amplitude and/or frequency. However, because the oscillations reflect synchronized changes in calcium in large populations of mushroom body cells, it has not been possible to determine the contribution of intrinsic membrane properties versus intercellular communication in generation of this activity. Assessment of the biophysical mechanisms involved in generating the oscillations requires the ability to resolve calcium dynamics in single cells.
Cameleons and camgaroos, two other calcium sensors that have also been transgenically expressed in the fly, have provided increased spatial resolution in analysis of calcium dynamics. These have been used to discriminate between odorevoked responses in pre-and postsynaptic regions, or between acetylcholine (ACh)-evoked responses in different lobes, of the mushroom bodies (Fiala et al. 2002;Yu et al. 2003). However, neither of these calcium sensors has been used to follow intracellular calcium levels at the resolution of individual Kenyon cells in the intact fly brain. In contrast, it has been possible to examine calcium levels in single Drosophila neurons using membrane permeant fluorescent dyes. Calcium green-1 reveals calcium responses to odorant stimuli in one or a small number of Kenyon cells in the semi-intact fly . Fura-2 imaging shows changes in intracellular calcium in response to depolarization and other stimuli, including application of acetylcholine, in "giant neurons" in embryonic Drosophila cultures treated with cytochalasin B (Alshuaib et al. 2004;Berke and Wu 2002). However, neither the study of Kenyon cells in the brain nor giant embryonic neurons in culture have reported detection of spontaneous fluctuations in intracellular calcium.
To explore the cellular mechanisms involved in regulating calcium levels in individual mushroom body Kenyon cells, we employed fura-2 imaging and electrophysiological analysis of neurons harvested from the central brain region of late pupal stage Drosophila. When grown in dissociated cell culture, these neurons regenerate processes and form functional synaptic connections, and Kenyon cells can be identified by cellspecific expression of green fluorescent protein (GFP) using the GAL4 system . Here we report that Kenyon cells generate spontaneous, transient increases in intracellular calcium, mediated by Plectreurys toxin (PLTX)sensitive voltage-gated calcium channels. The ability to gen-erate transients is up regulated during pupal development, transient frequency is similar to the frequency of calcium oscillations in vivo, but spontaneous calcium transients are not specific to Kenyon cells. This suggests that, while the transients are likely to represent cellular events that contribute to calcium oscillations in the mushroom bodies, they may also play a more general role in adult brain neurons.

Fly strains
Pupae were obtained from the mating of males from the homozygous enhancer trap line, OK107-GAL4 (Connolly et al. 1996), and females homozygous for a UAS-GFP transgene (both lines kindly provided by Y. Zhong). The OK107-GAL4;UAS-GFP pupae exhibit high levels of GFP expression in the cell bodies and processes of the mushroom body Kenyon cells in late stage pupae . Kenyon cells in cultures were identified as GFPϩ neurons with small soma diameters (3-6 m).

Primary pupal cultures
Cultures were prepared as described previously with minor modifications noted below. Heads were removed from animals at ϳ55-78 h after pupation (Bainbridge and Brownes 1981). Immediately after decapitation, heads were placed in sterile dissecting saline (DS) containing (in mM) 137 NaCl, 5.4 KCl, 0.17 NaH 2 PO 4, 0.22 KH 2 PO 4, 33.3 glucose, 43.8 sucrose, and 9.9 HEPES, pH 7.4. The brains were removed, and the optic lobes were discarded. Central brain regions from 5 to 15 animals were incubated in dissecting saline containing 50 U/ml papain activated by L-cysteine (1.32 mM) for 15 min at room temperature. Tissue was washed (3 times with 1.5 ml) with DS and (2 times with 1.5 ml) with Drosophila-defined culture medium, DDM2 composed of Ham's F-12 DMEM (high glucose; Irvine Scientific, Santa Ana, CA) supplemented with 1 mg/ml sodium bicarbonate, 20 mM HEPES, 100 M putrescine, 30 nM sodium selenite, 20 ng/ml progesterone, 50 g/ml insulin, 100 g/ml transferrin, and 1 g/ml of 20-hydroxyecdysone . Each culture was prepared from a single brain transferred to a 5-l drop of DDM2 culture medium on a ConA-laminin coated glass coverslip mounted to the bottom of a petri dish. The tissue was mechanically dissociated and the cells were allowed to settle to the substrate for 15 min. Dishes were flooded with 1.5 ml of DDM2 and maintained in a 23°C humidified 5% CO 2 incubator. After 24 h, 0.5 ml of a 3:1 mixture of DDM2 and conditioned Neurobasal medium (cNBM) was added to each dish. cNBM is neurobasal medium ϩ B27 supplements (Life Technologies) conditioned for 24 h by non-neuronal mouse brain feeder cell cultures (Hilgenberg et al. 1999). Cultures were fed every 4 -5 days by removing 1 ml of media from the dish and adding 1 ml of the 3:1 mixture of DDM2 and cNBM.
Culture dishes were prepared by punching 9-mm holes in the bottom of 35-mm petri dishes. Round 12-mm-diam glass coverslips were attached to the bottom of the dish with Sylgard. After fabrication, dishes were immersed in 70% EtOH for 5 min and allowed to air dry. Before coating, dishes were exposed to ultraviolet light for a minimum of 15 min. Coverslips were coated with ConA-laminin and used for culturing within 1 mo of coating (Kraft et al. 1998).

Calcium imaging
Cultured neurons were rinsed in a HEPES-buffered salt solution (HBSS) containing (in mM) 120 NaCl, 5.4 KCl, 1.8 CaCl 2 , 0.8 MgCl 2 , 15 glucose, 20 HEPES, 10 NaOH, and 0.001% phenol red, pH 7.2-7.4. They were loaded with calcium indicator dye by incubating cultures in HBSS containing 5 M fura-2 AM and 0.1% Pluronic F-127 (Molecular Probes) for 40 min, in the dark, at room tempera-ture. To allow for complete hydrolysis of the AM ester before imaging, cultures were washed three times with HBSS and incubated in fresh HBSS for another 45 min, in the dark, at room temperature. Fresh HBSS was added to the culture dish, and it was transferred to the stage of an inverted Nikon microscope for imaging with a ϫ40, 1.3 numerical aperture oil immersion objective. Fluorescent illumination was provided by a 150-W xenon arc lamp, and band-specific filters were used for excitation, alternating between 340 and 380 nm. All images were acquired at the emission wavelength 510 nm and recorded with a digital CCD camera (Orca-100, Hamamatsu). The 340/380 ratio images and their pseudocolor representations were generated digitally using Metafluor 4.0 imaging software (Universal Imaging). Data were collected at 5-s intervals.
Estimates of intracellular calcium concentrations were based on the following formula where K d is 225 nM (dissociation constant for fura-2) and R is the ratio of emission intensity at 510 nm after excitation at 340 and 380 nm (Grynkiewicz et al. 1985). R min is the ratio in a calcium-free solution, and R max is the ratio at saturating free Ca (39 M). F o and F s are the emission intensities at 510 nm after excitation at 380 nm in calcium-free and calcium-saturated solutions, respectively. Background levels of fluorescence in calcium-free, fura-free solution were determined at each wavelength and subtracted from signals detected. Values for R min (0.43), R max (7.89), and F o /F s (8.85) were determined using the Fura-2 Calcium Imaging Calibration Kit (F-6774, Molecular Probes). Because the calibration is based on a cell-free system, the values for intracellular calcium levels reported for the cultured neurons are viewed as estimates that are primarily useful in comparison of changes in calcium levels under different experimental conditions.

Data analysis
Intracellular calcium levels were evaluated by determining the average signal over the soma region of all GFPϩ Kenyon cells (7-40) in a single field of view in each OK107-GAL4;UAS-GFP culture. Neurons were included in the data set if a stable baseline, Յ150 nM calcium, was recorded for a minimum of 10 min, and an increase in the 340/380 ratio, Ͼ600 nM, was induced by exposure to 5 M ionomycin at the end of the experiment. A calcium transient was defined as an increase in 340/380 ratio, Ͼ75 nM (Ͼ5 times baseline noise level), that reaches its peak within 20 s and declines by Ն50% from the peak within 3 min. In a single field of view, the percentage of active Kenyon cells reflects the number of Kenyon cells exhibiting one or more transients in the first 10-min recording period, divided by the total number of Kenyon cells. The transient frequency in each cell was calculated as the number of transients divided by the time of recording. The amplitude of each transient was measured from baseline to peak, and an average value for each cell was calculated. An average transient frequency and amplitude were determined for each culture. The mean percentage of active cells, transient frequency, and transient amplitude were calculated from three or more cultures (n) at each age/condition in which seven or more individual neurons were analyzed in each culture. Only one field of view was examined/culture and each culture was prepared from the brain of a single animal.

Electrophysiology
To examine calcium currents in isolation, the pipette solution contained (in mM) 120 D-gluconic acid, 120 cesium hydroxide, 20 NaCl, 0.1 CaCl 2 , 2 MgCl 2 , 1.1 EGTA, and 10 HEPES, pH 7.2. The external recording saline was composed of HBSS containing TTX, curare, PTX (TCP). To examine voltage-gated potassium currents in isolation, the pipette contained (in mM) 120 K-gluconate, 20 NaCl, 0.1 CaCl 2 , 2 MgCl 2 , 1.1 EGTA, 4 ATP, and 10 HEPES, pH 7.2. In addition, 2 mM CoCl 2 was added to the external recording saline (TCP) to block the calcium currents and calcium-activated potassium currents. All data shown are corrected for the 5-mV liquid junction potential generated in these solutions. Data were acquired with a List EPC7 amplifier, a Digidata 1320A D-A converter (Axon Instruments), a Dell (Dimension 4100) computer, and pClamp 8 (Axon Instruments) software. Recordings were made at room temperature.

Cultured Kenyon cells generate spontaneous calcium transients
To explore the cellular mechanisms involved in regulation of calcium in neurons from the mushroom body region, intracellular calcium levels were monitored in Kenyon cells grown in primary culture. Each culture was prepared from a single central brain region harvested from a late stage OK107-GAL4; UAS-GFP pupa. The mushroom body Kenyon cells in these cultures were identified as small-diameter (3-6 m), GFPϩ neurons .
Fura-2 AM, a ratiometric calcium-imaging dye, was used to monitor calcium levels in the neurons. Fura-2 loaded neurons appeared healthy based on overall morphology and GFPϩ Kenyon cells, representing 10 -15% of the total population, were clearly identifiable when viewed with fluorescent illumination (Fig. 1A). A pseudo-color representation of the 340/380 ratio image of a single field, at two time-points during a 15-min recording in standard HBSS saline, revealed differences in the intracellular calcium levels in the cell bodies of the two indicated GFPϩ Kenyon cells (Fig. 1, B and C). Changes in calcium levels were also observed in the processes of some Kenyon cells. However, due to the small diameter of the neuritic processes in these neurons, analysis of calcium fluctuations was limited to the soma region.
The 340/380 ratios in each cell were converted to estimates of intracellular calcium concentrations based on a standard calcium curve generated in a cell-free system (see METHODS for details). The highly dynamic nature of calcium levels in one Kenyon cell (indicated by red arrow) is shown in a plot of the intracellular calcium concentration as a function of time (Fig.  1D). Simultaneous analysis of a second Kenyon cell (indicated by purple arrowhead) that appeared to have no physical connection with the first, showed that these cells exhibit independent calcium fluctuations that occur with different frequencies and patterns (Fig. 1D). No evidence of synchronized activity between Kenyon cells was observed, even when the cell bodies were in direct physical contact. Of 14 pairs of Kenyon cells that were physically adjacent to each other, transients were observed in one Kenyon cell but not the other in five cases. In the other nine pairs, both Kenyon cells exhibited calcium transients, but there was no synchrony in the timing of the transients (Fig. 1E). Finally, examination of Kenyon cells in low-density cultures within 1-2 days of plating showed that physically isolated Kenyon cells can also generate calcium transients (Fig. 1F).
FIG. 1. Spontaneous calcium transients in cultured mushroom body Kenyon cells. A: pupal neurons loaded with calcium indicator dye fura-2 exhibit normal morphology. Kenyon cells are identified as small diameter, green fluorescent protein (GFP)ϩ neurons in a 2-day-old culture prepared from an OK107-GAL4;UAS-GFP pupa 72 h after puparium formation (APF). Fluorescent mask projected on the Nomarski image. B and C: color-coded, ratiometric images (340/380) of same field of view, examined in standard HBSS recording saline, shows a difference in calcium levels in the soma region of 2 Kenyon cells (arrows, arrowheads) viewed at different times. D: intracellular calcium concentration as a function of time, determined over the soma region of the indicated Kenyon cells. E: physically connected pair of Kenyon cells exhibit nonsynchronous spontaneous calcium transients. F: calcium transients in a physically isolated Kenyon cell that has not yet extended neurites. Values plotted in D-F represent the background corrected average 340/380 ratios, converted to estimates of calcium levels based on a standard calcium curve generated in a cell-free system. Calcium levels evaluated in a circular field placed over the soma region of the neurons, sampled at 5-s intervals.
The actual calcium waveforms vary considerably between cells and even within a cell ( Fig. 2A). For the purposes of this study, a single transient was defined as a rapid increase (Ͻ20 s from baseline to peak) in intracellular calcium, 75 nM or more above the basal calcium level that declines by Ն50% within 3 min. A Kenyon cell with one or more transients during the first 600 s of recording was defined as active. The number of active Kenyon cells in each field was expressed as a percentage of all Kenyon cells in each field. In 1-to 4-day-old cultures, made from 55-to 78-h pupae, the majority (61 Ϯ 5%, n ϭ 14 cultures) of Kenyon cells were spontaneously active. While active Kenyon cells were observed in cultures as long as 12 days after plating, the percent that were active decreased with increasing age in culture.
Spontaneous calcium oscillations in the mushroom body region in whole brains first appeared 55-69 h after pupation and persisted into adulthood (Rosay et al. 2001). To determine if the ability of Kenyon cells to generate transients is also developmentally regulated, cultures were prepared from brains obtained from two earlier pupal stages, ϳ5 and 10 -30 h after puparium formation (APF). Cultures were observed at 1-4 days in vitro. Calcium transients were present in some Kenyon cells in the cultures prepared from pupae at 5 and 10 -30 h APF, showing that Kenyon cells are capable of generating calcium transients before oscillations are observed in the intact mushroom bodies. However, the number of active Kenyon cells, expressed as a percentage of the total number of Kenyon cells, was significantly lower in cultures made from pupae at 5 h compared with 55-78 h APF. The percentage of Kenyon cells active, in cultures made from pupae 10 -30 h APF, was intermediate (Fig. 2B). This pattern of developmental regula-tion is consistent with the calcium transients in the cultured Kenyon cells representing cellular events that contribute to generation of the calcium oscillations observed in the intact mushroom bodies.
The majority of the subsequent experiments focused on analysis of Kenyon cells in cultures prepared from late stage pupae (55-78 h APF) at 1-4 days in vitro.

Spontaneous calcium transients are an intrinsic property of Kenyon cells
Addition of 1 M TTX to the HBSS recording saline, a concentration previously shown to block sodium currents in Drosophila neurons (O'Dowd 1995), did not block the calcium transients, showing that they can occur independently of sodium channel activity (Fig. 3A). Spontaneous synaptic activity, mediated by nicotinic acetylcholine receptors (nAChRs) and GABA receptors, has also been observed in the cultured Kenyon cells . This activity is not necessary for generation of the transients either, because calcium transients persist in saline containing curare and PTX at concentrations that completely block the nAChR and GABA receptor-mediated synaptic currents (Fig. 3B). Additionally, application of APV and CNQX reveals that endogenous activity of ionotropic glutamate receptors is not required for generation of the calcium transients (Fig. 3C). To quantitatively assess the effect of specific blockers on transient activity, the percentage of active cells, transient frequency, and transient amplitude were compared in Kenyon cells examined in different external solutions. No significant differences were apparent between Kenyon cells examined in control saline (HBSS), TTX, TCP, or TCPAC (TTX, curare, PTX, APV, CNQX; Fig.  3D). Thus the majority of the calcium transients in the cultured Kenyon cells are independent of sodium-dependent action potentials and neurotransmitter receptor activity.
These data suggest that spontaneous calcium transients are generated in a cell autonomous fashion. However, it was still possible that there were receptors or channels mediating intercellular communication in cultured neurons that have yet to be identified. Therefore to more rigorously test this hypothesis, cultures were imaged at 1-2 days in culture, facilitating analysis of physically isolated Kenyon cells, in some cases before neurite outgrowth. These Kenyon cells were also able to generate spontaneous calcium transients (Fig. 1F). Quantitative analysis indicated that the properties of the transients in physically isolated Kenyon cells and transients in Kenyon cells that were in contact with the processes or cell bodies of other neurons were not significantly different (Fig. 3D). Together these data show that spontaneous calcium transients can be generated in a cell autonomous manner.

Calcium transients are modulated by nicotine and GABA
Blockade of nAChR and GABA receptor-mediated synaptic transmission did not inhibit generation of calcium transients. However, to determine whether activation of these receptors by exogenous agonists affects intrinsically generated calcium transients, cultures were exposed to nicotine or GABA. Bath application of 0.5 M nicotine caused a rapid and reversible increase in baseline calcium to levels higher than the peak of the average spontaneous calcium transient (Fig. 4A). Few if any transients were seen to rise above the elevated calcium levels in the presence of nicotine (Fig. 4A). The nicotine-induced elevation in intracellular calcium was blocked by preincubation of the culture in 0.1 M ␣-BTX, a nAChR antagonist (Fig. 4B). In contrast, application of exogenous GABA (100 M) resulted in a reversible decrease in basal calcium levels and a blockade of the calcium transients (Fig. 4, C-E). The GABA-induced changes were inhibited by preincubation of the cultures in 20 M PTX, a GABA receptor antagonist (Fig. 4, D and E). These data show that activation of excitatory receptors increases intracellular calcium levels, whereas activation of inhibitory receptors decreases basal calcium levels and blocks the transients. This suggests that basal calcium levels and calcium transients are regulated by membrane potential.

Transients are mediated by PLTX-sensitive calcium channels
To explore the role of calcium channels in generation of the transients, we examined intracellular calcium levels when extracellular calcium (normally 1.8 mM) was removed from the recording saline. This reversibly blocked the spontaneous calcium transients in Kenyon cells (Fig. 5, A and D). Similarly, addition of cobalt (2 mM), a general voltage-gated calcium channel blocker previously shown to block calcium currents in embryonic Drosophila neurons (O'Dowd 1995), reversibly inhibited spontaneous calcium transients in Kenyon cells (Fig.  5, B and D). These data show that calcium flux into the cell, through calcium channels, is necessary for generation of the calcium transients. Similar to the effect of GABA, there is also a reversible reduction in the baseline calcium levels after removal of calcium or addition of cobalt (Fig. 5, A, B, and E). This shows that calcium channels are active and regulate basal calcium levels in the absence of external stimulation, even when sodium channels and neurotransmitter receptors are blocked.
The pharmacological profile of the calcium channels underlying the calcium transients was further examined through bath application of several specific calcium channel antagonists. Calcium transient frequency (19 Ϯ 1/h) in control saline (TCP) was not significantly reduced by the vertebrate L-type calcium channel antagonist, verapamil (20 Ϯ 2 transients/h, 5 M), or the N-type blocker, -conotoxin GVIA (24 Ϯ 1 transient/h, 2 M). In sharp contrast, the insect calcium channel antagonist PLTX blocked the majority of the calcium transients in Kenyon cells within 3 min of application (Fig. 5, C and D). There was no recovery during a 10-min wash, consistent with the irreversible blockade of calcium currents by PLTX previously reported in embryonic Drosophila neurons (Leung et al. 1989). In addition, the basal calcium level was also significantly reduced by PLTX treatment (Fig. 5, C and E), though the magnitude of the reduction was smaller than that induced by 0 Ca 2ϩ and 2 mM Co 2ϩ . Thus calcium influx through calcium channels sensitive to PLTX but not to vertebrate Lor N-type channel blockers is involved in generation of spontaneous calcium transients in Kenyon cells.

Transients do not require calcium release from intracellular stores
In some vertebrate neurons, spontaneous increases in intracellular calcium are triggered by the influx of calcium across the plasma membrane, but the bulk of the calcium increase is due to release from intracellular stores (Gu et al. 1994;Owens and Kriegstein 1998). To determine if release of calcium from intracellular stores could be detected in Kenyon cells, cultures were exposed to caffeine during blockade of the voltage-gated calcium channels by cobalt. An increase in intracellular calcium typically induced by application of 10 mM caffeine is shown in Fig. 6A. Caffeine triggered an increase in intracellular calcium levels in 73% (16/22) of the active Kenyon cells with a mean amplitude of 297 Ϯ 33 nM.
To determine if calcium release from these intracellular stores is involved in generation of the transients in Kenyon cells, the effect of blocking release was monitored. Cultures were pretreated with caffeine to deplete intracellular stores, followed by thapsigargin to block reuptake. Kenyon cells in caffeine/thapsigargin-treated cultures were not only capable of generating robust spontaneous calcium transients (Fig.  6B), but the transient frequency was significantly higher in treated versus control cultures (26.5 Ϯ 2.4 and 18.2 Ϯ 1.4/h, respectively; P Ͻ 0.01, Student's t-test). In addition, basal FIG. 5. Spontaneous transients mediated by calcium influx across the membrane through Plectreurys toxin (PLTX)-sensitive calcium channels. A: removal of calcium from TCP (TTX, curare, PTX) recording saline (0 Ca 2ϩ ) reversibly blocks the calcium transients and reduces basal calcium levels in a Kenyon cell. B: addition of 2 mM cobalt (Co 2ϩ ) also reversibly blocks the calcium transients and reduces the basal calcium levels in a Kenyon cell. C: bath application of 50 nM PLTX irreversibly inhibits calcium transients and reduces basal calcium levels. D and E: mean transient frequency and basal calcium levels are significantly reduced by removal of calcium (0 Ca 2ϩ , n ϭ calcium levels were also significantly higher in the caffeine/ thapsigargin-treated versus control cultures (59 Ϯ 5 and 40 Ϯ 2.3 nM, respectively; P Ͻ 0.01, Student's t-test). These data show that calcium transients do not require calcium release from intracellular stores but are instead up-regulated by depletion of this calcium reservoir.

Voltage-gated calcium currents in Kenyon cells are PLTX sensitive
Based on previous studies in embryonic Drosophila neurons (Leung et al. 1989;O'Dowd 1995), the blockade of calcium transients by cobalt and PLTX suggests that the calcium channels involved in regulating the spontaneous transients in Kenyon cells are voltage-dependent. To test this hypothesis directly, whole cell voltage-clamp recordings were employed to examine isolated calcium currents. Potassium currents were blocked using cesium in place of potassium in the internal solution, and TCP was used as the control external recording solution.
A series of step depolarizations, from holding potentials of -75 or Ϫ80 mV, elicited a family of inward currents carried through voltage-gated calcium channels in 76% (19/25) of the Kenyon cells examined in control solution (Fig. 7A). Bath perfusion of 50 nM PLTX dramatically reduced the inward calcium currents in five of five Kenyon cells (Fig. 7A). To estimate the PLTX-induced blockade of the current and minimize the potential contribution of rundown, calcium currents recorded from Kenyon cells in control conditions (TCP) were compared with those preincubated for Ն10 min in 50 nM PLTX. All Kenyon cells in which a stable whole cell recording was obtained were included in the analysis. A value of zero was assigned to cells without detectable calcium currents. Peak calcium current amplitudes were normalized to the whole cell capacitance and expressed as current densities to control for variation in size of individual Kenyon cells (range, 1-8 pF).
Comparison of the current density versus voltage curves confirmed that 50 nM PLTX results in a dramatic and significant reduction in peak calcium current (Fig. 7B). The nonzero current density voltage curve in 50 nM PLTX reflects the presence of a residual voltage-gated calcium current in 5/12 cells. The blockade of calcium channels was specific in that voltage-gated potassium currents were not significantly altered by perfusion with (Fig. 7C), or preincubation ( Fig. 7D) in, 50 nM PLTX. These data confirm that Kenyon cells express voltage-gated calcium channels, the majority of which are FIG. 6. Release of calcium from intracellular stores is not required for generation of calcium transients. A: caffeine (10 mM) induces an increase in intracellular calcium in a single Kenyon cell in the presence of cobalt to block voltage-gated calcium transients. B: spontaneous calcium transients recorded in a Kenyon cell in which intracellular calcium stores have been depleted. Addition of 10 mM caffeine at 900 s into the recording saline, in the presence of cobalt to block the transients, did not induce an increase in intracellular calcium, confirming the depletion of intracellular calcium stores. Culture was pre-exposed to caffeine for 1 min and incubated in 5 nM thapsigargin for 45 min to block reuptake of calcium. blocked by 50 nM PLTX. In combination with the imaging data, these findings indicate that calcium influx through PLTXsensitive voltage-gated calcium channels mediates the spontaneous calcium transients and regulates resting levels of calcium in Kenyon cells.

Calcium transients are not limited to Kenyon cells
The calcium oscillations reported in the intact Drosophila brain, visualized with apoaequorin imaging, are restricted to the mushroom body region. In contrast, GFP negative (GFPϪ) neurons in the OK107-GAL4;UAS-GFP cultures, representing a variety of neuronal subtypes found within the central brain region but outside the mushroom bodies, also exhibit spontaneous PLTX-sensitive calcium transients (Fig. 8A). There was no significant difference in the percentage of active cells or the transient frequency in GFPϩ Kenyon cells and GFPϪ neurons examined in the same cultures (Fig. 8B). Although it should be noted that the amplitude of the transients in the GPFϪ neurons was significantly larger than in the GFPϩ Kenyon cells, the mechanism underlying this difference is unknown. These data show that cell autonomous calcium transients, unlike the synchronous calcium oscillations recorded in the intact brain, are not unique to Kenyon cells.

D I S C U S S I O N
Spontaneous transient changes in intracellular calcium levels have been shown to play an important role in maturation of cells in the nervous system of many animals. Our results show that the principal neurons in Drosophila mushroom bodies, Kenyon cells, are capable of generating spontaneous calcium transients in a cell autonomous fashion. The transient increases in intracellular calcium are mediated by calcium flux through a PLTX-sensitive voltage-gated calcium channels and do not require release of calcium from intracellular stores. The frequency of the Kenyon cell calcium transients in vitro is similar to the mushroom body calcium oscillations in vivo. In addition, the ability of Kenyon cells to generate spontaneous calcium transients is up-regulated during pupal development. These data suggest that cell autonomous calcium transients mediated by PLTX-sensitive calcium channels in the cultured Kenyon cells represent cellular events that contribute to generation of the calcium oscillations in the intact mushroom bodies, activity that has been implicated in memory consolidation (Rosay et al. 2001). However, the presence of calcium transients in the majority of non-Kenyon cells, as well Kenyon cells, suggests they may play a more general role in maturation or function of neurons in the CNS.
Why are the calcium oscillations imaged with apoaequorin localized specifically in mushroom bodies in the whole brain, when the majority of neurons from the central brain region have the ability to generate spontaneous calcium transients? One possibility is that only mushroom body Kenyon cells are synchronously activated in the intact brain preparations described previously (Rosay et al. 2001). The imaging system used to visualize the apoaequorin signal is capable of detecting the calcium oscillations in groups of coordinately activated cells but is not sensitive enough to measure calcium transients in single cells (Davis 2001). It is clear that the calcium oscillations in the intact mushroom bodies are the result of synchronized activation of the Kenyon cell population since drugs that alter electrical excitability or synaptic transmission can modulate the oscillation frequency and amplitude. If neurons in the central complex and antennal lobes generate calcium transients but are not coordinately activated, this could account for the reported absence of calcium oscillations in these regions (Rosay et al. 2001). A second possibility is that synchronized oscillations occur in other neuronal populations in the brain but were not detected due to technical limitations. One potential problem with the apoaequorin assay is that it relies on formation of a complex between transgenically expressed apoaequorin and a cofactor, coelenterazine, present in the recording saline (Rosay et al. 2001). Limited access of the cofactor to neurons in the central complex and antennal lobes could affect the ability to detect synchronous oscillations in these regions. It should be possible to test this hypothesis by imaging whole brains in which specific neuronal groups express camgaroos, calcium reporters that do not require cofactors (Yu et al. 2003).
Influx of calcium through voltage-gated channels is necessary for expression of synchronous calcium oscillations in vivo and calcium transients in vitro, based on blockade of both by removal of external calcium or addition of cobalt. However, drugs that block nAChRs, GABA receptors, and sodium channels alter calcium oscillations in vivo (Rosay et  sodium channels would explain why cholinergic miniature excitatory postsynaptic current and GABAergic miniature inhibitory postsynaptic current frequencies in cultured Drosophila neurons are strongly regulated by external calcium levels (Lee and O'Dowd 1999;Lee et al. 2003). A recent study also links activity of Dmca1A calcium channels to regulation of synaptic growth at the Drosophila NMJ (Rieckhof et al. 2003). Analysis of cultures made from Dmca1A mutants and/or addition of PLTX to the growth media will be used to explore the role of these channels in regulating spontaneous neurotransmitter release and synaptic growth.