- Email: [email protected]
Measurement of intracellular Ca2+ in cerebellar purkinje neurons in culture: Resting distribution and response to glutamate
Brain Resrcwh
Bulktin,
Vol. 21, pp.
3%361.
C Pergamon Press
plc, 1988.
0361-9230188 $3.00 + .OO
Printed in the U.S.A.
Measurement of Intracellular Ca2+ in Cerebellar Purkinje Neurons in Culture: Resting Distribution and Response to Glutamate JO~NA.CONNORAN~HS~U-YUTSENG AT&T Bell Laboratories,
Murray Hill, NJ 07974
neurons in culture: CONNOR, J. A. AND H.-Y, TSENG. ~~~~~s~~~~~~~~~~~~ of’ intvr~cell&r Co 2+ irr crrrhellnr Pdinje Rmitzg distrihrrtion cwd wspon.sr to ~lutrrmrte. BRAIN RES BULL 21(3) 353-361, 1988.--Ca ion levels in cerebellar Purkinje neurons, in culture, have been measured using the fluorescent indicator, fura-?, and digital imaging. Cells were loaded with the indicator both by injecting the free acid form and by allowing the membrane permeant form (/AM) become deesterified and trapped. The two methods gave significantly different results in that the /AM loaded cells showed localized regions of high Ca2+ in the soma whereas the injected cells did not. Resting levels in the remainder of the cytoplasm were similar however, as were the excursions in Ca?+ induced by electrical or chemical stimulation. Comparison of the data from the two methods suggests that qualitative measures of Ca in intracellular stores can be derived from the /AM loading method. Injected cells showed high Ca *+ levels in the soma that persisted for 3-8 minutes following removal of the injection electrode. The dendrites of these cells however maintained low CaZ+levels and differences of several hundred nM in Ca*+ were maintained between the soma and initial dendrite segment, demonstrating directly the large Ca pumping capacity of the dendrites. Localized regions of high Ca Zt in dendrites could be generated by applying glutamate from a microelectrode in TTX-Krebs saline. When studied in culture media with 4.7 mM K, the Purkinje neurons showed a bimodal distribution of Ca*+ with 35 to 40% showing stable Ca*+ levels between 250 and 350 nM, and the remainder 80 to 130 nM Ca*+. Granule neurons on the same coverslips had CaZ+ level in the lower range in >95% of the examples observed. Stimulus of the low Ca2+ Purkinje neurons with 1-3 set iontophoretic applications of glutamate from an extracellular microelectrode triggered increases in intracellular Ca” + that lasted for periods of several minutes. Similar Ca2+ changes induced by K depolarizations recovered in less than I min. Intracellular CaZ+
Purkinje neuron
Digital imaging
AT some point or another, nearly all of the current hypotheses of cellular “memory” mechanisms involve biochemical
Glutamate
imaging techniques [5]. Data presented here are from cerebellar Purkinje neurons grown in tissue culture. This type of preparation has proved extremely useful for electrophysiological studies (c.f., [lo, 11, 13, 201) but has not previously been exploited for optical measurements. We report differences in the localization of the indicator in cells depending upon the loading method used and also report measurements of Ca gradients produced by the excitatory amino acid, glutamate. Some of these changes are exceptionally long, with time courses of Ca elevations running to periods of minutes following a stimulus of I to ? sec. These long responses may reflect activation of other second messenger systems.
changes triggered by Ca 2+. The list includes ideas on how long term potentiation of synaptic inputs in the hippocampus works 14, 18, 25, 261, long-term depression in cerebellar Purkinje neurons (91 and long lasting effect of conditioning paradigms in molluscan neurons (c.f., 13,161). This convergence of theory is not particularly surprising given the large influxes of Ca ions that have been demonstrated during activity in some neurons (c.f., [l, 17, 191) and the large number of enzymes that are activated by Ca2+. There are, however, very few reliable measurements of the CaZ+ concentration changes that might actually be occurring in mammalian neurons under conditions where these mechanisms might be operative. Indeed, techniques and preparations whereby these difficult measurements can be made are only presently being developed. We describe a system for measuring Ca2+ concentrattons in dendrites and other processes of mammalian neurons using the flourescent Ca indicator, fura- [ 12,231 and digital
METHOD
Culttrrrs and Cdl Preparation Explant and trypsin dissociated monolayer cultures of rat cerebellum were prepared either from embryos (E20) or pups (PI-P4) as described previously [2,14]. Cells were grown in a
353
CONNOR AND TSENG
CCD CHIP: PHOTODIODES + SHIFT REGISTER
t
SERIAL READOUT OF STORED CHARGE
320
MICROSCOPE N. A. > 1
PRECISION SHUTTER 100-500 ms
FIG. 1. Diagram of the essential parts of imaging system. Charged coupled device camera is represented as a rectangular array of photodiodes and storage capacitors. The chip used employs a 320x512 array. The values of stored charge on elements are read out serially, digitized and stored as an array in the computer memory. A DEC LSI-I 1173computer is used in collecting and analyzing data. Changing the wavelength of the UV excitation was done manually during the time interval required to read the first of two images into the computer (-0.5 set).
media containing Minimal Essential Medium with Earle’s salts but without glutamine (MEM, Gibco, Grand Island, NY), supplemented with KC1 to a concent~tion of 25 mM. glucose (total, 6 g/l), NaHCO,, (total, 3.7 g/l), 2 mM glutamine, heat-inactivated horse serum (10% volivol, Gibco), and at 37” C in a humidified environment of 90% air/l@% COz (NAPCO, model 4600, Portland, OR). Culture medium was replaced 3 times per week. The best survival of all cells, and neurite development of the Purkinje neurons in particular was observed in horse serum (HS) supplemented media (MEM) containing 25 mM KCl. Data in this report were taken exclusively from cells grown in this medium, hereafter referred to as high K culture medium. Cells were grown on 18 mm, round coverslips coated with polylysine that served as the bottom of the recording chamber. Experiments were run at 31-33 degrees C in serum-free culture media with 4.7 mM K or in standard Krebs saline. Cells were loaded with the Ca indicator, fura-2, either by incubation with the acetoxymethylester (/AM) form of the dye as described previously [6,7], or by injecting the free acid form into single cells through a microelectrode. Microelectrodes were pulled from thin walled omega dot tubing (Haer, #30-30-O) and had resistances of 60-80 megohm when fdled with 3 M KCl. Tips of these electrodes were filled with 3 mM fura- dissolved in water and the barrel with 3 M KCI.
Although unreliable for precise electrical measurements, it was possible to record a negative voltage shift upon penetration of a cell. Generally, sufficient indicator leaked out of the electrode in a 20 to 40 set period to achieve adequate concentrations for measurement. During this period fluorescence was monitored intermittently, and the electrode removed after injection. Following experimental manipulations, a small circle (- I mm diam.) was scribed around the cell or cells examined and the tissue fixed and immunocytochemicahy stained as described previously [14], using PEP19 antibody [28]. In the cerebellum this antibody is specific for Purkinje neurons. Comparisons of cell shape, orientation of near neighbor cells was used to establish the identity of experimental cells as Purkinje neurons. Measurement
System
Ca levels were determined by obtaining pairs of images of the trapped indicator fluorescence (central wavelength 500 nm) using 340 nm and 380 nm excitation and then ratioing the two images [5,12]. The key elements of the photometer measurement system are a high numerical aperture UV objective and a good quality, low light level camera. The system used here is summarized in Fig. 1. The primary ele-
CA2+ IN CEREBELLAR
355
PURKINJE NEURONS
800nM
FIG. 2. A: Combined fluorescence-transmitted light image of Purkinje neuron filled with fura- and its nontilled neighbors. The excitation wavelength was 340 nm. Soma diameter of granule cell near the arrow is 5 Frn. B: Ratio image (3401380 nm excitation) of the Purkinje neuron 30 set after removal of the injection level at bifurcation -350 nM. C: 1 min after microelectrode. Soma Caz+ level - 750 nM. Dendrite electrode removal; soma Ca2+ level -5.50, dendrite - 300 nM D: 1.5 min: soma Ca*+ level - 400, dendrite -200 nM. E: 3 min: soma Ca2+ level 2.50 nM, dendrite -100 nM. Grey scale has been adjusted to give maximum contrast between cell regions.
356
CONNOR AND TSENG
FIG. 3. A: Ratio image of Purkinje neuron dendrite made 10 min after injection of indicator showing a gradient of Ca’+. B: Ratio image of hame neuron after moving the cell about one-half field to the left to show soma and second dendrite. The cell was bathed in serum-free Culture medium with 4.7 nM K for the records of A and B. C: Dendritic Ca 2+ level 3 min after replacing the culture medium with Kreb\ saline.
ment, a charge coupled device (CCD) camera is represented as a rectangular array of light sensing diodes each with a storage capacitor. In principle, each diode generates electrons
as it absorbs
light
and these
are stored
on the
capacitors. At the end of the capture period, or exposure, the amount of charge on each capacitor (voltage) is read out serially through a high impendance buffer amplifier. It is thus a “still camera” rather than a “video” camera and the preparation is illuminated only during the actual period that light is collected. It is a linear detector, and if a good grade chip is selected, free from nonuniformities in sensitivity across the field. Present day versions of CCD cameras are nearly identical to prototypes made at Bell Laboratories in the 1970’s [22]. In order to reduce thermal noise the CCD chip is cooled to approximately -40 degrees C. The voltage levels from the camera are digitized and stored as arrays in computer memory. It requires from 1 to I.5 set to collect and store a pair of images, depending upon their size and the exposure time. These image arrays may then be manipulated in some way or another, as by subtracting background or spatial filtering, or may be written directly to mass storage, a hard disk in this system. The arrays are displayed as grey levels or false color on a video monitor. For the indicator ratioing technique, pictures are taken using 340 and 380 nm excitation and then the numerical value of each picture element (pixel) at 340 nm excitation is divided by the value of the corresponding element in the 380 nm picture. This array of ratios is then displayed on a monitor by making higher ratio values brighter or by showing them on a color continuum as has become standard for the various tomography maps. RESULTS
Figure 2A show a composite of fura- injected into a Purkinje cells in the culture as they appear The cells had been growing in
picture of the fluorescence neuron and the surrounding in transmitted illumination. a monolayer culture for 25
days. The excitation wavelength was 340 nm. Ratio determinations of Ca” are made by dividing two such fluorescence pictures without the transmitted light being present. Fluorescence levels in the individual frames for this amount of injected indicator (approximately 200 PM) were approximately 40 times the camera noise and intrinsic fluorescence of underlying cells in secondary dendrites such as the one labeled by an arrow in Fig. 2A, to >_500times at the soma (see legend, Fig. 2). Indicator fluorescence from secondary dendrites was generally at least a factor of 10 greater than the background where the indicator was injected. Where the acetoxymethylester (/AM) form of fura- was used (see the Method section and Figs. 5 and 6 below), and all cells on the coverslip were loaded to one extent or another, the dendrite to background ratio was generally lower. It is of course desirable to have this ratio as high as possible in order to eliminate the need for uneven background corrections in Ca determination from ratio measurements, and to minimize the effects of spurious changes in the intrinsic fluorescence of underlying cells. The saving factors for the /AM loading technique are that there are many regions in a culture where the glial cells are very flat and therefore contain little indicator, and many glial cells appear to retain fura- more poorly than some neurons [7]. Thus dye injection is to be preferred over AM loading on some technical grounds, but is less desirable because of the trauma to the injected cell. Figure 2B shows the ratio image of the injected cell. The data were taken just after removal of the injection microelectrode, and the Ca levels in the soma were very high due to slight injury at the penetration site. The subsequent decrease in Ca*+ is followed over the next 5 min in 2B-E. Five minutes after the injection nearly uniform conditions had been established. In cells that survived injection (about 70%), the soma Ca levels gradually decreased over a period of 3 to 8 min and both soma and dendrites reached lower, more nearly uniformed levels as shown in the figure. What is most striking about records such as Fig. 2ED is the large concentration
CA*+ IN CEREBELLAR
PURKINJE
NEURONS
351
Fig. 4. A: Ratio image of Purkinje neuron loaded with membrane permeant fura-Z/AM. Ca Z+ level in the nuclear area (upper left portion of the soma) was -80 nM while the bright regions of the soma to the right of the nucleus registered around 140 nM. Levels further out the dendrite were approximately the same as the nuclear region. B: Flourescence picture of the neuron using 340 nm excitation. C and D: Fluorescence pictures made approximately IO seconds and 4tl secands after rupturing the soma membrane with a microelectrode.
of Ca2+ between the soma and its dendrites and that these gradients exist for minutes. Rupture of the cell membrane causes loss of all indicator within seconds, implying that the indicator, and the Ca’ L it is sensing, are free to move about in the cytoplasm. A sharp gradient like that of Fig. 2D then means that the dendrites are able to bind, sequester, and pump out CY+ at a rate so high that even large perturbations at the soma cause only relatively small increases 10 to 20 pm out the dendrite. Soma-dendrite gradients such as this were seen in every neuron injected (n>70), making it clear that large increases in Ca ion can occur in the soma and have little direct effect on the levels in dendrites. Ca” + levels are a sensitive function of the type of medium the cells are bathed in (Ca and K levels fixed), an observation that has important implications for experiments dealing with internal second messengers. For example, the events described in Fig. 2 were observed in a purely inorganic saline (Krebs). When experiments were carried out in culture media with 4.7 mM K, CaSi levels and distributions were much different. Figure 3 shows a portion of the soma and apical dendrite of a second neuron that had been injected 10 min previously and bathed in air/CO, equilibrated media during and after the injection. The cell had recovered beyond the point of high soma Ca”+ and had stabilized with the dendrite at a higher level. This type of distribution was seen in at least 5@%of the cells examined in culture media. For the figure it has been necessary to show the dendrite in two frames (Fig. 3A, B) because of Iimited imaging field size. Figure 3A shows the more distal end of the dendrite while 3B includes a portion of the soma. It can be seen that the Ca’? gradient
levels increase with distance from the soma. Figure 3C shows the Ca” distribution in this same cell 3 min after Krebs saline was substituted for the media. The levels had dropped and the concentration had become uniform, implying that a source of steady Ca” influx from the distal dendrite had been stopped. Even in cases where there were no obvious gradients of Ca”, a change from media to Krebs saline brought about a decrease in internal Ca”‘. These decreases occur at constant external temperature. pH, K, and Ca levels, and do not depend on the method of indicator loading. A similar observation was reported for rapidly growing cells early on in culture [5]. The composition of the bathing medium also greatly influences the duration of transmitter induced changes in Ca?- (see below, Fig. 6). Purkinje neurons loaded with the /AM form of the indicator showed a more complex distribution of Ca’+ levels than did cell where the dye was injected. Figure 4 shows an experiment illustrating this complexity and a probable underlying cause. Figure 4A shows a typical ratio image observed for /AM loading. The nuclear region is very low compared with the remaining soma (bright area) and the dendritic regions compare with the nuclear region. Where indicator was injected, there was no such distribution after the initial injection trauma has subsided (Fig. 2). Figure 4B shows indicator fluorescence using 340 nm excitation in the intact cell. Figure 4C and D are fluorescence pictures (340 nm excitation), taken 10 and 40 set after rupturing the cell membrane with a microelectrode. Brightness of the remaining indicator in these two panels relative to 4B has been electronically enhanced in addition to the increase in brightness expected
200
iopm I
53onM
FIG. 5. Timecourse of the response to iontophoretic application of glutamate from a microelectrode. Culture medium was flowing over the preparation. Position of electrode is shown in small bottom panel at right. Fluorescence of trapped indicator at the lower left. A: Ca*+ distribution in the cell before stimulus. B: Immediately after a 2 set glutamate application. Electrode removed from bath after the pulse. C: 50 set after glutamate pulse. D: 5 min after stimulus. E: 10 min after stimulus and 2 min after exchanging Krebs saline for culture medium in the bath. E: Ca2+ levels after 5 min in Krebs saline. Note localized regions of apparently high Ca *+ levels in the soma. Color coding for Ca 2c levels shown at bottom center.
55
,
CA*+ IN CEREBELLAR
FIG.
6. Localized
PURKINJE
359
NEURONS
response
fluorescence-transmitted B: Prestimulus baseline.
to glutamate stimulus applied to upper end of dendrite. Cell bathed in TTX-Krebs saline. A: Combined light picture of cell and surroundings. B, C. D: Ca 2t levels expressed as changes from prestimulus values (see text). C: Immediately after a 3 set application glutamate.D: Poststimulus recovery.
of
from high Ca2+ conditions in the damaged cell. It can be seen that the indicator quickly leaves the cell in the nuclear and dendrite regions but is trapped in other areas of the soma. Comparison of Fig. 4A and 4C shows that the areas of indicator trapping overlay with the bright (high Ca’+) regions in the ratio image. Thus with /AM loading the indicator is reporting Ca”+ levels in both the cytoplasm and intracellular compartments (see also [8,27]). Prolonged incubation with Ca-free media containing 2 I_LM ionomycin abolished the nonuniformities in the ratio image (such as Fig. 4A) as one would expect if the ionomycin, a Ca iontophore, were allowing Ca2+ to be drained from all intracellular compartments to the outside media. The above results indicate that under some conditions it is possible to follow both cytoplasmic levels of Caz+ and to monitor, at least in a relative way, the levels in intracellular compartments. Purkinje neurons studied under conditions more nearly simulating the long-term maintenance conditions; that is, bathed in culture medium at normal K levels (4.7 mM) and loaded noninvasively (/AM loading), presented a much more complex situation than was present in simple inorganic salines. The presence or absence of serum during experiments had no discernable effect. First there were large differences in the levels of Ca2+ between unstimulated cells. Many of the large neurons on a given coverslip showed stable intracellular Ca2+ levels of 250 to 350 nM. This was
particularly prevalent during the period from I2 to 25 days in when as many as 30 to 50% of the large neurons showed high Ca levels. A total of 25 cells showing high Ca’+ levels were monitored for ongoing spike activity by extracellular patch recording. Only 2 or 3 of these cells were active; thus it is not clear what the immediate causes of the elevated Ca” levels are. In a high percentage of these cells the Ca”+ levels dropped to under 100 nM when the media was replaced with Krebs saline. Cells that showed Ca2+ levels in the range of 50-120 nM in growth medium were tested for their responses to glutamate. Characteristically, these neurons showed a prolonged elevation in Ca2+ in response to iontophoretic applications. Figure 5 illustrates an example where a 3 set application of glutamate elicited a response that lasted for 10 min before recovery was induced by replacing the media with Krebs saline. Several washes with media were made in the initial 10 min period. Figure 5A shows Ca2+ distribution in the cell before stimulus. Figure 5B was made immediately after a 3 set iontophoretic pulse of glutamate was applied to the proximal apical dendrite (see Fig. 5H for electrode location). This caused a clear rise in Ca2+ in the dendrite but large effects were not seen in the rest of the cell until later. Figure 5C, 50 set after the stimulus, shows the peak of the response, with the largest Ca2+ increases occurring in portions of the soma. Ca*+ levels remained elevated over the next 8 min even though the glutamate electrode had been removed imculture
360
CONNOR AND TSENG
mediately after the stimulus and the bathing media exchanged several times (Fig. 5D, E). Only after the media was replaced with Krebs saline did the levels return towards rest over the next severai minutes (Fig. SF). In Fig. 5F clearly defined regions of elevated Ca2+ can be seen in the soma. We consider it likely that these are intracellular storage regions that have become loaded as a result of the long period of elevated Cazi. Similar responses, lasting from 4 to I2 min have been measured in approximately 30 neurons from 4 different primary preparations. The response described above is abolished in Ca-free saline as rapidly as the medium can be exchanged, in contrast to the reversal at normal external Ca illustrated in Fig. 5 which generally required several minutes. There was no Ca response to glutamate in Ca-free saline. Stimulation with high external K (50 mM applied from a 4 ,um pipette) generated Ca2+ conc~nt~tion changes that were as large as the glutamate induced changes but these changes were always transient with complete recovery occurring in 30 to 60 sec. The persisting levels of high Ca*+ were not maintained by extended periods of spiking since the addition of TTX to the medium did not bring about a rapid reduction of Ca. and in 3 cells monitored with external recording electrodes, spike activity was undetectable. Also, we have carried out an extensive set ofexperiments in which the effectsofglutamate stimuli were monitored by the whoie cell patch recording technique (Hockberger, Connor and Tseng, in preparation). Cells almost never fired for periods of more than a few seconds after glutamate iontophoresis, and in TTX saline were unable to generate self-sustaining Ca-spike activity. The experiment illustrated in Fig. 6 shows that undet specialized conditions the effect on Ca2’ of applying glutamate from a microelectrode could be localized to the dendrite. To prevent widespread increases in Ca’.. the cell was bathed in Krebs saline with I FM TTX to block propagating Na spike activity that was generally elicited in response to glutamate and to restrict presynaptic inputs to the cell. Krebs saline by itself promoted rapidly recovering responses as demonstrated above. Figure 6A shows the local area of cells with the fura- loaded neuron highlighted by its flourescence. For the remaining panels the ratio images have all been divided by the control ratio image so that only the regions where Ca 2+ has increased significantly brighten. Thus the image of the cell in Fig. 6B is absolutely uniform. because the frame is simply divided by itself. The CY+ level was -60 nM. The record of Fig. 6C was made immediately
after the application of glutamate (2 set iontophoretic pulse) and illustrates a large rise in Cazt in the vicinity of the application (to -400 nM) but a much smaller change elsewhere. even in the proximal end of the stimulated dendrite. This is shown in the figure by the fact that the soma and dendrite do not change shading much from Fig. 6B. Figure 6D shows Ca2+ levels 30 set after the stimulus when the initial levels had nearly been reestablished. Under these experimental conditions, Ca ion levels generally recovered to their initial values within 1 to 2 min after a stimulus.
These results demonstrate the exquisite ability of Purkinje neurons grown in culture to partition changes in intracellular Ca”. While it was not particularly unexpected to see localized changes in the dendrites in response to focal stimulation (Fig. 6), the ability of the dendrites to control a large flux from the soma for such long time periods was surprising (Fig.2). The falloff in Ca2+ concentration from the soma is probably a better indicator of the ability of dendrites to localize Ca2+ changes than the glutamate stimuli, since the radius of transmitter dispersion even from a microelectrode is large. Thus the rather general increase in the neurite tip in Fig. 4 may partially reflect external action of glutamate as well as internal diffusion of CaZi. The mechanism of generating the long Ca’+ responses (Fig. 5) in the Purkinje neurons is unclear at the present time. While it requires the presence of elements contained in growth media (other than serum), we do not know which ones. It is also unclear at this point whether the primary action of the glutamate is on the Purkinje neurons or on glia, analogous to the findings of Johnson and Ascher (ISI on giycine release in cultures. Leaving the mechanism as an unresolved problem, we regard it as likely that glutamate stimulation simply flips the neurons from a stable low Ca*+ state to a high Ca2+ state. a condition in which we observed many of the neurons without stimulation. Thus the interesting possibility exists that at least 2 different setpoints for internal Ca2+ may exist in these neurons, afthough we must emphasize that the conditions that develop in culture may not always reflect the properties of adult neurons in vivo. In view of the large number of enzymatic systems that are reguthis bistable characteristic could have lated by Ca”. profound effects upon the responses of Purkinje neurons under different conditions.
REFERENCES I. Ahmed, Z.; Connor, J. A. Measurement of calcium influx under 2. 3. 4. 5.
voltage clamp in molluscan neurons using the metallochromic dye Arsenazo III. J. Physiol. 285:61-82; 1979. Ahmed, Z.; Connor, J. A.; Tank, D.; Fellows, R. Expression of membrane currents in rat diencenhalic neurons in serum free culture. Dev. Brain Res. 28:221-231; 1986. Alkon, D. L. Calcium-mediated reduction of ionic currents: a biophysical memory trace. Science 226:1037-1045; 1984. Collindridge, G. L. Long term potentiation in the hippocampus: mechanisms of initiation and modulation by neurotransmitters. Trends Neurosci. 8:407-411; 198.5. _ Connor, J. A. Digital imaging of free calcium changes and of spatial gradients in growing processes in single, mammalian central nervous system cells. Proc Natl. Acad. Sci. USA 83:61796183; 1986.
6. Connor, J. A.; Cornwall, M. C.; Williams, G. II. Spatially resolved cytosolic calcium response to angiotensin II and potassium in rat glomerulosa cells measured by digital imaging techniques. J. Biol. Chem. 262:2919-2927; 1987. 7. Connor, J. A.; Tseng, H-Y.; Hackberger, P. E. Depolarization and transmitter induced changes in intracellular Ca of rat cerebellar granule cells in explant cultures. J. Neurosci. 7: 13841400; 1987. 8. Connor, J. A. Fluorescence imaging applied to the measurement of Ca2+ in mammalian neurons. In: Grinnel, A.; Armstrong, D.: Jackson. M., eds. New York: Plenum; 1988:395-406. 9. Ekerot, C. F.; Kano, M. Long-term depression of parallel fibre synapses following stimulation of climbing fibres. Brain Res. 342:357-360: 1983.
361
CA2+1NCEREBELLARPURKlNJENEURONS IO. Geller.
I I.
12
I3
I4
15.
16. 17.
IX. 19.
H.:
Woodward,
D.
Responses
of cultured
cerebellar
neurons to iontophoretically applied amino acids. Brain Res. 74:67-80: 1974. Gruel. D.: Franklin, C. L. Morphological and physiological differentiation of Purkinje neurons in cultures of rat cerebellum. J. Neurosci. 7:1271-1293: 1987. Grynkiewicz. G.: Poenie, M.: Tsien. R. Y. A new generation of Ca indicators with greatly improved fluorescence properties. J. Biol. Chem. 260:3440-3450: 1985. Hirano, A.: Ohmori, H. Voltage-gated and synaptic currents in rat Purkinje cells in dissociated cultures. Proc. Natl. Acad. Sci. USA 83: 1945-1949: 19X6. Hockberger, P. E.: Tseng. H-Y.; Connor, J. A. Immunocytochemical and electrophysiological differentiation of rat cerebellar granule cells in explant cultures. J. Neurosci. 7: l37& 1383: 1987. Johnson, J. W.; Ascher, P. Glycine potentiatea the NMDA response in cultured mouse brain neurons. Nature 325:579-531: 19x7. Kandel. E. R. Calcium and the control of synaptic strength by learning. Nature 293:697-700: 1981. Llinas, R.; Sugimori. M. Electrophysiological properties of in vitro Purkinje cell dendrites in mammalian cerebellar slices. J. Physiol. 305: 197-213: 1980. Lynch, G.: Baudry. M. The biochemistry of memory:a new and specific hypothesis. Science 224: 1057-1063: 1984. MacDermott. A. B.: Mayer, M. L.; Westbrook, G. L.: Smith. S. J.: Barker. J. L. NMDA-receptor activation increases cytoplasmic calcium concentration in cultured spinal cord neurones. Nature 311:519-522: 19X6.
Nelson,
P.: Peacock,
cultures
from
fetal
J. Electrical
mouse
activity
cerebellum.
in dissociated
Brain
cell
Res. 61:163-164:
1973. Romijn,
H.:
Habets,
A.:
Mud,
M.;
Walters,
P. Nerve
out-
growth. synaptogenesis and bioelectric activity in fetal rat cerebral cortex tissue cultured in serum-free. chemically defined medium. Dev. Brain Res. 2:583-589: 1982. Sequin, C. A.: Tompsett. F. Charge transfer devices. New York: Academic Press; 1975. Tsien, R. Y.: Poenie, M. Fluorescence ratio imaging: a new window into intracellular ionic signaling. Trends Biochem. Sci.
I I:45tk455: 1986. Tsien. R. Y.: Rink. T. J.: Poenie, M. Measurement of cytosolic free Ca in individual small cells using fluorescence microscopy with dual excitation wavelengths. Cell Calcium 6: 14.5-157: 1985. Turner. R. W.: Baimbridge. K. G.; Miller. J. Calcium processes of learning and memory in the hippocampus. Neuroscience 7:14l l-1416: 1981. Wigstrom. H.: Swarm. J. W.; Anderson. P. Calcium dependency of synaptic long-lasting potentiation in the hippocampal slice. Acta Physiol. Stand. 105: 116128: 1979. Williams. D. A.; Fogarty, K. E.: Tsien. R. Y.: Fay. F. S. Calcium gradients in single smooth muscle cells revealed by the digital imaging microscope using fura-2. Nature 318:55%S61; 19x5. /.iai. R.: Pan. Y.-C. E.: Hulmes, J. D.: Sangameswaran. L.; Morgan. J. I. Isolation. sequence, and developmental profile of a brain-specitic polypeptide, PEP-19. Proc. Natl. Acad. Sci. USA X3:842&8423: 1986.
Bulktin,
Vol. 21, pp.
3%361.
C Pergamon Press
plc, 1988.
0361-9230188 $3.00 + .OO
Printed in the U.S.A.
Measurement of Intracellular Ca2+ in Cerebellar Purkinje Neurons in Culture: Resting Distribution and Response to Glutamate JO~NA.CONNORAN~HS~U-YUTSENG AT&T Bell Laboratories,
Murray Hill, NJ 07974
neurons in culture: CONNOR, J. A. AND H.-Y, TSENG. ~~~~~s~~~~~~~~~~~~ of’ intvr~cell&r Co 2+ irr crrrhellnr Pdinje Rmitzg distrihrrtion cwd wspon.sr to ~lutrrmrte. BRAIN RES BULL 21(3) 353-361, 1988.--Ca ion levels in cerebellar Purkinje neurons, in culture, have been measured using the fluorescent indicator, fura-?, and digital imaging. Cells were loaded with the indicator both by injecting the free acid form and by allowing the membrane permeant form (/AM) become deesterified and trapped. The two methods gave significantly different results in that the /AM loaded cells showed localized regions of high Ca2+ in the soma whereas the injected cells did not. Resting levels in the remainder of the cytoplasm were similar however, as were the excursions in Ca?+ induced by electrical or chemical stimulation. Comparison of the data from the two methods suggests that qualitative measures of Ca in intracellular stores can be derived from the /AM loading method. Injected cells showed high Ca *+ levels in the soma that persisted for 3-8 minutes following removal of the injection electrode. The dendrites of these cells however maintained low CaZ+levels and differences of several hundred nM in Ca*+ were maintained between the soma and initial dendrite segment, demonstrating directly the large Ca pumping capacity of the dendrites. Localized regions of high Ca Zt in dendrites could be generated by applying glutamate from a microelectrode in TTX-Krebs saline. When studied in culture media with 4.7 mM K, the Purkinje neurons showed a bimodal distribution of Ca*+ with 35 to 40% showing stable Ca*+ levels between 250 and 350 nM, and the remainder 80 to 130 nM Ca*+. Granule neurons on the same coverslips had CaZ+ level in the lower range in >95% of the examples observed. Stimulus of the low Ca2+ Purkinje neurons with 1-3 set iontophoretic applications of glutamate from an extracellular microelectrode triggered increases in intracellular Ca” + that lasted for periods of several minutes. Similar Ca2+ changes induced by K depolarizations recovered in less than I min. Intracellular CaZ+
Purkinje neuron
Digital imaging
AT some point or another, nearly all of the current hypotheses of cellular “memory” mechanisms involve biochemical
Glutamate
imaging techniques [5]. Data presented here are from cerebellar Purkinje neurons grown in tissue culture. This type of preparation has proved extremely useful for electrophysiological studies (c.f., [lo, 11, 13, 201) but has not previously been exploited for optical measurements. We report differences in the localization of the indicator in cells depending upon the loading method used and also report measurements of Ca gradients produced by the excitatory amino acid, glutamate. Some of these changes are exceptionally long, with time courses of Ca elevations running to periods of minutes following a stimulus of I to ? sec. These long responses may reflect activation of other second messenger systems.
changes triggered by Ca 2+. The list includes ideas on how long term potentiation of synaptic inputs in the hippocampus works 14, 18, 25, 261, long-term depression in cerebellar Purkinje neurons (91 and long lasting effect of conditioning paradigms in molluscan neurons (c.f., 13,161). This convergence of theory is not particularly surprising given the large influxes of Ca ions that have been demonstrated during activity in some neurons (c.f., [l, 17, 191) and the large number of enzymes that are activated by Ca2+. There are, however, very few reliable measurements of the CaZ+ concentration changes that might actually be occurring in mammalian neurons under conditions where these mechanisms might be operative. Indeed, techniques and preparations whereby these difficult measurements can be made are only presently being developed. We describe a system for measuring Ca2+ concentrattons in dendrites and other processes of mammalian neurons using the flourescent Ca indicator, fura- [ 12,231 and digital
METHOD
Culttrrrs and Cdl Preparation Explant and trypsin dissociated monolayer cultures of rat cerebellum were prepared either from embryos (E20) or pups (PI-P4) as described previously [2,14]. Cells were grown in a
353
CONNOR AND TSENG
CCD CHIP: PHOTODIODES + SHIFT REGISTER
t
SERIAL READOUT OF STORED CHARGE
320
MICROSCOPE N. A. > 1
PRECISION SHUTTER 100-500 ms
FIG. 1. Diagram of the essential parts of imaging system. Charged coupled device camera is represented as a rectangular array of photodiodes and storage capacitors. The chip used employs a 320x512 array. The values of stored charge on elements are read out serially, digitized and stored as an array in the computer memory. A DEC LSI-I 1173computer is used in collecting and analyzing data. Changing the wavelength of the UV excitation was done manually during the time interval required to read the first of two images into the computer (-0.5 set).
media containing Minimal Essential Medium with Earle’s salts but without glutamine (MEM, Gibco, Grand Island, NY), supplemented with KC1 to a concent~tion of 25 mM. glucose (total, 6 g/l), NaHCO,, (total, 3.7 g/l), 2 mM glutamine, heat-inactivated horse serum (10% volivol, Gibco), and at 37” C in a humidified environment of 90% air/l@% COz (NAPCO, model 4600, Portland, OR). Culture medium was replaced 3 times per week. The best survival of all cells, and neurite development of the Purkinje neurons in particular was observed in horse serum (HS) supplemented media (MEM) containing 25 mM KCl. Data in this report were taken exclusively from cells grown in this medium, hereafter referred to as high K culture medium. Cells were grown on 18 mm, round coverslips coated with polylysine that served as the bottom of the recording chamber. Experiments were run at 31-33 degrees C in serum-free culture media with 4.7 mM K or in standard Krebs saline. Cells were loaded with the Ca indicator, fura-2, either by incubation with the acetoxymethylester (/AM) form of the dye as described previously [6,7], or by injecting the free acid form into single cells through a microelectrode. Microelectrodes were pulled from thin walled omega dot tubing (Haer, #30-30-O) and had resistances of 60-80 megohm when fdled with 3 M KCl. Tips of these electrodes were filled with 3 mM fura- dissolved in water and the barrel with 3 M KCI.
Although unreliable for precise electrical measurements, it was possible to record a negative voltage shift upon penetration of a cell. Generally, sufficient indicator leaked out of the electrode in a 20 to 40 set period to achieve adequate concentrations for measurement. During this period fluorescence was monitored intermittently, and the electrode removed after injection. Following experimental manipulations, a small circle (- I mm diam.) was scribed around the cell or cells examined and the tissue fixed and immunocytochemicahy stained as described previously [14], using PEP19 antibody [28]. In the cerebellum this antibody is specific for Purkinje neurons. Comparisons of cell shape, orientation of near neighbor cells was used to establish the identity of experimental cells as Purkinje neurons. Measurement
System
Ca levels were determined by obtaining pairs of images of the trapped indicator fluorescence (central wavelength 500 nm) using 340 nm and 380 nm excitation and then ratioing the two images [5,12]. The key elements of the photometer measurement system are a high numerical aperture UV objective and a good quality, low light level camera. The system used here is summarized in Fig. 1. The primary ele-
CA2+ IN CEREBELLAR
355
PURKINJE NEURONS
800nM
FIG. 2. A: Combined fluorescence-transmitted light image of Purkinje neuron filled with fura- and its nontilled neighbors. The excitation wavelength was 340 nm. Soma diameter of granule cell near the arrow is 5 Frn. B: Ratio image (3401380 nm excitation) of the Purkinje neuron 30 set after removal of the injection level at bifurcation -350 nM. C: 1 min after microelectrode. Soma Caz+ level - 750 nM. Dendrite electrode removal; soma Ca2+ level -5.50, dendrite - 300 nM D: 1.5 min: soma Ca*+ level - 400, dendrite -200 nM. E: 3 min: soma Ca2+ level 2.50 nM, dendrite -100 nM. Grey scale has been adjusted to give maximum contrast between cell regions.
356
CONNOR AND TSENG
FIG. 3. A: Ratio image of Purkinje neuron dendrite made 10 min after injection of indicator showing a gradient of Ca’+. B: Ratio image of hame neuron after moving the cell about one-half field to the left to show soma and second dendrite. The cell was bathed in serum-free Culture medium with 4.7 nM K for the records of A and B. C: Dendritic Ca 2+ level 3 min after replacing the culture medium with Kreb\ saline.
ment, a charge coupled device (CCD) camera is represented as a rectangular array of light sensing diodes each with a storage capacitor. In principle, each diode generates electrons
as it absorbs
light
and these
are stored
on the
capacitors. At the end of the capture period, or exposure, the amount of charge on each capacitor (voltage) is read out serially through a high impendance buffer amplifier. It is thus a “still camera” rather than a “video” camera and the preparation is illuminated only during the actual period that light is collected. It is a linear detector, and if a good grade chip is selected, free from nonuniformities in sensitivity across the field. Present day versions of CCD cameras are nearly identical to prototypes made at Bell Laboratories in the 1970’s [22]. In order to reduce thermal noise the CCD chip is cooled to approximately -40 degrees C. The voltage levels from the camera are digitized and stored as arrays in computer memory. It requires from 1 to I.5 set to collect and store a pair of images, depending upon their size and the exposure time. These image arrays may then be manipulated in some way or another, as by subtracting background or spatial filtering, or may be written directly to mass storage, a hard disk in this system. The arrays are displayed as grey levels or false color on a video monitor. For the indicator ratioing technique, pictures are taken using 340 and 380 nm excitation and then the numerical value of each picture element (pixel) at 340 nm excitation is divided by the value of the corresponding element in the 380 nm picture. This array of ratios is then displayed on a monitor by making higher ratio values brighter or by showing them on a color continuum as has become standard for the various tomography maps. RESULTS
Figure 2A show a composite of fura- injected into a Purkinje cells in the culture as they appear The cells had been growing in
picture of the fluorescence neuron and the surrounding in transmitted illumination. a monolayer culture for 25
days. The excitation wavelength was 340 nm. Ratio determinations of Ca” are made by dividing two such fluorescence pictures without the transmitted light being present. Fluorescence levels in the individual frames for this amount of injected indicator (approximately 200 PM) were approximately 40 times the camera noise and intrinsic fluorescence of underlying cells in secondary dendrites such as the one labeled by an arrow in Fig. 2A, to >_500times at the soma (see legend, Fig. 2). Indicator fluorescence from secondary dendrites was generally at least a factor of 10 greater than the background where the indicator was injected. Where the acetoxymethylester (/AM) form of fura- was used (see the Method section and Figs. 5 and 6 below), and all cells on the coverslip were loaded to one extent or another, the dendrite to background ratio was generally lower. It is of course desirable to have this ratio as high as possible in order to eliminate the need for uneven background corrections in Ca determination from ratio measurements, and to minimize the effects of spurious changes in the intrinsic fluorescence of underlying cells. The saving factors for the /AM loading technique are that there are many regions in a culture where the glial cells are very flat and therefore contain little indicator, and many glial cells appear to retain fura- more poorly than some neurons [7]. Thus dye injection is to be preferred over AM loading on some technical grounds, but is less desirable because of the trauma to the injected cell. Figure 2B shows the ratio image of the injected cell. The data were taken just after removal of the injection microelectrode, and the Ca levels in the soma were very high due to slight injury at the penetration site. The subsequent decrease in Ca*+ is followed over the next 5 min in 2B-E. Five minutes after the injection nearly uniform conditions had been established. In cells that survived injection (about 70%), the soma Ca levels gradually decreased over a period of 3 to 8 min and both soma and dendrites reached lower, more nearly uniformed levels as shown in the figure. What is most striking about records such as Fig. 2ED is the large concentration
CA*+ IN CEREBELLAR
PURKINJE
NEURONS
351
Fig. 4. A: Ratio image of Purkinje neuron loaded with membrane permeant fura-Z/AM. Ca Z+ level in the nuclear area (upper left portion of the soma) was -80 nM while the bright regions of the soma to the right of the nucleus registered around 140 nM. Levels further out the dendrite were approximately the same as the nuclear region. B: Flourescence picture of the neuron using 340 nm excitation. C and D: Fluorescence pictures made approximately IO seconds and 4tl secands after rupturing the soma membrane with a microelectrode.
of Ca2+ between the soma and its dendrites and that these gradients exist for minutes. Rupture of the cell membrane causes loss of all indicator within seconds, implying that the indicator, and the Ca’ L it is sensing, are free to move about in the cytoplasm. A sharp gradient like that of Fig. 2D then means that the dendrites are able to bind, sequester, and pump out CY+ at a rate so high that even large perturbations at the soma cause only relatively small increases 10 to 20 pm out the dendrite. Soma-dendrite gradients such as this were seen in every neuron injected (n>70), making it clear that large increases in Ca ion can occur in the soma and have little direct effect on the levels in dendrites. Ca” + levels are a sensitive function of the type of medium the cells are bathed in (Ca and K levels fixed), an observation that has important implications for experiments dealing with internal second messengers. For example, the events described in Fig. 2 were observed in a purely inorganic saline (Krebs). When experiments were carried out in culture media with 4.7 mM K, CaSi levels and distributions were much different. Figure 3 shows a portion of the soma and apical dendrite of a second neuron that had been injected 10 min previously and bathed in air/CO, equilibrated media during and after the injection. The cell had recovered beyond the point of high soma Ca”+ and had stabilized with the dendrite at a higher level. This type of distribution was seen in at least 5@%of the cells examined in culture media. For the figure it has been necessary to show the dendrite in two frames (Fig. 3A, B) because of Iimited imaging field size. Figure 3A shows the more distal end of the dendrite while 3B includes a portion of the soma. It can be seen that the Ca’? gradient
levels increase with distance from the soma. Figure 3C shows the Ca” distribution in this same cell 3 min after Krebs saline was substituted for the media. The levels had dropped and the concentration had become uniform, implying that a source of steady Ca” influx from the distal dendrite had been stopped. Even in cases where there were no obvious gradients of Ca”, a change from media to Krebs saline brought about a decrease in internal Ca”‘. These decreases occur at constant external temperature. pH, K, and Ca levels, and do not depend on the method of indicator loading. A similar observation was reported for rapidly growing cells early on in culture [5]. The composition of the bathing medium also greatly influences the duration of transmitter induced changes in Ca?- (see below, Fig. 6). Purkinje neurons loaded with the /AM form of the indicator showed a more complex distribution of Ca’+ levels than did cell where the dye was injected. Figure 4 shows an experiment illustrating this complexity and a probable underlying cause. Figure 4A shows a typical ratio image observed for /AM loading. The nuclear region is very low compared with the remaining soma (bright area) and the dendritic regions compare with the nuclear region. Where indicator was injected, there was no such distribution after the initial injection trauma has subsided (Fig. 2). Figure 4B shows indicator fluorescence using 340 nm excitation in the intact cell. Figure 4C and D are fluorescence pictures (340 nm excitation), taken 10 and 40 set after rupturing the cell membrane with a microelectrode. Brightness of the remaining indicator in these two panels relative to 4B has been electronically enhanced in addition to the increase in brightness expected
200
iopm I
53onM
FIG. 5. Timecourse of the response to iontophoretic application of glutamate from a microelectrode. Culture medium was flowing over the preparation. Position of electrode is shown in small bottom panel at right. Fluorescence of trapped indicator at the lower left. A: Ca*+ distribution in the cell before stimulus. B: Immediately after a 2 set glutamate application. Electrode removed from bath after the pulse. C: 50 set after glutamate pulse. D: 5 min after stimulus. E: 10 min after stimulus and 2 min after exchanging Krebs saline for culture medium in the bath. E: Ca2+ levels after 5 min in Krebs saline. Note localized regions of apparently high Ca *+ levels in the soma. Color coding for Ca 2c levels shown at bottom center.
55
,
CA*+ IN CEREBELLAR
FIG.
6. Localized
PURKINJE
359
NEURONS
response
fluorescence-transmitted B: Prestimulus baseline.
to glutamate stimulus applied to upper end of dendrite. Cell bathed in TTX-Krebs saline. A: Combined light picture of cell and surroundings. B, C. D: Ca 2t levels expressed as changes from prestimulus values (see text). C: Immediately after a 3 set application glutamate.D: Poststimulus recovery.
of
from high Ca2+ conditions in the damaged cell. It can be seen that the indicator quickly leaves the cell in the nuclear and dendrite regions but is trapped in other areas of the soma. Comparison of Fig. 4A and 4C shows that the areas of indicator trapping overlay with the bright (high Ca’+) regions in the ratio image. Thus with /AM loading the indicator is reporting Ca”+ levels in both the cytoplasm and intracellular compartments (see also [8,27]). Prolonged incubation with Ca-free media containing 2 I_LM ionomycin abolished the nonuniformities in the ratio image (such as Fig. 4A) as one would expect if the ionomycin, a Ca iontophore, were allowing Ca2+ to be drained from all intracellular compartments to the outside media. The above results indicate that under some conditions it is possible to follow both cytoplasmic levels of Caz+ and to monitor, at least in a relative way, the levels in intracellular compartments. Purkinje neurons studied under conditions more nearly simulating the long-term maintenance conditions; that is, bathed in culture medium at normal K levels (4.7 mM) and loaded noninvasively (/AM loading), presented a much more complex situation than was present in simple inorganic salines. The presence or absence of serum during experiments had no discernable effect. First there were large differences in the levels of Ca2+ between unstimulated cells. Many of the large neurons on a given coverslip showed stable intracellular Ca2+ levels of 250 to 350 nM. This was
particularly prevalent during the period from I2 to 25 days in when as many as 30 to 50% of the large neurons showed high Ca levels. A total of 25 cells showing high Ca’+ levels were monitored for ongoing spike activity by extracellular patch recording. Only 2 or 3 of these cells were active; thus it is not clear what the immediate causes of the elevated Ca” levels are. In a high percentage of these cells the Ca”+ levels dropped to under 100 nM when the media was replaced with Krebs saline. Cells that showed Ca2+ levels in the range of 50-120 nM in growth medium were tested for their responses to glutamate. Characteristically, these neurons showed a prolonged elevation in Ca2+ in response to iontophoretic applications. Figure 5 illustrates an example where a 3 set application of glutamate elicited a response that lasted for 10 min before recovery was induced by replacing the media with Krebs saline. Several washes with media were made in the initial 10 min period. Figure 5A shows Ca2+ distribution in the cell before stimulus. Figure 5B was made immediately after a 3 set iontophoretic pulse of glutamate was applied to the proximal apical dendrite (see Fig. 5H for electrode location). This caused a clear rise in Ca2+ in the dendrite but large effects were not seen in the rest of the cell until later. Figure 5C, 50 set after the stimulus, shows the peak of the response, with the largest Ca2+ increases occurring in portions of the soma. Ca*+ levels remained elevated over the next 8 min even though the glutamate electrode had been removed imculture
360
CONNOR AND TSENG
mediately after the stimulus and the bathing media exchanged several times (Fig. 5D, E). Only after the media was replaced with Krebs saline did the levels return towards rest over the next severai minutes (Fig. SF). In Fig. 5F clearly defined regions of elevated Ca2+ can be seen in the soma. We consider it likely that these are intracellular storage regions that have become loaded as a result of the long period of elevated Cazi. Similar responses, lasting from 4 to I2 min have been measured in approximately 30 neurons from 4 different primary preparations. The response described above is abolished in Ca-free saline as rapidly as the medium can be exchanged, in contrast to the reversal at normal external Ca illustrated in Fig. 5 which generally required several minutes. There was no Ca response to glutamate in Ca-free saline. Stimulation with high external K (50 mM applied from a 4 ,um pipette) generated Ca2+ conc~nt~tion changes that were as large as the glutamate induced changes but these changes were always transient with complete recovery occurring in 30 to 60 sec. The persisting levels of high Ca*+ were not maintained by extended periods of spiking since the addition of TTX to the medium did not bring about a rapid reduction of Ca. and in 3 cells monitored with external recording electrodes, spike activity was undetectable. Also, we have carried out an extensive set ofexperiments in which the effectsofglutamate stimuli were monitored by the whoie cell patch recording technique (Hockberger, Connor and Tseng, in preparation). Cells almost never fired for periods of more than a few seconds after glutamate iontophoresis, and in TTX saline were unable to generate self-sustaining Ca-spike activity. The experiment illustrated in Fig. 6 shows that undet specialized conditions the effect on Ca2’ of applying glutamate from a microelectrode could be localized to the dendrite. To prevent widespread increases in Ca’.. the cell was bathed in Krebs saline with I FM TTX to block propagating Na spike activity that was generally elicited in response to glutamate and to restrict presynaptic inputs to the cell. Krebs saline by itself promoted rapidly recovering responses as demonstrated above. Figure 6A shows the local area of cells with the fura- loaded neuron highlighted by its flourescence. For the remaining panels the ratio images have all been divided by the control ratio image so that only the regions where Ca 2+ has increased significantly brighten. Thus the image of the cell in Fig. 6B is absolutely uniform. because the frame is simply divided by itself. The CY+ level was -60 nM. The record of Fig. 6C was made immediately
after the application of glutamate (2 set iontophoretic pulse) and illustrates a large rise in Cazt in the vicinity of the application (to -400 nM) but a much smaller change elsewhere. even in the proximal end of the stimulated dendrite. This is shown in the figure by the fact that the soma and dendrite do not change shading much from Fig. 6B. Figure 6D shows Ca2+ levels 30 set after the stimulus when the initial levels had nearly been reestablished. Under these experimental conditions, Ca ion levels generally recovered to their initial values within 1 to 2 min after a stimulus.
These results demonstrate the exquisite ability of Purkinje neurons grown in culture to partition changes in intracellular Ca”. While it was not particularly unexpected to see localized changes in the dendrites in response to focal stimulation (Fig. 6), the ability of the dendrites to control a large flux from the soma for such long time periods was surprising (Fig.2). The falloff in Ca2+ concentration from the soma is probably a better indicator of the ability of dendrites to localize Ca2+ changes than the glutamate stimuli, since the radius of transmitter dispersion even from a microelectrode is large. Thus the rather general increase in the neurite tip in Fig. 4 may partially reflect external action of glutamate as well as internal diffusion of CaZi. The mechanism of generating the long Ca’+ responses (Fig. 5) in the Purkinje neurons is unclear at the present time. While it requires the presence of elements contained in growth media (other than serum), we do not know which ones. It is also unclear at this point whether the primary action of the glutamate is on the Purkinje neurons or on glia, analogous to the findings of Johnson and Ascher (ISI on giycine release in cultures. Leaving the mechanism as an unresolved problem, we regard it as likely that glutamate stimulation simply flips the neurons from a stable low Ca*+ state to a high Ca2+ state. a condition in which we observed many of the neurons without stimulation. Thus the interesting possibility exists that at least 2 different setpoints for internal Ca2+ may exist in these neurons, afthough we must emphasize that the conditions that develop in culture may not always reflect the properties of adult neurons in vivo. In view of the large number of enzymatic systems that are reguthis bistable characteristic could have lated by Ca”. profound effects upon the responses of Purkinje neurons under different conditions.
REFERENCES I. Ahmed, Z.; Connor, J. A. Measurement of calcium influx under 2. 3. 4. 5.
voltage clamp in molluscan neurons using the metallochromic dye Arsenazo III. J. Physiol. 285:61-82; 1979. Ahmed, Z.; Connor, J. A.; Tank, D.; Fellows, R. Expression of membrane currents in rat diencenhalic neurons in serum free culture. Dev. Brain Res. 28:221-231; 1986. Alkon, D. L. Calcium-mediated reduction of ionic currents: a biophysical memory trace. Science 226:1037-1045; 1984. Collindridge, G. L. Long term potentiation in the hippocampus: mechanisms of initiation and modulation by neurotransmitters. Trends Neurosci. 8:407-411; 198.5. _ Connor, J. A. Digital imaging of free calcium changes and of spatial gradients in growing processes in single, mammalian central nervous system cells. Proc Natl. Acad. Sci. USA 83:61796183; 1986.
6. Connor, J. A.; Cornwall, M. C.; Williams, G. II. Spatially resolved cytosolic calcium response to angiotensin II and potassium in rat glomerulosa cells measured by digital imaging techniques. J. Biol. Chem. 262:2919-2927; 1987. 7. Connor, J. A.; Tseng, H-Y.; Hackberger, P. E. Depolarization and transmitter induced changes in intracellular Ca of rat cerebellar granule cells in explant cultures. J. Neurosci. 7: 13841400; 1987. 8. Connor, J. A. Fluorescence imaging applied to the measurement of Ca2+ in mammalian neurons. In: Grinnel, A.; Armstrong, D.: Jackson. M., eds. New York: Plenum; 1988:395-406. 9. Ekerot, C. F.; Kano, M. Long-term depression of parallel fibre synapses following stimulation of climbing fibres. Brain Res. 342:357-360: 1983.
361
CA2+1NCEREBELLARPURKlNJENEURONS IO. Geller.
I I.
12
I3
I4
15.
16. 17.
IX. 19.
H.:
Woodward,
D.
Responses
of cultured
cerebellar
neurons to iontophoretically applied amino acids. Brain Res. 74:67-80: 1974. Gruel. D.: Franklin, C. L. Morphological and physiological differentiation of Purkinje neurons in cultures of rat cerebellum. J. Neurosci. 7:1271-1293: 1987. Grynkiewicz. G.: Poenie, M.: Tsien. R. Y. A new generation of Ca indicators with greatly improved fluorescence properties. J. Biol. Chem. 260:3440-3450: 1985. Hirano, A.: Ohmori, H. Voltage-gated and synaptic currents in rat Purkinje cells in dissociated cultures. Proc. Natl. Acad. Sci. USA 83: 1945-1949: 19X6. Hockberger, P. E.: Tseng. H-Y.; Connor, J. A. Immunocytochemical and electrophysiological differentiation of rat cerebellar granule cells in explant cultures. J. Neurosci. 7: l37& 1383: 1987. Johnson, J. W.; Ascher, P. Glycine potentiatea the NMDA response in cultured mouse brain neurons. Nature 325:579-531: 19x7. Kandel. E. R. Calcium and the control of synaptic strength by learning. Nature 293:697-700: 1981. Llinas, R.; Sugimori. M. Electrophysiological properties of in vitro Purkinje cell dendrites in mammalian cerebellar slices. J. Physiol. 305: 197-213: 1980. Lynch, G.: Baudry. M. The biochemistry of memory:a new and specific hypothesis. Science 224: 1057-1063: 1984. MacDermott. A. B.: Mayer, M. L.; Westbrook, G. L.: Smith. S. J.: Barker. J. L. NMDA-receptor activation increases cytoplasmic calcium concentration in cultured spinal cord neurones. Nature 311:519-522: 19X6.
Nelson,
P.: Peacock,
cultures
from
fetal
J. Electrical
mouse
activity
cerebellum.
in dissociated
Brain
cell
Res. 61:163-164:
1973. Romijn,
H.:
Habets,
A.:
Mud,
M.;
Walters,
P. Nerve
out-
growth. synaptogenesis and bioelectric activity in fetal rat cerebral cortex tissue cultured in serum-free. chemically defined medium. Dev. Brain Res. 2:583-589: 1982. Sequin, C. A.: Tompsett. F. Charge transfer devices. New York: Academic Press; 1975. Tsien, R. Y.: Poenie, M. Fluorescence ratio imaging: a new window into intracellular ionic signaling. Trends Biochem. Sci.
I I:45tk455: 1986. Tsien. R. Y.: Rink. T. J.: Poenie, M. Measurement of cytosolic free Ca in individual small cells using fluorescence microscopy with dual excitation wavelengths. Cell Calcium 6: 14.5-157: 1985. Turner. R. W.: Baimbridge. K. G.; Miller. J. Calcium processes of learning and memory in the hippocampus. Neuroscience 7:14l l-1416: 1981. Wigstrom. H.: Swarm. J. W.; Anderson. P. Calcium dependency of synaptic long-lasting potentiation in the hippocampal slice. Acta Physiol. Stand. 105: 116128: 1979. Williams. D. A.; Fogarty, K. E.: Tsien. R. Y.: Fay. F. S. Calcium gradients in single smooth muscle cells revealed by the digital imaging microscope using fura-2. Nature 318:55%S61; 19x5. /.iai. R.: Pan. Y.-C. E.: Hulmes, J. D.: Sangameswaran. L.; Morgan. J. I. Isolation. sequence, and developmental profile of a brain-specitic polypeptide, PEP-19. Proc. Natl. Acad. Sci. USA X3:842&8423: 1986.