LOCALIZATION AND FUNCTION OF ELECTROGENIC Na/BICARBONATE COTRANSPORTER NBCe1 IN RAT BRAIN DEBESHI MAJUMDAR

Transcription

1 LOCALIZATION AND FUNCTION OF ELECTROGENIC Na/BICARBONATE COTRANSPORTER NBCe1 IN RAT BRAIN by DEBESHI MAJUMDAR MARK O. BEVENSEE, COMMITTEE CHAIR DALE J. BENOS INYEONG CHOI LORI L. MCMAHON HARALD W. SONTHEIMER A DISSERTATION Submitted to the graduate faculty of The University of Alabama at Birmingham, in partial fulfillment of the requirements for the degree of Doctor of Philosophy BIRMINGHAM, ALABAMA 2009

2 LOCALIZATION AND FUNCTION OF ELECTROGENIC Na/BICARBONATE COTRANSPORTER NBCe1 IN RAT BRAIN DEBESHI MAJUMDAR CELLULAR AND MOLECULAR PHYSIOLOGY ABSTRACT Na-Coupled Bicarbonate Transporters (NCBTs) are members of the bicarbonate transporter superfamily that play important roles in regulating intracellular ph (ph i ) and extracellular ph (ph o ) in the central nervous system. Electrogenic Na/bicarbonate Cotransporter 1 (NBCe1) is an NCBT that is expressed in different mammalian tissues including the brain. NBCe1 has three splice variants NBCe1-A, -B and -C that differ in the amino and carboxy termini. We have first performed a systematic characterization of the localization profiles of the three NBCe1 splice variants at mrna and protein levels in rat brain. In these studies, we have used anti-sense probes and C-terminal antibodies that distinguish between the different NBCe1 splice variants. Using in situ hybridization technique, we have found that NBCe1-A is absent in rat brain whereas NBCe1-B and NBCe1-C are the brain-specific variants. Our results from immunohistochemistry were confirmed by those from immunoelectron microscopy demonstrating intracellular expression of NBCe1-A/B in neurons, and plasma membrane expression of NBCe1-C in glial cells more so than the surrounding neurons. Therefore, NBCe1 splice variants have different expression profiles in rat brain. Next, we examined whether the electrogenic NBC expressed in neurons is functional. We performed electrophysiology experiments in cultured rat hippocampal neurons bathed in either 5% CO 2 /24mM HCO 3 or 20% CO 2 /96mM HCO 3 using two different assays in whole-cell and perforated-patch modes. Based on the results from both assays, we found a population of rat hippocampal neurons ii

3 that exhibited functional electrogenic NBC activity. At the immunohistochemistry level, using C-terminal antibodies to NBCe1-A/B and -C we found that all neurons in culture expressed NBCe1-A/B and -C variants. Also, performing double labeling studies on neurons using antibodies to NBCe1-A/B and -C and an antibody to glutamic acid decarboxylase 67 (marker of inhibitory neurons), we found that NBCe1-A/B and -C were expressed in both excitatory and inhibitory neurons. Therefore, the heterogeneity in NBC function was not due to differences in NBCe1 expression. Keywords: expression, excitatory, functional, nervous system, protein, variants iii

4 DEDICATION I dedicate this dissertation to my parents and my husband. My parents, Subrata and Gopa Majumdar, who have shown me the light of this world, have always been my source of inspiration and support. Their affection and guidance have motivated me to persevere through the most challenging and difficult situations. Even though my parents are geographically far away from me, I could never feel their absence for a single day. My husband and best friend, Susanta has gifted me a world of happiness. He has been my source of constant encouragement and motivation, and has given me the love and emotional support to pull through during the happy as well as the difficult times of both life and research. It is true that, without my parents and my husband, I would never be able to accomplish my goals. iv

5 ACKNOWLEDGEMENTS First and foremost, I would like to thank God for always being with me. His blessings have helped me succeed in the difficult times of my life. I thank my mentor Dr. Mark O. Bevensee for guiding me in each and every step of this training process. He has been an excellent mentor, extremely supportive and helpful all through, and has taught me how to become an independent, conscientious, and thoughtful scientist. I thank all my committee members Dr. Mark O. Bevensee, Dr. Dale J. Benos, Dr. Inyeong Choi, Dr. Lori L. McMahon, and Dr. Harald W. Sontheimer for their extremely valuable suggestions, guidance and support all through these years. I would also like to thank Dr. Dale J. Benos, Department Chair, and Dr. Peter R. Smith, Graduate Program Director, for their continued support all through my training. I also extend my thanks to Dr. Lisa M. Schwiebert, the former Graduate Program Director, for her support during the initial years of my training. The Bevensee laboratory has been a wonderful place to work in. I thank both present and past members of the laboratory who have helped me in some way or the other during my graduate training. I wish to thank Dr. Carmel M. McNicholas, Dr. Jianping Wu and Ms. Jennifer B. Williams for their help in training me in the Bevensee laboratory. I would like to acknowledge the love and support of all my friends and relatives both here in the USA and in India. Finally, I would like to thank my grandparents in India who are always in my heart; their unconditioned love and support have made my dreams come true. v

6 TABLE OF CONTENTS Page ABSTRACT... ii DEDICATION... iv ACKNOWLEDGEMENTS...v LIST OF TABLES... viii LIST OF FIGURES... ix INTRODUCTION...1 Overview of ph Regulation and NCBT Activity...2 General Importance of ph Regulation in the Nervous System...2 Cellular Physiology and Energetics of Acid-base Transporters...5 Molecular Physiology of NCBTs...6 Functional Evidence for NCBTs in the Nervous System...11 Neurons...11 Glia...13 Epithelial Cells of Choroid Plexus...15 Localization of NCBTs in the Nervous System...17 Electrogenic Na/HCO 3 Cotransporter, NBCe Electrogenic Na/HCO 3 Cotransporter, NBCe Electroneutral Na/HCO 3 Cotransporter, NBCn Electroneutral Na/HCO 3 Cotransporter, NBCn2/NCBE...21 Electroneutral Na-driven Cl-HCO 3 Exchanger, NDCBE...24 Role of NCBTs in Nervous System Function...25 General Neuronal Excitability...25 Somatosensory Function...28 Cerebrospinal Fluid Secretion...30 Role of NCBTs in Energy-deficient Neurologic Conditions...32 Ischemia...32 Hypoxia...33 Goals of the Dissertation...35 Chapter 2 Overview...36 Chapter 3 Overview...37 Chapter 4 Overview...37 vi

7 TABLE OF CONTENTS (CONTINUED) Page LOCALIZATION OF ELECTROGENIC Na/BICARBONATE COTRANSPORTER NBCe1 VARIANTS IN RAT BRAIN...44 EVIDENCE FOR A FUNCTIONAL ELECTROGENIC Na/BICARBONATE COTRANSPORTER IN A POPULATION OF NEURONS CULTURED FROM RAT HIPPOCAMPUS...88 DISCUSSION Different Localization Profiles of NBCe1 Splice Variants in Rat Brain NBCe1 Expression in Neurons and Glia Significance of the Expression of Different NBCe1 Splice Variants in Brain Functional Electrogenic NBC in a Population of Rat Hippocampal Neurons Significance of an Electrogenic NBC in Neurons Future Directions Electrophysiology Experiments ph Experiments GENERAL LIST OF REFERENCES APPENDIX: INSTITUTIONAL ANIMAL CARE AND USE APPROVAL FORM vii

8 LIST OF TABLES Table Page INTRODUCTION 1 Molecular Characterization and Localization of Na-coupled Bicarbonate Transporters (NCBTs) in Tissues...38 viii

9 LIST OF FIGURES Figure Page LOCALIZATION OF ELECTROGENIC Na/BICARBONATE COTRANSPORTER NBCe1 VARIANTS IN RAT BRAIN 1 In situ Hybridization Probes and Rabbit Polyclonal Antibodies for Localizing NBCe1-A, -B, and -C Variants in Rat Brain Localization of mrna Encoding NBCe1 Variants in Rat Brain Expression of NBCe1 Variants in Tissue Homogenates from Different Regions of Rat Brain NBCe1-A/B Expression in Rat Cerebellum NBCe1-C Expression in Rat Cerebellum NBCe1-A/B Expression in Rat Hippocampus NBCe1-C Expression in Rat Hippocampus NBCe1-A/B and -C Expression in the Rat Cortex NBCe1-C and -A/B Expression in the Purkinje Layer of Rat Cerebellum Supplemental Figure1. NBCe1-A/B Expression in Rat Brain Perfusion Fixed with 4% PFA Supplemental Figure 2. NBCe1-C Expression in Rat Brain Perfusion Fixed with 4% PFA...86 ix

10 LIST OF FIGURES (CONTINUED) Figure Page EVIDENCE FOR A FUNCTIONAL ELECTROGENIC Na/BICARBONATE COTRANSPORTER IN A POPULATION OF NEURONS CULTURED FROM RAT HIPPOCAMPUS 1 HCO 3 -induced Outward Current Assay: Absence of a Functional Electrogenic NBC in a Cultured Hippocampal Neuron Voltage-clamped at -10 mv HCO 3 -induced Outward Current Assay: Presence of a Functional Electrogenic NBC in a Cultured Hippocampal Neuron Voltage-clamped at -10 mv HCO 3 -induced Outward Current Assay: Summary Data from Fig.1- and Fig.2- type Experiments HCO 3 -induced Outward Current Assay: Presence of a Functional Electrogenic NBC in a Cultured Hippocampal Neuron Voltage-clamped at -10 mv HCO 3 -induced Outward Current Assay: DIDS-sensitive, HCO 3 -induced Outward Current from a Cultured Hippocampal Neuron Voltage-clamped at -10 mv HCO 3 -induced Outward Current Assay: Summary Data from Fig.4- and Fig.5- type Experiments Na Removal-induced Inward Current Assay: Absence of a Functional Electrogenic NBC in a Cultured Hippocampal Neuron Voltage-clamped at -80 mv Na Removal-induced Inward Current Assay: Presence of a Functional Electrogenic NBC in a Cultured Hippocampal Neuron Voltage-clamped at -80 mv NBCe1-A/B, -C, and GAD67 Expression in Cultured Hippocampal Neurons x

11 LIST OF FIGURES (CONTINUED) Figure Page DISCUSSION 1 Significance of an Electrogenic NBC that Functions as an Acid Extruder in a Neuron xi

12 1 INTRODUCTION 1 The regulation of intracellular ph (ph i ) is critical for proper cell function because many cellular processes are sensitive to changes in ph (Roos and Boron, 1981; Bevensee and Boron, 2007). ph-sensitive processes include fertilization (Johnson and Epel, 1976), cell coupling (O'Beirne et al., 1987; Roos and Boron, 1981), alterations in cell structure (Parton et al., 1991), and the activity of metabolic enzymes such as phosphofructokinase (Trivedi and Danforth, 1966). In the central nervous system (CNS), which is comprised of cells with high surface-to-volume ratios and an extracellular space that is quite tortuous, the movement of acid-base equivalents across plasma membranes (e.g., by transporter-mediated ph i regulation and passive diffusion) not only changes ph i, but also has the additional consequence of changing extracellular ph (ph o ) in the opposite direction. Changes in ph i and/or ph o have important functional consequences in the CNS, particularly because the activity of many voltage-sensitive and ligand-gated ion channels associated with neuronal activity are sensitive to changes in ph i /ph o. The regulation and homeostasis of ph i is controlled primarily by acid-base transporters localized at the plasma membrane. As described in more detail below, acid-base transporters are characterized as HCO 3 -dependent or -independent acid extruders or acid loaders. Many Na-Coupled Bicarbonate Transporters (NCBTs) including 1 Portions of this Introduction have been adapted from a review that is in preparation for submission to Neuroscience: Majumdar D and Bevensee MO: Na-Coupled Bicarbonate Transporters (NCBTs) in the nervous system: function, localization and relevance to neurologic function.

13 2 electrogenic and electroneutral Na/Bicarbonate Cotransporters (NBCs) and Na-Driven Cl-Bicarbonate Exchangers (NDCBEs) are dominant contributors to the ph physiology of cells, especially those in the CNS. NCBTs in conjunction with anion exchangers (AEs) comprise the bicarbonate transporter (BT) superfamily the members of which are encoded by SoLute Carrier 4 (SLC4) genes. Our understanding of the physiologic importance of NCBTs especially in the CNS has expanded tremendously over the past decade following the cloning of the first NCBT cdna encoding the electrogenic NBC from salamander kidney (Romero et al., 1997). Recent data from studies on various NCBT knockout mice have highlighted the importance of cation-coupled bicarbonate transporters in neurologic processes, including neuronal activity and somatosensory function. In this introductory chapter, I first provide a general overview of the importance of ph regulation, the energetics of proton movement and acid-base transporter activity, and the molecular physiology of the BT superfamily. Next, I present both historical and contemporary information on the function and localization of NCBTs in the CNS. Then, I discuss the latest advances in understanding the role of NCBTs in the functioning nervous system and in energy-deficient neurologic conditions. Finally, I present the goals of this dissertation, and how the results from my work expand our understanding of the expression and function of NCBTs in the CNS. Overview of ph Regulation and NCBT Activity General Importance of ph Regulation in the Nervous System Changes in ph can have a powerful influence on brain function, for example by altering neuronal activity and synaptic transmission. Although there are exceptions, an

14 3 increase in ph o generally stimulates neuronal firing, whereas a decrease has the opposite effect (Chesler and Kaila, 1992; Balestrino and Somjen, 1988; Ransom, 2000). In fact, an increase in ph can trigger epileptiform activity in both humans and rodents (Cohen and Kassirer, 1982; Aram and Lodge, 1987; Woodbury et al., 1984; Jarolimek et al., 1989; Lee et al., 1996; Marshall and Engberg, 1980; Velisek et al., 1994). A decrease in ph has the opposite effect. For example, lowering ph o from 7.4 to 6.7 reversibly reduced the amplitude, increased the interval, and slowed the propagation of low extracellular Mg 2+ - induced epileptiform discharges in combined entorhinal cortex-hippocampal slices (Velisek et al., 1994). Lowering ph o to 6.2 inhibited the discharges completely. ph o -induced increases in neuronal activity are often associated with increased activity of N-methyl-Daspartate (NMDA) receptors that display pronounced ph sensitivity in the physiologic range (Tang et al., 1990; Traynelis and Cull-Candy, 1990) with increased open probability at high ph. Perhaps it is not surprising that changes in ph can alter neuronal firing in light of the myriad of ligand- and voltage-gated ion channels sensitive to changes in ph i and/or ph o (Traynelis, 1998; Tombaugh and Somjen, 1998). Many of these voltage-gated channels generate and propagate action potentials. ph changes presumably alter ion-channel activity by modifying electrostatic interactions between charged amino acids and inducing structural alterations. In isolated rat CA1 hippocampal neurons, a moderate extracellular acidosis (ph 6.4) depresses the peak Na + current by ~15%, whereas alkalosis has the opposite effect (Tombaugh and Somjen, 1996). K + channels are more sensitive to changes in ph i than ph o in the physiologic range. In general, a decrease in ph i inhibits K +

15 4 channels, although there are exceptions such as the delayed K + channel in perfused snail neurons that is stimulated by a decrease in ph i (Byerly and Moody, 1986). Compared to Na + and K + channels, Ca 2+ channels are typically more sensitive to changes in ph. In neurons, an increase in ph o reversibly enhances both high voltageactivated Ca 2+ currents (e.g., from L-, N-, P-, Q-, and R-type Ca 2+ channels) (Tombaugh and Somjen, 1996) and low voltage-activated Ca 2+ currents (e.g., from the T-type Ca 2+ channel) (Tombaugh and Somjen, 1997). Regarding effects of changes in ph i, the activity of high voltage-activated Ca 2+ channels in chick dorsal root ganglion neurons is reversibly enhanced by applying a weak base (which raises ph i ) and depressed by applying a weak acid (which lowers ph i ) (Mironov and Lux, 1991). Similar results have been obtained on rat CA1 neurons (Tombaugh and Somjen, 1997). Channels in the CNS that are activated by changes in ph o can influence neurotransmission. For example, DeVries has presented evidence that transient acidification at the synaptic cleft can inhibit a presynaptic voltage-gated Ca 2+ current in retinal cone photoreceptors due to a ph-mediated shift in the voltage dependence of the Ca 2+ channel (DeVries, 2001). Furthermore, the regulation of ph o by horizontal cells may modulate or contribute to the feedback inhibition of these Ca 2+ channels in the cone photoreceptors (Vessey et al., 2005). A change in ph is the primary stimulus for activating some channels such as acid-sensing ion channels (ASICs), which are stimulated by a decrease in ph o (Waldmann et al., 1999). The importance of these ph-sensitive channels in neurotransmission is highlighted by studies on ASIC1 knockout mice that have impaired hippocampal long-term potentiation, as well as reduced excitatory postsynaptic potentials and NMDA receptor activation (Wemmie et al., 2002). Neurons cultured from the hippo-

16 5 campus of these ASIC1 knockout mice fail to elicit acid (ph 5)-evoked cation currents. The ASIC1-knockout mice also display defects in both performance on the Morris water maze test and eye-blink conditioning findings that implicate these channels in synaptic plasticity, learning, and memory. ph changes in the brain can also regulate neurotransmission by modulating the release and/or uptake of neurotransmitters and neuromodulators. In studies performed on isolated rat-brain synaptosomes, the depolarization-induced release of total glutamate was 15% less at ph o 6.0 than 7.4, although the uptake of glutamate was unaltered in the ph o range (Fedorovich et al., 2003). However, such ph effects are dependent on the neurotransmitter/neuromodulator. For instance, in the same study, uptake of the acetylcholine precursor choline displayed a ph o dependence, with uptake being 80% less at ph o 6.0 compared to 7.4. In a different study on superfused rat hypothalamic synaptosomes, both an extracellular and an intraterminal acidification stimulated 3 H-dopamine release through a Ca 2+ -dependent exocytotic process (Cannizzaro et al., 2003). Cellular Physiology and Energetics of Acid-base Transporters For a cell with a resting membrane potential of -60 mv, and bathed in a ph-7.3 solution, the electrochemical gradient favors the passive entry of protons (or conjugate weak acid) into the cell, and the passive exit of weak bases such as bicarbonate out of the cell, thereby leading to a decrease in ph i (towards the equilibrium ph of 6.3). The extent of such passive movement depends on the cell membrane s permeability to protons or the acid-base species. Thus, most cells are continually exposed to a chronic acid load that tends to lower ph i. Probably because so many important cellular processes are sensitive

17 6 to ph (as described above), cells have evolved a system of acid-base transporters to regulate ph i above the equilibrium ph. Acid-base transporters are classified as acid loaders or acid extruders. Acid loaders transport protons into or base equivalents out of cells, whereas acid extruders transport those ions in the opposite direction. Acid-base transporters are further classified as being HCO 3 -independent or -dependent, and for many cells, the HCO 3 -dependent transporters are the more powerful of the two classes in regulating ph i. NCBTs usually function as acid extruders by transporting HCO 3 into cells and increasing ph i. However, as described below, some NCBTs can transport HCO 3 out of cells (depending on the prevailing ion gradients and membrane potential) and therefore function as acid loaders. Every NCBT is Na + -dependent. However, as detailed in the next section, NCBTs represent a diverse family of transporters that are either electroneutral or electrogenic (with more than one Na:HCO 3 stoichiometry), chloride-independent or chloride-dependent, and stilbene-sensitive or -insensitive. Molecular Physiology of NCBTs We have made significant advances in understanding the importance of NCBTs in the functional nervous system with the cdna cloning, molecular characterization, and localization of these proteins. Below, we describe the general characteristic features of the different NCBTs and their splice variants. A more extensive review of the molecular nature of NCBTs is presented in Boron et al. (Boron et al., 2009).

18 7 NBCe1. Historically, the first cdna encoding a NCBT was identified by Romero et al. who expression cloned the electrogenic NBC (NBCe1) from the proximal tubule of the salamander kidney (Romero et al., 1997) a preparation in which the first NBC was functionally identified (Boron and Boulpaep, 1983). Cloned NBCe1 expressed in Xenopus oocytes elicits a DIDS-sensitive, Na-dependent increase in ph i following a CO 2 - induced acidification when oocytes are exposed to a CO 2 /HCO 3 solution. The electrogenicity of the transporter is evident by a hyperpolarization when oocytes are exposed to the HCO 3 solution, and a depolarization when external Na + is removed (Romero et al., 1997). NBCe1 was subsequently cloned and identified from other preparations (Table I). Three splice variants of NBCe1 (-A, -B, and -C) have been identified, and these variants differ only at their N and/or C termini. The importance of NBCe1 in physiologic functions is highlighted by results from studies on human patients with NBCe1 gene mutations and data from a more recent NBCe1 knockout mouse. At present, investigators have identified human patients with one of seven missense mutations, three non-sense mutations, and a frameshift mutation in the SLC4A4 gene. All these patients exhibit proximal renal tubule acidosis, and many of them also have ocular abnormalities, including glaucoma and/or cataract formation (Igarashi et al., 1999; Demirci et al., 2006; Dinour et al., 2004). Although neurological abnormalities are not a consistent finding in all patients, two of them are mentally retarded (Igarashi et al., 1999; Igarashi et al., 2001) and another one displays developmental disabilities (Demirci et al., 2006). These mutations can lead to loss of NBC-mediated HCO 3 transport by either decreasing activity of the transporter (Igarashi et al., 1999; Dinour et

19 8 al., 2004) or reducing expression at the plasma membrane (Toye et al., 2006; Li et al., 2005). Gawenis et al. developed an NBCe1 null-mutant mouse by disrupting the SLC4A4 gene. The NBCe1 null-mutant mice, which died early before weaning, had severe metabolic acidosis, abnormalities in dentition, growth retardation, splenomegaly, hyponatremia, and impaired HCO 3 secretion in the colon (Gawenis et al., 2007). The mice had very thin and transparent skulls due to defects in bone mineralization an effect likely due to the metabolic acidosis. Although the null-mutant mice did not exhibit any overt neurological defects, further neurologic studies may reveal alterations in learning and memory and susceptibility to seizures. Of course, the compensatory expression of other acid-base transporters in these knockout mice during development may have masked an important role of NBCe1s in the CNS. NBCe2. From human heart and testis, Ira Kurtz s group cloned cdnas encoding two variants of another member of the NBC family NBCe2 or NBC4 that exhibit sequence similarity to NBCe1 (Pushkin et al., 2000a; Pushkin et al., 2000b). Subsequently, four additional NBCe2 variants were identified. The six variants (NBCe2-A to -F) differ at their C termini and/or last third of the transmembrane domain region. The A through D variants, but not E and F, are reported in the human genome database. At present, only the C variant of the cloned NBCe2s has been functional characterized and determined to be electrogenic (Virkki et al., 2002). Similar to the aforementioned studies on NBCe1, NBCe2-C expressed in oocytes and studied with voltage- and ph-sensitive microelectrodes elicits a DIDS-sensitive, Na-dependent increase in ph i following a CO 2 -induced

20 9 acidification. The electrogenicity of the transporter is evident by a hyperpolarization when oocytes are exposed to the HCO 3 solution, a depolarization when external Na + is removed, and Na + - and HCO 3 -induced currents under voltage-clamp conditions. These electrogenic signals are blocked by DIDS. NBCe2 at the RNA and protein levels has been detected in rodent and human tissues (Table I). NBCn1. cdna encoding the transporter (NBCn1-A) was first cloned from human skeletal muscle (Pushkin et al., 1999a), and then three additional variants (NBCn1-B, -C, and -D) were cloned shortly thereafter from rat smooth muscle (Choi et al., 2000). When expressed in oocytes, NBCn1-A (originally named NBC3) stimulated Cl -independent, DIDS-insensitive 22 Na uptake (Pushkin et al., 1999a). In subsequent ph i experiments on NBCn1-expressing oocytes, removing external Na + in the presence of CO 2 /HCO 3 produced an acidification, while raising external K + to depolarize the membrane had no effect on ph i. Thus, the transporter appears to be electroneutral. The electroneutrality of NBCn1 was particularly evident from microelectrode studies on oocytes expressing NBCn1-B cloned from rat smooth muscle (Choi et al., 2000). NBCn1-B-expressing oocytes displayed a Na + and HCO 3 -dependent, yet Cl - independent ph i recovery following a CO 2 -induced acid load. This ph i recovery was only modestly inhibited by DIDS. While the transporter is electroneutral based on the absence of a CO 2 /HCO 3 -induced hyperpolarization, a Na + conductance is associated with the transporter. NBCn1 is ubiquitously expressed (Table I). In additional to the aforementioned four splice variants, four additional ones have been identified. These eight variants

21 10 (NBCn1-A to -H) have an identical transmembrane domain region, but different splice cassettes at their cytoplasmic N and/or C termini. Bok et al. have developed an NBCn1 knockout mouse that is blind and displays defects in hearing (Bok et al., 2003). Details of this study are presented in a subsequent section. NBCn2. cdna encoding the transporter was first cloned from an insulin-secreting cell line MIN6 cdna library and called a Na-driven Chloride/Bicarbonate Exchanger (NCBE) (Wang et al., 2000). Based on 22 Na +, 36 Cl, and ph i studies with oocytes or HEK293 cells heterologously expressing the transporter, the authors concluded that the transporter exchanged extracellular Na + and HCO 3 for intracellular Cl and H +. However, according to more recent studies with voltage- and ph-sensitive microelectrodes, the transporter expressed in oocytes is electroneutral and responsible for a DIDSsensitive, Na + - and HCO 3 -dependent, but Cl -independent ph i recovery following a CO 2 -induced acidification (Parker et al., 2008a). Using Cl -sensitive microelectrodes to measure intracellular Cl, as well as 36 Cl uptake and efflux assays, the authors determined that Cl flux across the membrane is not coupled to Na + and HCO 3 transport, but rather, is linked to CO 2 /HCO 3 -stimulated, but Na + -independent, Cl self exchange (Parker et al., 2008a). Four splice variants (NBCn2-A through -D) have been identified with different internal cytoplasmic N terminal splice cassettes and/or different cytoplasmic C termini. NBCn2 displays pronounced expression in brain (Table I), and the -B and -C variants were cloned from rat brain (Giffard et al., 2003).

22 11 Jacobs et al. developed an NBCn2 (NCBE) knockout mouse that displays a high seizure threshold and a reduced brain ventricular volume (Jacobs et al., 2008). Details of the study are described in a subsequent section. NDCBE. cdna encoding the transporter was cloned from human brain (Grichtchenko et al., 2001). In functional studies using voltage- and ph-sensitive microelectrodes and a 36 Cl efflux assay, NDCBE expressed in oocytes induced a DIDS-sensitive, Na + and HCO 3 -dependent ph i recovery from a CO 2 -induced acidification. Na + and HCO 3 efflux from oocytes required extracellular Cl. Also, CO 2 /HCO 3 -induced stimulation of the transporter elicited a DIDS-sensitive and Na + -dependent increase in 36 Cl efflux. Four variants of NDCBE have been identified that differ at the cytoplasmic N and/or C termini. Similar to NBCn2, NDCBE also exhibits pronounced expression in brain (Table I). Functional Evidence for NCBTs in the Nervous System Neurons The Na-driven Cl-HCO 3 exchanger was the first ph i -regulating transporter functionally identified and shown to be the major acid extruding mechanism in classic neuronal preparations including the squid axon (Boron and De Weer, 1976a; Boron and De Weer, 1976b; Russell and Boron, 1976; Russell and Boron, 1981; Boron and Russell, 1983) and the snail neuron (Thomas, 1976a; Thomas, 1976b; Thomas, 1977). Work on these two preparations was performed simultaneously by Walter Boron and Roger Thomas two pioneers in the field of ph i regulation.

23 12 In both preparations, the recovery of ph from an acid load required intracellular Cl (Russell and Boron, 1976; Thomas, 1977), intracellular ATP (Boron and Russell, 1983; Russell and Boron, 1976), external Na + (Boron and Russell, 1983; Thomas, 1977), and external HCO 3 (Boron and De Weer, 1976a; Thomas, 1977). The transporter was also inhibited by SITS (Russell and Boron, 1976; Boron and Russell, 1983; Thomas, 1977). Based on ph i measurements, as well as 22 Na + and 36 Cl efflux/influx data, the transporter moves 1 Na + and 2 HCO 3 (or equivalent species) in one direction and 1 Cl in the opposite direction (Boron and Russell, 1983). The acid extrusion rate was ph i dependent and stimulated by a decrease in ph i (Boron and Russell, 1983). In the squid axon, intra-axonal ATP is required for a functional transporter, although not as an energy source (Boron et al., 1988) because the ATP effect can be mimicked by non-hydrolysable ATP S. The transporter s K M for ATP is 124 M (Davis et al., 2008). The Na-driven Cl-HCO 3 exchanger also appears to be the major HCO 3 - dependent acid extruder mammalian neurons. In CA1 neurons acutely isolated from the hippocampus of neonatal rats, applying CO 2 /HCO 3 elicited a ph i recovery from the initial CO 2 -induced acidification that was Na + -dependent and DIDS-sensitive (Schwiening and Boron, 1994). Cl dependence of the transporter was evident by a progressive decrease in the rate of the HCO 3 -induced ph i increase when the neurons were exposed several times to the CO 2 /HCO 3 solution in the continued absence of external Cl. In other studies, results from similar ph i experiments implicated the presence of a Na-driven Cl-HCO 3 exchanger in other mammalian neurons (Bevensee and Boron, 1998). According to conventional wisdom, neurons possess electroneutral NCBTs (see below). However, as described in more detail in Chapters 2 and 3 of this dissertation,

24 13 electrogenic NBC expression and function is also evident in neurons. Similar to electrogenic NBC, there is no functional evidence of NBCn1 or NBCn2 in any neuronal preparation. As discussed in the localization section below, using immunohistochemical techniques, Cooper et al. reported the expression of NBCn1 in the synapses and somatodendrites in embryonic rat hippocampal neurons (Cooper et al., 2005). They proposed that the possible role of NBCn1 is to regulate ph i at the synapses and therefore modulate neuronal activity. Glia Historically, Boron and Boulpaep working on perfused proximal tubules from the salamander kidney functionally identified an electrogenic NBC as the first cation-coupled HCO 3 transporter (Boron and Boulpaep, 1983). The basolateral transporter was responsible for a SITS-sensitive decrease in ph i and Na + i, as well as depolarization of the basolateral membrane upon lowering basolateral ph or HCO 3. Removing Cl had no effect on these changes in ph i or basolateral membrane potential. The electrogenicity of the transporter was particularly evident from a now-recognized hallmark assay lowering/removing external Na + elicited a simultaneous depolarization and decrease in ph i. This transporter in basolateral membrane vesicles from rabbit kidney cortex has a 1:3 Na + :HCO 3 stoichiometry based on a thermodynamic analysis (Soleimani et al., 1987). In the CNS, an electrogenic NBC was first characterized by Deitmer and Schlue working on the giant neuropile glial cells of the leech (Deitmer and Schlue, 1987; Deitmer and Schlue, 1989; Deitmer, 1992). The transporter was responsible for a CO 2 /HCO 3 -induced alkalization and hyperpolarization that was DIDS-sensitive and Na + -dependent

25 14 (Deitmer and Schlue, 1989). The HCO 3 -induced alkalization was not affected by either removing extracellular Cl or depleting the cells of intracellular Cl. The HCO 3 solution also elicited an expected increase in intracellular Na +. The electrogenic nature of the transporter was also evident from a simultaneous depolarization and HCO 3 -dependent ph i decrease elicited by removing external Na + (Deitmer, 1992). In leech glial cells, the transporter appears to have a stoichiometry of 1:2 Na:HCO 3 (Deitmer and Schlue, 1989; Deitmer, 1991). Using similar approaches, the transporter was subsequently identified and/or studied in other glial cell preparations including connective glial cells from leech (Szatkowski and Schlue, 1992), astrocytes from the Necturus optic nerve (Astion and Orkand, 1988), and mammalian astrocytes cultured from rat forebrain (Boyarsky et al., 1993), hippocampus (Bevensee et al., 1997a; Bevensee et al., 1997b; Pappas and Ransom, 1994), cerebellum (Brune et al., 1994) and cortex (Shrode and Putnam, 1994). In at least rat hippocampal astrocytes, the mammalian electrogenic NBC appears to have a 1:2 Na + :HCO 3 stoichiometry (Bevensee et al., 1997a, Bevensee et al., 1997b). Electrogenic NBC activity has also been functionally characterized in mammalian brain slices. In gliotic hippocampal slices, electrogenic NBC activity contributes to a depolarization-induced alkalization and an extracellular acid shift (Grichtchenko and Chesler, 1994a; Grichtchenko and Chesler, 1994b). Applying Ba 2+ or raising extracellular K + produces a Na + -dependent intracellular alkalization and an extracellular acidification that are both enhanced in the presence of CO 2 /HCO 3. In contrast to results from in vitro studies on cultured astrocytes, the in situ NBC is not sensitive to stilbene derivatives.

26 15 A few groups have reported evidence for Na-driven Cl-HCO 3 exchanger activity in astrocytes, including those cultured from rat brain (Mellergard et al., 1993; Shrode and Putnam, 1994). For example, in rat astrocyte and C6 glioma cell cultures, exposing cells to a CO 2 /HCO 3 solution elicited a ph i recovery following the initial CO 2 -induced acidification. This HCO 3 -induced ph i recovery was dependent on external Na +, blocked by DIDS, and inhibited by incubating the cells for 2 h in the absence of external Cl (Shrode and Putnam, 1994). There is evidence for a Na + - and HCO 3 -dependent, and perhaps Cl - independent transporter in cultured oligodendrocytes from embryonic mouse spinal cord (Kettenmann and Schlue, 1988) and adult rat cerebellum (Boussouf et al., 1997). However, the electrogenicity/electroneutrality of an oligodendrocyte NBC has yet to be determined. Support for a functional electrogenic NBC in oligodendrocytes comes from immunolocalization data in which an NBCe1-A/B antibody labels dendrites of oligodendrocytes cultured from mouse and rat (Ro and Carson, 2004). Epithelial Cells of Choroid Plexus As reviewed in more detail by McAlear and Bevensee (McAlear and Bevensee, 2003), according to early studies on ph-regulating transporters in the choroid epithelium, a Na-H exchanger and an anion-exchanger (AE) on the basolateral membrane in conjunction with a possible NBC or HCO 3 -conducting anion channel on the apical membrane contribute to both the ionic composition and ph of the cerebrospinal fluid (CSF). According to more recent functional data, multiple Na-coupled HCO 3 transporters appear to be involved.

27 16 Using the ph-sensitive dye BCECF, Bouzinova et al. measured ph i of epithelial cells from the rat choroid plexus to characterize HCO 3 transporters involved in both the ph i recovery from an acid load and the ph i increase elicited by applying CO 2 /HCO 3 at resting ph i in the nominal absence of the physiologic buffer (Bouzinova et al., 2005). In both assays, the HCO 3 -induced ph i increase was Na + dependent, but only the ph i increase elicited by applying CO 2 /HCO 3 at resting ph i was inhibited by DIDS. Cl -free media had no effect on the ph i recoveries. In light of additional immunologic evidence presented that the cells express NBCe2, as well as previous studies demonstrating the expression of NBCn1 and NBCn2/NCBE (see below), the authors conclude that rat choroid plexus epithelial cells have both a DIDS-insensitive NCBT (probably NBCn1) and a DIDS-sensitive NCBT (probably either NBCe2 and/or NBCn2/NCBE). In more recent studies, Millar and Brown measured small DIDS-sensitive, Na + - dependent, HCO 3 -induced outward currents from I-V plots obtained from mouse choroid plexus epithelial cells (Millar and Brown, 2008). From the reversal potentials of these I-V plots, the authors concluded that the NBC which may be NBCe2 according to the previous discussion has a 1:3 Na + :HCO 3 stoichiometry. With the membrane potential of the choroid plexus epithelial cell estimated to be -35 mv to -60 mv, this NBC would be expected to mediate HCO 3 secretion into the CSF given the Na + and HCO 3 gradients across the apical membrane.

28 17 Localization of NCBTs in the Nervous System Electrogenic Na/HCO 3 Cotransporter, NBCe1 NBCe1 was the first cloned NCBT localized in the CNS not surprising because it was the first member of the bicarbonate transporter superfamily cloned and subsequently used to generate molecular probes including PCR primers, crna probes, and antibodies. Schmitt et al. and Giffard et al. performed the early NBCe1 localization studies on rat brain using antibodies and polynucleotide probes nearly all of which did not distinguish among the NBCe1 splice variants (Schmitt et al., 2000; Giffard et al., 2000). Both groups found wide-spread expression of NBCe1 throughout the brain in a pattern consistent with glial-cell expression. According to double-labeling studies using an antibody to the glial fibrillary acidic protein (GFAP) and an in situ hybridization probe to NBCe1, the transporter was evident in astrocytes and Bergmann glia in the cerebellum (Giffard et al., 2000). Schmitt et al. performed in situ hybridization co-localization studies with probes to the astrocytic glutamate transporter 1 (GLT1) and NBCe1 (Schmitt et al., 2000). The authors observed NBCe1 mrna expression in the astrocytes of the cortex, dentate gyrus and brainstem (Schmitt et al., 2000). In double-labeling studies using antibodies to GFAP and NBCe1, the authors found expression of NBCe1 in astrocytes in hippocampus and cerebellum (Schmitt et al., 2000). NBCe1 mrna was also evident in neurons of the piriform and entorrhinal cortex, cerebellum, olfactory bulb, dentate gyrus and striatum. Using an antibody to the microtubule associated protein 2 (MAP2), a neuronal marker, Schmitt et al. found NBCe1 expression in Purkinje neurons of the cerebellum, as well as neurons in the dentate gyrus, pyramidal and molecular cell layers, and stratum oriens of the hippocampus.

29 18 NBCe1 has also been identified in human brain. Using human NBCe1 primers that did not discriminate among the NBCe1 splice variants, Damkier et al. found NBCe1 mrna in human cerebellum, cerebrum, hippocampus, and choroid plexus (Damkier et al., 2007). As mentioned earlier, NBCe1 has three splice variants. At the time I was exploring the localization of the three variants in rat brain (Chapter 2), Rickmann et al. reported on the expression of NBCe1-A and B in mouse brain using immunohistochemistry and immunoelectron microscopy (Rickmann et al., 2007). The authors found expression of NBCe1-A throughout brain, particularly in neuronal populations in the cerebellum, hippocampus, cerebral cortex, and olfactory bulb. In contrast, they reported the expression of NBCe1-B in astroglia and Bergmann glia. However, there are a few concerns with this study. First, the cdna encoding the A variant of NBCe1 has not been identified in a homology cloning study of NBCe1 from rodent brain (Bevensee et al., 2000). Second, Giffard et al. reported preliminary data that an in situ hybridization probe specific to the A variant did not appreciably label cells in rat brain. Finally, the antibody used by Rickmann et al. to localize NBCe1-B also recognized the C variant. As described in Chapter 2, I have performed a detailed localization study of all three NBCe1 splice variants in rat brain at both the mrna and protein levels. NBCe1 mrna and protein levels are developmentally regulated. In an in situ hybridization study on rat brain, Giffard et al. first detected NBC mrna at embryonic day- 17 (E17) in the spinal cord (Giffard et al., 2000). In rat forebrain, NBC mrna was first evident at postnatal day-0 (P0). The authors found that NBC mrna persisted into adulthood and the level was highest at P15. Using immunoblotting techniques with a nonspe-

30 19 cific NBCe1 antibody, Douglas et al. found that NBCe1 expression increases gradually from E16 to P105 in the cerebral cortex, cerebellum, and brain-stem diencephalon, but to a much lesser degree in the cortex (Douglas et al., 2001). Electrogenic Na/HCO 3 Cotransporter, NBCe2 NBCe2 mrna in brain has been localized to rat and mouse choroid plexus epithelial cells (Praetorius et al., 2004b). In immunofluorescence studies with an antibody to the N-terminal 22 residues of NBCe2, Bouzinova et al. identified expression in the apical membrane of mouse and rat choroid plexus (Bouzinova et al., 2005). Such apical expression was reinforced by results from electron microscopy studies with immunogold labeling of the apical microvillar projections from the mouse choroid plexus. Although the specific variant of NBCe2 expressed in choroid plexus is not known, it is likely to be NBCe2-C (Virkki et al., 2002) NBCe2 is also present in human brain. Using RT-PCR and human NBCe2 primers that are expected to amplify all splice variants of NBCe2 cdna, Damkier et al. found mrna expression in human cerebellum, cerebrum, hippocampus, and choroid plexus (Damkier et al., 2007). Electroneutral Na/HCO 3 Cotransporter, NBCn1 In a preliminary study, Risso Bradley et al. used immunohistochemistry and found NBCn1 expression in nerve fibers of the hippocampal slice (Risso Bradley et al., 2001). In a more recent study, Cooper et al. used RT-PCR techniques and found NBCn1- B in hippocampal neurons cultured from embryonic rats and NBCn1-E in neurons cul-

31 20 tured from adult rat hippocampal slices (Cooper et al., 2005). According to single-cell RT-PCR data with primers to NBCn1, glutamic acid decarboxylase 65 (GAD65), and GAD67, the authors found that NBCn1-B was present in both excitatory (GAD-negative) and inhibitory (GAD-positive) embryonic neurons. In immunocytochemistry studies, the authors found expression of NBCn1 at the plasma membrane of both the soma and dendrites of embryonic neurons. According to double-labeling antibody studies, NBCn1 partially colocalized with post synaptic density protein-95 (PSD-95) at excitatory synapses. Boedtkjer et al. used an antibody-independent technique to perform a more systemic analysis of NBCn1 localization in mouse brain (Boedtkjer et al., 2008). This immunoreactive-independent technique involves generating a transgenic mouse with the LacZ gene under the control of the NBCn1 promoter. Staining of -galactosidase activity can be used as a marker for NBCn1 transcriptional activity. In brain, -galactosidase staining was seen in several regions including the cortex, hippocampus, cerebellum, and epithelial cells of the choroid plexus. Staining was evident in all the layers of the hippocampus including the pyramidal cells layers and the dentate gyrus. Staining was also evident in the dentate nucleus and cortical Purkinje cells of the cerebellum. NBCn1 mrna has also been detected in the cerebrum and cerebellum of rat, and the choroid plexus of both mouse and rat (Praetorius et al., 2004b). Based on immunohistochemistry, the authors reported NBCn1 expression in the basolateral membranes of the epithelial cells of both rat and mouse choroid plexus. NBCn1 is also present in human brain. Using RT-PCR and human NBCn1 primers that are expected to amplify all splice variants of NBCn1 cdna, Damkier et al. found

32 21 mrna expression in human cerebellum, cerebrum, hippocampus, and choroid plexus (Damkier et al., 2007). Chen et al. subsequently reported the developmental expression of NBCn1-B in cortex, hippocampus, subcortex, and cerebellum of P16, P30, P104, and P118 mice. NBCn1-B expression was highest in cortex and hippocampus of P16 mice (Chen et al., 2007). Postnatal changes in NBCn1-B expression were brain region dependent. For example, NBCn1-B expression with age gradually increased in the hippocampus, decreased in the cortex, and remained the same in the subcortex and cerebellum. Changes in NBCn1 expression may be required for proper development of the various brain regions, and may highlight developmental changes in the ph physiology of the regions. Electroneutral Na/HCO 3 Cotransporter, NBCn2/NCBE In addition to NBCe2 as describe above, NBCn2 is also present in the epithelial cells of the choroid plexus (Praetorius et al., 2004b). In RT-PCR studies, Praetorius et al. detected NBCn2 mrna in the cerebrum and cerebellum of rat, and the choroid plexus of both mouse and rat. In immunochemical studies on the choroid plexus, NBCn2-B protein was evident on the basolateral membrane of the epithelial cells. Giffard et al. cloned two NBCn2 variants from rat brain: rb1ncbe (NBCn2-B) and rb2ncbe (NBCn2-C). Subsequently, they reported developmental and regional mrna expression patterns of NBCn2 in rat and mouse brain (Giffard et al., 2003). NBCn2 mrna in rat brain was evident from Northern blot analysis. According to in situ hybridization data, NBCn2-B mrna expression was widespread in the brain and spinal cord of rats as early as E19, and such expression persisted until adulthood. At higher

33 22 magnifications, NBCn2-B mrna was detected in the Purkinje cells of the cerebellum and CA3 principal neurons of the hippocampus. According to results from RT-PCR studies, mrna expression of the C variant was greater than that of the B variant in mouse cultured astrocytes, whereas both variant mrnas expressed similarly in mouse cultured neurons. The authors report that the C variant compared to B variant mrna was greater in cortex, striatum, and hippocampus of mouse brain, while the reverse was true for these regions from rat brain. To examine the cellular expression profiles of the two variants in mouse brain, the authors transfected cultured mouse astrocytes with retroviral vectors encoded FLAG-tagged NBCn2 constructs and subsequently performed immunocytochemistry with anti-flag antibodies. Although both variants were expressed at the plasma membrane, the B variant was more localized to intracellular vesicles, whereas the C variant was more localized to the actin cytoskeleton. NBCn2 is also present in human brain. Using RT-PCR and human NBCn2 primers that are expected to recognize all splice variants of NBCn2 cdna, Damkier et al. found higher NBCn2 mrna levels in human cerebrum, hippocampus, and choroid plexus than in the cerebellum (Damkier et al., 2007). Using immunoblotting techniques and a polyclonal antibody that recognizes the N-terminal 135 amino acids of NBCn2, Chen et al. found abundant expression of NBCn2 in both rat and mouse hippocampus, cerebellum, cerebral cortex, and subcortex to a lesser extent (Chen et al., 2008c). Interestingly, NBCn2 expression was evident in neurons cultured from rat hippocampus or freshly dissociated from mouse hippocampus, but not dissociated astrocytes from mouse. NBCn2 was localized to both the plasma membrane of the soma and the processes of cultured rat hippocampal neurons. Similar to the aforemen-

34 23 tioned results by Praetorius et al., both Bouzinova et al. and Chen et al. found basolateral expression of NBCn2 in the choroid plexus from E18 and adult rats. Similar to the results by Chen et al., Jacobs et al. used a different polyclonal antibody and found NBCn2 protein in lysates from mouse brain, as well as mixed neuron/glia cultures, but not pure glia cultures (Jacobs et al., 2008). In immunohistochemical studies, the authors observed broad NBCn2 expression in cortex, cerebellum, and olfactory bulb. Based on results from colocalization immunohistochemical studies with antibodies to cell-specific proteins, NBCn2 expression was evident in the CA3 layer of the hippocampus, some GAD-positive interneurons in the neocortex, and parvalbumin-positive Purkinje neurons in the cerebellum (but not interneurons in the molecular layer). In a developmental study using in situ hybridization techniques, Hubner et al. first detected NBCn2 mrna as early as E12.5 in mouse brain regions, including cerebellum and cortex (Hubner et al., 2004). mrna levels increased from E15.5 to E18.5 in hippocampus, and from E12.5 to E18.5 in cortex. The transcript expression followed a neuronal pattern and was not detected in large fiber tracts. mrna was found in the neuronal cell layer of the retina, the retinal pigment epithelium, as well as in certain regions of the central auditory pathway including the geniculate nucleus and auditory brain stem nuclei. NCBE mrna was also detected in the choroid plexus as expected from the aforementioned studies. Chen et al. used immunoblotting to examine and compare expression levels of NBCn2 in the cortex, cerebellum, hippocampus, and subcortex of mouse brain with development from P16 to P118 (Chen et al., 2007). In general, expression was highest in the cortex and hippocampus at the different age groups, and expression in any given re-

35 24 gion did not change with age. This developmental expression profile is different than with NBCn1, which displayed increased levels with age (Chen et al., 2007). Electroneutral Na-driven Cl-HCO 3 Exchanger, NDCBE In a preliminary immunohistochemistry study, Risso Bradley et al. reported NDCBE1 expression in the soma and dendrites of rat cerebellar Purkinje neurons (Risso Bradley et al., 2001). In a subsequent RNA study on rat brain, Praetorius et al. used RT- PCR techniques and found NDCBE1 mrna in cerebrum and cerebellum, but not in the choroid plexus (Praetorius et al., 2004b). In a more extensive protein study, Chen et al. recently used immunochemical approaches and an antibody to the N-terminus of human NDCBE-B to localize protein expression in mouse and rat brain (Chen et al., 2008b). In immunoblotting studies, the authors assessed NDCBE-B expression in membrane proteins from the cortex, subcortex, cerebellum, and hippocampus from mouse brain. NDCBE-B expression was higher in the cortex, subcortex, and cerebellum compared to the hippocampus. According to results from immunohistochemistry studies NDCBE-B is widely distributed in mouse and /or rat brain, with particular expression in cortex, hippocampus, cerebellum, substiantia nigra, brainstem, and olfactory bulb, In further studies with cell-specific antibodies, the authors found that NDCBE-B is expressed more so in neurons than astrocytes in the hippocampus and cerebellum of mouse, and neuronal expression is evident in both the soma and processes. Similar findings were obtained with neurons and astrocytes cultured from rat/mouse hippocampus. NDCBE-B protein was detected in the basolateral membrane of choroid plexus epithelial cells from E18, but not adult rat brains.

36 25 NDCBE is also expressed in human brain. Using human NDCBE1 primers specific for NDCBE-B, Damkier et al. found -B mrna expression in human cerebrum, hippocampus, and choroid plexus (Damkier et al., 2007). In addition, in immunohistochemistry studies using a human NDCBE1 antibody that is expected to recognize NDCBE-B, the authors showed NDCBE-B expression in pyramidal cells of human hippocampus. Role of NCBTs in Nervous System Function General Neuronal Excitability Neuron-glia interactions. As discussed above, a decrease in ph o typically inhibits neuronal firing, whereas an increase in ph o typically stimulates activity. Repetitive neuronal firing or depolarization as seen in spreading depression or epilepsy is often associated with a decrease in ph i. As mentioned earlier, a predominant acid-extrusion mechanism in astrocytes is an electrogenic NBC. Because an electrogenic NBC when active alters both ph i and the ph o by moving HCO 3 across the plasma membrane of cells, this transporter links changes in ph with neuronal activity. According to a classic model proposed by Chesler (Chesler, 1990) and Ransom (Ransom, 1992), an electrogenic NBC in astrocytes modulates neuronal excitability through changes in ph. According to this classic model, when a neuron fires an action potential, there is an increase in extracellular K + (K + o) that depolarizes neighboring astrocytes. This depolarization leads to stimulation of an electrogenic NBC activity in astrocytes and transport of Na +, HCO 3 and net-negative charge into the cell. The ensuing decrease in ph o tends to dampen further neuronal activity by inhibiting many ph-sensitive voltage- and ligand-gated channels. This negativefeedback model is predicted to be neuroprotective under pathophysiological conditions

37 26 associated with excessive and repetitive neuronal firing such as with epilepsy and spreading depression. In addition to regulating ph and modulating neuronal activity, the electrogenic NBC may also contribute to changes in the extracellular space (ECS) associated with neuronal activity. As mentioned by Ostby et al., neuronal stimulation causes a ~30% shrinkage of the ECS in brain (Ostby et al., 2009). To examine the impact of specific pathways involved in such shrinkage, the authors generated five models that incorporated the following traditional ion/water pathways: both the efflux of K + out of and the influx of Na + into firing neurons, the influx of Na +, K +, and Cl via channels into astrocytes, the influx of K + and the efflux of Na + via the Na-pump into astrocytes, and the osmotic movement of water into/out of astrocytes. In addition, the authors also evaluated the impact of two additional transporter-mediated pathways: i) the influx of Na + and HCO 3 via the NBC into astrocytes, and ii) the influx of Na +, K +, and Cl via the Na/K/Cl cotransporter 1 (NKCC1) into astrocytes. Although shrinkage could be reasonably well modeled in the absence of these additional transporter-mediated pathways, their presence provided better, more physiologically relevant results in agreement with the literature. Thus, an astrocytic NBC may act in concert with NKCC1 and additional ion/water channels to mediate neuronal activity-evoked decreases in the ECS. NBCn2 knockout mouse elevated seizure threshold. Recent knockout studies have highlighted the functional importance of NCBTs in the CNS. Jacobs et al. developed an NBCn2/NCBE knockout mouse by targeted disruption of exon 12 of SLC4A10 gene (Jacobs et al., 2008). In measuring ph i with fluorescence imaging and the ph-

38 27 sensitive dye BCECF, the authors found that CA3 pyramidal neurons from the knockout mice displayed slower ph i recoveries from propionate-induced acid loads than neurons from the wild-type mice. The slower ph i recoveries in the knockout mice correlated with propionate-induced decreases in the frequency of 4-aminopyridine (4-AP)-elicited interictal events in the CA3 region. Furthermore, in the continued presence of propionate, the frequency of interictal-like events in this region recovered to control levels in slices from the wild-type mice, but not the knockout mice. The lack of frequency recovery correlated with the incomplete ph i recovery in neurons from the knockout mice. Thus, NBCn2- mediated ph i regulation appears to influence excitability of the CA3 region. The knockout mice also have a high seizure threshold that may be due to a decrease in ph i caused by reduced NBCn2-mediated acid extrusion. Interestingly, in contrast to results from the knockout mice studies by Jacobs et al. described above, disruption of the SLC4A10 gene in a human patient is associated with partial complex epilepsy, in addition to loss of cognitive function, and moderate mental retardation (Gurnett et al., 2008). Therefore the role of NBCn2 in epilepsy and seizure disorders has yet to be fully elucidated. Gerbil seizure model elevated NBCe1 expression. The expression of NCBTs in the CNS may also be regulated by neuronal activity. Kang et al. compared NBCe1 expression in the hippocampus of gerbils that were grouped into seizure-resistant and seizure-sensitive groups depending on the degree of motor arrest induced by stroking the back of the neck (Kang et al., 2002). In seizure-sensitive animals, NBCe1 expression by immunohistochemistry increased in the CA1-3 regions of the hippocampus at the 30-min

39 28 time point following seizure induction. At the 3-h time point, expression increased further in the CA2-3 regions and also increased in the dentate gyrus (granule cells). At the 6-h time point, NBCe1 expression in all hippocampal regions decreased to pre-seizure levels. In the aforementioned correlative study, the functional impact of an increase NBCe1 expression on seizure activity is not entirely clear, but likely depends on HCO 3 - transport direction, which will be influenced by the transporter s Na + :HCO 3 stoichiometry, as well as the membrane potential (V m ) of the NBC-expressing brain cell. For an electrogenic NBC with a 1:2 Na + :HCO 3 stoichiometry, the electrochemical gradient would likely favor the transport of HCO 3 into a neuron or astrocyte with either a typical resting V m or a depolarized V m (e.g., during an action potential), thereby decreasing ph o and inhibiting neuronal activity. However, for an electrogenic NBC with a 1:3 Na + :HCO 3 stoichiometry, the electrochemical gradient would likely favor the transport of HCO 3 out of a neuron or astrocyte at a typical resting V m, thereby increasing ph o and promoting neuronal activity. Further studies are required to elucidate the physiologic significance of increased NBC expression/function and neuronal activity. Somatosensory Function NBCn1 knockout mouse visual impairments. NBCn1, which is expressed in both the eye and ear, is necessary for somatosensory function (Bok et al., 2003; Lopez et al., 2005). We will initially focus on the eye, and then discuss the ear in the next paragraph. Mice with a disruption of the SLC4A7 gene that encodes for NBCn1 are blind and have hearing defects (Bok et al., 2003). Based on results from confocal immunofluorescence studies on wild-type mice, NBCn1 is expressed in the outer plexiform layer of ret-

40 29 ina, and more specifically in photoreceptor synaptic terminals. In knockout compared to wild-type mice, there was a progressive deterioration of the photoreceptor region of the retina beginning at two months of age and leading to complete loss at eleven months. At four months of age, the morphology of the fundi region of knockout mice had markedly deteriorated as demonstrated by attenuation of blood vessels and diffuse granularity. The authors propose that eye defects in the NBCn1- knockout mice may result from interrupted NBC-mediated buffering of H + that would lower ph i and inhibit Ca 2+ efflux via the Ca 2+ -pump and/or Ca 2+ influx via ph-sensitive L-type voltage-gated Ca 2+ channels at the photoreceptor synaptic terminal. NBCn1 knockout mouse auditory impairments. In the same above study, Bok et al. also found hearing defects in the NBCn1 knockout mice (Bok et al., 2003). Based on results from cochlear immunohistochemical studies on wild-type mice, NBCn1 is expressed in fibrocytes of the spiral ligament. The authors found that one-month old knockout mice displayed degenerating inner and outer hair cells of cochlea (which progressed with age), and abnormal morphology of the stria vascularis and spiral ligament. The Reissner s membrane was collapsed a finding consistent with a role for NBCn1 in endolymph formation. In knockout compared to wild-type mice at three months of age, there was also a decrease in the amplitude of auditory brainstem responses elicited by acoustic clicks to the ear. Thus, NBCn1 plays an important role in the development and function of the auditory system. According to Bok et al., the NBCn1 knockout mouse serves as a model for Usher syndrome, which is a human disorder associated with similar hearing and vision disabili-

41 30 ties. SLC4A7 appears responsible for type 2B of Usher syndrome (Pushkin et al., 1999b), which has been mapped in a consanguineous Tunisian family with vision and hearing defects (Hmani-Aifa et al., 2002; Hmani et al., 1999). More recently, Lopez et al. performed a detailed characterization of the time course of cochlear hair cell degeneration and loss from postnatal P2 to P90 NBCn1 knockout mice (Lopez et al., 2005). Based on results from light and transmission electron microscopy studies, the authors observed degeneration of the hair cells as early as P21. The degeneration progressed with age until complete degeneration at P90. There was atrophy in the organ of Corti beginning at P21. At P90, both the organ of Corti and the myelin that surrounded spiral ganglia neurons were degenerated. In addition, there was formation of vacuoles in the cytoplasm of the spiral ganglia neurons. Apoptotic cells were detected in the cochlea as early as P8. Cerebrospinal Fluid Secretion NBCn2 knockout mouse. In the same study described above on NBCn2 knockout mice, Jacobs et al. used magnetic resonance imaging and measured an ~80% decrease in the brain ventricular volume in knockout vs. wild-type mice (Jacobs et al., 2008). The decreased volume was not the result of edema or increased intracranial pressure, but instead due to enlarged lateral intercellular spaces and reduced apical microvilli in the choroid plexus according to data from histological and electron microscopy studies. To examine NCBE-mediated CSF production, the authors performed fluorescence imaging experiments with the ph-sensitive dye BCECF to characterize HCO 3 -dependent ph i regulation in the epithelial cells of choroid plexus from the knockout mice. Na + - and

42 31 HCO 3 -dependent ph i recoveries from NH + 4 prepulse-induced acid loads were reduced in the knockout vs. wild-type mice. Given the localization of NBCn2 on the basolateral membrane of the choroid plexus epithelium (see above), the authors suggest that NBCn2 on the basolateral membrane contributes to Na + - and HCO 3 -dependent CSF production. Decreased HCO 3 -dependent CSF production is likely responsible for the reduced brain ventricular volume in the NBCn2 knockout mice. There is evidence that NCBT knockout mice may express compensatory acid-base transporters to regulate ph i. For example, compensatory ph i regulating mechanisms appear in the choroid plexus epithelial cells of the NBCn2 knockout mice. According to results from immunofluorescence studies, Damkier et al. found that the Na-H exchanger 1 (NHE1), which is usually localized to the apical membrane in the choroid plexus, is expressed in the basolateral membrane in the choroid plexus epithelial cells from the knockout mice (Damkier et al., 2009). In functional studies on the choroid plexus epithelium from knockout mice, there was an increase in the activity of an apical EIPAinsensitive NHE responsible for Na + -dependent ph i recoveries from NH + 4 -induced acid loads. The apical NHE was sensitive to EIPA in the wild-type mice. According to the authors, NBCn2 knockout increases the expression of an EIPA-sensitive NHE (NHE1) from the apical to basolateral membrane, and also increases the activity of an EIPA-insensitive NHE on the apical membrane.

43 32 Role of NCBTs in Energy-deficient Neurologic Conditions Ischemia Jung et al. examined the expression profiles of four sodium transporters, including NBCe1, in the ischemic penumbra in a rat model of focal cerebral ischemia induced by occlusion of the left middle cerebral artery (Jung et al., 2007). By immunoblotting procedures, the authors reported increased NBCe1 expression in ischemic penumbra tissues 3 and 6 h post surgery in the ischemic model vs. sham-operated controls. Although the reason for this ischemia-induced increase in NBCe1 expression is not known, the accompanying increase in electrogenic NBC activity may protect against associated decreases in ph i. Furthermore, ischemia-induced increases in extracellular K + (Leis et al., 2005) may be involved. According to the model of Jung et al., an ischemia-induced increase in extracellular K + will depolarize brain cells (e.g., astrocytes) and stimulate electrogenic NBC activity. The increased cellular influx of HCO 3 will buffer ischemiainduced decreases in ph i. On the other hand, the accompanying Na + influx is expected to promote cellular edema. One of the hallmarks of ischemia is both intracellular and extracellular acidification. Furthermore, McKee et al. have demonstrated a decrease in Mg 2+ concentration in both the CSF and serum early on in ischemic stroke (McKee et al., 2005). Cooper et al. recently examined the expression of NBCn1-B and its contribution to the cytotoxicity of rat hippocampal neurons under such ischemic-like conditions such as low ph o and 0 extracellular Mg 2+ (Cooper et al., 2009). Using immunoblot procedures on cultured rat hippocampal neurons, the authors reported increased expression of NBCn1-B in cultured neurons under the ischemic-like conditions. Regarding the effects of low ph o, the authors

44 33 found an increase in NBCn1-B protein expression by immunofluorescence in the cell body, plasma membrane, and dendrites of neurons cultured at ph o 6.5, but not 7.4 for 1-3 h. According to results from subsequent cytotoxicity studies, glutamate cytotoxicity was decreased in neurons from NBCn1 sirna knockdown animals after incubation in Mg 2+ - free media for 2 and 6 h, but not when the cells were incubated at a low ph o of 6.3 for 6 h. The authors speculate that 0 Mg 2+ stimulates NMDA receptor activity, which in turn induces cytotoxicity. The cytotoxicity appears to be promoted by NBCn1-mediated acid extrusion. Hypoxia Hypoxia influences both the function and expression of NCBTs. In ph i experiments with BCECF on astrocytes cultured from the rat hippocampus, Bevensee et al. observed that acute hypoxia (3% O 2 ) stimulated total acid extrusion during ph i recoveries from NH + 4 -induced acid loads in the nominal absence of CO 2 /HCO 3, but actually inhibited SITS-sensitive acid extrusion during ph i recoveries in the presence of CO 2 /HCO 3 (Bevensee and Boron, 2008). The data are consistent with acute hypoxia stimulating the activity of the Na-H exchanger activity, but inhibiting the activity of the SITS-sensitive electrogenic Na/HCO 3 cotransporter in the astrocytes. Hypoxia also stimulated an acidloading mechanism responsible for the ph i recovery from an alkali load in the astrocytes. Thus, acute hypoxia stimulated a HCO 3 -independent acid extruder and loader, but inhibited a HCO 3 -dependent acid extruder in the astrocytes. To examine hypoxia-induced changes in the expression of NDCBE-B in mouse brain, Chen et al. performed immunoblotting of membrane proteins from different re-

45 34 gions of the brain from adult and neonatal mice subjected to chronic continuous hypoxia (11% O 2 ) for 14 or 28 d (Chen et al., 2008a). At both the 14- and 28-d time points, chronic hypoxia decreased NDCBE-B expression by 20-50% in cortex, subcortex, hippocampus, and cerebellum of adults. In the neonates, the NDCBE-B protein expression was decreased by ~25% in both the hippocampus after 14 d and the subcortex after 28 d hypoxic treatments. The authors proposed that the general reduced effect of chronic hypoxia on NDCBE expression in neonates vs. adults might be due to a smaller compensated respiratory alkalosis and/or reduced hypoxia sensitivity of the neonates. Chen et al. performed similar studies on NBCn1-B and NBCn2 (Chen et al., 2007). At both the 14- and 28-d time points, chronic hypoxia decreased the expression of NBCn1-B and NBCn2 by 15-50% in cortex, subcortex, hippocampus, and cerebellum of both neonates and adults. The authors proposed that the hypoxia-induced decrease in protein expression of NBCn1-B and NBCn2 may result in decreased energy consumption in the brain by reducing either protein synthesis, the requirement for Na + -pump mediated Na + extrusion, and/or neuronal activity caused by associated changes in ph i /ph o. Unlike Chen et al. who examined the effect of chronic continuous hypoxia on expression of NCBTs in mouse brain, Douglas et al. examined the effect of chronic intermittent hypoxia on the expression of NBCe1 in different regions of the postnatal mouse CNS (Douglas et al., 2003). The authors performed immunoblotting in membrane fractions from cerebellum, cortex, hippocampus, and brainstem-diencephalon using three different NBCe1 antibodies. One antibody recognized all three NBCe1 variants, one recognized the C terminus of the NBCe1-A and -B variants, and the final one recognized the different C terminus of the C variant (see details in Chapter 2). Using the antibody that

46 35 recognized all three variants, Douglas et al. observed a 33% decrease in NBCe1 expression in the cerebellum of hypoxia-exposed mouse compared to controls. Using the antibody to the C-terminus of NBCe1-A/B, the authors reported a decrease in the expression of NBCe1-A/B in the hippocampus of hypoxic mice. However, with the antibody to the C-terminus of the C variant, NBCe1-C did not appear to show any appreciable change in expression in brain regions from hypoxic and normoxic mice. Therefore, the effects of chronic intermittent hypoxia on NBCe1 expression are both transporter variant and brain region dependent. Goals of the Dissertation The focus of this dissertation is on NBCe1. Although early localization studies as described above have detected NBCe1 in brain, there has been no detailed characterization of the localization profiles of the different NBCe1 variants. Therefore it is important to examine the expression profiles of the three different NBCe1 variants in brain. The overall goal of this dissertation is to understand the localization and function of NBCe1 in rat brain. We hypothesized that NBCe1 splice variants have different localization profiles in rat brain and neurons cultured from the rat hippocampus exhibit electrogenic NBC activity. This dissertation has two specific aims. The first specific aim is to study the localization profiles of all three NBCe1 splice variants at mrna and protein levels and identify cell types that express different NBCe1 splice variants (Chapter 2). Based on the results from experiments for the first specific aim we found that while NBCe1-B was expressed intracellularly in neurons, NBCe1-C was expressed in the plasma membranes of glia

47 36 more so than the surrounding neurons. The second specific aim was to examine whether the electrogenic NBC expressed in neurons is functional (Chapter 3). Studies for the second aim of the dissertation provide an evidence for a functional electrogenic NBC in any neuronal preparation for the first time. As described earlier, in the CNS an electrogenic NBC in astrocytes modulates neuronal activity through changes in ph i /ph o. However, the functional significance of an electrogenic NBC in neurons is not known. Also, the activity of a neuronal NBC will be governed by ph, ion, and/or voltage dependencies that may differ for an astrocytic NBC. Thus, it is critical to characterize and compare the activities of these electrogenic NBCs. Chapter 2 Overview We have performed detailed and systematic localization studies of all three NBCe1 variants at the mrna and the protein levels using a combination of in situ hybridization and immunochemistry. At the mrna level, we found that while NBCe1-A is nearly absent in brain, the two brain specific variants are -B and -C. At the protein level, both NBCe1-A/B and -C were expressed in brain. To identify the cell types that express these variants, we performed immunohistochemistry using cell specific markers for both neurons and glial cells. The results from these studies showed that NBCe1-C was expressed in the plasma membranes of primarily glial cells than neurons, while NBCe1-A/B was expressed intracellularly in neurons. These results were further confirmed by immunoelectron microscopy.

48 37 Chapter 3 Overview We have used a combination of electrophysiological and immunohistochemistry techniques to examine whether rat hippocampal neurons exhibit an electrogenic NBC both at the functional and the molecular levels. For the electrophysiology experiments, we have performed a bicarbonate-induced outward current assay and a Na removalinduced inward current assay in both whole-cell and perforated patch-clamp modes. In both assays, our results show that a population of rat hippocampal neurons exhibit electrogenic NBC activity. However results from immunohistochemistry studies show that all neurons express NBCe1. Therefore, the heterogeneity seen in NBC function was not due to differences in NBCe1 expression. Moreover, NBCe1 expression was not restricted to inhibitory or excitatory neurons as revealed by colocalization studies using GAD67, a marker of inhibitory neurons. Chapter 4 Overview Chapter 4 of this dissertation presents a general discussion of the findings from studies described in Chapters 2 and 3.

49 38 Table I. Molecular characterization and localization of Na-coupled bicarbonate transporters (NCBTs) in tissues NCBT Variants Alternate names SLC nomenclature Tissue distribution References NBCe1 A NBC1, rknbc SLC4A4 Salamander kidney (cloning) Romero et al.,1997 knbc, anbc, Human kidney (cloning), knbc-1, human kidney, pancreas (mrna) Burnham et al.,1997 Rat kidney (cloning), rat tissues (mrna) Romero et al.,1998 Burnham et al.,1998 Rat kidney (mrna) Abuladze et al., 2004 Xu et al., 2003 Human cornea (mrna) Usui et al., 1999 Human kidney (mrna) Abuladze et al.,1998 Rat kidney, brain (mrna) 1 Giffard et al., 2000 Rat heart (mrna) 2 Yamamoto et al., 2007 Rat kidney (mrna) 2a Praetorius et al., 2004b Murine colon, other tissues (mrna) Bachmann et al., 2003 Rat pancreas (mrna, protein) Soyfoo et al., 2009 Murine duodenum (mrna, protein) Praetorius et al., 2001 Human tissues (mrna), human duodenum (protein) 3 Damkier et al., 2007 Rat brain (mrna, protein) 4 Schmitt et al., 2000 Rat epididymis (mrna, protein) 5 Jensen et al., 1999 Human pancreas (mrna, protein) Marino et al., 1999 Rat kidney, pancreas (mrna, protein) Satoh et al., 2003 Rat, ambystoma kidney (protein) Maunsbach et al., 2000 Mouse brain (protein) Rickmann et al., 2007 Mouse brain (protein) 5a Douglas et al., 2003 Rat, salamander, rabbit kidney (protein) Schmitt et al., 1999 Rat salivary gland (protein) Roussa et al., 1999 Rat conjunctiva (protein) Bok et al., 2001 Mouse kidney (protein) Pushkin et al., 2004 Rat kidney (protein) Kwon et al., 2002 Rat pancreas (protein) 6 Thévenod et al.,1999 Human,bovine cornea (protein) 7 Sun and Bonanno, 2003 Sun et al., 2000 Zebrafish eye, gill (protein) 8 Sussman et al., 2009 Rat brain (protein) 8a,8b Douglas et al., 2001 Jung et al., 2007 Gerbil brain (protein) 8c Kang et al., 2002

50 39 Table I. cont d NCBT Variants Alternate names SLC nomenclature Tissue distribution References B NBC1, SLC4A4 Rat pancreas (cloning) Thévenod et al.,1999 rb1nbc Human pancreas (cloning), pnbc-1 human tissues and mouse pancreas (mrna) Abuladze et al., 1998 Human heart (cloning), human pancreas (mrna) Choi et al., 1999 Zebrafish (cloning), zebrafish tissues (mrna) 9 Sussman et al., 2009 Rat brain (cloning), rat tissues (protein) 9a Bevensee et al., 2000 Murine duodenum, mouse pancreas (mrna) Praetorius et al., 2001 Human cornea (mrna) Usui et al., 1999 Sun and Bonanno, 2003 Rat pancreas (mrna) Abuladze et al., 2004 Rat tissues (mrna) 9b Giffard et al., 2000 Bovine cornea (mrna, protein) Sun et al., 2000 Rat brain (mrna, protein) Majumdar et al Rat pancreas (mrna, protein) Soyfoo et al., 2009 Rat eye (protein) Bok et al., 2001 Rat, human pancreas (protein) Satoh et al., 2003 Mouse brain (protein) 10 Rickmann et al., 2007 C NBC1, SLC4A4 Rat brain (cloning), rb2nbc rat tissues (protein) Bevensee et al., 2000 Rat brain (mrna, protein) Majumdar et al., 2008 Mouse brain (protein) Douglas et al., 2003 NBCe2 A NBC4 SLC4A5 Human testis (cloning), human tissues (mrna) 11 Rat, mouse choroid plexus (mrna) 11a Pushkin et al., 2000b Praetorius et al., 2004b Rat heart (mrna) 11b Yamamoto et al., 2007 Human tissues (mrna), human kidney (protein) 12 Damkier et al., 2007 Rat tissues (mrna) 12a Xu et al., 2003 Mouse, rat choroid plexus (protein) 13 Bouzinova et al.,2005 Damkier et al., 2009 B NBC4 SLC4A5 Human heart (cloning) Pushkin et al., 2000a,b C NBC4 SLC4A5 Human tissues (cloning) Virkki et al., 2002 Human testis (cloning), human tissues (mrna) Sassani et al., 2002 Rat liver (cloning), rat testis (mrna) rat liver, kidney (mrna, protein) Abuladze et al., 2004 D NBC4 SLC4A5 Rat testis (mrna) Abuladze et al., 2004 E F NBC4 SLC4A5 Rat liver (cloning) Xu et al., 2003

51 40 Table I. cont d NCBT Variants Alternate names SLC nomenclature Tissue distribution References NBCn1 A NBC3 SLC4A7 Human skeletal muscle (cloning), human skeletal muscle, heart (mrna) Pushkin et al.,1999a Rat heart (mrna) 13a Yamamoto et al.,2007 Rat tissues, aorta (mrna) 14 Choi et al., 2000 Rat aorta, murine duodenum (mrna), murine duodenum (protein) 15 Praetorius et al., 2001 Human salivary glands, kidney (mrna), rat, human salivary glands, kidney (protein) 16 Gresz et al., 2002 Rat kidney (mrna, protein) 17 Praetorius et al., 2004a Rat tissues (mrna, protein) 18 Damkier et al., 2006 Rat kidney (mrna, protein) 18a Odgaard et al., 2004 Mouse tissues (mrna, protein) 19 Boedtkjer et al., 2008 Rat brain (mrna), rat, mouse choroid plexus (mrna, protein) 19a Praetorius et al., 2004b Human tissues (mrna), human kidney, duodenum (protein) 20 Damkier et al., 2007 Rat kidney (protein) 21 Kwon et al., 2002 Vorum et al., 2000 Human choroid plexus (protein) 21a Praetorius and Nielsen, 2006 Mouse ear (protein) 21b Lopez et al., 2005 Mouse ear, eye (protein) 21c Bok et al., 2003 Mouse choroid plexus (protein) 21d Damkier et al., 2009 Rat brain (protein) 21e Cooper et al., 2005 Risso Bradley et al., 2001 B NBC3 SLC4A7 Rat smooth muscle (cloning) Choi et al., 2000 Rat salivary glands, kidney (mrna) Gresz et al., 2002 Rat brain (mrna) Cooper et al., 2005 Mouse brain (protein) Chen et al., 2007 Rat brain (protein) Cooper et al., 2009 C NBC3 SLC4A7 Rat smooth muscle (cloning) Choi et al., 2000 Rat salivary gland, kidney (mrna) Gresz et al., 2002 D NBC3 SLC4A7 Rat smooth muscle (cloning) Choi et al., 2000 Rat salivary gland, kidney (mrna) Gresz et al., 2002 E NBC3 SLC4A7 Rat brain (cloning) Cooper et al., 2005 F- H NBC3 SLC4A7

52 41 Table I. cont d NCBT Variants Alternate names SLC nomenclature Tissue distribution References NBCn2 A NCBE SLC4A10 Insulin-secreting cell line (cloning), mouse pancreas, rat tissues, insulin-secreting cell line (mrna) 22 Wang et al., 2000 Human brain, choroid plexus, other tissues (mrna) 22a Damkier et al., 2007 Human brain (cloning, mrna 23 ) Parker et al., 2008a Mouse brain, other tissues (mrna) 24 Hubner et al., 2004 Rat tissues (mrna) 25 Giffard et al., 2003 Mouse brain (mrna 25a, protein 25b ) Jacobs et al., 2008 Rat,mouse brain (protein) 26 Chen et al., 2007 Chen et al., 2008c B rb1ncbe SLC4A10 Human brain (cloning) Parker et al., 2008a Rat brain (cloning), rat, mouse brain (mrna) mouse brain (protein) Giffard et al., 2003 Rat brain (mrna), rat, mouse choroid plexus (mrna, protein) Human choroid plexus (protein) Praetorius et al., 2004b Praetorius and Nielsen, 2006 Mouse brain (protein) Damkier et al., 2009 C rb2ncbe SLC4A10 Rat brain (cloning), rat, mouse brain (mrna) mouse brain (protein) Giffard et al., 2003 Mouse choroid plexus (mrna, protein) Human choroid plexus (protein) Praetorius et al., 2004b Praetorius and Nielsen, 2006 D NCBE SLC4A10 NDCBE A knbc-3 SLC4A8 Human brain (cloning, mrna) Parker et al., 2008b Mouse kidney (cloning), mouse tissues (mrna) Wang et al., 2001 Squid giant fiber lobe (cloning), squid tissues (mrna) 27 Virkki et al., 2003

53 42 Table I. cont d NCBT Variants Alternate names SLC nomenclature Tissue distribution References B NDCBE1 SLC4A8 Human brain (cloning), human brain, other tissues (mrna) Grichtchenko et al., 2001 Rat brain (mrna) Praetorius et al., 2004b Rat and mouse brain (protein) Chen et al., 2008a,b Human brain, choroid plexus, other tissues (mrna), human brain (protein) Damkier et al., 2007 Rat brain (protein) Risso Bradley et al., 2001 C SLC4A8 Human brain (cloning), human tissues (mrna) D, D SLC4A8 Human brain (cloning), human tissues (mrna) Parker et al., 2008b Parker et al., 2008b

54 43 Footnotes for Table I. 1 In situ hybridization studies were performed using a probe that is expected to recognize both NBCe1-A and -B. This probe labeled NBCe1 mrna in brain. 2,2a The primers that were used are expected to recognize other NBCe1 variants 3 The primers and antibody used for the study are expected to recognize other NBCe1 variants. 4,5 Polynucleotide probes and antibody used for the studies are expected to recognize other NBCe1 variants. 5a One antibody used in the study is expected to recognize other NBCe1 variants. The other antibody used recognized NBCe1-A/B variants. 6,7,8,8a,8b,8c The antibody used is expected to recognize other NBCe1 variants. 9 There is 82% sequence identity of zebrafish cdna with that of NBCe1-B. 9a The C-terminal antibody recognized both NBCe1-A and -B. 9b The probe used is expected to recognize NBCe1-B and -C. 10 The N-terminal antibody used for the study recognized both NBCe1-B and -C variants. 11 A probe was used that recognized NBCe2-A and -B. 11a The probe used that was is expected to recognize other NBCe2 variants. 11b The primers used are expected to recognize other NBCe2 variants. 12 The PCR products in the human tissues recognized NBCe2-A and -C variants. The antibody is expected to recognize other NBCe2 variants. 12a The probe used is expected to recognize all NBCe2 variants. 13 The antibodies used in the study are expected to recognize other NBCe2 variants. 13a The primers used are expected to recognize other NBCn1 variants. 14 A probe and primers were used that are expected to recognize other NBCn1 variants. 15 The primers and antibody used for the study are expected to recognize other NBCn1 variants. 16 Antibodies used in the study are expected to recognize other NBCn1 variants. 17,18, 18a The primers and antibody used in the study are expected to recognize other NBCn1 variants. 19 The labeling procedure is expected to recognize other NBCn1 variants. 19a The primers and antibody used is expected to recognize other NBCn1 variants 20 The primers and antibody used are expected to recognize other NBCn1 variants. 21,21a,21b,21c, 21d, 21e The antibody used in the study is expected to recognize other NBCn1 variants. 22 The cdna and primers used for cloning and mrna analysis are expected to recognize other NBCn2 variants. 22a The primers used is expected to recognize other NBCn2 variants. 23 A probe common to NBCn2-A and -B was used. 24,25,25a The probe used in the study are expected to recognize other NBCn2 variants. 25b The antibody used in the study is expected to recognize other NBCn2 variants. In addition, some immunolabeling was also performed using the diaminobenzidine (DAB) staining protocol. 26 The antibody used in the study is expected to recognize other NBCn2 variants. 27 Cloning and mrna studies are expected to recognize other NDCBE variants.

55 44 LOCALIZATION OF ELECTROGENIC Na/BICARBONATE COTRANSPORTER NBCe1 VARIANTS IN RAT BRAIN by DEBESHI MAJUMDAR, ARVID B. MAUNSBACH, JOHN J. SHACKA, JENNIFER B. WILLIAMS, URS V. BERGER, KEVIN P. SCHULTZ, LUALHATI E. HARKINS, WALTER F. BORON, KEVIN A. ROTH, AND MARK O. BEVENSEE Neuroscience 155 (2008) Copyright 2008 by Elsevier Ltd. Used by permission Format adapted for dissertation

56 45 ABSTRACT The activity of HCO 3 transporters contributes to the acid-base environment of the nervous system. In the present study, we used in situ hybridization, immunoblotting, immunohistochemistry, and immunogold electron microscopy to localize electrogenic Na/bicarbonate cotransporter NBCe1 splice variants (-A, -B, and -C) in rat brain. The in situ hybridization data are consistent with NBCe1-B and -C, but not -A, being the predominant NBCe1 variants in brain, particularly in the cerebellum, hippocampus, piriform cortex, and olfactory bulb. An antisense probe to the B and C variants strongly labeled granule neurons in the dentate gyrus of the hippocampus, and cells in the granule layer and Purkinje layer (e.g., Bergmann glia) of the cerebellum. Weaker labeling was observed in the pyramidal layer of the hippocampus and in astrocytes throughout the brain. Similar, but weaker labeling was obtained with an antisense probe to the A and B variants. In immunoblot studies, antibodies to the A and B variants ( A/B) and C variant ( C) labeled ~130-kDa proteins in various brain regions. From immunohistochemistry data, both A/B and C exhibited diffuse labeling throughout brain, but A/B labeling was more intracellular and punctate. Based on co-localization studies with antibodies to neuronal or astrocytic markers, A/B labeled neurons in the pyramidal layer and dentate gyrus of the hippocampus, as well as cortex. C labeled glia surrounding neurons (and possibly neurons) in the neuropil of the Purkinje cell layer of the cerebellum, the pyramidal cell layer and dentate gyrus of the hippocampus, and the cortex. According to electron microscopy data from the cerebellum, A/B primarily labeled neurons intracellularly and C labeled astrocytes at the plasma membrane. In summary, the B and

57 46 C variants are the predominant NBCe1 variants in rat brain and exhibit different localization profiles. KEYWORDS Cerebellum, glia, hippocampus, neuron, ph, transporter ABBREVIATIONS BT, bicarbonate transporter; DIG, digoxigenin; ER, endoplasmic reticulum; HRP, horseradish peroxidase; K + o, extracellular K + ; MAP2, microtubule-associated protein; NBC, Na/bicarbonate cotransporter; NBCe1 (or 2), electrogenic Na/bicarbonate cotransporter 1 (or 2); NeuN, neuronal nuclei; OB, olfactory bulb; ORF, open reading frame; PBS, phosphate-buffered saline; PFA, paraformaldehyde; ph i, intracellular ph, ph o, extracellular ph; RT, room temperature; UAB, University of Alabama at Birmingham. INTRODUCTION It is well established that ph is an important cellular parameter because alterations in ph can affect a variety of cellular processes including the activity of enzymes and ion channels. In the nervous system, changes in ph can influence neuronal activity. In general, a decrease in extracellular ph (ph o ) inhibits neuronal activity, whereas an increase in ph stimulates activity (Chesler and Kaila, 1992; Ransom, 2000; McAlear and Bevensee, 2003; Chesler, 2003). Changes in ph are also seen with neuropathologic conditions. For example, decreases in brain ph are associated with ischemia, hypoxia, and epileptic events leading to neuronal necrosis and overall brain damage (Katsura and Siesjo, 1998).

58 47 The regulation of brain ph is complex because the movement of acid-base equivalents across the plasma membranes of neurons and glia (e.g. astrocytes and oligodendrocytes) changes both ph o and intracellular ph (ph i ). Cells use a system of acid-base transporters in their plasma membranes to regulate ph i, and subsequently ph o. Some of the most powerful acid-base transporters in brain cells include HCO 3 -dependent ones, which are now classified as members of a bicarbonate transporter (BT) superfamily following their molecular identification over the past decade (Romero et al., 2004). The BT superfamily is comprised of anion exchangers (AEs), electrogenic and electroneutral Na/bicarbonate cotransporters (NBCs), and Na-driven Cl-bicarbonate exchangers (NDCBEs). The role of an electrogenic NBC in the nervous system has received considerable attention because the transporter couples changes in ph with neuronal activity (Chesler, 2003). In the nervous system, the activity of an electrogenic NBC was first characterized in invertebrate leech glial cells (Deitmer and Schlue, 1987, 1989; Deitmer, 1992), and subsequently identified in a number of mammalian astrocytes (see (Rose and Ransom, 1998; McAlear and Bevensee, 2003)). In at least rat hippocampal astrocytes, the transporter has a stoichiometry of 1 Na + :2 HCO 3 (Bevensee et al., 1997a,b) and functions as an acid extruder. According to the model proposed by Chesler (1990) and Ransom (1992), astrocytes modulate neuronal excitability through NBC-mediated changes in ph o. More specifically, an elevation in extracellular K + (K + o) during neuronal activity depolarizes astrocytes, thereby stimulating electrogenic NBC activity to transport Na +, HCO 3, and net negative charge into the astrocytes. The ensuing decrease in ph o will tend to dampen further neuronal activity. In support of this model, high K + o

59 48 depolarizes astrocytes in the gliotic hippocampal slice and elicits a Na + - and HCO 3 - dependent decrease in ph o due to the activity of an NBC (Grichtchenko and Chesler, 1994a,b). Our understanding of NBCs in the brain has advanced with the molecular identification of cdnas encoding variants of electrogenic NBCs (NBCe1 and NBCe2) and electroneutral NBCs (NBCn1). NBCe1 the first NBC to be cloned (Romero et al., 1997) has three splice variants: NBCe1-A, -B, and -C (Romero et al., 2004). In early cloning studies, the A variant was identified in rat and human kidney (Burnham et al., 1997; Romero et al., 1998), the B variant in human heart (Choi et al., 1999), human pancreas (Abuladze et al., 1998a), and rat brain (Bevensee et al., 2000), and the C variant in rat brain (Bevensee et al., 2000). These three variants are identical to one another except at the amino and/or carboxy termini. The amino-terminal 85 residues of the B and C variants replace 41 residues of the A variant, whereas the carboxy-terminal 61 residues of the C variant replace 46 residues of the A and B variants. This C variant found predominantly in brain (Bevensee et al., 2000) arises from a 97-base pair deletion near the 3 open reading frame (ORF), thereby shifting the stop codon further 3. Shortly after NBCe1 was identified, two groups (Schmitt et al., 2000; Giffard et al., 2000) examined the localization of NBCe1 in rat brain primarily using polynucleotide probes and/or polyclonal antibodies that did not distinguish among the three splice variants. Somewhat expectedly, both groups found NBCe1 widely distributed throughout the brain and localized to glial cells. However, a surprise finding was the presence of NBCe1 in some neuronal populations, including granule cells of the hippocampus and neurons of the cortex (Schmitt et al., 2000).

60 49 Clearly, it would be informative to know which of the three NBCe1 variants are expressed in brain, and in what cell types. In the aforementioned Giffard et al. study on rat brain (Giffard et al., 2000), the authors noted that in situ hybridization with the nonspecific NBCe1 probe was mimicked by a more specific probe to the B variant (which should have also recognized the C variant), but not by a probe to the A variant, which displayed minimal labeling. Somewhat surprisingly, Rickmann et al. (2007) recently reported the expression of both NBCe1-A and NBCe1-B in mouse brain based on immunohistochemistry and immunoelectron microscopy. The authors found expression of the A variant throughout brain, particularly in neuronal subpopulations in the cerebellum, hippocampus, cerebral cortex, and olfactory bulb (OB). In contrast, the authors reported the expression of NBCe1-B in astroglia and Bergmann glia. However, it is important to point out that the NBCe1-B antibody is expected to recognize not only NBCe1-B, but also NBCe1-C, which is predominantly expressed in brain of at least rat. Therefore, the expression profiles of the B and C variants in brain have yet to be eludicated. In the present study, we used mrna and protein localization techniques to examine all three NBCe1 splice variants in rat brain. According to in situ hybridization data, NBCe1-A mrna is minimal in brain, whereas NBCe1-B and possibly -C mrnas are predominantly found in cerebellum (granule and Purkinje layers), hippocampus (dentate gyrus and pyramidal layers), piriform cortex, and OB. In immunoblot studies, we used previously characterized antibodies (Bevensee et al., 2000) that distinguish between the C terminus of either the C variant ( C) or the A/B variant ( A/B). Both antibodies recognized a ~130-kDa protein from all brain regions. Based on immunohistochemistry studies, A/B displayed weak diffuse labeling throughout the brain and punctate labeling of

61 50 pyramidal neurons in the hippocampus and neurons in the cortex. In contrast, C displayed stronger diffuse labeling throughout the brain. C appeared to label glia more so than the surrounding neurons in the cerebellum, hippocampus, and cortex. Immunoelectron microscopy data from the cerebellum support the plasma-membrane expression of NBCe1-C in astrocytes and intracellular expression of NBCe1-B in neurons. Portions of this work have been published in abstract form (Majumdar et al., 2006). EXPERIMENTAL PROCEDURES Animals All animals were housed in approved animal care facilities at the University of Alabama at Birmingham (UAB). The animal facilities are under the direction of full-time veterinarians and are fully accredited by the American Association for Accreditation of Laboratory Animal Care, and meet all standards prescribed by the Guide for the Care and Use of Laboratory Animals. All protocols involving animals were approved by the Institutional Animal Care and Use Committee at UAB. Efforts were made to minimize the required number of animals and their suffering in this study. In Situ Hybridization Designing NBC Probes Sense and antisense probes were created by PCR using primers that flanked the following regions: i) the unique 5' ORF encoding the unique amino-terminal 41 residues of the A variant, ii) the different 5' ORF encoding the different amino-terminal 85 residues of the B and C variants, and iii) the 97-bp region near the 3' end of the ORF found

62 51 exclusively in the A and B variants (Fig.1). PCR products were purified and directionally subcloned into PCRII TOPO vector (Invitrogen, Inc., Carlsbad, CA, USA). crna probes were generated by transcription using the T7 or SP6 promoter. Performing In Situ Hybridization Non-radioactive in situ hybridization was performed as previously described using digoxigenin (DIG)-labeled crna probes (Berger and Hediger, 2001). Frozen rat brain or kidney sections (10 m), were captured onto Superfrost plus microscope slides (Fisher Scientific, Pittsburgh, PA, USA). Sections were fixed, acetylated, and incubated in the NBC probes (~100 ng/ml) at 70 C for 3 days. Hybridized probes were visualized using alkaline phosphatase-conjugated anti-dig Fab fragments (Roche, Indianapolis, IN, USA) and 5-bromo-4-chloro-3-indolyl-phosphate/nitro blue tetrazolium substrate (Kierkegard and Perry Laboratories, Gaithersburg, MD, USA). Sections were rinsed several times in a ph-9.5 solution containing 150 mm NaCl, 100 mm Tris, and 20 mm EDTA, and then coverslipped with glycerol gelatin. Control sections were incubated in the same concentration of the sense probe. Immunoblotting NBCe1 Antibodies Previously characterized polyclonal antibodies (Bevensee et al., 2000) distinguish between the C terminus of either the C variant ( C) or the A/B variant ( A/B) (Fig.1). A/B and C labeling are specific based on competition immunoblot experiments whereby antibody labeling of microsomal protein from total rat brain or oocytes injected with

63 52 with a cdna encoding an NBCe1 variant was inhibited by preabsorbing the antibodies with the corresponding antigenic fusion protein (Bevensee et al., 2000; Williams and Bevensee, unpublished observations). Labeling Protein from Rat Brain Regions We used two commercially available adult rat brain blots (Chemicon International, Inc., Temecula, CA, USA) containing ~25-30 μg protein from different brain regions of Sprague-Dawley wk-old male rats (~150 g). The membranes were preincubated in BLOTTO (1 phosphate-buffered saline (PBS) + 5% dry-milk powder % Tween-20) for 30 min, and then BLOTTO containing either 1:500 rabbit A/B, 1:2K rabbit C, or 1:5K mouse -actin (Sigma-Aldrich, St. Louis, MO, USA) for 1 h. Membranes were subsequently incubated in BLOTTO containing horseradish peroxidase (HRP)-conjugated secondary anti-igg (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA, USA) for 1 h. Labeling was detected using the SuperSignal West Pico chemiluminescence kit (Pierce Biotechnology, Inc., Rockford, IL, USA) and recorded by exposure to HXR film (Hawkins X-Ray Supply, Oneonta, AL, USA). The blot was stripped of bound antibody by first incubating the membrane in stripping buffer (100 mm -mercaptoethanol, 2% SDS, 62.5 mm Tris-HCl, ph ~6.8) at 50 C for 30 min with occasional rocking, and then washing the membrane in PBS containing 0.05% Tween-20. Densitometry analysis was performed using Biorad Quantity One software (version 4.5.1).

64 53 Immunohistochemistry Antibody Labeling Excised brains from adult male Sprague-Dawley rats were fixed in Bouin s fixative (71% saturated picric acid, 24% formalin, and 5% glacial acetic acid) overnight at 4 C, and then rinsed with 70% ethanol (or 67% ethanol/3.5% isopropanol). The brains were subsequently incubated in 70% ethanol (or 67% ethanol/3.5% isopropanol) for at least 6 h at 4 C before being embedded in paraffin. Five-micrometer sagittal or coronal sections were cut using a HM 355S microtome (Richard-Allan Scientific, Kalamazoo, MI, USA) and mounted onto glass slides. Sections were de-paraffinized by incubating the sections first in Citrisolv (Fisher Scientific) for 5 and 10 min, and then in isopropanol (Fisher Scientific) 3 5 min. For optimal C labeling, the sections were then subjected to an antigen-retrieval protocol, which involved incubating the sections in a 10 mm citrate buffer (ph 6.0) in a steam-set rice cooker for 30 min. Sections were then oxidized with 3% H 2 O 2 for 10 min at room temperature (RT) to remove endogenous peroxidase activity. For optimal A/B labeling, antigen retrieval was not required. For antibody labeling, sections were incubated first in a blocking buffer containing PBS supplemented with 1% bovine serum albumin (BSA) (Fisher Scientific), 0.2% evaporated milk (Nestle, Solon, OH, USA), and 0.3% Triton X-100 for 15 min at RT, and then in the blocking buffer with primary antibody overnight at 4 C. In competition experiments, fusion protein was added to the blocking buffer plus primary antibody and rocked for 1 h at RT before the buffer was applied to sections overnight. Sections were next washed with 1 PBS, and then incubated in blocking buffering containing 1:1K or 1:2K HRP-conjugated donkey anti-rabbit, donkey anti-mouse, or donkey anti-guinea-pig

65 54 antibody (Jackson Laboratories) for 1 h at RT. Labeling was detected using tyramide signal amplification, which involved incubating sections in a 1:500 dilution of Cy3- or FITC-conjugated tyramide dissolved in Plus Amp buffer (PerkinElmer LAS, Inc., Boston, MA, USA) for 30 min at RT. Sections were next labeled with 0.2 μg/ml bis benzimide (Hoechst 33,258) nuclear counter stain for 10 min. The slides were coverslipped using 1:1 PBS:glycerol. In negative controls, labeling was very weak in the absence of primary antibody. For double antibody labeling protocols, results were similar regardless of the order of antibody labeling. A relative lack of residual HRP activity remaining from the first round of labeling was confirmed in control studies where the second primary antibody was omitted. All immunohistochemical labeling was performed a minimum of two times. Additional antibodies used in this study were anti-mouse neuronal nuclei (NeuN) (Chemicon), anti-mouse microtubule-associated protein (MAP2) (Sigma-Aldrich), antimouse Calbindin (Sigma-Aldrich), GFAP (Chemicon), GLAST (Chemicon), and antimouse 1 of the Na + pump (Upstate Biotechnology, Inc. Lake Placid, NY, USA). Visualizing Antibody Labeling For some images, antibody labeling of sections was examined with an Axioskop microscope (Zeiss, Thornwood, NY, USA) equipped with epifluorescence capabilities, and images were captured using Axiovision software. For other images, antibody labeling was examined with a Leica inverted DMIRBE confocal microscope (High Resolution Imaging Facility, University of Alabama at Birmingham). The following excitation lasers

66 55 were used: i) Argon (488 nm) for FITC (green), ii) Krypton (568 nm) for Cy3 (red), and iii) UV (351 nm) for nuclear labeling (blue). Electron Microscopy Four rats ( g) were anesthetized by halothane (Halocarbon, River Edge, NJ, USA) and perfusion-fixed through the heart for 3 min with a 4% paraformaldehyde (PFA) solution containing 0.1 M sodium cacodylate buffer (ph 7.2). The initial perfusion rate was ~230 ml/min. The brains were excised and post-fixed in the same fixative for 4 h. Small tissue blocks were trimmed from the cerebellum, rinsed in buffer, infiltrated with 2.3 M sucrose, mounted on holders, and frozen in liquid nitrogen. The tissue was then cryosubstituted in a Reichert AFS freeze-substitution apparatus (Leica, Germany) and embedded in Lowicryl HM20 as previously described (Maunsbach et al. 2000). Briefly, samples were sequentially equilibrated over 3 days in methanol containing 0.5% uranyl acetate at temperatures gradually increasing from 90 ºC to 70 ºC, rinsed in pure methanol, infiltrated with Lowicryl HM20 at 45 o C, and finally, UV-polymerized for 2 days at 45 o C and 2 days at 0 ºC. Ultrathin Lowicryl sections were cut at ambient temperature on a Leica Reichert Ultracut S ultramicrotome (Leica, Vienna, Austria) and first blocked by incubation in PBS containing 0.05 M glycine and 0.1% skimmed milk powder. The sections were then incubated overnight at 4ºC with 1:100 C or 1:25 A/B in PBS containing 0.1% skimmed milk powder. The primary antibodies were visualized using goat anti-rabbit IgG conjugated to 10-nm colloidal gold particles (GAR.EM1O, Bio- Cell Research Laboratories, Cardiff, UK) diluted 1:50 in PBS with 0.1% skimmed milk powder and polyethyleneglycol (5 mg/ml). The sections were stained with uranyl acetate

67 56 and lead citrate before examination with a Morgagni electron microscope (Eindhoven, The Netherlands). Immunolabeling controls consisted of replacing the primary-antibody serum with either pre-immune rabbit IgG or preabsorbed antibody serum containing the appropriate maltose binding protein (MBP)-fusion protein used to generate the antibody. All controls showed absence of specific labeling. Chemicals Unless specified otherwise, all chemicals were obtained from Sigma-Aldrich. RESULTS Localization of Different mrnas Encoding NBCe1 Variants in Adult Rat Brain Probe Specific to NBCe1-A We performed in situ hybridization using probes to the 5' ORF encoding the unique 41 amino-terminal residues of the A variant (Fig.1). The antisense A probe displayed very little if any labeling throughout a sagittal section of adult rat brain (Fig.2A), including the separate OB from a different section. The antisense A probe did not label the cerebellum (Fig.2B) or hippocampus (Fig.2C). We performed a positive control by examining hybridization of the antisense probe in rat kidney, where NBCe1-A mrna and protein are present in the proximal tubule (Romero et al., 1998; Abuladze et al., 1998b; Schmitt et al., 1999; Maunsbach et al., 2000; Aalkjær et al., 2004). The antisense A probe primarily labeled the outer cortex of rat kidney where proximal tubules are present. As seen at higher magnification (Fig.2D), the probe clearly labeled both proximal tubules (PT, see arrows near the glomerulus (G) in the center of the image) and some dis-

68 57 tal nephron segments. Therefore, the antisense A probe can detect NBCe1-A mrna, which appears absent in rat brain. Probe to NBCe1-B or -C We next performed in situ hybridization using probes to the 5' ORF encoding the different 85 amino-terminal residues of the B and C variants (Fig.1). In contrast to the absence of antisense A-variant labeling shown in Fig.2A, we found strong labeling with the antisense B/C probe throughout a sagittal section of adult rat brain (Fig.2E), but particularly in specific brain regions such as the cerebellum, hippocampus, piriform cortex, as well as OB. As a negative control, the sense B/C probe did not exhibit appreciable labeling (Fig.2F). The punctuate labeling throughout the brain is consistent with NBCe1- B/C mrna in astrocytes (Schmitt et al., 2000). In the cerebellum, the antisense-b/c probe labeled both the neuron-rich granule layer and Purkinje layer (Fig.2G), with labeling in the Purkinje layer consistent with NBCe1-B/C mrna in Bergmann glia (Fig.2H). In the hippocampus, the antisense-b/c probe strongly labeled granule neurons in the dentate gyrus (Fig.2I) and weakly labeled cells in the pyramidal layer (Fig.2J). Probe to NBCe1-A/B According to the data presented in Fig.2E-J, mrna encoding the B or C variant is present in brain. To distinguish between these two variants at the mrna level, we designed probes to the 97-bp sequence found in the 3' ORF of the A and B variants, but not in the C variant (Fig.1). Because we established that A variant mrna is nearly absent in brain, this 97-bp probe should primarily detect B-variant mrna. The antisense A/B

69 58 probe (Fig.2K), but not the sense probe (Fig.2L), exhibited a labeling pattern similar to that seen with the antisense B/C probe, although with lower intensity. One notable exception was labeling of the OB by the antisense B/C probe (Fig.2E) but not the antisense A/B probe (Fig.2K). The antisense A/B probe labeled the granule and Purkinje layers of the cerebellum (Fig.2M), and the dentate gyrus and pyramidal layer of the hippocampus (Fig.2N). The intensity of labeling by the A/B probe was less than that by the B/C probe a finding consistent with less B than C variant mrna. However, the different intensities could also reflect different probe lengths. Overall, the data are consistent with the presence of NBCe1-B, and possibly NBCe1-C mrna in rat brain, with pronounced localization in cerebellum and hippocampus. Expression of NBCe1 Variants in Tissue Homogenates from Different Brain Regions To examine the protein expression of NBCe1-B and -C in rat brain, we performed immunoblot analyses using two previously characterized NBCe1 antibodies (Bevensee et al., 2000) that distinguish between the unique 61 carboxy terminal residues of the C variant ( C), and the different 46 residues of the A and B variants ( A/B) (Fig.1). Because A-variant mrna appears to be absent or present in low levels in brain (see above), an A/B signal reflects B-variant expression predominantly. However, we can not definitively exclude NBCe1-A expression if NBCe1-A mrna in low abundance is undetectable by our probe. In immunoblot studies, we used either 1:500 A/B (Fig.3A) or 1:2K C (Fig.3B) to probe a commercially available membrane containing protein lysates from 12 different brain regions of adult rat. We also probed the membrane with an anti-

70 59 body to -actin ( -actin, 1:5K), and normalized the intensities of either A/B or C to -actin labeling ( A/B: -actin or C: -actin, respectively). Both NBC antibodies recognized proteins of the expected size for the NBCe1 variants (~130 kda), as well as some lower-molecular weight proteins that may be degraded NBCe1 (Schmitt et al., 1999; Bevensee et al., 2000). In addition, faint C labeling of proteins greater than 199 kda (Fig.3B) may represent multimers of NBCe1-C. Among the brain regions, the intensity of A/B labeling was more variable than that of C. We obtained similar results on a second commercially available rat-brain membrane. Although there was variability in protein loading per lane based on the -actin labeling, consistent findings in both membranes were relatively high levels of A/B labeling in the cerebellum and spinal cord (Fig.3A), and relatively high levels of C labeling in the hippocampus, medulla, and spinal cord (Fig.3B). In summary, both NBCe1-B and -C are expressed in many brain regions of adult rat. Localization of NBCe1 Variants in Cerebellum, Hippocampus, and Cortex Based on our in situ hybridization and immunoblot data that are consistent with NBCe1-B and -C expression in rat brain, we next used immunohistochemistry on brain slices from adult rat to explore localization of the proteins in the cerebellum, hippocampus, and cortex. In general, both A/B and C exhibited diffuse labeling throughout the brain and in the aforementioned three regions. In addition, however, each antibody also displayed some distinct labeling as described below. As described in the Experimental Procedures, we fixed rat brains in Bouin s fixative. Compared to other fixation and antibody labeling protocols that we tested, we found

71 60 our Bouin s protocol to yield the most consistent labeling patterns with minimum background labeling. However, we obtained similar antibody labeling patterns with A/B (Supplemental Fig. 1) and C (Supplemental Fig. 2) using brains from 4% PFA-perfused adult male rats. NBCe1 Expression in Cerebellum A/B (1:100) only weakly labeled cell layers in the cerebellum. Labeling in the molecular layer was slightly greater than that seen in the granule layer where the nuclei of granule cells are counterstained with blue bis-benzimide (Fig.4A,B). No labeling was observed in the absence of primary antibody as shown in the negative control image that is merged with the bis-benzimide labeling (-C (Merge), Fig.4C). A/B did not appreciably label calbindin-positive Purkinje neurons in the cerebellum (Fig.4D-F). However, we occasionally observed weak A/B labeling of some cells in the Purkinje layer of some lobes (not shown). In contrast, C (1:500) labeled cells throughout the cerebellum (Fig.5A,B). No labeling was observed in the absence of primary antibody (Fig.5C). C labeling was particularly apparent around the calbindin-positive Purkinje neurons (Fig.5D-F). The labeling was specific because preabsorbing C with 100 μg/ml of the corresponding antigenic fusion protein used to generate the antibody (Bevensee et al., 2000) greatly reduced the labeling intensity (Fig.5G-I). Such competition was not observed with A/B fusion protein (not shown). The rim-like labeling of C is consistent with NBC expression either at the plasma membrane of the Purkinje neurons, or in cell bodies/processes from surrounding glial

72 61 glial cells (e.g., Bergmann glia). To explore these possibilities further, we performed colocalization studies with C and an antibody to the glial cell-specific glutamate transporter GLAST ( GLAST). As shown at both low and higher magnifications (Fig.5J-O), there was a strong correlation between C and GLAST labeling in both the molecular and Purkinje layers of the cerebellum. Both antibodies displayed rim-like labeling around the Purkinje neurons. Thus, the C labeling in the Purkinje cell layer is likely that of Bergmann glia data consistent with results from the in situ hybridization study (Fig.2H). On the other hand, we can not exclude the possibility that at least some of the C labeling in the Purkinje cell layer is neuronal. Indeed, we found some colocalized labeling (not shown) of C and an antibody to the neuron-localized 1 subunit of the Na + pump (Watts et al., 1991; Peng et al., 1997). NBCe1 Expression in Hippocampus Compared with the cerebellar results, A/B (1:100) weakly labeled cells in the hippocampus, although in a more distinct pattern. More specifically, the antibody labeled cells within all CA pyramidal layers and the dentate gyrus (Fig.6A-D). As seen at higher magnification of the CA1 layer, A/B labeled neurons in an intracellular and punctate pattern (Fig.6E-H). A similar A/B labeling pattern was seen in other CA layers, as well as the dentate gyrus. As we saw with C labeling, A/B labeling was specific because preabsorbing A/B with 50 μg/ml 1 of the corresponding antigenic fusion protein used to generate the antibody (Bevensee et al., 2000) greatly reduced the labeling intensity 1 We performed A/B and C competition experiments using 20, 50, and 100 μg/ml of the appropriate fusion protein. Overall, the appropriate fusion protein reduced A/B labeling to a greater extent than C la-

73 62 (Fig. 6I-L) 2. A/B competition was not observed with the C fusion protein (not shown). The A/B labeling was clearly neuronal because the cells co-labeled with an antibody to the neuronal protein MAP2 (Fig.6M-P), but not with an antibody to the glial protein GFAP (Fig.6Q-T). In other colocalization studies (not shown), the A/B-labeled neurons were also labeled with an antibody to the neuronal nuclear protein NeuN. C (1:500) also labeled the pyramidal layers and the dentate gyrus of the hippocampus, although the overall labeling was more intense (Fig.7A-D) than the labeling observed with A/B. As seen at higher magnification, the C labeling pattern in the pyramidal layers and dentate gyrus was different than the A/B labeling pattern. In particular, C displayed a rim-like labeling pattern around neurons in the pyramidal cell layer (Fig.7E-H), in contrast to the intracellular, punctate labeling pattern by A/B (Fig.6E-H). In the dentate gyrus, C predominantly labeled cells closer to the center of the hilus (Fig.7I-K), in contrast to general labeling in the granule cell layer. To evaluate the cell type(s) labeled with C, we performed colocalization studies with either an antibody to the neuronal protein NeuN or GLAST. As shown in Fig.7L-N, NeuN-positive neurons displayed rim-like C labeling. Based on colocalization studies with MAP2 (not shown), C labeling was also evident around the neurons. However, the C- NeuN colocalization was moderate, and some of the neurons displayed no or weak C labeling (see arrows). The colocalization of GLAST and C was much stronger in both the pyramidal layers (Fig.7O-Q) and the dentate gyrus (Fig.7R-T). The data are consistent with beling. A/B labeling was somewhat reduced by 20 μg/ml and nearly eliminated by 50 and 100 μg/ml fusion protein. However, C labeling was considerably reduced only at 100 μg/ml fusion protein. 2 Curiously, there appeared to be more labeling in the cerebellar white matter (not shown) with A/B and its competitor.

74 63 C labeling of glial cells predominantly in the pyramidal layers and dentate gyrus of the hippocampus. Similar hippocampal results were obtained with triple labeling with C, MAP2, and GLAST. NBCe1 Expression in Cortex Similar to that seen in the hippocampus, A/B (1:100) exhibited punctate, intracellular labeling in the cortex (Fig.8A,D). The labeled cells appeared to be neurons because a majority of them co-labeled with the antibody to the neuronal protein MAP2 (Fig.8B, C, see arrows). In contrast, C (1:500) labeling was more diffuse and more apparent around neuronal cell bodies (Fig.8E,H,I,L) that were NeuN-positive (Fig.8F,G). The labeling was more glial than neuronal because of strong co-localization with the glial marker GLAST (Fig.8J,K). High-resolution NBCe1 Expression in Cerebellum Immunoelectron microscopy showed the subcellular localization of C and A/B in the Purkinje layer of cerebellum. Colloidal gold particles representing C labeling were preferentially associated with the plasma membrane of glial cells (Fig.9A,B), identified through their shape or occasional apposition to the outer surface of capillaries. Although C primarily labeled the plasma membrane of glial cells, we cannot exclude the possibility that C may have also labeled the plasma membrane of neurons in some instances. However, C labeling followed the very complex periphery of the glial cells and included long stretches of membrane. Numerous small nerve processes nearby did not express C. Occasional gold particles close to the periphery of some thicker nerve fiber

75 64 appeared related to adjacent glial cells. A/B on the contrary was expressed in the cytoplasm of some nerve cell fibers (Fig.9C), but many large and all small fibers were unlabeled. Gold particles in the labeled fibers were often in filamentous areas in the cytoplasm and occasionally in the endoplasmic reticulum (ER). DISCUSSION In the present study, we examined the localization and expression of the three NBCe1 variants (A, B, and C) in rat brain. According to in situ hybridization data, A variant mrna appears to be absent, whereas B and possibly C variant mrna is present throughout brain, particularly in the cerebellum and hippocampus. However, we cannot exclude the possibility that A-variant mrna is present at a very low level, which may be difficult to detect with our short probe that is unavoidably near the detection threshold. In an important control experiment however, the A probe does detect NBC mrna in kidney. In support of our in situ hybridization results, we have had difficulty PCR amplifying NBCe1-A cdna from rat brain. Based on the near absence or low level of A-variant mrna in brain, NBCe1- A/B labeling in our immunohistochemistry studies reflects B variant expression predominantly. A/B clearly labeled neurons, whereas C appeared to label glial cells more so than neurons in the hippocampus and cortex. Furthermore, A/B labeling was intracellular and punctate compared to C labeling that surrounded neurons. Intracellular expression of NBCe1-B in neurons and plasma-membrane expression of NBCe1-C in glial cells was supported by immunogold electron microscopy data obtained from the cerebellum.

76 65 An intriguing observation was the intracellular localization profile of NBCe1 in neurons. According to the electron microscopy data (Fig.9C), A/B labeled cytoplasmic regions containing filamentous structures, and in some cases, the ER. Although it is not clear if the filaments and ER are labeled, the transporter may be in transit by itself or as some complex. While the functional role of a cytoplasmic NBC is unknown, an NBC in the membrane of an organelle such as the ER may play a role in organellar ph regulation similar to that seen with NHE6-9 (Nakamura et al., 2005). On the other hand, it is somewhat surprising if NBCe1 is targeted to a cytoplasmic compartment in neurons because NBCe1-A/B is targeted to basolateral membranes of other cells types, including the kidney proximal tubule and the pancreatic ducts (Romero et al., 2004). Alternatively, the plasma membrane may be the target site in neurons, and NBCe1 expression is normally low and/or requires a stimulus for increased surface expression. It is important to point out that the level or mode of NBCe1 activity may be different for the transporter in the cytoplasm vs. plasma membrane. For instance, polycystin-2 (PKD2) functions as an intracellular Ca 2+ -release channel in the ER, where it predominantly resides, but as a receptor-operated nonselective cation channel in the plasma membrane (Tsiokas et al., 2007). In comparing results from the in situ hybridization and immunohistochemistry data, the similarities exceed the differences. The similarities support the expression of both NBCe1-B and NBCe1-C in brain. For example, the major brain regions (e.g., cerebellum and hippocampus) labeled with the mrna probes were also labeled with the two antibodies. In fact, each region labeled with the B/C probe was also labeled with one or both of the antibodies. For example, labeling of Bergmann glia in the cerebellum by the B/C

77 66 B/C probe (Fig.2H) is consistent with C labeling around the Purkinje neurons (Fig.5D) that co-localizes with GLAST labeling (Fig.5M). Labeling of hippocampal cells in the dentate gyrus by the B/C mrna probe (Fig.2I) and to a lesser extent the CA layers (Fig.2J) is consistent with A/B (Fig.6) and/or C labeling (Fig.7). In all likelihood, B/C probe-positive cells in the internal portion of the dentate gyrus hilus are glial precursor cells based on the pattern of C labeling and co-localization with the glial marker GLAST (Fig.7I-K,R-T). For the A/B probe (which detects NBCe1-B primarily in brain), labeling in the dentate gyrus and CA layers was also evident (Fig.2K,N) and consistent with A/B labeling (Fig.6). Regarding differences between the mrna and protein localization data, both the B/C and A/B probes weakly label the pyramidal layer compared to the dentate gyrus (Fig.2I,J,N), whereas both antibodies show pronounced labeling (Fig.6A and Fig.7A). Also, both probes labeled the granule layer of the cerebellum more so than the molecular layer (Fig.2E,H,K,M), whereas the two antibodies labeled the two layers to the same extent or the molecular layer to a greater extent (Fig.4 and Fig.5). Mismatches between mrna and protein levels are not uncommon, and may be due to differences in translation efficiency, protein turnover, etc. It is important to point out that there appears to be differences in NBC localization in situ vs. in vitro. For example, according to immunoblot data on microsomal protein from cultured cortical cells, the A/B variant is predominant in astrocytes vs. neurons, and the opposite is true for the C variant (Bevensee et al., 2000). We have more recently obtained similar results on cultured hippocampal cells (Majumdar, Williams, Bevensee;

78 67 unpublished observations). Thus, NBC expression profiles are different for cells in culture. Electrogenic NBC activity has been identified and characterized in mammalian astrocytes in situ (Grichtchenko and Chesler, 1994a,b), as well as cultured from several brain regions (O Connor et al.,1994; Brune et al., 1994; Bevensee et al., 1997a,b). Our data are consistent with NBCe1-C being responsible for such activity in situ, although NBCe1-B activity may play a larger role when astrocytes are cultured. To date, electrogenic NBC activity has not been reported in neurons, although we have recently found a population of cultured hippocampal neurons that exhibit such activity (Majumdar, Wu, McNicholas-Bevensee, Bevensee, unpublished observations). Based on our in situ hybridization and immunochemistry data, as well as the aforementioned immunoblot data, it would not be surprising to find NBCe1 expression at the plasma membrane of cultured neurons. As described in the introduction, results from early molecular and localization studies stemming from the cloning of NBCs are consistent with the expression of NBCe1 variants in brain. Both NBCe1-B and NBCe1-C but not NBCe1-A were cloned by homology from a rat-brain cdna library (Bevensee et al., 2000). Using polynucleotide probes and/or polyclonal antibodies that did not distinguish among the three NBCe1 variants, Schmitt et al. (2000) and Giffard et al. (2000) found NBCe1 widely distributed throughout the rat brain especially the cerebellum and hippocampus and localized to glial cells and some neurons. In addition, Giffard et al. (2000) found minimal expression of A variant mrna. Our results complement and extend these findings. For example, we also found minimal expression of A-variant mrna in our in situ hybridization studies,

79 68 but pronounced B- and/or C-variant mrna expression throughout the brain, particularly in the cerebellum and hippocampus. At the protein level, neurons were labeled with A/B, whereas glial cells were primarily labeled with C. More recently, Rickmann et al. (2007) reported on the expression profiles of the A and B variants in adult mouse brain using N-terminal antibodies. In contrast to our findings, as well as those of Giffard et al. (2000) working on rat, Rickmann et al. (2007) described the localization of NBCe1-A mrna in several brain regions (e.g. OB, cerebral cortex, hippocampus, and cerebellum) and expression of the protein in neuronal populations/regions including i) Purkinje cell bodies and dendrites of the cerebellum, ii) granule cells, non-pyramidal neurons, and apical cell dendrites in the hippocampus, and iii) pyramidal neurons (and some astroglial lamellae) in the cerebral cortex. The discrepancy between our A-variant data and those of the Rickmann group is not clear, although a species difference may be responsible. Rickmann et al. (2007) also reported that NBCe1-B mrna is present in the same brain regions as described above for NBCe1-A. However, the profile of protein expression is different. For example, the group s NBCe1-B antibody labeled i) Bergmann glia and Purkinje cell bodies (weakly) of the cerebellum, ii) non-pyramidal neurons, apical cell dendrites, and perivascular astroglial compartments in the hippocampus, and iii) pyramidal neurons (weakly) and astroglial processes (weakly) in the cerebral cortex. However, because this NBCe1-B antibody is also expected to recognize the C variant as well, the labeling patterns likely reflect B and/or C variant expression. Therefore, it is not surprising that one or both of our NBCe1 antibodies recognize most of these regions.

80 69 Consistent with our findings, the authors reported localization of NBCe1-B in the rough ER of cerebellar neurons. As described in the introduction, electrogenic NBC activity has been extensively studied in astrocytes, where the transporter can modulate neuronal activity through changes in ph i /ph o. However, the functional significance of electrogenic NBC activity in neurons is poorly understood. It is well known that neurons exhibit a decrease in ph i upon firing (in some preparations due to Ca 2+ -pump mediated exchange of intracellular Ca 2+ and extracellular H + following an elevation of intracellular Ca 2+ ) (McAlear and Bevensee, 2003; Chesler, 2003). The firing-evoked ph i decrease would be expected to stimulate the activity of an electrogenic NBC in neurons, thereby reducing the magnitude of the ph i decrease. Neuronal depolarization would also stimulate electrogenic NBC activity, and further retard the ph i decrease. An obvious question pertains to the importance of multiple NBCe1 variants in brain. Although the physiological significance of the three NBCe1 variants with different amino and/or carboxy termini is not fully known, these regions may influence expression at the plasma membrane or regulation by signaling molecules or binding proteins. One of the distinguishing features of the C variant compared to the A and B variants is a PDZbinding motif at the carboxy terminus (Bevensee et al., 2000). This motif may confer C variant-specific regulation by PDZ proteins in certain cells. Changes in NBCe1 expression have been associated with development and disease, although little is known about the involvement of specific variants. For example, general NBCe1 expression in rat brain increases progressively from embryonic day 16 until postnatal day 105, particularly in the brainstem-diencephalon and cerebellum

81 70 compared to the cortex (Douglas et al., 2001). NBCe1 expression has also been correlated to neuropathologies such as epilepsy. For example, general NBCe1 immunoreactivity is elevated in the hippocampus of the gerbil from 30 min to 3 h after the onset of seizure, but then returns to the pre-seizure level at 6 h (Kang et al., 2002). In light of the results herein, further studies are required to assess the influence of different NBCe1 variants on ph i regulation during development or diseases such as epilepsy. CONCLUSION In conclusion, we have examined mrna localization and protein expression of the three NBCe1 variants in rat brain. According to our studies, NBCe1-B and NBCe1-C, but not NBCe1-A are the predominant NBCe1 variants expressed in brain, particularly in the cerebellum, hippocampus, and cortex. In the hippocampus and cortex, NBCe1-B is localized intracellularly in neurons. NBCe1-C is expressed primarily in glia surrounding neurons and possibly neurons within the neuropil of the pyramidal cell and granule layers of the hippocampus, the Purkinje cell layer of the cerebellum, and the cortex. The general role of NBCs in the brain is to regulate ph i and ph o. Distinct expression profiles of NBCe1 variants may contribute to specific ph physiologies seen with development or disease state.

82 71 ACKNOWLEDGMENTS This work was supported by NIH/NINDS NS (M.O.B), AHA predoctoral fellowship B (D.M.), and the UAB Neuroscience Core Facilities NS This work was supported by NIH Neuroscience Blueprint Core Grant NS57098 to the University of Alabama at Birmingham. We thank Cecelia B. Latham for her expertise in immunohistochemistry, Albert Tousson for assistance with the confocal microscopy at the high resolution imaging facility at UAB, Dr. Lori L. McMahon for evaluating our immunolabeling and providing valuable suggestions, Dr. Michael J. Caplan (Yale University) for providing thoughtful comments regarding our electron microscopy data, Dr. Peter R. Smith and his laboratory for use of his fluorescence microscope, Dr. Buffie Clodfelder- Miller for taking pictures of the in situ labeling, Dr. Jim Schafer for evaluating our in situ hybridization data from kidney, the laboratory of Dr. John J. Hablitz for providing the GLAST antibody, the laboratory of Dr. David F. Crawford for providing the -actin antibody, and laboratory of Dr. Marcas Bamman for help with the densitometry analysis, and Dr. Carmel M. McNicholas-Bevensee for valuable comments regarding the manuscript. REFERENCES Aalkjær C, Frische S, Leipziger J, Nielsen S, Praetorius J (2004) Sodium coupled bicarbonate transporters in the kidney, an update. Acta Physiol Scand 181: Abuladze N, Lee I, Newman D, Hwang J, Boorer K, Pushkin A, Kurtz I (1998a) Molecular cloning, chromosomal localization, tissue distribution, and functional expression of the human pancreatic sodium bicarbonate cotransporter. J Biol Chem 273: Abuladze N, Lee I, Newman D, Hwang J, Pushkin A, Kurtz I (1998b) Axial heterogeneity of sodium-bicarbonate cotransporter expression in the rabbit proximal tubule. Am J Physiol 274:F628-F633.

83 72 Berger UV, Hediger MA (2001) Differential distribution of the glutamate transporters GLT-1 and GLAST in tanycytes of the third ventricle. J Comp Neurol 433: Bevensee MO, Apkon M, Boron WF (1997a) Intracellular ph regulation in cultured astrocytes from rat hippocampus. II. Electrogenic Na/HCO 3 cotransport. J Gen Physiol 110: Bevensee MO, Weed RA, Boron WF (1997b) Intracellular ph regulation in cultured astrocytes from rat hippocampus. I. Role of HCO 3. J Gen Physiol 110: Bevensee MO, Schmitt BM, Choi I, Romero MF, Boron WF (2000) An electrogenic Na/HCO 3 cotransporter (NBC) with a novel C terminus, cloned from rat brain. Am J Physiol Cell Physiol 278:C1200-C1211. Brune T, Fetzer S, Backus KH, Deitmer JW (1994) Evidence for electrogenic sodiumbicarbonate cotransport in cultured rat cerebellar astrocytes. Pflügers Arch 429: Burnham CE, Amlal H, Wang Z, Shull GE, Soleimani M (1997) Cloning and functional expression of a human kidney Na + :HCO 3 cotransporter. J Biol Chem 272: Chesler M (1990) The regulation and modulation of ph in the nervous system. Prog Neurobiol 34: Chesler M (2003) Regulation and modulation of ph in the brain. Physiol Rev 83: Chesler M, Kaila K (1992) Modulation of ph by neuronal activity. Trends Neurosci 15: Choi I, Romero MF, Khandoudi N, Bril A, Boron WF (1999) Cloning and characterization of a human electrogenic Na + -HCO 3 cotransporter isoform (hhnbc). Am J Physiol 276:C576-C584. Deitmer JW (1992) Evidence for glial control of extracellular ph in the leech central nervous system. Glia 5: Deitmer JW, Schlue W-R (1987) The regulation of intracellular ph by identified glial cells and neurones in the central nervous system of the leech. J Physiol (Lond ) 388: Deitmer JW, Schlue W-R (1989) An inwardly directed electrogenic sodium-bicarbonate cotransport in leech glial cells. J Physiol (Lond ) 411: Douglas RM, Schmitt BM, Xia Y, Bevensee MO, Biemesderfer D, Boron WF, Haddad GG (2001) Sodium-hydrogen exchangers and sodium-bicarbonate co-transporters: ontogeny of protein expression in the rat brain. Neuroscience 102:

84 73 Giffard RG, Papadopoulos MC, van Hooft JA, Xu L, Giuffrida R, Monyer H (2000) The electrogenic sodium bicarbonate cotransporter: developmental expression in rat brain and possible role in acid vulnerability. J Neurosci 20: Grichtchenko II, Chesler M (1994a) Depolarization-induced acid secretion in gliotic hippocampal slices. Neuroscience 62: Grichtchenko II, Chesler M (1994b) Depolarization-induced alkalinization of astrocytes in gliotic hippocampal slices. Neuroscience 62: Kang TC, An SJ, Park SK, Hwang IK, Suh JG, Oh YS, Bae JC, Won MH (2002) Alterations in Na + /H + exchanger and Na + /HCO 3 cotransporter immunoreactivities within the gerbil hippocampus following seizure. Brain Res Mol Brain Res 109: Katsura K, Siesjo B (1998) Acid-base metabolism in ischemia. In: ph and Brain Function (Kaila K, Ransom B, eds), pp New York: Wiley-Liss, Inc. Majumdar D, Shacka JJ, Williams JB, Berger UV, Schultz K, Harkins LE, Roth KA, Bevensee MO (2006) Localization of electrogenic Na/bicarbonate cotransporter NBCe1 variants in rat brain. Soc.Neurosci.Abst. 33, # Maunsbach AB, Vorum H, Kwon TH, Nielsen S, Simonsen B, Choi I, Schmitt BM, Boron WF, Aalkjær C (2000) Immunoelectron microscopic localization of the electrogenic Na/HCO 3 cotransporter in rat and Ambystoma kidney. J Am Soc Nephrol 11: McAlear SD, Bevensee MO (2003) ph regulation in non-neuronal brain cells and interstitial fluid. In: Non-neuronal cells of the nervous system: Function and dysfunction (Hertz L, ed), pp Amsterdam: Elsevier. Nakamura N, Tanaka S, Teko Y, Mitsui K, Kanazawa H (2005) Four Na + /H + exchanger isoforms are distributed to Golgi and post-golgi compartments and are involved in organelle ph regulation. J Biol Chem 280: O Connor ER, Sontheimer H, Ransom BR (1994) Rat hippocampal astrocytes exhibit electrogenic sodium-bicarbonate cotransport. J Neurophysiol 72: Peng L, Martin-Vasallo P, Sweadner KJ (1997) Isoforms of Na, K - ATPase and subunits in the rat cerebellum and in granule cell cultures. J Neurosci 17: Ransom BR (1992) Glial modulation of neural excitability mediated by extracellular ph: A hypothesis. Prog Brain Res 94: Ransom BR (2000) Glial modulation of neural excitability mediated by extracellular ph: a hypothesis revisited. Prog Brain Res 125:

85 74 Rickmann M, Orlowski B, Heupel K, Roussa E (2007) Distinct expression and subcellular localization patterns of Na + /HCO 3 cotransporter (SLC 4A4) variants NBCe1-A and NBCe1-B in mouse brain. Neuroscience 146: Romero MF, Fong P, Berger UV, Hediger MA, Boron WF (1998) Cloning and functional expression of rnbc, an electrogenic Na + -HCO 3 cotransporter from rat kidney. Am J Physiol 274:F425-F432. Romero MF, Fulton CM, Boron WF (2004) The SLC4 family of HCO 3 transporters. Pflügers Arch 447: Romero MF, Hediger MA, Boulpaep EL, Boron WF (1997) Expression cloning and characterization of a renal electrogenic Na + /HCO 3 cotransporter. Nature 387: Rose CR, Ransom BR (1998) ph regulation in mammalian glia. In: ph and Brain Function (Kaila K, Ransom B, eds), pp New York: Wiley-Liss, Inc. Schmitt BM, Berger UV, Douglas RM, Bevensee MO, Hediger MA, Haddad GG, Boron WF (2000) Na/HCO 3 cotransporters in rat brain: expression in glia, neurons, and choroid plexus. J Neurosci 20: Schmitt BM, Biemesderfer D, Romero MF, Boulpaep EL, Boron WF (1999) Immunolocalization of the electrogenic Na + /HCO 3 cotransporter in mammalian and amphibian kidney. Am J Physiol 276:F27-F36. Tsiokas L, Kim S, Ong EC (2007) Cell biology of polycystin-2. Cell Signal 19: Watts AG, Sanchez-Watts G, Emanuel JR, Levenson R (1991) Cell-specific expression of mrnas encoding Na +, K +, - ATPase - and - subunit isoforms within the rat central nervous system. Proc Natl Acad Sci USA 88:

86 Fig.1. In situ hybridization probes and rabbit polyclonal antibodies for localizing NBCe1- A, -B, and -C variants in rat brain. The cdna and protein diagrams are not drawn to scale. Antisense and sense A probes recognize the unique 5' ORF (bp 1-123) that encodes the unique amino-terminal 41 residues of the A variant. Antisense and sense B/C probes recognize the different 5' ORF (bp 1-255) that encodes the different amino-terminal 85 residues of the B and C variants. Antisense and sense A/B probes recognize the 97-bp region near the 3' end of the ORF found in the A and B variants, but not the C variant. Two different rabbit polyclonal antibodies previously characterized (Bevensee et al., 2000) distinguish between the C-terminal 46 residues of the A and B variants ( A/B antibody) and the unique C-terminal 61 residues of the C variant ( C antibody). 75

87 Fig.2. Localization of mrna encoding NBCe1 variants in rat brain. (A-D) A specific antisense probe to the 5' ORF of NBCe1-A. Little NBCe1-A mrna was detected throughout brain (A), including the cerebellum (B) and the hippocampus (C). As a positive control, the antisense probe labeled the outer cortex and proximal tubules (PT) near the glomerulus (G) of rat kidney (D). (E-J) A specific antisense probe to the 5' ORF of NBCe1-B/C variants. NBCe1-B/C mrna was present throughout brain, particularly in the cerebellum and hippocampus, as well as the OB shown in the inset (E). Minimal labeling was obtained with the sense probe (F). NBCe1-B/C mrna was prominent in (i) the granular and Purkinje layers of the cerebellum (G, H), especially Bergmann glia in the Purkinje layer (H), and (ii) the dentate gyrus of the hippocampus (I). NBCe1-B/C mrna was less pronounced in the pyramidal layer of the hippocampus (J). (K-N) A specific antisense probe to the 3' ORF of NBCe1-A/B variants. NBCe1-A/B mrna was present throughout brain (K), as well as cerebellum (M) and hippocampus (N), in a pattern similar to that seen with the B/C antisense probe. Minimal labeling was obtained with the sense probe (L). Numbers above scale bars in panels are in micrometers. 76

88 Fig. 3. Expression of NBCe1variants in tissue homogenates from different regions of rat brain. (A, top) A commercially available blot containing total protein from various brain regions probed with 1:500 A/B. NBCe1-A/B is ~130 kda. (A, bottom) The same commercial blot was reprobed with 1:5000 -actin. Intensity of A/B labeling normalized to that of -actin is shown as A/B: -actin. (B) The blot probed with 1:2000 C. NBCe1-C is ~130 kda. Intensity of C labeling normalized to that of -actin is shown as C: -actin. 77

89 Fig. 4. NBCe1-A/B expression in rat cerebellum. (A-C) Fluorescence microscopy image of weak A/B labeling in the cerebellum of a sagittal brain section. A/B labeling of the molecular layer was slightly greater than the granule layer (A). Images of A/B labeling and nuclear bis benzimide staining are overlaid (B). No labeling was seen in the negative control without primary antibody (C). Scale bar = 20 μm for A-C. (D-F) Fluorescence microscopy image of double labeling by A/B (D) and Calbindin (1:1K) (E) in the cerebellum of a sagittal brain section. There was minimal colocalization seen in the merged figure (F). Scale bar = 20 μm for D-F. 78

90 Fig. 5. NBCe1-C expression in rat cerebellum. (A-C) Fluorescence microscopy image of C labeling in the cerebellum of a sagittal brain section. C labeled the granule and the Purkinje cell layers (A). Images of C labeling and nuclear bis benzimide staining are overlaid (B). No labeling was seen in the negative control without primary antibody (C). Scale bar = 20 μm for A-C. (D-F) Fluorescence microscopy image of double labeling of C and Calbindin (1:1K) in the cerebellum of a sagittal brain section. Colocalization of C (D) and Calbindin (E) was consistent with NBCe1-C expression around the Purkinje neurons as seen in the merged image (F). Scale bar = 10 μm for D-F. (G-I) Fluorescence microscopy image showing the specificity of C labeling in the cerebellum of a sagittal brain section. C labeling (G) was reduced by preabsorption with 100 μg/ml fusion protein (H,I). Scale bar = 20 μm for G-I. (J-O) Fluorescence microscopy images of double labeling by C and GLAST (1:500) at lower (J-L) and higher (M-O) magnifications in the cerebellum of a sagittal brain section. Colocalization of C (J,M) and GLAST (K,N) was consistent with NBCe1-C expression in glia within the Purkinje cell layer as seen in the merged image (L, O). Scale bars = 20 μm for J-L, and 10 μm for M-O. 79

91 Fig. 6. NBCe1-A/B expression in rat hippocampus. (A-D) Fluorescence microscopy image of intracellular A/B labeling in the hippocampus of a sagittal brain section. A/B (A) exhibited punctate intracellular labeling in the pyramidal cell layer and the dentate gyrus of the hippocampus. Images of A/B labeling (A) and nuclear bis benzimide staining (B) are overlaid (C). No labeling was seen in the negative control without primary antibody (D). Scale bar = 100 μm for A-D. (E-H) Fluorescence microscopy image of intracellular A/B labeling in the CA1 layer of a sagittal brain section. Images of A/B labeling (E) and nuclear bis benzimide staining (F) are overlaid (G). No labeling was seen in the negative control without primary antibody (H). Scale bar = 10 μm for E-H. (I- L) Fluorescence microscopy image showing the specificity of A/B labeling in the CA1 layer of a sagittal brain section. A/B labeling (I,J) was reduced by preabsorption with 50 μg/ml fusion protein (K,L). Scale bar = 20 μm for I-L. (M-P) Fluorescence microscopy image of intracellular A/B labeling in the CA3a layer of a sagittal brain section. Colocalization of A/B (M) and MAP2 (1:100K) (N) was consistent with NBCe1-A/B expression in the pyramidal cells as seen in the merged image (O). No labeling was seen in the negative control without primary antibody (P). Scale bar = 20 μm for M-P. (Q-T) Fluorescence microscopy image of distinct A/B and GFAP (1:5K) labeling in the CA3a layer of sagittal brain section. A/B (Q) and GFAP (R) did not colocalize a finding consistent with the absence of NBCe1-A/B in astrocytes as seen in the merged image (S). No labeling was seen in the negative control without primary antibody (T). Scale bar = 20 μm for Q-T. 80

92 Fig. 7. NBCe1-C expression in rat hippocampus. (A-D) Fluorescence microscopy image of C labeling in the hippocampus of a sagittal brain section. C (A) labeled the pyramidal cell layer and dentate gyrus. Images of C labeling (A) and nuclear bis benzimide staining (B) are overlaid (C). No labeling was seen in the negative control without primary antibody (D). Scale bar = 100 μm for A-D. (E-H) Fluorescence microscopy image of rim-like C labeling in the CA1 layer of the hippocampus of a brain section. Images of C labeling (E) and nuclear bis benzimide staining (F) are overlaid (G). No labeling was seen in the negative control without primary antibody (H). Scale bar = 10 μm for E-H. (I- K) Confocal microscopy image of C labeling in the inner hilar portion of the dentate gyrus of a brain section. Images of C labeling (I) and nuclear bis benzimide staining (J) are overlaid (K). Scale bar = 100 μm for I-K. (L-N) Confocal microscopy image of double labeling by C and NeuN (1:5K) in the CA1 layer of the hippocampus of a sagittal brain section. Colocalization of C (L) and NeuN (M) was consistent with NBCe1-C expression around the pyramidal neurons as seen in the merged image (N). Scale bar = 50 μm for L-N. (O-T) Fluorescence microscopy images of double labeling of C and GLAST (1:500) in the hippocampus of a sagittal brain section. Colocalization of C (O, R) and GLAST (P, S) in the CA1/CA2 layer (O-Q) and granule layer of the dentate gyrus (R-T) was consistent with NBCe1-C expression in glia as seen in the merged images (Q, T). Scale bar = 10 μm for (O-Q) and 20 μm for (R-T). 81

93 Fig. 8. NBCe1-A/B and -C expression in the rat cortex. (A-D) Fluorescence microscopy image of A/B labeling in the cortex of a sagittal brain section. Colocalization of A/B (A) and MAP2 (1:100K) (B) was consistent with NBCe1-A/B expression in the cortical neurons as seen in the merged image (C). Strong colocalization was evident in several neurons identified by the white arrows. No labeling was seen in the negative control without primary antibody (D). Scale bar = 20 μm for A-D. (E-H) Fluorescence microscopy image of C labeling in the cortex of a sagittal brain section. Colocalization of C (E) and NeuN (1:5K) (F) was consistent with NBCe1-C expression around the cortical neurons as seen in the merged image (G) No labeling was seen in the negative control without primary antibody (H). Scale bar = 20 μm for E-H. (I-L) Fluorescence microscopy image of distinct C and GLAST (1:500) labeling in the cortex of a sagittal brain section. C (I) and GLAST (J) colocalized as seen in the merged image (K) a finding consistent with the presence of NBCe1-C in glia surrounding neurons. No labeling was seen in the negative control without primary antibody (L). Scale bar = 20 μm for I-L. 82

94 Fig. 9. NBCe1-C and -A/B expression in the Purkinje layer of rat cerebellum. (A-C) Immunoelectron micrographs demonstrating the subcellular localization of NBCe1-C (A,B) and NBCe1-A/B (C) at the subcellular level. The colloidal gold particles conjugated to the secondary antibody (arrowheads) mark C labeling of the plasma membrane of a glial cell and are located either directly over or immediately adjacent to the glial membrane profile (A,B). The inset in the bottom, right corner of panel A shows a zoomed-in view of the region identified by the dotted rectangle. This glial cell (G) has a complex shape as illustrated in the cross-section (B) where the cytoplasm is labeled blue. Little or no C label is associated with adjacent nerve cell profiles (N). In contrast, A/B is preferentially located in the cytoplasm of some, but not all, nerve fibers (N) observed in the section (C). Scale bar = 1 μm for A-C. 83

95 Supplemental Figure 1. NBCe1-A/B expression in rat brain perfusion fixed with 4% PFA. (A-D) Fluorescence microscopy image of intracellular A/B labeling in the hippocampus of a sagittal brain section. A/B (A) exhibited punctate intracellular labeling in the pyramidal cell layer and the dentate gyrus of the hippocampus. Images of A/B labeling (A) and nuclear bis benzimide staining (B) are overlaid (C). No labeling was seen in the negative control without primary antibody (D). Scale bar = 100 μm for A-D. (E-H) Fluorescence microscopy image of intracellular A/B labeling in the CA1 layer of a sagittal brain section. Images of A/B labeling (E) and nuclear bis benzimide staining (F) are overlaid (G). No labeling was seen in the negative control without primary antibody (H). Scale bar = 10 μm for E-H. (I-L, M-P) Fluorescence microscopy image of A/B labeling in the cerebellum of a sagittal brain section seen at lower (I-K) and higher (M-O) magnifications. A/B labeling of the molecular layer was slightly greater than the granule layer. Also occasional labeling in the Purkinje cell layer was seen (I,M). Images of A/B labeling and nuclear bis benzimide staining are overlaid (K,O). No labeling was 84

96 seen in the negative control without primary antibody (L,P). Scale bar = 20 μm for I-L and M-P. (Q-T) Fluorescence microscopy image of A/B labeling in the cortex of a sagittal brain section. A/B labeling (Q) was consistent with NBCe1-A/B expression in cortical neurons based on the MAP2 labeling pattern shown in the manuscript. No labeling was seen in the negative control without primary antibody (T). Scale bar = 20 μm for Q- T. Overall, A/B labeling of PFA-fixed tissue was very similar to A/B labeling of Bouin s- fixed tissue described in the manuscript, with one exception. In the Purkinje cell layer of the cerebellum, we observed a higher percentage of A/B-labeled cells in the PFA-fixed vs. Bouin s-fixed tissue. Methods. The exposed heart of an anesthetized adult male Sprague-Dawley rat ( g) was pierced with an 18-gauge needle attached to perfusion tubing. The systemic circulation was flushed first with ~300 mls of 0.9% saline containing the vasodilator NaNO 2, following by ~300 mls of 4% PFA/PBS, which stiffened the rat. The excised brain was cut in half and then placed in a phosphate-buffered 4% PFA, (diluted from a 20% stock of formalin from Fisher Scientific, Fair Lawn, NJ, USA) and kept overnight at 4ºC. Prior to the overnight incubation, both the external and internal surfaces of the perfusion-fixed brain were white and free of blood. Subsequently, the brain was rinsed with 70% ethanol before embedding in paraffin. Five-micrometer sagittal sections were cut using a HM 355S microtome (Walldorf, Germany), and then mounted onto glass slides. Immunohistochemical labeling with A/B was then performed as described in the manuscript. 85

97 86 Supplemental Figure 2. NBCe1-C expression in rat brain perfusion fixed with 4% PFA. (A-D) Fluorescence microscopy image of C labeling in the hippocampus of a sagittal brain section. C (A) labeled the pyramidal cell layer and dentate gyrus. Images of C labeling (A) and nuclear bis benzimide staining (B) are overlaid (C). No labeling was seen in the negative control without primary antibody (D). Scale bar = 100 μm for A-D. (E-H) Fluorescence microscopy image of rim-like C labeling in the CA1 layer of the hippocampus. Images of C labeling (E) and nuclear bis benzimide staining (F) are overlaid (G). No labeling was seen in the negative control without primary antibody (H). Scale bar = 10 μm for E-H. (I-L, M-P) Fluorescence microscopy image of C labeling in the cerebellum of a sagittal brain section at lower (I-K) and higher (M-O) magnifications. C labeled the molecular, granule and the Purkinje cell layers (I,M). Images of C labeling and nuclear bis benzimide staining are overlaid (K, O). No labeling was seen in the negative control with-

98 out primary antibody (L,P). Scale bar = 20 μm for I-L, and 10 μm for M-P. (Q-T) Fluorescence microscopy image of C labeling in the cortex of a sagittal brain section. C (Q) labeling was consistent with NBCe1-C expression around the cortical neurons based on NeuN and GLAST labeling patterns shown in the manuscript. Images of C labeling and nuclear bis benzimide staining (R) are overlaid (S). No labeling was seen in the negative control without primary antibody (T). Scale bar = 20 μm for Q-T. Overall, C labeling was very similar in PFA- vs. Bouin s- fixed tissue. Methods. See Supplemental Fig

99 88 EVIDENCE FOR A FUNCTIONAL ELECTROGENIC Na/BICARBONATE COTRANSPORTER IN A POPULATION OF NEURONS CULTURED FROM RAT HIPPOCAMPUS DEBESHI MAJUMDAR, JIANPING WU, CARMEL M. MCNICHOLAS AND MARK O. BEVENSEE In preparation for (Journal of Neuroscience) Format adapted for dissertation

100 89 ABSTRACT In the central nervous system (CNS), regulation of intracellular ph (ph i ) and extracellular ph (ph o ) by acid-base transporters is crucial for proper brain function. The predominant HCO 3 -dependent acid-base transporter in glial cells is an electrogenic Na/Bicarbonate Cotransporter (NBC) that contributes to the recovery of ph i from an acid load. According to molecular data, neurons also express electrogenic NBC (e.g., NBCe1). However, a functional electrogenic NBC has not been identified or characterized in any neuronal preparation. In the present study, we used the patch-clamp technique in both whole-cell and perforated-patch modes to assess the function of an electrogenic NBC in neurons cultured from the rat hippocampus. According to our results, a population of neurons exhibited electrogenic NBC activity as evident by a Na + -dependent, HCO 3 - induced mean outward current of ~7.3 pa, and a HCO 3 -dependent, Na + -removal induced mean inward current of 2.1 pa for neurons bathed in either 5% CO 2 /24 mm HCO 3 or 20% CO 2 /96 mm HCO 3 solutions. 500 M DIDS inhibited the Na + -dependent, HCO 3 - induced outward current by ~60%. According to results from immunohistochemistry studies with antibodies to NBCe1 variants, all the cultured neurons expressed both NBCe1- A/B and -C in excitatory and inhibitory neurons. Thus, the heterogeneity of NBC activity was not due to differences in NBCe1 expression. An electrogenic NBC in neurons is expected to regulate both ph i and ph o, and may be neuroprotective under excessive firing events such as epilepsy and spreading depression. KEYWORDS Intracellular ph, acid-base, patch-clamp, neuron, bicarbonate

101 90 ABBREVIATED TITLE Functional electrogenic NBC in a population of rat hippocampal neurons INTRODUCTION The regulation of ph in the CNS is crucial because many processes such as ion channel activity and neuronal firing are sensitive to changes in both ph i and ph o (Tombaugh and Somjen, 1996; Chesler, 2003). In order to regulate ph i, cells in the CNS have evolved plasma-membrane acid-base transporters, which can be categorized as either acid loaders or acid extruders. The first electrogenic NBC in the CNS was identified in the giant neuropile leech glial cells (Deitmer and Schlue, 1989a; Deitmer and Schlue, 1989b; Deitmer and Schlue, 1987), and this transporter is thought to function as the major acid extruding mechanism in both invertebrate and mammalian glia. In contrast, the major HCO 3 -dependent acid extruder in neuronal preparations is the Na-driven Cl-HCO 3 Exchanger (NDCBE), which was first identified in the squid giant axon (Boron and De Weer, 1976; Boron and Russell, 1983; Boron, 1985) and the snail neuron (Thomas, 1977). After the first electrogenic NBC (NBCe1) was cloned by Romero et al. (Romero et al., 1997) and subsequent members of the bicarbonate transporter superfamily were identified, investigators used molecular tools (e.g., crna probes and antibodies) to characterize the localization profiles of the transporters. As described in the Introduction, NBCe1 has three splice variants that differ at the amino and/or carboxy termini (Romero et al., 1998; Bevensee et al., 2000; Choi et al., 1999). According to early localization stu-

102 91 studies with crna probes and antibodies that were expected to recognize all three NBCe1 splice variants, in regions such as the rat hippocampus, cortex, and striatum, NBCe1 was not only found in glia as expected, but also in neurons (Schmitt et al., 2000). Neuronal expression of NBCe1 was reinforced by our more recent data with antibodies and polynucleotide probes that distinguish among the different NBCe1 splice variants (Majumdar et al., 2008). We found that NBCe1-B is expressed intracellularly in neurons, whereas NBCe1-C is expressed primarily in the plasma membrane of glia and perhaps neurons of rat brain. According to results from another study somewhat at odds with our findings, Rickmann et al. detected expression of NBCe1-A in neuronal populations from hippocampus, cerebellum, olfactory bulb (periglomerular and mitral cells), and cortex of mouse brain slices (Rickmann et al., 2007). Based on the aforementioned localization studies, there is clear evidence for NBCe1 expression in both glial cells and neurons of mouse and rat brain. To date, there is no functional evidence for electrogenic NBC in any neuronal preparation. Perhaps the activity of an electrogenic NBC is masked by the activity of an electroneutral NDCBE, which appears to be the predominant HCO 3 -dependent acid extruder in neuronal preparations, particularly in ph i experiments where the two transporters are likely to function in a similar fashion and Cl-dependence is difficult to determine. In ph i experiments on acutely isolated rat hippocampal neurons, Schwiening and Boron (Schwiening and Boron, 1994) measured ph i changes with the ph-sensitive dye BCECF to identify a Na + - and Cl -dependent, DIDS-sensitive HCO 3 transporter that is mainly responsible for ph i recoveries from acid loads when neurons are bathed in CO 2 /HCO 3. However, according to more recent data, cultured hippocampal neurons appear to have

103 92 another Cl -independent, HCO 3 -dependent transporter that may be an NBC. Using fluorescence imaging of ph i with BCECF, Williams and Bevensee (unpublished data) found that exposing neurons to a 0 Na + solution caused acid loading that was dependent on HCO 3, independent of external Cl (ruling out an NDCBE), and inhibited by niflumic acid. Determining the electrogenicity of this presumed NBC requires the use of patchclamp techniques to measure transporter-mediated currents or voltage changes. Historically, investigators have used two general assays in electrophysiologic experiments to identify and characterize electrogenic NBC activity. The first assay involves exposing a cell to a CO 2 /HCO 3 solution and monitoring a HCO 3 -induced outward current due to the movement of Na +, HCO 3, and net negative charge into the cell. A HCO 3 - induced current due to an electrogenic NBC should require external Na + and be sensitive to inhibitors of HCO 3 transporters such as stilbene derivatives (e.g., DIDS). The second assay involves removing external Na + while a cell is bathed in a CO 2 /HCO 3 solution and monitoring a 0 Na + -induced inward current due to the movement of Na +, HCO 3, and net negative charge out of the cell. A 0 Na + -induced inward current due to an electrogenic NBC should require HCO 3 and be sensitive to inhibitors of HCO 3 transporters. In this second assay, Na + appears as an anion; the 0 Na + -induced inward current is a hallmark signal that was used to identify the first electrogenic NBC by function (Boron and Boulpaep, 1983), and to clone the transporter by expression (Romero et al., 1997). Using whole-cell and perforated patch-clamp techniques, we used the aforementioned bicarbonate-induced outward current and Na removal-induced inward current assays to explore the functional presence of electrogenic NBC in cultured rat hippocampal neurons. Our results are consistent with a population of neurons expressing a functional

104 93 electrogenic NBC. According to immunohistochemistry data, expression of the most established electrogenic NBC, NBCe1, is evident in all neurons, including both excitatory neurons (glutamic acid decarboxylase 67 (GAD67) negative) and inhibitory neurons (GAD67 positive), and therefore does not appear responsible for the functional heterogeneity. MATERIALS AND METHODS Animals Sprague-Dawley rats were housed in approved animal care facilities at the University of Alabama at Birmingham (UAB). The animal care facilities at UAB are under the supervision of full-time veterinarians and are fully accredited by the American Association for Accreditation of Laboratory Animal Care and meet all standards prescribed by the Guide for the Care and Use of Laboratory Animals. All the protocols followed were approved by the Institutional Animal Care and Use Committee at UAB. We have made efforts to minimize the required number of animals and their suffering in this study. Isolation of Cultured Hippocampal Neurons P0 Hippocampal Neurons Hippocampal neurons from postnatal-day 0 (P0) Sprague-Dawley pups were cultured by Dr. Susan Lyons (Sontheimer laboratory, Department of Neurobiology, UAB) using a modified protocol from Banker (Banker and Goslin, 1998). After postnatal pups were decapitated, the hippocampi were quickly removed and transferred to Hank s Buffered Salt Solution (HBSS, Invitrogen). The HBSS solution was replaced with fresh HBSS

105 94 containing papain (Worthington) for min at 37 C in the 5% CO 2 incubator. The papain-containing HBSS was then removed and 2-3 mls of neuronal astrogliaconditioned media was added and removed to rinse the tissue. Neurobasal A medium (NBM-A, Invitrogen) supplemented with B-27 (Invitrogen) and 1% Penicillin/Streptomycin (P/S, Invitrogen) was then added and the tissue triturated until the cell suspension became cloudy. The density of neurons was determined with a hemacytometer, and neurons were then plated at ~400 cells/mm 2 onto poly-ornithine coated coverslips. Neurons were cultured in NBM-A media containing B-27 and 0.5 mm glutamine (Invitrogen). The media was replaced hours after of cell plating to remove unattached cells. Subsequently, half the media was replaced every 7 days. E16-20 Hippocampal Neurons Hippocampal neurons from embryonic-day (E16-20) Sprague-Dawley pups were isolated using one of the following two culture protocols. Similar electrophysiologic results were obtained using neurons from either protocol. Culture protocol 1. After a pregnant rat was anesthetized, the embryos were removed and placed in a culture dish containing HBSS+ (HBSS containing 1% P/S). The hippocampi were removed from the decapitated head of each embryo, and temporarily stored in a conical tube containing HBSS+ on ice. Hippocampi were minced with a scalpel and the suspension centrifuged for 2 min at 75 g. The supernatant was aspirated and the hippocampi were incubated in 0.05% trypsin (Mediatech) for 9 min at 37 C. The tissue was then triturated with glass Pasteur pipettes 3-4 times before adding Soybean Tryp-

106 95 sin Inhibitor (STI) and triturating an additional ten times gently. The cells in suspension were allowed to settle briefly before adding the suspension as a layer on top of a 4% solution of bovine serum albumin (BSA) in HBSS+. The suspension was centrifuged for 5 min at 63 g and the supernatant was aspirated. The pellet was suspended in NBM++ (Neurobasal Media (NBM, Invitrogen) containing 25 μm glutamic acid, 500 μm glutamine, 25 μm of -mercaptoethanol, and a 1:50 dilution of B-27 supplement). The cells were counted using a hemacytometer and then plated in NBM++ at ~700 neurons/mm 2 on poly-ornithine coated coverslips. Every fourth day, half the media was replaced with NBM+ (NBM++ media without glutamic acid). Culture protocol 2. Some of the hippocampal neurons from this protocol were obtained either from the laboratory of Dr. Lynn Dobrunz (Department of Neurobiology, UAB) or from IDDRC Core-B Recombinant Technologies (UAB). After the pregnant rat was anesthetized, the embryos were removed, decapitated, and the brains placed in HBSS. Hippocampi were subsequently removed and temporarily stored in a conical tube containing NBM, B-27, and P/S on ice. The supernatant was aspirated and the hippocampi were incubated in papain (28 U/ml) in HBSS for 15 min at 37 C in the 5% CO 2 incubator. Next, 3 mls of astrocyte-conditioned media were added and removed to rinse the tissue. NBM supplemented with B-27, P/S, and 1.25 μm L-glutamine was then added and the tissue was triturated for 1 min with a P1000 pipette tip until the cell suspension became cloudy. The density of neurons was determined with a hemacytometer, and neurons were then plated at ~700 cells/mm 2 onto poly-ornithine coated coverslips using NBM supplemented with B-27, P/S, and 1.25 μm L-glutamine. One half of the media

107 96 was replaced 24 hours after of cell plating to remove unattached cells. Half the media was replaced every 7 days thereafter. Electrophysiology Experiments on cultured rat hippocampal neurons were performed at room temperature (RT) using the whole-cell or perforated (Amphotericin B) patch-clamp technique in voltage-clamp mode in a flow-through chamber on the stage of an inverted Leica DMIRB microscope (Leica Microsystems, Heidelberg, Germany) as previously described (McAlear et al., 2006). Bath solution was changed using a VC-6 perfusion valve control system (Warner Instruments) converging onto an 8-port manifold (Warner Instruments). Before each experiment, each solution line was primed with fresh solution. Micropipettes (3-5 M ) were pulled from G85165T-3 borosilicate glass capillaries (Warner Instruments) using a micropipette puller (PC-10, Narishige, Tokyo, Japan). Signals were obtained with an Axopatch 200B amplifier (Axon Instruments), and then digitized with a 1321A Digidata interface (Axon Instruments). Signals were filtered either at 1 or 2 khz using an 8-pole Bessel filter. pclamp 8.2 software (Axon Instruments) was used to control data acquisition and analysis. For the figures shown and analyzed, current recordings were further subjected to data reduction by substitute averaging (either by 200 or 1000) using Clampfit software (Axon Instruments). The data were exported to Microsoft Excel 2002 and subsequently transferred to Origin 7.5 (OriginLab Corporation) for graphing.

108 97 Solutions Whole-cell Experiments The standard HEPES solution contained (in mm): 125 NaCl, 3 KCl, 1 CaCl 2, 1.2 MgSO 4, 2 NaH 2 PO 4, 32 HEPES, 10.5 Glucose, and 18.8 NaOH, and titrated with NaOH or HCl to ph 7.3. The 5% CO 2 /24 mm NaHCO 3 solution contained (in mm): NaCl, 3 KCl, 1 CaCl 2, 1.2 MgSO 4, 2 NaH 2 PO 4, 10.5 Glucose, and titrated with NaOH or HCl to ph 7.3 before adding 24 NaHCO 3 and equilibrating the solution with 5% CO 2 /balance O 2. The standard patch-pipette solution contained (in mm): 10 NaCl, 120 NMDG-cyclamate, 10 HEPES, 10 EGTA, and 1 MgCl 2, and titrated with N-methyl-Dglucammonium (NMDG + ) or cyclamic acid to ph 7.3. The pipette solution was K + free to to minimize any contaminating K + current. In some experiments, the patch-pipette solution i) was titrated to ph 7.0 or 7.4, ii) contained ATP and GTP, and/or iii) contained a higher or lower concentration of NMDG-cyclamate (140 or 96 mm), a lower concentration of EGTA (1 or 3 mm), a lower concentration of HEPES (5 mm). Similar results were obtained with these different pipette solutions. For Na + -free solutions, Na + was replaced with equimolar NMDG +. Perforated-patch Experiments The same HEPES and 5% CO 2 /24 mm HCO 3 solutions described above were used in perforated-patch experiments. The 20% CO 2 /96 mm NaHCO 3 solution contained (in mm): 54.4 NaCl, 3 KCl, 1 CaCl 2, 1.2 MgSO 4, 2 NaH 2 PO 4, 10.5 Glucose, and titrated with NaOH or HCl to ph 7.3 before adding 96 NaHCO 3 and equilibrating the solution with 20% CO 2 /balance O 2. The pipette solution previously described (Bevensee et al.,

109 ) contained (in mm): 105 KCl, 45 NaCl, 1 MgCl 2, 0.2 CaCl 2, 10 EGTA, 1 HEPES, and titrated to ph M Amphotericin B was added to the pipette solution, which was made every ~3 h and stored on ice. Patch pipettes were back-filled with pipette solution without antibiotic. A 30 mg/ml stock solution of Amphotericin B in DMSO was stored at -20 C. Immunohistochemistry E16-20 rat hippocampal neurons cultured on coverslips were washed with phosphate-buffered saline (PBS) and fixed with ice-cold 4% paraformaldehyde (PFA) for 20 min at RT. PFA was discarded and a permeabilization solution supplemented with 4% PFA, PBS, and 0.25% Triton X-100 was added for 10 min at RT. The neurons were then washed 3 with cold phosphate-buffered saline (PBS) before incubating the cells for 1h at 37 C in a blocking buffer comprised of PBS supplemented with 10% BSA. The neurons were incubated in two primary antibodies (1:100 NBCe1-A/B or 1:100 NBCe1-C and 1:100 GAD67) in blocking buffer comprised of PBS and 3% normal horse serum overnight at 4 C. Cells were then washed 3 with PBS before being incubated in two secondary antibodies (1:200 Alexa anti-rabbit 594 and Alexa anti-mouse 488, Jackson Immunoresearch) in a blocking buffer comprised of PBS and 3% normal horse serum for 1-2 h in the dark at RT. Neurons were washed 3 with PBS and then treated with 0.2 μg/ml bis benzimide (Hoechst 33,258) counter stain for 10 min. Neurons were subsequently washed 3 with PBS and 1 with ddh 2 O, and then mounted onto glass slides using permafluor mountant (Thermo Scientific). Labeling was extremely weak in negative controls in which the primary antibodies were omitted in the aforementioned protocol.

110 99 Chemicals All chemicals were obtained from Sigma-Aldrich unless specified. Statistics Means between groups of data were compared using paired and unpaired Student s t test (Microsoft Excel 2002). A P value of less than 0.05 was considered significant. RESULTS Functional Assay: CO 2 /HCO 3 -induced Outward Currents 5% CO 2 /24 mm HCO 3 As described in the Introduction, there is both molecular evidence and preliminary functional data consistent with an active electrogenic NBC in rat hippocampal neurons. In our first series of experiments, we performed the aforementioned bicarbonate-induced outward current assay for electrogenic NBC activity in the neurons. In these experiments, we voltage-clamped the neurons at a very depolarized potential of -10 mv to increase the electrical driving force for the influx of Na +, HCO 3, and net-negative charge via an electrogenic NBC. We observed one of two responses when neurons were switched from a nominally CO 2 /HCO 3 -free solution to one containing 5% CO 2 /24 mm HCO 3 (same ph of 7.3). An example of one response is shown in Fig. 1 in which the CO 2 /HCO 3 solution failed to elicit any current. The absence of a CO 2 /HCO 3 -induced current was observed in approximately half the neurons (8/17). An example of the second response is shown in

111 100 Fig. 2A in which the CO 2 /HCO 3 solution elicited outward currents (a and b) consistent with electrogenic NBC activity. Such a CO 2 /HCO 3 -induced outward current was observed in the other nine neurons. If this HCO 3 -induced outward current is due to an electrogenic NBC, then the current should be inhibited by removing external Na +. Although the Fig.-2A experiment was performed in the presence of the epithelial Na + -channel inhibitor amiloride, removing external Na + in the nominal absence of CO 2 /HCO 3 caused a large outward current, presumably due to an amiloride-insensitive Na + conductance. The 0 Na + -induced outward current was evident in both the presence and absence of amiloride. In the absence of external Na +, returning to the CO 2 /HCO 3 solution only caused a small inward current (c). Thus, the CO 2 /HCO 3 -induced outward currents seen earlier in the experiment (a and b) required the presence of external Na + a finding consistent with the presence of an electrogenic NBC. In five experiments similar to that shown in panel A, the mean CO 2 /HCO 3 -induced outward current was ~8-fold larger in the presence vs. absence of external Na + (Fig. 2B). CO 2 /HCO 3 -induced currents in the presence and absence of external Na + from the 17 experiments presented in Figs. 1 and 2 are compiled in Fig. 3A. In the 17 experiments, CO 2 /HCO 3 elicited i) small or modest inward currents in eight neurons (greyfilled diamonds) with a mean of 2.3 ± 0.9 pa (n=8), or ii) small or modest outward currents (black-filled diamonds) with a mean of 5.1 ± 2.2 pa (n=9). In five of the latter nine neurons, the CO 2 /HCO 3 -induced current was reduced from a mean of 8.6 ± 3.3 pa in the presence of external Na + to 1.2 ± 2.1 pa in the absence of external Na + (open diamonds). Thus, the mean Na + -dependent, HCO 3 -induced outward current in these five neurons

112 101 was 7.4 ± 2.2 pa, and this apparent NBC current was evident in ~50% of all the neurons examined. As shown in the summary data presented in Fig. 3A, HCO 3 -induced currents were small (0 ± 5 pa) in many neurons including ones displaying Na + -dependent, HCO 3 -induced outward currents. We attempted to amplify or unmask any HCO 3 - induced outward current using the following three approaches. First, we added 10 μm phosphatidylinositol 4,5-bisphosphate (PIP 2 ) to the pipette solution to stimulate electrogenic NBC activity. Cytosolic PIP 2 stimulates electrogenic NBC activity in both macropatches from oocytes expressing NBCe1-A (Wu et al., 2009a), and whole oocytes expressing NBCe1-B or -C (Wu et al., 2009b). However, patch-pipette PIP 2 did not appear to increase the size of the CO 2 /HCO 3 -induced outward currents in our aforementioned whole-cell experiments (data not shown). Second, we performed the same HCO 3 -induced outward current assay described above using the perforated patch-clamp technique to avoid dialyzing out any important cytosolic factor that may be necessary for NBC activity. However, as shown in Fig. 3, the size pattern of CO 2 /HCO 3 -induced outward currents from perforated-patch experiments (panel B) were similar to that obtained from whole-cell experiments (panel A). The mean CO 2 /HCO 3 -induced outward current of 4.0 ± 1.8 pa in 4/5 neurons was comparable to the corresponding 5.1 ± 2.2 pa in 9/17 neurons seen in whole cell experiments. Our third approach to amplify the electrogenic NBC current involved a more extensive series of experiments using higher levels of CO 2 and HCO 3 as described in the following section.

113 102 20% CO 2 /96 mm HCO 3 To amplify an electrogenic NBC signal, we performed the same HCO 3 -induced outward current assay as described above, but with solutions containing a higher concentration of HCO 3 (96 mm) and equilibrated with a higher percentage of CO 2 (20%) to keep ph constant at 7.3. These experiments were performed using the perforated patchclamp technique. In contrast to our findings with the physiologic CO 2 /HCO 3 solutions, we observed only one response when neurons were switched from a nominally CO 2 /HCO 3 -free solution to one containing 20% CO 2 /96 mm HCO 3. An example is shown in Fig. 4A in which the CO 2 /HCO 3 solution elicited an outward current (a) consistent with electrogenic NBC activity. Such a CO 2 /HCO 3 -induced outward current was observed in 15/16 neurons. Removing external Na + in the nominal absence of CO 2 /HCO 3 elicited a small outward current consistent with a small Na + conductance. However, in the absence of external Na +, returning to the CO 2 /HCO 3 solution caused an outward current (b) that was smaller than the outward current observed in the presence of external Na + (a). In eight of ten experiments similar to that shown in panel A, the mean CO 2 /HCO 3 -induced outward current was ~1.6-fold larger in the presence vs. absence of external Na + (Fig. 4B). The mean Na + -dependent, HCO 3 -induced outward current in these eight neurons was 7.3 ± 1.7 pa. In summary, HCO 3 -induced outward currents were larger with the 96 vs. 24 mm HCO 3 solutions, and there was a Na + -dependent component to the 96 mm HCO 3 -induced outward current. If at least part of the HCO 3 -induced outward current seen in Fig. 4A is due to an electrogenic NBC, then it should be inhibited by DIDS. As shown in Fig. 5, switching to the 20% CO 2 /96 mm HCO 3 solution elicited the expected outward current (a). Subse-

114 103 quently, the neuron was exposed to a HEPES-buffered solution containing 500 μm DIDS and the current was allowed to stabilize. Applying the 96 mm HCO 3 solution but this time in the continued presence of DIDS elicited an outward current that was smaller (b) than observed in the absence of DIDS at the beginning of the experiment (a). In two experiments, DIDS inhibited the outward current by 44% and 71%, and the mean DIDSsensitive, HCO 3 -induced current (9.2 pa) was comparable to the mean Na + -dependent, HCO 3 -induced current (7.3 pa) obtained from Fig.-4 type experiments. 20% CO 2 /96 mm HCO 3 -induced currents in the presence and absence of external Na + from the 16 experiments presented in Figs. 4 and 5 are compiled in Fig. 6. In 15/16 experiments, CO 2 /HCO 3 elicited large outward currents (black-filled diamonds) with a mean of 16.9 ± 2.8 pa. In eight out of ten of these neurons, the CO 2 /HCO 3 -induced current was reduced from a mean of 20.3 ± 4.1 pa in the presence of external Na + to 13.0 ± 3.3 pa in the absence of external Na + (open diamonds). Thus, the mean Na + -dependent, HCO 3 -induced outward current in these eight neurons was 7.3 ± 1.7 pa. For unknown reasons, the HCO 3 -induced outward current was actually greater in the absence vs. presence of external Na + in two neurons (striped diamonds). Functional Assay: 0 Na + -induced Inward Currents In our second series of experiments, we performed the Na-removal assay for electrogenic NBC activity as described in the Introduction above. Recall that an electrogenic NBC will manifest itself as an inward current elicited by removing external Na + in the presence vs. absence of CO 2 /HCO 3 as the transporter moves Na +, HCO 3, and netnegative charge out of the cell. In these experiments, we voltage-clamped neurons at a

115 104 hyperpolarized potential of -80 mv to increase the electrical driving force for the efflux of NBC-mediated net-negative charge. Experiments were performed using the perforated patch-clamp configuration, and with 20% CO 2 /96 mm HCO 3 solutions. As shown in Fig. 7, removing external Na + in the nominal absence of CO 2 /HCO 3 induced an outward current (a), possibly due to a Na + -conductance. In the presence of 20% CO 2 /96 mm HCO 3, removing external Na + also elicited an outward current (b) that was of similar magnitude. Such 0 Na + -induced outward currents were observed in seven out of nine neurons. However, different results were obtained in the remaining two neurons one of which is shown in Fig. 8. Removing external Na + elicited an outward current that was smaller in the presence (a) vs. absence (b) of the 96 mm HCO 3 solution. The smaller 0 Na + -induced outward current in the presence of the HCO 3 solution is consistent with an NBC-mediated, 0 Na + -induced inward current masked by a larger NBCindependent, 0 Na + -induced outward current. A similar result was obtained in another neuron. Thus, 2/9 or 22% of the neurons exhibited a HCO 3 -dependent, Na + removalinduced inward current (1.1 and 1.7 pa) consistent with electrogenic NBC activity. Similar results were obtained with 5% CO 2 /24 mm HCO 3 solutions in which one out of four neurons displayed a HCO 3 -dependent, Na + removal-induced current of 3.5 pa (data not shown). When combining all the CO 2 /HCO 3 data, the mean HCO 3 -dependent, Na + removal-induced current was 2.1 pa (n=3).

116 105 Molecular Assay: Expression of NBCe1 Variants Based on results above from two functional assays, we found that a population of rat hippocampal neurons that appear to exhibit electrogenic NBC activity (see summary data in Figs. 3 and 6). We next used immunohistochemistry to determine if the neurons in culture express NBCe1 the most established electrogenic NBC in neurons and glia in rat brain (Majumdar et al., 2008). As previously reported (Majumdar et al., 2008) and described in Chapter 2, NBCe1-B and -C, but not -A are the predominant NBCe1 variants in slices of rat brain. More specifically, an antibody to the C terminus of the A and B variants ( A/B) labels neurons intracellularly (e.g., in the hippocampus), whereas an antibody to the different C terminus of the C variant ( C) labels the plasma membrane of glial cells and possibly neurons (e.g., in the hippocampus). As shown in Fig. 9, A/B appeared to label all hippocampal neurons in culture, including both GAD67-positive and - negative neurons in double-labeling experiments with GAD67 (A-C). No labeling was observed in the absence of both A/B and GAD67 (D). Similar results were obtained from double-labeling experiments using NBCe1-C (E-G). Overall then, NBCe1-A/B and -C appear to be expressed in both inhibitory and excitatory neurons. DISCUSSION In the present study, we used whole-cell and perforated patch-clamp techniques to explore the presence of a functional electrogenic NBC in rat hippocampal neurons using both bicarbonate-induced outward current and Na + removal-induced inward current assays described in the Introduction. According to our results, a population of rat hippocampal neurons exhibits a functional electrogenic NBC. We examined this population

117 106 further by performing immunohistochemistry with antibodies to NBCe1 variants. All neurons appeared to express both NBCe1-A/B and NBCe1-C, and expression was evident in both excitatory (GAD67-negative) and neurons and inhibitory (GAD67-positive) neurons. In a similar fashion, Cooper et al. found that electroneutral NBC1 (NBCn1) mrna was present in both excitatory and inhibitory neurons cultured from rat hippocampus (Cooper et al., 2005). In summary, the heterogeneity of electrogenic NBC activity was not due to differences in NBCe1 expression. Bicarbonate-induced Outward Current Assay Approximately half of the hippocampal neurons displayed a Na + -dependent, HCO 3 -induced outward current when the neurons were bathed in a physiologic level of CO 2 /HCO 3. Although most of the experiments were performed in whole-cell mode, similar results were obtained in the perforated-patch mode. Notably, HCO 3 -induced outward currents in many of these neurons was very small (0 ± 5 pa) and only 50% of the neurons displayed an inward current. To amplify these currents, we performed experiments with solutions containing higher levels of CO 2 (20%) and HCO 3 (96 mm). According to the summary data shown in Fig. 6, the mean HCO 3 -induced outward current was enhanced ~3.3 fold in 20% CO 2 /96 mm HCO 3 solutions under perforated-patch conditions. In addition, a greater percentage of neurons displayed a CO 2 /HCO 3 -induced outward current (94%). However, the size of the Na + -dependent component of the HCO 3 -induced outward current (mean of 7.3 pa) was the same as observed in 5% CO 2 /24 mm HCO 3 (mean of 7.4 pa). Thus, high CO 2 /HCO 3 stimulated an NBC-independent outward current.

118 107 CO 2 /HCO 3 solutions can influence electrogenic processes through CO 2 -induced intracellular acidification or molecular CO 2 per se (independent of changes in ph) that can alter the expression and/or activity of ion channels. For example, exposing postnatal mice to 7.5-8% CO 2 for two weeks increased the expression of the Na + channel subtype I protein in the hippocampus (Gu et al., 2004). Similarly, raising extracellular CO 2 from 5 to 10% increased L-type Ca 2+ currents in rabbit carotid body glomus cells (Summers et al., 2002). CO 2 may also alter electrogenic processes by carbamylating proteins, similar to that seen with hemoglobin. In our studies, the NBC-independent current activated by high CO 2 /HCO 3 could be due to either a CO 2 -induced ph i decrease that stimulates a K + channel, or perhaps a HCO 3 current through an anion channel. The following additional finding supports the conclusion that the Na + -dependent, HCO 3 -induced outward current of ~7.3 pa obtained from neurons bathed in either physiologic and high levels of CO 2 /HCO 3 is due to an electrogenic NBC. DIDS inhibited ~60% of the high CO 2 /HCO 3 -induced current (Fig. 5), and the DIDS-sensitive current of 9.2 pa was comparable in size to the Na + -dependent, HCO 3 -induced current of ~7.3 pa. Na-removal Assay Examining a 0 Na + -induced inward current due to an electrogenic NBC was difficult because removing Na + elicited an outward current in both the presence and nominal absence of CO 2 /HCO 3. There are at least a few possible explanations for this 0 Na + - induced outward current. The most likely explanation is a Na + conductance. However, we were unsuccessful in inhibiting the 0 Na + -induced outward current with the voltagedependent, Na + -channel blockers TTX or capsaicin. If a Na + -channel is responsible, then

119 108 it is not a voltage-dependent one. Other possible explanations for the outward current include 0 Na + -induced reversal of the 3Na-Ca exchanger and increased activity of the Napump. We suspect that this 0 Na + -induced outward current masked the expected NBCmediated inward current in our NBC assay. In fact, a small fraction of neurons (~25%) exhibited a smaller 0 Na + -induced outward current in the presence of CO 2 /HCO 3 than in the nominal absence. The calculated mean HCO 3 -dependent, 0 Na + -induced inward current of 2.1 pa in these neurons is consistent with electrogenic NBC activity. One might ask if a small HCO 3 -dependent inward current of 2.1 pa in our Na + - removal assay is an expected size for an electrogenic NBC activity. NBC current can be calculated if one knows either the ratio of the Na + :HCO 3 flux through the transporter, or the flux of one of the ions, as well as the Na + :HCO 3 stoichiometry of the transporter. With some preliminary ph i data mentioned in the Introduction, the latter approach is currently easier. If the transporter in neurons is similar to the electrogenic NBC in hippocampal astrocytes, then the stoichiometry is 1:2 Na + :HCO 3. According to data from fluorescence ph i imaging experiments (Williams and Bevensee, unpublished data), removing external Na + from cultured hippocampal neurons bathed in a 0 Cl, 5% CO 2 /17 mm HCO 3 solution (ph 7.3) elicited a HCO 3 -dependent acid-loading rate of 66 M/s. This HCO 3 flux (really a pseudoflux because the units include cell volume rather than surface area) corresponds to a charge movement of mol/liter cell volume/s 10 5 C/mol = 6.6 C/L/s. For the transporter having a 1:2 Na + :HCO 3 stoichiometry with one net negative charge moving for each pair of HCO 3 ions, the NBC-mediated charge movement is 3.3 C/L/s. For a neuron with a cell volume of 1 pl, the total charge movement is therefore 3.3 C/L/s L = C/s = 3.3 pa. This small current is in the range of

120 109 our measured HCO 3 -dependent, 0 Na + -induced current of 2.1 pa. The difference could easily be explained by different experimental conditions. For example, the ph i experiments were performed at 37 C, whereas the patch-clamp experiments were performed at room temperature. Because biological processes are usually faster at higher temperatures, our 2.1-pA current would likely be higher and closer to 3.3 pa had we performed our patch-clamp experiments at 37 C. On the other hand, it is not clear why the Na + - dependent, HCO 3 -induced outward current in our first assay is considerably larger (i.e., 7.3 pa). Importantly though, all these currents are in an expected range. Functional Heterogeneity of Electrogenic NBC Activity in Hippocampal Neurons In both of our NBC assays, only a fraction of hippocampal neurons exhibited electrogenic NBC activity. The reason for such heterogeneity is not known, but it is not due to differences in the expression of the most well-known NBC, NBCe1. Although it is possible that the functional electrogenic NBC is not NBCe1, the only other known electrogenic NBC is NBCe2. However, there is not much evidence for NBCe2 expression in neuronal or glial cells of the brain; the transporter is primarily expressed in the epithelial cells of the choroid plexus (Millar and Brown, 2008). Functional NBC heterogeneity may result from neurons with different sets of NBC regulators. Some NBC regulators include classic ones such as protein kinases (Perry et al., 2006), camp (Gross et al., 2003; Gross et al., 2001), ATP (Heyer et al., 1999), Ca 2+ (Muller-Berger et al., 2001) and PIP 2 (Wu et al., 2009a; Wu et al., 2009b) as well as less conventional ones such as protein binding partners. For example, the IP 3 -receptor binding protein released with IP 3 (IRBIT) binds to the B variant of NBCe1 and increases

121 110 increases the transporter current when expressed in oocytes (Shirakabe et al., 2006). If NBC regulators are responsible for the functional NBC heterogeneity among neurons, then NBC-mediated ph i regulation would appear to be dynamic among a seemingly identical group of neurons. Furthermore, NBC activity in a neuron may be silent one moment, and yet active the next following activation of a regulatory pathway. Such dynamic ph i regulation may have important ramifications on neuronal activity. Further studies are required to determine if such neuronal heterogeneity is also present in vivo. ACKNOWLEDGEMENTS This work was supported by NIH/NINDS NS (M.O.B). We thank Dr. Susan Lyons (previous member of the Sontheimer laboratory, UAB) for providing us with P0 rat hippocampal neurons. Dr. Lynn Dobrunz s laboratory (UAB), and the IDDRC Core-B Recombinant Technologies (UAB) also kindly provided some of the embryonic rat hippocampal neurons. REFERENCES Banker G, Goslin K (1998) Culturing Nerve Cells. Cambridge, MA: MIT Press. Bevensee MO, Apkon M, Boron WF (1997) Intracellular ph regulation in culture astrocytes from rat hippocampus. II. Electrogenic Na/HCO 3 cotransport. J Gen Physiol 110: Bevensee MO, Schmitt BM, Choi I, Romero MF, Boron WF (2000) An electrogenic Na + - HCO 3 cotransporter (NBC) with a novel COOH-terminus, cloned from rat brain. Am J Physiol Cell Physiol 278: C1200-C1211. Boron WF (1985) Intracellular ph-regulating mechanism of the squid axon. Relation between the external Na + and HCO 3 dependences. J Gen Physiol 85: Boron WF, Boulpaep EL (1983) Intracellular ph regulation in the renal proximal tubule of the salamander. Basolateral HCO 3 transport. J Gen Physiol 81:

122 111 Boron WF, De Weer P (1976) Intracellular ph transients in squid giant axons caused by CO 2, NH3, and metabolic inhibitors. J Gen Physiol 67: Boron WF, Russell JM (1983) Stoichiometry and ion dependencies of the intracellularph-regulating mechanism in squid giant axons. J Gen Physiol 81: Chesler M (2003) Regulation and modulation of ph in the brain. Physiol Rev 83: Choi I, Romero MF, Khandoudi N, Bril A, Boron WF (1999) Cloning and characterization of a human electrogenic Na + -HCO 3 cotransporter isoform (hhnbc). Am J Physiol 276: C576-C584. Cooper DS, Saxena NC, Yang HS, Lee HJ, Moring AG, Lee A, Choi I (2005) Molecular and functional characterization of the electroneutral Na/HCO 3 cotransporter NBCn1 in rat hippocampal neurons. J Biol Chem 280: Deitmer JW, Schlue WR (1987) The regulation of intracellular ph by identified glial cells and neurones in the central nervous system of the leech. J Physiol 388: Deitmer JW, Schlue WR (1989a) An inwardly directed electrogenic sodium-bicarbonate co-transport in leech glial cells. J Physiol 411: Deitmer JW, Schlue WR (1989b) Functional role of the sodium-bicarbonate cotransport of glial cells in the leech central nervous system. Acta Physiol Scand Suppl 582: 31. Gross E, Fedotoff O, Pushkin A, Abuladze N, Newman D, Kurtz I (2003) Phosphorylation-induced modulation of pnbc1 function: distinct roles for the amino- and carboxy-termini. J Physiol 549: Gross E, Hawkins K, Pushkin A, Sassani P, Dukkipati R, Abuladze N, Hopfer U, Kurtz I (2001) Phosphorylation of Ser(982) in the sodium bicarbonate cotransporter knbc1 shifts the HCO 3 : Na + stoichiometry from 3 : 1 to 2 : 1 in murine proximal tubule cells. J Physiol 537: Gu XQ, Xue J, Haddad GG (2004) Effect of chronically elevated CO 2 on CA1 neuronal excitability. Am J Physiol Cell Physiol 287: C691-C697. Heyer M, Muller-Berger S, Romero MF, Boron WF, Fromter E (1999) Stoichiometry of the rat kidney Na + -HCO 3 cotransporter expressed in Xenopus laevis oocytes. Pflugers Arch 438: Majumdar D, Maunsbach AB, Shacka JJ, Williams JB, Berger UV, Schultz KP, Harkins LE, Boron WF, Roth KA, Bevensee MO (2008) Localization of electrogenic Na/bicarbonate cotransporter NBCe1 variants in rat brain. Neuroscience 155:

123 112 McAlear SD, Liu X, Williams JB, McNicholas-Bevensee CM, Bevensee MO (2006) Electrogenic Na/HCO3 cotransporter (NBCe1) variants expressed in Xenopus oocytes: functional comparison and roles of the amino and carboxy termini. J Gen Physiol 127: Millar ID, Brown PD (2008) NBCe2 exhibits a 3 HCO 3 :1 Na + stoichiometry in mouse choroid plexus epithelial cells. Biochem Biophys Res Commun 373: Muller-Berger S, Ducoudret O, Diakov A, Fromter E (2001) The renal Na-HCO 3 cotransporter expressed in Xenopus laevis oocytes: change in stoichiometry in response to elevation of cytosolic Ca 2+ concentration. Pflugers Arch 442: Perry C, Blaine J, Le H, Grichtchenko II (2006) PMA- and ANG II-induced PKC regulation of the renal Na + -HCO 3 cotransporter (hknbce1). Am J Physiol Renal Physiol 290: F417-F427. Rickmann M, Orlowski B, Heupel K, Roussa E (2007) Distinct expression and subcellular localization patterns of Na + /HCO 3 cotransporter (SLC 4A4) variants NBCe1-A and NBCe1-B in mouse brain. Neuroscience 146: Romero MF, Fong P, Berger UV, Hediger MA, Boron WF (1998) Cloning and functional expression of rnbc, an electrogenic Na + - HCO 3 cotransporter from rat kidney. Am J Physiol 274: F425-F432. Romero MF, Hediger MA, Boulpaep EL, Boron WF (1997) Expression cloning and characterization of a renal electrogenic Na + / HCO 3 cotransporter. Nature 387: Schmitt BM, Berger UV, Douglas RM, Bevensee MO, Hediger MA, Haddad GG, Boron WF (2000) Na/HCO 3 cotransporters in rat brain: expression in glia, neurons, and choroid plexus. J Neurosci 20: Schwiening CJ, Boron WF (1994) Regulation of intracellular ph in pyramidal neurones from the rat hippocampus by Na + -dependent Cl - HCO 3 exchange. J Physiol 475: Shirakabe K, Priori G, Yamada H, Ando H, Horita S, Fujita T, Fujimoto I, Mizutani A, Seki G, Mikoshiba K (2006) IRBIT, an inositol 1,4,5-trisphosphate receptor-binding protein, specifically binds to and activates pancreas-type Na + / HCO 3 cotransporter 1 (pnbc1). Proc Natl Acad Sci U S A 103: Summers BA, Overholt JL, Prabhakar NR (2002) CO 2 and ph independently modulate L-type Ca 2+ current in rabbit carotid body glomus cells. J Neurophysiol 88: Thomas RC (1977) The role of bicarbonate, chloride and sodium ions in the regulation of intracellular ph in snail neurones. J Physiol 273:

124 113 Tombaugh GC, Somjen GG (1996) Effects of extracellular ph on voltage-gated Na +, K + and Ca 2+ currents in isolated rat CA1 neurons. J Physiol 493 ( Pt 3): Wu J, McNicholas CM, Bevensee MO (2009a) Phosphatidylinositol 4,5-bisphosphate (PIP 2 ) stimulates the electrogenic Na/HCO 3 cotransporter NBCe1-A expressed in Xenopus oocytes. Proc Natl Acad Sci U S A 106: Wu J, Liu X, Thornell IM, Bevensee MO (2009b) Phosphatidylinositol 4,5-bisphosphate (PIP 2 ) stimulation of electrogenic Na/bicarbonate cotransporter NBCe1 variants expressed in Xenopus laevis oocyes. FASEB J. 23:800.13

125 Fig. 1. HCO 3 -induced outward current assay: Absence of a functional electrogenic NBC in a cultured hippocampal neuron voltage-clamped at -10 mv. Switching the bath solution from a nominally HCO 3 free, HEPES-buffered solution to one containing 5% CO 2 /24 mm HCO 3 failed to elicit any current. Similar results were obtained in 8/17 neurons. 114

126 Fig. 2. HCO 3 -induced outward current assay: Presence of a functional electrogenic NBC in a cultured hippocampal neuron voltage-clamped at -10 mv. A. Switching the bath solution from a the HEPES-buffered solution to one containing 5% CO 2 /24 mm HCO 3 elicited an outward current (a,b) that was reversible upon returning to the HEPES-buffered solution. In the absence of external Na +, applying the HCO 3 solution only caused a small inward current (c). HCO 3 - induced outward currents were obtained in 9/17 neurons, and in 5/5 of those neurons, the HCO 3 - induced current was inhibited by removing external Na +. B. Summary data from the aforementioned five experiments. 115

127 Fig. 3. HCO 3 -induced outward current assay: Summary data from Fig. 1- and Fig. 2-type experiments. A. Whole-cell experiments were performed using solutions containing 5% CO 2 /24 mm HCO 3, and CO 2 /HCO 3 -induced currents in the presence and absence of Na + are shown for all experiments. CB effect is defined as a CO 2 /HCO 3 -induced outward current expected from an electrogenic NBC, whereas No CB effect is defined as a CO 2 /HCO 3 -induced inward current inconsistent with an electrogenic NBC. B. Perforated patch-clamp experiments were performed using solutions containing 5% CO 2 /24 mm HCO 3, and CO 2 /HCO 3 -induced currents in the presence of Na + are shown for all experiments. 116

128 Fig. 4. HCO 3 -induced outward current assay: Presence of a functional electrogenic NBC in a cultured hippocampal neuron voltage-clamped at -10 mv. A. Switching the bath solution from a the HEPES-buffered solution to one containing 20% CO 2 /96 mm HCO 3 elicited an outward current (a) that was reversible upon returning to the HEPES-buffered solution. In the absence of external Na +, applying the HCO 3 solution only caused a smaller outward current (b). HCO 3 -induced outward currents were obtained in 15/16 neurons, and in 8/10 of those neurons, the HCO 3 -induced current was reduced by removing external Na +. B. Summary data from the aforementioned ten experiments. 117

129 Fig. 5. HCO 3 -induced outward current assay: DIDS-sensitive, HCO 3 -induced outward current from a cultured hippocampal neuron voltage-clamped at -10 mv. Switching the bath solution from a the HEPES-buffered solution to one containing 20% CO 2 /96 mm HCO 3 elicited an outward current (a) that was reversible upon returning to the HEPESbuffered solution. In the presence of 500 μm DIDS, the HCO 3 solution caused a smaller outward current (b). 118

130 Fig. 6. HCO 3 -induced outward current assay: Summary data from Fig. 4- and Fig. 5-type experiments. A. Perforated patch-clamp experiments were performed using solutions containing 20% CO 2 /96 mm HCO 3, and CO 2 /HCO 3 -induced currents in the presence and absence of Na + are shown for all experiments. CB effect is defined as a CO 2 /HCO 3 -induced outward current expected from an electrogenic NBC, whereas No CB effect is defined as a CO 2 /HCO 3 -induced inward current inconsistent with an electrogenic NBC. 119

131 Fig. 7. Na removal-induced inward current assay: Absence of a functional electrogenic NBC in a cultured hippocampal neuron voltage-clamped at -80 mv. Removing external Na + in a nominally HCO 3 free, HEPES-buffered solution elicited an outward current (a) that was smaller than the one observed on removing external Na + in a 20% CO 2 /96 mm HCO 3 solution (b). Similar results were obtained in 7/9 neurons. 120

132 Fig. 8. Na removal-induced inward current assay: Presence of a functional electrogenic NBC in a cultured hippocampal neuron voltage-clamped at -80 mv. Removing external Na + in a 20% CO 2 /96 mm HCO 3 solution elicited an outward current (a) that was smaller than the one observed on removing external Na + in a nominally HCO 3 free, HEPESbuffered solution (b). Similar results were obtained in 2/9 neurons. 121

133 122 A/B GAD67 Merge No 1 antibody A B C D C GAD67 Merge E F G Fig. 9. NBCe1-A/B, -C, and GAD67 expression in cultured hippocampal neurons. (A-D) Fluorescence microscopy image of A/B and GAD67 labeling in cultured hippocampal neurons. Colocalization of A/B (1:100) (A) and GAD67 (1:100) (B) is seen in the merged image (C). No labeling was seen in the negative control without primary antibody (D). (E-G) Fluorescence microscopy image of C and GAD67 labeling in cultured hippocampal neurons. Colocalization of C (1:100) (E) and GAD67 (1:100) (F) is seen in the merged image (G). Scale bar = 10 μm for A-G.