INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Intracellular free Ca²⁺ plays crucial roles in adult and developing nervous systems. Dendritic Ca²⁺ signals are important for information processing (Borst and Egelhaaf 1992; Hirsch et al. 1995; Sobel and Tank 1994) and synaptic plasticity (Bliss and Collingridge 1993; Yuste and Tank 1996). The somatic Ca²⁺ concentration can influence gene transcription (Hardingham et al. 1997). Internal Ca²⁺ affects aspects of neuronal differentiation such as axon extension and growth-cone motility (Gomez and Spitzer 1999; Gomez et al. 1995; Gu and Spitzer 1995; Kater and Mills 1991; Kater et al. 1988; Lnenicka et al. 1998). Elevations of free intracellular Ca²⁺ may occur by a variety of mechanisms, including release from intracellular stores (Berridge 1998; Libscombe et al. 1988; Wang and Augustine 1995), activity of the Na⁺-Ca²⁺ exchanger operating in the reverse mode (Blaustein 1988), and through ligand-gated and voltage-dependent channels (MacDermott et al. 1986; Malinow et al. 1994; Regehr and Tank 1992). Dendritic Ca²⁺ spikes as occurring in cerebellar Purkinje cells (Llinas and Sugimori 1980) cause Ca²⁺ influx into dendrites (Tank et al. 1988). Furthermore, backpropagating Na⁺ action potentials can cause dendritic Ca²⁺ influx through voltage-dependent calcium channels (Christie et al. 1995; Spruston et al. 1995), thereby mediating activity-dependent influences on dendritic plasticity (Magee and Johnston 1997). The responses to Ca²⁺ elevations differ among neurons and depend on the spatial distribution of Ca²⁺ accumulation. This study addresses the mechanisms and the distribution of Ca²⁺ accumulation in different cell compartments of an identified motoneuron by in situ Ca²⁺-imaging experiments at critical stages of postembryonic development.

During the metamorphosis of the moth, Manduca sexta, many motoneurons undergo dramatic structural and functional modifications for the acquisition of new adult behavior (Consoulas et al. 2000). For example, changes in the dendritic morphology and membrane currents allow an identified flight motoneuron, MN5, to change its behavioral role (Duch and Levine 2000). The larval Ca²⁺ currents become negligible during early pupal stages but increase dramatically during mid-pupal life and then remain unchanged until adulthood (Duch and Levine 2000). The loss of most larval dendrites is followed by the formation of dendritic growth cones, rapid dendritic growth, and new branch formation until about 50% of adult development (Duch and Levine 2000). During the second half of adult development, growth cones are no longer present and branching is limited to the perimeter of the dendritic field, which continues to increase in size (F. Libersat and C. Duch, unpublished observations). This "switch" in the mode of dendritic growth coincides temporally with the occurrence of prominent Ca²⁺ spikes, which are allowed by a delay of several days between the development of the adult Ca²⁺ currents and the subsequent increase in K⁺ currents (Duch and Levine 2000). Because activity-dependent Ca²⁺ influx can affect neuronal differentiation (Cohan et al. 1987; Fields et al. 1990; Gomez and Spitzer 1999; Kater and Mills 1991; Kater et al. 1988; Mattson et al. 1988), the Ca²⁺ spikes may influence dendritic growth.

This study tested whether activity-dependent Ca²⁺ influx occurred in the dendrites of MN5, whether Ca²⁺ influx occurred in the growth cones during early pupal life, how the Ca²⁺ spikes during later stages affected the internal Ca²⁺ concentrations, and whether the Ca²⁺ signaling changed during metamorphosis. The results are consistent with a regulatory role of Ca²⁺ in postembryonic dendritic growth.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals

Manduca sexta obtained from the laboratory culture of the Division of Neurobiology at the University of Arizona were reared on artificial diet (Bell and Joachim 1976) under a long-day photoperiod regimen (17:7 h light/dark cycle) at 26°C and ~60% humidity. Both chronological and morphological criteria were used for staging of animals (Bell and Joachim 1976; Consoulas et al. 1996; Nijhout and Williams 1974; Tolbert et al. 1983). In summary, L5 represents the fifth larval instar, W0-W4 signify the 5 days of wandering, P0 indicates the day of the pupal molt, and P1-P18 are the following 18 days of pupal development. In this study, pupal stages P4 and P8 were used for a developmental comparison of cytosolic Ca²⁺ signaling during dendritic growth. P4 is the stage where prominent growth cones are formed, and rapid dendritic growth and new branch formation takes place (Fig. 1A). P8 is a stage where growth cones are no longer observed and further branching is limited to the perimeter of the dendritic field (Fig. 1B) (Duch and Levine 2000).

View larger version (157K): [in this window] [in a new window]   Fig. 1. A: stack of optical sections projected into 1 focal plane taken of the dendritic field of MN5 at pupal stage P4 after intracellular injection of rhodamine dextran, fixation, clearing, and mounting of the preparation. At stage P4, the tips of all dendrites show growth-cone-like structures, some of which are indicated with white arrowheads. One growth cone is shown as a selective enlargement in the bottom right corner. B: stack of optical sections taken from the dendritic region of MN5 at pupal stage P8. Same staining procedure as in A. C: single focal plane of the dendritic region of MN5 in situ at pupal stage P4 taken at 380-nm excitation wavelength with a cooled CCD camera after intracellular injection with fura-2. Growth cones are marked with white arrowheads. One growth cone is shown at higher magnification in the bottom right corner. D: single focal plane of the dendritic region of MN5 in situ at pupal stage P8 taken at 380-nm excitation wavelength with a cooled CCD camera after intracellular injection with fura-2. The black ovals in C and D indicate region 1 that was used for intensity measurements in subsequent figures. Scale is 50 µM.

Dissection for intracellular recordings

The animals were anesthetized by chilling on ice for 15 min. Animals were dissected along the dorsal midline and superfused with saline. The thoracic and the first two abdominal ganglia were removed, transferred into a silicone-elastomer (Sylgard)-coated petri dish, and pinned down at the cut ends of their lateral nerves in saline. The ganglionic sheath was removed mechanically with a fine pair of forceps.

Nerve 1 of the mesothoracic ganglion was left intact toward its specific peripheral branch, which contains only the axons of the five dorsal longitudinal flight muscle motoneurons. MN5 is the only motoneuron in the mesothoracic ganglion, which carries an axon in this particular nerve branch (Duch et al. 2000). Antidromic stimulation from this nerve branch was used to identify MN5 during intracellular recordings.

Injections of fura-2

Fura-2 pentapotassium salt (Molecular Probes, Eugene, OR) was used for calcium-imaging experiments. The tips of thin-walled borosilicate electrodes (resistances, 35-40 M) were filled with 12 mM fura-2 in 100 mM potassium acetate. The electrode shafts were filled with 1 M potassium acetate, and an air bubble was left between the tip and the shaft to prevent dye dilution. Following intracellular penetration and antidromic identification of MN5, dye was injected iontophoretically by applying hyperpolarizing current of 1 nA amplitude for 10 min in preparations of pupal stage P4 and for 20 min in preparations of pupal stage P8. Pupal stage P8 was injected for a longer duration because the volume of MN5 increases considerably between the pupal stages P4 and P8 (Duch and Levine 2000). Following dye injection, ganglia were left in saline for an additional 15 min to allow dye diffusion.

Electrophysiology

Subsequently, ganglia were transferred to the imaging setup and MN5 was re-impaled with a thin-walled borosilicate electrode (resistances, 100-120 M) filled with 2 M potassium acetate. The re-impalement was conducted for two reasons. First, it prevented further uncontrolled dye filling of the cells during the imaging experiments. Second, simultaneous intracellular and optical recordings required very shallow electrode angles and the low-resistance electrodes that were used for dye injections injured the cells in such recording conditions. An Axoclamp 2B amplifier (Axon Instruments, Foster City, CA) was used for all intracellular recordings. For antidromic stimulation of MN5, an extracellular suction electrode was placed at nerve 1. Antidromic spike initiation was confirmed by simultaneous intracellular recordings. The stimulation amplitude was set just above firing threshold. The manipulators for the electrodes were mounted to the microscope stage so that the preparation could be moved to image different fields of view from the same neuron. The intra- and extracellular recordings were synchronized with the imaging sequences by reading them simultaneously with a trigger trace provided by the imaging system into Clampex 8 (Axon Instruments) software. The trigger trace defined the time point of the start of each acquired frame.

Imaging

Fura-2 was excited in single (380 nm) or dual (340 nm/380 nm) wavelength illumination mode, and fluorescence images on the basis of emission light passing a 530-nm filter were captured with a cooled CCD camera (Hamamatsu 4742-95), mounted on a fluorescence microscope (Olympus BX50Wi). The camera and a filter wheel were controlled with "Simple PCI" software (Compix). The same software was used for data acquisition and fluorescence intensity measurements. The intensity measurement data were further analyzed with Excel 4.0 (Microsoft) and Clampfit8 (Axon Instruments). The imaging speed in dual wavelength illumination mode was limited to 1 Hz due to the slow rotation speed of the filter wheel. Therefore this mode was used predominantly to determine the resting ratio of MN5 prior to and during the time course of the experiments and during drug applications. The imaging speed in single wavelength mode was 50 Hz when leaving the shutter open during the acquisition of an entire sequence. The chip resolution was 1,024 × 1,024 pixels. All experiments using single wavelength mode were conducted at 8 × 8 binning, meaning that pictures of 128 × 128 pixels were transferred to the computer. With these settings, the exposure times per frame were 10 ms for the dendritic region and 2 ms for the cell body. All single-wavelength data shown were acquired under these conditions. In all experiments, Ca²⁺ responses of MN5 were measured as responses to either antidromic stimulation via nerve 1 or spikes elicited orthodromically with an electrode placed in the cell body. Regions of interest were placed at defined locations of MN5, as indicated in the figures, and the fluorescence intensities in each region were measured off-line with Simple PCI. Background was routinely subtracted. For all dendritic measurements, background was determined 50 µm anterior of the site where the major dendrite branches off the axon of MN5 because no arbors of MN5 are located in this region. For measurements from the soma or the link segment, background was measured 50 µm anterior of the soma, or 50 µm anterior of the respective site along the link segment. Photobleaching was not corrected. In all experiments using dual wavelength mode, changes in the F₃₄₀/F₃₈₀ ratio were taken to indicate changes in intracellular Ca²⁺ concentration ([Ca²⁺]_i) (Grynkiewicz et al. 1985). However, the absolute Ca²⁺ concentrations were not determined, given uncertainties concerning accurate calibration in situ. In all experiments, the resting ratio was determined after the re-impalement and prior to single-wavelength imaging. Preparations with a resting ratio >1.5 were likely to have been injured during the two intracellular impalements and were discarded. Similarly, the resting ratio was determined routinely between subsequent single-wavelength acquisitions to ensure the continued health of the cells. Experiments were terminated as soon as the resting ratio increased by >0.3 as compared with the value observed at the beginning of the experiment. In all experiments using single-wavelength mode, the index F₍₃₈₀₎/F_mean(380) was defined to account for relative changes in Ca²⁺ concentration. F_mean(380) is the average fluorescence intensity at 380 nm excitation obtained from 50 frames that were acquired during the last second directly before a stimulation. F₍₃₈₀₎ represents the difference from F_mean(380) to fluorescence intensities excited at 380 nm at any given time. Data were expressed as means ± SD. Student's t-test was applied for statistical significance.

In some experiments calcium green 5 (9 mM in 10 mM HEPES/100 mM potassium acetate, Molecular Probes) was injected into MN5 to compare the time constant of decay to measurements obtained with fura-2. This experiment was performed to check whether the high-Ca²⁺ affinity of fura-2 changed the time course of the observed Ca²⁺ signals artificially as was found for fly visual interneurons (Haag and Borst 2000). Although differences in the time course of the Ca²⁺ signals were observed between the two dyes, with the time constant of decay faster with calcium green 5 (n = 6; P = 0.032), differences between developmental stages were still observed.

Solutions

External saline for dissection and recording consisted of (in mM): 140 NaCl, 5 KCl, 4 CaCl₂, 28 D-glucose, and 5 HEPES, pH was adjusted to 7.4 using 1 M NaOH. Zero-Ca²⁺ saline contained 140 NaCl, 5 KCl, 4 MgCl₂, 28 D-glucose, 5 HEPES, and 0.5 EGTA. Cyclopiazonic acid (CPA, Sigma) was used at 20 µM concentration in normal saline.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

At pupal stage P4 (Fig. 1A), MN5 is characterized by dendritic growth cones, rapid dendritic growth, and branch formation. At pupal stage P8 (Fig. 1B), MN5 has no dendritic growth-cones and rapid growth ceases. At both developmental stages, fura-2 injections allowed reliable in situ visualization of all major compartments of MN5 with a cooled CCD camera (Fig. 1, C and D). A prominent landmark that could easily be identified and compared between developmental stages was the region where the major primary dendrite branches of the primary neurite linking the dendritic field to the soma (see black oval in Fig. 1, C for P4, and D for P8). This region will be referred to as region 1 in subsequent figures.

At pupal stage P4, recordings from the soma of MN5 show small, passively invading, Na⁺-based spikes, which are actively generated in the proximal axon (Duch and Levine 2000), as is the usual case for insect motoneurons (Gwilliam and Burrows 1980). Single Na⁺ spikes evoked orthodromically by current injection into the soma did not evoke a detectable Ca²⁺ response in region 1 (Fig. 2C), or in any other region of the neuron (data not shown). Similarly, bursts of four to five orthodromic Na⁺ spikes at stage P4 did not evoke any detectable Ca²⁺ response in MN5 (Fig. 2E). In contrast, individual Ca²⁺ spikes, which occur spontaneously, or can be elicited orthodromically at pupal stage P8 (Duch and Levine 2000) led to a pronounced Ca²⁺ response in region 1 of MN5 (Fig. 2D). During multiple Ca²⁺ spikes, the amplitudes of the intracellular Ca²⁺ increases summed in an approximately linear fashion (Fig. 2F). At pupal stage P4, prolonged positive current injections of 3 nA led to high-frequency bursts of Na⁺ spikes, which caused slow Ca²⁺ responses of ~5% amplitude in region 1 (Fig. 2G). In contrast, at pupal stage P8 current injection of 3 nA for only 250 ms induced a brief burst of four Ca²⁺ spikes that led to rapid Ca²⁺ responses of >15% amplitude (Fig. 2H). In summary, Na⁺ spikes at pupal stage P4 evoked only minor increases in intracellular Ca²⁺ in MN5 that were only detectable during strong bursts, whereas individual Ca²⁺ spikes at pupal stage P8 evoked large elevations of the intracellular Ca²⁺ concentrations in region 1. Although the Na⁺ spikes that occur at pupal stage P4 cause only small depolarizations in the soma, the level of depolarization in region 1 must be quite high because action potentials are propagated actively along the axon. Therefore the difference in the Ca²⁺ responses at this site between the pupal stages P4 and P8 was probably not due simply to the level of depolarization, but to the density of Ca²⁺ channels in region 1. This is further supported by subthreshold current injections into MN5 at pupal stage P8 (Fig. 3). Single Ca²⁺ spikes elicited by current injections just over threshold led to Ca²⁺ responses of >5% amplitude. In contrast, the same cell showed weak Ca²⁺ responses of ~2% amplitude on subthreshold current injections (Fig. 3). In some preparations, it was possible to evoke a Na⁺ spike while staying subthreshold for the induction of a Ca²⁺ spike (Fig. 3). Like subthreshold current injections, single Na⁺ spikes evoked weak Ca²⁺ responses of ~2% amplitude at pupal stage P8 (Fig. 3). In contrast, at pupal stage P4, when Ca²⁺ spikes cannot be elicited (Duch and Levine 2000), single Na⁺ spikes and current injections just below the firing threshold did not cause a detectable Ca²⁺ response (Fig. 2), although the firing threshold is roughly 40 mV at both stages (Duch and Levine 2000). Therefore at similar membrane potentials, Ca²⁺ signals were induced at pupal stage P8 but not at pupal stage P4. At pupal stage P4, only much stronger depolarizations, such as those induced by trains of orthodromic (Fig. 2) or antidromic spikes (see following text) led to Ca²⁺ responses in region 1. 

View larger version (37K): [in this window] [in a new window]   Fig. 2. MN5 was filled intracellularly with fura-2 at pupal stages P4 and P8. A: representative image of stage P4 MN5 at the region where the major dendrite branches off the axon (region 1), obtained with a CCD camera at an exposure time of 10 ms. B: representative image of the same region of MN5 at pupal stage P8. In A and in B, the white ovals indicate the region where changes in the emission intensity of fura-2 at an excitation wavelength of 380 nm were measured as responses to intracellularly elicited spikes. C, E, and G: taken from simultaneous intracellular and optical recordings of the cell shown in A (stage P4). D, F, and H: taken from simultaneous intracellular and optical recordings of the cell shown in B (stage P8). C: 3 representative sweeps of the intracellular Ca2+ response (top) at stage P4 to single Na+ spikes (bottom), that were elicited orthodromically by current injections of 0.3-nA amplitude. D: representative Ca2+ response of MN5 (top) at stage P8 to a single orthodromically elicited Ca2+ spike. E: representative Ca2+ responses to a single Na+ spike and to bursts of 4 and 5 Na+ spikes at stage P4 that were elicited orthodromically by current injections of 1- and 1.5-nA amplitude. F: Ca2+ responses in MN5 at stage P8 to 1-3 orthodromically elicited Ca2+ spikes. The intracellular voltage recordings were offset for clarity. The numbers 1-3 indicate which Ca2+ response belongs to which intracellular recording. G: Ca2+ response of MN5 at stage P4 to a high-frequency burst of spikes that was elicited by orthodromic current injection of 3-nA amplitude. H: representative Ca2+ response of MN5 at stage P8 to a brief burst of 4 Ca2+ spikes that was elicited orthodromically. Scale is 30 µm.

View larger version (31K): [in this window] [in a new window]   Fig. 3. Ca2+ responses of the primary dendrite (region 1) of MN5 at pupal stage P8 to a single orthodromically elicited Ca2+ spike (1), to a single orthodromically elicited Na+ spike (2), and to a subthreshold current injection (3). The intracellular voltage recordings were offset for clarity.

To test whether Ca²⁺-spike-induced Ca²⁺ elevations were distributed evenly throughout different regions of the neuron at pupal stage P8, different compartments of MN5 were analyzed (Fig. 4). The same eight regions of interest were defined and imaged in eight different preparations of pupal stage P8. Region 1 was defined as described in the preceding text. Region 2 was set in a distance of 200 µm from region 1 along the link segment toward the soma (Fig. 4A). Region 3 was set in a distance of 100 µm from region 1 along the axon toward nerve 1 (Fig. 4A). Region 4 was set in distance of 150 µm from region 1 along the primary dendrite of MN5 (Fig. 4A). The somatic Ca²⁺ responses of the same cells were measured in a different field of view (region 8, Fig. 4B). In an additional field of view, three regions along the link segment with distances of 100 µm relative to each other were analyzed (Fig. 4C, regions 5-7). Region 5 was set 50 µm to region 1 on the link segment. Individual Ca²⁺ spikes evoked very similar Ca²⁺ responses in all 8 regions analyzed (Fig. 4, D-F). The time constants of the decay of the signal were not significantly different among any of the regions shown in Fig. 4A (P  0.1; Fig. 4, D and G). The time constants measured from the somata of the same preparations were considerably longer (2.9 ± 0.8; P = 0.042). This difference could reflect a true difference in calcium dynamics but may have been due to the different volume to surface ratio of the soma at the two stages, or the excessive dye concentrations in the soma when the cells were filled to visualize dendritic and axonal regions. Therefore in three preparations of pupal stage P8, MN5 was filled for only 5 min. In these preparations, the time constant of decay in the soma was similar to that observed in all other regions (1.4 ± 0.7; P  0.2). This also explains the slightly longer time constant in region 2 in Fig. 4G because this region was closest to the soma. Therefore the time constant of decay of the intracellular Ca²⁺ elevations in response to a single Ca²⁺ spike was similar in all compartments of MN5. The same was true for the signal amplitude (P  0.1; Fig. 4, H and I).

View larger version (51K): [in this window] [in a new window]   Fig. 4. Ca2+ responses to single orthodromically elicited Ca2+ spikes in different parts of MN5 at stage P8. A: regions of MN5 at stage P8 that were imaged at 380-nm excitation. The white circles, 1-4, indicate the 4 regions in which the intracellular Ca2+ responses to single orthodromic spikes were measured. B: the soma of the same neuron that is shown in A. C: the primary neurite linking the soma and the dendritic field of MN5 from the same neuron as in A and B. The white circles, 5-7, indicate the 3 regions in which the intracellular Ca2+ responses to single spikes were measured. Region 5 lies ~50 µm right of region 1 defined in A. D: the Ca2+ responses (top) to a single orthodromically elicited Ca2+ spike (bottom) in the 4 different regions indicated in A (regions 1-4). The responses were similar in time course and amplitude. E: the Ca2+ responses (top) in 3 regions indicated in C to a single Ca2+ spike (regions 5-7). F: a representative somatic Ca2+ response to a single Ca2+ spike (region 8). G: mean time constant of the decay of the Ca2+ response in the 4 regions shown in A and in D. The values were obtained from 8 different preparations, for each of which the Ca2+ responses to 3 single Ca2+ spikes were evaluated. The error bars represent the SD. No statistical differences were detected among the 4 regions. H: mean amplitude of the Ca2+ response to single Ca2+ spikes in the 4 different regions indicated in A. The values were obtained from the Ca2+ responses to 3 spikes each in 8 different animals. The error bars represent the SD. I: mean amplitude of the Ca2+ response to single Ca2+ spikes in the soma (average of 3 preparations in which the soma was filled with fura-2 for 5 min only) and in the 3 different regions of the primary neurite indicated in C. The values were obtained from the Ca2+ responses to 3 spikes each in 6 different animals. The error bars represent the SD. No statistical difference was found between any combination of the means shown in H and in I (paired t-test for all values obtained from the same imaging sequences, i.e., among the 4 regions shown in A and among the 3 regions shown in C; unpaired t-test for all comparisons of values obtained from the same preparation but from different imaging sequences, i.e., the somatic values to all other regions, and the 3 regions shown in C to the 4 regions shown in A). Scale is 50 µm.

In summary, the Ca²⁺ spikes that occur at pupal stage P8 led to strong elevations of the internal Ca²⁺ concentration throughout all compartments of MN5. In contrast, single Na⁺ spikes at pupal stage P4 produced no detectable signal with the Ca²⁺ imaging techniques used in this study. However, high-frequency bursts at pupal stage P4 led to small elevations of the internal Ca²⁺ concentration (Fig. 2). Therefore it appears likely that at least some parts of the membrane of MN5 support a low Ca²⁺ channel density at pupal stage P4. Although single-electrode voltage-clamp experiments from the soma of MN5 in situ did not reveal Ca²⁺ currents at pupal stage P4 (Duch and Levine 2000), other membrane regions might allow small Ca²⁺ currents that were not detectable in somatic recordings due to space-clamp problems. The following experiments were conducted to explore this possibility and to further compare Ca²⁺ signals at the two developmental stages.

Antidromic stimulation of the axon of MN5 allows the motoneuron to be driven at high frequencies. Furthermore it does not evoke Ca²⁺ spikes at pupal stage P8 (Duch and Levine 2000). Thus antidromic stimulation is a useful method for comparing the intracellular Ca²⁺ responses at equivalent levels of depolarization at both stages. Antidromic stimulation with 60 Hz for 10 s led to Ca²⁺ responses in region 1 in both developmental stages (Fig. 5A). The amplitudes of the Ca²⁺ responses were significantly higher in preparations of pupal stage P8 (Fig. 5, A and B). Fitted curves for the decay of the Ca²⁺ signals are shown in the inset in Fig. 5A (P4 gray lines, P8 black lines). The time constant was significantly longer for pupal stage P4 (Fig. 5C), indicating that either the Ca²⁺ buffering was stronger or the Ca²⁺ extrusion was slower at this developmental stage. Thus Ca²⁺ responses during high-frequency Na⁺ spikes that were evoked by antidromic stimulation were of higher amplitude and had a faster decay at pupal stage P8 than at pupal stage P4. This quantitative comparison of the Ca²⁺ dynamics prompted the concern that the concentrations of free Ca²⁺ indicator differed in MN5 between the two developmental stages, attributing to the different volumes and possibly also to different protein compositions. To test whether the higher signal amplitude measured in MN5 at pupal stage P8 as compared with pupal stage P4 was simply due to buffering effects of the Ca²⁺ indicator (Fig. 5, A and B), some preparations were measured in dual-wavelength mode. Although this compromised the temporal resolution, dual-wavelength recordings are far less dependent on the dye concentration and thus serve as a useful comparison with the single-wavelength recordings that were conducted with high time resolution. As in single-wavelength mode, dual-wavelength mode revealed that antidromic stimulation with 60 Hz for 10 s led to Ca²⁺ responses in region 1 in both developmental stages (Fig. 5D). Furthermore the amplitudes of the Ca²⁺ responses were about twice as large in preparations of pupal stage P8 (Fig. 5, D and E), thus confirming the data obtained in single-wavelength mode (Fig. 5B). We could not definitively test whether the time constant might have been influenced artificially by differences in the free indicator concentration among the stages. However, varying the dye injection times by 30% for each stage (n = 2 for each stage), and test experiments with calcium green 5 (n = 3 for each stage) always showed that the time constant at pupal stage P4 was about twice as long as compared with pupal stage P8 (data not shown). This strongly suggests that the differences in time course found between developmental stages were not artificially produced by different buffering effects of the indicator dye but rather due to real differences in the calcium dynamics. Although the absolute Ca²⁺ concentrations were not determined, the ratio of resting fluorescence at 340- and 380-nm excitation was significantly higher at pupal stage P8 than at pupal stage P4 (P  0.01; Fig. 5, D and F), indicating a higher resting intracellular Ca²⁺ level.

View larger version (63K): [in this window] [in a new window]   Fig. 5. Ca2+ responses of MN5 at pupal stages P4 and P8 to antidromically elicited Na+ spikes at a stimulation frequency of 60 Hz. A: 3 representative Ca2+ responses to 60-Hz stimulation from 3 different preparations at pupal stage P8 (black dots interconnected with black lines) and 3 representative Ca2+ responses to 60-Hz stimulation from three different preparations at pupal stage P4 (gray triangles), as recorded in single wavelength mode. Insets: single exponential fits of the decay for the 3 preparations from pupal stage P8 (black lines) and the 3 preparations from pupal stage P4 (gray lines). The Simplex function that is provided by Clampfit 8 software was used for the standard exponential fits. The maximum number of iterations was set to 5,000, and the sum of square errors was selected as minimization method. B: mean amplitude of the Ca2+ responses to 60-Hz antidromic stimulation for P4 (n = 7) and for P8 (n = 9). The error bars represent the SD. The asterisk indicates that the difference is statistically significant (unpaired t-test, P  0.01). C: mean time constant () of the decay of the Ca2+ signal for P4 (n = 7) and for P8 (n = 9). The error bars represent the SD. The asterisk indicates that the difference is statistically significant (unpaired t-test, P  0.001). D: 3 representative Ca2+ responses to 60-Hz stimulation from 3 different preparations at pupal stage P8 (black dots interconnected with black lines) and 3 representative Ca2+ responses to 60-Hz stimulation from 3 different preparations at pupal stage P4 (black squares interconnected with black lines) as recorded in dual wavelength mode. As in single-wavelength mode, the signal amplitude is significantly higher at pupal stage P8. This ratiometric method addressed the possibility that differences in signal amplitude were due to different calcium indicator concentrations at P4 as compared with P8. E: mean amplitude of the Ca2+ responses to 60-Hz antidromic stimulation for P4 (n = 5) and for P8 (n = 8), expressed as the ratio of fluorescence intensity at 340 nm/380 nm. The error bars represent the SD. The asterisk indicates that the difference is statistically significant (unpaired t-test, P  0.01). F: mean Ca2+ resting ratios obtained with dual wavelength measurements at the pupal stages P4 (n = 10) and P8 (n = 10). The error bars represent the SD. The asterisk indicates that the difference is statistically significant (unpaired t-test, P  0.001).

To test whether the Ca²⁺ responses that were evoked antidromically were caused by Ca²⁺ influx through the cell membrane, Ca²⁺ was replaced by Mg²⁺ in the extracellular solution (Fig. 6). The control stage P4 Ca²⁺ response in normal saline (Fig. 6, gray curve) was abolished after 10 min in Ca²⁺ free saline (Fig. 6). Washing in normal saline led to 50% recovery of the initial response amplitude after 5 min and to a full recovery after 10 min. Therefore the Ca²⁺ responses were dependent on external Ca²⁺. At pupal stage P8, the Ca²⁺ response to antidromic stimulation could also be abolished in Ca²⁺-free saline (data not shown). However, this does not exclude the possibility that additional Ca²⁺ release from intracellular stores contributes to the higher amplitude of the Ca²⁺ response at pupal stage P8.

View larger version (31K): [in this window] [in a new window]   Fig. 6. Ca2+ responses of MN5 at pupal stage P4 to antidromically elicited Na+ spikes at a stimulation frequency of 60 Hz in normal and in Ca2+-free saline. The gray line shows a control stimulation in normal saline. The response was almost completely abolished after 10 min in Ca2+-free saline. After 5 min in normal saline 50% of the response amplitude were recovered. After 10 min in normal saline the full response amplitude was recovered.

Pharmacological experiments were conducted to examine whether the Ca²⁺ influx through the cell membrane caused Ca²⁺ release from intracellular stores. The Ca²⁺ ATPase inhibitor cyclopiazonic acid (CPA) depletes intracellular Ca²⁺ stores (Lohr and Deitmer 1999). CPA (20 µM) clearly acted on the intracellular Ca²⁺ stores of MN5 in situ at both stages, P4 and P8, as shown by ratiometric imaging of the resting Ca²⁺ concentration. For example, in a stage P8 MN5, the resting ratio at region 1 increased significantly without depolarization after 90 s of CPA bath application (Fig. 7B). After an additional 90 s, the resting ratio again declined (Fig. 7B) probably due to the activity of Ca²⁺ pumps in the cytoplasmatic membrane. Imaging the Ca²⁺ responses to antidromic stimulation before and after this effect of CPA revealed no significant differences in the Ca²⁺ responses (Fig. 7C). In summary, MN5 showed CPA induced Ca²⁺ release from intracellular stores at both developmental stages, but this had no effect on the depolarization induced intracellular Ca²⁺ rise at stage P8.

View larger version (30K): [in this window] [in a new window]   Fig. 7. Ca2+ responses of MN5 at pupal stage P8 to antidromically elicited Na+ spikes at a stimulation frequency of 60 Hz before and after applying CPA. A: intensity measurements were executed at region 1 (white oval). Scale is 50 µM. B: the resting ratio between excitation at 340 and at 380 nm. The time point when CPA was added to the bath solution (20 µM) is indicated with an arrow. One hundred fifty seconds after CPA application, the resting Ca2+ ratio increased, suggesting release from Ca2+ stores; 200 s after CPA application, the resting Ca2+ ratio decreased again probably because Ca2+ was pumped out of the cell. C: the intracellular Ca2+ response to 60-Hz antidromic stimulation for 10 s was imaged at a single wavelength (380 nm) before CPA application (control, black line) and 600 s after CPA application (gray line). No differences were observed between the control response and the response after CPA application.

To determine whether the Ca²⁺ influx through voltage-dependent calcium channels (VDCCs) that was induced by antidromic stimulation was distributed evenly throughout different compartments of MN5, different regions were analyzed while applying the same antidromic stimulation protocol. The regions are indicated by the white ovals in Fig. 8, A (pupal stage P4) and B (pupal stage P8). The Ca²⁺ responses differed significantly among the different regions at both stages (Fig. 8, C and D). At pupal stage P4, the strongest Ca²⁺ response was observed in the dendrites (region 4). In some preparations, the resolution was high enough to image individual growth cones (Fig. 8E). There the Ca²⁺ responses were as high as those observed in the dendrites. The Ca²⁺ responses observed in the axon ~100 µm away from region 1 were 50% of the amplitude observed in region 1 and only 15% of the amplitude observed in the growth cones and the dendrites (Fig. 8C). Antidromic stimulation evoked no Ca²⁺ signals in the link segment (Fig. 8C) or in the soma (Fig. 8F) at pupal stage P4.

View larger version (71K): [in this window] [in a new window]   Fig. 8. Ca2+ responses of different regions within MN5 at pupal stages P4 and P8 to antidromically elicited Na+ spikes at a stimulation frequency of 60 Hz. A: the regions of MN5 at stage P4 that were imaged at 380-nm excitation. B: the regions of MN5 at stage P8 that were imaged at 380-nm excitation. In A and B, the white circles 1-4 indicate the 4 regions in which the intracellular Ca2+ responses to antidromically evoked Na+ spikes at 60 Hz for 10 s were measured. Scale is 50 µM. C: representative Ca2+ responses in the 4 regions indicated in A at pupal stage P4. In the regions 1-3, the Ca2+ concentration increased between 3 and 5%. In the main dendrite (region 4), the same stimulation led to Ca2+ concentration increases of 15-20% to 60-Hz antidromic stimulation. D: representative Ca2+ responses in the 4 regions indicated in B at pupal stage P8. In regions 1, 2, and 4, the Ca2+ concentration increased between 10 and 20% in response to 60-Hz antidromic stimulation. In the axon (region 3) Ca2+ increases of 50% were observed. E: representative Ca2+ responses to antidromic stimulation with 60 Hz for 10 s in the distal part of the primary dendrite and in a representative growth cone at pupal stage P4. Right: the regions of the intensity measurements. Both the distal part of the primary dendrite and the growth cones show similar Ca2+ increases of almost 20% in response to the stimulation protocol. Scale is 20 µm. F: no Ca2+ response occurred to antidromic stimulation with 60 Hz for 10 s in the soma at pupal stage P4. The white oval indicates the region of the intensity measurement in a representative preparation of pupal stage P4. Scale is 50 µM.

In contrast, at pupal stage P8, a very different distribution of the Ca²⁺ responses within different parts of MN5 was found (Fig. 8D). The strongest response amplitudes were observed in the axon ~100 µm away from region 1 (Fig. 8D) in contrast to pupal stage P4 where the Ca²⁺ elevations were most pronounced in the dendrites. At pupal stage P8, the dendritic Ca²⁺ responses were ~60% of the amplitude that was observed in region 1. Furthermore at pupal stage P8, Ca²⁺ responses were observed in the link segment. The amplitude of these signals was similar to those observed in the dendritic region (Fig. 8D). In summary, at pupal stage P4, Ca²⁺ elevations in response to antidromically evoked Na⁺ spikes were mostly restricted to the dendritic regions, whereas at pupal stage P8 Ca²⁺ elevations were observed throughout all central compartments of the cell, with the highest amplitudes in the axon, where the antidromic spikes would be of greatest amplitude. Therefore at pupal stage P8, the largest Ca²⁺ influx was observed where the largest depolarization occurred. In contrast, at P4 the largest Ca²⁺ influx was observed in the dendrites despite the larger depolarization in the axonal region.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The key finding of this study is that activity-dependent Ca²⁺ influx via VDCCs occurs in the dendrites of MN5 but changes during development. The changes in Ca²⁺ signaling coincide with distinct phases of dendritic growth, suggesting that the developmental modifications in ionic currents are not only important for the new adult behavioral role of MN5, but also for specific aspects of the dendritic modifications that occur during the integration into the flight motor network.

Mechanisms of Ca²⁺ accumulation in MN5 at different developmental stages

At pupal stage P8, single Ca²⁺ spikes caused elevations of the internal Ca²⁺ concentration of >5% in all central compartments of MN5. This Ca²⁺ accumulation was likely to be caused entirely by influx through VDCCs because multiple Ca²⁺ spikes led to approximately linear increases in intracellular free Ca²⁺. Ca²⁺-dependent Ca²⁺ release from internal stores did not appear necessary, and removal of extracellular Ca²⁺ abolished the responses. Although substantial Ca²⁺ responses occurred in response to propagating Ca²⁺ spikes, subthreshold depolarizations alone could induce low levels of Ca²⁺ influx, probably through low-voltage-activated Ca²⁺ channels, as in the dendrites of fly visual interneurons (Oertner et al. 2001). No significant differences in Ca²⁺ response amplitudes were observed among soma, axon, and dendrites of MN5 following the Ca²⁺ spike, indicating a rather even distribution of VDCCs throughout the neuron.

Even at stage P4, when whole cell Ca²⁺ currents are negligible (Duch and Levine 2000) and no Ca²⁺ spikes can be elicited, slow elevations in intracellular Ca²⁺ levels occurred in the dendrites in response to trains of action potentials. Surprisingly, the largest Ca²⁺ influx on antidromic stimulation did not occur in the axon where depolarizations caused by the Na⁺ spikes would be highest but rather in the dendrites and even in the distal growth cones. The activity-dependent increase of internal Ca²⁺ at P4 depended on external Ca²⁺ as shown by Ca²⁺ replacement experiments. Furthermore, Ca²⁺ release from intracellular stores was probably not involved, although intracellular stores were present in the dendrites at both developmental stages as shown by elevations in internal free Ca²⁺ after CPA application. Second-messenger-induced Ca²⁺ release from internal stores might be an additional regulatory mechanism in dendritic Ca²⁺ signaling (David and Pitman 1996; Wand and Augustine 1995).

Voltage-dependent dendritic Ca²⁺ influx has recently been reported to serve different functions among different types of neurons such as cricket giant interneurons (Ogawa et al. 2000), tangential cells of the fly visual system (Oertner et al. 2001), and hippocampal pyramidal cells (Christie et al. 1995; Isomura and Kato 1999). In pyramidal cells, dendritic Ca²⁺ influx through VDCCs plays critical roles in the induction of LTP (Magee and Johnston 1997), whereas in fly tangential cells, it might serve local adaptation to visual motion stimulation (Oertner et al. 2001). Its functional relevance for the adult MN5 remains to be investigated. In MN5, the net Ca²⁺ currents is unchanged from pupal stage P8 through adulthood (Duch and Levine 2000). Thus Ca²⁺ influx in the adult MN5 might be important for dendritic signal integration and for modulating the firing properties.

The activity-dependent Ca²⁺ signals at stage P4 in the absence of Ca²⁺ spikes might occur via voltage-dependent calcium channels in the dendritic regions. Action potentials cause large elevations of internal Ca²⁺ in hippocampal neurons by influx through VDCCs (Christie et al. 1995; Spruston et al. 1995). This mechanism would require that Na⁺ spikes depolarize distal dendritic regions of MN5. Action potentials ordinarily do not propagate actively into the dendritic region of insect motoneurons, but the distal dendrites are probably depolarized passively, especially in the reduced dendritic field at stage P4. The distal dendrites and the growth cones show the largest Ca²⁺ response to trains of Na⁺ spikes either because the calcium channel density is highest there or because of restricted diffusion. The fact that we did not observe Ca²⁺ currents in voltage-clamp recordings from the soma at P4 (Duch and Levine 2000) is consistent with the former possibility. Alternatively, increases in internal Ca²⁺ at pupal stage P4 might reflect influx through the Na⁺-Ca²⁺ exchanger stimulated by a rise in internal Na⁺ (Blaustein 1988). In contrast, at pupal stage P8, the Na⁺-spike-induced dendritic Ca²⁺ signals are smaller than the axonal ones. This is consistent with the hypothesis that VDCCs are evenly distributed throughout MN5 but that Na⁺ fails to depolarize distal dendrites in the more complex dendritic field at this stage.

Putative role of Ca²⁺ spikes for dendritic growth

The Ca²⁺ spikes that occur transiently during the developmental modifications of MN5 strongly affect the dendritic Ca²⁺ concentration. As shown previously, these spikes occur exclusively during pupal stages P7-P9 and thus correlate temporally with a switch in the mode of dendritic growth (Duch and Levine 2000). The extensive dendritic growth-cone branching that takes place in MN5 during earlier pupal stages ceases during the stages when these Ca²⁺ spikes occur. Growth cones are not observed during the second half of pupal development (Duch and Levine 2000; Libersat and Duch, unpublished morphometric analysis). Because the Ca²⁺ spikes occur spontaneously and can be evoked by sensory stimulation in the isolated ganglion preparation, it is likely that they also occur during normal development (Duch and Levine 2000). In cultured neurons, activity-dependent Ca²⁺ influx can operate as a growth-stopping signal (Baird et al. 1996), and high levels of Ca²⁺ influx suppress motile growth-cone structures (Mattson and Kater 1987). Membrane-potential-dependent Ca²⁺ influx into the growth cones and adjacent dendrites inhibits elongation and finally leads to a collapse of the growth cone (Kater and Mills 1991). The present study shows that strong Ca²⁺ influx into the dendrites of MN5 in situ correlates temporally with the cessation of growth-cone branching, suggesting that these spikes serve to stop growth-cone branching during metamorphosis. A restriction of dendritic branch number might have important consequences for the generation of a functional flight motor circuit.

Ca²⁺ homeostatic mechanisms can override depolarization induced growth-cone inhibition (Fields et al. 1990, 1993). Therefore it is important to note that the resting ratio of F₃₄₀/F₃₈₀ in MN5 was significantly higher at pupal stage P8 than at pupal stage P4. Although the measurements were not calibrated to calculate the internal Ca²⁺ concentrations, they indicate a higher level at P8. This suggests that the higher Ca²⁺ influx at pupal stage P8 is not counteracted by developmental increases in the Ca²⁺ handling mechanisms. This is further supported by the relatively long-lasting increase in internal Ca²⁺ levels in response to a single Ca²⁺ spike.

Ultimately, the Ca²⁺ influx has to be translated into changes in the cytoskeleton to affect neuronal growth. Promising candidates via which Ca²⁺ signals might be read are the Ca²⁺-sensitive enzymes CAM-K II and calcineurin. In the developing Xenopus optical tectum, CAM-K II is required to limit the elaboration of neuronal arbors (Zou and Cline 1999). In hippocampal neurons, CAM-K-II-dependent phosphorylation of MAPK is critical for dendritic spine morphology (Wu et al. 2001). During embryonic spinal cord development, Ca²⁺ transients inhibit the rate of axon outgrowth (Gomez and Spitzer 1999) by mediating an increase in the activity of calcineurin (Lautermilch and Spitzer 2000). The functional role of these pathways in insect motoneuron remains to be investigated.

The Ca²⁺ spikes probably propagate actively into the soma because somatic and dendritic Ca²⁺ responses to single Ca²⁺ spikes were similar in amplitude and time course. Although the functional importance of the somatic Ca²⁺ spikes in MN5 remains unclear, nuclear Ca²⁺ signals can affect gene transcription (Hardingham et al. 1997) and thus play a regulatory role during neuronal differentiation. Therefore it will be important to examine whether the somatic Ca²⁺ signals in MN5, which occur only during specific times of postembryonic development, are translated into nuclear Ca²⁺ elevations.

Putative role for low levels of Ca²⁺ influx into distal dendrites and growth cones at P4

The effects of intracellular Ca²⁺ on growth-cone behavior are concentration dependent (reviewed in Kater and Mills 1991). Briefly, at optimal levels of internal Ca²⁺, outgrowth is profuse, but lower or higher levels result in decreased outgrowth. Therefore the low levels of Ca²⁺ influx at stage P4 might allow MN5 to maintain optimal Ca²⁺ levels and promote growth-cone branching. During normal development, MN5 is probably never active at the high frequencies that were used for antidromic stimulation in this study, and thus fast elevations of the Ca²⁺ concentrations by >20% are unlikely to occur. Nevertheless, Na⁺-spike-induced Ca²⁺ influx was also detectable at antidromic stimulation frequencies of 5 Hz, although the responses were more variable and less robust (data not shown). In cultured neurons, changes in the internal Ca²⁺ concentration as little as 30-50 nM reliably alter filopodia morphology (Rehder and Kater 1992). Therefore moderate activity at pupal stage P4 might allow low levels of Ca²⁺ influx within a permissive range for dendritic growth.

Putative role for Ca²⁺ influx in the adult MN5

The large net Ca²⁺ current in MN5 remains unchanged from pupal stage P8 through adulthood (Duch and Levine 2000). Voltage-dependent dendritic Ca²⁺ influx serves a variety of functions in different neurons (Christie et al. 1995; Isomura and Kato 1999; Oertner et al. 2001; Ogawa et al. 2000). In pyramidal cells, dendritic Ca²⁺ influx through VDCCs plays critical roles in the induction of LTP (Magee and Johnston 1997), whereas in fly tangential cells it might serve local adaptation to visual motion stimulation (Oertner et al. 2001). A testable possibility in the case of MN5 is that large Ca²⁺ currents, and perhaps elevations in internal Ca²⁺, are important for participation in flight. Ca²⁺ conductances play a major role in the generation of plateau potentials (Hancox and Pitman 1991, 1993; Hartline and Russel 1984; Kiehn 1991). Although the adult MN5 displayed no plateau potentials in the isolated ganglion preparation, prolonged Ca²⁺ action potentials could be evoked after applying TEA (Duch and Levine 2000). In many systems, plateau potentials can be induced by neuromodulators or any intervention that sufficiently reduces opposing outward currents (Hultborn and Kiehn 1992; Kiehn and Harris-Warrick 1992; Ramirez and Pearson 1991). Flight behavior, for example, is strongly influenced by the neuromodulator octopamine, which evokes plateau potentials in locust flight interneurons (Orchard et al. 1993). One possibility therefore is that the adult Ca²⁺ currents come into play because opposing outward currents are reduced by modulatory signals during flight. In particular, moths display a prolonged warm-up phase during which neuromodulators may readily prepare the CNS for flight behavior (Claassen and Kammer 1986). To explore the function of developmental changes in the Ca²⁺ signaling in identified neurons, both developmental and behavioral consequences will have to be considered. A challenging task will be to understand how changes in the Ca²⁺ currents that are important for the adult firing properties of a neuron might also be used as signals for developmental modifications, such as dendritic remodeling.