Neuron.
Vol. 8, 883-889,
May, 1992, Copyright
0 1992 by Cell Press
Recruitment of Ca2+ Channels by Protein Kinase C during Rapid Formation of Putative Neuropeptide Release Sites in Isolated Aplysia Neurons Ronald J. Knox,* Elizabeth A. Quattrocki,* John A. Connor,+ and Leonard K. Kaczmarek** *Department of Pharmacology *Department of Molecular and Cellular Physiology Yale University School of Medicine New Haven, Connecticut 06510 +Roche Institute of Molecular Biology Roche Research Center Nutley, New Jersey 07110
Summary Activation of protein kinase C (PKC) in Aplysia bag cell neurons causes the recruitment of voltage-dependent calcium channels. Using imaging techniques on isolated cells, we have now found that an activator of PKC, 12-Otetradecanoyl-phorbol-13-acetate (TPA), promotes the rapid appearance of new sites of calcium influx associated with a change in the morphology of neurite endings. In untreated cells, calcium influx triggered by action potentials occurs along neurites and in the central region of growth cones, but does not usually occur at the leading edge of lamellipodia. TPA produces extension of the lamellipodium, and action potentials now trigger calcium influx at the distal edge of the newly extended endings. Cotreatment with TPA and a cyclic AMP analog promotes movement of secretory organelles toward the new sitesof calcium influx. Our results suggest that these second messenger systems promote the rapid formation of morphological structures that contributeto thepotentiation of peptide release. Introduction Brief electrical stimulation of an afferent input to the bag cell neurons of Aplysia triggers a prolonged discharge (-30 min) of action potentials during which there occurs a stimulation of phosphoinositide turnover (Fink et al., 1988) and an increase in intracellular cyclic AMP levels (Kaczmarek et al., 1978; Kaczmarek and Kauer, 1983). This discharge leads to the release of several neuroactive peptides, including egg-laying hormone, which in vivo, triggers a sequence of reproductive behaviors (Conn and Kaczmarek, 1989). During the first IO-15 min of the discharge there is a progressive potentiation of the ability of action potentials to evoke the secretion of egg-laying hormone from the neurites of these cells (Loechner et al., 1990). The potentiation of peptide release during a discharge may result, in part, from the stimulation of protein kinase C (PKC), which leads to an enhancement of calcium influx. The height and width of action potentials of bag cell neurons are determined to a large extent by the amplitude of the voltage-dependent calcium currents (Kaczmarek et al., 1982). Over the first few minutes of the discharge there is a pro-
gressive increase in the amplitude of action potentials, which is blocked by inhibitors of PKC (Conn et al., 1988b). Microinjection of PKC or application of activators of this enzyme to isolated neurons produces an enhancement of action potentials that resembles the increase occurring during an evoked discharge (DeRiemer et al., 1985b). Patch-clamp experiments have shown that this occurs through the unmasking of a new species of voltage-dependent calcium channel on the plasma membrane. In untreated bag cell neurons, only calcium channelswith a unitary conductance of 12 pS are detected (Strong et al., 1987). In cells exposed to activators of PKC, however, a novel 24 pS calcium channel can also be recorded (Strong et al., 1987), and the net voltage-dependent calcium current is increased 2-fold or more (DeRiemer et al., 1985b). Stimulation of a discharge also evokes an increase in cyclic AMP levels in bag cell neurons (Kaczmarek et al., 1978). Although microinjection of the catalytic subunit of the cyclic AMP-dependent protein kinase, or application of cyclic AMP analogs, has been shown to produce spike broadening in isolated bag cell neurons, voltage-clamp analysis indicates that these act primarily on potassium conductances. No effect of cyclic AMP has been detected on the major voltagedependent calcium current in bag cell neurons. An elevation of cyclic AMP levels, however, also modifies several other properties of bag cell neurons that may contribute to potentiation of peptide secretion. These include increased synthesis of the peptide precursor (Bruehl and Berry, 1985) and enhancement of movement of secretory granules along neurites and into growth cones (Forscher et al., 1987; Azhderian et al., 1991, Sot. Neurosci., abstract). We have now used video-enhanced microscopy, coupled with digital imaging of intracellular calcium levels, to examine the actions of 12-O-tetradecanoylphorbol-13-acetate (TPA), an activator of PKC, and of a cyclic AMP analog on morphology and on distribution of calcium within the neuritesof bag cell neurons, where releaseof peptides such as egg-laying hormone is believed to occur (Dudek and Blankenship, 1976; Arch, 1976; Frazier et al., 1967; Kaczmarek et al., 1978). Results Phorbol Ester Treatment Induces Extension of Lamellipodia Bag cell neurons were isolated from the abdominal ganglion of Aplysia californica and plated on a polylysine substrate. Figure 1 and Figure 2a show the typical appearance of their neurites when viewed through fluorescence and video-enhanced Nomarski optics, respectively. After several hours or more, their growth cones are very large (30-100 Grn) and contain a central cytoplasmic region that is rich in organelles and mi-
--
22
SALINE
i
5.5
-
ka2+lPM
20pm
20 nM TPA Figure
1. Phorbol
Ester Alters
the Spatial
Organization
of Calcium
Entry
into
Neurite
Endings
Left panels (1 and 6): fluorescent images (380 nm excitation) of a bag cell neuron growth cone before and after exposure to TPA. Intensities are shown on a modified log scale and are normalized to the maximum intensity for each picture. After TPA treatment the growth cone becomes 2-3 times thicker in thecentral portion. Color pictures 2-5 show maps of internal calcium concentrations before and during the firing of a 3 Hz action potential train as indicated in the panels. Color pictures 7-10 show calcium concentrations for an identical firing pattern after exposure to TPA.
a
Control
(- 40 mins)
Control (0 mins)
20nM TPA (+ 40 mins)
al 40 d 5 E 4 30 5 L? 20
a .E
ko 5 5 s
0
Control
20 nM TPA 100,uM H-7 +TPA
the Mor-
(a) Video-enhanced differential interference contrast images of a bag cell neuron growth cone. In a 40 min period prior to TPA application there is little change in either the size or appearance of the growth cone. TPA treatment produces a marked extension of the lamellipodial membrane at the leading edge of the growth cone, resulting in a significant increase in the areas of lamellae. This morphological change could be prevented by preincubation of the neurons with 100 uM H-7, a selective inhibitor of PKC in bag cell neurons. Scale bar, 20 urn. (b) Summary of the action of TPA on 14 growth cones. Measurement of the area of lamellae was determined as described previously (Forscher et al., 1987).
mm
b
Figure 2. Phorbol Ester Changes phology of Growth Cones
100uM H-7
Cahum 085
Channels
-O--t + a1
in Growth
Cones
BEFORE STIMULATION DURING SPIKE TRAIN RECOVERY 4 B
Comparison of [Cal’], changes durrng firing of action potentials in a growth cone treated first with 200 nM pCPTcAMP and then with the cyclic AMP analog and 20 nM TPA. (A) Maximum [Cal’], changes during firing before either treatment. Locations of analysis boxes for the graphs are shown in the fluorescence pictures (380 nm excitation) immediately below. Note that changes are smallest in the leading edge of the growth cone. (B) [Cal+], changes after 20 min exposure to pCPTcAMP. Both the levels and concentration profiles are similar to those of the untreated neuron. (C) Following treatment with both kinase activators, [Ca2+], changes are much larger during firing, and the greatest change is now at the leading edge of the growth cone. To maximize clarity in black and white reproductions, slightly different scalings were used for each fluorescent picture.
crotubules (Forscher et al., 1987; Forscher and Smith, 1988). As in other cell types, this region is surrounded by a flat actin-rich lamellipodium that is virtually organelle free (Bridgeman et al., 1986). Under control conditions the percentage of total growth cone area occupied by organelles was 40% + 3% (SEM, N = 36; Figure 48). Growth and movement of intracellular organelles were monitored for at least 40 min before addition of activators of protein kinases. In general, relatively little growth occurred over this time period (Figures 2a and 2b). However, within minutes of bath application of IO-20 nM TPA, which activates PKC in bag cell neurons (DeRiemer et al., 1985a), there was a rapid extension of lamellae that was evident within several minutes. Although this morphological change could be observed after only a few minutes exposure to TPA, the effect peaked by 40 min. This resulted in a 23% f 6% increase (N = 14; Figure 2b) in the area of the growth cones. TPA treatment did not alter the distribution of organelles within the central domain of the growth cone (Figure 2a; see also Figure 4B). To test the specificity of action of TPA, cells were pretreated with 100 uM H-7, a selective inhibitor of PKC in these cells (Conn et al., 1988a, 1988b), for 15 min before addition of the phorbol ester. H-7 was found to block the actions of TPA on the neurites of the cells (N = 7; Figure 2b). In addition, the inactive phorbol ester 4a-phorbol 12,13-didecanoate did not produce extension of neurite endings (N = 3). Phorbol Ester Induces Recruitment of Calcium Channels To examine the distribution of voltage-dependent
cal-
cium channels at the tips of neurites, cells were injected with the calcium indicator dye fura-2, and digitized images were made of the ratio of fluorescence at 340 and 380 nm excitation wavelengths. This ratio gives a direct measure of the intracellular calcium ion concentration [Ca?‘],. Trains of action potentials were evoked by depolarizing current pulses applied to the somata of the neurons. The major sites of calcium entry through voltage-dependent calcium channels were visualized as sites where localized elevation of calcium occurred in response to the first few action potentials during a train. [Ca*-1, changes were absolutely dependent upon the presence of calcium in the external bathing medium, and action potentials fired in 0 calcium, 1 mM EGTA saline gave no [Ca*+], increases. During a train of action potentials (3 Hz, 5-15 s) in untreated cells in normal calcium-containing media, calcium levels generally rose uniformly along the length of the neurite, although “hot spots” of calcium entry could sometimes be detected along the central part of the neurites. The leading edges of the growth cones generally showed lower levels of calcium than the more central regions, making it likely that the influx occurred primarily in the central region and then spread by diffusion toward the lamellipodium (Figure 1, panels 2-5). This suggests that such lamellae had a low density of functional voltageactivated calcium channels at the time of observation. Figure 1 (panels 7-10) shows that, after only a few minutesexposuretoTPA,thepatternofcalciuminflux during action potentials changed such that the first detectable sign of calcium entry was now localized to a ring that corresponded anatomically to the leading
NC?UKl” 886
Control
TPA + pCPTcAMP
80 1
Control Figure4.
Invasion
TPA of Growth
Cones
TPA’pCPTcAMP by Secretory
Organelles
-
Control in Response
to a Cyclic
TPA AMP
Analog
and
TPA pCPTcAMP
a Phorbol
Ester
(A) In the presence of 20 nM TPA, pCPTcAMP induces movement of intracellular organelles into the extended lamellipodium. The directed movement of organelles along microtubules is detected as lines extending from the central region of the growth cone into the periphery. Scale bar, 20 urn. (ET) Histograms showing the effects of 20 nM TPA alone and of 20 nM TPA plus 200 PM pCPTcAMP on the area of bag cell lamellae and on the area of the growth cones occupied by organelles (N = 10 neurons).
edge of the newly extended lamellipodial membrane. The greatest initial increase in [Cal+], now occurred in the lamellipodium. The slight decrease in the overall amplitude of the [Ca2’], change within the central region after TPA treatment probably reflects the greatly expanded volume of the growth cone of this cell. Qualitatively similar results were seen in 22 neurons, with 7 neurons showing no significant changes in [Ca2’], after phorbol ester application. In addition, after phorbol estertreatment, increased calcium entry was also detected at the soma and at locations on the central neurite branches that had previously appeared dormant (note also the “bud”at the lower left of the neurite of Figure 1). Control experiments were carried out with an inactive phorbol ester, 4a-phorbol 12,13-didecanoate, which does not activate PKC. No changes in the amount or distribution of calcium influx or in the morphology of the neurite endings were observed in response to this agent (N = 4). Also there was no change in the distribution of calcium influx in cells that failed to respond to TPA. In particular, this occurred in cells with rapidly growing neurites. in these cases we often
saw large [CaL’], changes at the lamellae during firing prior to addition of phorbol ester (see also Cohan et al., 1987), and the application of phorbol ester produced little further enhancement. In these circumstances it is possible that PKC activity is constitutively activated by factors such as substrate contact (Bixby, 1989). The simplest interpretation of the appearance of new sites of calcium elevation is that, as at the soma (Strong et al., 1987), activation of PKC leads to the recruitment of inactive or “covert” calcium channels in the newly extended membrane at the margin of growth cones and other sites on the neurites. We found no evidence for caffeine-releasable calcium stores in the bag cell neurons, as previously reported for amphibian neurons (Lipscombe et al., 1988). However, we cannot entirely eliminate the possibility that other additional factors, such as changes in the spatial distribution of calcium buffering, also contribute in part to the observed changes. Effects of a Cyclic The calcium action
AMP Analog potentials
on Calcium Influx of bag cell neurons
are
Calcium 007
Channels
in Growth
Cones
also enhanced by cyclic AMP analogs or injection of the catalytic subunit of the cyclic AMP-dependent protein kinase (Kaczmarek et al., 1980). Cyclic AMP analogs have not been found to enhance calcium current (Kaczmarek and Strumwasser, 1984), but rather to decrease voltage-dependent potassium currents. We therefore examined the actions of the cyclic AMP analog pCPTcAMP (8+-chlorophenylthio) cyclic AMP, 200 PM) on calcium influx and morphology of the growth cones. In contrast to TPA, treatment with pCPTcAMP alone did not cause elongation of neurite endings (data not shown; see also Forscher et al., 1987). However, during a train of action potentials, a greater rateof [Ca*+], elevation often occurred (Figures 3A and 3B), presumably as a result of broader action potentials, but the spatial profile of calcium entry was the same.
Cotreatment of Bag Cell Neurons with Phorbol Ester and Cyclic AMP Analog Although cyclic AMP analogs do not unmask new calcium channels at the growth cone, they do produce a characteristic change in the appearance of the neurite endings. Previous work (Forscher et al., 1987) has shown that elevation of cyclic AMP levels transforms the lamellipodium of bag cell neurons into a thickened terminal engorged with organelles that are predominantly the peptide-containing neurosecretory granules. We have now found that this effect also occurs in newly formed lamellae that were induced by TPA. Cotreatment of cells with pCPTcAMP and TPA produced extension of the lamellae followed by directed movement of organelles from the central region to the distal edge of the growth cone (Figure 4B). Imaging of calcium levels in such cotreated cells shows that, as in cells treated with TPA alone, novel sites of calcium entry appear at the distal edge (see Figure 3C). The graphs in this figure show the changes in calcium levels during a train of action potentials before and after treatment with pCPTcAMP and TPA. The fluorescent pictures below, which do not represent calcium levels, show the locations at which the calcium levels were measured and were made while the growth of the neurite was stable. This example was chosen, in part, because there was relatively little morphologic change induced by these treatments to confound the interpretation of changes in calcium levels by changes in surface to volume ratio. In the presence of TPA, the distal tip of the neurite became the dominant site of calcium entry. Similar effects were observed in 6 neurons.
Discussion Our results indicate that in the bag cell neurons, coactivation of PKC and the cyclic AMP-dependent protein kinase produces a rapid change in the morphology at the distal tips of bag cell neurites and changes in the pattern of calcium entry. In particular, new sites
of calcium entry are formed and are coupled to the extension of growth cone membrane and to the movement of secretory organelles toward sites subjacent to the newly activated calcium entry sites. The simplest interpretation of these data is that, as at the soma (Strong et al., 1987), stimulation of PKC causes the recruitment of a new species of voltage-dependent calcium channel at the tips of bag cell neurites. It is highly unlikely that the changes in calcium levels during action potentials that we have described could result purely from changes in growth cone morphology, without the recruitment of new calcium channels. In particular, because the growth cones become thicker after treatment with TPA, the changes in the surface to volume ratio should reduce rather than increase calcium levels in response to a given influx of calcium. In the experiments of Strong et al. (1987), patch-clamp recordings on the somata of untreated bag cell neurons revealed only one species of calcium channel, which has a unitary conductance of 12 pS when measured with barium as the permeant ion. Activation of PKC by phorbol esters or synthetic diacylglycerols brought about the appearance of an additional novel calcium channel with a unitary conductanceof24pS.Similarunmaskingofthis24pSchannel in the extended famellipodium could explain the changes in calcium distribution that we have observed. Because of the difficulty of patch clamping the very thin famellipodial membrane, however, we cannot exclude the possibilitythat thechannelsat the terminals differ from either of the two types detected on the soma. It is also possible that, in addition to new sites of calcium influx, changes in parameters such as local calcium buffering may partly contribute to the observed changes in calcium distribution during trains of action potentials. Our results also show that exposure of bag cell neurons to TPA produces rapid extension of lamellipodial membrane, presumably through a dynamic change in the growth cone cytoskeleton. It is possible that the extension of the membrane is linked mechanistically to the appearance of the new calcium entry sites. For example, extension of membrane in other cell types is thought to result from fusion and incorporation of subplasmalemmal vesicles into the membrane. If such vesicles contained calcium channels, then a PKC-induced stimulation of fusion of vesicles into the plasma membrane could lead to the simultaneous expansion of the area of the growth cone and the emergence of the new sites of calcium entry. There is some circumstantial evidence that the appearance of the24 pS channel at the soma in response toactivators of PKC may also result from the insertion of new channels rather than the activation of previously “silent” channels preexisting in the plasma membrane (Strong et al., 1987; Conn and Kaczmarek, 1989). The latter possibility cannot however be ruled out for either the channels at the soma or those in the growth cones. Nevertheless, targeting of lipid- and ion channel-containing vesicles to the plasma membrane could repre-
sent a novel and efficient mechanism for coupling ion channel regulation to changes in cell structure. Structural changes in the morphology of neurotransmitter release sites in response to external stimulation have been reported for other Aplysia neurons. For example, the sensory neurons that mediate the gill and siphon withdrawal reflexes in Aplysia undergo marked structural changes within 48 hr of the application of strong sensitizing stimuli (Bailey and Chen, 1989). The synaptic modifications that occur during sensitization of these reflexes include an increase in size and vesicle content of active zones, as well as an increase in the total number of synaptic varicosities. In addition, in other invertebrate and vertebrate nervous systems, stimuli that evokechanges in the behavior of an animal have been shown to involve structural remodeling of synapses. Other workers have shown that electrical stimulation of isolated invertebrate neurons influences the structure of their growth cones, probably through regulation of intracellular calcium levels (Cohan and Kater, 1986; see also Jones et al., 1987; McCobb and Kater, 1986, Sot. Neurosci., abstract). Our results with isolated peptidergic bag cell neurons have shown that stimulation of second messenger systems may produce very rapid morphological changes, including the redistribution of secretory granules and calcium channels, that occur over a time course of 5-15 min. Because the effects we have reported were studied in isolated bag cell neurons, it is not yet known when and under what conditions similar changes in the structure and function of bag cell neurites occur in the intact nervous system. However, stimulation of afterdischarges in these cells causes an increase in phosphoinositide turnover, (Fink et al., 1988), cyclic AMP levels (Kaczmarek et al., 1978), and the activity of the cyclic AMP-dependent protein kinase system (Jennings et al., 1982)and leads to a significant potentiation of peptide release by individual action potentials (Loechner et al., 1990).Thus it is possible that a restructuring of release sites in the intact nervous system contributes to the potentiation of secretion. Direct imaging of neurites in situ will be required possibility. An alternative is that the growth endings induced by these second messengers utes to a progressive increase in the size neurons and their neurites, which occurs reproductive season of Aplysia. Experimental
to test this of neurite contribof bag cell during the
Procedures
Cell Culture Isolated bag ceil neurons were prepared in primary culture as previously described (Kaczmarek et al., 1979). The abdominal ganglion was removed from the animal and incubated for 18 hr in a neutral protease solution (Dispase, 40 mg/3 ml) at room temperature. Bag cell clusters were then dissected from their surrounding connective tissue, and cells were plated using a Pasteur pipetteintoculturedishescontainingartificial seawater. The cells were plated on #I glass microscope coverslips coated with polylysine and allowed to extend processes for at least 12 16 hr. Generally growth cones had reached a stable state or else one of very slow outgrowth at the time of examination.
Video-Enhanced Microscopy of Bag Cell Neurons Bag cell neuronswereviewed usingazeiss IM-35 inverted microscope fitted with Nomarski differential interference contrast optics. Zeiss objectives (16/0.5 neofluar and 40/0.9 neofluar) and an oil immersion condenser (numerical aperture 0.63) were used. Video-enhanced differential interference contrast images of bag cell neurons growing on #I glass coverslips were obtained with a silicon-intensified target video camera (Hamamatsu C2400) coupled to an Imaging Technologies Inc. 151 image processor whose frame grabber board was resident in a Compaq 386/20e host computer. High resolution, high contrast images were obtained by digital averaging of the incoming video signal. Hard copy video micrographs were made using a video printer (Sony UP 5000), and each example shown is a digitized average of 256 frames acquired over a 4 s time period.
Calcium Imaging of Growth Cones Fura-2-free acid (Crynkiewicz et al., 1985) was injected Into cell somatatoconcentrationsof50-IOOuM bypressureejectionfrom intracellular microelectrodes (electrical resistance >50 MS1 when filled with 3 M KCI). Injections required 20-40 s, and cells were allowed to equilibrate for at least 30 min after the injections. For electrical recording and stimulation, the neurons were repenetrated using microelectrodes filled with 1 M KCI. Fura- fluorescence increases maximallywith calcium binding at 340 nm excitation and decreases with calcium at 380 nm excitation. Calcium determinations were made by ratioing two images of fura- fluorescence using 340 and 380 nm excitation. Details of this method as well as the imaging apparatus have previously been described (Connor, 1986; Fink et al., 1988). Acquisition time for one frame pair was approximately 800 ms. Calcium influx was induced in bag cell neurons by firing the cells at three action potentials per second by injecting intracellular current pulses (0.5-I nA, 75 ms). Corrections for substrata fluorescence and camera dark current were carried out as follows. Proper focus of neurites was first determined visually under UV excitation, and then the field of excitation (-250 urn diameter) was moved to a nearby cell-free location, and exposures of the proper duration were taken at both 340 and 380 nm excitation. Since cells were filled individuallybyinjectingfura-2,itwasneverdifficulttofindsuitable,clean areas on the glass substrata, as is often a problem when ester loading of the indicator is used. These images were stored in computer RAM and subtracted from all subsequent cell images. With this correction, the mean background signal (measured in cell-free areas) was less than 1 arbitrary unit (AU), and the standard deviation of this background was ~0.5 AU. The latter number istheuseful noisefigureforthesystem,sincethecooled CCD camera is a linear transducer, and there is no effect on gain at the different mean light levels encountered here. For comparison, thecorrected 340and 380 nm fluorescence readings at the upper right hand tip of the growth cone of Figure 1 before TPA treatment (upper pictures) were 7 and 10.7 AU, respectively, at rest (values measured in a 4 x 4 pixel box covering ~2 pm). Mean backgrounds were 0.3 and 0.2 AU. During the electrical stimulation shown, these values changed to 10 and 2.8 AU. After TPA treatment (Figure 1, lower records), resting 340 and 380 nm values were 7.8 and 15.5 AU at the upper right-hand edge and went to 13.1 and 3.1 during the maximum electrical response. The analysis box lay within the “red” response area shown in Figure 1. Display data were masked such that areas where the 380 nm signal fell below 2 AU during the peak response did not appear in any of the ratio images. In making the final ratio images displayed in Figure 1, the individual 340 and 380 nm images were filtered (low pass, recursive, with nearest neighbor pixels weighted by 0.25) before division, enabling better use of color display by reducing extreme ratio pixel values during peak mlsponse. Data for tabulation were always checked for proper background correction by measuring residual signals in cell-free areas of images near the structures of interest. When filling the cells, fura- was injected until the autofluorescence of the soma comprised no more than about 5% of total fluorescence with 380 nm excitation.
Calcium 889
Channels
in Growth
Cones
We thank Dr. Laura Fink for her participation in the early stages of this work. This research was supported in pat-t by a grant from the National Institutes of Health to L. K. K. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisemeni’ in accordance with 18 USC Section 1734 solely to indicate this fact. Received
November
11, 1991;
revised
February
11, 1992.
References Arch, S. (1976). Neutoendoctine Aplysia califotnica. Am. Zool.
regulation 76, 167-175.
Bailey, C. H., and Chen, M. (1989). changes at identified sensory neuron sensitization in Aplysia. J. Neurosci. Bixby, J. L. (1989). Protein tion of neutite outgrowth.
of
egg
laying
in
Time course of structural synapses during long-term 9, 1774-1780.
kinase C is involved in laminin Neuron 3, 287-297.
stimula-
Bridgeman, 8. W., Kachat, B., and Reese, T. S. (1986). The sttuctuteof cytoplasm inditectlyftozencultutedcells. ii.Cytoplasmic domains associated with otganelle movements. J. Cell Biol. 702, 1510-1521. Btuehl, C. L., and Betty, R. W. (1985). Regulation of synthesis of the neutosectetory egg-laying hormone of Aplysia: antagonistic roles of calcium and cyclic adenosine 3’:5’-monophosphate. J. Neutosci. 5, 1233-1238. Cohan, C. S., and Katet, S. B. (1986). Suppression gation and growth cone motility by electrical 232, 1638-1640.
of neutiteelonactivity. Science
Cohan, C. S., Connot, J. A., and Katet, S. B. (1987). and chemically mediated increases in intracellular neutonal growth cones. J. Neurosci. 7, 3588-3599.
Electrically calcium in
Conn, P. J., and Kaczmatek, L. K. (1989). A model for the study of the molecular mechanisms involved in the control of prolonged animal behaviors. Mol. Neutobiol. 3, 237-273. Conn, P. J., Strong, J. A., Azhdetian, E. M., Naitn, A. C., Gteengard, P., and Kaczmatek, L. K. (1988a). Protein kinase inhibitors selectively block photbol-ester or fotskolin-induced changes in excitability of Aplysia neurons. J. Neutosci. 9, 473479. Conn, P. J., Strong, J. A., and Kaczmatek, of protein kinase C pteventenhancementof action potentials in peptidetgic neurons 9, 480.-487.
L. K. (1988b). Inhibitors calcium current and of Aplysia. J. Neutosci.
Connot, J. A. (1986). Digital imaging of free calcium changes and of spatial gradients in growing processes in single, mammalian central nervous system cells. Ptoc. Natl. Acad. Sci. USA83,61796183. DeRiemet, S. A., Gteengatd, P., and Kaczmatek, Calcium/phosphatidylsetine/diacylglycetol-dependen~ phosphotylation in the Aplysia nervous system. 2672-2676.
L. K. (1985a). protein J. Neutosci. 5,
DeRiemet, S. A., Strong, 1. A., Albert, K. A., Cteengatd, P., and Kaczmatek, L. K. (1985b). Enhancement of calcium current in Aplysia neurons by photbol ester and protein kinase C. Nature 373, 313-316. Dudek, F. E., and Blankenship, J. E. (1976). Neutoendocrine (bag) cells of Aplysia: spike blockade and a mechanism for potentiation. Science 792, 1009-1010. Fink, IL. A., Connot, J. A., and Kaczmatek, L. K. (1988). lnositol ttisphosphate releases intracellulatly stored calcium and modulates ion channels in molluscan neurons. J. Neutosci. 8, 2544 2555. Forschet, P., and Smith, S. J. (1988). Actions of cytochalasins on the organization of actin filaments and microtubules in a neutonal growth cone. J. Cell Biol. 707, 1505-1516.
Fotschet, P., Kaczmatek, L. K., Buchannan, I. A., and Smith, S. J. (1987). Cyclic AMP induces changes in the distribution and ttansport of otganelles within growth cones of Aplysia bag cell neurons. J. Neutosci. 7, 3600-3611. Frazier, W. T., Kandel, E. R., Kupfetmann, I., Waziti, R., and Coggeshall, E. (1967). Morphological and functional properties of identified neurons in the abdominal ganglion of Aplysia californica. J. Neutophysiol. 30, 1288-1351. Grynkiewicz, G., Poenie, M., and Tsien, R. Y. (1985).A ation of calcium indicators with greatly improved properties. J. Biol. Chem. 260, 3440-3448. Jennings, K. R., Kaczmatek, and Sttumwasset, F. (1982). tetdischatge in peptidetgic 158-168.
new genetfluorescent
L. K., Hewick, R. M., Dreyer, W. I., Protein phosphotylation during afneurons of Aplysia. J. Neutosci. 2,
Jones, P. G., Rosset, S. J., and Bulloch, A. (2. M. (1987). Glutamate suppression of feeding and the underlying output of affectot neurons in Helisoma. Brain Res. 437, 56-68. Kaczmatek, a prolonged J. Neutosci.
L. K., and Kauet, refractory period 3, 2230-2239.
J. A. (1983). Calcium entry causes in peptidetgic neurons of Aplysia.
Kaczmatek, L. K., and Strumwasset, F. (1984). A voltage-clamp analysis of currents underlying cyclic-AMP induced membrane modulation in isolated peptidetgic neurons of Aplysia. J. Neutophysiol. 52, 340-349. Kaczmatek, L. K., Jennings, K., and Sttumwasset, F. (1978). Neurotransmitter modulation, phosphodiestet inhibitor effects, and cyclic AMP correlates of afterdischarge in peptidetgic neurites. Ptoc. Natl. Acad. Sci. USA 75, 5200-5204. Kaczmatek, L. K., Finbow, M., Revel, J. P., and Sttumwasset, F. (1979). The morphology and coupling of Aplysia bag cells within the abdominal ganglion and in cell culture. J. Neutobiol. 70,535550. Kaczmarek, L. K., Jennings, K. R., Sttumwasset, Walter, U., Wilson, F. D., and Gteengatd, P. (1980). of the catalytic subunit of cyclic-AMP-dependent enhances calcium action potentials of bag cell culture. Ptoc. Natl. Acad. Sci. USA 77, 7487-7491.
F., Naitn, A. C., Mictoinjecton protein kinase neurons in cell
Kaczmatek, L. K., Jennings, K., and Sttumwasset, F. (1982). An early sodium and a late calcium phase in the afterdischarge of peptide secreting neurons of Aplysia. Brain Res. 283, 105-115. Kaczmarek, L. K., Strong, J. A., and DeRiemet, S. A. (1985). Biochemical mechanisms that regulate potassium and calcium curtents in peptidetgic neurons. In Neurosectetion and the Biology of Neutopeptides, H. Kobayashi, H. A. Bern, and A. Utano, eds. (Berlin: Springer-Vetlag), pp. 80-87. Kandel, learning.
E. R., and Schwartz, J. H. (1982). Science 278, 433-443.
Molecular
Katet, S. B., Mattson, M. P., Cohan, C. S., and (1988). Calcium regulation of the neutonal growth Neutosci. 77, 315-321.
biology
of
Connot, J. A. cone. Trends
Lipscombe, D., Madison, D. V., Poenie, M., Reutet, H.,andTsien, R. W. (1988). Spatial distribution of calcium channels and cytosolic calcium transients in growth cones and cell bodies of sympathetic neurons. Ptoc. Natl. Acad. Sci. USA 85, 2398-2402. Loechnet, K. J., Azhdetian, E. M., Dryer, R., and Kaczmatek, (1990). Progressive potentiation of peptide release during tonal discharge. J. Neutophysiol. 63, 738-744.
L. K. a neu-
Strong, J. A., Fox, A. P., Tsien, R. W., and Kaczmatek, L. K. (1987). Stimulation of protein kinase C recruits covert calcium channels in Aplysia bag cell neurons. Nature 325, 714-717.
Vol. 8, 883-889,
May, 1992, Copyright
0 1992 by Cell Press
Recruitment of Ca2+ Channels by Protein Kinase C during Rapid Formation of Putative Neuropeptide Release Sites in Isolated Aplysia Neurons Ronald J. Knox,* Elizabeth A. Quattrocki,* John A. Connor,+ and Leonard K. Kaczmarek** *Department of Pharmacology *Department of Molecular and Cellular Physiology Yale University School of Medicine New Haven, Connecticut 06510 +Roche Institute of Molecular Biology Roche Research Center Nutley, New Jersey 07110
Summary Activation of protein kinase C (PKC) in Aplysia bag cell neurons causes the recruitment of voltage-dependent calcium channels. Using imaging techniques on isolated cells, we have now found that an activator of PKC, 12-Otetradecanoyl-phorbol-13-acetate (TPA), promotes the rapid appearance of new sites of calcium influx associated with a change in the morphology of neurite endings. In untreated cells, calcium influx triggered by action potentials occurs along neurites and in the central region of growth cones, but does not usually occur at the leading edge of lamellipodia. TPA produces extension of the lamellipodium, and action potentials now trigger calcium influx at the distal edge of the newly extended endings. Cotreatment with TPA and a cyclic AMP analog promotes movement of secretory organelles toward the new sitesof calcium influx. Our results suggest that these second messenger systems promote the rapid formation of morphological structures that contributeto thepotentiation of peptide release. Introduction Brief electrical stimulation of an afferent input to the bag cell neurons of Aplysia triggers a prolonged discharge (-30 min) of action potentials during which there occurs a stimulation of phosphoinositide turnover (Fink et al., 1988) and an increase in intracellular cyclic AMP levels (Kaczmarek et al., 1978; Kaczmarek and Kauer, 1983). This discharge leads to the release of several neuroactive peptides, including egg-laying hormone, which in vivo, triggers a sequence of reproductive behaviors (Conn and Kaczmarek, 1989). During the first IO-15 min of the discharge there is a progressive potentiation of the ability of action potentials to evoke the secretion of egg-laying hormone from the neurites of these cells (Loechner et al., 1990). The potentiation of peptide release during a discharge may result, in part, from the stimulation of protein kinase C (PKC), which leads to an enhancement of calcium influx. The height and width of action potentials of bag cell neurons are determined to a large extent by the amplitude of the voltage-dependent calcium currents (Kaczmarek et al., 1982). Over the first few minutes of the discharge there is a pro-
gressive increase in the amplitude of action potentials, which is blocked by inhibitors of PKC (Conn et al., 1988b). Microinjection of PKC or application of activators of this enzyme to isolated neurons produces an enhancement of action potentials that resembles the increase occurring during an evoked discharge (DeRiemer et al., 1985b). Patch-clamp experiments have shown that this occurs through the unmasking of a new species of voltage-dependent calcium channel on the plasma membrane. In untreated bag cell neurons, only calcium channelswith a unitary conductance of 12 pS are detected (Strong et al., 1987). In cells exposed to activators of PKC, however, a novel 24 pS calcium channel can also be recorded (Strong et al., 1987), and the net voltage-dependent calcium current is increased 2-fold or more (DeRiemer et al., 1985b). Stimulation of a discharge also evokes an increase in cyclic AMP levels in bag cell neurons (Kaczmarek et al., 1978). Although microinjection of the catalytic subunit of the cyclic AMP-dependent protein kinase, or application of cyclic AMP analogs, has been shown to produce spike broadening in isolated bag cell neurons, voltage-clamp analysis indicates that these act primarily on potassium conductances. No effect of cyclic AMP has been detected on the major voltagedependent calcium current in bag cell neurons. An elevation of cyclic AMP levels, however, also modifies several other properties of bag cell neurons that may contribute to potentiation of peptide secretion. These include increased synthesis of the peptide precursor (Bruehl and Berry, 1985) and enhancement of movement of secretory granules along neurites and into growth cones (Forscher et al., 1987; Azhderian et al., 1991, Sot. Neurosci., abstract). We have now used video-enhanced microscopy, coupled with digital imaging of intracellular calcium levels, to examine the actions of 12-O-tetradecanoylphorbol-13-acetate (TPA), an activator of PKC, and of a cyclic AMP analog on morphology and on distribution of calcium within the neuritesof bag cell neurons, where releaseof peptides such as egg-laying hormone is believed to occur (Dudek and Blankenship, 1976; Arch, 1976; Frazier et al., 1967; Kaczmarek et al., 1978). Results Phorbol Ester Treatment Induces Extension of Lamellipodia Bag cell neurons were isolated from the abdominal ganglion of Aplysia californica and plated on a polylysine substrate. Figure 1 and Figure 2a show the typical appearance of their neurites when viewed through fluorescence and video-enhanced Nomarski optics, respectively. After several hours or more, their growth cones are very large (30-100 Grn) and contain a central cytoplasmic region that is rich in organelles and mi-
--
22
SALINE
i
5.5
-
ka2+lPM
20pm
20 nM TPA Figure
1. Phorbol
Ester Alters
the Spatial
Organization
of Calcium
Entry
into
Neurite
Endings
Left panels (1 and 6): fluorescent images (380 nm excitation) of a bag cell neuron growth cone before and after exposure to TPA. Intensities are shown on a modified log scale and are normalized to the maximum intensity for each picture. After TPA treatment the growth cone becomes 2-3 times thicker in thecentral portion. Color pictures 2-5 show maps of internal calcium concentrations before and during the firing of a 3 Hz action potential train as indicated in the panels. Color pictures 7-10 show calcium concentrations for an identical firing pattern after exposure to TPA.
a
Control
(- 40 mins)
Control (0 mins)
20nM TPA (+ 40 mins)
al 40 d 5 E 4 30 5 L? 20
a .E
ko 5 5 s
0
Control
20 nM TPA 100,uM H-7 +TPA
the Mor-
(a) Video-enhanced differential interference contrast images of a bag cell neuron growth cone. In a 40 min period prior to TPA application there is little change in either the size or appearance of the growth cone. TPA treatment produces a marked extension of the lamellipodial membrane at the leading edge of the growth cone, resulting in a significant increase in the areas of lamellae. This morphological change could be prevented by preincubation of the neurons with 100 uM H-7, a selective inhibitor of PKC in bag cell neurons. Scale bar, 20 urn. (b) Summary of the action of TPA on 14 growth cones. Measurement of the area of lamellae was determined as described previously (Forscher et al., 1987).
mm
b
Figure 2. Phorbol Ester Changes phology of Growth Cones
100uM H-7
Cahum 085
Channels
-O--t + a1
in Growth
Cones
BEFORE STIMULATION DURING SPIKE TRAIN RECOVERY 4 B
Comparison of [Cal’], changes durrng firing of action potentials in a growth cone treated first with 200 nM pCPTcAMP and then with the cyclic AMP analog and 20 nM TPA. (A) Maximum [Cal’], changes during firing before either treatment. Locations of analysis boxes for the graphs are shown in the fluorescence pictures (380 nm excitation) immediately below. Note that changes are smallest in the leading edge of the growth cone. (B) [Cal+], changes after 20 min exposure to pCPTcAMP. Both the levels and concentration profiles are similar to those of the untreated neuron. (C) Following treatment with both kinase activators, [Ca2+], changes are much larger during firing, and the greatest change is now at the leading edge of the growth cone. To maximize clarity in black and white reproductions, slightly different scalings were used for each fluorescent picture.
crotubules (Forscher et al., 1987; Forscher and Smith, 1988). As in other cell types, this region is surrounded by a flat actin-rich lamellipodium that is virtually organelle free (Bridgeman et al., 1986). Under control conditions the percentage of total growth cone area occupied by organelles was 40% + 3% (SEM, N = 36; Figure 48). Growth and movement of intracellular organelles were monitored for at least 40 min before addition of activators of protein kinases. In general, relatively little growth occurred over this time period (Figures 2a and 2b). However, within minutes of bath application of IO-20 nM TPA, which activates PKC in bag cell neurons (DeRiemer et al., 1985a), there was a rapid extension of lamellae that was evident within several minutes. Although this morphological change could be observed after only a few minutes exposure to TPA, the effect peaked by 40 min. This resulted in a 23% f 6% increase (N = 14; Figure 2b) in the area of the growth cones. TPA treatment did not alter the distribution of organelles within the central domain of the growth cone (Figure 2a; see also Figure 4B). To test the specificity of action of TPA, cells were pretreated with 100 uM H-7, a selective inhibitor of PKC in these cells (Conn et al., 1988a, 1988b), for 15 min before addition of the phorbol ester. H-7 was found to block the actions of TPA on the neurites of the cells (N = 7; Figure 2b). In addition, the inactive phorbol ester 4a-phorbol 12,13-didecanoate did not produce extension of neurite endings (N = 3). Phorbol Ester Induces Recruitment of Calcium Channels To examine the distribution of voltage-dependent
cal-
cium channels at the tips of neurites, cells were injected with the calcium indicator dye fura-2, and digitized images were made of the ratio of fluorescence at 340 and 380 nm excitation wavelengths. This ratio gives a direct measure of the intracellular calcium ion concentration [Ca?‘],. Trains of action potentials were evoked by depolarizing current pulses applied to the somata of the neurons. The major sites of calcium entry through voltage-dependent calcium channels were visualized as sites where localized elevation of calcium occurred in response to the first few action potentials during a train. [Ca*-1, changes were absolutely dependent upon the presence of calcium in the external bathing medium, and action potentials fired in 0 calcium, 1 mM EGTA saline gave no [Ca*+], increases. During a train of action potentials (3 Hz, 5-15 s) in untreated cells in normal calcium-containing media, calcium levels generally rose uniformly along the length of the neurite, although “hot spots” of calcium entry could sometimes be detected along the central part of the neurites. The leading edges of the growth cones generally showed lower levels of calcium than the more central regions, making it likely that the influx occurred primarily in the central region and then spread by diffusion toward the lamellipodium (Figure 1, panels 2-5). This suggests that such lamellae had a low density of functional voltageactivated calcium channels at the time of observation. Figure 1 (panels 7-10) shows that, after only a few minutesexposuretoTPA,thepatternofcalciuminflux during action potentials changed such that the first detectable sign of calcium entry was now localized to a ring that corresponded anatomically to the leading
NC?UKl” 886
Control
TPA + pCPTcAMP
80 1
Control Figure4.
Invasion
TPA of Growth
Cones
TPA’pCPTcAMP by Secretory
Organelles
-
Control in Response
to a Cyclic
TPA AMP
Analog
and
TPA pCPTcAMP
a Phorbol
Ester
(A) In the presence of 20 nM TPA, pCPTcAMP induces movement of intracellular organelles into the extended lamellipodium. The directed movement of organelles along microtubules is detected as lines extending from the central region of the growth cone into the periphery. Scale bar, 20 urn. (ET) Histograms showing the effects of 20 nM TPA alone and of 20 nM TPA plus 200 PM pCPTcAMP on the area of bag cell lamellae and on the area of the growth cones occupied by organelles (N = 10 neurons).
edge of the newly extended lamellipodial membrane. The greatest initial increase in [Cal+], now occurred in the lamellipodium. The slight decrease in the overall amplitude of the [Ca2’], change within the central region after TPA treatment probably reflects the greatly expanded volume of the growth cone of this cell. Qualitatively similar results were seen in 22 neurons, with 7 neurons showing no significant changes in [Ca2’], after phorbol ester application. In addition, after phorbol estertreatment, increased calcium entry was also detected at the soma and at locations on the central neurite branches that had previously appeared dormant (note also the “bud”at the lower left of the neurite of Figure 1). Control experiments were carried out with an inactive phorbol ester, 4a-phorbol 12,13-didecanoate, which does not activate PKC. No changes in the amount or distribution of calcium influx or in the morphology of the neurite endings were observed in response to this agent (N = 4). Also there was no change in the distribution of calcium influx in cells that failed to respond to TPA. In particular, this occurred in cells with rapidly growing neurites. in these cases we often
saw large [CaL’], changes at the lamellae during firing prior to addition of phorbol ester (see also Cohan et al., 1987), and the application of phorbol ester produced little further enhancement. In these circumstances it is possible that PKC activity is constitutively activated by factors such as substrate contact (Bixby, 1989). The simplest interpretation of the appearance of new sites of calcium elevation is that, as at the soma (Strong et al., 1987), activation of PKC leads to the recruitment of inactive or “covert” calcium channels in the newly extended membrane at the margin of growth cones and other sites on the neurites. We found no evidence for caffeine-releasable calcium stores in the bag cell neurons, as previously reported for amphibian neurons (Lipscombe et al., 1988). However, we cannot entirely eliminate the possibility that other additional factors, such as changes in the spatial distribution of calcium buffering, also contribute in part to the observed changes. Effects of a Cyclic The calcium action
AMP Analog potentials
on Calcium Influx of bag cell neurons
are
Calcium 007
Channels
in Growth
Cones
also enhanced by cyclic AMP analogs or injection of the catalytic subunit of the cyclic AMP-dependent protein kinase (Kaczmarek et al., 1980). Cyclic AMP analogs have not been found to enhance calcium current (Kaczmarek and Strumwasser, 1984), but rather to decrease voltage-dependent potassium currents. We therefore examined the actions of the cyclic AMP analog pCPTcAMP (8+-chlorophenylthio) cyclic AMP, 200 PM) on calcium influx and morphology of the growth cones. In contrast to TPA, treatment with pCPTcAMP alone did not cause elongation of neurite endings (data not shown; see also Forscher et al., 1987). However, during a train of action potentials, a greater rateof [Ca*+], elevation often occurred (Figures 3A and 3B), presumably as a result of broader action potentials, but the spatial profile of calcium entry was the same.
Cotreatment of Bag Cell Neurons with Phorbol Ester and Cyclic AMP Analog Although cyclic AMP analogs do not unmask new calcium channels at the growth cone, they do produce a characteristic change in the appearance of the neurite endings. Previous work (Forscher et al., 1987) has shown that elevation of cyclic AMP levels transforms the lamellipodium of bag cell neurons into a thickened terminal engorged with organelles that are predominantly the peptide-containing neurosecretory granules. We have now found that this effect also occurs in newly formed lamellae that were induced by TPA. Cotreatment of cells with pCPTcAMP and TPA produced extension of the lamellae followed by directed movement of organelles from the central region to the distal edge of the growth cone (Figure 4B). Imaging of calcium levels in such cotreated cells shows that, as in cells treated with TPA alone, novel sites of calcium entry appear at the distal edge (see Figure 3C). The graphs in this figure show the changes in calcium levels during a train of action potentials before and after treatment with pCPTcAMP and TPA. The fluorescent pictures below, which do not represent calcium levels, show the locations at which the calcium levels were measured and were made while the growth of the neurite was stable. This example was chosen, in part, because there was relatively little morphologic change induced by these treatments to confound the interpretation of changes in calcium levels by changes in surface to volume ratio. In the presence of TPA, the distal tip of the neurite became the dominant site of calcium entry. Similar effects were observed in 6 neurons.
Discussion Our results indicate that in the bag cell neurons, coactivation of PKC and the cyclic AMP-dependent protein kinase produces a rapid change in the morphology at the distal tips of bag cell neurites and changes in the pattern of calcium entry. In particular, new sites
of calcium entry are formed and are coupled to the extension of growth cone membrane and to the movement of secretory organelles toward sites subjacent to the newly activated calcium entry sites. The simplest interpretation of these data is that, as at the soma (Strong et al., 1987), stimulation of PKC causes the recruitment of a new species of voltage-dependent calcium channel at the tips of bag cell neurites. It is highly unlikely that the changes in calcium levels during action potentials that we have described could result purely from changes in growth cone morphology, without the recruitment of new calcium channels. In particular, because the growth cones become thicker after treatment with TPA, the changes in the surface to volume ratio should reduce rather than increase calcium levels in response to a given influx of calcium. In the experiments of Strong et al. (1987), patch-clamp recordings on the somata of untreated bag cell neurons revealed only one species of calcium channel, which has a unitary conductance of 12 pS when measured with barium as the permeant ion. Activation of PKC by phorbol esters or synthetic diacylglycerols brought about the appearance of an additional novel calcium channel with a unitary conductanceof24pS.Similarunmaskingofthis24pSchannel in the extended famellipodium could explain the changes in calcium distribution that we have observed. Because of the difficulty of patch clamping the very thin famellipodial membrane, however, we cannot exclude the possibilitythat thechannelsat the terminals differ from either of the two types detected on the soma. It is also possible that, in addition to new sites of calcium influx, changes in parameters such as local calcium buffering may partly contribute to the observed changes in calcium distribution during trains of action potentials. Our results also show that exposure of bag cell neurons to TPA produces rapid extension of lamellipodial membrane, presumably through a dynamic change in the growth cone cytoskeleton. It is possible that the extension of the membrane is linked mechanistically to the appearance of the new calcium entry sites. For example, extension of membrane in other cell types is thought to result from fusion and incorporation of subplasmalemmal vesicles into the membrane. If such vesicles contained calcium channels, then a PKC-induced stimulation of fusion of vesicles into the plasma membrane could lead to the simultaneous expansion of the area of the growth cone and the emergence of the new sites of calcium entry. There is some circumstantial evidence that the appearance of the24 pS channel at the soma in response toactivators of PKC may also result from the insertion of new channels rather than the activation of previously “silent” channels preexisting in the plasma membrane (Strong et al., 1987; Conn and Kaczmarek, 1989). The latter possibility cannot however be ruled out for either the channels at the soma or those in the growth cones. Nevertheless, targeting of lipid- and ion channel-containing vesicles to the plasma membrane could repre-
sent a novel and efficient mechanism for coupling ion channel regulation to changes in cell structure. Structural changes in the morphology of neurotransmitter release sites in response to external stimulation have been reported for other Aplysia neurons. For example, the sensory neurons that mediate the gill and siphon withdrawal reflexes in Aplysia undergo marked structural changes within 48 hr of the application of strong sensitizing stimuli (Bailey and Chen, 1989). The synaptic modifications that occur during sensitization of these reflexes include an increase in size and vesicle content of active zones, as well as an increase in the total number of synaptic varicosities. In addition, in other invertebrate and vertebrate nervous systems, stimuli that evokechanges in the behavior of an animal have been shown to involve structural remodeling of synapses. Other workers have shown that electrical stimulation of isolated invertebrate neurons influences the structure of their growth cones, probably through regulation of intracellular calcium levels (Cohan and Kater, 1986; see also Jones et al., 1987; McCobb and Kater, 1986, Sot. Neurosci., abstract). Our results with isolated peptidergic bag cell neurons have shown that stimulation of second messenger systems may produce very rapid morphological changes, including the redistribution of secretory granules and calcium channels, that occur over a time course of 5-15 min. Because the effects we have reported were studied in isolated bag cell neurons, it is not yet known when and under what conditions similar changes in the structure and function of bag cell neurites occur in the intact nervous system. However, stimulation of afterdischarges in these cells causes an increase in phosphoinositide turnover, (Fink et al., 1988), cyclic AMP levels (Kaczmarek et al., 1978), and the activity of the cyclic AMP-dependent protein kinase system (Jennings et al., 1982)and leads to a significant potentiation of peptide release by individual action potentials (Loechner et al., 1990).Thus it is possible that a restructuring of release sites in the intact nervous system contributes to the potentiation of secretion. Direct imaging of neurites in situ will be required possibility. An alternative is that the growth endings induced by these second messengers utes to a progressive increase in the size neurons and their neurites, which occurs reproductive season of Aplysia. Experimental
to test this of neurite contribof bag cell during the
Procedures
Cell Culture Isolated bag ceil neurons were prepared in primary culture as previously described (Kaczmarek et al., 1979). The abdominal ganglion was removed from the animal and incubated for 18 hr in a neutral protease solution (Dispase, 40 mg/3 ml) at room temperature. Bag cell clusters were then dissected from their surrounding connective tissue, and cells were plated using a Pasteur pipetteintoculturedishescontainingartificial seawater. The cells were plated on #I glass microscope coverslips coated with polylysine and allowed to extend processes for at least 12 16 hr. Generally growth cones had reached a stable state or else one of very slow outgrowth at the time of examination.
Video-Enhanced Microscopy of Bag Cell Neurons Bag cell neuronswereviewed usingazeiss IM-35 inverted microscope fitted with Nomarski differential interference contrast optics. Zeiss objectives (16/0.5 neofluar and 40/0.9 neofluar) and an oil immersion condenser (numerical aperture 0.63) were used. Video-enhanced differential interference contrast images of bag cell neurons growing on #I glass coverslips were obtained with a silicon-intensified target video camera (Hamamatsu C2400) coupled to an Imaging Technologies Inc. 151 image processor whose frame grabber board was resident in a Compaq 386/20e host computer. High resolution, high contrast images were obtained by digital averaging of the incoming video signal. Hard copy video micrographs were made using a video printer (Sony UP 5000), and each example shown is a digitized average of 256 frames acquired over a 4 s time period.
Calcium Imaging of Growth Cones Fura-2-free acid (Crynkiewicz et al., 1985) was injected Into cell somatatoconcentrationsof50-IOOuM bypressureejectionfrom intracellular microelectrodes (electrical resistance >50 MS1 when filled with 3 M KCI). Injections required 20-40 s, and cells were allowed to equilibrate for at least 30 min after the injections. For electrical recording and stimulation, the neurons were repenetrated using microelectrodes filled with 1 M KCI. Fura- fluorescence increases maximallywith calcium binding at 340 nm excitation and decreases with calcium at 380 nm excitation. Calcium determinations were made by ratioing two images of fura- fluorescence using 340 and 380 nm excitation. Details of this method as well as the imaging apparatus have previously been described (Connor, 1986; Fink et al., 1988). Acquisition time for one frame pair was approximately 800 ms. Calcium influx was induced in bag cell neurons by firing the cells at three action potentials per second by injecting intracellular current pulses (0.5-I nA, 75 ms). Corrections for substrata fluorescence and camera dark current were carried out as follows. Proper focus of neurites was first determined visually under UV excitation, and then the field of excitation (-250 urn diameter) was moved to a nearby cell-free location, and exposures of the proper duration were taken at both 340 and 380 nm excitation. Since cells were filled individuallybyinjectingfura-2,itwasneverdifficulttofindsuitable,clean areas on the glass substrata, as is often a problem when ester loading of the indicator is used. These images were stored in computer RAM and subtracted from all subsequent cell images. With this correction, the mean background signal (measured in cell-free areas) was less than 1 arbitrary unit (AU), and the standard deviation of this background was ~0.5 AU. The latter number istheuseful noisefigureforthesystem,sincethecooled CCD camera is a linear transducer, and there is no effect on gain at the different mean light levels encountered here. For comparison, thecorrected 340and 380 nm fluorescence readings at the upper right hand tip of the growth cone of Figure 1 before TPA treatment (upper pictures) were 7 and 10.7 AU, respectively, at rest (values measured in a 4 x 4 pixel box covering ~2 pm). Mean backgrounds were 0.3 and 0.2 AU. During the electrical stimulation shown, these values changed to 10 and 2.8 AU. After TPA treatment (Figure 1, lower records), resting 340 and 380 nm values were 7.8 and 15.5 AU at the upper right-hand edge and went to 13.1 and 3.1 during the maximum electrical response. The analysis box lay within the “red” response area shown in Figure 1. Display data were masked such that areas where the 380 nm signal fell below 2 AU during the peak response did not appear in any of the ratio images. In making the final ratio images displayed in Figure 1, the individual 340 and 380 nm images were filtered (low pass, recursive, with nearest neighbor pixels weighted by 0.25) before division, enabling better use of color display by reducing extreme ratio pixel values during peak mlsponse. Data for tabulation were always checked for proper background correction by measuring residual signals in cell-free areas of images near the structures of interest. When filling the cells, fura- was injected until the autofluorescence of the soma comprised no more than about 5% of total fluorescence with 380 nm excitation.
Calcium 889
Channels
in Growth
Cones
We thank Dr. Laura Fink for her participation in the early stages of this work. This research was supported in pat-t by a grant from the National Institutes of Health to L. K. K. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisemeni’ in accordance with 18 USC Section 1734 solely to indicate this fact. Received
November
11, 1991;
revised
February
11, 1992.
References Arch, S. (1976). Neutoendoctine Aplysia califotnica. Am. Zool.
regulation 76, 167-175.
Bailey, C. H., and Chen, M. (1989). changes at identified sensory neuron sensitization in Aplysia. J. Neurosci. Bixby, J. L. (1989). Protein tion of neutite outgrowth.
of
egg
laying
in
Time course of structural synapses during long-term 9, 1774-1780.
kinase C is involved in laminin Neuron 3, 287-297.
stimula-
Bridgeman, 8. W., Kachat, B., and Reese, T. S. (1986). The sttuctuteof cytoplasm inditectlyftozencultutedcells. ii.Cytoplasmic domains associated with otganelle movements. J. Cell Biol. 702, 1510-1521. Btuehl, C. L., and Betty, R. W. (1985). Regulation of synthesis of the neutosectetory egg-laying hormone of Aplysia: antagonistic roles of calcium and cyclic adenosine 3’:5’-monophosphate. J. Neutosci. 5, 1233-1238. Cohan, C. S., and Katet, S. B. (1986). Suppression gation and growth cone motility by electrical 232, 1638-1640.
of neutiteelonactivity. Science
Cohan, C. S., Connot, J. A., and Katet, S. B. (1987). and chemically mediated increases in intracellular neutonal growth cones. J. Neurosci. 7, 3588-3599.
Electrically calcium in
Conn, P. J., and Kaczmatek, L. K. (1989). A model for the study of the molecular mechanisms involved in the control of prolonged animal behaviors. Mol. Neutobiol. 3, 237-273. Conn, P. J., Strong, J. A., Azhdetian, E. M., Naitn, A. C., Gteengard, P., and Kaczmatek, L. K. (1988a). Protein kinase inhibitors selectively block photbol-ester or fotskolin-induced changes in excitability of Aplysia neurons. J. Neutosci. 9, 473479. Conn, P. J., Strong, J. A., and Kaczmatek, of protein kinase C pteventenhancementof action potentials in peptidetgic neurons 9, 480.-487.
L. K. (1988b). Inhibitors calcium current and of Aplysia. J. Neutosci.
Connot, J. A. (1986). Digital imaging of free calcium changes and of spatial gradients in growing processes in single, mammalian central nervous system cells. Ptoc. Natl. Acad. Sci. USA83,61796183. DeRiemet, S. A., Gteengatd, P., and Kaczmatek, Calcium/phosphatidylsetine/diacylglycetol-dependen~ phosphotylation in the Aplysia nervous system. 2672-2676.
L. K. (1985a). protein J. Neutosci. 5,
DeRiemet, S. A., Strong, 1. A., Albert, K. A., Cteengatd, P., and Kaczmatek, L. K. (1985b). Enhancement of calcium current in Aplysia neurons by photbol ester and protein kinase C. Nature 373, 313-316. Dudek, F. E., and Blankenship, J. E. (1976). Neutoendocrine (bag) cells of Aplysia: spike blockade and a mechanism for potentiation. Science 792, 1009-1010. Fink, IL. A., Connot, J. A., and Kaczmatek, L. K. (1988). lnositol ttisphosphate releases intracellulatly stored calcium and modulates ion channels in molluscan neurons. J. Neutosci. 8, 2544 2555. Forschet, P., and Smith, S. J. (1988). Actions of cytochalasins on the organization of actin filaments and microtubules in a neutonal growth cone. J. Cell Biol. 707, 1505-1516.
Fotschet, P., Kaczmatek, L. K., Buchannan, I. A., and Smith, S. J. (1987). Cyclic AMP induces changes in the distribution and ttansport of otganelles within growth cones of Aplysia bag cell neurons. J. Neutosci. 7, 3600-3611. Frazier, W. T., Kandel, E. R., Kupfetmann, I., Waziti, R., and Coggeshall, E. (1967). Morphological and functional properties of identified neurons in the abdominal ganglion of Aplysia californica. J. Neutophysiol. 30, 1288-1351. Grynkiewicz, G., Poenie, M., and Tsien, R. Y. (1985).A ation of calcium indicators with greatly improved properties. J. Biol. Chem. 260, 3440-3448. Jennings, K. R., Kaczmatek, and Sttumwasset, F. (1982). tetdischatge in peptidetgic 158-168.
new genetfluorescent
L. K., Hewick, R. M., Dreyer, W. I., Protein phosphotylation during afneurons of Aplysia. J. Neutosci. 2,
Jones, P. G., Rosset, S. J., and Bulloch, A. (2. M. (1987). Glutamate suppression of feeding and the underlying output of affectot neurons in Helisoma. Brain Res. 437, 56-68. Kaczmatek, a prolonged J. Neutosci.
L. K., and Kauet, refractory period 3, 2230-2239.
J. A. (1983). Calcium entry causes in peptidetgic neurons of Aplysia.
Kaczmatek, L. K., and Strumwasset, F. (1984). A voltage-clamp analysis of currents underlying cyclic-AMP induced membrane modulation in isolated peptidetgic neurons of Aplysia. J. Neutophysiol. 52, 340-349. Kaczmatek, L. K., Jennings, K., and Sttumwasset, F. (1978). Neurotransmitter modulation, phosphodiestet inhibitor effects, and cyclic AMP correlates of afterdischarge in peptidetgic neurites. Ptoc. Natl. Acad. Sci. USA 75, 5200-5204. Kaczmatek, L. K., Finbow, M., Revel, J. P., and Sttumwasset, F. (1979). The morphology and coupling of Aplysia bag cells within the abdominal ganglion and in cell culture. J. Neutobiol. 70,535550. Kaczmarek, L. K., Jennings, K. R., Sttumwasset, Walter, U., Wilson, F. D., and Gteengatd, P. (1980). of the catalytic subunit of cyclic-AMP-dependent enhances calcium action potentials of bag cell culture. Ptoc. Natl. Acad. Sci. USA 77, 7487-7491.
F., Naitn, A. C., Mictoinjecton protein kinase neurons in cell
Kaczmatek, L. K., Jennings, K., and Sttumwasset, F. (1982). An early sodium and a late calcium phase in the afterdischarge of peptide secreting neurons of Aplysia. Brain Res. 283, 105-115. Kaczmarek, L. K., Strong, J. A., and DeRiemet, S. A. (1985). Biochemical mechanisms that regulate potassium and calcium curtents in peptidetgic neurons. In Neurosectetion and the Biology of Neutopeptides, H. Kobayashi, H. A. Bern, and A. Utano, eds. (Berlin: Springer-Vetlag), pp. 80-87. Kandel, learning.
E. R., and Schwartz, J. H. (1982). Science 278, 433-443.
Molecular
Katet, S. B., Mattson, M. P., Cohan, C. S., and (1988). Calcium regulation of the neutonal growth Neutosci. 77, 315-321.
biology
of
Connot, J. A. cone. Trends
Lipscombe, D., Madison, D. V., Poenie, M., Reutet, H.,andTsien, R. W. (1988). Spatial distribution of calcium channels and cytosolic calcium transients in growth cones and cell bodies of sympathetic neurons. Ptoc. Natl. Acad. Sci. USA 85, 2398-2402. Loechnet, K. J., Azhdetian, E. M., Dryer, R., and Kaczmatek, (1990). Progressive potentiation of peptide release during tonal discharge. J. Neutophysiol. 63, 738-744.
L. K. a neu-
Strong, J. A., Fox, A. P., Tsien, R. W., and Kaczmatek, L. K. (1987). Stimulation of protein kinase C recruits covert calcium channels in Aplysia bag cell neurons. Nature 325, 714-717.
