Molecular and functional properties of the gamma-aminobutyric acid type A receptor

?-gamma-Aminobutyric acid (GABA) is the major inhibitory neurotransmitter in the mammalian central nervous system. The GABA-A receptor is a ligand gated chloride channel and is one of the major targets for GABA released in the synapse. The receptor is thought to be formed as a pentameric combination of different GABA-A receptor subunits. Currently 15 possible subunits have been cloned and sequenced, as well as a significant number of splice variants of those subunits. The subunit, or molecular, composition of the GABA-A receptor determines its functional properties. Although some of this relationship between molecular composition and functional properties has been studied, the physiological ramifications of seemingly limitless possible combinations of different subunits has much left to be defined. In this study, the molecular and functional properties of GABA-A receptors are studied using single cell reverse transcription polymerase chain reaction (RT-PCR) to determine the subunit expression in individual cells in combination with patch clamp electrophysiological recordings to define the functional receptor characteristics. The design, optimization, and examination of a single cell RT-PCR protocol for its effectiveness in detecting the expression of GABA-A subunit mRNA, and its use in conjunction with patch clamp recording are discussed herein. Briefly, it was found that a nested PCR approach using primers capable of amplifying families of subunits in the first round of PCR, and subunit specific primers in the second round, was necessary to achieve sufficient detection sensitivity. Multiple primer pairs are used in the first round of the nested PCR as a multiplex PCR reaction to amplify all of the GABA-A receptor subunits of interest. The type I and type II error rates for this single cell RT-PCR protocol are experimentally determined and methods of improving the statistical power are discussed. Single cell RT-PCR was also performed in conjunction with ultrafast ligand exchange patch clamp recordings in an effort to correlate molecular (subunit expression) and functional GABA-A receptor properties in cultured fetal mouse cortical neurons. Neurons that had similar functional responses also had similar RT-PCR results. From the data obtained, it appears that subunit mRNA expression and receptor function are related in neurons in primary culture.

MOLECULAR AND FUNCTIONAL PROPERTIES OF THE GAMMA-AMINOBUTYRIC ACID TYPE A RECEPTOR by Sterling Noble Sudweeks A dissertation submitted to the faculty of The University of Utah in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Pharmacology and Toxicology The University of Utah December 1997 Copyright © Sterling Noble Sudweeks 1997 All Rights Reserved THE UNIVERSITY OF UTAH GRADUATE SCHOOL SUPERVISORY COMMITTEE APPROVAL of a dissertation submitted by Sterling Noble Sudweeks This dissertation has been read by each member of the following supervisory committee and by majority vote has been found to be satisfactory. Louis R. BarrOllS Glen R. Hanson Scott W. Rogers H. Steve White THE UNIVERSITY OF UTAH GRADUATE SCHOOL FIN AL READING APPROV AL To the Graduate Council of the University of Utah: I have read the dissertation of Sterling Noble Sudweeks in its fmal form and have found that (1) its format, citations and bibliographic style are consistent and acceptable; (2) its illustrative materials including figures, tables, and charts are in place; and (3) the final manuscript is satisfactory to the supervisory committee and is ready for submission to The Graduate ~chool. :J/I!7~ f§~;i~ Date oJ~~~mmittee Approved for the Major Department ChairJDean Approved for the Graduate Council Dean of The Graduate School ABSTRACT y-Aminobutyric acid (GABA) is the major inhibitory neurotransmitter in the mammalian central nervous system. The GAB A-A receptor is a ligand gated chloride channel and is one of the major targets for GABA released in the synapse. The receptor is thought to be formed as a pentameric combination of different GABA-A receptor subunits. Currently 15 possible subunits have been cloned and sequenced, as well as a significant number of splice variants of those subunits. The subunit, or molecular, composition of the GABA-A receptor determines its functional properties. Although some of this relationship between molecular composition and functional properties has been studied, the physiological ramifications of seemingly limitless possible combinations of different subunits has much left to be defined. In this study, the molecular and functional properties of GABA-A receptors are studied using single cell reverse transcription polymerase chain reaction (RT -PCR) to determine the subunit expression in individual cells in combination with patch clamp electrophysiological recordings to define the functional receptor characteristics. The design, optimization, and examination of a single cell RT-PCR protocol for its effectiveness in detecting the expression of GABA-A subunit mRNA, and its use in conjunction with patch clamp recording are discussed herein. Briefly, it was found that a nested PCR approach using primers capable of amplifying families of subunits in the first round of PCR, and subunit specific primers in the second round, was neccessary to achieve sufficient detection sensitivity. Multiple primer pairs are used in the first round of the nested PCR as a multiplex PCR reaction to amplify all of the GABA-A receptor subunits of interest. The type I and type II error rates for this single cell RT-PCR protocol are experimentally determined and methods of improving the statistical power are discussed. Single cell RT -PCR was also performed in conjunction with ultrafast ligand exchange patch clamp recordings in an effort to correlate molecular (subunit expression) and functional GAB A-A receptor properties in cultured fetal mouse cortical neurons. Neurons that had similar functional responses also had similar RT-PCR results. From the data obtained, it appears that subunit mRNA expression and receptor function are related in neurons in primary culture. v TABLE OF CONTENTS ABSTRACT. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. iv LIST OF FIGURES .................................................. viii ACKNOWLEDGMENTS .............................................. ix Chapter 1. INTRODUCTION TO THE GABA-A RECEPTOR .................... 1 GABA Receptors in the CNS ...................................... 1 References ..................................................... 9 2. DEVELOPMENT OF A SINGLE CELL REVERSE TRANSCRIPTION POLYMERASE CHAIN REACTION TECHNIQUE FOR THE ANALYSIS OF GABA-A RECEPTORS ........................ 14 Summary ..................................................... 14 Introduction ................................................... 15 Results and Discussion .......................................... 17 Experimental Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 References .................................................... 40 3. AN ANALYSIS OF THE SINGLE CELL RT -PCR TECHNIQUE FOR TYPE I AND TYPE II ERROR RATES USING A MIXED POPULATION OF DEFINED GAB A-A RECEPTORS ................................ 44 Summary ..................................................... 44 Introduction ................................................... 45 Results ....................................................... 47 Discussion .................................................... 52 Experimental Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 References .................................................... 61 4. ULTRAFAST LIGAND EXCHANGE RECORDINGS IN CONJUNCTION WITH MRNA ANALYSIS .................... ' ..... 64 Summary ..................................................... 64 futroduction ................................................... 64 Results ....................................................... 68 Discussion .................................................... 73 Experimental Procedures ......................................... 75 References .................................................... 80 5. CONCLUSIONS ............................................... 81 Functional and Molecular Properties of GABA-A Receptors .. . . . . . . . . . . . 81 References .................................................... 84 vii LIST OF FIGURES 1. A simplified diagram of the GAB A-A receptor ......................... 6 2. Success rate of ~ detection using a semi-nested pair of PCR primers ...... 26 3. Effects of Mg2+ concentration on PCR product formation . . . . . . . . . . . . . . . . 28 4. Effects of changing the annealing temperature on PCR product formation .. 29 5. The effects of using increasing amounts of Taq polymerase with identical substrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 6. The effects of using various solutions in the aspiration pipette . . . . . . . . . . . . 35 7. Comparison of a l ~iY2s and ~~1 mean peak currents with the application of 10 JlM GABA and 10 J.LM GABA + 1 J.LM diazepam ................. 48 8. Comparison of mean peak currents (in pA) of al~2'Y2S and ~~l receptors for cells that were RT -PCR negative and RT -PCR positive .............. 51 9. Effects of decreasing the concentration of substrate when near the detection limit for PCR ....................................... 53 10. Diagram of the theta tube apparatus ................................ 66 11. Normalized traces of ultrafast ligand exchange patch clamp recordings . . . . . 70 12. A kinetic model of the GABA-A receptor ............................ 71 13. Superimposed figures of the decay phases and the obtained curve fits using two exponentials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 14. RT-PCR results of neuron A ...................................... 74 15. RT-PCR results of neuron B ...................................... 75 16. RT-PCR results of neuron C ...................................... 76 17. Superimposed images of the decay phases from neurons B and C ......... 77 18. Superimposed images of the decay phases from neurons A and B ......... 77 ix ACKNOWLEDGMENTS I want to acknowledge the members of my thesis committee and thank them for their support and guidance: Dr. Roy Twyman, Dr. Lou Barrows, Dr. Glen Hanson, Dr. Scott Rogers, and Dr. Steve White. I also want to thank Claudia, Reeve, Devin, and Sarah for their patience, love, and support. CHAPTER 1 INTRODUCTION TO THE GABA-A RECEPTOR GABA Receptors in the CNS Histoty of GABA y-Aminobutyric acid (GABA) was first identified in the CNS in 1950, and reported simultaneously by two independent groups (Roberts and Frankel, 1950; Awapara et al., 1950). Over the next 20 years, several groups performed experiments providing the evidence that GABA was a neurotransmitter, not simply another amino acid. In 1956 the anticonvulsant action of GABA was shown by direct application to mammalian cortex (Hayashi and Nagai, 1956), giving the first indication of its inhibitory function. Bazemore and colleagues provided more evidence for GABA's inhibitory role when they reported in 1957 that GABA attenuated stretch-receptor discharges in crayfish and lobster muscle preparations (Bazemore et aI., 1957). In that same year, it was also published that the in vivo levels of GABA were decreased in the brains of animals following semicarbazide induced seizures (Killam and Bain, 1957; Killam 1957). The following year, 1958, Kuffler and Edwards were able to demonstrate that GABA mimicks the then unknown natural crayfish inhibitory transmitter, again by showing attenuation of the stretch-receptor discharges (Kuffler 2 and Edwards, 1958). It was shown further, in 1962, that GABA was actually being released into the extracellular space specifically by the inhibitory neurons in the crustacean muscle preparations, providing more evidence that GABA was the natural inhibitory transmitter (Kravitz et al., 1962). This same group later reported that GABA was concentrated in inhibitory neurons 100: 1 in comparison to excitatory neurons (Kravitz et al., 1965). Adding strongly to the hypothesized neurotransmitter role of GABA, in 1969 it was shown that GABA is concentrated near the synaptic cleft in nerve terminals in rat cerebral cortex (Neal and Iversen, 1969). The release of GABA was described in 1970, and was found to be in response to depolarizing stimuli through a Ca2+ dependent mechanism (Bradford, 1970). In 1971 a specific uptake mechanism was reported for GABA showing that it is actively cleared from the synaptic cleft (Orkand and Kravitz, 1971). The distribution of GABA was found to be localized to the CNS (in contrast to the peripheral distribution in arthropods), with concentration differences across brain structures (Iversen and Bloom, 1972). Even though all of these data supported the role of GABA as an inhibitory neurotransmitter, very little was known about actual GABA receptors until pharmacological agents were found that modulated and differentiated the GABA responses. GABA Receptor Characterization Pharmacolo~ical modulation of GABA receptors. The first pharmacological agent described to alter the GABA response was picrotoxin (Takeuchi and Takeuchi, 1969). Picrotoxin was found to block the inhibitory action of GABA on crustacean 3 muscle preparations. Another GABA antagonist, bicuculline, was described in 1970 (Curtis et al., 1970). It was not until 1980, however, that different GABA responses were differentiated by bicuculline as either sensitive or insensitive (Bowery et aI., 1980). The bicuculline sensitive response became known as the action of the GABA-A receptor. The bicuculline insensitive response, which can be mimicked by baclofen, is now known to be through a G-protein linked second messenger system in conjunction with what is called the GABA-B receptor (reviewed in Mott and Lewis, 1994). During the 1980s a number of other pharmacological agents were discovered that acted on the GABA-A receptor, including many compounds in the following classes: benzodiazepines, barbiturates, ~-carbolines, and some anesthetics, neurosteroids, and insecticidal compounds (reviewed in Sieghart, 1995; Olsen et al., 1991). Binding studies with many of these compounds showed that marked receptor heterogeneity existed in the CNS, making the results difficult to interpret, especially with the overly simplistic structural model of the GAB A-A receptor at the time (reviewed in Burt and Kamatchi, 1991). Biochemical and molecular characterization of the GABA-A receptor. The GABA-A receptor was purified by benzodiazepine-affinity chromatography from bovine brain in 1982 (Sigel et al., 1982; Sigel et al., 1983). The biochemical characterization described two subunits, a (53 KD) and ~ (57 KD), and provided photo affinity labeling data that GABA bound primarily to the ~ subunit while benzodiazepines bound primarily to the a subunit. The initial hypothesis was that the GABA-A receptor was simply a tetramer of two a and two ~ .subunits. This was later shown to be incorrect (see below). Partial amino acid sequence data from these purified subunits were used in 1987 to clone and sequence the 0.1 and ~1 GABA-A 4 receptor subunits (Schofield et al., 1987). The initial cloning of the 0.1 and ~l subunits was quickly followed by the cloning of multiple other subunits, making the old idea of a simple receptor system somewhat obsolete. These sequences have been grouped together by homology into the 0.1-6, ~1-4' 11-3' 0, E, and PI-2 families (reviewed in Burt and Kamatchi, 1991; see also Cutting et al., 1992 ; Davies, et al., 1997). Protein sequence identity within a subunit family is about 70-80%, and between families is about 30-40%. Characterization of the Pl-2 subunits has shown them to be expressed almost exclusively in the retina, and they are currently thought to form homomeric receptors which some have classified as GABA-C receptors (Johnston, 1994). In addition to the nUlTlber of GABA-A receptor subunits cloned, splice variants have also been reported for the ~2' ~3' ~4' and 12 subunits (Harvey et al., 1994; Kirkness and Fraser, 1993; Bateson et al., 1991; Whiting et aI., 1990). The myriad of subunit and splice variant sequences currently cloned provide an ample base for seemingly limitless receptor heterogeneity. GABA-A Rectaltor Structure and Function. The cloning and sequencing of GABA-A receptor subunits showed that the GABA-A receptor was structurally similar to other ligand-gated ion channels like the glycine, 5HT-3, and nicotinic acetylcholine receptors (Schofield et al., 1987; Betz, 1990). The hydropathy plots of subunits from all these receptors suggest they have a similar tertiary structure: a long extracellular domain, three transmembrane regions, an intracellular loop, and a fourth transmembrane domain leading to a short extracellular carboxy terminus (see Figure 1). Structural data from studies of the other ligand-gated ion channel receptors suggested that the GABA-A receptor is a pentamer instead of a tetramer as had previously been proposed (Langosch et aI., 1988; Anand et aI., 1991; Unwin, 1995). The secohd proposed transmembrane region (TM2) shows the highest level of sequence identity within receptor families and across receptor types. This suggested that it was involved in some integral formation of an ion channel, and various studies have shown that it plays a large role in the formation of the ion channel pore (Unwin, 1993; Ffrench-Constant, 1993; reviewed in Snlith and Olsen, 1995). The pharmacological and functional properties of the GAB A-A receptor are determined by the combination of subunits used to form the ion channel complex. Subunit composition has been shown to affect the apparent binding affinity of GABA (Mohler et al., 1990), the single channel conductance and kinetics (Porter et al., 1992; Angelotti and Macdonald, 1993), the phosphorylation sites (Moss, 1995), and the pharmacological modulation of the receptor (Ltiddens and Wisden, 1991; Pritchett et al., 1989, Puia et al., 1991). Transfected cells and oocytes expressing known subunit combinations have made it possible for us to draw our current conclusions about the relationship of subunit composition and function; however, with the large number of subunit combinations available (Burt and Kamatchi estimated over 150,000 permutations, not including splice variants; Burt and Kamatchi, 1991), it would be a herculean task indeed to examine all possible subunit combinations this way. Some 5 •• I I I , , , , , , , , I _ ..... .. #. '.*. ...- ........~ -~-4It 4It~ ~~ ~ ~ ..• •• •• • • •• • ••••• • • • • 1 2 3 4 ••••••••• • • • • • • • • -~ 6 Figure 1. A simplified diagram of the GABA-A receptor. On the left is represented the pentameric GAB A-A receptor complex inserted into the cell membrane. The enlargement on the right shows one subunit pulled out with four transmembrane domains numbered. Note that the tertiary structure of the subunit sequence is greatly simplified to facilitate clarity. 7 attempts have been made to identify possible in vivo subunit combinations using in situ hybridization and immunohistochemistry studies to determine the expression profiles of the various GAB A-A receptor subunits (Laurie et al., 1992; Benke et aI., 1991). These studies have revealed that the regulation of subunit expression is very complex, with each subunit having its own distribution that changes across brain structures with time. In some brain regions where limited subunit expression occurs (i.e., cerebellum) these types of studies have suggested possible subunit combinations (Khan et al., 1994). For the most part, however, the molecular make-up of native GABA-A receptors is still unknown, leaving the function somewhat undefined. Single cell reverse transcription polymerase chain reaction (RT -PCR) is a relatively new technique that makes possible the analysis of mRNA expression in individual cells (Eberwine et al., 1992). Although a powerful technique on its own, it becomes even more informative when combined with patch clamp recording. These two techniques used in conjunction allow for the study of the mRNA expression of an individual cell, as well as the analysis of the physical properties of the ion channels expressed by that same cell. This allows the relationship between mRNA and functional protein expression to be examined. These techniques have been applied for the study of other ligand-gated ion channel systems with good success (Jonas et aI., 1994; Lambolez et al., 1992). Applying this type of analysis to the study of GABA-A receptors should help decipher the subunit combinations expressed by actual neurons, either in culture or in brain slices. Chapter 2 of this dissertation describes the development of a single cell RT- PCR protocol for the analysis of the GABA-A receptor. Chapter 3 contains an evaluation of the protocol developed, using known subunit combinations expressed by transfected cells to determine the type I and type II error rates for the single cell RT­PCR technique. Chapter 4 describes the attempts to apply single cell RT -PCR with ultrafast ligand exchange patch clamp recordings to coordinate subunit combinations (molecular properties) with functional neuronal responses. The concluding remarks, with possible future directions of study are given in Chapter 5. 8 References Anand, R., Conroy, W.G., Schoepfer, R., Whiting, P., and Lindstrom, J. (1991). Neuronal nicotinic acetylcholine receptors expressed in Xenopus oocytes have a pentameric quaternary structure. J. BioI. Chern. 266, 11192-11198. 9 Angelotti, T.P., and Macdonald, R.L. (1993). Assembly of GABAA receptor subunits: alpha 1 beta 1 and alpha 1 beta 1 gamma 2S subunits produce unique ion channels with dissimilar single-channel properties. J. Neurosci. 13, 1429-1440. Awapara, J., Landua, A., Fuerst, R., and Seale, B. (1950). Free gamma-aminobutyric acid in brain. J. BioI. Chern. 187, 65-69. Bateson, A.N., Lasham, A., and Darlison, M.G. (1991). 1-aminobutyric acidA receptor heterogeneity is increased by alternative splicing of a novel p subunit gene transcript. J. Neurochem.56, 1437-1440. Bazemore, A.W., Elliot, K.A.C., and Florey, E. (1957). Isolation of factor 1. J. Neurochem. 1,334-339. Benke, D., Mertens, S., Trzeciak, A., Gillessen, D., and Mohler, H. (1991). GABAA receptors display association of 12 subunit with a l and P2I3 subunits. J. BioI. Chern. 266, 4478-4483. Betz, H. (1990). Ligand-gated ion channels in the brain: the amino acid receptor superfamily. Neuron. 5, 383-392. Bowery, N.G., Hill, D.R., Hudson, A.L., Doble, A., Middlemiss D.N., Shaw, J., and Turnbull, M. (1980). (-)Baclofen decreases neurotransmitter release in the mammalian CNS by an action at a novel GABA receptor. Nature. 283, 92-94. Bradford, H.F. (1970). Metabolic response of synaptosomes to electrical stimulation: release of amino acids. Brain Res. 19, 239-247. Burt, D.R., and Kamatchi, G.L. (1991). GABAA receptor subtypes: from pharmacology to molecular biology. FASEB J. 5, 2916-2923. Curtis, D.R., Duggan, A.W., Felix, D., and Johnston, G.A. (1970). GABA, bicuculline and central inhibition. Nature. 226, 1222-1224. Cutting, G.R., Curristin, S., Zoghbi, H., O'Hara, B., Seldin, M.F., and Uhl, G.R. (1992). Identification of a putative gamma-aminobutyric acid (GAB A) receptor subunit rho2 cDNA and colocalization of the genes encoding rho2 (GABRR2) and rho1(GABRR1) to human chromosome 6q14-q21 and mouse chromosome 4. Genomics. 12,801-806. Davies, P.A., Hanna, M.C., Hales, T.G., and Kirkness, E.F. (1997). Insensitivity to anaesthetic agents conferred by a class of GABA(A) receptor subunit.Nature 385, 820-823. Eberwine, J., Yeh, H., Miyashiro, K., Cao, Y., Nair, S., Finnell, R., Zettel, M., and Coleman, P. (1992). Analysis of gene expression in single live neurons. Proc. Natl. Acad. Sci. USA 89, 3010-3014. 10 Ffrench-Constant, R.H., Rocheleau, T.A., Steichen, J.C., and Chalmers, A.E. (1993). A point mutation in a Drosophila GABA receptor confers insecticide resistance. Nature. 363, 449-451. Harvey, R.J., Chinchetru, M.A., and Darlison, M.G. (1994). Alternative splicing of a 51-nucleotide exon that encodes a putative protein kinase C phosphorylation site generates two forms of the chicken 'Y-aminobutyric acidA receptor ~2 subunit. J. Neurochem. 62, 10-16. Hayashi, T., and Nagai, K. (1956). Action ofw-amino acids on the motor cortex of higher animals. Especially gamma-aminobutyric acid as the real inhibitory principle in the brain. Proc. XXth Int. Physio!. Congr. 410. Iversen, L.L., and Bloom, F.E. (1972). Studies of the uptake of 3 H-gaba and (3H) glycine in slices and homogenates of rat brain and spinal cord by electron microscopic autoradiography. Brain Res. 41, 131-143. Johnston, G.A. (1994). GABAC receptors. Prog. Brain Res. 100,61-65. Jonas, P., Racca, C., Sakmann, B., Seeburg, P.H., and Monyer, H. (1994). Differences in Ca2+ permeability of AMP A-type glutamate receptor channels in neocortical neurons caused by differential GluR-B subunit expression. Neuron 12, 1281-1289. Khan, Z.U., Gutierrez, A., and De BIas, A.L. (1994). The subunit composition of a GABAAlbenzodiazepine receptor from rat cerebellum. J. Neurochem. 63, 371-374. Killam., K.F., and Bain, J.A. (1957). Convulsant hydrazides 1: in vitro and in vivo inhibition of vitamin B6 enzymes by convulsant hydrazides. J. Pharmacol. Exp. Ther. 119,225-262. Killam, K.F. (1957). Convulsant hydrazides 2: comparison of electrical changes and enzyme inhibition induced by the administration of thiosemicarbazide. J. Pharmacol. Exp. Ther. 119,263-271. Kirkness, E.F., Kusiak, J.W., Fleming, J.T., Menninger, J., Gocayne, J.D., Ward, D.C., and Venter, J.C. (1991). Isolation, characterization, and localization of human genomic DNA encoding the beta 1 subunit of the GABAA receptor (GABRB 1). Genomics. 10, 985-995. Kravitz, E.A., Potter, D.D., and van Gelder, N.M. (1962). Gamma-aminobutyric acid distribution in the lobster nervous system: CNS, peripheral nerves and isolated motor and inhibitory axons. Biochem Biophys. Res. Comnlun. 7, 231-236. Kravitz, E.A., and Potter, D.D. (1965). A further study of the distribution of gamma­aminobutyric acid between excitatory and inhibitory axons of the lobster. J. Neurochem. 12, 323-328. Kuffler, S., and Edwards, C. (1958). Mechanism of gamma aminobutyric acid inhibition. J. Neurophysiol. 21,589-610. Lambolez, B., Audinat, E., Bochet, P., Crepel, F., and Rossier, J. (1992). AMPA receptor subunits expressed by single Purkinje cells. Neuron 9, 247-258. Langosch, D., Thomas, L., and Betz, H. (1988). Conserved quaternary structure of ligand-gated ion channels: the postsynaptic glycine receptor is a pentamer. Proc. Natl. Acad. Sci. USA. 85,7394-7398. . Laurie, D.J., Wisden, W., and Seeburg, P.H. (1992). The distribution of thirteen GABAA receptor subunit mRNAs in the rat brain. III. Embryonic and postnatal development. J. Neurosci. 12,4151-4172. 11 Liiddens, H., and Wisden, W. (1991). Function and pharmacology of multiple GABAA receptor subunits.Trends Pharmacol. Sci. 12,49-51. Mohler, H., Mahlerbe, A., Draguhn, A., Sigel, E., Sequier, J.M., Persohn, E., and Richards, lG. (1990). GABAA-receptor subunits: functional expression and gene localisation. In GABA and Benzodiazepine Receptor Subtypes, G. Biggio and E. Costa, eds. (New York: Raven Press), pp.23-33. Moss, S.J., Gorrie, G.H., Amato, A., and Smart, T.G. (1995). Modulation ofGABAA receptors by tyrosine phosphorylation. Nature 377, 344-348. . Mott, D.D., and Lewis, D.V. (1994). The pharmacology and function of central GABAB receptors. Int. Rev. Neurobiol. 36, 97-223. Neal, M.J., and Iversen, L.L. (1969). Subcellular distribution of endogenous and (3H) gamma-aminobutyric acid in rat cerebral cortex. J. Neurochem. 16, 1245-1252. Olsen, R.W., Bureau, M., Endo, S., Smith, G., Deng, L., Sapp, D., and Tobin, A.J. (1991). GABAA-benzodiazepine receptors: demonstration of pharmacological subtypes in the brain. Adv. Exp. Med. BioI. 287, 355-364. Orkand, P., and Kravitz, E.A. (1971). Localization of the sites of gamma-butyric acid (GABA) uptake in lobster nerve muscle preparations. 1. Cell BioI. 49, 75-89. 12 Porter, N.M., Angelotti, T.P., Twyman, R.E., and Macdonald, R.L. (1992). Kinetic properties of al~l y-aminobutyric acidA receptor channels expressed in chinese hamster ovary cells: regulation by pentobarbital and picrotoxin. Mol. Pharmacol. 42, 872-881. Pritchett, D.B., Sontheimer, R., Shivers, B.D., Ymer, S., Kettenmann, R., Schofield, P.R., and Seeburg, P.R. (1989). Importance of a novel GABAA receptor subunit for benzodiazepine pharmacology. Nature 338, 582-585. Puia, G., Vicini, S., Seeburg, P.R. and Costa, E. (1991). Influence of recombinant y­aminobutyric acid-A receptor subunit composition on the action of allosteric modulators ofy-aminobutyric acid-gated cr currents. Mol. Pharmacol. 39, 691-696. Roberts, E., and Frankel, S. (1950). Gamma-aminobutyric acid in brain: its formation from glutamic acid. J. BioI. Chern. 187,55-63. Schofield, P.R., Darlison, M.G., Fujita, N., Burt, D.R., Stephenson, F.A., Rodriguez, R., Rhee, L.M., Ramachandran, J., Reale, V., Glencorse, T.A., Seeburg, P.R., and Barnard, E.A. (1987). Sequence and functional expression of the GABA A receptor shows a ligand-gated receptor super-family. Nature. 328,221-227. Sigel, E., Mamalaki, C., and Barnard, E.A. (1982). Isolation of a GABA receptor from bovine brain using a benzodiazepine affinity column. FEBS Lett. 147,45-48. Sigel, E., Stephenson, F.A., Mamalaki, C., and Barnard, E.A. (1983). A gamma-aminobutyric acidlbenzodiazepine receptor complex of bovine cerebral cortex. J. BioI. Chern. 258, 6965-6971. Sieghart, W. (1995). Structure and pharmacology of gamma-aminobutyric acidA receptor subtypes. Pharmacol. Rev. 47, 181-234. Smith, G.B., and Olsen, R.W. (1995). Functional domains of GABAA receptors. Trends Pharmacol. Sci. 16, 162-168. Takeuchi, A., and Takeuchi, N. (1969). A study of the action of picrotoxin on the inhibitory neuromuscular junction of the crayfish. J. PhysioI. 205, 377-391. Unwin, N. (1993). Nicotinic acetylcholine receptor at 9 A resolution. J. Mol. BioI. 229, 1101-1124. Unwin, N. (1995). Acetylcholine receptor channel imaged in the open state. Nature. 373,37-43. 13 Whiting, P., McKernan, R.M., and Iversen, L.L. (1990). Another mechanism for creating diversity in y-aminobutyrate type A receptors: RNA splicing directs expression of two forms ofY2 phosphorylation site. Proc. Natl. Acad. Sci. U.S.A. 87,9966-9970. CHAPTER 2 DEVELOPMENT OF A SlNGLE CELL REVERSE TRANSCRIPTION POLYMERASE CHAlN REACTION TECHNIQUE FOR THE ANALYSIS OF GABA-A RECEPTORS Summary GABA-A receptor subunit composition determines the functional characteristics of the receptor. However, because of the myriad of possible GABA-A receptor subunits, actual receptor composition in neurons is largely unknown. The development of a single cell reverse transcription polymerase chain reaction CRT -PCR) protocol for the study of GABA-A receptor subunit expression is detailed herein. Parameters for the RT reaction are examined. A nested PCR approach using primers that amplify families of GABA-A receptor subunits in the first round, and specific primers for each subunit in the second round was developed. Different methods for cell aspiration are also discussed and compared. The RT -PCR technique described is suitable for use in conjunction with patch clamp recording methods. Analysis of individual neurons using patch clamp recordings and single cell RT -PCR allows for the study of the functional receptor characteristics in conjunction with mRNA expression. This type of analysis should help unravel how ion channel properties and subunit composition relate in 15 actual neurons. Introduction With the advent of molecular biology cloning techniques, there has been a realization of how complicated many neurotransmitter receptor systems can be. Where once it was believed that only single receptors types existed, cloning has shown that multiple receptors, each with slightly different properties, actually exist. The GAB A-A receptor is a good example of this; what was thought to be a relatively simple receptor system has proven to be one with seemingly limitless diversity. The GAB A-A receptor is a multimeric transmembrane glycoprotein complex that functions as an anion selective pore when activated by ligand binding. The channel is thought to be formed by a pentameric combination of receptor subunits. Currently 15 different receptor subunits have been cloned and sequenced and have been shown to participate in GABA-A receptor formation. These have been grouped together into families by sequence homology as (X. 1-6 , ~1-4' "'{1-3' 0, and E (Burt and Kamatchi, 1991; Macdonald and Olsen, 1994; Stephenson 1995; Dunn et al., 1994; Davies et al., 1997). In addition, there are splice variants reported for the ~2' ~3' ~4' and "'{2 subunits that increase even further the possible diversity (Harvey et al., 1994; Kirkness and Fraser, 1993; Bateson et al., 1991; Whiting et al., 1990). The combination of subunits forming the receptor determines the receptor's physical properties. Subunit composition has been shown to affect receptor phosphorylation, pharmacological modulation, apparent GABA affinity, conductance, and single channel kinetics (Moss et al.,1995; Pritchett et al., 1989; 16 Liiddens and Wisden, 1991; Puiaet al., 1991; Mohleret al., 1990; Porter et al., 1992; Angelotti and Macdonald, 1993). The great level of possible diversity in channel molecular make-up and functional properties makes it difficult to understand all of the roles the receptor plays in vivo. By defining the receptor and its different subpopulations, we can gain a clearer view of its functional roles and gain insight into designing more specific therapeutic pharmacological agents for treatment of anxiety and epilepsy, and for anesthesia as well. The techniques that have given us most of our current understanding about receptor subpopulation expression patterns are in situ hybridization (Wisden et al., 1992; Laurie et al., 1992) and immunohistochemistry (Benke et al., 1991; Endo and Olsen, 1993). These measure the expression of mRNA and protein respectively. Both of these methods have shown that subunit expression patterns can be complicated, with each subunit showing its own neuronal distribution pattern spatially and developmentally. Another technique that has been used extensively to study ion channels is patch clamp recording (Hamill et al. 1981; Twyman et al. 1989). This technique allows the experimenter to measure the current flowing through ion channels and study the channel's functional properties. However, since no direct correlation of subunit mRNA expression (i.e., in situ hybridization) and subunit protein expression (I.e., immunohistochemistry) has been performed for the multitude of GABA-A receptor subunits, it has been difficult to define how the molecular properties and the physiological function of these ion channels relate in actual neurons. A relatively new technique, single cell reverse transcription polymerase chain 17 reaction (single cell RT-PCR; Eberwine et al., 1992; Lambolez, 1992; Jonas et aI., 1994), used in conjunction with patch clamp recording has shown great promise in increasing the understanding of how subunit mRNA expression and the formation of functional ion channels interplay. This combination of techniques allows the study of both the functional characteristics of the receptor and the subunit mRNA expression of a single cell. This chapter explains the development of a single cell RT -PCR protocol for the study of GABA-A receptors. Results and Discussion Description of the Problem Single cell RT -PCR has been used by other groups to analyze the expression of neurotransmitter receptor subunit mRNA (Lambolez, 1992; Jonas et al., 1994). However, these studies have analyzed relatively few targets per cell, such that the number of PCR reactions from an individual cell is quite low. One question we had was whether or not it was possible to run a large number of PCR reactions from an individual cell to look for the expression of a larger number of the GABA-A subunits that have been cloned. The feasability of this was not known because the sensitivity of single cell RT -PCR was not known. Was it so sensitive that one had to worry about amplifying genomic sequences vs. cDNA sequences, or was it not sensitive enough to detect multiple subunits if the RT reaction is divided into multiple PCR reactions? If low sensitivity is an issue (which it does tum out to be), what are the ways the sensitivity can be increased enough to allow the examination of all the subunits desired, and how reliable are the results obtained? These are the questions we had in mind when developing and analyzing the single cell RT -PCR technique described below. The Reverse Transcription Reaction 18 The initial step in detecting specific tuRN A sequences using PCR is to make cDNA copies of them. When starting with limited amounts of RNA (Le. single cells), it is important to get as much usable target cDNA as possible in order to have enough substrate to use in subsequent PCR reactions. Thus, it is important to optimize the RT reaction. Fortunately, there are relatively few parameters to optimize in the RT reaction. The main considerations we had were the type of primers used to initiate the RT reaction, and the type and concentration of reverse transcriptase enzyme to use. Primin(: the RT reaction. Three different priming methods are commonly used for RT reactions: oligo-dT, specific antisense primers (i.e. downstream PCR primers), and random hexamers. We decided to use random hexamer DNA oligos to prime the RT reaction. This decision was based on reports that this method minimizes possible complications and gives the most consistent results when followed by PCR (Veres et al., 1987; Noonan and Roninson, 1988; Kawasaki, 1990; Chelley and Kahn, 1994). Possible complications with the other two methods would be areas of mRNA secondary structure between the poly-A tail and the region to be amplified later by PCR; and the addition of a number of downstream primers in order to get all of the GABA-A sequences could complicate later PCR by causing primer concentration imbalances. Regions of mRNA secondary structure have been shown to cause the termination of 19 cDNA extension, by making the reverse transcriptase falloff the mRNA template. hnbalances in the upstream and downstream primer concentrations can cause smearing and non-specific products in the PCR reaction. Random hexamers, on the other hand, initiate reverse transcription at multiple, relatively random places and thus avoid some of the problems with having to reverse transcribe through difficult regions of mRNA secondary structure. They are also small enough that the high temperatures used in PCR keep them from initiating the amplification of nonspecific products. RT enzymes. The most well used reverse transcriptases are mouse Moloney leukemia virus (M-ML V) RT and avian myoblastosis virus (AMV) RT. The activity of these enzymes is similar at 37°C (however, AMV RT has a longer half-life than M­MLV at higher temperatures). We compared two dual activity enzymes, Tth and Retrotherm™ RT, with M-MLV RT to see which worked better in the RT portion of RT-PCR. Tth and M-MLV RTwere compared using the buffers and standard conditions given by the suppliers, random hexamers for priming, and 1 ng samples of fetal mouse cortex, adult mouse cortex, and adult mouse cerebellar RNA as substrates. PCR amplification using Taq polymerase and primers for ~-actin and the GABA-A ~2 subunit were performed from each sample. The M-ML V RT samples showed more PCR product from each of the samples tested than the Tth RT samples. M-ML V RT and Retrotherm™ RTwere similarly compared, looking at GABA-A al'~' and a3 subunits with PCR, and again M-ML V RT gave superior product formation. It should be noted that the Tth and Retrotherm ™ enzymes use much higher incubation temperatures (65-70°C) during the RT reaction than M-MLV RT (37°C). The priming 20 activity of random hexamers at these elevated temperatures is probably not very high and may be a major reason why a decrease in RT activity was seen. With this in mind, we also looked at using antisense primers for priming the RT reaction with the Retrotherm™ enzyme and still, the M-MLV RT enzyme with random hexamers at 37°C showed superior activity. A comparison of the DNA polymerase PCR activity of the Tth enzyme with Taq polymerase was also performed and is detailed later on. RT enzyme concentration. A series of M-ML V RT concentrations were compared to see if enzyme concentration had any effect on subsequent PCR product formation. 10.0 J.11 RT reactions using 0.1 ng fetal mouse cortex RNA as the substrate and 5,10,15,20,25, and 30 U/J.1l ofM-MLV RT were compared using product formation in subsequent PCR for the GABA-A (11 subunit as the endpoint. Very little difference in product formation was seen; however the 100 U reaction (10 U/J.1I) showed slightly more product than the others. This is also the concentration of RT enzyme suggested in the protocol included by the supplier. Polymerase Chain Reaction There are different methods reported for analyzing single cell cDNA libraries using PCR. One is to simply run a different PCR reaction for each target sequence of interest (Arkett et al., 1994). This approach works fine when there are few sequnces to be examined. However, when there are many targets of interest, the RT mix has to be divided into small aliquots, meaning a lower concentration of substrate is available for each individual PCR reaction. We found that this can severely affect the rate of false 21 negatives in the peR (see Effects of substrate volume on peR, below). Another peR method that has been reported gets around this problem by increasing the amount of target sequences that can be analyzed from a single peR reaction. This is done by using primers that amplify more than one target, and then distinguishing the different target sequences using restriction enzyme digests of the peR product mix. This has worked well for glutamate receptors where the sequence identity is high between subunits, allowing for the construction of primers that amplify more than one subunit sequence (Lambolez, 1992; Jonas et al., 1994). We have tried this approach for the GABA-A receptor subunits and have not been able to get enough peR product to perform restriction digests for the analysis, perhaps indicating the level of expression is lower for GABA than glutamate receptors. To overcome this problem we developed a nested peR approach. An initial round of peR uses a mix of primers that amplify all of the GABA-A receptor subunits of interest to us. This is followed by a second round of peR for each individual sequence where the initial peR reaction serves as the substrate. The development of this protocol is covered below. peR primers. The first consideration for primer design is the region to amplify. The region between putative transmembrane regions ill and N of the various GAB A-A receptor subunits was of interest to us. This area has been shown to contain many of the reported splice variants; it also has the greatest amount of sequence diversity, making it easier to design specific primers that distinguish between otherwise closely related sequences. With primer location in mind, a second consideration was to make the upstream and downstream primers bind to separate exons. This makes it possible to 22 distinguish by size the PCR products from spliced mRNA and those of possible genomic origin. It is generally believed that the intronlexon boundaries are highly conserved among GABA-A subunits (Kirkness et aI., 1991). To analyze primer position in relation to intronlexon boundaries, all sequences of interest were aligned with the GABA-A human ~l (Kirkness et al., 1991) and mouse 0 (Sommer et al., 1990) sequences for which the intronlexon boundaries are known. Since GAB A-A receptor subunit mRNAs are relatively short (they all average about 1500 bp), and there are 9 exons, it was fairly simple to obtain primers on separate exons by requiring primers to bind a few hundred basepairs apart. The exons in the region we were analyzing for PCR primers averaged only 125 bp, with known intron sequence lengths (Sommer et al., 1990) ranging from 103 bp to 10 kb. All of the primer pairs used in the first round of nested PCR spanned at least one intronlexon boundary. However, while keeping primers on separate exons is a good idea in theory, in reality we have found it is highly unlikely to get amplification from genomic sources when amplifying from a single cell (see Chapter 3; Johansen et al., 1995; Karrer et al., 1995). It appears that the sensitivity of single cell RT -PCR is not great enough to get amplification from genomic sequences. A much harder problem to solve has been increasing the sensitivity of single cell RT -PCR to a level where reliable detection of the expressed receptor subunits occurs. Possible primers were examined with the PCRPLAN module of the PC/GENETM (ver. 6.70) software package from IntelliGenetics. This program checks primer pairs for a number of parameters; the ones we were most interested in were: 23 primer self-complementarity, stem-loop formation, complementarity between primers, complementarity to other regions of the template, and compatible calculated annealing temperatures (calculated annealing temperatures within soe of each other). Primers that have self-complementarity (will bind to another prinler molecule of the same sequence) can form stem-loops (primer folds over and binds to itself), or are complementary to the other primer sequence can remove a large fraction of the primer pool from being available for peR. This can severely diminish formation of the desired product. Primers that have significant complementarity to other regions of the same template (or other unknown templates) will cause multiple peR products to be formed. Primer pairs that do not have annealing temperatures close to one another can cause effects similar to those previously described for primer concentration imbalances. However, the actual annealing temperature used for optimal peR needs to be determined empirically since it can not always be calculated correctly solely from sequence composition. Nested peR. We developed a nested peR approach to analyze the various GABA-A subunits we were interested in identifying. We found the nested approach neccessary to increase the sensitivity after our initial attempts at simply dividing the RT reaction into separate peR reactions for each subunit showed too much variability. We designed pairs of primers that would amplify families of subunits (hereafter called general primers). Specific primer pairs (hereafter referred to as specific primers) that bind to sequences in the products of the general primer reactions were used to perform a second round of peR amplification to identify the expression of specific subunits. 24 The nested PCR approach has been reported to greatly increase both the sensitivity and specificity ofPCR (Jackson et al., 1991), and we have found that it greatly increases the reliability of detection of GABA-A subunit sequences from single cells. The design of the general primers was accomplished by aligning families of GAB A-A receptor subunit nucleotide sequences and analyzing regions containing a high degree of sequence identity for possible PCR primers. All general primer pairs were also required to have compatible Mg2+ concentrations and annealing temperatures (see below for description of optimization). This allowed us to run them all together as a single multiplex PCR reaction (Chamberlain and Chamberlain, 1994). This makes it unneccessary to divide the RT reaction into separate PCR reactions when the initial substrate concentration is critically low. The design of specific PCR primers was performed by aligning the sequences of the regions amplified by the general primers and looking for regions of low sequence identity. Candidate primer sequences were screened against the GenBank DNA sequence database using the BLASTN computer program (Altschul et al., 1990). Primers that showed possible binding to other known sequences were rejected. Compatible primer pairs were again chosen using the PCRPLAN computer program. Seminested PCR. Initially, after discovering that reliable detection of subunits using individual PCR reactions was not possible with the high number of reactions required, we tried using a seminested PCR approach. With this technique, PCR using a pair of general primers is followed by a second round of PCR using one nested specific primer and one of the original general primers. We used the specific upstream primers 25 for each subunit and the general downstream a and ~ primers for the al - 3 and ~1-3 subunits. The advantages of using a seminested approach are that it increases sensitivity and specificity similar to the fully nested approach, and it requires the design and construction of fewer primers than fully nested peR. However, we found that the detection of one subunit could be affected by the presence of higher concentrations of other sequences that can anneal with the general primer. Using dilutions of plasmids containing the human GAB A-A a1 and ~ sequences, we tested the reliability of performing peR from low concentrations (typical of single cells) of DNA in the presence of 10 times higher concentrations of a competing sequence. When 1.0 fg of the a2 containing plasmid is used alone as the substrate for peR using an a2 specific upstream primer and the a general downstream primer, 15 of 15 reactions showed the a2 product. When 0.5 fg of the plasmid is used as substrate, 11 of 15 reactions showed the ~ product. When 0.1 fg was used as substrate, no ~ product formation was seen in 15 reactions. However, when 1.0 fg of the ~ plasmid is combined with 10 fg of the a l plasmid, 13 of 15 reactions showed the ~ product, a slight decrease in the success rate of using ~ plasmid alone. When 0.5 fg of ~ plasmid is combined with 5.0 fg of the a l plasmid, a sharp decrease in the success rate is seen as only 2 of 15 reactions gave the ~ product (see Figure 2). Interestingly, when equal concentrations of al and ~ plasmids are mixed together at 1.0 fg or 0.5 fg each, 15 of 15 and 12 of 15 reactions respectively produced the desired ~ product, showing virtually no change in the success rate of detection. We hypothesized that the 100%~r~---n~~--------~----D7~TI------------~--------__ ~ 80%-+---n; Alpha 2 alone Alpha 2 + 1 X Alpha 1 Alpha 2 + 10 X Alpha 1 1.0 fg Alpha 2 0.5 fg Alpha 2 Figure 2. Success rate of Clz detection using a seminested pair of peR primers (n=15). 10 times higher concentrations of 0,1 drastically decreases the success rate of Clz detection at low substrate concentrations (0.5 fg). higher concentration of competing sequence was usurping the downstream primer, decreasing amplification of the desired sequence.. To see if this effect could be eliminated, we repeated the above experiments using a pair of primers where both 26 upstream and downstream primers are specific for the Clz sequence (i.e., typical of fully nested peR). Using a pair of specific primers negated the decrease in detection in the presence of the other a sequence. For this reason, we designed and used fully nested peR primers for each of the GABA-A receptor subunits. M~2+ and annealin~ temperature optimization. Once putative primers have been designed, it is important to determine their optimal peR conditions. The two major variables involved in peR primer optimization are Mg2+ concentration and annealing temperature. The optimal Mg2+ concentration for primer pairs was determined by running peR on samples containing identical substrate with Mg2+ concentrations 27 ranging from 1.5 mM to 4.5 mM in 0.5 mM steps. When these products were examined on agarose gels, there was usually a specific Mg2+ concentration that produced either greater, or less nonspecific, PCR product formation than the others (see Figure 3). Some primers, however, worked well in a range of Mg2+ concentrations. The optimal annealing temperature for each primer pair was determined by running PCR on samples using a range of annealing temperatures in 2 0 C steps. Again, PCR products at each annealing temperature were compared to see where optimum product formation occurred (see Figure 4). The concentration of primer pairs used in PCR was also examined in a similar manner and was found to have relatively little effect on product formation in the range we examined (0.5 ~ to 2.5 ~ each primer). Primer sensitivity. After determining the above parameters, we tested the sensitivity of the primer pairs using serial dilutions of substrate. For some primer pairs (~1-3)' we were able to use dilutions of plasmids containing the desired GAB A-A receptor sequence. Specific primer pairs were deemed suitable for use if consistent amplification was obtainable from 1.0 fg plasmid used as the substrate. For the other sequences, however, we either did not have the desired GABA-A subunit in a plasmid, or the plasmids we had contained the sequence from a species other than mouse (for which our primers were designed), and were different enough to preclude good primer binding (Le., >2 mismatches) and PCR amplification. To analyze the sensitivity of these primer pairs, we used cDNA from RT reactions of serially diluted fetal mouse cortex RNA as the substrate. These primer pairs were deemed suitable if consistent amplification was obtained from the fetal cortex RNA RT samples for which 100 pg Figure 3. Effects of Mg2+ concentration on PCR product fonnation. Mg2+ concentration (in mM) is listed on each lane. PCR is for the GABA-A (X3 subunit. The optimal Mg2+ concentration in this figure is 1.5 mM. RNA had been used as the substrate (for a 10 JlI reaction). Sometimes several primer pairs for a specific subunit had to be examined in order to find one with suitable sensitivity. In order to simplify our single cell PCR analysis, we decided to exclude from our attention subunits that had been reported to have low expression levels in the developing cortex ((X6' 0, 11 and 13; Laurie et al., 1992). We also excluded (X4 and (xs because we could not detect them using RT-PCR from RNA extracts of the two week old cultured fetal mouse neurons we study, even when using very high RNA concentrations as the substrate (i.e. 1 Jlg RNA). The E subunit has only recently been 28 reported and was also not included in our studies. As such, the subunits we focused our search on were: (X1-3' ~1-3' and 12- The primer sequences we finally used, along with Figure 4. Effects of changing the annealing temperature on PCR product formation. The temperature is marked for each lane. At lower annealing temperatures (47°C-53°C) multiple PCR products are formed. The desired product is the upper band (904 bp) for the GABA-A ~ general primers. Optimal annealing temperature is 57°C. 29 their empirically determined optimal Mg2+ concentrations and annealing temperatures, are shown in Table 1. Effect of substrate volume on PCR. PCR appears to be very susceptible to variability when extremely low levels of substrate are used (see Chapter 3; Karrer et al., 1995; Surmeier et aI., 1996). We found that under these circumstances, the volume added from the RT reaction to PCR had a great effect on the probability of product formation. When 0.5 J..1l from a 10 J.ll RT reaction (using 10 pg of fetal mouse cortex RNA, approximating the level of a single cell; Kawasaki, 1990) is used as the substrate for PCR for the GABA-A ~ subunit, only 1 out of 8 (12.5%) reactions was positive. When 1.0 J.ll was added from the RT reaction to PCR, 4 out of 8 (50%) reactions were positive. When 2.5 J..1l aliquots were used, 6 out of 8 (75%) reactions were positive. To see if the same effect occurred with neurons, we did an RT reaction from a single Table I Primers, annealing temperatures and magnesium concentrations for single cell RT-PCR of GAB A-A receptor subunits SUBUNIT(S) UPSTREAM PRIMER DOWNSTREAM PRIMER(S) TEMP. Mg++ SIZE Alpha CTCCTGATACATTCTTCCACAATOG CAGACAGCTATGAACCAGTCCATOGC 57 3.75 mM 592 General CAGACGGCTATGAACCAGTCCAT* Alpha I GCTCCTGCGTATCACAGAOG GGAAGTGAGTCGTCATAACCAC 60 2.5mM 302 Alpha 2 GCTCCCGATOGCTCCAOG ACCTGTACTGGATTTAATTGTTTCC 63 2.0 to 3.0 mM 145 Alpha 3 CAACAAGCTGCTCAGACTGG ACCAAAGACTGTGCGGGCA 59 1.5 mM 433 Bcta TCAGGATCACAACCACAGC CGGGACCACCTGTCTATGG 57 3.75 mM 904 General CGGGACCACCGATCAATGG* Bcta I AGGAGCGAGCAAACAAGACC CTGCTCAGTGGCTTGCGGTA 51 1.5 mM 214 Beta 2 CAACAACGAGAAGATGCGCC CAGTTGAGAGGCACGTCTCCTC 61 3.5mM 265 Beta 3 CTAAGAGTGAAATAAACCOGGTGG AAGACCTCCTCCGTAGGTGGG 59 3.0mM 237 Gamma TCTGTGTGCTTCATCTTTGTGTTT ACCCAGTAAACAAGATTGAACAAG 57 3.75 mM 368 General Gamma 2 TTTGTCAGCAACCGGAAGC TGTCTCCAGGCTCCTGTTCG 61 1.5 to 3.5 mM 218/242 Beta-actin AAGATCCTGACCGAGCGTGG CAGCACTGTGTTGGCAT AGAGG** 50 3.0 to 4.5 mM 327 * The multiple alpha and beta general downstream primers are mixed at a I: 1:2 concentration with the upstream primers ** The same beta-actin primers are used in both rounds ofPCR (i.e. specific primers are not nested) w 0 31 cultured fetal mouse cortical neuron and used aliquots of two sizes as the substrate for repeated PCR reactions. When 0.5 fll of the 10 fll RT reaction was used as the substrate for PCR looking for ~-actin expression, only 4 out of 6 (66%) of the reactions showed product. However, when 1.0 fll aliquots were added to the PCR, 6 out of 6 (100%) of the reactions were positive. The variability of the PCR was seen to be related to the initial volume of the RT reaction used as substrate. In addition, since the level of GABA-A receptor subunit mRNA is most likely lower than that of ~-actin, it would be neccessary to use greater than 1 fll aliquots for the analysis of individual GAB A-A receptor subunits. If we increased the aliquot size substantially (assuming individual PCR reactions for each subunit) we would have to reduce the number of sequences in our analysis. For this reason we designed general primers for the nested PCR. Furthermore, we constructed general primers that had similar optimal PCR conditions. Running all of the general primers together in a single multiplex PCR reaction allowed us to minimize the variability of the PCR results by maximizing the amount of substrate added. PCR enzyme comparisons. As stated above, enzymes other than Taq polymerase have been used for RT-PCR. We compared Tth and Taq by performing PCR on aliquots from RT reactions (using M-MLV) from single neurons and from 1 ng fetal mouse cortex RNA. Primers for ~-actin and the GABA-A ~1 subunit were used for the single cell PCR, and primers for the GABA-A ~2 subunit were used for PCR from the RNA dilution. In all cases, more PCR product was seen from the Taq reactions. 32 Hot start PCR. An alteration in the standard PCR protocol that has been reported to increase product specificity is the "hot start" procedure (Chou et al., 1992). This protocol calls for the addition of a critical component of the reaction mix only after all other components have been added, and the reaction mix is brought to an elevated temperature (i.e., 70°C). The rational is that raising the temperature of the reaction mix before primer extension can begin keeps spurious products from forming in the PCR. In our hands, this approach was found to be very tedious when running multiple samples and gave no noticeable improvement in product formation; as such we decided not to use the hot start protocol. Enzyme concentration. The effects of various concentrations of Taq polymerase on PCR product formation were also examined. We looked at the effects of using 0.025,0.05,0.075 and 0.1 U/J,l1 Taq. The concentration recommended by the supplier is 0.025 U/J,ll. PCR amplification of ~-actin from 1.0 ng adult mouse cortex RNA with increasing concentrations ofTaq is shown in Figure 5. Increasing the Taq concentration gave more product, however, higher enzyme concentrations also caused an increase in nonspecific PCR products. Experience with the nested primer RT-PCR approach has led us to use a Taq concentration of 0.05 U/J,l1 for the initial round of multiplex PCR with the general primers. This maximizes the desired product formation without giving unwanted non-specific products. The recommended concentration of 0.025 U/J,l1 (which is more cost-effective, and no increase in sensitivity was seen using the higher enzyme concentration here) is used in the following specific PCR reactions. Cell Aspiration Figure 5. The effects of using increasing amounts of Taq polymerase with identical substrates. Each lane shows the concentration of Taq in U/J1l. Lanes were cut out and placed in order of increasing enzyme concentration for clarity. 33 There are two methods reported for obtaining individual cells for single cell RT-PCR analysis in conjunction with patch clamp recording (Eberwine et al., 1992; Grigorenko and Yeh, 1994). The first is to aspirate the cell into the same electrode used for whole cell recordings. After the recording is performed, suction is applied to the internal side of the electrode and the cytoplasm is aspirated directly, leaving behind an empty cell merrlbrane. The relatively small size of the opening of these electrodes requires a significant amount of time, on the order of ten minutes or more, for the aspiration to take place. A second approach is to perform the recording with one electrode, and then use a second much larger electrode to aspirate the entire cell, outer membrane and all. We decided to use the second approach, since it is relatively easy to aspirate single cells from the dissociated cultures we examine. This keeps the time 34 between initially perturbing the cell with the recording electrode and analyzing the mRNA contents of the cell to a minimum. Since mRNA is highly susceptible to degradation, time is of major importance. U sing a second electrode for aspiration also has some other advantages. It allows greater variability in the type of recording done from the cell, i.e., not just whole cell recordings can be performed but patch recordings can also be done. It also allows the use of other solutions in the aspiration electrode than the recording solution. We examined the effects various solutions had on the subsequent RT-PCR results. External recording solution, sterile pyrogen-free water, and 0.9% NaCl in water were compared. All solutions were tested and found negative for RNase activity using the RNase/Alert RNase detection kit (Ambion). One ng/J11 dilutions of fetal mouse cortex RNA were made in all the above solutions from 1 J1g/J11 RNA stocks (in water). 3 J1l of these dilutions were used as substrate in 10 J11 RT reactions. Nested PCR for the GAB A-A receptor ex3 subunit (using the ex3 specific upstream primer and the ex general downstream primers) was then performed from the RT samples. PCR product was only observed from the dilutions in water, indicating that the other solutions had inhibitory effects on the RT reaction (see Figure 6). The negative effects presumably occurred by altering the final ion concentrations and/or the pH in the RT reaction mix. One minor complication with using water as the aspiration solution is that the aspiration has to be performed quickly. If the aspiration electrode is left near the target cell too long, the cell will explode due to osmotic pressure. However, it is not desirable to have the aspiration electrode in the bath for very long to minimize the risk of Figure 6. The effects of using various solutions in the aspiration pipette are modeled using RNA diluted in (A) sterile water, (B) 0.9% NaCI, and (C) external recording solution as the substrate for RT­PCR. PCR products show that water is the superior aspiration solution. contamination, so a quick aspiration is a good idea anyway. The single cell RT-PCR protocol described above allows for the detection of expression of the GABA-A receptor subunits (X,1-3, ~1-3' "12, and ~-actin. When performed in conjunction with patch clamp recordings it provides a method for 35 obtaining detailed information about receptor functional properties and possible subunit composition. It is important to remember, however, that mRNA and protein expression do not always have a one to one correlation. Regulation of protein expression can occur by regulation of mRNA transcription, but there are also posttranscriptional regulatory mechanisms as well. In the GABA-A receptor system it has been reported 36 that protein levels and mRNA levels generally appear to be well correlated (Dunn et al., 1994; Stephenson 1995), suggesting that the expression of receptor subunits may be regulated at the level of transcription. However, only by performing studies where both the functional receptor properties and the mRNA expression patterns are examined together can we determine if this is in fact the case. This type of analysis will also help us decipher receptor subtype composition and how it relates to receptor functional characteristics. The single cell RT -PCR technique, in conjunction with patch clamp recordings, is a promising approach for unraveling this complicated puzzle. ExPerimental Procedures Reverse Transcription Reaction The M-MLV RT mix consisted of (in mM unless otherwise stated) 50 KCI, 10 Tris-HCI (pH 8.3), 5 MgCI2, 1 DIT, 1 each dNTP (Gibco-BRL), 2.5 J.IM random hexamers (Gibco-BRL), 0.33 U/J.1l Rnase inhibitor (5 Prime-3 Prime Inc., Boulder CO.), RNA (either from aspirated single cells or amounts given in text for RNA tissue extracts) and 10 U/JlI Superscript™ II reverse transcriptase (Gibco-BRL) in sterile nonpyrogenic water (Baxter). Nonpyrogenic water tested negative for Rnase activity using the RNaseAlert™ RNase detection kit (Ambion). Single cell RT reactions had a total volume of 10 J.1l each; other RT reactions (Le., from fetal cortex RNA dilutions) had volumes from 10 J.1l to 200 ,.d. RT reactions were held at 25°C for 10 minutes, 37°C for 60 minutes, 95°C for 5 minutes in a PTC-100 thermal cycler (MJ Research Inc., Watertown MA.), and then stored at 4°C until running PCR. The Tth (Boehringer Mannheim) RT reactions consisted of (in mM unless otherwise stated) 10 Tris-HCl (pH 8.9), 90 KCl, 0.9 MnC12, 200 JlM each dNTP, 2.5 JlM random hexamers, RNA sample, and 0.2 U/J . Ll Tth enzyme. Each reaction had a total volun1e of 10 J1l. Reactions were held at 25°C for 10 minutes, and 70°C for 20 minutes in a PTC-100 thermal cycler, and then stored at 4°C until running PCR. 37 Retrotherm™ RT (Epicentre Technologies, Madison WI.) reactions consisted of (in mM unless otherwise stated) 10 Tris-HCl (pH 8.3),50 KCl, 1.5 MgC12, 0.75 MnS04, 0.2 each dNTP (Gibco-BRL), RNA sample, 0.5 U/J.11 Retrotherm™ RT enzyme and either 2.5 JlM random hexamers or 5 nglJ1l of the GABA-A receptor subunit general minus 0'., ~, '¥, and ~-actin minus primers (see Table 1). Each reaction had a volume of 20 J1l. Reactions were held at 25°C for 5 minutes, and 70°C for 10 minutes in a PTC-100 thermal cycler, and then stored at 4°C until running PCR. Polymerase Chain Reaction (PCR) All primers were constructed by the University of Utah Oligo Synthesis Facility; see Table 1 for primer sequences and specific primer Mg2+ concentrations and annealing temperatures used for mouse sequences. The GABA-A ~ upstream primer used for amplification from plasmids containing the human ~ sequence was TGCATATACAACTTCAGAGGTCACT, the 0'.2 specific downstream primer was the same one listed in Table 1. The PCR for the first round of nested PCR was performed by mixing (in mM unless otherwise stated) 50 KCl, 10 Tris-HCl (pH 8.3), 3.75 MgC12, 38 0.25 each dNTP, 3.0 ~ upstream and downstream general primers, and 0.05 U/JlI Taq ploymerase (Fisher) in sterile nonpyrogenic water. As substrate, we used either the entire 10 JlI single cell RT reaction mix, or amounts stated in the text (for RT from various RNA dilutions). The PCR reactions were held at 95°C for 2.5 minutes then cycled 25 times. Each cycle consisted of 95°C for 10 seconds, 57°C for 25 seconds, and 72°C for 35 seconds. One JlI samples of the general PCR reaction were analyzed in the second round of nested PCR in 25 JlI reactions for each specific GAB A-A subunit. These reactions consisted of (in mM unless otherwise stated) 50 KCI, 10 Tris­HCI (pH 8.3), 1.5-4.0 MgCI2, 0.25 each dNTP, 1.0 ~ upstream and downstream specific primers, and 0.025 U/JlI Taq ploymerase in sterile pyrogen-free water. These reactions were held at 95°C for 2.5 minutes then cycled 40 times as follows: 95 °c for 10 seconds, annealing temperature for 15 seconds, 72°C for 20 seconds. 4.0 JlI samples were analyzed on 2% agarose gels using either ethidium bromide or SYBR® Green I (FMC BioProducts, Rockland ME.) and U. V. illumination. PCR using the Tth enzyme consisted of (in mM unless otherwise stated) 10 mM Tris-HCI (pH 8.9), 100 KCI, 1.5 MgCI2, 50 Jlg/ml bovine serum albumin, 0.05% (v/v) Tween® 20, 0.2 each dNTP, 1.0 ~ upstream and downstream primers, 0.02 U/Jll Tth, and 1 JlI substrate from RT reactions (using M-MLV RT) as stated in the text. Samples were held at 95°C for 2.5 minutes then cycled 40 times as follows: 95°C for 10 seconds, 50°C for 15 seconds, 72°C for 20 seconds. Samples were analyzed with ethidium bromide and U.V. illumination of 2% agarose gels. 39 Total RNA Tissue Extracts RNA extraction from mouse fetal cortex, adult cortex, and adult cerebellum were obtained by homogenization of the described tissues in and subsequent treatment with the TRIzolTM LS reagent system (Gibco-BRL). The RNA samples obtained were stored at -70°C as 1 J.!glJ.1l stocks in sterile nonpyrogenic water. Aspiration Electrodes Aspiration electrodes were pulled from soda lime glass microhematocrit capillary tubes (Baxter) using a P-87 micropipette puller (Sutter Instruments) to a tip diameter of approximately 3 J.!m. 40 References Altschul, S.F., Gish, W., Miller, W., Myers, E.W., and Lipman, DJ. (1990). Basic local alignment search tool. J. Mol. BioI. 215,403-410. Arkett, S.A., Dixon, J., Yang, J.N., Sakai, D.D., Minkin, C., and Sims, S.M. (1994). Mammalian osteoclasts express a transient potassium channel with properties of Kv1.3. Receptors Channels. 2, 281-293. Angelotti, T.P., and Macdonald, R.L. (1993). Asserrlbly of GABAA receptor subunits: alpha 1 beta 1 and alpha 1 beta 1 gamma 2S subunits produce unique ion channels with dissimilar single-channel properties. J. Neurosci. 13,1429-1440. Bateson, A.N., Lasham, A., and Darlison, M.G. (1991). y-aminobutyric acidA receptor heterogeneity is increased by alternative splicing of a novel ~ subunit gene transcript. J. Neurochem.56, 1437-1440. Benke, D., Mertens, S., Trzeciak, A., Gillessen, D., and Mohler, H. (1991). GABAA receptors display association of Y2 subunit with 0.1 and ~2J3 subunits. J. BioI. Chern. 266, 4478-4483. Burt, D.R., and Kamatchi, G.L. (1991). GABAA receptor subtypes: from pharmacology to molecular biology. FASEB J. 5, 2916-2923. Chamberlain, J.S., and Chamberlain, J.R. (1994). Optimization of Multiplex PCRs. In The Polymerase Chain Reaction, K.B. Mullis, F. Ferre and R.A. Gibbs, eds. (Boston, Mass.: Birkhauser), pp. 97-109. Chelley, J., and Khan, A. (1994). RT-PCR and mRNA quantitation. In The Polymerase Chain Reaction, K.B. Mullis, F. Ferre and R.A. Gibbs, eds. (Boston, Mass.: Birkhauser), pp. 97-109. Chou, Q., Russell, M., Birch, D.E., Raymond, 1., and Bloch, W. (1992). Prevention of pre-PCR mis-priming and primer dimerization improves low-copy-number amplifications. Nucleic Acids Res. 20,1717-1723. Davies, P.A., Hanna, M.C., Hales, T.G., and Kirkness, E.F. (1997). Insensitivity to anaesthetic agents conferred by a class of GABA(A) receptor subunit.Nature 385, 820-823. Dunn, S.M., Bateson, A.N., and Martin, I.L. (1994). Molecular neurobiology of the GABAA receptor. Int. Rev. Neurobiol. 36, 51-96. Eberwine, 1., Yeh, H., Miyashiro, K., Cao, Y., Nair, S., Finnell, R., Zettel, M., and Coleman, P. (1992). Analysis of gene expression in single live neurons. Proc. Natl. Acad. Sci. USA 89, 3010-3014. Endo, S., and Olsen, R.W. (1993) Antibodies specific for alpha-subunit subtypes of GABAA receptors reveal brain regional heterogeneity. J. Neurochem. 60, 1388-1398. Grigorenko, E.V., and Yeh, H.H. (1994). Expression profiling of GABAA receptor beta-subunits in the rat retina. Vis. Neurosci. 11, 379-387. 41 Hamill, D.P., Marty, A., Neher, E., Sakmann, B., and Sigworth, F.J. (1981). Improved patch-clanlp techniques for high-resolution current recording from cells and cell-free membrane patches. Pflugers Arch 391,85-100. Harvey, RJ., Chinchetru, M.A. and Darlison, M.G. (1994). Alternative splicing of a 51-nucleotide exon that encodes a putative protein kinase C phosphorylation site generates two forms of the chicken l'-aminobutyric acidA receptor ~2 subunit. J. Neurochem. 62, 10-16. Jackson, D.P., Hayden, J.D., and Quirke, P. (1991). Extraction of nucleic acid from fresh and archival material. In PCR A Practical Approach. MJ. McPherson, P. Quirke, and G.R. Taylor eds. (New York: Oxford Univ. Press), pp. 29-50. Johansen, F.F., Lambolez, B., Audinat, E., Bochet, P., and Rossier, J. (1995). Single cell RT-PCR proceeds without the risk of genomic DNA amplification. Neurochem. Int. 26, 239-243. Jonas, P., Racca, C., Sakmann, B., Seeburg, P.H., and Monyer, H. (1994). Differences in Ca2+ permeability of AMP A-type glutamate receptor channels in neocortical neurons caused by differential GluR-B subunit expression. Neuron 12, 1281-1289. Karrer, E.E., Lincoln, J.E., Hogenhout, S., Bennett, A.B., Bostock, R.M., Martineau, B., Lucas, W.J., Gilchrist, D.G., and Alexander, D. (1995). In situ isolation of mRNA from individual plant cells: creation of cell-specific cDNA libraries. Proc. Natl. Acad. Sci. USA 92,3814-3818. Kawasaki, E.S. (1990). Amplification of RNA. In PCR Protocols: A Guide to Methods and Applications, M.A. Innis ed. (San Diego, California: Academic Press), pp. 21-27. Kirkness, E.F., Kusiak, J.W., Fleming, J.T., Menninger, J., Gocayne, J.D., Ward, D.C., and Venter, J.C. (1991). Isolation, characterization, and localization of human genomic DNA encoding the beta 1 subunit of the GABAA receptor (GABRB 1). Genomics. 10, 985-995. Kirkness, E.F., and Fraser, C.M. (1993). A strong promoter element is located between alternative exons of a gene encoding the human l'-aminobutyric acid type A receptor ~3 subunit (GABRB3). 1. BioI. Chern. 268,4420-4428. Lambolez, B., Audinat, E., Bochet, P., Crepel, F. and Rossier, 1. (1992). AMPA receptor subunits expressed by single Purkinje cells. Neuron 9, 247-258. Laurie, OJ., Wisden, W., and Seeburg, P.R. (1992). The distribution of thirteen GABAA receptor subunit mRNAs in the rat brain. ITI. Embryonic and postnatal development. 1. Neurosci. 12,4151-4172. 42 Liiddens, R., and Wisden, W. (1991). Function and pharmacology of multiple GABAA receptor subunits.Trends PharmacoI. Sci. 12,49-51. Macdonald, R.L., and Olsen, R.W. (1994). GABAA receptor channels. Annu. Rev. Neurosci. 17, 569-602. Mohler, R., Mahlerbe, A., Draguhn, A., Sigel, E., Sequier, 1. M., Persohn, E., and Richards,1. G. (1990). GABAA-receptor subunits: functional expression and gene localisation. In GABA and Benzodiazepine Receptor Subtypes, G. Biggio and E. Costa, eds. (New York: Raven Press), pp.23-33. Moss, S.l., Gorrie, G.R., Amato, A., and Smart, T.G. (1995). Modulation of GABAA receptors by tyrosine phosphorylation. Nature 377, 344-348. Noonan, K.E., and Roninson, LB. (1988). mRNA phenotyping by enzymatic amplification of randomly primed cDNA. Nucleic Acids Res. 16, 10366 Porter, N.M., Angelotti, T.P., Twyman, R.E., and Macdonald, R.L. (1992). Kinetic properties of (Xl ~l 'Y-aminobutyric acidA receptor channels expressed in chinese hamster ovary cells: regulation by pentobarbital and picrotoxin. Mol. Pharmacol. 42, 872-881. Pritchett, D.B., Sontheimer, R., Shivers, B.D., Ymer, S., Kettenmann, R., Schofield, P.R., and Seeburg, P.R. (1989). Importance of a novel GABAA receptor subunit for benzodiazepine pharmacology. Nature 338,582-585. Puia, G., Vicini, S., Seeburg, P.R., and Costa, E. (1991). Influence of recombinant 'Y­aminobutyric acid-A receptor subunit composition on the action of allosteric modulators of 'Y-aminobutyric acid-gated cr currents. Mol. Pharmacol. 39, 691-696. Sommer, B., Poustka, A., Spurr, N.K., and Seeburg, P.R. (1990). The murine GABAA receptor delta-subunit gene: structure and assignment to human chromosome 1. DNA Cell BioI. 9, 561-568. Stephenson, F.A. (1995). The GABAA receptors. Biochem. 1.310, 1-9. Surmeier, DJ., Song, WJ., and Yan, Z. (1996). Coordinated expression of dopamine receptors in neostriatal medium spiny neurons. J. Neurosci. 16, 6579-6591. Twyman, R.E., Rogers, CJ., and Macdonald, R.L. (1989). Differential regulation of gamma-aminobutyric acid receptor channels by diazepam and phenobarbital. Ann. Neurol. 25, 213-220. 43 Veres, G., Gibbs, R.A., Scherer, S.E., and Caskey, C.T. (1987). The molecular basis of the sparse fur mouse mutation. Science 237,415-417 Whiting, P., McKernan, R.M., and Iversen, L.L. (1990). Another mechanism for creating diversity in y-aminobutyrate type A receptors: RNA splicing directs expression of two forms ofY2 phosphorylation site. Proc. Natl. Acad. Sci. U.S.A. 87,9966-9970. Wisden, W., Laurie, DJ., Monyer, R., and Seeburg, P.R. (1992). The distribution of 13 GABAA receptor subunit mRNAs in the rat brain. I. Telencephalon, diencephalon, mesencephalon. J. Neurosci.12, 1040-1062. CHAPTER 3 AN ANALYSIS OF THE SINGLE CELLRT-PCR TECHNIQUE FOR TYPE I ANDTYPE II ERROR RATES USING A MIXED POPULATION OF DEFINED GABA-A RECEPTORS Summary Single cell reverse transcription polymerase chain reaction (single cell RT -PCR) in conjunction with whole cell patch clamp recordings from stably and transiently transfected human embryonic kidney cells (HEK 293) expressing GABA-A receptors of known compositions was performed to analyze the detection properties of the single cell RT -PCR method. The main parameters of interest were the rates for type I (false +) and type II (false -) errors. The calculated error rates were 0.03 for type I and 0.29 for type II. Cells with functional GABA-A responses were divided into two populations, RT -PCR negative and RT -PCR positive, depending on whether GABA-A subunit expression was detected. RT -PCR negative cells showed statistically lower peak currents in response to 10 J.1M GABA than RT-PCR positive cells, suggesting that the GABA-A subunit mRNA concentration in these cells is directly proportional to the level of functional protein. Furthermore, there appears to be a threshold mRNA 45 concentration for detection using single cell RT-PCR, below which the proportion of false negatives increases dramatically, greatly increasing the overall type II error rate. Excluding the cells expressing low amounts of receptors (using electrophysiological criteria) from single cell RT-PCR analysis greatly reduces the type II error rate, making the analysis more powerful statistically. Introduction y-Aminobutyric acid (GAB A) is the major neurotransmitter involved in synaptic inhibition in the central nervous system (CNS). The GABA-A receptor is a multimeric protein complex that forms a ligand-gated chloride selective ion channel, similar to other receptors in the ligand-gated receptor superfamily. This superfamily includes: GAB A-A, glycine, neuronal nicotinic acetylcholine, and 5-HT3 (5- hydroxytryptamine type 3; Schofield et al. 1987). The molecular structure of the GABA-A receptor is thought to be a pentamer formed by the combination of various subunits, inserted as a transmembrane glycoprotein complex into the outer membrane of the cell. Currently, there are at least fifteen distinct subunits of the GABA-A receptor for which the protein and nucleotide sequences are known. These subunits are grouped into families according to sequence homology as follows: "1-6' Pl-4' y 1-3' 0, and £ (Burt and Kamatchi, 1991; Macdonald and Olsen, 1994; Stephenson 1995; Dunn et al., 1994; Davies et aI., 1997). In addition, splice variants have also been described for the 12 (Whiting et aI., 1990), ~2 (Harvey et al., 1994), ~3 (Kirkness and Fraser, 1993), and ~4 (Bateson et al., 1991) subunits. The molecular makeup of the receptor complex has been shown to have a pronounced effect on: the single channel conductance and kinetics (Porter et al., 1992; Angelotti and Macdonald, 1993), the apparent binding affinity of GABA (Mohler et aI., 1990), and the pharmacological modulation of the receptor (Liiddens and Wisden, 1991; Pritchett et al., 1989, Puia et al., 1991). 46 Currently, the most predominant methods used for mapping neurotransmitter receptor subunit expression in the brain are in situ hybridization, radioligand binding, and immunohistochemistry (Benke et aI., 1991, Wisden et al., 1992). The first of these measures mRNA expression and the other two measure protein levels. These methods have been invaluable in providing the data leading up to our current understanding of the GABA-A and other neurotransmitter receptor systems. However, since in situ hybridization has not been performed simultaneously on the same preparation as either of the other two, some speculation has occurred as to how GABA-A receptor mRNA and protein levels are actually related. Single cell reverse transcription polymerase chain reaction (RT-PCR) is a powerful and relatively new technique used to analyze the mRNA expression of individual cells (Eberwine et al., 1992; Lambolez, 1992; Jonas et al., 1994). One of the most powerful features of this technique is the mRNA analysis can be combined with the functional analysis of ion channels using patch clanlp recording. This allows the experimenter to analyze both the mRNA and functional protein expression of a particular ion channel in the same cell, and draw direct inferences between the two. However, a concern has been the undefined parameters of the single cell RT -PCR method, such as the occurrences of type I errors (false +), and 47 type IT errors (false -). With these rates unknown, conclusions drawn from the analysis of obtained single cell RT -PCR data could be questionable. The purpose of the following experiments was to determine the accuracy and reliability of the single cell RT -PCR technique. We used cultures expressing mixtures of cells expressing known GABA-A subunit combinations that could be electrophysiologically distinguished to experimentally determine the rates for type I and type IT errors for the single cell RT -PCR protocol explained herein. By defining these parameters using a defined system, it becomes easier to interpret the RT-PCR data obtained in other, more complex experiments. Results Predefined GABA-A Receptor Populations in Mixed Cultures A stably transfected HEK 293 cell line expressing a,1~lY2:t (generously provided by Dr. Donald Carter, Upjohn; Hamilton et al., 1993) and transiently transfected HEK 293 cells expressing <lz~1 GABA-A subunits were mixed together in culture. The a,1~2'Y2S receptor combination shows a large enhancement when submaximal GABA concentrations are co-applied with diazepam (Figure 7). Since the <lz~l receptors lack a 'Y subunit, they show no enhancement of GABA evoked currents in the presence of diazepam (Figure 7; Pritchett, 1989). Thus, we could expect the two receptor populations to be clearly distinguishable by their pharmacology when mixed together. Whole cell recordings were taken from culture dishes containing the mixed populations 48 :I! 300 IeIn! 250 ...I . -+ 200 ID 10 microM GABA II( -Q. 150 -cco: 100 [2 10 microM GABA + .•... 1 microM Diazepam! ::I 50 .C.J 0 CI A•. Alpha1 Alpha2 cC:I Beta2 Beta1 :•IE Gamma2S Figure 7. Comparison of al~2Y2S and ~~1 mean peak currents with the application of 10 J.LM GABA and 10 J.LM GABA + 1 J.LM diazepam (currents are from the RT-PCR positive groups only). Receptor types are easily distinguished electrophysiologically by the presence (al ~2 Y2S) or absence (~~l) of diazepam enhancement. of cells and the peak current responses to 10 J.LM GABA, 10 J.LM GABA + 1 J.1M diazepam, and 1 mM GABA were measured as an assesment of functional GAB A-A receptor protein characteristics. Single cell RT -PCR performed on the cells after recording was used as an assesment of GABA-A receptor subunit mRNA expression. By using the electrophysiological characterizations of the cells to establish the actual subunit combination, we could calculate what the occurrences were for false positives and false negatives in the RT-PCR reactions and determine the type I and type IT error rates. Thirty cells were recorded from and analyzed for correlation of electrophysiological characteristics and single cell RT-PCR for the at and ~ GABA-A subunits. Twenty of the cells showed diazepam enhancement, 8 cells showed no diazepam enhancement, and 2 cells showed no GABA current. In addition, as further 49 negative controls, RT-PCR was performed from 2 samples where only external solution was aspirated. Sixteen of the 20 cells (80%) that showed a diazepam enhancement (al~2'Y2S) also showed the correct RT-PCR results, the remaining 4 showed no RT-PCR products. Four of the 8 cells (50%) lacking the diazepam enhancement (a2~1) gave the correct ~ subunit RT -PCR results, 1 of which was also positive for a l . The remaining 4 showed no detectable RT-PCR products. Both the 2 external solution samples and the 2 cells with recordings that showed no GABA response were negative for RT-PCR products. The type I error rate is calculated by dividing the number of false positive PCR reactions (1) by the total number of reactions that should have had negative RT­PCR results based on the cells' electrophysiological responses. The 20 cells with diazepam enhancement should all be negative for ~, the 8 cells without diazepam enhancement should all be negative for aI' and the 2 cells without a GABA response and the 2 external only samples should be negative for both a l and ~ (8 negative control reactions), equaling 36 total expected negative reactions. This gives a calculation of 2.80/0 (1/36) false positives, for a type I error rate of 0.03. The type II error rate is calculated by taking the total number of false negative reactions (8) and dividing by the total number of reactions that should have been positive (i.e. 20 a l + 8 ~ = 28 expected positive reactions). This gives a calculation of 28.6% (8/28) false negatives, so the type II error rate = 0.29. 50 Comparison of RT-PCR Positive and RT-PCR Ne~ative Cells In order to understand the basis for false negative RT -PCR results, we analyzed the peak currents (a direct measure of functional protein levels) to the application of 10 J.1M GABA for those cells that were RT-PCR positive and those that were RT-PCR negative. Since the two receptor populations have different efficacies to 10 J.1M GABA we analyzed the ~~l responses separately from the a l ~iY2S responses. The mean peak current for pharmacologically identified al~tf2s cells with a negative RT-PCR identification was 31 pA (S.D. = 16.5) whereas the mean peak current for positively identified RT-PCR cells was 88pA (S.D. = 86.5; Figure 8). Comparison of the mean peak currents of the two populations for statistically significant difference was performed using a Student's t-test for samples assuming unequal variance (used because the two samples were of greatly different sizes, n = 4 and 16; Moser and Stevens, 1992) giving a two-tailed P value of 0.024. The mean peak current in response to 10 J.1M GABA for the ~~l cells without an RT-PCR identification was 65 pA (S.D. = 29.9) while the mean peak current for ~~l RT -PCR positively identified cells was 178 pA (S.D. = 80.8; Figure 8). A comparison of these two groups using a Student's t-test for samples assuming equal variance (n = 4 and 4) gave a two-tailed P value of 0.039. These results suggest that HEK cells with low levels of functional protein expression (i.e. low peak currents) also have low levels of mRNA expression. With cells showing weak whole cell current responses, the mRNA levels are often low enough to be at or below the detection limit of our single cell RT -PCR protocol. 51 . :& 250 w U) ...I. . 200 -+ c( -.A., . 150 D RT- PCR Negative .!c. 100 RT- PCR Positive ::::I (J .lIII: 50 II a•. c 0 II :•& Alpha1 Alpha2 Beta2 Beta1 Gamma2S Figure 8. Comparison of mean peak currents (in pA) of al~lY2S and a2~1 receptors for cells that were RT-PCR negative (n=4) and RT-PCR positive (n=16) from the application of 10 mM GABA. The difference in peak current between RT -PCR positive and RT -PCR negative cells was statistically significant for both receptor types (P<0.05 Student's t-test, two-tailed). RT -PCR Detection Threshold To study how limiting concentrations of template can affect the type II error rate of the PCR reaction we used limiting dilutions of plasmids containing the human a l or Cl:z GABA-A receptor subunits. A set of 1:10 serial dilutions of the plasmids from 1 flglfll stocks was made down to 0.1 fglJll. PCR reactions were then run using 1 Jll samples from these stocks as the substrate. The limiting DNA concentration was the last dilution from which PCR product was obtained. By using 0.5 Jll samples from the limiting dilution as the substrate and running multiple PCR reactions at the same concentrations, we observed a concentration dependent effect on the rate of false negatives. Using the al plasmid as a template, 100% (15/15) of the reactions using 1.0 52 fg substrate were positive, 46.7% (7115) of the reactions using 0.5 fg substrate were positive, and 0% (0115) of the reactions using 0.1 fg substrate were positive. For the ~ subunit the results were 100% (15115) for 1.0 fg substrate, 73.3% (11/15) for 0.5 fg substrate, and 0% (0/15) for 0.1 fg substrate (Figure 9). Thus we can see that by decreasing the substrate added, the occurrence of false negatives in the peR reaction increases in a concentration dependent fashion, eventually passing the limit for detection. This supports the idea that cells with low levels of target mRNA might not give reliable single cell RT -peR results. Similar findings on the occurrence of false negatives with limiting amounts of single cell RT-PCR substrate have also been reported by Karrer et al. (1995) and Surmeier et al. (1996). Discussion The single cell RT -peR technique used in combination with patch clamp recording is rapidly becoming a mainstream method used for the investigation of mRN A and functional protein expression. However, detailed analysis of the properties and limitations of this method has been lacking. Such knowledge is important to have a better understanding of how to interpret the obtained results. We wanted to calculate what the experimentally determined type I (false +, statistically called a) and type II (false -, statistically called ~) error rates were. Figure 9 shows that the a 1 and a.z primer pairs have similar limits of detection (close to 0.5 fg of DNA). The type IT error rates near the detection limit also appear to be similar, with ~ actually having a slightly lower type IT error rate than a1• Because 100%--.--- 80%-+--- 60%-+--- 40%-+--- 20%--+---- 0% --'------ Alpha 1 Success rate Alpha 2 Success rate 1.0 fg 0.5fg II 0.1 fg Figure 9. Effects of decreasing the concentration of substrate when near the detection limit for PCR. Each bar represents the success rate for 15 reactions. As the concentration decreases, the success rate decreases as well, giving a corresponding increase in the type II error rate. 53 the error rates at the detection limit are similar, we chose to pool together the RT-PCR data obtained from the transfected cells for both subunits to calculate the overall type I and type II error rates of the RT-PCR procedure. The differences seen in the ratio of false negatives to true positives in the two transfected populations are likely due to the differences in the way the two populations were transfected. The, a l ~{f2S cells are a stably transfected cell line, which one would expect to have a more uniform level of mRNA expression of the transfected subunits than transiently transfected cells such as the ~~1 population, where extremely variable levels of expression can be seen. This could cause a higher false negative rate for detection of ~ in the transiently transfected cells, even when the type II error rate for the ~ primers would be expected to be 54 slightly lower than for the a l primers (Figure 9). By convention, a result is usually considered significant only if the chance of making a type I error is less than 5% (i.e. P<0.05). Power is defined as the probability that a statistically significant result will be found, and the statistical power of a test is inversely related to the chance of making a type IT error. The formula for power is: 1- p, where ~ is the false negative probability rate (Hays, 1988). This means that the higher the type IT error rate is, the lower the power of the test (and the lower the probability of actually detecting a subunit that is present). For the reliable interpretation of data, it is preferable to have the power of a test as high as possible, so the false negative rate should be as low as possible. Our results show that the actual type I error rate for our single cell RT-PCR analysis is less than 0.05 (a = 0.03), but the type IT error rate is somewhat high, and would not normally be considered acceptable (~ = 0.29, power = 0.71). These findings could lead one to doubt the validity of excluding the expression of certain receptor subunits based on their nondetection using single cell RT -PCR due to the possibility of obtaining a false negative result. For example, if the power of detecting a subunit is 0.71, the probability of obtaining a false negative result for a subunit would be 0.29. One way to increase confidence in interpreting negative RT -PCR results would be to pool together data from cells that had identical responses (assuming it is possible to sufficiently distinguish different responses functionally and that identical functional responses are given only by identical subunit mRNA expression patterns). The probablilty of getting a false negative for the same subunit from two cells expressing the same subunits would be 0.29 X 0.29, or 0.08. Similarly, 55 if it were possible to group together three cells with identical patterns of expression (as measured by identical function) the probability of getting the same three false negatives would be 0.293 , or 0.024. However, it may not always be possible to sufficiently resolve similarities and differences in responses in order to properly group cells together. In this case, it may be easier to increase the statistical power (giving more confidence in the interpretation of negative results) by decreasing the the type II error rate. In the study given above, the type II error rate can be reduced considerably by examining the differences in the populations that gave positive RT -PCR results and those that did not and using the observed criteria to exclude from analysis cells that might give ambiguous results. The RT-PCR positive and RT-PCR negative populations had statistically different mean peak currents. By removing 95% of the al~t(2S RT-PCR negative population by excluding all cells with currents below two standard deviations above the mean (31 pA + [2 X 16.5 pAl = 64 pA, assuming the currents form a normal distribution) the proportion of false negatives in the al~2'Y2S cells drops from 4/20 to 0/7, to give a type II error rate of 0.0 instead of 0.20. If the same calculations are performed on the ~~l cells as well, removing from consideration the RT-PCR results of all cells with currents less than 124.8 pA (65.0 pA + [2 X 29.9] = 124.8 pA), the proportion of false negatives drops from 4/8 to 0/3. Taken together with the al~2'Y2S data, this changes the overall type II error rate from 0.29 to 0.00. This means a corresponding increase in the calcuated power from 0.71 to 1.0. This gives a substantial increase in the statistical credibility of the .obtained RT -PCR data, allowing one to more reliably exclude the expression of a certain subunit when a negative RT -PCR result is given. The one drawback is a decrease in the value of n, the number of cells used for evaluation. 56 The levels of mRNA and functional protein appear to be correlated in both the stably and transiently transfected HEK 293 cells. If this is also true for neurons, then single cell RT -PCR from neurons would be considerably more reliable in determining subunit composition from cells with good current responses. Recent reviews containing comparisons of published reports of in situ hybridization and immunohistochemistry for the GABA-A receptor system generally show good agreement in subunit expression patterns (Stephenson, 1995; Dunn et aI., 1994). This suggests that at least for the GABA-A system, the levels of protein appear to be similar to the levels of mRNA. As such, it may be neccessary to exclude cells with weak current responses when trying to correlate patch clamp recordings with single cell RT­PCR in order to confidently interpret the data. Experimental Procedures Stably and Transiently Transfected HEK 293 Cells The stably transfected HEK 293 cell line expressing the al~21'2s GABA-A subunit combination was obtained from Dr. Donald Carter, at the Upjohn Company. Plasmids containing the human ~~l GABA-A receptor subunit combination were obtained from Dr. Dolan Pritchett. Transient transfections of HEK 293 cells (A TCC CRL 1579) were performed using the standard C8.:2P04 transfection protocol (Chen and Okayama, 1987) using 20 J,Lg of ~~l plasmid per 10 em dish at approximately 40% 57 confluency. The 10 cm dishes of transformed cells were washed with MEM (Sigma) 4 hours after transfection, and then replated as mixed cultures with the stably transfected cells in 35 mm dishes. Recordings were performed 48 hours following transfection. Whole Cell Patch Clamp Recordings Recordings were performed using an Axopatch 200A amplifier (Axon Instruments) in whole cell voltage clamp mode, with the holding potential set at -75 mY, at room temperature (20-25°C). Recordings were documented on a Gould 2400S chart recorder (Gould Instrument Systems, Inc.) and as computer files using Axotape software (Axon Instruments) on a 486DX2-66 PC compatible computer (Gateway 2000). Electrodes were pulled from soda lime glass micro hematocrit capillary tubes (Baxter) using a P-87 micropipette puller (Sutter Instruments) to a resistance of approximately 2.5 Mn in the recording solutions. External solution consisted of (in mM) 142 NaCI, 1.5 KCI, 1 CaCI2, 1 MgCI2, 10 glucose, 30 sucrose, 10 Na-HEPES at pH 7.4, and adjusted to 320 mOsm with sucrose. Internal solution consisted of (in mM) 140 CsCI, 4 MgCI2, 10 Na-HEPES, 5 EGTA at pH 7.4 and adjusted to 290 mOsm with sucrose. Both internal and external were made from nonpyrogenic sterile water for irrigation USP (Baxter) and found free of RNase contamination after testing with RNaseAlert™ (Ambion Inc.). 1mM and 10 J.1M GABA solutions were diluted from 1M GABA stock in distilled water (Sigma). 1J.1M diazepam was diluted from 1mM diazepam stocks in DMSO (Sigma) giving a final DMSO concentration of 0.1 %, which has previously been shown to not affect GAB A-A receptors (Nakahiro et al., 1992). GABA and GABA + diazepam solutions were applied in 2 second applications from a gravity driven multi-valve (valves from The Lee Co., Essex CT.) perfusion system using a ValveBank™ 8IT (Automate Scientific, Inc.) controller. SiOide Cell RT-PCR 58 In order to minimize RNase contamination, gloves were worn during the entire recording/aspiration process. Gloves were also worn when handling RT and PCR samples. After recordings were performed, cells were aspirated by replacing the recording electrode with a larger (approximately 1 Mil resistance with the recording solutions given) electrode containing 2 III of the above mentioned sterile water (Baxter) using gentle suction from an attached 5 cc syringe. The contents of the aspiration electrode were expelled into RnaselDNase free 0.2 ml thin wall tubes (Life Science Products, Inc., Denver CO.) containing 8 Jll of the RT reaction mix. The RT mix consisted of (in mM unless otherwise stated): 50 KCl, 10 Tris-HCl pH 8.3, 5 MgCI2, 1 DTT, 4 dNTP mix (Gibco), 2.5 JlM random hexamers (Gibco), 0.33 VIllI Rnase inhibitor (5 Prime-3 Prime Inc., Boulder CO.), and 10 VIllI Superscript™ IT reverse transcriptase (Gibco) in sterile water (Baxter). RT reactions were held at 25°C for 10 minutes, 37°C for 60 minutes, 95°C for 5 minutes in a PTC-100 thermal cycler (MJ Research Inc., Watertown MA), and then stored at 4°C until running PCR. A general PCR reaction to amplify all GABA-A a subunits present was performed using CTCCTGATACATTCTTCCACAATGG as the.upstream primer and 59 CAGACAGCAA TGAACCA(GA)TCCAT for the downstream primer. All primers were constructed by the University of Utah Oligo Synthesis Facility. The entire 10 J.lI RT reaction mix above was added to the PCR reaction mix, consisting of (in mM unless otherwise stated) 50 KCI, 10 Tris-HCI pH 8.3, 3.75 MgCI2, ImM dNTP mix (Gibco), 0.5 J.1.M each primer, and 0.05 U/J.lI Taq ploymerase (Fisher) in sterile water (Baxter). The PCR reactions were held at 95°C for 2.5 minutes then cycled 25 times. Each cycle consisted of 95°C for 10 seconds, 57°C for 25 seconds, and 72°C for 35 seconds. 1 J.lI samples of the general PCR reaction were then analyzed in 25 J.lI specific nested PCR reactions for the 0.1 and ~ GABA-A ~ubunits. The 0.1 primers used were GCTCCTGCGTATCACAGAGG upstream and GGAAGTGAGTCGTCATAACCAC downstream. The ~ primers used were TGCATATACAACTTCAGAGGTCACT upstream and ACCTGTACTGGATTTAATTGTTTCC downstream. These reactions were held at 95°C for 2.5 minutes then cycled 40 times as follows: 95°C for 10 seconds, 60°C for 15 seconds, 72°C for 20 seconds. Subsequently, 4.0 J.1.l samples were analyzed on 2% agarose gels using ethidium bromide and U.V. illumination. PCR from Plasmid DNA Dilutions of plasmids expressing the human GAB A-A 0.1 and ~ subunits were made in TE buffer as explained in the text. PCR reactions were mixed up in groups of 15 and then aliquoted into separate tubes to run in the cycler. The same specific at and ~ upstream primers were used as listed above, and the a general downstream primer listed above was used for both subunits. Reactions were cycled 40 times using the 60 same protocol as listed above for the specific 0.1 and ~ reactions. Statistical Analysis Microsoft Excel version 5.0 for Windows™ was used to perform the Student's t -test analysis of the data, and to calculate the mean, standard error of the mean (S.E.M.), and standard deviation of the peak currents for the different populations. 61 References Angelotti, T.P., and Macdonald, R.L. (1993). Assembly of GABAA receptor subunits: alpha 1 beta 1 and alpha 1 beta 1 gamma 2S subunits produce unique ion channels with dissimilar single-channel properties. J. Neurosci.13, 1429-1440. Bateson, A.N., Lasham, A., and Darlison, M.G. (1991). y-aminobutyric acidA receptor heterogeneity is increased by alternative splicing of a novel ~ subunit gene transcript. J. Neurochem. 56, 1437-1440. Benke, D., Mertens, S., Trzeciak, A., Gillessen, D., and Mohler, H. (1991). GABAA receptors display association ofY2 subunit with (Xl and ~213 subunits. J. BioI. Chern. 266, 4478-4483. Burt, D.R., and Kamatchi, G.L. (1991). GABAA receptor SUbtypes: from pharmacology to molecular biology. FASEB J. 5, 2916-2923. Chen, C., and Okayama, H. (1987). High-efficiency transformation of mammalian cells by plasmid DNA. Mol. Cell. BioI. 7,2745-2752. Davies, P.A., Hanna, M.C., Hales, T.G., and Kirkness, E.F. (1997). Insensitivity to anaesthetic agents conferred by a class of GABA(A) receptor subunit.Nature 385, 820-823. Dunn, S.M., Bateson, A.N., and Martin, I.L. (1994). Molecular neurobiology of the GABAA receptor. Int. Rev. NeurobioI. 36, 51-96. Eberwine, J., Yeh, H., Miyashiro,K., Cao, Y., Nair, S., Finnell, R., Zettel, M., and Coleman, P. (1992). Analysis of gene expression in single live neurons. Proc. NatI. Acad. Sci. USA 89,3010-3014. Hamilton, BJ., Lennon, DJ., 1m, H.K., 1m, W.B., Seeburg, P.H., and Carter, D.B. (1993). Stable expression of cloned rat GABAA receptor subunits in a human kidney cell line. Neurosci. Lett. 153, 206-209. Harvey, R.J., Chinchetru, M.A., and Darlison, M.G. (1994). Alternative splicing of a 51-nucleotide exon that encodes a putative protein kinase C phosphorylation site generates two forms of the chicken y-aminobutyric acidA receptor ~2 subunit. J. Neurochem. 62, 10-16. Hays, W.L. (1988) Hypothesis testing. In Statistics. (Orlando FL. Holt, Rinehart and Winston), pp263-266. Jonas, P., Racca, C., Sakmann, B., Seeburg, P.H., and Monyer, H. (1994). Differences in Ca2+ permeability of AMP A-type glutamate receptor channels in neocortical neurons caused by differential GluR-B subunit expression. Neuron 12, 1281-1289. Karrer, E.E., Lincoln, J.E., Hogenhout, S., Bennett, A.B., Bostock, R.M., Martineau, B., Lucas, W.J., Gilchrist, D.G., and Alexander, D. (1995). In situ isolation of mRNA from individual plant cells: creation of cell-specific cDNA libraries. Proc. Natl. Acad. Sci. USA 92,3814-3818. 62 Kirkness E. F., and Fraser C. M. (1993). A strong promoter element is located between alternative exons of a gene encoding the human 'Y-aminobutyric acid type A receptor ~3 subunit (GABRB3). 1. BioI. Chern. 268,4420-4428. Lambolez B., Audinat E., Bochet P., Crepel F. and Rossier J. (1992). AMPA receptor subunits expressed by single Purkinje cells. Neuron 9, 247-258. Liiddens, H., and Wisden, W. (1991). Function and pharmacology of multiple GABAA receptor subunits. Trends Pharmacol. Sci. 12, 49-51. Macdonald, R.L., and Olsen, R.W. (1994). GABAA receptor channels. Annu. Rev. Neurosci. 17, 569-602. Mohler, H., Mahlerbe, A., Draguhn, A., Sigel, E., Sequier, J. M., Persohn, E., and Richards, J. G. (1990). GABAA-receptor subunits: functional expression and gene localisation. In GABA and Benzodiazepine Receptor Subtypes, G. Biggio and E. Costa, eds. (New York, Raven Press), pp.23-33. Moser, B.K., and Stevens, G.R. (1992). Homogeneity of variance in the two-sample means test. American Statistician. 46, 19-21. Nakahiro, M., Arakawa, 0., Narahashi, T., Ukai, S., Kato, Y., Nishinuma, K., and Nishimura, T. (1992). Dimethyl sulfoxide (DMSO) blocks GABA-induced current in rat dorsal root ganglion neurons. Neurosci. Lett. 138, 5-8. Porter, N.M., Angelotti, T.P., Twyman, R.E., and Macdonald, R.L. (1992). Kinetic properties of (Xl ~l 'Y-aminobutyric acidA receptor channels expressed in chinese hamster ovary cells: regulation by pentobarbital and picrotoxin. Mol. Pharmacol. 42, 872-881. Pritchett, D.B., Sontheimer, H., Shivers, B.D., Ymer, S., Kettenmann, H., Schofield, P.R., and Seeburg, P.H. (1989). Importance of a novel GABAA receptor subunit for benzodiazepine pharmacology. Nature 338, 582-585. Puia, G., Vicini, S., Seeburg, P.H., and Costa, E. (1991). Influence of recombinant 'Y­aminobutyric acid-A receptor subunit composition on the action of allosteric modulators of'Y-aminobutyric acid-gated cr currents. Mol. Pharmacol. 39, 691-696. 63 Scofield, P.R., Darlison, M.G., Fujita, N., Burt, D.R., Stephenson, F.A., Rodriguez, H., Rhee, L.M., Ramachandran, J., Reale, V., Glencorse, T.A., Seeburg, P.H., and Barnard, B.A. (1987). Sequence and functional expression of the GABAA receptor shows a ligand-gated receptor super-family. Nature 328,221-227. Stephenson, F.A. (1995). The GABAA receptors. Biochem. J. 310, 1-9. Surmeier, D.J., Song, W.J., and Yan, Z. (1996). Coordinated expression of dopamine receptors in neostriatal medium spiny neurons. J. Neurosci. 16, 6579-6591. Whiting, P., McKernan, R.M., and Iversen, L.L. (1990). Another mechanism for creating diversity in 1-aminobutyrate type A receptors: RNA splicing directs expression of two forms of 12 phosphorylation site. Proc. Natl. Acad. Sci. U.S.A. 87, 9966-9970. Wisden, W., Laurie, D.J., Monyer, H., and Seeburg, P.H. (1992). The distribution of 13 GABAA receptor subunit mRNAs in the rat brain. I. Telencephalon, diencephalon, mesencephalon. J. Neurosci.12, 1040-1062. CHAPTER 4 ULTRAFAST LIGAND EXCHANGE RECORDINGS IN CONJUNCTION WITH MRNA ANALYSIS SUmmary Ultrafast ligand exchange patch clamp recordings were performed from fetal mouse cortical neurons in primary culture to determine GABA-A receptor functional properties. Single cell reverse transcription polymerase chain reaction was done on the cells recorded from to analyze the mRNA expression of GABA-A subunits. Efforts to correlate the molecular expression and functional properties of GABA-A receptors using the data provided by these two techniques used in conjunction are discussed. Introduction Determination of Functional and Molecular Properties Patch clamp recording is a powerful analytical tool. It provides detailed information about how ion channel protein complexes function, and has been used to determine how pharmacological agents modulate the functional properties of ion channels. However, in order to resolve subtle distinctions in channel kinetics due to differences in subunit composition, many recordings need to be grouped together to 65 overcome the signal to noise ratio. This is not possible when trying to correlate the single channel patch clamp recordings of the GABA-A receptor from a single cell with single cell RT -PCR. In addition, most patch clamp studies to date have been limited to studying receptor populations under equilibrium conditions, which is not what occurs physiologically. For example, the study of diazepam's effects on the GABA-A receptor was performed under conditions of prolonged GABA applications, using relatively low concentrations of GABA (Twyman et aI., 1989). In the synapse, agonist is not available under steady state conditions. The release of a vesicle of neurotransmitter from the presynaptic bouton typically leads to a localized high concentration of agonist at the postsynaptic receptor, which is available only briefly before being rapidly cleared away by various active mechanisms. To what extent the information determined under steady state conditions can be extrapolated to synaptic conditions is relatively unkown. Ultrafast liiand exchanie. Ultrafast ligand exchange patch clamp recording is a newly developed technique for performing patch clamp recording studies under more synaptic-like conditions (Akaike et al., 1986; Brett et al., 1986; Franke et al., 1987; Maconchie and Knight, 1989; Colquhoun et al., 1992; Lavoie and Twyman, 1996) and is capable of distinguishing differences in receptor subunit composition at the single cell level (Lavoie, 1996). A theta tube (a double lumen glass tube, with a thin dividing septum) is pulled to a tip diameter of approximately 100 J,lm. One side of the theta tube is used to apply external recording solution and the other side contains external plus GABA (or other solution of interest). The tube is attached to a piezo-electric transducer capable of causing a rapid lateral translation of 40 J,lm. An outside-out patch 66 is pulled and placed within the stream of external about 200 Jlm from the mouth of the theta tube and about 20 J.Lm from the externallexternal+GABA interface (see Figure 10). Thus, when the piezo-electric transducer is triggered, the theta tube quickly moves sideways and exposes the patch to a brief "pulse" of GABA. The movement of the theta tube back into its original position moves the stream of external solution (after a programmed duration) back over the recording electrode, washing away the applied GABA and mimicking the synaptic conditions of having ligand present briefly at a high concentration and then rapidly cleared away. The time it takes for this application technique to reach full agonist concentration is about 100 JlS, while functionally the time it takes the channels to get from 10% peak current to 90% peak current is typically 500 to 1100 JlS, depending on .. Theta tube in translated position Piezo-electric: ·.1 •••••••••••••••• , •••••• transducer I " .:. 1-". .. . Recording Electrode moves theta : ••.• 11 ••••••••••••••••••••• tubel~e~I~~~~~~~~~~~~~~~~~~~_ •••••••• I I • • I • • • • • • • • • • • I Theta tube in original position Figure 10. Diagram of the theta tube apparatus (heavy lines, bottom left) used for ultrafast ligand exchange patch clamp recording. In the above diagram, external solution is coming out of the top half of the theta tube, and external plus GABA is coming out of the bottom half (solutions represented as thin lines coming out of theta tube). When a piezo-electric transducer is triggered, it causes the theta tube to shift laterally (represented as the theta tube in dotted lines) such that the tip of the recording electrode is exposed to the GABA containing solution. The duration of the exposure can be programmed by the user, for our experiments we used pulses giving GABA exposures 1 ms in duration. 67 molecular composition (Lavoie, 1996). Because the application of agonist occurs more rapidly than the physical opening of the channel, this technique allows the determination of receptor activation kinetics which have previously been unresolvable. Also, since the agonist is rapidly cleared away, the resolution of receptor relaxation kinetics from an open, fully agonist bound state to a closed, no agonist bound state is also possible. The activation and deactivation kinetics provide a fingerprint (termed a system response) for receptor functional properties. Since channel functional properties have previously been shown to be dependent on the molecular composition of the ion channel/receptor complex (reviewed in Burt and Kamatchi, 1991; Macdonald and Olsen, 1994), each kinetically different system response is thought to be typical of a different combination of receptor subunits. Sinile cell mRNA expression analysis. The ultrafast ligand exchange patch clamp recording technique can be performed together with single cell reverse transcription polymerase chain reaction (RT-PCR) analysis. This not only allows for the acquisition of highly detailed information about receptor functional properties, but also provides information about the receptor subunit mRNA combinations expressed by the cell the patch was obtained from. The relationship between mol~cular and functional properties of GAB A-A receptors has been studied in transfected cells and injected oocytes, but has been too difficult to examine in actual neurons because of the tremendous heterogeneity of receptors expressed and unknown molecular compositions of native receptors. We attempted to see if using the ultrafast ligand exchange patch clamp recording technique in conjunction with single cell RT-PCR could provide us 68 with information concerning the subunit composition of GABA-A receptors in 'cultured fetal mouse neurons. Results Our experimental design was to obtain ultrafast ligand exchange patch clamp recordings from 12 to 14 day old fetal mouse cortex neurons in primary culture and then analyze the neuron the patch was obtained from for the expression of GABA-A receptor (l1-3, ~1-3' and 12 subunit mRNAs using single cell RT -peR. Ultrafast Li~and Exchan~e Patch Clamp Recordinl:s In order to obtain good ultrafast ligand exchange recordings, it is neccessary to have many receptors (hundreds) present in the patch. To get as many receptors as possible, patches were pulled slowly from the neurons, making the patches as physically large as possible. However, even with very large patches, ultrafast ligand exchange recordings of GABA-A receptors require high receptor density in order to obtain a big enough current response to analyze. The great majority of neurons recorded from did not have a high enough receptor density to give useful results. A second complication with this technique is that high agonist concentrations (i.e., 10 mM GABA) are required. Since the streams of solutions coming out of the theta tube are continuous, the culture dish is exposed to relatively high concentrations of GABA during the patch recording process. This can cause the phenomenon known as desensitization to occur. When the receptors are in a desensitized state, they no longer 69 pass current with the application of agonist. Functionally, this makes repeated recordings from the same culture dish difficult to obtain. As such, even after many attempts, we were able to acquire only 7 ultrafast ligand exchange recordings. Of these, 4 had current responses that were too noisy to be useful because of low currents (either from desensitization or low receptor density), leaving only 3 analyze able system responses. The system responses obtained are given in Figure 11 with their peak currents normalized, labeled A, B, and C. The 10% to 900/0 rise-times of A, B and C were: 900 Ils, 500 Ils and 700 J.lS respectively. The current decay phases were best fit with two exponentials, reported as the time constants ('tshon and 'tlong) with their respective relative initial amplitudes (Amp.). The presence of two exponentials suggests two different open states of the channel. When modeling the channel's activity (see Figure 12) the't values from the curve fits are proportional to the opening rates for each open state, ~l and ~2 , over K-2 , the rate of the transition from the fully GABA bound closed state to the singly GABA bound closed state. Thus, assuming ~l is the opening rate of the longer lived open state, and ~2 is the opening rate of the shorter lived closed state, 't1ong is proportional to ~/K-2 and 'tshon is proportional to ~21K -2' The computed 't values are given in Table 2, along with the peak currents for each trace, and the RT -PCR results for each neuron. The curve fits are shown superimposed on the inverted decay traces in Figure 13. A B c 20 ms Figure 11. Normalized traces of ultrafast ligand exchange patch clamp recordings. Recordings are labeled as A, B, and C. Recordings were obtained from patches pulled from fetal mouse cortical neurons, with 1 ms applications of 10 mM GABA. 70 G G C~ C~ K- K- 1 2 Figure 12. A kinetic model of the GABA-A receptor with two open states (0) and three closed states (C). GABA binding steps are shown as G. Table 2 Functional and molecular properties of neuronal GABA-A receptors Neuron Rise-time 'tshort Amp. 'tlon~ Amp. Peak RT-PCR A 900 ~tS 24.10 ms 0.663 94.23 ms 0.320 42pA CLl~CL3Y2S B 500 Jls 8.78 ms 0.458 64.60 ms 0.460 37 pA PI C 700 Jls 12.26 ms 0.536 60.77 ms 0.472 193 pA ~Pl 71 20 ms Figure 13. Superimposed figures of the decay phases and the obtained curve fits using two exponentials (t]ong and'tshort)' Note that the traces have been flipped vertically (as a requirement of the curve fitting software used). 72 73 Sinl:le Cell RT -PCR This technique is described in detail in Chapter 2 and Chapter 3. Single cell RT -PCR analysis for the GABA-A (l1-3' ~1-3' and "12 receptor subunits was performed on all 3 neurons giving useable recordings. The gels containing the PCR products are shown in Figures 14, 15, and 16. As stated previously, the PCR results are listed combined with the system response parameters in Table 2. Discussion The combination of ultrafast ligand exchange patch clamp recording with single cell RT-PCR is a very difficult process. The two techniques are technically challenging on their own, and combining them makes the procedure even more so. This makes it hard to get a large pool of data to analyze, making conclusions difficult to draw. However, there are certain observations that can be made, even from the limited data obtained. When system responses B and C are normalized and superimposed (Figure 17) the kinetic resemblance is apparent, and indeed their time constants and relative initial amplitudes are also very similar. The neurons giving system responses Band C also showed similar RT -PCR expression profiles. For B the neuron expressed detectable levels of ~l' while C expressed detectable levels of ~ and ~l' In Chapter 3 it was shown that InRNA detection probability is related to the peak currents. It is interesting to note that the peak current of trace B is significantly smaller than that for trace C, suggesting that there may also be ~ expressed in the neuron giving trace B, albeit at Figure 14. RT -PCR results of neuron A. Lanes are labeled at top. The far left lane is a 100 bp marker, 200 bp is the bottom band. Image is the inverse of an ethidium bromide stained 2% agarose gel. Band sizes are (L to R): a l, 302; a2, 145; a3, 433; no ~ bands are present; 128,218 bp. levels below the limit of detection for our RT-PCR protocol. In contrast, Figure 18 shows the superimposition of traces A and B, where the 74 longer time constants of A (see Table 2) are apparent, showing the functional diversity between the two responses. The neuron giving trace A also showed a much different RT-PCR expression profile than the other two neurons: a l , ~,a3' and 12S (the short form of 12)' This is in agreement with the hypothesis that different subunit expression profiles (leading to different molecular properties) give different functional results. In conclusion, it should be noted that while ultrafast ligand exchange recordings and single cell RT -PCR are compatible techniques for analyzing the same neuron for its Figure 15. RT-PCR results of neuron B. Lanes are lebeled at the top. Far right lane is a 100 bp marker, bottom band is 200 bp. The only visible product is ~l' 214 bp. Image is the inverse on an ethidium bromide stained 2% agarose gel. 75 molecular and functional properties, it can be very difficult to successfully perform the two together in practice. However, the results obtained this far suggest that molecular composition, as determined from subunit mRNA expression, is related to receptor functional properties. Experimental Procedures Neuronal Cultures Neurons were dissociated from fetal mouse (embryonic days 15-17) cortex and maintained in minimum essential media (MEM) at 5% CO2 with 10% horse serum (Hyclone, Logan UT), glutamine (Sigma), and glucose (Sigma) added. At days 4 to 6, Figure 16. RT-PCR results of neuron C. Lanes are lebeled at the top. Far right lane is a 100 bp marker, bottom band is 100 bp. Image is inverse of an ethidium bromide stained 2% agarose gel. Products visible are (L to R): ~, 145 (light band); ~l' 214 bp. cultures were treated for 24 hours with 10 nM cytosine arabinoside (Sigma) to inhibit 76 growth of nonneuronal cells. Recordings were performed after 12 to 15 days in culture. Ultrafast Uiand Exchan~ Patch ClamP Recordini The theta tube is constructed by pulling double lumen, thin septum glass (R&D Scientific Glass) on a P-3 horizontal puller (Narishige) and sanded to give a tip size of approximately 100 J..U1l. The theta tube is attached to a LSS-3100 piezo-electric transducer (Burleigh Instruments). The transducer was controlled and traces were recorded by pClamp 6.0 software (Axon Instruments). Pulse duration (i.e. GABA application duration) was 1 ms, and sets of 10 pulses were performed with 1.5 s intervals between pulses. Recording electrodes were pulled from soda lime glass Figure 17. Superimposed images of the decay phases from neurons Band C. Notice the kinetic similarity in the two traces. Figure 18. Superimposed images of the decay phases from neurons A and B. Notice the much longer time constants of A are apparent in comparison to B, separating the two traces for both the initial decay phase (bottom left, 'tshort) and the tail (top, 'tlonJ. 77 78 microhematocrit capillary tubes (Baxter) using a P-87 micropipette puller (Sutter Instruments) to a resistance of approximately 2.5 MO in the recording solutions. External solution consisted of (in mM) 142 NaCI, 1.5 KCI, 1 CaCI2, 1 MgCI2, 10 glucose, 30 sucrose, 10 Na-HEPES at pH 7.4, and adjusted to 320 mOsm with sucrose. Internal solution consisted of (in mM) 140 CsCI, 4 MgCI2, 10 Na-HEPES, 5 EGTA at pH 7.4 and adjusted to 290 mOsm with sucrose. Both internal and external were made from nonpyrogenic sterile water for irrigation USP (Baxter). Solutions containing 10 mM GABA were diluted from 1 M GABA stock in distilled water (Sigma). Sin.:le Cell RT -PCR After recordings were performed, cells were aspirated into an approximately 1 Mil electrode containing 2 fll of sterile water (Baxter) using gentle suction from an attached 5 cc syringe. The contents of the aspiration electrode were expelled into RnaselDNase free 0.2 ml thin wall tubes (Life Science Products, Inc., Denver CO.) Containing 8 fll of the RT reaction mix. The RT mix consisted of (in mM unless otherwise stated) 50 KCI, 10 Tris-HCI pH 8.3, 5 MgCI2, 1 DTT,4 dNTP mix (Gibco), 2.5 f..1M random hexamers (Gibco), 0.33 U/fll Rnase inhibitor (5 Prime-3 Prime Inc., Boulder CO.), and 10 U/fll Superscript™ IT reverse transcriptase (Gibco) in sterile water (Baxter). RT reactions were held at 25°C for 10 minutes, 37°C for 60 minutes, 95°C for 5 minutes in a PTC-l00 thermal cycler (MJ Research Inc., Watertown MA), and then stored at 4°C until running PCR. A general PCR reaction to amplify GABA-A receptor 0;1-3, ~1-3 and Y2 subunits 79 present was performed by adding the (l, ~, and 1 general upstream and downstream primers together in a single multiplex PCR reaction. See Table 1, Chapter 2 for primer sequences. The entire 10 Jll RT reaction mix above was added to the PCR reaction mix, consisting of (in mM unless otherwise stated) 50 KCI, 10 Tris-HCI pH 8.3, 3.75 MgCI2, ImM dNTP mix (Gibco), 0.5 JlM each primer, and 0.05 V/J.1L Taq ploymerase (Fisher) in sterile water (Baxter). The PCR reactions were held at 95°C for 2.5 minutes, then cycled 25 times. Each cycle consisted of 95°C for 10 seconds, 57°C for 25 seconds, and 72°C for 35 seconds. One Jll samples of the general PCR reaction were analyzed in 25 Jll specific nested PCR reactions for the (l1-3 ~1-3 and 12 GABA-A receptor subunits. The specific reactions were held at 95°C for 2.5 minutes then cycled 40 times as follows: 95°C for 10 seconds, 60°C for 15 seconds, 72°C for 20 seconds. Subsequently, 4.0 Jll samples were analyzed on 2% agarose gels using ethidium bromide and V.V. illumination. Exponential Curve Fittini Analysis Curve fitting of decay curves was performed using SigmaPlot@ software version 2.0 (Jandel Scientific) using the standard two exponential transform. References Akaike, N., Inoue, M., and Krishtal, O.A. (1986). "Concentration-clamp" study of 'Y­aminobutyric- acid-induced chloride current kinetics in frog sensory neurones. J. Physiol. 379, 171-185. 80 Brett, R.S., Dilger, J.P., Adams, P.R., and Lancaster, B. (1986). A method for the rapid exchange of solutions bathing excised membrane patches. Biophys. 1. 50, 987-992. Burt, D.R., and Kamatchi, G.L. (1991). GABAA receptor subtypes: from pharmacology to molecular biology. FASEB J. 5,2916-2923. Calquhoun, D., Jonas, P., and Sakmann, B. (1992). Action of brief pulses of glutamate on AMP Alkainate receptors in patches from different neurones of rat hippocampal slices. J. Physiol. 458, 261-287. Franke, C., Hatt, H., and Dudel, J. (1987). Liquid filament switch for ultra-fast exchanges of solutions at excised patches of synaptic membrane of crayfish muscle. Neurosci. Lett. 77, 199-204. Lavoie, A.M. (1996) Determinants of kinetic function for recombinant and native 'Y­aminobutyric acid type A receptors using ultrafast ligand exchange. (SLC Utah: University of Utah) pp. 28-63. Lavoie, A.M., and Twyman, R.E. (1996). Direct evidence for diazepam modulation of GABAA receptor microscopic affinity. Neuropharmacology. 35, 1383-1392 Macdonald, R.L., and Olsen, R.W. (1994). GABAA receptor channels. Annu. Rev. Neurosci. 17, 569-602. Chen, C., and Okayama, H. (1987). High-efficiency transformation of mammalian cells by plasmid DNA. Mol. Cell. BioI. 7, 2745-2752. Maconochie, D.1., and Knight, D.E. (1989