Abstract: Carbon-fiber microelectrodes were used to electrochemically measure concentrations of oxidizable neurotransmitters, serotonin, epinephrine, and norepinephrine. The electrodes consisted of a carbon filament surrounded by a glass capillary tube. The carbon filament was connected to a patch clamp amplifier to set the filament at a certain voltage and to measure current flow, which increased during oxidation. This was done both with a triangle wave, cyclic voltammetry, and a set potential of +5 Calibration curves were done with known and .6V, amperometry. concentrations of both epinephrine and serotonin in vitro. In vivo recording were performed in Pleurobranchea pedal ganglion and NIE-115 Neuroblastoma cells. Epinephrine behavior was close to theory with a one to one relation between increase in I and [Epil, but the response to 5-HT was not as strong. The electrodes measured 90 uM (400 pA by amperometry release of serotonin from the Pleurobranchea pedal ganglion. Readings were also taken from NIE-115 cells which showed 1 to 100 uM (1.6nA) norepinephrine release after stimulation with carbachol, a non-oxidizable analog of acetylcholine. Introduction: Carbon-fiber microelectrodes are, and have been for years, used in electrochemistry. The method has great potential for looking at biological problems because it has fantastic speed, within milliseconds, and sensitivity, detection down to 100 nM (Wightman, 1981). It has even be used to measure single vesicle release (Chow et al., 1992 and Wightman, 1991). The technique takes advantage of the fact that amine-comtaining neurotransmitters are easily oxidizable. All catecholamines for instance, oxidize in a two-electron process, shown for epinephrine in Figure 1. Because the two released electrons are constant, they are, theoretically, a measure of the amount of oxidation that has occured. The carbon-fiber electrode takes advantage of this on a small scale and uses a 8 micron diameter carbon-fiber. As shown in Figure 2, when the voltage potential of the electrode, as locked in by a patch clamp amplifier, is above the oxidation potential of the neurotransmitters, which is different for each substance, the carbon filament oxidizes the neurotransmitter. The higher above this the oxidation point the potential is set, the greater the rate of oxidation at the electrode surface. When a range of set voltages for the carbon-fiber is scanned, a leak and capacitance base current causes a fluctuation of current, higher when it increases and lower when it decreases. Above +1 V water is oxidized causing a large peak. To determine the exact current for a disk shaped electrode which is totally irreversible (assuming low concentration so that the oxidized product diffuses away before being reduced) the equation for the current can be solved to: i= Covl2k v1/2 is the scan rate and because i = C dv/dt for any capacitor, like this electrode, the higher the scan rate the higher the background current from capacitance. Co is the concentration of oxidized substances. K is a factor that takes into account a number of different factors, such as the geometry of the carbon-fiber surface and the nature of the oxidation reaction occurring. The relation to current and concentration in this equation is direct for any given voltage (Bard and Faulkner, 1980, 222). When the carbon-fiber is set at a certain potential, the situation is similar to when one particular voltage of the whole range scanned in cyclic voltammetry is examined. Concentration is important because the only way new oxidizable molecules get to the carbon- fiber surface is by diffusion which is driven by the concentration. Not all catecholamines are the same, and serotonin presents a special difficulty because its oxidized form creates a film over the electrode which reduces the electrodes sensitivity (Wightman, 1988). It can be reduced by scanning at very high scan rates (200V/s) which unfortunately also involves a lot of background current, making it difficult not to saturate the patch clamp amplifier. The next part of this study was to use the microelectrodes to measure neurotransmitter release in biological systems. Two systems were chosen. First, The pedal ganglion of the Pleurobranchea is believed to secrete serotonin (Rhanor Gillette, personal communication) which is supported by the production of serotonin by the pedal and abdominal ganglia (Wilbur, 1985), and the NIE-115 mouse neuroblastoma cells, which contain norepinephrine (Herken, 1982). Materials and Methods: Electrode Fabrication: The carbon-fiber electrodes were made by a standard method. (Fishman, 1992). The carbon-fibers were spread onto a clean piece of paper. Figure 3A is a picture of carbon-fibers. Smoother paper worked better. A single fiber was separated with the finger tips. Multiple fibers were be fairly easily visually distinguished from single ones. Two fibers together can be separated by rolling at the tip or the middle. Fibers were selected that were as long as or longer than the capillary tubes. A vacuum pump (Air Cadet, Model No. 7530-40) was attached to a glass capillary tube (Work Precision Instruments, Cat No TWISOF) and used to aspirate a single carbon filament whose tip was overhanging the paper 2mm. The far end of the fiber was held to prevent the whole fiber from being sucked in. Using the vacuum pressure, the fiber was brought all the way through the capillary tube. The capillary tube was then pulled in an electrode puller (David Kopf Instrument Vertical Pipette Puller, Model 700C, modified by Professor William Gilly) with heat set at 47. The carbon-fiber that separated the two electrodes was cut and the electrodes were removed. The next step was to dip the tips into epoxy (Epo-tek 301) for ten minutes while it filled the tip. Four hours in the oven (NAPCO, Model 5830) at 65°C was adequate to dry the epoxy, making a working electrode, Figure 3B. There was a differences between different electrode pullers. The Narishiga PE-2 would not pull electrodes that would fill with epoxy after pulling. Electrodes were painted with silgard and then heated until dry. The electrodes made this way worked normally and were made much more quickly. The last step was to cut the carbon-filament very close to the end of the glass of the capillary tube. A razor blade was used to cut the carbon fiber. This was done immediately before experimental use. Testing the microelectrodes: The electrodes were tested in vitro. Baselines were tested in distilled water with 150 mM NaCl. A ground, a bare silver/silver chloride wire, was put into the solution. The electrode was connected to a headstage (Warner Instrument Corporation, 5C-501) of patch clamp amplifier. The output current was filtered (Frequency Device, 902) at 20 hz. The command voltage was provided by Exact, model 119 Function Generator. Constant potentials for amperometry were provided by a de current source. The current and voltage were viewed on a Nicolet 3091 digital oscilloscope. Electrodes were simply put into the pools with varying known concentrations of neurotransmitter for testing. To prevent premature oxidation, the solid compounds were stored in a freezer in a light opaque container. When the epinephrine and serotonin solutions were measured out, they were added to 150 mM NaCl solution that had been bubbled for 1 minute with 99.999% N. (Liquid Carbonic) and sealed with Parafilm. As soon as the compound was added it was again bubbled with N2 and sealed. For a portable N2 source, balloons were filled with N2 and fitting with a rubber tubing adaptor and a needle. Epinephrine dose response calibration: The electrodes was positioned in the bath and solutions were changed by using the profusion device. The carbon filament was kept at +.5V and tested at a range of concentrations. The order was: baseline with O mM epinephrine, then stepping up concentrations, and then ending with another wash at 0 mM epinephrine. The current and the command voltage were measure on tape recorder (Tanberg Instrumentation Recorder, Series 115). The information was then digitized. Measurements in the Pleurobranchea pedal ganglion: Before the experiment, a dose response calibration was done with creatine sulfate serotonin. Concentrations of 10-8 to 10-3 serotonin were tested. Between each serotonin concentration, the baseline was again established by returning the electrode to a pool of without serotonin. The difference was measured directly off of the oscilloscope screen. A suction electrode was used to attach to a nerve bundle so that Grass SD9 Stimulator could stimulate the whole cluster. A microelectrode was placed in one of the ganglion's nerves to measure activity. The carbon-fiber microelectrode was placed under a group of cells. The four track tape recorder was used to record the responses of the current and command voltage for the carbon fiber microelectrode and the membrane potential which were transferred to a chart recorder for analysis (Gould Recorder 220). During the experiment the table was tapped. Testing NIE-115 neuroblastoma cells: This experiment was done on a patch clamp set-up (List-Medical Patch Clamp LM-EPC 7 amplifier and LIM-EpC 7 5/n 583644 headstage) and was filtered at 3khz (Frequency Device, 902 LPF). A DAT tape recorder (Sony High Density Linear A/D D/A tape recorder) was used to record the amperometry data at +.6V. The same type of oscilloscope was used. A gravity perfusion system was used to change concentrations. A calibration was done for epinephrine both at the beginning and the end of the experiment. The effects of 1 mM carbachol alone were also tested. The electrode was positioned with a very long carbon filament so the carbon filament was resting on a group of NIE-115 mouse neuroblastoma cells which had been raised in culture (Mathes, 1992). The cells were kept normal external saline (Mathes, 1992). Carbachol was then added. Results: Testing electrodes: The behavior of typical carbon-filament microelectrode under cyclic voltammetry is shown in Figure 4 if it is viewed with both current and voltage on the Y axis and time on the Xaxis. The voltage trace is the triangle wave. The current is the trace showing the curve. Each microelectrode will show a different current trace, but the basic shape should be the same. It should show extremes at both ends of the current trace. It should also show another peak as the current increases past zero in the presence of an oxidizable substance, as it does in the lower trace of Figure 4. A working electrode in cyclic voltammetry should look like Figure 5A if it is graphed xy with current the y axis and voltage the x axis, a voltammogram. Each electrode again has a different basic shape. The bettei the electrode and the slower the scan rate the thinner the leak and capacitance is, which causes open middle to be reduced. A functional electrode also has to show another peak when an oxidizable material is added that should be at the positive side of the increasing voltage, the upper right of the curve. The peak increases and moves to the right as the concentration increases. At high concentrations, as the 25mM in Figure 5A, other differences can be seen. There are two troughs, the first smaller than the other. Figure 5B is graph of computer subtraction the baseline from the after oxidation trace of three different epinephrine concentration's voltammogram and shows these other peaks very well. To eliminate leak and capacitance currents a baseline was subtracted from the voltammetric traces. Voltage and current were digitized (the computer programs used were Clampex version 4.10 with Amp and Clampfit Version 5.5 from Axon Instruments on an ZTS 386 computer with the A/D input designed by Stuart Thompson) and the base OM epinephrine was subtracted from the curve with the various epinephrine concentrations. This was done by starting with the minimum point taking one whole cycle of the curve to get the two traces in the same phase. The level of the current corresponds almost directly to the concentration. Figure 6 shows this with the current trace both before and after a SmM increase in concentration. The epinephrine calibration was done this way. The plot of concentration and current is Figure 7A. The plot of the log of concentration and log of current is Figure 7B. The best fit linear fit for this line (Cricket Graph 1.3) was y-5.43 + 0.86x which makes the power law for this relationship y- 269153.47x0.86 (x-M, y-PA). Figure 7C is a plot of log of concentration and log of current for serotonin. This line has a slope of y=3.17 +0.14x (Cricket Graph 1.3). This was the best serotonin calibration that was obtained. It was preformed before the Pleurobranchea experiment. Figure 9A shows the cyclic voltammetry during the Pleurobranchea experiment. This change occurred without stimulation. Figure 9B shows a trace from the same experiment. The jump in amperometry corresponds to a physical shock. The jump was 400 pA and decayed back down to the baseline. Figure 9A does not give quantitative data, because a voltammetry calibration was not done before the experiment. However it does raise and answer some interesting questions. First it does not correspond to any of the nerve stimulation that were given to the ganglion. It was a completely spontaneous event, independent of activity in the one cell of the cluster whose membrane potential is measured on the lower trace. However the current trace is not ambiguous. It shows a dramatic increase in current and a very definite peak at the raising positive voltage. No apparent harm was done to the cells when shaken for the cell with the current clamped microelectrode continued to spike normally. Figure 9B gives more quantitative data but has more questions about it's validity. The signal was triggered by shaking the table, but the signal is what was expected for release during amperometry. Based on the calibration in Figure 7C 90 uM serotonin would give this response. The mouse neuroblastoma showed this same pattern, but in a more controlled fashion. After the carbachol, which is not oxidizable, was added to the NIE-115 cells the current rose dramatically from baseline by 1.6nA and then slowly decayed as seen in Figure 10. When carbachol was added without cells there was no change in the baseline. What is remarkable is the slow time of the rise if what is measured is vesicle release. However, The carbon-filament used was long and lay over a cluster of the cells. Therefore the data represents the sum of the release from a number of cells which would account the slow overall release. The calibrations were off and epinephrine, not norepinephrine, was used, but the concentrations recorded should correspond to norepinephrine levels between 1 and 100 UM. Discussion: The major finding of this study is that electrochemical detection of the catecholamine neurotransmitters serotonin, in Pleurobranchea pedal ganglion, and norepinephrine, in NIE-115 neuroblastoma cells, is possible. The first lessons is that there are four basic types of electrodes. Diagrams of the first two types are in Figure 8. Electrodes of the first type were broken. The current saturated in the direction of the voltage, positive for positive voltages and negative for negative voltages, except around zero where they changed. The second group (8B) were electrodes that had no carbon-filament exposed. They showed a flat trace until the oscilloscope gain was increased to the maximum and then showed a parallelogram trace with a negative current. The third group showed normal voltammograms, but showed no response to epinephrine. The fourth group were functional electrodes which showed normal voltammograms and showed an increase of current with epinephrine. Occasionally the functional electrodes would show a trace like the third trace on Figure 8. This was simply a problem of the gain being too high (.5pA/mV instead of .IpAlmV) for that particulai electrode. Higher gains often gave even more problems because of mismatched impedance between the electrode and the headstage. The calibration for epinephrine is very promising. Figure 7B has a slope near 1,.86. This is close to theory. Even so I would advocate a dose response calibration before and after each experiment to see any deviations from normal because each electrode is different and use changes its response. As the electrodes are used they get dirty (Fishman, 1992) which reduces their sensitivity. The response of a dirty electrode would be less current change for a given concentration change which is what was observed. 10 Figure 7C is more troubling. The response to serotonin is much worse than expected. The slope is a seventh of the theoretical value. It has been stated that after being oxidized serotonin forms a film over the carbon-fiber (Wightman, 1988). Again this dirtying of the carbon-filament surface results in the slope of the log graph less than one. The value, .14, indicates a very dirty electrode. This loss of sensitivity with corresponds with experience with serotonin during other attempted calibrations. For example, an electrode started with strong response of 900 pA from 10-3M, and yet when the calibration was finished, the electrode showed no change at 10-2M. It is difficult to know exactly what compound's oxidation is being recorded using this technique. But I feel secure in the correct identification of the compounds in these experiments. In the Pleurobranchea experiment, the first trace is serotonin. It is released without any outside disturbance of the cell, and it is a single clean peak in the oxidation which suggest that only one oxidizable species is involved. The amperometric measurement is more difficult to identify as secreted serotonin. Although is seems very unlikely, it could be simply the result of cells being lysed and spilling their oxidizable contents onto the electrode. However, the cyclic voltammetric measurements clearly show that the cells are able to release serotonin. If the electrode really had broken into a cell, the oxidation would occur more rapidly, and the decay should have been slower if it was trapped in the cell's contents. Norepinephrine occurs in NIE-115 cells ten times as much as the other catecholamines found, dopamine and serotonin. Therefore it seems likely that it is the major component of the release. The release that is linked to carbachol is without action potentials, so this release occured without spikes. Carbachol is detected by the muscarinic acetylcholine receptors which in this preparations does not cause action 11 potentials. The conventional pathway-action potential, calcium influx, and then secretion—is bypassed. This shows that the calcium influx may be sufficient to get norepinephrine release. The concentration of transmitters released in these systems has not been measured before. Both concentrations are in uM range and localized in a very small area. This is a very powerful signal even taking into account the artificial nature of the stimulus. Wightman reported total concentration of catecholamine release from bovine chromaffin cells of 10 uM (1991). Serotonin receptors have very high affinities, upper nM concentrations, for serotonin (Wang and Peroutka, 1988). This means that there would be a huge detection of these signals. Again I must state that these concentrations are high, but the exact level I am unsure of, because the calibration for the two experiments were poor. I would like to stress that these results are largely suggestive. They raise some interesting questions about the biological systems which were observed. However, they were done in this study to examine the behavior of the carbon-fiber microelectrodes in biological systems. In this respect, they were complete successes. The phenomena exhibited, whatever their cause, were very much what was expected for both voltammetry and more important amperometry in two different systems. There several areas yet to be explored. Reliable dose-response calibrations is needed before anything else. The technique of cutting the carbon-fiber very close to the glass or polishing the surface also needs to be improved to improve detection. Now that this procedure is working, the options are almost endless. 12 Acknowledgements: I am going to make this section all together too long because I will attempt the impossible and thank all those who deserve it. I will undoubtedly forget some of those who need it (like Mom who is always forgotten at times like this) I apologize profusely, the mistake was not because I have forgotten your contribution but is entirely the fault of my sleep debt. First for the scientific thanks, both Rhanor Gillette, Rob Swezey, and Harvey Fishman gave me scientific aid when I requested it. In fact, Harvey Fishman saved my life and gave me the key that allowed me to get it working. I would like to thank them all for helping a lil' undergrad as he floundered his way through his project. I would also like to thank Gilly for the stream of equipment he lent me (even after I started breaking his stuff) The first person who needs to be thank is Stuart, who if he had any sense, would have kick me out of his lab the first time I walked in the door or at least when I screwed up with the electrode puller after he spent 15 minutes telling me not to screw up. Thanks for giving me a change to screw up again (and again, and again....). Stuart was also "totally key" in getting both the technical and theoretical aspects of this project working. Sorry I didn’t listen and follow more than I did. Sam Wang needs to be thanked not only for scientific help, but more importantly for trying to give me an education in bad music and helping me keep my sanity. I also would like to thank you for not just doing this project yourself (although you would certainly have done a better job). I pray that you'Il find your blonde (my addition) tortured artist very soon. Iwould also like to thank Chris for getting me to experience first hand how gross coffee really is and showing me the terrors of married life and those totally killer NIE-115 cells. 13 Now for the random thanks. My housemates and lab partners need to be thanked to putting up with my insanity-"overwork leads to all forms of 5 s. 1 d like to thank Coke giving me the calories and caffeine to metal disorder: survive this quarter. I'd like to thank my body for putting up with all the terrible things l've been feeding it. Most important l’d like to thank 107.1 rocks and 94.5 who I listed to almost continually (except 5 hours for sleeping some Garth Brooks, Right Said Fred, and a little Debbie Gibson). To them I owe for holding what little sanity I have left together. I shall miss 107.1 more than anything else about Monterey. My final thanks is to my father. All those nights of reading me to sleep from Crucible: The Story of Chemistry had some effect after all. Thanks for making me what little of a scientist I am. Literature Cited: Bard, Allen and Larry Faulkner. Electrochemical Methods: Eundamentals and Applications. John Wiley & Sons, Inc.:New York, 1980. John E. Baur, Eric W. Kristensen, Leslie J. may, Donna J. Wiedemann, and R. Mark Wightman. 1988. "Fast-Scan Voltammetry of Biogenic Amines." Analytiçal Chemistry. Vol 60, 1268-1272. Chow, Robert H., Ludolf von Ruden, & Erwin Neher. 1992. "Delay in vesicle fusion revealed by electrochemical monitoring of single secretory events in adrenal chromaffin cells." Nature. Vol 356, 60-63. Fishman, Harvey. Personal Communication. 1992. Stanford University Chemistry Department. Herken, Hans. et al. 1982. "Mouse Neuroblastoma clone NIE-115: A suitable model for studying the action of dopamine agonist on tryosine hydroxylase activity." Biological Pharmacology. Vol 31. No. 7, 1279-1282. Mathes, Chris. M. 1992. "Intracellular calcium release in NIE-115 neuroblastoma cells is mediated by the MI muscarinic receptor subtype and is antagonized by MCN-A-343." Unpublished. Accepted Brain Research. Peroutka, Stephen and Samuel Wang. The Serotonin Receptors. The Humana Press: Clifton, NJ. 1988. Wightman, R. Mark. 1981. "Microvoltammetric Electodes. Analytical Chemistry. Vol 53, 1125-1134. Wightman, R., J. Jankowski, R. Kenedy, K. Kawagoe, T. Schroeder, D. Leszczyszyn, J. Near, E. Diliberto, Jr., and O.Viverous. 1991. "Temporally resolved catecholamine spikes correspond to single vesicle release from individual chromaffin cells. Proc. Natl. Acad. Sci. U.S.A. Vol. 88, 10754¬ 10758. 15 1985. Wilbur, Karl. The Mollusca. Vol 8. Academic Press, Inc: Orlando Fl. Figure Legend: Figure 1-The oxidation of the catecholamines occurs by this general two electron process which is shown here with epinephrine. Figure 2-The microelectrode is given a potential by the amplifier. At an adequate potential, the fiber oxidizes the epinephrine in contact with it. This results in two electrons being transferred to the carbon filament which is detected by the amplifier and recorded as higher current. Figure 3- A) A photograph of the carbon-fibers. B) A photograph of a microelectrode. Figure 4-A picture from a carbon-fiber microelectrode of a display of the voltage and current on the y axis and time on the x axis. The voltage is a triangle wave which correspond to the potential of the carbon-fiber. The current is the current flowing through the carbon¬ fiber. The upper trace shows the background current. The lower trace shows the current while oxidation is occurring. Figure 5-A) An xy plot of voltage and current, voltammogram. The inner trace is the background. The outer trace is in the presence of 25mM epinephrine. The additional current is from oxidation of epinephrine. B) This is a voltamogram with the background subtracted. There are three different concentrations here. The electrode shows no change at O.0lmM epinephrine. O.ImM shows a 17 very distinct peak. ImM has a very large response. In fact, the current has saturated the amplifier so the top of the peak is flattened. Figure 6-Amperometric response to a 5mM increase in epinephrine concentration. Figure 7-A) A graph of the amperometric current vs. concentration of epinephrine. B) A graph of the log the concentration and log of the current for epinephrine with the baseline subtracted. It has a slope of .86. C) This is a graph of the concentration and log of the change in current for serotonin. It has a slope of .14. Figure 8—A) is the xy plot of voltage and current for a broken electrode. The current saturates the amplifier for almost the entire trace. B) An xy plot for an electrode with the carbon-fiber missing or covered. The current is very small and negative, and the curve it makes is a very angular shape. C) A normal electrode at too high gain. When the gain is reduced the trace looks like Figure 6A. Figure 9—Electrochemical measurement of serotonin in pedal ganglion of Pleurobrachea A) The trace is the current with cyclic voltammetry. The first half shows the baseline current flow. Then, a change occurs and the current increases and changes its shape. Next, the gain of the amplifier was turned down to prevent saturation. After the change, an additional peak can be seen that corresponds to oxidation at the carbon filament. B) The increase in amperometic 18 current corresponds to a mechanical stimulation. The voltage of the electrode was +.6V. Figure 10—Two digitized scans overlaid. The lower is the baseline current done with the electrode in the bath resting on the NIE-115 neuroblastoma cells with a holding voltage of +.6V. The second, higher, is current after carbachol was added. CH5 HO-CCh NH OH OH Epinephrine Figure CH — NH HO-C-CH. o- Electron Acceptor 0- CH5 HO-C—CH-NH Oxidized Epinephrine 1. CI Na epi Carbon Filament at +5V epi Na Depletion Lager Figure 2 To Amplifier Ag/AgCl Na' CI epi epi Na An epinephrine molecule being oxidized Na CI Ground Ag/AgCl C1 B) Figure 3 OM 25mM Epi igure 2 5 sec 50 n4 B, A) 18.0 16.0 - 14.0 - 12.0 - 10.0 - 8.0 - 6.0 - 4.0 - 2.0 - O.0 -2.0 - -4.0 - -6.0 -8.0 - -10.0 - -12.0 - -14.0 — -16.0 — -1.2 Figure 5 10nA 1V Soførtst —0.4 -0.8 0.4 O.8 Voltoge (V 1mM Epi 1mM OimM 1.2 Figure 6 200 pA sed 0 A C Figure 7 3+ 2.8 — 2.6 — 2.4 - 2.2 - 1.8 - 1.6 - 1.4 - 1.2 - 0.8 - 0.6 - 0.4 - 0.2 - o Of+O0f-OBE-OTE-ORE-OSE-OSE-ORE-OSE-OSE-OOE-OGE-OAE-OREOSE-O3-OD-00 Concentration on Epl (M) 3.5 - 2.5 - 1.5 - 0.3- o + 95 4 -6.7 -8.3 -8 -5.7 -5.3 -3 -4.7 -4.3 —4 -3.7 -3.3 -3 -2.7 -2.3 -2 Log of Concentration of Epl (M) 274 — 272 - 2.7 - 268 - 266 - 264 262 258 - 236 - 2.54 252 - 248 - 246 244 + A — o B Figure 8 — Voltage 120 zero current 100my below zero current C J ( 2 Current (pA) S