- Email: [email protected]
Effects of inorganic mercury on [3H]dopamine release and calcium homeostasis in rat striatal synaptosomes
TOXICOLOCIY
AND
APPLIED
PHARMACOLOGY
102,3
16-330
(1990)
Effects of Inorganic Mercury on [3H]Dopamine Release and Calcium Homeostasis in Rat Striatal Synaptosomes’ F. HARE,‘,~ S. MEHDI REZAZADEH,~ GARY P. COOPER, DANIEL J. MINNEMA. AND I. ARTHUR MICHAELSON
MICHAEL
Effects of Inorganic Mercury on [3H]Dopamine Release and Calcium Homeostasis in Rat Striatal Synaptosomes. HARE, M. F.. REZAZADEH, S. M., COOPER, G. P.. MINNEMA. D. J.. AND MICHAELSON, 1. A. ( 1990). 7i,\-icol. .4p/~/. Plrurmuc~ol. 102. 3 16-330. Inorganic mercury (Hg”) in vitro increases spontaneous transmitter release from nerve terminals. The mechanisms of action are not well understood but may involve alterations in intraterminal Ca” dynamics, In this study we describe actions of Hg’+ in vitro on isolated mammalian CNS striatal nerve terminals (synaptosomes). Cobalt (2 mM) completely blocked the effect of 2 +M Hg’+ on spontaneous [‘Hldopamine release. Cadmium (100 FM) was equipotent to Co*’ in blocking depolarization-dependent [3H]dopamine release, but did not alter the 2 PM Hg’+-induced spontaneous [‘Hldopamine release. Depolarization-dependent [‘Hldopamine release was not altered by 5 MM Hg’+. It appears that the site of action of Hg’+ on spontaneous [3H]dopamine release is not the Ca” channel. The effects of Hg” on intraterminal ionized Calf ([Ca”],) were evaluated using the Ca”-specific fluorescent probe. fura-2. Hg” (1-8 KM) had no effect on [Ca”], in 1.2 mM Ca”-containing buffers. In nominal Ca”’ media. 4 and 8 pM Hg’+ significantly decreased [Ca”],. Following exposure to 4 and 8 NM Hg’+ the quenching of extrasynaptosomal fura-? by Mn” was increased. suggesting that Hg’+ facilitated the leakage offura-2. This apparent leakage was probably due to a nonspecific increase in membrane permeability since 2 pM Hg*’ produced a Co”-insensitive increase in [3H]deoxyglucose phosphate efflux. Hg” (2-8 KM) did not increase the leakage of either lactate dehydrogenase or soluble protein from synaptosomes. Hg’+ produced a concentration-dependent ( l-8 pM) increase in 45Ca’ ’ efflux from superfused synaptosomes which was insensitive to blockade either by 2 mM Co’+ or by 100 NM Cd’+. These data suggest that the transmitter releasing action of Hg’+ involves interactions with sites that also interact with Co” but not with Cd’+. Furthermore, Hg” may have direct transmitter releasing actions (i.e.. Ca’+-mimetic properties). as well as nonspecific actions on plasma membrane permeability which may not necessarily be linked to [3H]dopamine release. $, IWO Academ,c Prers. Inc.
Inorganic mercury (Hg”+) increases spontaneous transmitter release from both the frog neuromuscular junction and the rat brain synaptosomes (Manalis and Cooper, 1975; Juang, 1976; Hare et al., 1989). The mechanism(s) responsible for these effects has been hypothesized to result from either (a) alterations in intracellular calcium ion (Ca2+) reg-
’ Portions ofthis manuscript have been reported in abstract form in To\-icohgist. 1989, 9(I). 97 [Abstract No. 3851. Supported by ES-03992. ’ To whom correspondence should be addressed. ’ Present address: Department of Pharmacology & Toxicology, Michigan State University, East Lansing.
MI 48824- I3 17. ’ Division of Pharmacology, University of Texas at Austin.
004 1-008x/90
College of Pharmacy. Austin. TX 787 12.
$3.00
Copyright Q 1990 hy Academic Prrss, Inc. All rights of rcproductmn in any form reserved
316
MERCURY-INDUCED
DOPAMINE
RELEASE
ulation (Binah et al., 1978; Cooper and Manalis, 1983); and/or (b) Ca*+-mimetic actions of the mercuric ion (Cheung, 1984). It is well documented that alterations in the concentration of free ionized Ca*+ ([Ca*‘]i) in nerve terminals influence neurotransmitter release (Augustine et al., 1987). Perturbations in intraterminal Ca2+ homeostasis leading to increases in [Ca*+]i could mediate the observed increases in spontaneous neurotransmitter release following Hg*+. It is known that Hg’+ can alter intracellular Ca2’ homeostasis in a variety of cell types and organelles. For instance, Hg’+ (31 PM) decreases 45Ca’t uptake into isolated rat brain mitochondria (Binah et al., 1978). Mercuric chloride (10 FM) inhibits ATP-dependent “Ca2+. uptake by rat brain nonmitochondrial intrasynaptosomal particles, but has minimal effects on depolarization-dependent 4sCa’t uptake into intact rat brain synaptosomes (Hood and Harris, 1980). Mercuric chloride (a 10 PM) produces a dose-dependent increase in the efflux of s6Rb+ from rat erythrocytes (Sneddon, 1987). Presumably, the *6Rb+ efflux is through a Ca’+-sensitive K+ channel that is activated by elevated [Ca’+]i. Direct measurements of Hg”-induced [Ca”]i increases in cultured renal tubular cells with the Ca’j-specific fluorescent probe fura(Smith et al., 1987) suggest that Hg*+ may likewise increase [Ca”] in synaptosomes. Increases in transmitter release could then be mediated by Ca” -dependent vesicular exocytotic machinery. We present data that only partially support a hypothesis of altered intrasynaptosomal Ca’+ homeostasis by Hg”. It is likely that the effects of Hg’+ on nerve terminals are complex and involve multiple mechanisms that are both concentration and time dependent. Mercury-induced transmitter release in synaptosomes may involve a direct action of Hg’+ on transmitter release sites, whereas alterations in Ca’+ homeostasis may occur via other mechanisms which are not necessarily
AND
CALCIUM
HOMEOSTASIS
317
linked to the observed changes in transmitter release.
tein .4ssay
.4nimals and animal cure. Adult male Long-Evans rats ( 150-250 g, Charles River, Wilmington, MA) were housed two per cage. fed Laboratory Chow 5001 (Ralston Purina, St. Louis, MO), and provided with tap water ad libitum. All animal care procedures were in accordance with NIH guidelines. Preparation of.s.~frupfosonzes. 411 synaptosomal isolation procedures were performed at 3°C. “Purified” synaptosomes were prepared as described by Minnema and Michaelson ( 1986). Briefly. rat sttiatal homogenates were centrifuged at IOOOg for IO mitt, and the supemates layered onto a discontinuous 0.8/ I .2 M sucrose gradient and centrifuged at 223,0001: for 25 min. The synaptosomes were removed from the 0.8- 1.2 M sucrose interphase. diluted with 0.32 M sucrose. and kept on ice until used (~4 hr from termination). Protein was determined using the Coomassie blue method of Bradford (1976). Synaptosomes (240 pg protein adjusted to 200 ~1 with 0.32 M sucrose) were added to 800 (~1 of physiological Hepes buffer. In the 45Ca” efilux studies, the Hepes buffer contained 20 mM Na’ (replaced isoosmotically with choline chloride). The Hepes buffer consisted of (mM). I40 NaCI. 5 KCI, 1.2 MgS04, I I. I glucose. 8.5 Hepes, and 5 Trizma Base. In experiments using [3H]dopamine. 0.5 mM nialamide, a monoamine oxidase inhibitor was included. For the uptake of [3H]deoxyglucose. the Hepes buffer contained 1 mM glucose (replaced isoosmotically with sucrose). Following the addition of a 25-~1 aliquot of0.5 PCi of [3H]dopamine (26.7 Ci/mmol). 6.25 &i of [3H]deoxyglucose (30.2 Ci/mmol). or 2 PCi of 4’CaCI, (1.6 mCi/mmol) the synaptosomes were incubated for I5 min at 37°C under Oz/C02 (95%/5%) in a Dubnoff metabolic shaker. The “loaded” synaptosomes were layered onto a 0.65-pm pore-size cellulose acetate filter in a filter support chamber, connected to a “push-pull” design superfusion system (Minnema and Michaelson. 1985). and superfused (2.0 ml/min) for 14 min to remove extrasynaptosomal [-‘H]dopamine and [‘Hldeoxyglucase. In the 45Ca’+ efflux experiments, synaptosomes were first superfused for IO min with Hepes buffer containing 100 pM EGTA and then for 4 min in normal Hepes buffer. Fractions were then collected directly into liquid scintillation vials at timed intervals. Radioactivity was assessed in a Packard Model 460 liquid scintillation spectrometer.
318
HARE
Lactare dehydrogenase activity attdproteitl d~~tertnrnutirms. Synaptosomes (70 pg protein adjusted to IO0 ~1) were added to 380 ~1 Hepes buffer that contained either no added CaCI, (nominal Ca’+) or I .2 mM CaCl? and placed in a Dubnoff metabolic shaker (37°C) for 5 min. Hepes buffer (20 ~1) that contained [Hg’+] to give 2.4, or 8 pM (final concentration) was added. For controls, 30 pl of Hepes buffer only was added. Synaptosomes were then incubated for an additional 5 min. the reaction was terminated by centrifugation (8000x, 5 min). and the supernate was removed and stored (2°C) for later analysis. To determine total synaptosomal lactate dehydrogenase activity (LDH) and soluble protein. synaptosomes were exposed to Hg2+ as above for 5 min. lysed with 100 ~1 of I .4% Triton X- 100 (0.23% final concentration), and centrifuged (8OOOg; 5 min) and the supernatant was removed. Lactate dehydrogenase activity was determined using a modification of the method of Bergmeyer and Bernt (1974). The reaction mix contained (final concentration. mM) 41.2 KZHP04, 6.78 KH,PO,, 0.58 Na pyruvate. 0. I2 NADH. I.10 NaHCO,, and 0.04 L-cysteine (free base). Briefly. a 167-p] aliquot of supernatant was added to cuvettes containing the reaction mix. The final volume was 2.7 I ml. The conversion of NADH to NAD+ was monitored by a decrease in light absorption at 340 nm. The reaction proceeded for IO min at 37°C and specific activity of LDH was calculated. Extrasynaptosomal protein was determined using the Coomassie blue method (Bradford. 1976).
Experiments involving the measurement of intrasynaptosomal ionized Ca’+ ([Ca’+],) were conducted in the laboratory of Dr. Steven Leslie. Division of Pharmacology. College of Pharmacy. University of Texas at Austin. Adult male Sprague-Dawley rats (200-300 g. University of Texas, Animal Resources Center) were used in all experiments. Synaptosomes were prepared from striata according to the method of Cotman ( 1974). The synaptosomes were resuspended in 4 ml of buffer which consisted of (mM) I36 NaCI, 5 KCI, I .3 MgC& ~ IO glucose, and 20 Trizma Base, and the pH was adjusted to 7.5 with I M maleic acid. The protein concentration ranged from 6 to 7 mg/ml. The synaptosomal suspension was divided into two 2ml aliquots. poured into two glass vials, placed in a Dubnoff metabolic shaker. and equilibrated for IO min at 37°C. Fura-2/AM (I mM), dissolved in dimethyl sulfoxide (DMSO), was added to one ofthe vials containing the synaptosomal suspension (final concentration = IO PM). An equal volume of DMSO (I % final concentration) was added to the other vial. designated as unloaded preparation. and served as control. The suspensions were incu-
ET AL. bated for 20 min, diluted fivefold with 37°C incubation medium. and transferred to two 50-ml centrifuge tubes for an additional IO-min incubation. The tubes were then centrifuged (I 3,OOOg; IO min) and the pellets resuspended in 2 ml of ice-cold incubation medium and placed on ice. The protein concentration of the suspension varied between I .6 and 2.5 mg/ml. Aliquots (100 ~1) of fura-2-loaded synaptosomal suspensions or unloaded suspensions were resuspended in 2 ml (37°C) of buffer containing either nominal CaCl? or I.2 mM CaC& in polystyrene cuvettes. The [Ca’+]” in nominal Cal+ buffer was determined using free furaand ranged between 0.5 and I .O pM. The samples were gently vortexed and incubated for 5 min. For studying the effects of Hg*+ on the resting [Ca’+],. Hg*+ (l-8 PM) or buffer (for control samples) was added and incubated for an additional 5 min. Baseline fluorescence was then measured at 340- and 380-nm excitation wavelengths with the emission wavelength set at 505 nm, using a Spex Fluorolog Spectrofluorometer (Spex Industries. Edison. NJ). The slitwidth was set at 1 mm (3.6 nm bandwidth) for both excitation and emission. The temperature ofthe cuvette compartment was maintained at 37’C. The minimum fluorescence (F,,,,,) was determined by adding 50 ~1 of a 4% solution of sodium dodecyl sulfate (SDS) and 20 ~1 of 0.5 M EGTA (in 3 M Trizma Base). R,,, was calculated as the ratio of FAilO/FiIo. To determine F,,,a,,,. using a different sample, the tissue was lysed with SDS and then 20 ni of0.4 M Ca” solution (final concentration = 4 IIIM) was added. Under these conditions, the ratio of ~‘&~~x,, is known as R,,,. Dye leakage was assessed by adding 20 ~1 of4 tnM Mn2+ (final concentration = 40 FM) to the cuvettes after determining baseline fluorescence. Fluorescence signals were also corrected for autofluorescence. The [Ca’+], was calculated according to the method of Grynkiewicz rt al. ( 1985) [Ca”],
= &X
(R - R,,,/R,,,
- R) X (Sf,/Sb&
where Sfi is fluorescence of furaat minimal Ca’+ and SbZ is that at Cal+ saturation at the excitation wavelength 380 nm. k;, the dissociation constant of the fura-2-Ca’+ complex. is 224 nM (Grynkiewicz PI al.. 1985). Da/a hatzdlin!:at~dsfatis/ics. The amount ofradiolabel released per fraction is expressed as a percentage of the total radioactivity:
(total
dpm released per fraction X 100 dpm released + dpm remaining on filter)
The numberoftrials (n) is indicated in each ligure legend. One animal was used per trial. In the furaexperiments. the pooled striata from two animals were used per trial (n). One-way analysis of variance (ANOVA) was used to analyze the data from superfusion experiments when appropriate. Data from the LDH. protein. and furaexperiment were analyzed using either one-way or two-way
MERCURY-INDUCED
DOPAMINE
RELEASE
ANOVA (Steel and Torrie. 1960). Post hoc analyses were done using the Newman-Keuls test at p < 0.05 signiiicance level. In some experiments the number of replications per treatment varies. In these cases, appropriate alterations were made for both ANOVA and post hoc tests. c%emi~&. [jH]Dopamine (3.4-[8-3H(A’)]dihydroxyphenylethylamine, NET-094). [3H]deoxyglucose (2-I 1,2‘H]deoxy-D-glucose. NET-549A). 45CaC12 (NEZ-013). and Aquasol-? were obtained from New England Nuclear (Boston, MA). Ultrapure sucrose was obtained from Beckman (Palo Alto, CA). Fura-Z/AM (furaacetoxymethyl ester) was purchased from Molecular Probes, Inc. (Eugene, OR). Hepes, Trizma Base. nialamide. EGTA, NADH, pyruvate. HgCl?. CdCL. and CoC& were obtained from Sigma (St. Louis, MO). All other chemicals were of reagent grade and were obtained from available commercial sources. All solutions were prepared with double distilled water.
RESULTS Eflect of Hg’+ [‘H]Dopamine
on Depolarization-Evoked Release
Striatal synaptosomes superfused in physiological Hepes buffer containing 0.25 mM Ca’+ released approximately 1% of the total [3H]dopamine per 15set fraction (fractions l-4; Fig. 1). During fraction 6 synaptosomes were exposed to a 1 S-set pulse of 6 1 mM G (solid line: Fig. 1). This [K+], is sufficient to depolarize striatal synaptosomes which, in the presence of [Ca2+10 > 0.025 mM, produces an increase in the rate of [3H]dopamine release above baseline (Minnema and Michaelson, 1985). Depolarization-dependent [3H]dopamine release returned to baseline within 45 sec. Addition of 5 PM Hg*+ to the Hepes buffer (dashed line) beginning with fraction 5 followed 15 set later (fraction 6) by a depolarizing pulse of K’ resulted in a release of [3H]dopamine indistinguishable from that in the absence of Hg” (solid line). Hg’+ (5 /*M) alone produced an increase in spontaneous [3H]dopamine release with an onset latency of approximately 30 set (dotted line).
AND
CALCIUM
HOMEOSTASIS
Efect of Co” and Cd2’ on Hg’+-Induced creases in Spontaneous [3H]Dopamine lease
319 InRe-
In nominal Ca’+ media, 2 PM Hg’+ increased spontaneous [3H]dopamine release (Fig. 2). When 2 mM Co” was included in the buffer along with Hg2+ there was a total block of the Hg*‘-induced spontaneous [3H]dopamine release (Fig. 2A). Co*+ (2 mM) alone produced a gradual increase in spontaneous [3H]dopamine release. Recognizing that Co’+ may have other effects in addition to its Ca*+ channel blocking action, we chose Cd’+, a more potent and presumably specific Ca*+ channel blocker, to determine if the Co’+ block of Hg’+ occurred at the voltagedependent Ca’+ channel. To this end, we titrated the Cd2+ concentration to give a similar depolarization-dependent block of [3H]dopamine release as was observed with 2 mM Co2+ (approximately 95%; data not shown). However, this concentration of Cd2+ ( 100 PM) failed to block the release of [3H]dopamine induced by 2 PM Hg2+ (Fig. 2B). Cadmium alone had no effect on the spontaneous release of [3H]dopamine (Fig. 2B).
Efect of Hg”
on Fura-
Fluorescence
Synaptosomes incubated with 10 PM fura2/AM for 45 min hydrolyze 90-95% of the ester (Daniel1 et al., 1987). This finding was confirmed in this laboratory by incubating synaptosomes for various times and determining R,,, . The rise in R,,,, corresponds to fura- formed. R,,, increased with time and plateaued after about 45 min. Addition of NaOH after R,,, stabilized produced no further increases in R,,,. The mean R,,, in these experiments was 5.9 and ranged between 4.17 and 7.8 1. These values are similar to those reported to give stable [Ca*+]i readings in synaptosomes (Komulainen and Bondy, 1987).
320
HARE ET AL
I
1
I
I
I
I
II
I
II
7 8 9 10 FRACTIONNUMBER(15 set each) 2
3
4
5
6
FIG. 1. The effect of Hg” on spontaneous and depolarization-dependent [3H]dopamine release. Release of [‘Hldopamine is expressed as a percentage of total release plus that remaining on the filter. Beginning with fraction 5. the superfusion buffer contained 5 pM Hg*+ (dashed and dotted lines). During fraction 6. synaptosomes received a I S-set pulse of superfusion buffer containing 6 I mM KCI (solid and dashed lines). In one run, 5 WM Hg” only was given without a high K’ pulse (dotted line). The Hepes buffer contained 0.25 mM Ca”. Values represent the means f SEM of five or six animals.
In synaptosomes incubated for 10 min in 1.2 mM Ca2+, the [Ca”]i was significantly elevated relative to synaptosomes incubated in nominal Cal+ buffers (p < 0.001) (Table 1). Exposure to 1,2,4, and 8 FM Hg2+ for 5 min produced no statistically significant effect on [Ca2+]i irrespective of [Ca2+lo (p > 0.05) when analyzed using two-way ANOVA (Table 1). However, when [Ca*+]i of synaptosomes in nominal Ca’+ media was analyzed using oneway ANOVA a significant decrease in [Ca2+]i was revealed (p < 0.001). Post hoc analysis revealed that both the 4 and the 8 I.LM Hg*+ groups were significantly different from controls (p < 0.05). Fluorescence of fura- in synaptosomal suspensions originates from both intrasynaptosomal and extrasynaptosomal sources.
Generally, extrasynaptosomal fura- fluorescence is differentiated by quenching the signal with Mn*+ (Komulainen and Bondy, 1987). The remaining fluorescence is presumed to be intrasynaptosomal. In nominal Ca2’ media, 40 PM Mn*+ quenched 37% of the fura- fluorescence at the 340-nm excitation wavelength (Fig. 3A). In the presence of 1.2 mM Ca*+, the Mn2’ quench of fura- fluorescence at 340 nm was reduced to 24%. Likewise, at all concentrations of Hg*+ in 1.2 mM Cal+-containing buffer, the percentage of fluorescence quenched by Mn*+ was reduced relative to the nominal Ca2+ buffer (p < 0.0 1). Exposure to 8 VM Hg2+ increased the percentage fluorescence that was quenched by Mn*+ irrespective of the [Ca’+],, (a < 0.01) (Fig. 3A). When [Ca2’10 was nominal, both 4 and
MERCURY-INDUCED 6-
I I
I
I
I
I
I
I,
DOPAMINE
I
I
I
RELEASE AND CALCIUM
I
I
A
FRACTION NUMBER (30 set
I
!
I
I
I
321
HOMEOSTASIS
I
f
II
11
1
6
each)
FRACTION NUMBER (30 set
each)
FIG. 2. The effect of Hg”, Co’+. and Cd”, alone or in combination, on spontaneous [lH]dopamine release from superfused striatal synaptosomes. Exposure to the various treatments began with fraction 3 and continued through fraction 12. Data are expressed as in Fig. I, except the fractions are 30 set each. Buffers contained no added Ca”. (A) The effect of 2 pM Hgzc on spontaneous release in the presence and absence of 2 mM Co’+. (B) The effect of 2 KM Hg’+ in the presence and absence of 100 PM Cd’+. Values represent the means +_SEM of four or five animals. The curves for control and 2 PM Hg’+ in A and Bare identical.
8 PM Hg’+ produced statistically significant increases in the percentage of Mn*+ quenching relative to controls (p < 0.0 1). It is assumed that the synaptosomal membrane is relatively impermeable to Mn*+ during the time the fluorescent measurements are taken. Thus, the fluorescence that is quenched by Mn2+ is limited to extrasynaptosomal fura-2. However, it must be considered that Hg*’ may permeabilize the synaptosomal membranes, not only allowing fur-a-2 to leak out, but also increasing the accessibility of intrasynaptosomal fura- to Mn2+. Consequently, the increase in Mn2+ quenching could be intrasynaptosomal. To test this possibility a separate set of experiments was conducted. Synaptosomes were exposed to Hg2+ as before but instead of quenching with Mn2+ at the end of exposure, 1-ml aliquots of the synaptosomes were centrifuged (7OOOg; 2
min) in a Fisher microcentrifuge. An aliquot of the supematant (0.5 ml) was then added to 1.5 ml of 1.2 mM Ca2+ incubation buffer and the fluorescence was determined as described previously. Fluorescence at 340 nm increased relative to controls at Hg2+ concentrations of 4 and 8 PM (p < 0.01) independent of the [Ca2+10 (Fig. 3B). It should be noted that 8 PM HgZf increased the Mn2+ quench in nominal and 1.2 mM Ca2+ media by 81 and 58%, respectively, over controls (Fig. 3A). However, this concentration of Hg2+ increased extrasynaptosomal fura- in nominal and 1.2 mM Ca2+ media by only 52 and 39%, respectively, over controls (Fig. 3B). That is, the increase in Mn2+ quench observed after exposure to 8 PM Hg2’ is only partially accounted for by an increase in extrasynaptosomal fura-2. This discrepancy is likewise apparent for other Hg’+ concentrations and suggests that a com-
322
HARE ET AL. TABLE I
LDH present in the assay. Alterations in LDH activity therefore reflect alterations in EFFECTOF Hg” ON~NTKASYNAPTOSOMAL~ONIZED the quantity of LDH. Since Hg2+ is a potent Ca’+ INRATSTRIATALSYNAPTOSOMES inhibitor of a variety of enzymes, this cannot [Ca’+], nM be assumed a priori. Therefore, Hg2+-inW+ 1 duced decreases in LDH activity, even in the I.2 mM Ca’+ ” Nominal Cal+ ” PM presence of 40 PM cysteine in the assay mix, 0 176k 5(12=7) 474 * 42. (n = 3) must be considered suspicious. For this reaI 185~l:!(rz=3) 533 * 70 (n = 3) son, protein leakage from synaptosomes was 3 178k 13(n=3) 458 + 14 (n = 3) also examined. The quantitation of cytosolic 4 117 k 16 (II = 3)’ 425 i 72 (n = 3) protein using the Coomassie blue method 471 +37(r)= 3) 8 108k 8(n=4)’ was insensitive to inhibition by up to 10 PM Hg’+ (data not shown). Approximately 23% Note Values represent the means f SEM for three to of the total Triton X- 100 releasable protein seven trials(n). 0 Nominal Ca’+ refers to the residual buKer [Ca”] was released from synaptosomes incubated since EGTA was not used. Measurements using furafor 10 min in physiologic Hepes buffer (Fig. reveal this concentration to be between 0.5 and 1 PM. 4B). Mercury had no statistically significant ‘Statistically significant from nominal Ca” (11 effect on synaptosomal leakage of cytosolic < 0.001). proteins (Fig. 4B). However, total protein re” Statistically significant from nominal Ca” controls (a < 0.001) in a one-way ANOVA. leased by membrane disruption with Triton X-100 was significantly decreased by 8 pM Hg”+ both in nominal (p < 0.05) and in 1.2 ponent of the Mn2+ quench is intrasynaptomM Ca” (p < 0.0 1) buffers. somal. These findings do not eliminate the possibility that Hg” increases the leakiness of the plasma membrane to smaller molecules such Determination qf Synaptosomal Integritli as [3HJdopamine and fura-2. To address this possibility the effect of Hg” on [3H]deoxyThe Hg’+-induced leakage of fura- from glucose 6-phosphate (t3H]dgluP) leakage synaptosomes suggests that Hg’+ may profrom superfused synaptosomes was assessed. duce nonspecific membrane damage. To test This molecule is sensitive to toxicant-infor this we monitored the leakage of LDH duced alterations in membrane permeability and protein from synaptosomes exposed to (Walum, 1982) and has a molecular weight Hg2+ for 5 min. In control synaptosomes ap- comparable to dopamine and fura-2. [3H]proximately 20% of the total Triton X- 100 re- dgluP was spontaneously released from synleasable LDH was extrasynaptosomal, irre- aptosomes at the rate of approximately 1% of spective of the [Ca’+lo. Exposure to 8 PM the total synaptosomal [3H]dgluP pool per 30 Hg’+ for 5 min produced a statistically sig- set (Fig. 5). When 2 mM Co’+ was added to nificant decrease in the leakage of LDH, but the superfusate beginning at fraction 3. there only in nominal CG’ (p < 0.05) (Fig. 4A). was no change in [3H]dgluP efflux. However, Synaptosomes exposed to Hg’+ for 5 min and addition of 2 PM Hg*+ increased [3H]dgluP resubsequently lysed with Triton X-100 re- lease by almost sixfold. Hg” (2 mh/r) and Co?-’ leased approximately fivefold more LDH (2 mM) together increased [3H]dgluP efflux simthan nonlysed samples. There was no sta- ilar to that produced by 2 pM Hg2+ alone. tistically significant effect of Hg’+ or Ca2+ IZjflux of 4-‘Ca2‘.fiom Supe&sed Synaptosomes (Fig. 4A). Mercury caused a concentration-depenThe specific activity of LDH normally is assumed to be proportional to the quantity of dent increase in the efflux of preloaded 45Ca’i
,*
4**
m
Mn QUENCH (% total
fluorescence)
324
8
600 -
= 5J
400 -
v) 5 d
HARE
ET AL.
500-
300 200 100
Intact \
_
I 0
,
I 2
,
I
4
/
/
/
8
0
HgC$ Q-M/l,
I
I
2
I
I
I
4
1
8
HgC12 (PM)
FIG. 4. Effects of Hg’+ on synaptosomal LDH activity and protein concentration in nominal Ca2+ (solid lines) and I .2 mM Ca*+ (dashed lines) buffers. (A) Extrasynaptosomal LDH after 5-min exposures to Hg*’ (solid circles) or Hg’+ exposure followed by 0.23% (final concentration) Triton X-100 (solid squares). U = one International unit. (B) Extrasynaptosomal protein concentration after exposure to Hg’+ for 5 min (solid circles) or Hg2+ followed by 0.23% Triton X-100 (solid squares). Values represent the means 2 SEM of three to five animals. Statistically significant from controls: *p < 0.05: **p -c 0.0 I.
from striatal synaptosomes superfused in nominal Ca2+ buffers (Fig. 6). Removal of the Hg2+ resulted in a decrease in 45Ca’+ release rates but at no concentration did the release rates return to preexposure levels within the time frame employed. To determine if the stimulation of 45Ca2+ efflux due to Hg*+ was relevant to the observed increases in transmitter release, we tested the effects of the inorganic Ca2+ channel blockers Co*+ (2 mM) and Cd2+ (100 PM). Synaptosomes exposed to 2 PM Hg2+ released 45Ca2+ at an accelerated rate (Fig. 7). Cobalt alone decreased 45Ca2t efflux to less than that of controls and the efhux did not return to control levels as long as the Co*+ was present (Fig. 7A). Mercury and Co2+ together increased “Ca2+ efflux to a level less than that observed for Hg2+ alone. Analysis of area under the curve revealed that the decrease in 45Ca2+ efflux produced by Co*+ alone was of the same magnitude as the attenuation of the Hg2’-induced 4sCa2+efflux by Co2+ (p > 0.05). That is, Co*+ decreased
45Ca’+ efflux in both instances by the same amount. Cadmium, on the other hand, increased 45Ca2+ efflux which persisted throughout exposure (Fig. 7B). The presence of Cd’+ and Hg*+ together exacerbated the increase in 45Ca2+ efflux relative to that produced by Hg’+ alone (Fig. 7B). As in Fig. 7A, analysis of area under the curve revealed that the increase in 45Ca2’ efflux by the combination of Hg’+ and Cd2’ relative to that produced by Hg*+ alone was of the same magnitude as the increase in 45Ca2+efflux produced by Cd*+ alone relative to controls (a > 0.05) (Fig. 7B). DISCUSSION Inorganic mercury increases [3H]dopamine release from rat striatal synaptosomes. The mechanisms mediating this effect are not likely due to an action of Hg2+ on either the plasma membrane Na+,K+-ATPase or the Nat-dependent high-affinity dopamine
MERCURY-INDUCED
DOPAMINE
RELEASE AND CALCIUM
trec4tmant
I I
1
I
HOMEOSTASIS
1
I
I
I
I
I
I
I
I
I
I
:
325
.
2 3 4 5 6 7 8 9 10 11 12 FRACTION NUMBER (30 set each)
FIG. 5. The effects of2 pM Hg’+ (Hg) and 2 mM Co*+ (Co) on the release of [3H]deoxyglucose 6-phosphate from rat striatal synaptosomes. The metals were given during the phase marked “treatment” (fractions 312) except for control. The superfusing buffer contained no added Ca*+. Other experimental conditions as described under Methods. Values represent the means ? SEM of four to five animals.
transporter (Hare et al., 1989). In the present report we suggest that Hg2’ may have direct effects on [3H]dopamine release presumably near presynaptic “active zones” where exocytosis of neurotransmitter originates. Depolarization-dependent release of [3H]dopamine from rat striatal synaptosomes is tightly coupled to [Ca2+],, below 1 mM (Drapeau and Blaustein, 1983). That is, decreases in [Ca2+10 result in parallel decreases in depolarization-dependent [3H]dopamine release. Moreover, a number of inorganic cations decrease the depolarization-dependent uptake of 45Ca2+ and [3H]dopamine release (Drapeau and Blaustein, 1983; Nachshen, 1984). Therefore, depolarization-dependent [3H]dopamine release from striatal synaptosomes in low [Ca2+10 is limited primarily by Ca2+ en-
try. Since 5 PM Hg2+ had no effect on depolarization-dependent [3H]dopamine release when [Ca2’10 was 0.25 mM, the fast Ca2+ channels probably are not involved in the Hg2+-induced increases in spontaneous [3H]dopamine release. This observation is in agreement with others who have observed that low concentrations of Hg2+ (~10 PM) have no effect on the depolarization-dependent uptake of 45Ca2+ by synaptosomes (Nachshen, 1984). It was surprising then that the Ca2+ channel blocker Co2+ was effective in blocking Hg2+induced [3H]dopamine release. Co2+ (2 mM) blocked approximately 95% of the depolarization-dependent [3H]dopamine release. However, Co’+ is known to have other effects on nerve terminals, particularly on the so-
HARE ET AL.
326 I
I
I
I
I
I
I
1
2
3
I
I
I
I
I1
I
I
I
I
I
I
4
S
6
7
8
3
I
Ill-i
I
III
87-
I
10 11 12 13 14
FRACTION NUMBER (30 set each) 6. Mercuric ion-induced efflux ofpreloaded 45Caz+ from superfused striatal synaptosomes. 45Ca’i efflux is expressed as a percentage of total 45Ca2+released plus that remaining on the filter. There was no added Ca’+ in the superfusing buffers. There was a concentration-dependent increase in 4sCa2+efflux when Hg’+ was given during fractions 3-5. Values represent the means + SEM of four to six animals. FIG.
called slow Ca’+ uptake during depolarization (Nachshen, 1984). Cadmium (100 PM; which also blocked 95% of the depolarization-dependent [3H]dopamine release) failed to block the Hg2+-induced spontaneous release of [3H]dopamine. This effectively eliminates the involvement of the fast Ca2+ channels in Hg*+-induced spontaneous [‘Hldopamine release. Cadmium also is ineffective in inhibiting Pb’+-induced increases in spontaneous acetylcholine release from rat brain synaptosomes (Suszkiw er al., 1984). These authors suggest that Pb*+ acts at the plasma membrane to augment the exocytotic release of acetylcholine since Pb2+ has minimal effect on 45Ca’+ movements in intracellular organelles. Also, membrane impermeant chelators such as EDTA and dimercaptosuccinic acid
rapidly reverse the increase in MEPP frequency produced by low [Pb*‘] (Kolton and Yaari, 1982). Conceivably, Pb2+ and Hg*+, which have similar effects on spontaneous [3H]dopamine release and 4SCa2+efflux from synaptosomes, have common sites of action. Increases in [Ca2+]i are considered obligatory for the exocytotic release of [3H]dopamine from synaptosomes (Drapeau and Blaustein, 1983). However, there was no increase in [Ca2+]i in synaptosomes exposed to Hg*+. Therefore, if the [“Hldopamine released by Hg2+ originated from synaptic vesicles, Hg’+ may act directly to stimulate this release. That is, Hg*+ may bind to sites at or near the plasma membrane in the so-called active zone. Upon binding, Hg*+ may alter the configuration of plasma membrane pro-
MERCURY-INDUCED
DOPAMINE
RELEASE AND CALCIUM
327
HOMEOSTASIS
2 mM Co trsotmont
I
12
3
4 5
6
7
4
8 9 101112
FRACTION NUMBER (30 set
each)
12
3
4
5 6
7 8
5 101112
FRACTION NUMBER (30 set
each)
FIG. 7. ‘%a’+ efflux from preloaded, superfused synaptosomes exposed to Hg’+, Co”, or Cd”. Experimental conditions as in Fig. 6 except synaptosomes were exposed to the metals during fractions 3- 12. The superfusate contained no added Ca’*. (A) Effect of 2 mM Co’+ and 2 MM Hg’+ alone or in combination on 4sCa’i efflux. (B) Effect of 100 ,uM Cd’+ and 2 PM Hg’+ alone or in combination on 45Ca’+ etIIux. Values represent the means f SEM ofthree to five animals
teins to such an extent that transmitter molecules are released from juxtaposed synaptic vesicles. It has been postulated that sulfhydryls may directly regulate the release of acetylcholine from neuromuscular junctions (Kosower and Werman, 1971). Likewise, in synaptosomes, plasma membrane impermeant sullhydryl reagents release [3H]acetylcholine independent of [Ca’+10 (Baba et al., 1979). The observation that Hg2+ decreased [Ca’+]i in synaptosomes in nominal Ca’+ media (Table 1) is analogous to the effect of (Ylatrotoxin observed in guinea pig synaptosomes (Meldolesi et al., 1984). This effect was thought to be due to the net efflux of Ca2+ from the synaptosomes, since there was a concomitant increase in quin-2 efflux suggesting plasma membrane damage had occurred. The net efflux of Ca’+, presumably responsible for the decline in [Ca”]i, occurred because of the presence of 100 PM EGTA in
the extrasynaptosomal media, which is known to lower resting [Ca’+]i in synaptosomes. However, in the present experiments, [Ca’+10 was between 0.5 and 1.O PM and it is therefore unlikely that Ca2’ would exit synaptosomes via this mechanism. The interpretation of this observation therefore remains unclear. Mercury (4 and 8 PM) increased the efflux of fura- from synaptosomes. This effect likely was not due to gross membrane damage since no increases in LDH or protein release from synaptosomes were observed. However, Hg2’ did increase the permeability of the plasma membrane to [3H]dgluP. Evidently, the increased permeability of the plasma membrane induced by Hg*+ is selective for low molecular weight compounds (i.e., <140,000 Da, the molecular weight of LDH). These membrane permeabilizing effects of IIg2+ are apparently independent of its trans-
328
HARE
mitter releasing action. This is based on the observation that Co’+, which totally blocked Hg2+-induced [3H]dopamine release, did not alter Hg*+-induced 4sCa2+ and [3H]dgluP efflux. The membrane permeabilizing effect of Hg2+ would predict an increase in the leakage of cytosolic [3H]dopamine since a sizable component of synaptosomal t3H]dopamine resides in the cytosol, particularly when its metabolism is inhibited by monoamine oxidase inhibitors. Furthermore, the Naf-dependent, high-affinity dopamine transporter, which mediates a component of ouabainand low NG-induced [3H]dopamine release (Raiteri et al., 1979; Hare et al., 1989) provides functional evidence of a cytosolic [3H]dopamine pool in striatal synaptosomes. Alternatively, it could be that Hg” produces at least two types of plasma membrane lesions, one specific for t3H]dopamine, the other specific for [3H]dgluP. Co2+ could be effective in blocking only the t3H]dopamine specific lesion with no action of the lesions responsible for [3H]dgluP leakage. In this scenario, Co” acts to block the efflux of [“Hldopamine rather than hinder access of Hg’+ to release sites. As noted under Results, the Hg2+-induced increase in Mn2+ quench is greater than can be accounted for by the leakage of fura-2. It is likely that Mn’+ quenches some intrasynaptosomal fura-2. The ability of Mn2+ to permeate the plasma membrane of synaptosomes (Komulainen and Bondy, 1987) and of cultured rat cerebellar granule cells (Connor et al., 1987) to quench fura- fluorescence is recognized. However, this is not a serious dilemma since [Ca”]i can be calculated independently of the concentration of furainside the synaptosomes. In the presence of 1.2 mM Ca’+, the Mn2+ quench was less than that observed in nominal Ca’+ (Fig. 3A). Superficially, this suggests an antagonistic effect of Cd+ on fura- efflux. However, fluorescence intensity from supernatant fura- was similar, irrespective of the [Ca2+10 during incubation (Fig. 3B). Altematively, in the presence of 1.2 mM Ca’+, the
ET AL.
quenching efficiency of Mn2+ could be reduced since it competes with Ca2+ for binding to fura-2. However, higher concentrations of Mn2+ failed to significantly increase quenching in 1.2 mM Ca2+ media (data not shown). It seems unlikely that cation competition for binding to fura- could completely explain these data. Another possible explanation could be that in 1.2 mM Cd+, Mn2+ is less effective in permeating the plasma membrane than it is in nominal Ca2+ media since it competes with Ca’+ for entry into synaptosomes. This would effectively reduce the intrasynaptosomal quench in 1.2 mM Cd’. Most likely, the decrease in Mn2+ quenching in 1.2 tnM Ca2+ media results from a combination of these factors. Ca’+ has membrane stabilizing properties and competes with Mn” for binding to fura- and entry into synaptosomes. Mercury produced a concentration-dependent increase in 45Ca2+ efflux from striatal synaptosomes. The efflux of preloaded 45Ca2t from cells has been used as an indirect indicator of changes in Ca” status (Pounds, 1984). The inference is that the observed increase in 45Ca2t efflux induced by Hg’+ reflects 45Ca2+ mobilization from intraterminal stores resulting in an increase in resting [Ca2+]i. Presumably, the synaptosome responds to changes in [Ca2+li by extruding Ca2+ from the terminal (Nachshen et al., 1986). However, on the basis of our measurements using fura2, Hg” did not increase [Ca2+]i and, in nominal Ca2+ media, actually decreased [Ca’+]i at Hg’+ concentrations of 4 and 8 pM. The fact that 2 pM Hg’+ increased 45Ca2+ efflux without measurable changes in [Ca’+]i can be explained as follows. In the Hepes buffers used in our experiments, in which no Ca” was added (nominal Ca’+), the [Ca’+k, was likely greater than 1 PM. Under these conditions, the [Ca’+]i was less than 0.2 pM. Hg2+-induced membrane permeabilization could have increased the influx of 40Ca2f in the direction of the Ca” electrochemical gradient. This increased influx of 40Ca2+ was apparently efficiently buffered since no increase
MERCURY-INDUCED
DOPAMINE
RELEASE AND CALCIUM
in resting [Ca*+]i occurred. Since the kinetic balance between Ca2+ influx and Ca2+ efflux regulates synaptosomal [Ca2’], (Nicholls, 1986), the increased efflux of 45Ca2+ by Hg2+ suggests that the influx of 40Ca2t was accelerated, mixed with intrasynaptosomal 45Ca2+ and both 40Ca2f and 45Ca2+ extruded via plasma membrane extrusion mechanisms. The observation that Co2+ was ineffective at inhibiting the Hg2+-induced efflux of both [3H]dgluP and 45Ca’f is congruous with this hypothesis. In conclusion, Hg” increased [3H]dopamine release from rat striatal synaptosomes. This effect is hypothesized to be due to an action of Hg” at or near the plasma membrane at a site that also interacts with Co” but not Cd*+. The increase in spontaneous [‘Hldopamine release occurred without concomitant elevations in [Ca’+]i. Furthermore, Hg*+ had other effects, in particular an increase in the permeability of the plasma membrane to small molecules. The increase in [3H]dopamine release was apparently not mediated via this nonspecific leakage pathway since it was blocked by Co”+ which had no effect on the leakage of other small molecules tested. ACKNOWLEDGMENTS The authors express their appreciation to Dr. Steven Leslie. Department of Pharmacology, University of Texas at Austin. for allowing M.F.H. the use of laboratory space, equipment, and supplies during the conduct of experiments measuring [Ca*‘],, Robert Greenland and Marlene Jaeger for their technical expertise, and Annette Townsley for preparation of the manuscript. Funds for this work were provided by NIEHS Grant ES-03992.
REFERENCES AUGUSTINE, G. J.. CHARLTON, M. P.. AND SMITH, S. J. (1987). Calcium action in synaptic transmitter release. Annu. Rev. Neurosci. 10, 633-693. BABA, A., FISHERMAN, J. S., AND COOPER, J. R. (1979). Action of sulfhydryl reagents on cholinergic mechanisms in synaptosomes. Biochnn. Pharmacol. 28, 1879-1883. BERGMEYER, H. U.. AND BERNT, E. (1974). Lactate dehydrogenase: UV-assay with pyruvate and NADH. In
HOMEOSTASIS
329
Methods of Enzymatic Analysis (H. U. Bergmeyer. Ed.), pp. 574-576. Academic Press, New York. BINAH, O., MEIRI, U.. AND RAHAMIMOFF, H. (1978). The effects of HgClz and mersalyl on mechanisms regulating intracellular calcium and transmitter release. Eur. J. Pharmacol. 51,453-457. BRADFORD, M. (I 976). A rapid and sensitive method for the quantification of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72,248-254. CHEUNG, W. Y. (I 984). Calmodulin: Its potential role in cell proliferation and heavy-metal toxicity. Fed. Proc. 43,2995-2999. CONNOR, J. A.. TSENG, H.. AND HOCKBERGER, P. E. ( 1987). Depolarizationand transmitter-induced changes in intracellular Ca*+ of rat cerebellar granule cells in explant cultures. J. Neurosci. 7, l384- 1400. COOPER, G. P., AND MANALIS. R. S. (1983). Influence of heavy metals on synaptic transmission: A review. Nerrroto.xico[ogy, 4,69-84. COTMAN. C. W. (1974). Isolation of synaptosomal and synaptic plasma membrane fractions. In Methods in Enzymology 31A (L. Fleischer and L. Packer, Eds.), pp. 445-452. Academic Press. New York. DANIELL, L. C., BRASS,E. P., AND HARRIS, R. A. (1987). Effect ofethanol on intracellular ionized calcium concentrations in synaptosomes and hepatocytes. Mol. Pharmacol. 32,83 l-837. DRAPEAU, P., AND BLAUSTEIN, M. P. (1983). Initial release of [3H]dopamine from rat striatal synaptosomes: Correlation with calcium entry. J. Newosci. 3. 703713. GRYNKIEWICZ, G., POENIE, M.. AND TSIEN, R. Y. (1985). A new generation of Ca” indicators with greatly improved fluorescence properties. J. Biol. Chem. 260,3440-3450. HARE, M. F., MINNEMA, D. J., COOPER. G. P.. AND MICHAELSON. I. A. (I 989). Effects of mercuric chloride on [‘Hldopamine release from rat brain striatal synaptosomes. Tosicol. Appl. Pharmucol. 99,266-275. HOOD, W. F., AND HARRIS, R. A. ( 1980). Effects of depressant drugs and sulfhydryl reagents on the transport of calcium by isolated nerve endings. Biochem. Pharmacol. 29,957-959. JUANG. M. S. (1976). An electrophysiological study of the action of methylmercury chloride and mercuric chloride on the sciatic-sartorius muscle preparation of the frog. Toxicol. Appl. Pharmacol. 37,339-348. KOLTON. L., AND YAARI, Y. (1982). Sites of action of lead on spontaneous transmitter release from motor nerve terminals. IX. J. Med. Sci. 18, 165-170. KOMULAINEN, H., AND BONDY, S. C. (I 987). The estimation offree calcium within synaptosomesand mitochondria with fura-2: comparison to quin-2. Neurothem. Int. 10,55-64.
330
HARE ET AL
KOSOWER, E. M., AND WERMAN. R. (197 I). New step in transmitter release at the myoneural junction. Narwe New Bid 233 . 13 - 1- 122. MANALIS, R. S., AND COOPER, G. P. (1975). Evoked transmitter release increased by inorganic mercury at the frog neuromuscular junction. Nature (London) 251,690-69 1. MELDOLESI, J., HUTTNER, W. B., TSIEN, R. Y.. ANI> POZZAN, T. ( 1984). Free cytoplasmic Ca*+ and neurotransmitter release: Studies on PC 12 cells and synaptosomes exposed to cy-latrotoxin. Ptw. Nafl. .-lead. Sci. US.4 81,620-624.
MINNEMA. D., AND MICHAELSON. 1.A. (1985). A superfusion apparatus for the examination of neurotransmitter release from synaptosomes. J. Neworr~i. Mrthads 14, 193-206.
MINNEMA. D.. AND MICHAELSON, 1. A. (1986). Differential effects of inorganic lead and &aminolevulinic acid in vilro on synaptosomal &aminobutyric acid release. To.ricol. .dppl. Phartnc~ol. 86, 437-447. NACHSHEN. D. A. (1984). Selectivity of the Ca binding site in synaptosome Ca channels: Inhibition of Ca influx by multivalent metal cations. J. Cm. Phv.sio/. 83, 941-967. NACHSHEN, D. A.. SANCHEZ-ARMASS. S.. AND WEINSTEIN, A. M. (1986). The regulation of cytosolic calcium in rat brain synaptosomes by sodium-dependent calcium efflux. J. Ph?:yio/. 381, 17-28.
NICHOLLS, D. G. (1986). Intracellular calcium homeostasis. Brit. Med. Bull. 42.353-358. POUNDS.J. (I 984). Effect oflead intoxication on calcium homeostasis and calcium-mediated cell function: A review. Ntwotoxicology 5,295-332. RAITERI, M., CERRITO, F.. CERVONI, A. M., AND LEVI. G. (I 979). Dopamine can be released by two mechanismsdifferentially affected by the dopamine transport inhibitor nomifesine. .I. Pharmacol. Ekp. Ther. 208, 195-202.
SMITH, M. W.. AMBUDKAR, 1. S., PHELPS, P. C.. REGFC. A. L.. AND TRUMP, B. F. (1987). Hgr&induced changes in cytosolic Ca’+ of cultured rabbit renal tubular cells. Biochim. Biophys. &la 931, I 30- 142. SNEDDON,J. (1987). Action of di- and tri-valent cations on calcium-activated K’-efflux in rat erythrocytes. Biwhetn.
Phurma&.
36, 3723-3730.
STEEL, R. G. D.. AND TORRIE. J. H. (1960). Principles and Procedures of .Qulis/ics. McGraw-Hill. New York. SUSZKIW. J., TOTH, G.. MURAWSKY, M.. AND COOPER, G. P. ( 1984). Effects of Pb2+ and Cd’+ on acetylcholine release and Ca2+ movements in synaptosomes and subcellular fractions from rat brain and Torpedo electric organ. Brain Res. 323, 3 l-46. WALUM, E. ( 1982). Membrane lesions in cultured mouse neuroblastoma cells exposed to metal compounds. To.~i~olog~
25,67-74.
AND
APPLIED
PHARMACOLOGY
102,3
16-330
(1990)
Effects of Inorganic Mercury on [3H]Dopamine Release and Calcium Homeostasis in Rat Striatal Synaptosomes’ F. HARE,‘,~ S. MEHDI REZAZADEH,~ GARY P. COOPER, DANIEL J. MINNEMA. AND I. ARTHUR MICHAELSON
MICHAEL
Effects of Inorganic Mercury on [3H]Dopamine Release and Calcium Homeostasis in Rat Striatal Synaptosomes. HARE, M. F.. REZAZADEH, S. M., COOPER, G. P.. MINNEMA. D. J.. AND MICHAELSON, 1. A. ( 1990). 7i,\-icol. .4p/~/. Plrurmuc~ol. 102. 3 16-330. Inorganic mercury (Hg”) in vitro increases spontaneous transmitter release from nerve terminals. The mechanisms of action are not well understood but may involve alterations in intraterminal Ca” dynamics, In this study we describe actions of Hg’+ in vitro on isolated mammalian CNS striatal nerve terminals (synaptosomes). Cobalt (2 mM) completely blocked the effect of 2 +M Hg’+ on spontaneous [‘Hldopamine release. Cadmium (100 FM) was equipotent to Co*’ in blocking depolarization-dependent [3H]dopamine release, but did not alter the 2 PM Hg’+-induced spontaneous [‘Hldopamine release. Depolarization-dependent [‘Hldopamine release was not altered by 5 MM Hg’+. It appears that the site of action of Hg’+ on spontaneous [3H]dopamine release is not the Ca” channel. The effects of Hg” on intraterminal ionized Calf ([Ca”],) were evaluated using the Ca”-specific fluorescent probe. fura-2. Hg” (1-8 KM) had no effect on [Ca”], in 1.2 mM Ca”-containing buffers. In nominal Ca”’ media. 4 and 8 pM Hg’+ significantly decreased [Ca”],. Following exposure to 4 and 8 NM Hg’+ the quenching of extrasynaptosomal fura-? by Mn” was increased. suggesting that Hg’+ facilitated the leakage offura-2. This apparent leakage was probably due to a nonspecific increase in membrane permeability since 2 pM Hg*’ produced a Co”-insensitive increase in [3H]deoxyglucose phosphate efflux. Hg” (2-8 KM) did not increase the leakage of either lactate dehydrogenase or soluble protein from synaptosomes. Hg’+ produced a concentration-dependent ( l-8 pM) increase in 45Ca’ ’ efflux from superfused synaptosomes which was insensitive to blockade either by 2 mM Co’+ or by 100 NM Cd’+. These data suggest that the transmitter releasing action of Hg’+ involves interactions with sites that also interact with Co” but not with Cd’+. Furthermore, Hg” may have direct transmitter releasing actions (i.e.. Ca’+-mimetic properties). as well as nonspecific actions on plasma membrane permeability which may not necessarily be linked to [3H]dopamine release. $, IWO Academ,c Prers. Inc.
Inorganic mercury (Hg”+) increases spontaneous transmitter release from both the frog neuromuscular junction and the rat brain synaptosomes (Manalis and Cooper, 1975; Juang, 1976; Hare et al., 1989). The mechanism(s) responsible for these effects has been hypothesized to result from either (a) alterations in intracellular calcium ion (Ca2+) reg-
’ Portions ofthis manuscript have been reported in abstract form in To\-icohgist. 1989, 9(I). 97 [Abstract No. 3851. Supported by ES-03992. ’ To whom correspondence should be addressed. ’ Present address: Department of Pharmacology & Toxicology, Michigan State University, East Lansing.
MI 48824- I3 17. ’ Division of Pharmacology, University of Texas at Austin.
004 1-008x/90
College of Pharmacy. Austin. TX 787 12.
$3.00
Copyright Q 1990 hy Academic Prrss, Inc. All rights of rcproductmn in any form reserved
316
MERCURY-INDUCED
DOPAMINE
RELEASE
ulation (Binah et al., 1978; Cooper and Manalis, 1983); and/or (b) Ca*+-mimetic actions of the mercuric ion (Cheung, 1984). It is well documented that alterations in the concentration of free ionized Ca*+ ([Ca*‘]i) in nerve terminals influence neurotransmitter release (Augustine et al., 1987). Perturbations in intraterminal Ca2+ homeostasis leading to increases in [Ca*+]i could mediate the observed increases in spontaneous neurotransmitter release following Hg*+. It is known that Hg’+ can alter intracellular Ca2’ homeostasis in a variety of cell types and organelles. For instance, Hg’+ (31 PM) decreases 45Ca’t uptake into isolated rat brain mitochondria (Binah et al., 1978). Mercuric chloride (10 FM) inhibits ATP-dependent “Ca2+. uptake by rat brain nonmitochondrial intrasynaptosomal particles, but has minimal effects on depolarization-dependent 4sCa’t uptake into intact rat brain synaptosomes (Hood and Harris, 1980). Mercuric chloride (a 10 PM) produces a dose-dependent increase in the efflux of s6Rb+ from rat erythrocytes (Sneddon, 1987). Presumably, the *6Rb+ efflux is through a Ca’+-sensitive K+ channel that is activated by elevated [Ca’+]i. Direct measurements of Hg”-induced [Ca”]i increases in cultured renal tubular cells with the Ca’j-specific fluorescent probe fura(Smith et al., 1987) suggest that Hg*+ may likewise increase [Ca”] in synaptosomes. Increases in transmitter release could then be mediated by Ca” -dependent vesicular exocytotic machinery. We present data that only partially support a hypothesis of altered intrasynaptosomal Ca’+ homeostasis by Hg”. It is likely that the effects of Hg’+ on nerve terminals are complex and involve multiple mechanisms that are both concentration and time dependent. Mercury-induced transmitter release in synaptosomes may involve a direct action of Hg’+ on transmitter release sites, whereas alterations in Ca’+ homeostasis may occur via other mechanisms which are not necessarily
AND
CALCIUM
HOMEOSTASIS
317
linked to the observed changes in transmitter release.
tein .4ssay
.4nimals and animal cure. Adult male Long-Evans rats ( 150-250 g, Charles River, Wilmington, MA) were housed two per cage. fed Laboratory Chow 5001 (Ralston Purina, St. Louis, MO), and provided with tap water ad libitum. All animal care procedures were in accordance with NIH guidelines. Preparation of.s.~frupfosonzes. 411 synaptosomal isolation procedures were performed at 3°C. “Purified” synaptosomes were prepared as described by Minnema and Michaelson ( 1986). Briefly. rat sttiatal homogenates were centrifuged at IOOOg for IO mitt, and the supemates layered onto a discontinuous 0.8/ I .2 M sucrose gradient and centrifuged at 223,0001: for 25 min. The synaptosomes were removed from the 0.8- 1.2 M sucrose interphase. diluted with 0.32 M sucrose. and kept on ice until used (~4 hr from termination). Protein was determined using the Coomassie blue method of Bradford (1976). Synaptosomes (240 pg protein adjusted to 200 ~1 with 0.32 M sucrose) were added to 800 (~1 of physiological Hepes buffer. In the 45Ca” efilux studies, the Hepes buffer contained 20 mM Na’ (replaced isoosmotically with choline chloride). The Hepes buffer consisted of (mM). I40 NaCI. 5 KCI, 1.2 MgS04, I I. I glucose. 8.5 Hepes, and 5 Trizma Base. In experiments using [3H]dopamine. 0.5 mM nialamide, a monoamine oxidase inhibitor was included. For the uptake of [3H]deoxyglucose. the Hepes buffer contained 1 mM glucose (replaced isoosmotically with sucrose). Following the addition of a 25-~1 aliquot of0.5 PCi of [3H]dopamine (26.7 Ci/mmol). 6.25 &i of [3H]deoxyglucose (30.2 Ci/mmol). or 2 PCi of 4’CaCI, (1.6 mCi/mmol) the synaptosomes were incubated for I5 min at 37°C under Oz/C02 (95%/5%) in a Dubnoff metabolic shaker. The “loaded” synaptosomes were layered onto a 0.65-pm pore-size cellulose acetate filter in a filter support chamber, connected to a “push-pull” design superfusion system (Minnema and Michaelson. 1985). and superfused (2.0 ml/min) for 14 min to remove extrasynaptosomal [-‘H]dopamine and [‘Hldeoxyglucase. In the 45Ca’+ efflux experiments, synaptosomes were first superfused for IO min with Hepes buffer containing 100 pM EGTA and then for 4 min in normal Hepes buffer. Fractions were then collected directly into liquid scintillation vials at timed intervals. Radioactivity was assessed in a Packard Model 460 liquid scintillation spectrometer.
318
HARE
Lactare dehydrogenase activity attdproteitl d~~tertnrnutirms. Synaptosomes (70 pg protein adjusted to IO0 ~1) were added to 380 ~1 Hepes buffer that contained either no added CaCI, (nominal Ca’+) or I .2 mM CaCl? and placed in a Dubnoff metabolic shaker (37°C) for 5 min. Hepes buffer (20 ~1) that contained [Hg’+] to give 2.4, or 8 pM (final concentration) was added. For controls, 30 pl of Hepes buffer only was added. Synaptosomes were then incubated for an additional 5 min. the reaction was terminated by centrifugation (8000x, 5 min). and the supernate was removed and stored (2°C) for later analysis. To determine total synaptosomal lactate dehydrogenase activity (LDH) and soluble protein. synaptosomes were exposed to Hg2+ as above for 5 min. lysed with 100 ~1 of I .4% Triton X- 100 (0.23% final concentration), and centrifuged (8OOOg; 5 min) and the supernatant was removed. Lactate dehydrogenase activity was determined using a modification of the method of Bergmeyer and Bernt (1974). The reaction mix contained (final concentration. mM) 41.2 KZHP04, 6.78 KH,PO,, 0.58 Na pyruvate. 0. I2 NADH. I.10 NaHCO,, and 0.04 L-cysteine (free base). Briefly. a 167-p] aliquot of supernatant was added to cuvettes containing the reaction mix. The final volume was 2.7 I ml. The conversion of NADH to NAD+ was monitored by a decrease in light absorption at 340 nm. The reaction proceeded for IO min at 37°C and specific activity of LDH was calculated. Extrasynaptosomal protein was determined using the Coomassie blue method (Bradford. 1976).
Experiments involving the measurement of intrasynaptosomal ionized Ca’+ ([Ca’+],) were conducted in the laboratory of Dr. Steven Leslie. Division of Pharmacology. College of Pharmacy. University of Texas at Austin. Adult male Sprague-Dawley rats (200-300 g. University of Texas, Animal Resources Center) were used in all experiments. Synaptosomes were prepared from striata according to the method of Cotman ( 1974). The synaptosomes were resuspended in 4 ml of buffer which consisted of (mM) I36 NaCI, 5 KCI, I .3 MgC& ~ IO glucose, and 20 Trizma Base, and the pH was adjusted to 7.5 with I M maleic acid. The protein concentration ranged from 6 to 7 mg/ml. The synaptosomal suspension was divided into two 2ml aliquots. poured into two glass vials, placed in a Dubnoff metabolic shaker. and equilibrated for IO min at 37°C. Fura-2/AM (I mM), dissolved in dimethyl sulfoxide (DMSO), was added to one ofthe vials containing the synaptosomal suspension (final concentration = IO PM). An equal volume of DMSO (I % final concentration) was added to the other vial. designated as unloaded preparation. and served as control. The suspensions were incu-
ET AL. bated for 20 min, diluted fivefold with 37°C incubation medium. and transferred to two 50-ml centrifuge tubes for an additional IO-min incubation. The tubes were then centrifuged (I 3,OOOg; IO min) and the pellets resuspended in 2 ml of ice-cold incubation medium and placed on ice. The protein concentration of the suspension varied between I .6 and 2.5 mg/ml. Aliquots (100 ~1) of fura-2-loaded synaptosomal suspensions or unloaded suspensions were resuspended in 2 ml (37°C) of buffer containing either nominal CaCl? or I.2 mM CaC& in polystyrene cuvettes. The [Ca’+]” in nominal Cal+ buffer was determined using free furaand ranged between 0.5 and I .O pM. The samples were gently vortexed and incubated for 5 min. For studying the effects of Hg*+ on the resting [Ca’+],. Hg*+ (l-8 PM) or buffer (for control samples) was added and incubated for an additional 5 min. Baseline fluorescence was then measured at 340- and 380-nm excitation wavelengths with the emission wavelength set at 505 nm, using a Spex Fluorolog Spectrofluorometer (Spex Industries. Edison. NJ). The slitwidth was set at 1 mm (3.6 nm bandwidth) for both excitation and emission. The temperature ofthe cuvette compartment was maintained at 37’C. The minimum fluorescence (F,,,,,) was determined by adding 50 ~1 of a 4% solution of sodium dodecyl sulfate (SDS) and 20 ~1 of 0.5 M EGTA (in 3 M Trizma Base). R,,, was calculated as the ratio of FAilO/FiIo. To determine F,,,a,,,. using a different sample, the tissue was lysed with SDS and then 20 ni of0.4 M Ca” solution (final concentration = 4 IIIM) was added. Under these conditions, the ratio of ~‘&~~x,, is known as R,,,. Dye leakage was assessed by adding 20 ~1 of4 tnM Mn2+ (final concentration = 40 FM) to the cuvettes after determining baseline fluorescence. Fluorescence signals were also corrected for autofluorescence. The [Ca’+], was calculated according to the method of Grynkiewicz rt al. ( 1985) [Ca”],
= &X
(R - R,,,/R,,,
- R) X (Sf,/Sb&
where Sfi is fluorescence of furaat minimal Ca’+ and SbZ is that at Cal+ saturation at the excitation wavelength 380 nm. k;, the dissociation constant of the fura-2-Ca’+ complex. is 224 nM (Grynkiewicz PI al.. 1985). Da/a hatzdlin!:at~dsfatis/ics. The amount ofradiolabel released per fraction is expressed as a percentage of the total radioactivity:
(total
dpm released per fraction X 100 dpm released + dpm remaining on filter)
The numberoftrials (n) is indicated in each ligure legend. One animal was used per trial. In the furaexperiments. the pooled striata from two animals were used per trial (n). One-way analysis of variance (ANOVA) was used to analyze the data from superfusion experiments when appropriate. Data from the LDH. protein. and furaexperiment were analyzed using either one-way or two-way
MERCURY-INDUCED
DOPAMINE
RELEASE
ANOVA (Steel and Torrie. 1960). Post hoc analyses were done using the Newman-Keuls test at p < 0.05 signiiicance level. In some experiments the number of replications per treatment varies. In these cases, appropriate alterations were made for both ANOVA and post hoc tests. c%emi~&. [jH]Dopamine (3.4-[8-3H(A’)]dihydroxyphenylethylamine, NET-094). [3H]deoxyglucose (2-I 1,2‘H]deoxy-D-glucose. NET-549A). 45CaC12 (NEZ-013). and Aquasol-? were obtained from New England Nuclear (Boston, MA). Ultrapure sucrose was obtained from Beckman (Palo Alto, CA). Fura-Z/AM (furaacetoxymethyl ester) was purchased from Molecular Probes, Inc. (Eugene, OR). Hepes, Trizma Base. nialamide. EGTA, NADH, pyruvate. HgCl?. CdCL. and CoC& were obtained from Sigma (St. Louis, MO). All other chemicals were of reagent grade and were obtained from available commercial sources. All solutions were prepared with double distilled water.
RESULTS Eflect of Hg’+ [‘H]Dopamine
on Depolarization-Evoked Release
Striatal synaptosomes superfused in physiological Hepes buffer containing 0.25 mM Ca’+ released approximately 1% of the total [3H]dopamine per 15set fraction (fractions l-4; Fig. 1). During fraction 6 synaptosomes were exposed to a 1 S-set pulse of 6 1 mM G (solid line: Fig. 1). This [K+], is sufficient to depolarize striatal synaptosomes which, in the presence of [Ca2+10 > 0.025 mM, produces an increase in the rate of [3H]dopamine release above baseline (Minnema and Michaelson, 1985). Depolarization-dependent [3H]dopamine release returned to baseline within 45 sec. Addition of 5 PM Hg*+ to the Hepes buffer (dashed line) beginning with fraction 5 followed 15 set later (fraction 6) by a depolarizing pulse of K’ resulted in a release of [3H]dopamine indistinguishable from that in the absence of Hg” (solid line). Hg’+ (5 /*M) alone produced an increase in spontaneous [3H]dopamine release with an onset latency of approximately 30 set (dotted line).
AND
CALCIUM
HOMEOSTASIS
Efect of Co” and Cd2’ on Hg’+-Induced creases in Spontaneous [3H]Dopamine lease
319 InRe-
In nominal Ca’+ media, 2 PM Hg’+ increased spontaneous [3H]dopamine release (Fig. 2). When 2 mM Co” was included in the buffer along with Hg2+ there was a total block of the Hg*‘-induced spontaneous [3H]dopamine release (Fig. 2A). Co*+ (2 mM) alone produced a gradual increase in spontaneous [3H]dopamine release. Recognizing that Co’+ may have other effects in addition to its Ca*+ channel blocking action, we chose Cd’+, a more potent and presumably specific Ca*+ channel blocker, to determine if the Co’+ block of Hg’+ occurred at the voltagedependent Ca’+ channel. To this end, we titrated the Cd2+ concentration to give a similar depolarization-dependent block of [3H]dopamine release as was observed with 2 mM Co2+ (approximately 95%; data not shown). However, this concentration of Cd2+ ( 100 PM) failed to block the release of [3H]dopamine induced by 2 PM Hg2+ (Fig. 2B). Cadmium alone had no effect on the spontaneous release of [3H]dopamine (Fig. 2B).
Efect of Hg”
on Fura-
Fluorescence
Synaptosomes incubated with 10 PM fura2/AM for 45 min hydrolyze 90-95% of the ester (Daniel1 et al., 1987). This finding was confirmed in this laboratory by incubating synaptosomes for various times and determining R,,, . The rise in R,,,, corresponds to fura- formed. R,,, increased with time and plateaued after about 45 min. Addition of NaOH after R,,, stabilized produced no further increases in R,,,. The mean R,,, in these experiments was 5.9 and ranged between 4.17 and 7.8 1. These values are similar to those reported to give stable [Ca*+]i readings in synaptosomes (Komulainen and Bondy, 1987).
320
HARE ET AL
I
1
I
I
I
I
II
I
II
7 8 9 10 FRACTIONNUMBER(15 set each) 2
3
4
5
6
FIG. 1. The effect of Hg” on spontaneous and depolarization-dependent [3H]dopamine release. Release of [‘Hldopamine is expressed as a percentage of total release plus that remaining on the filter. Beginning with fraction 5. the superfusion buffer contained 5 pM Hg*+ (dashed and dotted lines). During fraction 6. synaptosomes received a I S-set pulse of superfusion buffer containing 6 I mM KCI (solid and dashed lines). In one run, 5 WM Hg” only was given without a high K’ pulse (dotted line). The Hepes buffer contained 0.25 mM Ca”. Values represent the means f SEM of five or six animals.
In synaptosomes incubated for 10 min in 1.2 mM Ca2+, the [Ca”]i was significantly elevated relative to synaptosomes incubated in nominal Cal+ buffers (p < 0.001) (Table 1). Exposure to 1,2,4, and 8 FM Hg2+ for 5 min produced no statistically significant effect on [Ca2+]i irrespective of [Ca2+lo (p > 0.05) when analyzed using two-way ANOVA (Table 1). However, when [Ca*+]i of synaptosomes in nominal Ca’+ media was analyzed using oneway ANOVA a significant decrease in [Ca2+]i was revealed (p < 0.001). Post hoc analysis revealed that both the 4 and the 8 I.LM Hg*+ groups were significantly different from controls (p < 0.05). Fluorescence of fura- in synaptosomal suspensions originates from both intrasynaptosomal and extrasynaptosomal sources.
Generally, extrasynaptosomal fura- fluorescence is differentiated by quenching the signal with Mn*+ (Komulainen and Bondy, 1987). The remaining fluorescence is presumed to be intrasynaptosomal. In nominal Ca2’ media, 40 PM Mn*+ quenched 37% of the fura- fluorescence at the 340-nm excitation wavelength (Fig. 3A). In the presence of 1.2 mM Ca*+, the Mn2’ quench of fura- fluorescence at 340 nm was reduced to 24%. Likewise, at all concentrations of Hg*+ in 1.2 mM Cal+-containing buffer, the percentage of fluorescence quenched by Mn*+ was reduced relative to the nominal Ca2+ buffer (p < 0.0 1). Exposure to 8 VM Hg2+ increased the percentage fluorescence that was quenched by Mn*+ irrespective of the [Ca’+],, (a < 0.01) (Fig. 3A). When [Ca2’10 was nominal, both 4 and
MERCURY-INDUCED 6-
I I
I
I
I
I
I
I,
DOPAMINE
I
I
I
RELEASE AND CALCIUM
I
I
A
FRACTION NUMBER (30 set
I
!
I
I
I
321
HOMEOSTASIS
I
f
II
11
1
6
each)
FRACTION NUMBER (30 set
each)
FIG. 2. The effect of Hg”, Co’+. and Cd”, alone or in combination, on spontaneous [lH]dopamine release from superfused striatal synaptosomes. Exposure to the various treatments began with fraction 3 and continued through fraction 12. Data are expressed as in Fig. I, except the fractions are 30 set each. Buffers contained no added Ca”. (A) The effect of 2 pM Hgzc on spontaneous release in the presence and absence of 2 mM Co’+. (B) The effect of 2 KM Hg’+ in the presence and absence of 100 PM Cd’+. Values represent the means +_SEM of four or five animals. The curves for control and 2 PM Hg’+ in A and Bare identical.
8 PM Hg’+ produced statistically significant increases in the percentage of Mn*+ quenching relative to controls (p < 0.0 1). It is assumed that the synaptosomal membrane is relatively impermeable to Mn*+ during the time the fluorescent measurements are taken. Thus, the fluorescence that is quenched by Mn2+ is limited to extrasynaptosomal fura-2. However, it must be considered that Hg*’ may permeabilize the synaptosomal membranes, not only allowing fur-a-2 to leak out, but also increasing the accessibility of intrasynaptosomal fura- to Mn2+. Consequently, the increase in Mn2+ quenching could be intrasynaptosomal. To test this possibility a separate set of experiments was conducted. Synaptosomes were exposed to Hg2+ as before but instead of quenching with Mn2+ at the end of exposure, 1-ml aliquots of the synaptosomes were centrifuged (7OOOg; 2
min) in a Fisher microcentrifuge. An aliquot of the supematant (0.5 ml) was then added to 1.5 ml of 1.2 mM Ca2+ incubation buffer and the fluorescence was determined as described previously. Fluorescence at 340 nm increased relative to controls at Hg2+ concentrations of 4 and 8 PM (p < 0.01) independent of the [Ca2+10 (Fig. 3B). It should be noted that 8 PM HgZf increased the Mn2+ quench in nominal and 1.2 mM Ca2+ media by 81 and 58%, respectively, over controls (Fig. 3A). However, this concentration of Hg2+ increased extrasynaptosomal fura- in nominal and 1.2 mM Ca2+ media by only 52 and 39%, respectively, over controls (Fig. 3B). That is, the increase in Mn2+ quench observed after exposure to 8 PM Hg2’ is only partially accounted for by an increase in extrasynaptosomal fura-2. This discrepancy is likewise apparent for other Hg’+ concentrations and suggests that a com-
322
HARE ET AL. TABLE I
LDH present in the assay. Alterations in LDH activity therefore reflect alterations in EFFECTOF Hg” ON~NTKASYNAPTOSOMAL~ONIZED the quantity of LDH. Since Hg2+ is a potent Ca’+ INRATSTRIATALSYNAPTOSOMES inhibitor of a variety of enzymes, this cannot [Ca’+], nM be assumed a priori. Therefore, Hg2+-inW+ 1 duced decreases in LDH activity, even in the I.2 mM Ca’+ ” Nominal Cal+ ” PM presence of 40 PM cysteine in the assay mix, 0 176k 5(12=7) 474 * 42. (n = 3) must be considered suspicious. For this reaI 185~l:!(rz=3) 533 * 70 (n = 3) son, protein leakage from synaptosomes was 3 178k 13(n=3) 458 + 14 (n = 3) also examined. The quantitation of cytosolic 4 117 k 16 (II = 3)’ 425 i 72 (n = 3) protein using the Coomassie blue method 471 +37(r)= 3) 8 108k 8(n=4)’ was insensitive to inhibition by up to 10 PM Hg’+ (data not shown). Approximately 23% Note Values represent the means f SEM for three to of the total Triton X- 100 releasable protein seven trials(n). 0 Nominal Ca’+ refers to the residual buKer [Ca”] was released from synaptosomes incubated since EGTA was not used. Measurements using furafor 10 min in physiologic Hepes buffer (Fig. reveal this concentration to be between 0.5 and 1 PM. 4B). Mercury had no statistically significant ‘Statistically significant from nominal Ca” (11 effect on synaptosomal leakage of cytosolic < 0.001). proteins (Fig. 4B). However, total protein re” Statistically significant from nominal Ca” controls (a < 0.001) in a one-way ANOVA. leased by membrane disruption with Triton X-100 was significantly decreased by 8 pM Hg”+ both in nominal (p < 0.05) and in 1.2 ponent of the Mn2+ quench is intrasynaptomM Ca” (p < 0.0 1) buffers. somal. These findings do not eliminate the possibility that Hg” increases the leakiness of the plasma membrane to smaller molecules such Determination qf Synaptosomal Integritli as [3HJdopamine and fura-2. To address this possibility the effect of Hg” on [3H]deoxyThe Hg’+-induced leakage of fura- from glucose 6-phosphate (t3H]dgluP) leakage synaptosomes suggests that Hg’+ may profrom superfused synaptosomes was assessed. duce nonspecific membrane damage. To test This molecule is sensitive to toxicant-infor this we monitored the leakage of LDH duced alterations in membrane permeability and protein from synaptosomes exposed to (Walum, 1982) and has a molecular weight Hg2+ for 5 min. In control synaptosomes ap- comparable to dopamine and fura-2. [3H]proximately 20% of the total Triton X- 100 re- dgluP was spontaneously released from synleasable LDH was extrasynaptosomal, irre- aptosomes at the rate of approximately 1% of spective of the [Ca’+lo. Exposure to 8 PM the total synaptosomal [3H]dgluP pool per 30 Hg’+ for 5 min produced a statistically sig- set (Fig. 5). When 2 mM Co’+ was added to nificant decrease in the leakage of LDH, but the superfusate beginning at fraction 3. there only in nominal CG’ (p < 0.05) (Fig. 4A). was no change in [3H]dgluP efflux. However, Synaptosomes exposed to Hg’+ for 5 min and addition of 2 PM Hg*+ increased [3H]dgluP resubsequently lysed with Triton X-100 re- lease by almost sixfold. Hg” (2 mh/r) and Co?-’ leased approximately fivefold more LDH (2 mM) together increased [3H]dgluP efflux simthan nonlysed samples. There was no sta- ilar to that produced by 2 pM Hg2+ alone. tistically significant effect of Hg’+ or Ca2+ IZjflux of 4-‘Ca2‘.fiom Supe&sed Synaptosomes (Fig. 4A). Mercury caused a concentration-depenThe specific activity of LDH normally is assumed to be proportional to the quantity of dent increase in the efflux of preloaded 45Ca’i
,*
4**
m
Mn QUENCH (% total
fluorescence)
324
8
600 -
= 5J
400 -
v) 5 d
HARE
ET AL.
500-
300 200 100
Intact \
_
I 0
,
I 2
,
I
4
/
/
/
8
0
HgC$ Q-M/l,
I
I
2
I
I
I
4
1
8
HgC12 (PM)
FIG. 4. Effects of Hg’+ on synaptosomal LDH activity and protein concentration in nominal Ca2+ (solid lines) and I .2 mM Ca*+ (dashed lines) buffers. (A) Extrasynaptosomal LDH after 5-min exposures to Hg*’ (solid circles) or Hg’+ exposure followed by 0.23% (final concentration) Triton X-100 (solid squares). U = one International unit. (B) Extrasynaptosomal protein concentration after exposure to Hg’+ for 5 min (solid circles) or Hg2+ followed by 0.23% Triton X-100 (solid squares). Values represent the means 2 SEM of three to five animals. Statistically significant from controls: *p < 0.05: **p -c 0.0 I.
from striatal synaptosomes superfused in nominal Ca2+ buffers (Fig. 6). Removal of the Hg2+ resulted in a decrease in 45Ca’+ release rates but at no concentration did the release rates return to preexposure levels within the time frame employed. To determine if the stimulation of 45Ca2+ efflux due to Hg*+ was relevant to the observed increases in transmitter release, we tested the effects of the inorganic Ca2+ channel blockers Co*+ (2 mM) and Cd2+ (100 PM). Synaptosomes exposed to 2 PM Hg2+ released 45Ca2+ at an accelerated rate (Fig. 7). Cobalt alone decreased 45Ca2t efflux to less than that of controls and the efhux did not return to control levels as long as the Co*+ was present (Fig. 7A). Mercury and Co2+ together increased “Ca2+ efflux to a level less than that observed for Hg2+ alone. Analysis of area under the curve revealed that the decrease in 45Ca2+ efflux produced by Co*+ alone was of the same magnitude as the attenuation of the Hg2’-induced 4sCa2+efflux by Co2+ (p > 0.05). That is, Co*+ decreased
45Ca’+ efflux in both instances by the same amount. Cadmium, on the other hand, increased 45Ca2+ efflux which persisted throughout exposure (Fig. 7B). The presence of Cd’+ and Hg*+ together exacerbated the increase in 45Ca2+ efflux relative to that produced by Hg’+ alone (Fig. 7B). As in Fig. 7A, analysis of area under the curve revealed that the increase in 45Ca2’ efflux by the combination of Hg’+ and Cd2’ relative to that produced by Hg*+ alone was of the same magnitude as the increase in 45Ca2+efflux produced by Cd*+ alone relative to controls (a > 0.05) (Fig. 7B). DISCUSSION Inorganic mercury increases [3H]dopamine release from rat striatal synaptosomes. The mechanisms mediating this effect are not likely due to an action of Hg2+ on either the plasma membrane Na+,K+-ATPase or the Nat-dependent high-affinity dopamine
MERCURY-INDUCED
DOPAMINE
RELEASE AND CALCIUM
trec4tmant
I I
1
I
HOMEOSTASIS
1
I
I
I
I
I
I
I
I
I
I
:
325
.
2 3 4 5 6 7 8 9 10 11 12 FRACTION NUMBER (30 set each)
FIG. 5. The effects of2 pM Hg’+ (Hg) and 2 mM Co*+ (Co) on the release of [3H]deoxyglucose 6-phosphate from rat striatal synaptosomes. The metals were given during the phase marked “treatment” (fractions 312) except for control. The superfusing buffer contained no added Ca*+. Other experimental conditions as described under Methods. Values represent the means ? SEM of four to five animals.
transporter (Hare et al., 1989). In the present report we suggest that Hg2’ may have direct effects on [3H]dopamine release presumably near presynaptic “active zones” where exocytosis of neurotransmitter originates. Depolarization-dependent release of [3H]dopamine from rat striatal synaptosomes is tightly coupled to [Ca2+],, below 1 mM (Drapeau and Blaustein, 1983). That is, decreases in [Ca2+10 result in parallel decreases in depolarization-dependent [3H]dopamine release. Moreover, a number of inorganic cations decrease the depolarization-dependent uptake of 45Ca2+ and [3H]dopamine release (Drapeau and Blaustein, 1983; Nachshen, 1984). Therefore, depolarization-dependent [3H]dopamine release from striatal synaptosomes in low [Ca2+10 is limited primarily by Ca2+ en-
try. Since 5 PM Hg2+ had no effect on depolarization-dependent [3H]dopamine release when [Ca2’10 was 0.25 mM, the fast Ca2+ channels probably are not involved in the Hg2+-induced increases in spontaneous [3H]dopamine release. This observation is in agreement with others who have observed that low concentrations of Hg2+ (~10 PM) have no effect on the depolarization-dependent uptake of 45Ca2+ by synaptosomes (Nachshen, 1984). It was surprising then that the Ca2+ channel blocker Co2+ was effective in blocking Hg2+induced [3H]dopamine release. Co2+ (2 mM) blocked approximately 95% of the depolarization-dependent [3H]dopamine release. However, Co’+ is known to have other effects on nerve terminals, particularly on the so-
HARE ET AL.
326 I
I
I
I
I
I
I
1
2
3
I
I
I
I
I1
I
I
I
I
I
I
4
S
6
7
8
3
I
Ill-i
I
III
87-
I
10 11 12 13 14
FRACTION NUMBER (30 set each) 6. Mercuric ion-induced efflux ofpreloaded 45Caz+ from superfused striatal synaptosomes. 45Ca’i efflux is expressed as a percentage of total 45Ca2+released plus that remaining on the filter. There was no added Ca’+ in the superfusing buffers. There was a concentration-dependent increase in 4sCa2+efflux when Hg’+ was given during fractions 3-5. Values represent the means + SEM of four to six animals. FIG.
called slow Ca’+ uptake during depolarization (Nachshen, 1984). Cadmium (100 PM; which also blocked 95% of the depolarization-dependent [3H]dopamine release) failed to block the Hg2+-induced spontaneous release of [3H]dopamine. This effectively eliminates the involvement of the fast Ca2+ channels in Hg*+-induced spontaneous [‘Hldopamine release. Cadmium also is ineffective in inhibiting Pb’+-induced increases in spontaneous acetylcholine release from rat brain synaptosomes (Suszkiw er al., 1984). These authors suggest that Pb*+ acts at the plasma membrane to augment the exocytotic release of acetylcholine since Pb2+ has minimal effect on 45Ca’+ movements in intracellular organelles. Also, membrane impermeant chelators such as EDTA and dimercaptosuccinic acid
rapidly reverse the increase in MEPP frequency produced by low [Pb*‘] (Kolton and Yaari, 1982). Conceivably, Pb2+ and Hg*+, which have similar effects on spontaneous [3H]dopamine release and 4SCa2+efflux from synaptosomes, have common sites of action. Increases in [Ca2+]i are considered obligatory for the exocytotic release of [3H]dopamine from synaptosomes (Drapeau and Blaustein, 1983). However, there was no increase in [Ca2+]i in synaptosomes exposed to Hg*+. Therefore, if the [“Hldopamine released by Hg2+ originated from synaptic vesicles, Hg’+ may act directly to stimulate this release. That is, Hg*+ may bind to sites at or near the plasma membrane in the so-called active zone. Upon binding, Hg*+ may alter the configuration of plasma membrane pro-
MERCURY-INDUCED
DOPAMINE
RELEASE AND CALCIUM
327
HOMEOSTASIS
2 mM Co trsotmont
I
12
3
4 5
6
7
4
8 9 101112
FRACTION NUMBER (30 set
each)
12
3
4
5 6
7 8
5 101112
FRACTION NUMBER (30 set
each)
FIG. 7. ‘%a’+ efflux from preloaded, superfused synaptosomes exposed to Hg’+, Co”, or Cd”. Experimental conditions as in Fig. 6 except synaptosomes were exposed to the metals during fractions 3- 12. The superfusate contained no added Ca’*. (A) Effect of 2 mM Co’+ and 2 MM Hg’+ alone or in combination on 4sCa’i efflux. (B) Effect of 100 ,uM Cd’+ and 2 PM Hg’+ alone or in combination on 45Ca’+ etIIux. Values represent the means f SEM ofthree to five animals
teins to such an extent that transmitter molecules are released from juxtaposed synaptic vesicles. It has been postulated that sulfhydryls may directly regulate the release of acetylcholine from neuromuscular junctions (Kosower and Werman, 1971). Likewise, in synaptosomes, plasma membrane impermeant sullhydryl reagents release [3H]acetylcholine independent of [Ca’+10 (Baba et al., 1979). The observation that Hg2+ decreased [Ca’+]i in synaptosomes in nominal Ca’+ media (Table 1) is analogous to the effect of (Ylatrotoxin observed in guinea pig synaptosomes (Meldolesi et al., 1984). This effect was thought to be due to the net efflux of Ca2+ from the synaptosomes, since there was a concomitant increase in quin-2 efflux suggesting plasma membrane damage had occurred. The net efflux of Ca’+, presumably responsible for the decline in [Ca”]i, occurred because of the presence of 100 PM EGTA in
the extrasynaptosomal media, which is known to lower resting [Ca’+]i in synaptosomes. However, in the present experiments, [Ca’+10 was between 0.5 and 1.O PM and it is therefore unlikely that Ca2’ would exit synaptosomes via this mechanism. The interpretation of this observation therefore remains unclear. Mercury (4 and 8 PM) increased the efflux of fura- from synaptosomes. This effect likely was not due to gross membrane damage since no increases in LDH or protein release from synaptosomes were observed. However, Hg2’ did increase the permeability of the plasma membrane to [3H]dgluP. Evidently, the increased permeability of the plasma membrane induced by Hg*+ is selective for low molecular weight compounds (i.e., <140,000 Da, the molecular weight of LDH). These membrane permeabilizing effects of IIg2+ are apparently independent of its trans-
328
HARE
mitter releasing action. This is based on the observation that Co’+, which totally blocked Hg2+-induced [3H]dopamine release, did not alter Hg*+-induced 4sCa2+ and [3H]dgluP efflux. The membrane permeabilizing effect of Hg2+ would predict an increase in the leakage of cytosolic [3H]dopamine since a sizable component of synaptosomal t3H]dopamine resides in the cytosol, particularly when its metabolism is inhibited by monoamine oxidase inhibitors. Furthermore, the Naf-dependent, high-affinity dopamine transporter, which mediates a component of ouabainand low NG-induced [3H]dopamine release (Raiteri et al., 1979; Hare et al., 1989) provides functional evidence of a cytosolic [3H]dopamine pool in striatal synaptosomes. Alternatively, it could be that Hg” produces at least two types of plasma membrane lesions, one specific for t3H]dopamine, the other specific for [3H]dgluP. Co2+ could be effective in blocking only the t3H]dopamine specific lesion with no action of the lesions responsible for [3H]dgluP leakage. In this scenario, Co” acts to block the efflux of [“Hldopamine rather than hinder access of Hg’+ to release sites. As noted under Results, the Hg2+-induced increase in Mn2+ quench is greater than can be accounted for by the leakage of fura-2. It is likely that Mn’+ quenches some intrasynaptosomal fura-2. The ability of Mn2+ to permeate the plasma membrane of synaptosomes (Komulainen and Bondy, 1987) and of cultured rat cerebellar granule cells (Connor et al., 1987) to quench fura- fluorescence is recognized. However, this is not a serious dilemma since [Ca”]i can be calculated independently of the concentration of furainside the synaptosomes. In the presence of 1.2 mM Ca’+, the Mn2+ quench was less than that observed in nominal Ca’+ (Fig. 3A). Superficially, this suggests an antagonistic effect of Cd+ on fura- efflux. However, fluorescence intensity from supernatant fura- was similar, irrespective of the [Ca2+10 during incubation (Fig. 3B). Altematively, in the presence of 1.2 mM Ca’+, the
ET AL.
quenching efficiency of Mn2+ could be reduced since it competes with Ca2+ for binding to fura-2. However, higher concentrations of Mn2+ failed to significantly increase quenching in 1.2 mM Ca2+ media (data not shown). It seems unlikely that cation competition for binding to fura- could completely explain these data. Another possible explanation could be that in 1.2 mM Cd+, Mn2+ is less effective in permeating the plasma membrane than it is in nominal Ca2+ media since it competes with Ca’+ for entry into synaptosomes. This would effectively reduce the intrasynaptosomal quench in 1.2 mM Cd’. Most likely, the decrease in Mn2+ quenching in 1.2 tnM Ca2+ media results from a combination of these factors. Ca’+ has membrane stabilizing properties and competes with Mn” for binding to fura- and entry into synaptosomes. Mercury produced a concentration-dependent increase in 45Ca2+ efflux from striatal synaptosomes. The efflux of preloaded 45Ca2t from cells has been used as an indirect indicator of changes in Ca” status (Pounds, 1984). The inference is that the observed increase in 45Ca2t efflux induced by Hg’+ reflects 45Ca2+ mobilization from intraterminal stores resulting in an increase in resting [Ca2+]i. Presumably, the synaptosome responds to changes in [Ca2+li by extruding Ca2+ from the terminal (Nachshen et al., 1986). However, on the basis of our measurements using fura2, Hg” did not increase [Ca2+]i and, in nominal Ca2+ media, actually decreased [Ca’+]i at Hg’+ concentrations of 4 and 8 pM. The fact that 2 pM Hg’+ increased 45Ca2+ efflux without measurable changes in [Ca’+]i can be explained as follows. In the Hepes buffers used in our experiments, in which no Ca” was added (nominal Ca’+), the [Ca’+k, was likely greater than 1 PM. Under these conditions, the [Ca’+]i was less than 0.2 pM. Hg2+-induced membrane permeabilization could have increased the influx of 40Ca2f in the direction of the Ca” electrochemical gradient. This increased influx of 40Ca2+ was apparently efficiently buffered since no increase
MERCURY-INDUCED
DOPAMINE
RELEASE AND CALCIUM
in resting [Ca*+]i occurred. Since the kinetic balance between Ca2+ influx and Ca2+ efflux regulates synaptosomal [Ca2’], (Nicholls, 1986), the increased efflux of 45Ca2+ by Hg2+ suggests that the influx of 40Ca2t was accelerated, mixed with intrasynaptosomal 45Ca2+ and both 40Ca2f and 45Ca2+ extruded via plasma membrane extrusion mechanisms. The observation that Co2+ was ineffective at inhibiting the Hg2+-induced efflux of both [3H]dgluP and 45Ca’f is congruous with this hypothesis. In conclusion, Hg” increased [3H]dopamine release from rat striatal synaptosomes. This effect is hypothesized to be due to an action of Hg” at or near the plasma membrane at a site that also interacts with Co” but not Cd*+. The increase in spontaneous [‘Hldopamine release occurred without concomitant elevations in [Ca’+]i. Furthermore, Hg*+ had other effects, in particular an increase in the permeability of the plasma membrane to small molecules. The increase in [3H]dopamine release was apparently not mediated via this nonspecific leakage pathway since it was blocked by Co”+ which had no effect on the leakage of other small molecules tested. ACKNOWLEDGMENTS The authors express their appreciation to Dr. Steven Leslie. Department of Pharmacology, University of Texas at Austin. for allowing M.F.H. the use of laboratory space, equipment, and supplies during the conduct of experiments measuring [Ca*‘],, Robert Greenland and Marlene Jaeger for their technical expertise, and Annette Townsley for preparation of the manuscript. Funds for this work were provided by NIEHS Grant ES-03992.
REFERENCES AUGUSTINE, G. J.. CHARLTON, M. P.. AND SMITH, S. J. (1987). Calcium action in synaptic transmitter release. Annu. Rev. Neurosci. 10, 633-693. BABA, A., FISHERMAN, J. S., AND COOPER, J. R. (1979). Action of sulfhydryl reagents on cholinergic mechanisms in synaptosomes. Biochnn. Pharmacol. 28, 1879-1883. BERGMEYER, H. U.. AND BERNT, E. (1974). Lactate dehydrogenase: UV-assay with pyruvate and NADH. In
HOMEOSTASIS
329
Methods of Enzymatic Analysis (H. U. Bergmeyer. Ed.), pp. 574-576. Academic Press, New York. BINAH, O., MEIRI, U.. AND RAHAMIMOFF, H. (1978). The effects of HgClz and mersalyl on mechanisms regulating intracellular calcium and transmitter release. Eur. J. Pharmacol. 51,453-457. BRADFORD, M. (I 976). A rapid and sensitive method for the quantification of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72,248-254. CHEUNG, W. Y. (I 984). Calmodulin: Its potential role in cell proliferation and heavy-metal toxicity. Fed. Proc. 43,2995-2999. CONNOR, J. A.. TSENG, H.. AND HOCKBERGER, P. E. ( 1987). Depolarizationand transmitter-induced changes in intracellular Ca*+ of rat cerebellar granule cells in explant cultures. J. Neurosci. 7, l384- 1400. COOPER, G. P., AND MANALIS. R. S. (1983). Influence of heavy metals on synaptic transmission: A review. Nerrroto.xico[ogy, 4,69-84. COTMAN. C. W. (1974). Isolation of synaptosomal and synaptic plasma membrane fractions. In Methods in Enzymology 31A (L. Fleischer and L. Packer, Eds.), pp. 445-452. Academic Press. New York. DANIELL, L. C., BRASS,E. P., AND HARRIS, R. A. (1987). Effect ofethanol on intracellular ionized calcium concentrations in synaptosomes and hepatocytes. Mol. Pharmacol. 32,83 l-837. DRAPEAU, P., AND BLAUSTEIN, M. P. (1983). Initial release of [3H]dopamine from rat striatal synaptosomes: Correlation with calcium entry. J. Newosci. 3. 703713. GRYNKIEWICZ, G., POENIE, M.. AND TSIEN, R. Y. (1985). A new generation of Ca” indicators with greatly improved fluorescence properties. J. Biol. Chem. 260,3440-3450. HARE, M. F., MINNEMA, D. J., COOPER. G. P.. AND MICHAELSON. I. A. (I 989). Effects of mercuric chloride on [‘Hldopamine release from rat brain striatal synaptosomes. Tosicol. Appl. Pharmucol. 99,266-275. HOOD, W. F., AND HARRIS, R. A. ( 1980). Effects of depressant drugs and sulfhydryl reagents on the transport of calcium by isolated nerve endings. Biochem. Pharmacol. 29,957-959. JUANG. M. S. (1976). An electrophysiological study of the action of methylmercury chloride and mercuric chloride on the sciatic-sartorius muscle preparation of the frog. Toxicol. Appl. Pharmacol. 37,339-348. KOLTON. L., AND YAARI, Y. (1982). Sites of action of lead on spontaneous transmitter release from motor nerve terminals. IX. J. Med. Sci. 18, 165-170. KOMULAINEN, H., AND BONDY, S. C. (I 987). The estimation offree calcium within synaptosomesand mitochondria with fura-2: comparison to quin-2. Neurothem. Int. 10,55-64.
330
HARE ET AL
KOSOWER, E. M., AND WERMAN. R. (197 I). New step in transmitter release at the myoneural junction. Narwe New Bid 233 . 13 - 1- 122. MANALIS, R. S., AND COOPER, G. P. (1975). Evoked transmitter release increased by inorganic mercury at the frog neuromuscular junction. Nature (London) 251,690-69 1. MELDOLESI, J., HUTTNER, W. B., TSIEN, R. Y.. ANI> POZZAN, T. ( 1984). Free cytoplasmic Ca*+ and neurotransmitter release: Studies on PC 12 cells and synaptosomes exposed to cy-latrotoxin. Ptw. Nafl. .-lead. Sci. US.4 81,620-624.
MINNEMA. D., AND MICHAELSON. 1.A. (1985). A superfusion apparatus for the examination of neurotransmitter release from synaptosomes. J. Neworr~i. Mrthads 14, 193-206.
MINNEMA. D.. AND MICHAELSON, 1. A. (1986). Differential effects of inorganic lead and &aminolevulinic acid in vilro on synaptosomal &aminobutyric acid release. To.ricol. .dppl. Phartnc~ol. 86, 437-447. NACHSHEN. D. A. (1984). Selectivity of the Ca binding site in synaptosome Ca channels: Inhibition of Ca influx by multivalent metal cations. J. Cm. Phv.sio/. 83, 941-967. NACHSHEN, D. A.. SANCHEZ-ARMASS. S.. AND WEINSTEIN, A. M. (1986). The regulation of cytosolic calcium in rat brain synaptosomes by sodium-dependent calcium efflux. J. Ph?:yio/. 381, 17-28.
NICHOLLS, D. G. (1986). Intracellular calcium homeostasis. Brit. Med. Bull. 42.353-358. POUNDS.J. (I 984). Effect oflead intoxication on calcium homeostasis and calcium-mediated cell function: A review. Ntwotoxicology 5,295-332. RAITERI, M., CERRITO, F.. CERVONI, A. M., AND LEVI. G. (I 979). Dopamine can be released by two mechanismsdifferentially affected by the dopamine transport inhibitor nomifesine. .I. Pharmacol. Ekp. Ther. 208, 195-202.
SMITH, M. W.. AMBUDKAR, 1. S., PHELPS, P. C.. REGFC. A. L.. AND TRUMP, B. F. (1987). Hgr&induced changes in cytosolic Ca’+ of cultured rabbit renal tubular cells. Biochim. Biophys. &la 931, I 30- 142. SNEDDON,J. (1987). Action of di- and tri-valent cations on calcium-activated K’-efflux in rat erythrocytes. Biwhetn.
Phurma&.
36, 3723-3730.
STEEL, R. G. D.. AND TORRIE. J. H. (1960). Principles and Procedures of .Qulis/ics. McGraw-Hill. New York. SUSZKIW. J., TOTH, G.. MURAWSKY, M.. AND COOPER, G. P. ( 1984). Effects of Pb2+ and Cd’+ on acetylcholine release and Ca2+ movements in synaptosomes and subcellular fractions from rat brain and Torpedo electric organ. Brain Res. 323, 3 l-46. WALUM, E. ( 1982). Membrane lesions in cultured mouse neuroblastoma cells exposed to metal compounds. To.~i~olog~
25,67-74.