436
Anal. Chem. 1989, 6 1 , 436-441
Quantitative Analysis of Individual Neurons by Open Tubular Liquid Chromatography with Voltammetric Detection Robert T. Kennedy and James W. Jorgenson' Department of Chemistry, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599-3290
The abllity to analyze lndlvlduai cells is often important In biology because of the heterogeneity of tissue; this Is especlally true In the area of neurobldogy. A method is described for the determlnatlon of trace levels of organlc compounds in individual cells by open tubular llquld chromatography wlth voitammetrlc detection. I n the method, a cell Is Isolated, an Internal standard is added, the cell is homogenized and centrlfuged, and the supernatant is injected dlrectly onto the chromatography cdumn. Slnce data are collected In both the electrochemical and chromatographic domains, the resolution of the method Is better than that obtalned by using amperometric detection. The comblnatlon of voltammetry and chromatography also alds In the MentHkation of compounds. By use of thls method three dmerent neurons, D2, E4, and F1, from the land snall Helix aspersa are analyzed. The data show that the cells give certain unique and repeatable chemlcal profiles. Dopamlne, serotonin, tyrosine, and tryptophan were Mentifled and quantlfled In two of the celb at the femtomole level. I n the third cell, only the two amlno aclds were observed and measured. The quantltatlve data indlcate that the method Is at least as reliable as other methods that have been applied to slngie cells and considerably more sensltive. The comblnatlon of qualltatlve and quantltatlve information allows for the chemlcal mapping of cells.
INTRODUCTION In a number of areas of biology it is useful to know the chemical contents of individual cells and how cells differ chemically. Such knowledge is valuable in determining a cell's function. This is especially true in the area of neurobiology where cell heterogeneity is well-known. An area of active research concerns the role of individual neurons in memory and learning and how biochemical plasticity may be involved in these processes ( I ) . Another, older question concerns the use of neurotransmitters and individual neurons. Dale's hypothesis, as expounded by Eccles, states that individual neurons contain and release only one neurotransmitter at all of its branches and terminals. This concept, as stated, appears to be too limited in light of modern evidence but is still a subject of interest (2). It is apparent that in both of these areas it would be valuable to be able to analyze cells individually. Because of the interest in the chemistry of individual neurons, a number of analytical methods have been developed for them. These methods include, micro thin-layer chromatography (TLC) (3), gas chromatography/mass spectrometry (GC/MS) ( 4 ) , high-performance liquid chromatography (HPLC) with amperometric detection (5), and enzymatic radiolabeling (6). All of these methods provide good information; however they have important limitations. Micro TLC is not sensitive enough to actually analyze individual cells and therefore requires the pooling of four to six giant cells for
* Author
to whom correspondence should be addressed.
analysis. Also the method requires an extensive sample preparation, including the derivatizing of compounds for detection. GC/MS of individual cells is limited, because of a lack of sensitivity, to using the mass spectrometer in the selected ion monitoring mode; therefore a limited number of compounds can be determined at one time. This limitation is important because it means that the compounds to be determined must be chosen before the experiment. Also this method often requires sample derivatization in order to improve volatility of analytes. HPLC with amperometric detection is also limited by sensitivity and can only be used for cells that contain high levels of analyte. Enzymatic radiolabeling is sensitive; however it requires that the analyte be selected before the experiment, and the specificity of the method is limited by cross-reactivities. We have recently begun to develop methods for the analysis of individual cells by using open tubular liquid chromatography (OTLC) with voltammetric detection (7). The separation is performed inside of a capillary tube which typically has an internal diameter of 15-20 pm. The stationary phase is bound to the inner wall of the capillary rather than to particles packed inside the column as in conventional liquid chromatography. With voltammetric detection, the potential on the detector electrode is scanned so that voltammograms can be obtained on compounds as they elute from the column. A number of properties of this method appear to make it a good choice for the analysis of single cells. First of all, OTLC is a high-resolution separation method. Columns with inner diameters of 15 pm, such as those used in this report, equal or exceed the resolving power of modern HPLC columns. Another important aspect of the columns is that because of their small dimensions, they require small samples. A 15 p m i.d. column that is 2 m long has a total volume of 390 nL and a typical injection volume of 5 nL. Finally, the detectors that have been developed for OTLC have good mass sensitivity. The voltammetric detector used in these experiments has a limit of detection of 0.1 fmol for hydroquinone. This same detector, used in an amperometricmode, has a detection limit of 1amol. In both cases the detection limit is well below the detection limit of any other method that has been applied to individual cells. Because of this combination of properties, OTLC allows for the simultaneous determination of a variety of compounds at trace levels in samples of nanoliter volumes. The original method which we developed using OTLC had some drawbacks with respect to quantification and therefore was not as reliable as desired (7). First of all, there was no way to account for variables associated with the sample preparation and injection. In the method, the cell was homogenized and centrifuged and the supernatant removed and injected onto the chromatography column. Unfortunately it was not possible to remove all of the supernatant or to know how much of the supernatant was removed. Analyte could also be lost due to adsorption to glassware that the sample came in contact with. These problems resulted in artificially low values for compounds and an extra source of variability. Quantification was also limited because reponse of the detector electrode could change from run to run because of adsorption to the electrode surface and other forms of electrode fouling.
0003-2700/89/0361-0436$01.50/0Q 1989 American Chemical Society
ANALYTICAL CHEMISTRY, VOC. 61.
NO. 5.
MARCH 1. 1989
437
In this report we deacrite a new method, also based on open tubular liquid chromatography with voltammetric detection, for analyzing individual cells which takes into account these problems and apply it to the analysis of three, identified neurons from the land snail Helix aspersa. Variables in the sample preparation are monitored hy using an internal standard. The quantification by the detector is improved by using an electrochemical cleaning procedure and by using daily calibrations. The data obtained indicate that cells have unique and repeatable chromatograms that allow for the chemical mapping of cells. The data also show that the putative neurotransmitters dopamine and Frhydroxyhyptamine (serotonin) coexist in two of the neurons. EXPERIMENTAL SECTION Sample Preparation. The cells used in this study were three giant neurons (diameters 120-140 pm) from the land snail Helix uspersu. Specifically,the neurons were those labeled F1 (right parietal ganglion), E4 (visceral ganglion), and D2 (left parietal ganglion), accordii to the map of Kerkut et al. (8).These neurons were used hecaw their positions were well estahlished in the maps and so could he readily identifed from preparation to preparation. To perform the analysis, the Helix brain was dissected out as previously described and mounted on a microscope slide under Ringer solution (9). The cells were isolated by using micrcdissection techniques (10). Only the cell body, not the axon, was used for the analysis. Once isolated, the cells were rinsed in Ringer solution and then transferred, with approximately 5-10 nL of Ringer, to a 5oo-nL microvial by using a pipet which had an inner diameter of approximately 120 pm. Next 1-2 nL of intemal standard solution, which was 0.1 mM 3,4-dhydmxyhenzylamine (DHBA) dissolved in the mobile p h with 2 mM ascorhic acid, was added with a pneumatic microsyringe. The ascorbic acid was added to prevent the oxidation of analytes. The syringe used in this step is similar in design to other syringes that were developed to perform intracellular injections ( 1 1 , l Z ) . The syringe waa calibrated hy injectingaqueous solutions into mineral oil and determining the volume dispensed from the size of the resultant spherical droplet. The size of the droplet was determined by measuring its diameter by use of a microscope fitted with a reticle. The details of the syringe and its use as a dispenser of internal standard have heen described previously (13). Once the intemal standard was added, the cell was homogenizes inside the microvial with a miniature glass rod. The resultant solution was then centrifuged at 3000g for 10 min and the supematant removed and injected directly into the column by using the microinjector which has heen described before (7).All of the sample preparation and injections were done with the aid of a Wolfe Selectra I1 stereomicroscope, a Narishiga MM-33 micromanipulator, and an Oriel micropositioner. Chromatography. The chromatography columns were borosilicate glass capillaries with inner diameters of 15-19 Mm and lengths of 220-250 em and were fabricated in the lab (14). Dimetylwtadecylsilanewas chemically bound to the inner wall of the capillary to serve as the stationary phase. The mobile phase was 0.1 M citrate buffer adjusted to pH 3.1 with sodium hydroxide, The mobile phase also contained 0.21 mM dimethyloctylamine and 0.624.95 mM sodium octyl sulfate. The octyl sulfate was varied as needed to keep the chromatographic retention of the biogenic amines constant. This was necessary due to changes in capacity factors for the analytes as the columns aged. Detection. The detector used in these experiments was a carbon fiber microelectrode which has heen described before (15-18). The fiber had a diameter of 9 pm and a length of 1mm and was inserted into the outlet end of the column by using mimpdtioners. The potential on the electrode was ramped f" 0.0 to +1.3 V versus Ag/AgCl at 1.0 V/s. After a ramp was completed, the potential was stepped hack to 0.0 V and held for 1.3 s before initiating the next voltage ramp. Current measure. menta were made only during the forward scan and the data were collected at 100 points per second. The effect of charging current wm reduced by using a background subtraction technique. The applied voltage, data acquisition,and data manipulation were all done with a microcomputer.
. 240
-
o:4
TIME (MINUTES)
mure 1. Chromatovoltammogram of cell F1 obtained by using the memod described in ihe text, the column had a 19 pm 1.d. The peak labels are as follows: DA is dopamine. DHBA is 3.4dlhydroxy-
benzylamine, 5-HT Is serotonin, TYR is for tyrosine. and TRP is for tryptophan. The voltammgram fw ssroionh has two peaks due to the two functionalitieswhich are oxidizable at the electrode. namely the phenol group and the indole nitrogen. Chemicals. All standards, reagents, and mobile phase constituents were used as received from Sigma Chemical Co. except for the dimethyloctylamine,which was from Aldrich Chemical co. RESULTS AND DISCUSSION An example of the data that is obtained by using this method is shown in surface plot form in Figure 1. In the plot, each line parallel to the time axis corresponds to a cbromatogram taken a t the voltage indicated on the potential axis. Each line parallel to the potential axis represents a voltammogram taken at the time indicated. All of the voltammograms were subtracted from a scan obtained prior to the elution of any compounds. The chromatovoltammogram is shown to begin just after the elution of the first compound because there was a large unretained peak due to the added ascorhic acid which interferes with the other peaks. When the plot is studied, it is important to realize that it is from an individual cell and that each of larger peaks represents hundreds of femtomoles and the smaller peaks represent just a few femtomoles injected onto the column. Another important aspect of these data is the added resolution that is obtained as a result of combining voltammetry with chromatography. A number of peaks that would not have been resolved chromatographically were resolved in the electrochemical domain. The voltammetry also adds qualitative information to the data which improves the reliability of the identification of compounds. Unknown compounds can be identified by matching their chromatographicretention times and voltammelzicpeaks with standards. With this method, four compounds, tyrosine, tryptophan, dopamine, and serotonin, were identified in the cell F1 as shown. The same four compounds were found in the cell E4,but only the two amino acids were present in D2. The fact that serotonin and dopamine were not detected in cell D2 indicates that they are present a t less than 0.14 and 0.15 fmol respectively, based on the detection limit for those compounds. Other compounds that were tested for, but not observed, include 3,4-dihydroxyphenylalanine(DOPA), 3,4dihydroxyphenylacetic acid (DOPAC), 5-hydroxytryptophan (5-HTP), and 5-hydroxyindolacetic acid (HIAA). The detection limits for these compounds set their upper limits in the cells as follows: DOPA,
