The effects of intracellular application of various concentrations of adenine nucleoside phosphates and nucleotide analogs on the M-type K current (I M) of single neurons isolated from sympathetic ganglia were studied. With 1 mmMgATP intracellularly I M decreased to 25% of its initial level 39 min after the start of whole-cell recording. In the absence of ATP the current decreased more rapidly. Addition of glucose and pyruvate extracellularly was equivalent to adding 1 mm MgATP intracellularly. AMP-PNP, a nonhydrolyzable ATP analog, at a concentration of 1 or 3 mm was unable to maintain I M in the absence of ATP. When ATP and AMP-PNP were combined in the pipette, however, the maintenance ofI M was prolonged. A series of nucleotides and analogs have been combined with ATP to test for their ability to maintain I M and to alter calcineurin phosphatase activity. There was a positive correlation between the ability of a nucleotide to prevent the rundown ofI M and its ability to inhibit calcineurin phosphatase activity. These findings show that the amplitude ofI M is dually regulated by cellular levels of adenine nucleotide diphosphates and triphosphates. A hydrolyzable form of ATP is necessary to maintain the M current. The maintenance ofI M is further enhanced by the simultaneous presence of ADP or other adenine nucleotides that alter calcineurin activity, but not by higher concentrations of ATP alone. These results are consistent with regulation of I M by phosphorylation events that maintain I M and dephosphorylation events that lead to current rundown.
- sympathetic neuron
- potassium current
- nucleotide phosphates
The M-current (I M) is a voltage- and time-dependent potassium current that has been observed in a variety of neuronal cell types. I M was first recorded from sympathetic neurons (Brown and Adams, 1980) and has subsequently been observed in recordings from brain (Halliwell and Adams, 1982; Halliwell, 1986), spinal cord (Murase et al., 1986), and smooth muscle (Sims et al., 1985). I M is activated at membrane potentials positive to −60 mV and consists of an outward K current that does not inactivate. Compounds that act as agonists at gonadotropin-releasing hormone receptors, muscarinic receptors, substance P receptors, or purinergic receptors inhibitI M. The precise signal transduction mechanism by which these compounds decrease I M is not known, but the data clearly suggest that the inhibition ofI M by these receptors follows activation of guanine nucleotide-binding proteins (Pfaffinger, 1988; Brown et al., 1989; Lopez and Adams, 1989; Simmons and Mather, 1992).
Intracellular levels of ATP have been suggested to be involved in the control of I M; however, the nature of this regulation remains unclear. There are conflicting findings with regard to the role of intracellular ATP in modulatingI M. Pfaffinger (1988) found that whenI M was recorded with electrodes containing a solution that lacked ATP, I M declined much more rapidly than if the solution suffusing the cell contained ATP. When ATP was replaced by AMP-PNP, the M-current was lost. Tokimasa and Akasu (1990) found that I M in dorsal root ganglion cells disappeared or was reduced to 20% of its initial value after 15 min when ATP was replaced by AMP-PNP in the whole-cell recording electrode. Our results (Simmons et al., 1990), however, showed that when intracellular ATP was replaced by AMP-PNP, the current was maintained at 84% of the initial value after 13 min. Chen and Smith (1992) also found that I M was not affected when ATP was left out of the pipette solution and that AMP-PCP did not affect the rundown of I M.
Direct comparisons of previous studies on I M are difficult, because different laboratories have used different solutions, both intracellularly and extracellularly. Of particular concern are the concentrations of MgCl2, ATP, and GTP added to pipette solutions and the presence or absence of metabolic substrates such as glucose and pyruvate in the extracellular solutions and in the pipette. Here, we examine the role of intracellular ATP, a variety of other nucleotide phosphate analogs, and extracellular pyruvate and glucose in the maintenance of I M in single neurons.
MATERIALS AND METHODS
Cell isolation. Bullfrogs (Rana catesbeiana) were obtained from Charles D. Sullivan Co. (Nashville, TN) and handled according to national and institutional guidelines with the approval of the Institutional Animal Use and Care Committee. To obtain sympathetic neurons, a frog was chilled in ice for 1 hr, rapidly decapitated, and double-pithed. Sympathetic ganglia were dissected and incubated for 90 min in dissociation medium with collagenase A (1 mg/ml) and trypsin (0.5 mg/ml), for 90 min in dissociation medium with collagenase B (0.75 mg/ml), and then stored in growth medium at 4°C for up to 3 d. On the day of the experiment, one or two ganglia at a time were triturated 25–50 times with a long-shanked glass Pasteur pipette to free single cells from the ganglia. The cells were placed in a Petri dish on the stage of an inverted microscope and continuously perfused with extracellular solution.
Whole-cell recordings. All experiments were conducted at room temperature, ∼20°C. Whole-cell recordings (Hamill et al., 1981) were made with electrodes with resistances of 0.25–1 MΩ when filled with intracellular solution. The extracellular solution was controlled by a single-cell superfusion system as described previously (Simmons and Mather, 1991). In selected experiments, intracellular solutions were changed by perfusing the interior of the whole-cell recording electrode as described by Fischmeister and Hartzell (1987). The recordings were filtered at 1 kHz and stored on magnetic tape.
Measurement of IM . The membrane potential was maintained at a holding potential of −30 mV. Every 8 sec the voltage was stepped to −50 mV for 500 msec followed by return to −30 mV. After a voltage step from −30 to −50 mV, the outward current decreases because of the time- and voltage-dependent closing of M-channels (Fig. 1) (Brown and Adams, 1980). After returning to −30 from −50 mV, the outward current returns exponentially (Fig. 1). The amplitude of this exponential relaxation of current is termed I M(−30) and provides a measure of I M that is independent of changes in other currents or “leak” currents (Simmons et al., 1994). Instantaneous and steady-state current amplitudes at −30 and −50 mV were determined by computer.
Quantification of the rundown of IM .To quantify the time-dependent changes in the amplitude ofI M during an experiment, the current was measured every 8 sec, and the amplitude ofI M(−30) was plotted versus time. The initial current is defined as the maximum current reached after obtaining the whole-cell configuration. The rundown of I Munder various experimental conditions is expressed as the time it took for the amplitude of I M to decrease to 25% of this initial current value, T(25%). If the current did not reach 25% of the initial value after 1 hr, the experiment was halted, and the T(25%) for that cell assigned a value of 60 min.
Solutions for electrophysiology. Composition of solutions is in mm unless otherwise noted. Extracellular solution: NaCl 128, KCl 2.4, CaCl2 1.8, MgCl2 1.8, HEPES 10, and TTX 0.0003, pH 7.4 (with NaOH). Dissociation medium: NaCl 98, KCl 2.4, NaH2PO4 0.6, MgCl2 1.8, sodium pyruvate 5, creatine 5.7, glucose 5, M199 10 ml/l, penicillin 100 U/ml, streptomycin 100 μg/ml, and HEPES 20, pH 7.4 (with NaOH). Growth medium: NaCl 118, KCl 2.4, creatine 5.7, glucose 5, sodium pyruvate 5, 100× MEM vitamins 10 ml/l, penicillin 100 U/ml, streptomycin 100 μg/ml, 50× MEM essential amino acids 20 ml/L, 100× MEM nonessential amino acids 10 ml/l, bovine serum albumin 1 mg/ml, and HEPES 20, pH 7.4 (with NaOH). Intracellular (or recording electrode) solution: KCl 120, BAPTA 1, and HEPES 10, pH 6.8 (with KOH), plus MgCl2 and nucleotides.
The amounts of MgCl2 and nucleotides to be added to the pipette solutions were determined by a computer program developed by Robert E. Godt (Medical College of Georgia, Augusta, GA) (Godt and Lindley, 1982) with modifications by H. Criss Hartzell (Emory University, Atlanta, GA). Calculations were made to give a free [Mg] of 0.5 mm and the desired concentrations of MgATP. This program takes into account the stability constants of the various components in a manner similar to that described in Fabiato (1988). In addition to accounting for the concentrations of the various constituents, temperature, pH, and ionic strength are also considered in the calculations. The stability constants for ATP and ADP analogs were taken to be the same as for ATP and ADP, respectively.
Most chemicals were obtained from Sigma Chemical (St. Louis, MO) or Fisher Scientific (Pittsburgh, PA). Proteolytic enzymes, nucleotides, and leupeptin were obtained from Boehringer Mannheim (Indianapolis, IN).
Measurement of calcineurin activity. Calcineurin phosphatase activity was assayed by measuring the release ofpara-nitrophenol from para-nitrophenyl phosphate spectrophotometrically at 405 nm (Pallen and Wang, 1983) in a Vmax kinetic microplate reader from Molecular Devices Corp. (Menlo Park, CA). The assay was performed at 30°C in a 200 μl reaction volume containing 2 μg of purified bovine brain calcineurin (Upstate Biotechnologies, Lake Placid, NY) and 500 nm purified bovine brain calmodulin (Sigma) in buffer consisting of (in mm): 120 KCl, 10 HEPES, 2 dithiothreitol, and 0.35 ATP, pH 8.1. Varying amounts of MgCl2 and adenine nucleotides were added according to the method described above to give buffers containing (in mm): 0.3 ADP, 3 ADP, 1 AMP-PNP, 3 AMP-PNP, or 3 sodium pyrophosphate. Blank wells were prepared containing (in mm) 120 KCl, 0.8 MgCl2, and 10 HEPES, pH 8.1. To monitor background activity, control wells were prepared that contained buffer solutions without calcineurin and calmodulin. The microtiter plate was equilibrated for 5 min at 30°C in a precision convection oven. The reaction was initiated by the addition of 20 mm para-nitrophenyl phosphate (Sigma). After incubation at 30° for 60 min, the absorbance values for each well were measured and the background value subtracted to obtain the final optical density values.
Data analysis and statistics. Data are expressed as mean ± SD. Tests for differences among experimental groups were determined by one-way or two-way ANOVA as appropriate followed by Student–Newman–Keuls test for multiple comparisons.
Dependence of IM on intracellular ATP
Typical cellular levels of MgATP are in the range of 1 mm. Previous studies of I M have typically used solutions in the recording electrode that would yield approximately this amount of MgATP. The time-dependent changes inI M observed throughout a recording period with this concentration of MgATP (1 mm) or without ATP in the recording pipette are shown in Figure 1. In this experiment, when the electrode contained 1 mm MgATP, the current ran down 43%, from a peak of 460 to 265 pA in the first 25 min after the start of whole-cell recording (Fig. 1 A). When ATP was omitted from the pipette solution, I M declined more rapidly (Fig. 1 B), decreasing 81% from 520 to 100 pA within ∼13 min. The amplitude of I M(−30) for the duration of these two experiments is plotted as a function of time in Figure 2. This graph shows that when ATP (1 mm) was included in the pipette (Fig. 2,filled circles), the maintenance ofI M was prolonged compared with the amplitude ofI M when ATP was omitted from the pipette (Fig.2, open triangles). In 12 cells tested with 1 mmMgATP in the pipette, the time it took to decline to 25% of the initial current level, the T(25%), was 39.0 ± 19.0 min. In 10 cells with no ATP added, the T(25%) was 12.8 ± 3.5 min.
Maintenance of IM requires ATP hydrolysis
There are two obvious mechanisms by which ATP may be used to prevent rundown of I M: ATP could serve as a phosphate donor in a phosphorylation reaction that maintains the M-current, or the energy from ATP hydrolysis could be required to maintain I M. Either of these mechanisms would require the availability of a hydrolyzable form of ATP. To determine whether hydrolyzable ATP was required to maintainI M, the ATP in the pipette was replaced by AMP-PNP, a nonhydrolyzable ATP analog. I Mdisappeared more rapidly when the electrode contained MgAMP-PNP at a concentration of 1 mm (Fig. 2, open squares) than when it contained MgATP (1 mm) (Fig. 2, filled circles). The average T(25%) with 1 mmMgAMP-PNP in the cell was 8.9 ± 5.3 min (n = 6). Increasing the intracellular concentration of AMP-PNP to 3 mm resulted in a rapid rundown of the current with an average T(25%) of 7.6 ± 3.5 min (n = 20).
Extracellular glucose and pyruvate are equivalent to intracellular ATP
The finding that AMP-PNP resulted in a rapid disappearance ofI M was in contrast to our earlier finding thatI M was maintained at 84% of control after 13 min when AMP-PNP replaced ATP in the pipette (Simmons et al., 1990). In the previous study, 5 mm pyruvate and 5 mmglucose had been included in the extracellular solution, whereas in the present study glucose and pyruvate were omitted. Thus, it was possible that in our earlier studies the cells were utilizing the available pyruvate and glucose to synthesize enough ATP to maintainI M in the presence of intracellular AMP-PNP. To test this possibility, glucose and pyruvate were added to the bath solution, and the effect of AMP-PNP was reexamined. When glucose and pyruvate (5 mm each) were present, the decrease inI M observed with 1 mm AMP-PNP in the cell was slowed from a T(25%) of 8.9 ± 5.3 min (n = 6) to a T(25%) of 21.2 ± 14.7 min (n = 7) (Fig. 3).
When either one of these energy substrates alone was added to the extracellular perfusate, the decrease of I M with time was intermediate between that seen with both present and that seen with neither present. With 5 mm glucose theT(25%) was 11.3 ± 5.2 min (n = 7), and with 5 mm pyruvate the T(25%) was 11.7 ± 5.9 min (n = 6). With pyruvate and 2-deoxy-d-glucose the T(25%) was similar to the result seen with pyruvate alone (11.2 ± 3.4 min;n = 13).
Concentration–effect curve for ATP
To determine the dependence of I M on the concentration of intracellular ATP, electrodes were filled with solutions containing different amounts of MgATP. The T(25%) with various [MgATP] in the pipette is plotted in Figure4 (open circles). The maintenance of I M was prolonged as [MgATP] was increased from 0 to 1 mm. With a further increase of [MgATP] to 3 mm, however, the decrease ofI M with time was unchanged.
A combination of intracellular AMP-PNP with intracellular ATP slows the rundown of IM
In contrast to the findings with AMP-PNP alone, when AMP-PNP and ATP were combined in the pipette a different story emerged. In Figure 4the filled symbols show the concentration-effect curve for MgATP in the presence of a constant concentration of 1 mmMgAMP-PNP (filled circles) or 3 mmMgAMP-PNP (filled squares).
The concentration–effect curve for ATP exhibited a downward shift in the presence of 1 mm MgAMP-PNP. At a concentration of 3 mm, however, MgAMP-PNP in the simultaneous presence of MgATP at concentrations of ≥0.3 mm resulted in a sustained maintenance of I M. With 1 mm MgATP and 3 mm MgAMP-PNP, there was essentially no rundown of the current during a 1 hr recording period. This compares to a rundown to 25% in 39 min with 1 mm MgATP alone or in 7.6 min with 3 mm MgAMP-PNP alone.
The rapid rundown observed with 3 mm AMP-PNP could be reversed by intracellular perfusion of 1 mm ATP along with 3 mm AMP-PNP. This experiment was repeated in five cells. In all cases the rundown normally observed with 3 mmAMP-PNP was halted when ATP was perfused intracellulary. In four of the five cells the current rundown was not only halted but was also reversed, as illustrated in Figure 5.
These results suggest that there are dual effects of adenine nucleotides on I M. Maintenance ofI M requires a hydrolyzable form of ATP, because a nonhydrolyzable analog alone is insufficient to maintainI M. In the presence of ATP,I M is further regulated by the simultaneous presence of nonhydrolyzable adenine nucleotides.
To see whether other nucleotides or analogs might be more effective than AMP-PNP in producing this effect, we have tested the effects of a variety of other nucleotide analogs on I M.
Effects of adenine nucleoside diphosphate analogs on the rundown ofIM
Concentration–effect curves for MgATP have been constructed in the presence of a constant concentration of ADP or AMP-CP (3 mm). In these experiments, Mg and ATP were added at concentrations calculated to provide 0.5 mm free Mg and the desired amount of MgATP. Of the 3 mm added ADP or AMP-CP, 1 mm is predicted to be free ion, 0.8 mmcomplexed to Mg ions, and the remainder complexed to other metals. Figure 6 shows a comparison of the effects of ATP alone (open circles) and in the presence of 3 mm ADP (open triangles) or 3 mmAMP-CP (open squares). The ability of ATP to slow the rundown of I M was enhanced in the simultaneous presence of ADP. In contrast to this, the ability of ATP to slow the rundown of I M was inhibited by 3 mmAMP-CP for ATP concentrations of ≤1 mm but was enhanced when ATP was present at a concentration of 3 mm. A concentration–effect curve for ADP in the presence of 0.3 mm ATP is shown in Figure7.
From the dose–response curves shown, it is apparent that the ability of compounds to alter the rundown of I M is most easily observed with 0.3 mm MgATP in the pipette. With 0.3 mm MgATP, the T(25%) for AMP-CP (3 mm) was 6.4 min, for AMP-PNP (1 mm) was 14.7 min, for MgATP alone was 19.4 min, for AMP-PNP (3 mm) was 31.2 min, and for ADP (3 mm) was 58.3 min. The actions of a variety of other analogs on I M rundown in the presence of 0.3 mm MgATP have also been tested. TheT(25%) for these compounds is shown in Table1. The data are listed in order of decreasing T(25%).
Effects of adenine nucleotides and analogs on other electrophysiological properties
In addition to measuring I M, several other parameters have been measured to examine the specificity of the effects with respect to I M. For statistical comparisons, t tests have been made between the data obtained under the various experimental conditions to the data obtained with 0.3 mm ATP. When recording I M a run-up of current is sometimes observed during the first 30–60 sec of recording (Figs. 2, 3). The time from obtaining whole-cell recording to reaching this initial peak current was not altered by the different nucleotides. Nor was the amplitude of I M at this initial time point affected. The leak current at this time and at theT(25%) were also measured and did not differ from the values obtained with 0.3 mm ATP.
It has been shown previously that the activity of the calcium- and calmodulin-dependent phosphatase 2B calcineurin modulatesI M in neurons from rat sympathetic ganglia (Marrion, 1996). It has also been shown that phosphates modulate calcineurin activity (King and Huang, 1984). To test the hypothesis that the effects of intracellular nucleotide phosphates onI M may be accounted for by the effects of these nucleotides on calcineurin activity, the effects of concentrations of selected nucleotides that were shown to alter I Mabove have been tested on calcineurin phosphatase activity. The results of the calcineurin phosphatase assays are shown in Table2.
Effects of a calcineurin inhibitor
Cyclosporin A is an inhibitor of calcineurin activity (Liu et al., 1991). When applied intracellularly to single neurons with 0.3 mm ATP, cyclosporin A at a concentration of 500 nm increased the T(25%) from 19.4 ± 14.9 to 30.5 ± 10.7 and to 41.9 ± 15.8 min when added at a concentration of 1 μm. Cyclosporin A (500 nm) had no effect when applied outside the cell [T(25%) = 9.1 ± 1.8, n = 2].
The results clearly show that nucleotide phosphates play a critical role in the regulation of I M. Furthermore, there are at least two ways by which adenine nucleotides can modulate I M. First, hydrolyzable ATP must be available for I M to be maintained. Second, above and beyond this requirement for a basal level of ATP, other nucleotide phosphates and analogs act to further modulate the level ofI M. These results are consistent with regulation of I M by kinases and phosphatases.
Several previous studies have suggested that I Mis regulated by phosphorylation catalyzed by myosin light chain kinase. Application of myosin light chain kinase inhibitors decreasesI M, whereas intracellular perfusion of catalytic subunits of myosin light chain kinase enhancesI M (Akasu et al., 1993; Tokimasa et al., 1995). The result presented here, that ATP is required to prevent the rundown of I M, extends these findings to suggest that myosin light chain kinase-catalyzed phosphorylation events not only alter the amplitude of I M but may be required for the maintenance of I M.
In addition to the positive regulation of I M by myosin light chain kinase-catalyzed phosphorylation events,I M has also been shown to be negatively regulated by phosphatase-catalyzed dephosphorylation events (Marrion, 1996). Marrion (1996) showed that I M was decreased when a catalytically active form of calcineurin was applied intracellularly to rat sympathetic neurons. This suggests thatI M is regulated by the calcium and calmodulin-dependent phosphatase type 2B, calcineurin, and that this phosphatase activity results in a decrease in the amplitude ofI M.
In the present study, several compounds have been shown to prevent the rundown of I M. These compounds include ADP, PPi, and cyclosporin A. The results of the phosphatase assay show that ADP and PPi, at concentrations that prevent the rundown ofI M, are effective inhibitors of calcineurin phosphatase activity. Cyclosporin A is well established as a calcineurin inhibitor (Liu et al., 1991). These data support the previous findings that I M is regulated by phosphatase activity (Marrion, 1996) and further illustrate that this phosphatase activity can be regulated by endogenous compounds such as ADP and PPi. Pyrophosphate and ADP at a concentration of 3 mm significantly inhibited calcineurin phosphatase activity and slowed I M rundown. A 0.3 mmconcentration of ADP changed neither calcineurin activity nor the rundown of I M.
This conclusion is further supported by the results obtained with AMP-PNP. The effect of AMP-PNP depended on both its concentration and on the concentration of ATP present. As intracellular ATP was increased, the maintenance of I M was prolonged. When AMP-PNP at a concentration of 1 mm was added along with ATP, the ability of ATP to maintainI M was decreased. When AMP-PNP at a concentration of 3 mm was added along with ATP at concentrations of ≥0.3 mm, however, the maintenance ofI M was enhanced. These results can be interpreted in accordance with the effects of AMP-PNP on calcineurin activity. At a concentration of 1 mm, AMP-PNP increased calcineurin activity and speeded I M rundown. At 3 mm, AMP-PNP decreased calcineurin activity and slowedI M rundown.
Although there is qualitative agreement between the effects of the various nucleotides in the calcineurin assays and their effects onI M rundown, there are quantitative differences. For example, 3 mm ADP and 3 mm AMP-PNP have equivalent, less than PPi, effects in the calcineurin assay. On I M rundown, the effect of ADP is greater than that of PPi, whereas the effect of AMP-PNP is less than PPi. This difference might be accounted for by the different situations under which these two measurements were made. The bovine brain calcineurin in vitro is clearly under different conditions than the calcineurin in a bullfrog sympathetic ganglion cell undergoing whole-cell recording.
Importance of considering the role of pyruvate and glucose in physiological solutions on ionic currents
Many studies of ionic currents in single cells use whole-cell recordings. With this technique, it is generally assumed that the electrode contents effectively suffuse the cell and thereby control the cellular contents. Metabolic substrates such as glucose and pyruvate are routinely added to either the extracellular medium or to the pipette solution during such experiments. The data presented here demonstrate that these additives can have important modulatory effects on ionic currents. These effects may occur even when the pyruvate and glucose are only added extracellularly and recordings are made with large electrodes that would be expected to provide good control of intracellular constituents.
The ability of extracellularly applied pyruvate and glucose to substitute for intracellular ATP shows that enzymatically dissociated single neurons maintain active metabolic processes. TheT(25%) in the presence of extracellular pyruvate and glucose and intracellular AMP-PNP was 21.2 min. A similarT(25%) of 24.6 min was obtained with 1 mm MgATP and 1 mm AMP-PNP. Thus, adding 5 mm glucose and 5 mm pyruvate to the extracellular medium is equivalent to putting 1 mm MgATP in the pipette.
This can account for some of the discrepancies observed in previous work from different laboratories. Those studies that had reportedI M was maintained without ATP had 5 mm pyruvate and 5 mm glucose in the extracellular solution (Simmons et al., 1990) or 10 mmglucose in the extracellular solution and 10 mm glucose in the pipette solution (Chen and Smith, 1992). Those studies that had reported that I M disappeared in the absence of ATP (Pfaffinger, 1988; Tokimasa and Akasu, 1990) did not add either pyruvate or glucose to their solutions.
These data also show that ATP is required for the maintenance ofI M as suggested previously (Pfaffinger 1988;Tokimasa and Akasu 1990). The finding that the M current was maintained in the presence of intracellular ATP, but not with AMP-PNP alone, supports previous work showing modulation of I Mby kinases and phosphatases (Akasu et al., 1993; Marrion, 1996).
In conclusion, the data presented here show that the levels of endogenous compounds such as ATP, ADP, and PPi play an important role in modulating the level of I M in single neurons. The present results, when taken in perspective with other studies on I M, are consistent with a dual regulation of I M by kinases and phosphatases. In the presence of ATP, myosin light chain kinase-catalyzed phosphorylation enhances I M. Calcineurin-catalyzed dephosphorylation inhibitsI M. High levels of ADP or PPidecrease the activity of calcineurin and enhanceI M. The ultimate result of an increasedI M is a decrease in neuronal excitability. Thus, the intracellular levels of adenine nucleotides provide an important mechanism for cellular regulation of neuronal excitability by influencing the level of I M.
This work was supported by Grant NS-25999 from the National Institute of Neurological Disorders and Stroke. We thank Dr. James B. Becker and Mr. Robert J. Mather for participating in some of the early experiments.
Correspondence should be addressed to Dr. Mark A. Simmons, Department of Pharmacology, Marshall University, Huntington, WV 25704-9388.