Journal of Neuroscience Methods, 38 (1991) 81-88
81
© 1991 Elsevier Science Publishers B.V. 0165-0270/91/$03.50
NSM01244
Applications of piezoelectric fluid jetting devices to neuroscience research Gary L. Bernardini 1, B. Allan R a m p y 1, Gailyn A. Howell ~, Donald J. Hayes 2 and Christopher J. Frederickson 1 I Laboratory ofNeurobiology Unit'ersity of Texas at Dallas, Richardson, TX 75080 (U.S.A.) and -" MicroFab Technologies, bw., Piano, TX 75074 (U.S.A.) (Received 2 July 1990) (Revised version received 15 March 1991) (Accepted 19 March 1991)
Key words: Fluid dispensing; Drug application; Pumps; Histochemistry; Neurophysiology; Brain slices Piezoelectric p u m p s or "jets" are used in industry for precise dispensing of small volumes of fluids. In the present work we have tested the feasibility of using these piezoelectric devices for dispensing fluids in neurobiological research. In one experiment, 70 picoliter (pl) droplets of histochemical reagent were jetted onto discrete targets of frozen tissue sections, and qualitative and quantitative histochemical studies were done on the small (170 tzm diameter) circle of tissue wet by the droplets. In the second experiment, 70 pl droplets of neuroactive drugs were jetted onto brain tissue slices while recording single neurons extracellularly in vitro, and the effects of the drugs were found to vary systematically as a function of the number of drops and the distance between drop application and the recorded neuron. The results indicate that piezoelectric jets could have wide application for dispensing fluids in neurobiological research.
Introduction
Biomedical research often involves dispensing small amounts of reagents, a task to which there are presently three approaches: (i) mechanical pumps, (ii) pressure ejection (Baranyi et al., 1987), and (iii) iontophoresis (Zieglansberger et al., 1974). Each of these methods has limitations peculiar to it, and problems such as limited range of ejection or pumping rates, clogged plumbing lines, dead volumes, erratic dispensing, and (with iontophoresis) effects of ejection currents, are familiar to users. In the present work we have done preliminary
Correspondence." Christopher J. Frederickson, Laboratory for Neurobiology, University of Texas Richardson, T X 75080, U.S.A.
at
Dallas,
Box 688,
evaluation of piezoelectric pumps, or "jets" for fluid dispensing. These devices are used as jets in ink-jet printing, where they typically dispense droplets onto paper from remote, computer-controlled heads, with the placement of the fused droplets forming the printed characters. Droplet rates up to 20000 per s and placement accuracy of approx. 5 /xm are typical in these commercial applications (Bogy and Talke, 1984; Dijksman, 1984). Two applications of piezoelectric devices were tested: (i) dispensing of "microdrops" of histochemical reagents onto cryostat sections of brain tissue for histochemical studies of discrete cytoarchitectonic brain regions, and (ii) dispensing of microdrops of drugs onto live brain slices in vitro while recording from single neurons for studies of the physiological localization of drug-sensitive sites.
82 Methods
Piezopumps and dispensing apparatus The piezoelectric dispensing jets were made by fusing a piezoelectric crystal to the side of a capillary pipette. A small section of polyethylene tubing connected the back end of the pipette to a 0.5 mL fluid reservoir, and the front end of the pipette was pulled to a small diameter. The entire assembly was encased in an aluminum shell, and the exposed tip of the pipette was polished to give a final orifice with a 5 0 / z m inner diameter (Fig. 1). Typical crystal lengths are 10 mm and capillary lengths are 25 ram. When a voltage is applied between the inner and outer surfaces of the crystal, the thickness of the annulus changes, thus producing a change in diameter of the annulus. This change in diameter causes the glass tube to deform and produces a volume change in the fluid inside the glass tube. This volumetric change causes p r e s s u r e a n d velocity transients to occur in the fluid and these are directed so as to produce a drop that issues from an orifice. The nominal
voltage required to produce a drop is 61) V, and a typical waveform is a pulse of 40 ~zs duration. For jetting fluid drops onto tissue, the piezoelectric device and fluid reservoir were mounted on an X-Y-Z micropositioner which was attached to the body of a stereomicroscope. The jet was then aligned relative to the microscope body so that fluid drops dispensed from the piezoelectric jet would land on whatever target was positioned in the microscope field of view, in the plane of focus, directly under the cross hairs of a viewing reticule. To select a new target for sequential drop application, the jet/cross hair was simply aimed at a new spot on the target by moving either the target or the microscope head, both of which were mounted on separate X-Y positioners. In the typical configuration, the jet was positioned about 10 mm above the target. To drive the piezoelectric jet, monophasic pulses of 50-70 V, 1-20 gs rise times, 10-30/xs duration, and 5-20 /xs fall time were generated by Techtronix pulse formers, amplified by a video amplifier, and delivered to the piezoelectric crys-
Fig. 1. Upper panel: assembled piezoelectric jet (without fluid reservoir) ready for use. The entire unit is approximately 25 mm in length. Lower panel: drops in flight exiting the tip of a jet: The drops emerge as cylinders, sometimes assume a "barbell" shape, then coalesce into spheres. The two spheres at the right are about 70 # m in diameter.
83 tal. Single pulses were used to dispense single droplets, and trains of 0.01-1000 Hz were used to apply multiple drops.
Microdrop histology To test the suitability of piezoelectric jets for histologic applications, unfixed cryostat sections of mouse brain were placed on the stage of the microscope, and positioned so that a specific cytoarchitectonic region of interest lay under the sighting cross hair of the microscope. Next, 1-50 drops of staining reagent were applied on a single spot at slow enough rates (0.2 Hz) so that each drop dried completely before the next was applied. By this method from 1- to 50-fold multiples of reagent molecules could be applied to the spot on the tissue that was wetted by the ejected fluid droplet. For the histology testing, the histofluorescent zinc probe, toluene sulfonamide quinoline (TSQ) (Frederickson et al., 1987) was dissolved in 0.1 M universal buffer in deionized water (15 /~g T S Q / m l of solution) and jetted onto either the hilus of the dentate gyrus or field CA1 of horizontal 22 ~ m cryostat sections of the mouse hippocampal formation. The tissue sections were then transferred to a fluorescence microscope, illuminated with 365 nm ultraviolet light, and the intensity of fluorescent emissions was measured with a microspectrophotometer (Farrand) attached to the microscope. Microdrop chemical stimulation of brain slice in L,itro Coronal slices (500 ~m) of mesencephalon from young (15-20 g) mice were conventionally prepared (Bernardini et al., 1986; Howell et al., 1984) and maintained in a slice chamber in artificial CSF medium. The tissue was held on a net with the upper surface of the slice at the interface between the liquid medium and the warm, moist gas (95% 0 2, 5%CO2). Fresh medium was perfused continuously through the 2.0 mL tissue chamber at approximately 0.5 ml per min. Single neurons in the vicinity of dopamine-containing groups A9 and A10 were recorded extracellularly via glass pipettes (2-3 M~Q) and were monitored on an oscilloscope and recorded on magnetic
tape. Only neurons with the slow firing, longduration, action potentials characteristic of dopaminergic cells (Bunney et al., 1983; also see Fig. 3) were studied. The piezoelectric jet was positioned about 10 mm from the surface of the brain slice and aligned so that ejected droplets hit the slice surface directly under the cross hairs of the stereomicroscope through which the tissue was viewed. Thc piezojet and microscope were mounted together on an X-Y positioner, so that the site of drug application could be changed by simply moving the microscope/jet assembly to a new position. Microdrops of deionized water, 2-50 mM sodium glutamate and 2-10 mM dopamine (in deionized water) ejected at rates from 0.01 to 1000 Hz were tested. Each of these solutions was filtered before use.
Results
Dispensing histologic stains and neuroactil'e drugs In the present experiments, piezoelectric jets with orifices of approx. 50 # m in diameter were selected. Droplets emitted by these jets emerged from the orifice as cylinders of fluid, approx. 50 ~ m in diameter, which then coalesced in flight to form spherical drops about 70 ~ m in diameter (Fig. 1). To determine drop volume, 120 000 drops of isopropanol (at 2000 drops per s) were collected in pre-weighed microfuge tubes, and the accumulated fluid was weighed. For 3 separate jets tested, 11.4, 12.3 and 12.5 mg of fluid were emitted, indicating that the mass of single drops was from 0.095 to 0.104 p,g. Time-lapse photography (Fig. 1) indicated that the velocity of the drops in flight was approximately 3 m / s . When the distance between the jet orifice and the target was in the range of 1-2 cm, successive individual drops landed on the target in virtually exact (5-10 /.tm) superposition, (Fig. 2). The several different aqueous solutions (e.g. TSQ in buffer, sodium glutamate, and dopamine in deionized water) used for neurobiological experiments were all found to be suitable for dispensing; with the appropriate electrical pulse height and waveform, uniform drops were consistently obtained at ejection rates from 0.01 to 1000
84 4
-3 iii Z taJ
o _/ LI.
",,,'!1,: ".
/,,,
/.-
|
1
I
~
2
I
4
I
]
I
I
1
8
l
1
NUMBER of DROPS
Fig. 2. Top left image is a fluorescence photomicrograph of a section of mouse hippocampal formation: The fluorescent s p o t (arrow) was made by jetting three successive drops of TSQ onto the tissue, then illuminating the tissue section with UV: light. The drawing lower left is a tracing made from the upper image with the granule neuron stratum (G), and the outline of the dentate gyrus (thick line) indicated. Locations of granule and hilar neurons are shown schematically for scale. The data graphed on the right are the fluorescence intensity readings (arbitrary units) obtained when 1-8 drops of TSQ were jetted onto either the zinc-rich hilus of the dentate gyrus (as shown in the figures on t h e left) or onto the stratum radiatum of hippocampal field CA1, a region with only moderate zinc. Scale bar = 200/xm.
Hz. The pulse parameters used for dispensing deionized water were also found to be suitable for dispensing dilute (1-50 mM) aqueous solutions. Gravimetric determinations of drop size for one such solution (10 mM dopamine) indicated that the individual drops dispensed contained 72 pl of solution. (Because solutions differ in surface tension and viscoscity, calibrations would normally be done using the same solution to be used in an experiment.)
Histologic staining When drops of the histochemical reagent, TSQ, were jetted onto cryostat sections of brain tissue, the TSQ reacted with tissue zinc and formed the
fluorescent zinc/TSQ chelate complex. Individual drops of TSQ produced circles of fluorescence about 170 /zm in diameter on the tissue, and successive drops (each applied after the preceding drop had dried) produced a progressively brighter fluorescence within the same 170 tzm circle (Fig. 2). In these "microdrop" histochemical tests, an individual drop of the staining solution contained approx. 4 x 10-15 mol TSQ (70 pl drop of 60 IzM TSQ solution), and the cylindrical volume of tissue wet by a single drop (ca. 0.5 nl; 170 ~zm diameter x22 /~m thick) was estimated to contain about 150 x 10 ~5 mol chelatable zinc (assuming 250 IzM of zinc in the tissue; see Frederick-
85 son et al., 1983). Titration experiments were therefore undertaken to determine whether fluorescence from the tissue would increase monotonically with successive drops of reagent until a 1:1 molar ratio of T S Q / z i n c was achieved at approx. 37 (150/4) drops of TSQ solution. Figure 2 shows that fluorescent yield from the tissue did increase with successive drops. Moreover, the tissue zone with a high concentration of zinc that was studied (dentate hilus; see Frederickson et al., 1983, 1987) yielded much greater fluorescence than a comparison region (stratum radiatum of hippocampal field CA1) that has only moderate amounts of zinc. Attempts to carry these titration experiments to stoichiometric completion (i.e., to 1:1 molar ratios of zinc/TSQ), however, were not successful. Regardless of the total number of drops applied, peak fluorescence from the spot was never as high as that observed when the same region of tissue was saturated in TSQ by immersing the section in a stoichiometric excess of TSQ in a staining dish.
Focal application of drugs to brain slices during single neuron recording The piezojets proved entirely satisfactory for applying drugs to brain tissue slices during extracellular recording of neurons. The electrical pulse driving the piezoelectric crystal generated a transient artifact (ca. 50 msec) in the single unit trace, but because that transient gave a suitable marker for the time of drop ejection, no effort was made to remove the signal by shielding the leads a n d / o r housing of the piezojet. Aside from the electrical artifact, there was no other apparent effect of drop ejection per se upon the slice preparation or recording. Thus, when single drops of deionized water were jetted onto the tissue at the location of the recording pipette tip, no change in the action potential waveform or in the firing pattern of the recorded neurons could be detected. Apparently neither the dilution of extracellular fluids by the 70 pl volume nor the kinetic energy (about 6 × 10 -l° J / d r o p ) of the drop landing on the tissue was sufficient to perturb the recording pipette placement or the slice preparation. Stable single unit
recordings were routinely obtained from over 30 different neurons during drug application, even (in a few cases) while applying up to 1000 drops of fluid directly on the recording site at a rate of 1000 drops/s. When neuroactive compounds were jetted onto the slices, neuronal responses varied systematically according to the compound applied, the number of drops (drug dose), and the distance between the recorded neuron and the site of drug application. Glutamate, for example, either caused no change in firing rate or caused a brisk acceleration of firing in almost all neurons tested. In general, cells that were encountered more than 200 Ixm below the surface of the slice showed little or no response. For cells near the surface of the slice, the response to glutamate depended upon the number of drops applied and the proximity of the drop placement to the cell. Thus, drops placed directly at the point where the recording pipette penetrated the surface of the slice generally produced maximal acceleration whereas drops placed as little as 250 ~tm away from that point (Fig. 3b) typically had no detectable effect at all. One interesting exception to the latter result, which was observed in three different neurons, was a distinct slowing of the recorded neuron observed when glutamate drops were applied 200-400 # m off target. Presumably, interneurons inhibitory to the recorded neurons were excited by the glutamate in these cases. Application of dopamine to dopamine-containing neurons by conventional superfusion or iontophoresis methods produces relatively rapid, marked depression of firing rates (Bernardini et al., 1986; Bunney et al., 1983). In the present experiments, application of dopamine drops gave the same result. Two to 5 drops of dopamine (10 mM; 70 pl) applied directly to the recording site consistently depressed firing rates for 10-30 s (Fig. 4). Drops applied 200 to 400 g m off target generally had no detectable effect on the activity of the recorded cell. In one case, however, when drops were aimed at the distal dendrites of the dopaminergic neurons (which ramify in parts reticulata), the neuron being recorded (about 300 ~zm distant, in pars compacta) showed clear and long-lasting depression of firing.
86
,o
q I
//
m
JP'
,,, 5 0
~
,
I-.
81
© 1991 Elsevier Science Publishers B.V. 0165-0270/91/$03.50
NSM01244
Applications of piezoelectric fluid jetting devices to neuroscience research Gary L. Bernardini 1, B. Allan R a m p y 1, Gailyn A. Howell ~, Donald J. Hayes 2 and Christopher J. Frederickson 1 I Laboratory ofNeurobiology Unit'ersity of Texas at Dallas, Richardson, TX 75080 (U.S.A.) and -" MicroFab Technologies, bw., Piano, TX 75074 (U.S.A.) (Received 2 July 1990) (Revised version received 15 March 1991) (Accepted 19 March 1991)
Key words: Fluid dispensing; Drug application; Pumps; Histochemistry; Neurophysiology; Brain slices Piezoelectric p u m p s or "jets" are used in industry for precise dispensing of small volumes of fluids. In the present work we have tested the feasibility of using these piezoelectric devices for dispensing fluids in neurobiological research. In one experiment, 70 picoliter (pl) droplets of histochemical reagent were jetted onto discrete targets of frozen tissue sections, and qualitative and quantitative histochemical studies were done on the small (170 tzm diameter) circle of tissue wet by the droplets. In the second experiment, 70 pl droplets of neuroactive drugs were jetted onto brain tissue slices while recording single neurons extracellularly in vitro, and the effects of the drugs were found to vary systematically as a function of the number of drops and the distance between drop application and the recorded neuron. The results indicate that piezoelectric jets could have wide application for dispensing fluids in neurobiological research.
Introduction
Biomedical research often involves dispensing small amounts of reagents, a task to which there are presently three approaches: (i) mechanical pumps, (ii) pressure ejection (Baranyi et al., 1987), and (iii) iontophoresis (Zieglansberger et al., 1974). Each of these methods has limitations peculiar to it, and problems such as limited range of ejection or pumping rates, clogged plumbing lines, dead volumes, erratic dispensing, and (with iontophoresis) effects of ejection currents, are familiar to users. In the present work we have done preliminary
Correspondence." Christopher J. Frederickson, Laboratory for Neurobiology, University of Texas Richardson, T X 75080, U.S.A.
at
Dallas,
Box 688,
evaluation of piezoelectric pumps, or "jets" for fluid dispensing. These devices are used as jets in ink-jet printing, where they typically dispense droplets onto paper from remote, computer-controlled heads, with the placement of the fused droplets forming the printed characters. Droplet rates up to 20000 per s and placement accuracy of approx. 5 /xm are typical in these commercial applications (Bogy and Talke, 1984; Dijksman, 1984). Two applications of piezoelectric devices were tested: (i) dispensing of "microdrops" of histochemical reagents onto cryostat sections of brain tissue for histochemical studies of discrete cytoarchitectonic brain regions, and (ii) dispensing of microdrops of drugs onto live brain slices in vitro while recording from single neurons for studies of the physiological localization of drug-sensitive sites.
82 Methods
Piezopumps and dispensing apparatus The piezoelectric dispensing jets were made by fusing a piezoelectric crystal to the side of a capillary pipette. A small section of polyethylene tubing connected the back end of the pipette to a 0.5 mL fluid reservoir, and the front end of the pipette was pulled to a small diameter. The entire assembly was encased in an aluminum shell, and the exposed tip of the pipette was polished to give a final orifice with a 5 0 / z m inner diameter (Fig. 1). Typical crystal lengths are 10 mm and capillary lengths are 25 ram. When a voltage is applied between the inner and outer surfaces of the crystal, the thickness of the annulus changes, thus producing a change in diameter of the annulus. This change in diameter causes the glass tube to deform and produces a volume change in the fluid inside the glass tube. This volumetric change causes p r e s s u r e a n d velocity transients to occur in the fluid and these are directed so as to produce a drop that issues from an orifice. The nominal
voltage required to produce a drop is 61) V, and a typical waveform is a pulse of 40 ~zs duration. For jetting fluid drops onto tissue, the piezoelectric device and fluid reservoir were mounted on an X-Y-Z micropositioner which was attached to the body of a stereomicroscope. The jet was then aligned relative to the microscope body so that fluid drops dispensed from the piezoelectric jet would land on whatever target was positioned in the microscope field of view, in the plane of focus, directly under the cross hairs of a viewing reticule. To select a new target for sequential drop application, the jet/cross hair was simply aimed at a new spot on the target by moving either the target or the microscope head, both of which were mounted on separate X-Y positioners. In the typical configuration, the jet was positioned about 10 mm above the target. To drive the piezoelectric jet, monophasic pulses of 50-70 V, 1-20 gs rise times, 10-30/xs duration, and 5-20 /xs fall time were generated by Techtronix pulse formers, amplified by a video amplifier, and delivered to the piezoelectric crys-
Fig. 1. Upper panel: assembled piezoelectric jet (without fluid reservoir) ready for use. The entire unit is approximately 25 mm in length. Lower panel: drops in flight exiting the tip of a jet: The drops emerge as cylinders, sometimes assume a "barbell" shape, then coalesce into spheres. The two spheres at the right are about 70 # m in diameter.
83 tal. Single pulses were used to dispense single droplets, and trains of 0.01-1000 Hz were used to apply multiple drops.
Microdrop histology To test the suitability of piezoelectric jets for histologic applications, unfixed cryostat sections of mouse brain were placed on the stage of the microscope, and positioned so that a specific cytoarchitectonic region of interest lay under the sighting cross hair of the microscope. Next, 1-50 drops of staining reagent were applied on a single spot at slow enough rates (0.2 Hz) so that each drop dried completely before the next was applied. By this method from 1- to 50-fold multiples of reagent molecules could be applied to the spot on the tissue that was wetted by the ejected fluid droplet. For the histology testing, the histofluorescent zinc probe, toluene sulfonamide quinoline (TSQ) (Frederickson et al., 1987) was dissolved in 0.1 M universal buffer in deionized water (15 /~g T S Q / m l of solution) and jetted onto either the hilus of the dentate gyrus or field CA1 of horizontal 22 ~ m cryostat sections of the mouse hippocampal formation. The tissue sections were then transferred to a fluorescence microscope, illuminated with 365 nm ultraviolet light, and the intensity of fluorescent emissions was measured with a microspectrophotometer (Farrand) attached to the microscope. Microdrop chemical stimulation of brain slice in L,itro Coronal slices (500 ~m) of mesencephalon from young (15-20 g) mice were conventionally prepared (Bernardini et al., 1986; Howell et al., 1984) and maintained in a slice chamber in artificial CSF medium. The tissue was held on a net with the upper surface of the slice at the interface between the liquid medium and the warm, moist gas (95% 0 2, 5%CO2). Fresh medium was perfused continuously through the 2.0 mL tissue chamber at approximately 0.5 ml per min. Single neurons in the vicinity of dopamine-containing groups A9 and A10 were recorded extracellularly via glass pipettes (2-3 M~Q) and were monitored on an oscilloscope and recorded on magnetic
tape. Only neurons with the slow firing, longduration, action potentials characteristic of dopaminergic cells (Bunney et al., 1983; also see Fig. 3) were studied. The piezoelectric jet was positioned about 10 mm from the surface of the brain slice and aligned so that ejected droplets hit the slice surface directly under the cross hairs of the stereomicroscope through which the tissue was viewed. Thc piezojet and microscope were mounted together on an X-Y positioner, so that the site of drug application could be changed by simply moving the microscope/jet assembly to a new position. Microdrops of deionized water, 2-50 mM sodium glutamate and 2-10 mM dopamine (in deionized water) ejected at rates from 0.01 to 1000 Hz were tested. Each of these solutions was filtered before use.
Results
Dispensing histologic stains and neuroactil'e drugs In the present experiments, piezoelectric jets with orifices of approx. 50 # m in diameter were selected. Droplets emitted by these jets emerged from the orifice as cylinders of fluid, approx. 50 ~ m in diameter, which then coalesced in flight to form spherical drops about 70 ~ m in diameter (Fig. 1). To determine drop volume, 120 000 drops of isopropanol (at 2000 drops per s) were collected in pre-weighed microfuge tubes, and the accumulated fluid was weighed. For 3 separate jets tested, 11.4, 12.3 and 12.5 mg of fluid were emitted, indicating that the mass of single drops was from 0.095 to 0.104 p,g. Time-lapse photography (Fig. 1) indicated that the velocity of the drops in flight was approximately 3 m / s . When the distance between the jet orifice and the target was in the range of 1-2 cm, successive individual drops landed on the target in virtually exact (5-10 /.tm) superposition, (Fig. 2). The several different aqueous solutions (e.g. TSQ in buffer, sodium glutamate, and dopamine in deionized water) used for neurobiological experiments were all found to be suitable for dispensing; with the appropriate electrical pulse height and waveform, uniform drops were consistently obtained at ejection rates from 0.01 to 1000
84 4
-3 iii Z taJ
o _/ LI.
",,,'!1,: ".
/,,,
/.-
|
1
I
~
2
I
4
I
]
I
I
1
8
l
1
NUMBER of DROPS
Fig. 2. Top left image is a fluorescence photomicrograph of a section of mouse hippocampal formation: The fluorescent s p o t (arrow) was made by jetting three successive drops of TSQ onto the tissue, then illuminating the tissue section with UV: light. The drawing lower left is a tracing made from the upper image with the granule neuron stratum (G), and the outline of the dentate gyrus (thick line) indicated. Locations of granule and hilar neurons are shown schematically for scale. The data graphed on the right are the fluorescence intensity readings (arbitrary units) obtained when 1-8 drops of TSQ were jetted onto either the zinc-rich hilus of the dentate gyrus (as shown in the figures on t h e left) or onto the stratum radiatum of hippocampal field CA1, a region with only moderate zinc. Scale bar = 200/xm.
Hz. The pulse parameters used for dispensing deionized water were also found to be suitable for dispensing dilute (1-50 mM) aqueous solutions. Gravimetric determinations of drop size for one such solution (10 mM dopamine) indicated that the individual drops dispensed contained 72 pl of solution. (Because solutions differ in surface tension and viscoscity, calibrations would normally be done using the same solution to be used in an experiment.)
Histologic staining When drops of the histochemical reagent, TSQ, were jetted onto cryostat sections of brain tissue, the TSQ reacted with tissue zinc and formed the
fluorescent zinc/TSQ chelate complex. Individual drops of TSQ produced circles of fluorescence about 170 /zm in diameter on the tissue, and successive drops (each applied after the preceding drop had dried) produced a progressively brighter fluorescence within the same 170 tzm circle (Fig. 2). In these "microdrop" histochemical tests, an individual drop of the staining solution contained approx. 4 x 10-15 mol TSQ (70 pl drop of 60 IzM TSQ solution), and the cylindrical volume of tissue wet by a single drop (ca. 0.5 nl; 170 ~zm diameter x22 /~m thick) was estimated to contain about 150 x 10 ~5 mol chelatable zinc (assuming 250 IzM of zinc in the tissue; see Frederick-
85 son et al., 1983). Titration experiments were therefore undertaken to determine whether fluorescence from the tissue would increase monotonically with successive drops of reagent until a 1:1 molar ratio of T S Q / z i n c was achieved at approx. 37 (150/4) drops of TSQ solution. Figure 2 shows that fluorescent yield from the tissue did increase with successive drops. Moreover, the tissue zone with a high concentration of zinc that was studied (dentate hilus; see Frederickson et al., 1983, 1987) yielded much greater fluorescence than a comparison region (stratum radiatum of hippocampal field CA1) that has only moderate amounts of zinc. Attempts to carry these titration experiments to stoichiometric completion (i.e., to 1:1 molar ratios of zinc/TSQ), however, were not successful. Regardless of the total number of drops applied, peak fluorescence from the spot was never as high as that observed when the same region of tissue was saturated in TSQ by immersing the section in a stoichiometric excess of TSQ in a staining dish.
Focal application of drugs to brain slices during single neuron recording The piezojets proved entirely satisfactory for applying drugs to brain tissue slices during extracellular recording of neurons. The electrical pulse driving the piezoelectric crystal generated a transient artifact (ca. 50 msec) in the single unit trace, but because that transient gave a suitable marker for the time of drop ejection, no effort was made to remove the signal by shielding the leads a n d / o r housing of the piezojet. Aside from the electrical artifact, there was no other apparent effect of drop ejection per se upon the slice preparation or recording. Thus, when single drops of deionized water were jetted onto the tissue at the location of the recording pipette tip, no change in the action potential waveform or in the firing pattern of the recorded neurons could be detected. Apparently neither the dilution of extracellular fluids by the 70 pl volume nor the kinetic energy (about 6 × 10 -l° J / d r o p ) of the drop landing on the tissue was sufficient to perturb the recording pipette placement or the slice preparation. Stable single unit
recordings were routinely obtained from over 30 different neurons during drug application, even (in a few cases) while applying up to 1000 drops of fluid directly on the recording site at a rate of 1000 drops/s. When neuroactive compounds were jetted onto the slices, neuronal responses varied systematically according to the compound applied, the number of drops (drug dose), and the distance between the recorded neuron and the site of drug application. Glutamate, for example, either caused no change in firing rate or caused a brisk acceleration of firing in almost all neurons tested. In general, cells that were encountered more than 200 Ixm below the surface of the slice showed little or no response. For cells near the surface of the slice, the response to glutamate depended upon the number of drops applied and the proximity of the drop placement to the cell. Thus, drops placed directly at the point where the recording pipette penetrated the surface of the slice generally produced maximal acceleration whereas drops placed as little as 250 ~tm away from that point (Fig. 3b) typically had no detectable effect at all. One interesting exception to the latter result, which was observed in three different neurons, was a distinct slowing of the recorded neuron observed when glutamate drops were applied 200-400 # m off target. Presumably, interneurons inhibitory to the recorded neurons were excited by the glutamate in these cases. Application of dopamine to dopamine-containing neurons by conventional superfusion or iontophoresis methods produces relatively rapid, marked depression of firing rates (Bernardini et al., 1986; Bunney et al., 1983). In the present experiments, application of dopamine drops gave the same result. Two to 5 drops of dopamine (10 mM; 70 pl) applied directly to the recording site consistently depressed firing rates for 10-30 s (Fig. 4). Drops applied 200 to 400 g m off target generally had no detectable effect on the activity of the recorded cell. In one case, however, when drops were aimed at the distal dendrites of the dopaminergic neurons (which ramify in parts reticulata), the neuron being recorded (about 300 ~zm distant, in pars compacta) showed clear and long-lasting depression of firing.
86
,o
q I
//
m
JP'
,,, 5 0
~
,
I-.
