Citation
Biosynthesis, degradation, and intracellular targeting of EAAC1 glutamate/aspartate transporter in C6 glioma cells

Material Information

Title:
Biosynthesis, degradation, and intracellular targeting of EAAC1 glutamate/aspartate transporter in C6 glioma cells
Creator:
Yang, Wenbo, 1964-
Publication Date:
Language:
English
Physical Description:
xiii, 184 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Amino acid transport systems ( jstor )
Amino acids ( jstor )
Antibodies ( jstor )
Biodegradation ( jstor )
Biosynthesis ( jstor )
Cell membranes ( jstor )
Cultured cells ( jstor )
Glioma ( jstor )
Membrane proteins ( jstor )
Monomers ( jstor )
ATP-Binding Cassette Transporters -- biosynthesis ( mesh )
ATP-Binding Cassette Transporters -- metabolism ( mesh )
ATP-Binding Cassette Transporters -- pharmacokinetics ( mesh )
Aspartic Acid -- metabolism ( mesh )
Biological Transport -- physiology ( mesh )
Carrier Proteins -- biosynthesis ( mesh )
Carrier Proteins -- metabolism ( mesh )
Carrier Proteins -- pharmacokinetics ( mesh )
Cell Line ( mesh )
Department of Biochemistry and Molecular Biology thesis Ph.D ( mesh )
Dissertations, Academic -- College of Medicine -- Department of Biochemistry and Molecular Biology -- UF ( mesh )
Glioma ( mesh )
Glutamic Acid -- metabolism ( mesh )
Rats ( mesh )
Research ( mesh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1998.
Bibliography:
Bibliography: leaves 174-183.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Wenbo Yang.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
029904910 ( ALEPH )
51175823 ( OCLC )

Downloads

This item has the following downloads:


Full Text










BIOSYNTHESIS, DEGRADATION, AND INTRACELLULAR TARGETING OF EAAC1 GLUTAMATE/ASPARTATE TRANSPORTER IN C6 GLIOMA CELLS










By

WENBO YANG


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1998














ACKNOWLEDGMENTS


I would like to thank the members of my supervisory committee: Dr.

Charles Allen, Dr. Brian Cain, Dr. William Dunn, and Dr. Susan Frost. I wish to extend special thanks to my mentor, Dr. Michael Kilberg. Without him, this study would not be possible. I would also acknowledge the current and previous members of Dr. Kilberg's laboratory for their friendship and valuable support. Most of all, I would like to thank my husband, Ming, my daughter, Helen, as well as my parents and parents-in-law, whose endless love, support, and personal sacrifice have made my career dream come true.















TABLE OF CONTENTS


pge

ACKNOW LEDGM ENTS................. .............................................................. ii

LIST OF FIGURES ........................... .................... v

A B B R E V IA T IO N S .......................................................................................................vii

A B S T R A C T ..................................................................................................................xii

CHAPTERS
1 IN TR O D U C TIO N ............................ .............................. ......... .................... 1

Overview of Mammalian Amino Acid Transport ........................................................... 1
Anionic Amino Acid Transport ............................................................ ................... 2
Molecular Cloning of the SDHA Glutamate/Aspartate Transporters .............................. 3
Structural and Functional Characteristcs of the Five SDHA Glutamate/Aspartate Transporters .................................................................. ... ..... ................... 7
Tissue and the Cell Distribution of the Five SDHA Glutamate/Aspartate T ransporters.......................................................... ............................................. . . . . 9
Physiological Significance of the SDHA Glutamate/Aspartate Transporters.................. 10
Biosynthesis and Degradation of Membrane Glycoproteins........................................ 13
De Novo Biosynthesis and Intracellular Targeting of the Integral Membrane N -G lyco protein ............................................................................................................. 15
Degradation of the Plasma Membrane Proteins ........................................................... 19
C6 Glioma Cell and Its Utilization in SDHA Glutamate/Aspartate Transporter S tu d ie s .......................................................................................................................... 2 7

2 MATERIALS AND METHODS .................. ................................... 32

M a te ria ls ....................................................................................................................... 3 2
M etho d s ............... ............................................................ .................................... 3 3










3 ACTIVITY, PROTEIN CHARACTERISTICS, AND BIOSYNTHESIS OF EAAC1 GLUTAMATE/ASPARTATE TRANSPORTER IN C6 GLIOMA CELLS ..... 59 In tro d u ctio n .................................................................................................................. 5 9
R esults . . . . . .... ............. ............................ ........................................... 65
Discussion ..................................................... ... ................. 84


4 DETERMINATION OF THE DEGRADATION RATE, AND THE PM RESIDENCE TIME FOR EAAC1 TRANSPORTER IN C6 GLIOMA CELLS..........121 Introduction ..................................................................................... ........................ 12 1
R esu lts ............................................................................................. ........................ 126
Discussion ............................................ ............... ...................134

5 CONCLUSIONS AND FUTURE DIRECTIONS ...............................................159


C onclu sions................................................................................... .......................... 159
F utu re d irectio n s ..........................................................................................................17 1

LIST OF REFERENCES ........... ............... ...............................174

BIOGRAPHICAL SKETCH ......................................................................184













LIST OF FIGURES


Figure Pae

2-1 EAAC1 Peptide Competition For Immunoblotting ...................................49

2-2 Effect of Salt on Immunoprecipitation of EAAC1 Protein ........................51

2-3 Effectiveness of Depletion of EAAC1 Protein by Immunoprecipitation ....53 2-4 EAAC1 Immunoprecipitation Protein Concentration Curve .....................55

2-5 Determination of the Elution Conditions for Immunoprecipitated EAACI. 56 2-6 Specificity of Cell Surface Protein Biotinylation .......................................58

3-1 Na+-Dependent Glutamate/Aspartate Transport Activity of
C 6 G liom a C ells ...................................................................................... 91

3-2 RT-PCR Detection of Glutamate/Aspartate Transporters in
C 6 G liom a C ells ...................................................................................... 93

3-3 Effect of Cell Density/Growth on the XAG- Activity, EAAC1 Content,
and Differentiation of C6 Glioma Cells ....................................................95

3-4 Immunocytochemistry Staining of C6 Glioma Cells with Specific
EAAC1, M6PR, and 414 Antibodies ......................................................97

3-5 Sucrose Gradient Fractionation of Total Cellular Membranes ................99

3-6 Cell Surface Biotinylation of EAAC1 in C6 Glioma Cells .......................101

3-7 Oligomerization of the EAAC1 Transporter ...........................................103

3-8 Over-Expression of EAAC1 Transporter Protein in BNL CL.2 Cells ......105 3-9 De-Glycosylation of EAAC1 Protein with N-Glycosidase F ...................107

3-10 Effect of Tunicamycin on the EAAC1 Transporter Protein Content
and Transport A ctivity ........................................................................... 109

3-11 De Novo Biosynthesis of EAAC1 Transporter in C6 Glioma Cells ........111









3-12 Maturation of EAAC1 Protein from Lower MW Forms to Its Monomer... 113 3-13 Transition of Newly Synthesized EAAC1 from its Lower MW
Form to the Mature Monomer Form in C6 Glioma Cells ...................... 115

3-14 Tafficking of Newly Synthesized EAAC1 Protein to the Plasma
M em brane in C6 Gliom a Cells ............................................................ 117

3-15 Endoglycosidase H-Sensitivity of Newly Synthesized EAAC1
Protein in C6 G liom a Cells .................................................................. 119

4-1 Determination of the EAAC1 Protein Degradation Using Pulse-Chase
Labeling in C6 G liom a Cells ................................................................ 142

4-2 Degradation Rate of the Cell Surface-Biotinylated EAAC1 Protein ....... 144 4-3 Determining the Plasma Membrane Residence Time of Newly
Synthesized EAAC1 Protein ............................................................... 146
4-4 Comparison of the EAAC1 Degradation Rate with the Rate of
Its Disapperance from Cell Surface in C6 Glioma Cells ...................... 148

4-5 Effect of Various Inhibitors on the Degradation of EAAC1 Protein
in C 6 G liom a C ells . ............................................................................. 150

4-6 Effect of ALLN on the Degradation of EAAC1 in C6 Glioma Cells ........ 152 4-7 Effect of Leupeptin + NH4CI on the Degradation of EAAC1 in
C 6 G liom a C ells ................................................................................. 154

4-8 Effect of Leupeptin Alone on the Degradation of EAAC1 in
C 6 G liom a C ells ............................................ ..................................... 156

4-9 Effect of NH4CI Alone on the Degradation of EAAC1 ............................ 158















ABBREVIATIONS


PCi pg

1-l

pM 3ME [R]o [R]t ALLN

AS ATP BFA BSA C12E9

cDNA CFTR Chol-KRP

CNP CNS


microcurie

microgram

microliter

micromolar

-mercaptoethanol initial radiolabeled protein concentration the radiolabeled protein remaining after time t N-acetyl-leu-leu-nodrleucinal asparagine synthetase adenosine 5'-triphosphate brefeldin A

bovine serum albumin polyoxyethylene 9 lauryl ether complementary deoxyribonucleic acid cystic fibrosis tramsmembrane regulator protein sodium-free Krebs-Ringers phosphate buffer 2',3'-cyclic nucleotide-3'-phosphohydrolase central nervous system









cRNA complementary ribonucleic acid DMEM Dulbecco's modified Eagle medium DTT dithiothreitol EAAC1 rodent excitatory amino acid carrier 1 EAAT1-5 human excitatory amino acid transporters 1-5 EDTA ethylenediamine tetraacetic acid EGF epidermal growth factor EGTA ethylene glycol-bis(p-aminoethyl ether)-tetraacetic acid Endo H protein endoglycosidase H ER endoplasmic reticulum FBS fetal bovine serum FITC fluorescein isothiocynate GABA y-aminobutyric acid GDP guanosine 5'-diphosphate GFAP glial fibrillary acidic protein GLAST rodent glutamate/aspartate transporter GLT1 rodent glutamate transporter 1 Glut facilitated glucose transporter GLYT1 glycine transporter 1 GRP78 glucose regulated protein 78 GS glutamine synthetase GTP guanosine 5'-triphosphate HEPES 4-(2-hydroxyethyl)-l1-piperazineethanesulfonic acid









hr hour HRP horseradish peroxidase IgG immunoglubin G kd the fractional rate constant for protein degradation kDa kilodalton LDL low density lipoprotein M6PR mannose-6-phosphate receptor mA miliampere MBP maltose binding protein MEM Eagles minimal essential medium min minute mM milimolar mRNA messenger ribonucleic acid MW molecular weight NaKRB sodium-containing Krebs-Ringers bicarbonate buffer NaKRP sodium-containing Krebs-Ringers phosphate buffer NFDM non-fat dry milk NGS normal goat serum NOC nocodazole NSF N-ethylmaleimide-sensitive fusion protein PAGE polyacrylamide gel electrophoresis PBS phosphate buffered saline PCR polymerase chain reaction









PES buffer

PI3K PKA PKC

PM

PMSF PNGase F

RT RT-PCR

SD SDB SDHA

SDS SEB buffer SNAP SNARE

SRP

Sulfo-NHS-LC-biotin

Tu/2 TBS-T TCA

Tf

TLCK


detergent-containing protein solubilization buffer phosphatidylinositol 3-kinase protein kinase A protein kinase C plasma membrane phenylmethyl-sulfony fluoride protein N-glycosidase F room temperature reverse transcriptase-PCR standard deviation protein gel sample dilution buffer sodium-dependent high affinity sodium-dodecyl-sulfate sucrose-EDTA-containing Tris-buffer soluble NSF attachment protein SNAP receptor signal recognition particle sulfosuccinimidyl-6-(biotinamido)hexanoate half-life

Tris-buffered saline containing Tween trichloroacetic acid transferrin

Na-p-tosyl-L-lysine chloromethyl ketone









TPCK N-tosyl-L-phenylalanine chloromethyl ketone Tris tris(hydroxymethyl)aminomethane














Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

BIOSYNTHESIS, DEGRADATION, AND INTRACELLULAR TARGETING OF EAAC1 GLUTAMATE/ASPARTATE TRANSPORTER IN C6 GLIOMA CELLS

By

Wenbo Yang

December, 1998


Chairman: Dr. Michael S. Kilberg
Major Department: Biochemistry and Molecular Biology


Mammalian amino acid transport activities have been extensively studied over the past 30 years by describing their kinetics, substrate specificity, tissue distribution, and regulation. However, the "life cycle", that is, the biosynthesis, processing, and degradation of these transporter proteins, has not yet been documented. This study and others of its kind are beginning to make use of the newly available cDNAs and sequence-specific antibodies as tools to investigate transporter cell biology and function. In the present research, rat C6 glioma cells were used as a model system to study the activity, biosynthesis, intracellular targeting, and degradation of the EAAC1 glutamate/aspartate transporter protein. The EAAC1 transporter is the major sodium-dependent high-affinity glutamate/aspartate transporter expressed endogenously in C6 glioma cells.









Both the glutamate/aspartate uptake activity and the EAAC1 protein content were correlatively down-regulated by increased cell density or decreased cell growth. The EAAC1 transporter protein was co-translationally N-glycosylated with high-mannose oligosaccharide chains, which were processed into complextype sugar chains as the protein matured. The final maturation steps for EAAC1 protein coincided with its PM arrival, which started at about 45 min and finished about 3 hours after initial synthesis. The newly synthesized EAAC1 protein was protected from degradation during the maturation/PM-targeting process, as well as during the first 5 hours after PM arrival. Once started (i.e., 8 hours after the initial synthesis), the degradation of the newly synthesized EAAC1 protein followed first order kinetics with a decay rate of 11.5% per hour yielding a halflife of 6 hours. The EAAC1 transporter protein was degraded, at least partially, through the endocytosis-lysosome protein degradation pathway. Once the EAAC1 protein was endocytosed, it was degraded immediately in lysosomes or late endosomes, no significant intracellular/recycling pool was detected. These results illustrate some interesting and unique features regarding the processing and trafficking of the EAAC1 transporter.















CHAPTER 1
INTRODUCTION


Overview of Mammalian Amino Acid Transport

Amino acids not only serve as precursors for protein biosynthesis, but also function as carriers for nitrogen and carbon atoms, as metabolic fuels, as neurotransmitters, as biosynthetic precursors, and as sensitive regulators of cell osmolarity. They can be transferred across the plasma membrane of cells through the mediation of various types of protein carriers. This transferring movement is called "amino acid transport" and these carriers are referred to as "amino acid transporters". The study of amino acid transport into animal cells actually began in 1913 with the pioneering observations of Van Slyke and Meyer who demonstrated tissue accumulation of amino acids against a concentration gradient (Van Slyke and Meyer, 1913). In the early 1960s, the Christensen laboratory originated the description of individual transport activities based on the chemical properties, size. and conformation of the amino acid side chain (Oxender and Christensen, 1963). The transport activities were further classified into individual transporter carrier "systems" using substrate competition assays. Therefore, each of the transport systems exhibits distinct substrate specificity for certain classes of amino acids, for example, zwitterionic, anionic, or cationic amino acids.









Basically, there are two mechanisms by which the amino acids are

transported across the plasma membrane of cells. Some transport systems are secondary active transporters, energized by the Na'-electrochemical gradient, and result in net accumulation of amino acid against a concentration gradient. Others are Na -independent facilitated transporters and permit flux in either direction depending on the chemical principle of mass action.



Anionic Amino Acid Transport

The anionic amino acid glutamate is the predominant excitatory

neurotransmitter in the mammalian central nervous system (CNS) (Fagg and Foster, 1983; Robinson and Coyle, 1987). It is generally assumed that extracellular concentrations of the excitatory amino acids are regulated primarily by the clearance of these neurotransmitters from the synaptic clefts through their specific transporters with little or no extracellular metabolism (Iversen, 1975; Kuhar, 1973). An excessive amount of these excitatory transmitters can be toxic to the neurons that have glutamate receptors (Choi, 1992). Transport of the anionic amino acids across the plasma membrane can be mediated by either high-affinity (Km = 2 - 50 pM) or low-affinity (Km > 100 pM) Na'dependent transporters, or by a Na -independent facilitated transporter (Schousboe, 1981; Nicholls and Attwell, 1990; Kanai et al., 1993). However, the Na -dependent high-affinity (SDHA) glutamate/aspartate transporters, that mediate an activity referred to as System XAG', are believed to play an essential role in glutamatergic transmission in the CNS as well as in the









glutamate/aspartate metabolism in brain and peripheral tissues (Radian et al., 1990; Danbolt et al., 1992; Kanai, 1993). This activity transports both the L- and D-steroisomers of aspartate, but only L-glutamate (Christensen and Makowske,1983; Brew and Attwell, 1987), and is very sensitive to membrane potential (Gazzola et al., 1981). The mechanism for anionic amino acid uptake through system XAG is electrogenic and coupled to the co-transport of three Na* ions and one H , as well as the counter-transport of one K ion (Zerangue and Kavanaugh, 1996). This coupling mechanism allows glutamate to be concentrated to high levels in both neurons and epithelia.

Even before the cloning of the first SDHA glutamate transporter, the existence of several distinct activities was proposed because transporters in synaptosomes prepared from various brain regions can be differentiated pharmacologically using specific inhibitors (Bridges et al., 1991; Fletcher and Johnston, 1991; Rauen et al., 1992). Over the past decade, five cDNAs, all encoding for SDHA glutamate/aspartate transporter activity, have been cloned. They are designated as EAAT1-5 for human clones, but for the rodent counterparts, EAAT1-3 were originally named as GLAST, GLT1, and EAAC1, respectively (for review, see Malandro and Kilberg, 1996). The latter rodent nomenclature will primarily be used in this report.



Molecular Cloning of the SDHA Glutamate/Aspartate Transporters

In 1992, Storck and coworkers identified the first SDHA glutamate

transporter cDNA and named the encoded transporter as GLAST (Storck et al.,









1992). During the isolation of a galactosyltransferase from rat brain, they copurified a hydrophobic protein of about 66 kDa. Proteolytic fragments of this protein were sequenced, and then oligonucleotide probes were generated. Using these probes, they identified a cDNA clone from rat brain cDNA library, which encoded a protein of 543 amino acids with a predicated molecular weight of 60 kDa. This clone showed considerable sequence similarity to the previously identified glutamate and monocarboxylate transporters of bacteria (Tolner et al., 1992). The SDHA glutamate/aspartate transport (XAG) activity of GLAST was confirmed by expression in Xenopus oocytes after cRNA injection followed by transport assays using radiolabeled amino acid substrates (Klockner et al., 1993).

Pines and coworkers cloned the second member of the SDHA

glutamate/aspartate transporter family, GLT1 (Pines et al., 1992). The cloning of GLT1 essentially began several years before the isolation of the cDNA with the purification of the transporter protein to near-homogeneity (Danbolt et al., 1990, 1992). The antibody generated against the partially purified transporter protein was used to immuno-screen a rat cDNA expression library and isolate the GLT1 cDNA, which encoded a 573 amino acid protein with a predicted molecular mass of 63 kDa. The function of this transporter was confirmed by measuring the uptake of radiolabeled substrates following its over-expression in mammalian cells.








Also in 1992, Kanai and Hediger identified the third member of this

glutamate transporter family by oocyte expression cloning using fractionated mRNA from rabbit intestine and named it EAAC1 (Kanai and Hediger, 1992). This cDNA contained an open reading frame encoding 524 amino acids and a protein of predicted molecular weight of 57 kDa. Expression of the EAAC1 cDNA in Xenopus oocytes resulted in a 1000-fold increase in Na -dependent uptake of L-glutamate over that of water-injected oocytes. A rat EAAC1 homolog was later isolated, which shares 95% similarity with the rabbit clone (Velaz-Faircloth et al., 1996). A polyclonal antibody against a fusion protein consisting of the C-terminal 120 amino acids of the rat EAAC1 and maltose






C)



11111


524
COOH


(Predicted Rabbit EAAC1 Protein Structure)









binding protein was generated by our laboratory, which has been shown to specifically recognize the EAAC1 protein from brain and placenta by immunoblotting (Matthews et al., 1998).

In 1995, the fourth member of this family, EAAT4, was isolated by Fairman and colleagues using degenerate oligonucleotide primers corresponding to conserved sequences within the other members of the SDHA glutamate transporter family and low-stringency RT-PCR from human cerebellum mRNA (Fairman et al., 1995). Using an oocyte expression system, it was shown that the EAAT4 transporter functioned not only as a SDHA glutamate/aspartate transporter, but also as a chloride channel activated by substrate binding, that is, binding by glutamate/aspartate and sodium. Given this unique property of the EAAT4 transporter, it was postulated to be involved in not only the clearance of glutamate from the synaptic cleft, but also in reestablishment of membrane potential by influencing cellular chloride permeability.

Last year, Arriza and coworkers cloned EAAT5, the fifth member of this family, by screening a human retinal cDNA library with a glutamate transporter cDNA isolated from salamander retina (Arriza et al., 1997). Similarly with EAAT4, EAAT5 was also shown to be a chloride channel as well as a SDHA glutamate/aspartate transporter. The associated chloride conductance activity of EAAT5 was postulated to play a role in visual processing.











Structural and Functional Characteristics of the Five SDHA Glutamate/Aspartate Transporters

These five members of the SDHA glutamate/aspartate transporter family exhibit 36 to 55% amino acid identities with each other (51-55% identity between GLAST, GLT1, and EAAC1) (Kanai et al., 1997). The hydropathy profile for each of these transporters predicts a hydrophobic protein with 6 - 8 putative transmembrane domains at the N-terminus and several shorter hydrophobic regions at the C-terminus, which may compose additional transmembrane domains (Malandro and Kilberg, 1996). There are at least 2 - 3 consensus sites for N-glycosylation located within the second extracellular loop, as well as several consensus sequences for protein kinase A (PKA) and/or protein kinase C (PKC) phosphorylation within the intracellular loops. Three members of this family, GLAST, GLT1, and EAAC1, have been successfully detected by immunoblotting as well as immunocytochemistry using specific antibodies. It has been reported that they all form homo-multimers, even when detected using immunoblotting under reducing conditions (Haugeto et al., 1996). When detected by immunoblotting, the mature monomers of these transporters run as a broad band between 66 and 74 kDa, depending on the cell type, which is likely due to micro-heterogeneity of their N-glycosylation sites, and possibly other post-translational modifications such as phosphorylation. The same transporter protein prepared from different tissues or cells migrates differently on SDS-PAGE, that is, with different apparent molecular weights (Dowd et al., 1996). This result suggests they may be









modified differently by N-glycosylation and/or phosphorylation according to the individual cell or tissue type.

It has been shown that GLAST, GLT1, and EAAC1 are N-glycosylated in cultured cells and intact rat brain (Danbolt et al., 1992; Stork et al., 1992; Dowd et al., 1996). N-glycosylation has been shown to be important for the function of many plasma membrane proteins, such as the insulin receptor (Ronnett et al., 1984), EGF receptor (Slieker et al., 1986), LDL receptor (Edwards et al., 1989), and glycine amino acid transporter GLYT1 (Olivares et al., 1995). It is postulated that N-glycosylation is important for the correct folding and oligomerization of many membrane proteins and acts as a quality control for the export of newly synthesized proteins from the ER (Montreuil et al., 1995). Nglycosylation may also play a role in stabilizing glycoproteins against proteolytic degradation (Fiedler and Simons, 1995). More recently, it has been shown that the N-glycan chains may also serve as one form of the selective sorting determinant for the correct intracellular targeting of the glycoproteins (Aridor and Balch, 1996). However, it also has been shown that abolishment of the Nglycosylation of GLAST, using site-directed mutagenesis, had no effect on the function of this transporter (Conradt et al., 1995).

At least two members of this transporter family, GLT1 and EAAC1, have been shown to be regulated by protein phosphorylation through the PKC pathway, and maybe the PI3K pathway as well (Casado et al., 1993; Dowd and Robinson, 1996; Davis et al., 1998). It was suggested that the increase in transport activity associated with EAAC1 protein phosphorylation correlated









with an increase in cell surface expression of the transporter and could not be attributed to the biosynthesis of new transporter (Davis et al., 1998).


Tissue and Cell Distribution of the Five SDHA Glutamate/Aspartate Transporters

Among these five members of the SDHA glutamate/aspartate transporter family, the EAAC1 transporter appears to be the most ubiquitously expressed. Although quite abundant in brain, a significant level of expression of this transporter can also be detected outside the nervous system: in small intestine, kidney, heart, skeletal muscle, lung, liver (Kanai and Heidiger, 1992; VelazFaircloth et al., 1996), and placenta (Matthews et al., 1998). Although the expression of the GLT1, GLAST, and EAAT4 transporters is extremely high in the central nervous system, a low level of the proteins or their correspondent mRNAs are also found in other tissues and cells, including human fibroblasts, placenta, heart, lung, and skeletal muscle (reviewed by Malandro and Kilberg, 1996). EAAT4 is expressed exclusively in cerebellar Purkinje cell soma and dendritic trees (Furuta et al., 1997), whereas EAAT5 is believed to be mainly confined to the retina (Arriza et al., 1997).

Within the brain, EAAT4 mRNA is strictly confined to cerebellum (Fairman et al., 1995), whereas EAAC1, GLAST, and GLT1 mRNAs and proteins can be detected throughout the brain (Gegelashvili and Schousboe, 1998). However, when the brain is further dissected into distinct regions, each of these transporters is enriched in a particular area and among different neural cell types. Using in situ hybridization, immunocytochemistry, and









immunohistochemistry techniques, it has been shown that GLT1 and GLAST are exclusively expressed in astrocytes, whereas EAAC1 is expressed only in neurons in the brain (Rothstein et al., 1994; Torp et al., 1994; Lehre et al., 1995). Moreover, EAAC1 is primarily restricted to dendrosomatic compartment of the glutamatergic and nonglutamatergic neurons, though some immunoreactivity was also found in the presynaptic boutons of GABAergic neurons (Rothstein et al., 1994). This observation suggests that EAAC1 may not be a major contributor for the proper termination of the glutamateric neurotransmission through removal of glutamate from the synaptic clef. In fact, this is consistent with the data obtained from the knockout experiments which will be described below.



Physiological Significance of the SDHA Glutamate/Aspartate Transporters

SDHA glutamate/aspartate transporters are key components in the

synaptic termination for glutamatergic neurotransmission. Because there is no mechanism for enzymatically degrading the released excitatory neurotransmitters in the synaptic cleft, their immediate removal by neural cells is solely responsible for the termination of the neurotransmission. Both neurons and astrocytes exhibit high capacity for glutamate uptake, although, as described below, astroglial appeared to be the primary site for glutamate clearance (uptake and biotransformation) in brain areas with high glutamatergic activity (Rothstein et al., 1996; Schousboe, 1981; Schousboe and Divac, 1979). It is estimated that the glutamate transporters can concentrate glutamate more









that 10,000-fold across the cell membrane. Consistent with this proposal, the intracellular glutamate concentration in neurons is as high as 10 mM, whereas the synaptic concentration is kept at ~ 1 pM (Hediger, 1994). Therefore, SDHA glutamate transporters are believed to play an important role in protecting neurons from excitotoxicity caused by abnormally high concentrations of extracellular glutamate. In addition, each of these SDHA glutamate transporters has been shown to have a specific and unique distribution, spatially and temporally, during brain development, suggesting that they may also play important roles in brain maturation (Bar-Peled et al., 1997). Selective in vivo and in vitro knockout of the individual glutamate transporters using anti-sense oligonucleotides showed that reduction of GLT1, and to a much lesser extent GLAST, but not EAAC1, resulted in neurodegeneration (Rothstein et al., 1996). These results provided additional proof that astroglial uptake of glutamate by GLT1 and GLAST is the major mechanism in excitotoxicity and synaptic clearance of this neurotransmitter. There is still great uncertainty as to the major function of the neuronal EAAC1 transporter. It is postulated that the EAAC1 transporter may play a role in keeping the neuronal intracellular glutamate at high levels for use as a neurotransmitter, as a precursor for GABA synthesis, or for other metabolic reactions in the brain (Kanai et al., 1995a). In addition, consistent with its ubiquitous expression among different organs and tissues, EAAC1 transporter may function as the major glutamate transporter providing glutamate and aspartate for general metabolism and other intracellular functions. Glutamate has been shown to be important for a variety of cellular









functions including cell differentiation, proliferation, and migration (Pearce et al., 1987; Mattson et al., 1988).

As mentioned above, the transport activity of the SDHA

glutamate/aspartate transporters is coupled with the co-transport of three Na and one H and counter-transport of one K (Zerangue and Kavanaugh, 1996). As a consequence of this stoichiometry, the functional state of the SDHA glutamate transporter(s) may contribute to the rise of extracellular glutamate to neurotoxic levels in pathological conditions such as anoxia and ischemia after stroke (Kanai, et al., 1995b). In these pathologic conditions, the reduced cellular ATP level might cause a breakdown of electrochemical gradients across the membrane, a rise in extracellular K , a decrease in extracellular Na , and depolarization of the membrane (Szatkowski and Attwell, 1994). In this circumstance, the SDHA glutamate transporters are proposed to run in reverse, resulting in a non-vesicular release of glutamate to toxic levels into the synaptic cleft, and subsequently leading to neuron damage (Nicholls and Attwell, 1990; Kanai et al., 1995). Therefore, selective inhibitors of the neuronal glutamate transporter, EAAC1, may be of therapeutic interest for preventing reversed glutamate transport, without affecting the capacity of glial glutamate transporters to keep the synaptic glutamate concentration at low levels. The GLT1 mRNA and protein are down-regulated in the brain following transient ischemia, while the expression of GLAST and EAAC1 are not altered (Torp et al., 1995). Given that GLT1 appears to be the primary mechanism for clearance of glutamate from the synapse (Rothstein et al., 1996), it is possible that the









decrease of GLT1 activity is one of several factors that contribute to the high sensitivity of neurons to post-ischemic damage.

Some of the sporadic forms of amyotrophic lateral sclerosis (ALS), a

neurodegenerative disease caused by a slow loss of motor neurons, have been associated with a reduction of GLTI glutamate transporter (Rothstein et al., 1992; Rothstein et al., 1995). It was later shown that aberrant RNA processing of the GLT1 transporter was responsible for the 60 -70% reduction in GLT1 protein expression in the motor cortex and spinal cord of ALS patients (Lin et al., 1998). The presence of the aberrant GLT1 mRNA species in cerebrospinal fluid may have diagnosis utility for ALS. Decreased glutamate transporter protein in the cortex, especially GLT1, has also been implicated in the neurodegeneration that occurs in Alzheimer disease (Cowburn et al., 1988; Scott et al., 1995). In contrast, schizophrenia and other psychoses are thought to result, at least partially, from glutamatergic hypofunction, a condition caused by excessive glutamate uptake (Carlsson and Carlsson, 1990).



Biosynthesis and Degradation of Membrane Glycoproteins

As stated above, the activity of amino acid transporters has been

extensively studied over the past thirty years, but little is known about their biosynthesis, degradation, and intracellular trafficking due to the lack of antibodies. In the past 5 years, the cloning and expression of a number of mammalian amino acid transporters has led to the generation of sequencespecific antibodies from corresponding peptides and fusion proteins. Therefore,









studies of amino acid transporter "life cycle" can be performed, for the first time, with these specific antibodies. I will use the term "life cycle" to describe the collective process of biosynthesis, targeting, and then degradation which requires many steps of membrane vesicle trafficking carrying the protein from one subcellular compartment to another. In recent years, the intracellular trafficking of many membrane proteins, including hormone receptors, major histocompatibility complex, ion channels, and glucose transporters, has been widely studied in eukaryotic cells (Alberts et al., 1994). Many of the individual steps that contribute to the life cycle of membrane proteins have been documented. Given that all of the amino acid transporter sequences elucidated so far encode for N-glycosylated integral membrane proteins, it is likely that many similarities exist between the life cycle of amino acid transporters and other membrane N-glycoproteins. There are basically two major pathways involved. One is the biosynthesis/exocytosis pathway, which transfers newly synthesized membrane proteins to their different cellular destinations. The other is the endocytosis/recycling pathway, which internalizes plasma membrane proteins and then either stores them in recycling vesicles or targets them for degradation (Alberts et al., 1994). Transfer of proteins from one compartment to another is achieved through the sequential movement of membrane protein cargo between distinct compartments, a process mediated by budding of small membrane vesicles, migration and recognition by the next compartment, and subsequent membrane fusion. The core protein machinery that underlines vesicle transport includes coat proteins, which sculpt a vesicle out of a donor









membrane; the vesicle- and target-specific identifiers v-SNAREs and tSNAREs, which bind to each other and thereby dock the vesicles to the acceptor membrane; small GTP-binding proteins (GTPase), which hydrolyze bound GTP to GDP in regulation of vesicle budding and docking; and NSF and SNAP proteins, which bind to the SNARE complex and initiate fusion when NSF hydrolyzes ATP (Nuoffer and Balch, 1994). In most cases, these pathways are highly regulated to meet the needs of the cells.



De Novo Biosynthesis and Intracellular Targeting of the Integral Membrane NGlycoproteins

De novo biosynthesis of an integral membrane N-glycoprotein initiates with the co-translational insertion of the nascent polypeptide into the rough ER membrane (Alberts et al., 1994). In most cases, although the nutrient transporters cloned to date are an exception, the ER import step requires the newly synthesized polypeptide to possess an ER signal peptide, which is recognized by a signal recognition particle (SRP). SRP binds both the growing polypeptide chain and the ribosome and directs them to a receptor protein on the cytosolic surface of the rough ER membrane. This docking of the ribosomepolypeptide-SRP-SRP receptor complex on the membrane initiates the translocation process that threads a loop of polypeptide chain across the ER membrane through a hydrophilic pore in a protein translocator (Rapoport, 1991).









All of the amino acid transporter sequences that have been elucidated have predicated N-glycosylation site(s), that is, an Asn-X-Ser/Thr consensus sequence, where X can be any amino acid except possibly proline and asparagine. Having this sequence stretch does not guarantee N-glycosylation of the protein, because other features of the polypeptide also have an influence on whether such potential sites become glycosylated or not. The amino acid transporter proteins containing one or more of these sequences that have been investigated are all N-glycosylated. For the biogenesis of N-glycoproteins, the nascent polypeptide is co-translationally modified, in the ER, with a preformed high-mannose oligosaccharide chain [(GIcNAC)2(Man)9(GIc)3] transferred from dolichol-P-P to Asn in the peptide (Voet and Voet, 1990). This initial Nglycosylation step is identical for all N-glycoproteins, but the later processing in the ER and Golgi complex by a combination of trimming and addition of specific carbohydrate residues varies widely depending upon the type of the cell, the individual glycoprotein, and possibly the physiological conditions. Processing of the oligosaccharide begins immediately in the ER upon the attachment of the core sugar complex. Two specific glycosidases remove the first two glucose residues in sequence, generating Asn-GIcNAC2Man9 (Alberts et al., 1994). Reglucosylation of the core oligosaccharide in the ER can also happen. It is believed that these glucose residues can interact with the protein chaperones in the ER to retain the newly synthesized N-glycoprotein until it is correctly folded, that is, functioning as quality control.









After the polypeptide is correctly folded, the N-glycoprotein will be

transported from the ER to the Golgi compartment by small membrane vesicles budding from the specialized areas of the smooth ER, migrating, then recognizing/fusing with cis-Golgi network (Alberts et al., 1994). While in the Golgi complex, the attached oligosaccharide chain(s) will undergo further glycosylation modification, in particular, they are altered from high-mannose- to complex-type in the cis Golgi. This processing pathway is highly ordered and most of the enzymes involved are rather strictly compartmentalized. Therefore, different oligosaccharide residues can serve as hallmarks for individual steps as the protein passes through the ER and Golgi complex. The mature Nglycoprotein will be targeted from the trans Golgi network (TGN) to its functional site, for example, the plasma membrane, through clathrin-coated vesicles (Schmid, 1997).

Endoglycosidases are a class of enzymes that can catalyze the cleavage of oligosaccharide chains at specific sugar residues. These enzymes are often useful for characterizing the oligosaccarides on glycoproteins and elucidating the progress through the protein maturation pathway by examining the changes in the sensitivities of the glycoprotein to different enzymes. The best characterized and most widely used enzyme for this purpose is endoglycosidase H (Endo H). When a glycoprotein passes through the medial Golgi compartment, the attached oligosaccharide chain will be processed from high-mannose- to a complex-type chain. Endo H cleaves between the two GIcNAC residues adjacent to the Asn residue, but it specifically acts on the









high-mannose and some hybrid types of N-linked oligosaccharides (Robbins, 1984). Therefore, resistance to EndoH digestion can be used as a hallmark for the N-glycosylated proteins that have already proceeded beyond the medial Golgi compartment. In contrast, protein N-glycosidase F (PNGase F) cleaves between the first GlcNAC residue and the Asn residue on all N-glycosylated proteins, and can be used to determine whether a protein is N-glycosylated or not (Maley, 1989).

N-glycosylation has been shown to be important for the functions of

many plasma membrane proteins, including insulin receptor (Montreeuil et al., 1995; Ronnett et al., 1984), EGF receptor (Slieker et al., 1986), LDL receptor (Edwards et al., 1989), and glycine amino acid transporter GLYT1 (Olivares et al., 1995). N-glycosylation has been shown to be important for the correct folding and oligomerization of many proteins and acts as a quality control for the export from the ER (Ware et al., 1995; Nauseef et al., 1995). Also, it has been suggested that glycosylation may play an important role in stabilizing glycoproteins against proteolytic degradation (Fiedler and Simons, 1995). More recently, distinct N-glycan chains have been shown to comprise one form of a potential sorting signal for the selective transport of N-glycoproteins along their biosynthetic pathway through recognition and binding with specific lectins (Aridor and Balch, 1996). Nevertheless, abolishment of the N-glycosylation of the SDHA glutamate/aspartate transporter GLAST by site-directed mutagensis had no effect on its plasma membrane arrival and its transport activity (Conradt et al., 1995).









For the amino acid transporter proteins, little is known about how long it might take for an N-glycosylated membrane protein to be synthesized and then targeted to the PM, and even less is known about how long it will stay there. Cariappa and Kilberg (1990) elucidated the biosynthesis and intracellular targeting of an amino acid transporter, System A, using functional assays at a time when no antibodies were available for any amino acid transporter. The de novo biosynthesis of rat hepatic System A was up-regulated by the treatment with glucagon and dexmethasone. Using a Golgi subfractionation technique, they showed that the initial increased transport activity could be detected in the cis Golgi at approximately 45 min following hormone treatment, and in the remaining Golgi fractions and at the cell surface after 60 min. These results were consistent with whole cell transport data showing a 1 hr lag prior to the protein synthesis-dependent increase in plasma membrane System A transport activity following hormone treatment (Christensen and Kilberg, 1987). Studies also have shown that the hormone-induced System A activity has a half-life about 1.5 hr after hormone withdrawal (Handlogten and Kilberg, 1984). However, no individual integral membrane N-glycosylated amino acid transporter protein has been documented with regard to its de novo biosynthesis and targeting time, as well as its residence time at the PM.



Degradation of Plasma Membrane Proteins

After the biosynthesis and targeting of an integral membrane Nglycoprotein to its functional destination has been completed, the part of its life









cycle that is still to come is the degradation. "Birth" and "death" go hand-inhand, this is true for the fate of a protein as much as for any other "living" things. The homeostasis of membrane proteins inside cells is highly regulated through not only the rate of protein synthesis, but also protein degradation. In response to alterations in the environment, proteins that are no longer needed can be eliminated, and their amino acids can be reutilized. The continued turnover of cellular proteins may also prevent the accumulation of a variety of deleterious nonenzymatic modifications such as oxidation, deamination, and glycosylation. In fact, the reduced degradation rates of certain proteins in aged tissue may contribute to the accumulation of aberrant proteins in aging. Furthermore, certain pathways of proteolysis may have evolved to eliminate mistakes in protein biosynthesis and assembly. Rapid and selective degradation of abnormal and mutant proteins is crucial for cell survival (Doherty and Mayer, 1992).

To date, the mechanism(s) responsible for the degradation of amino acid transporters, as well as for any other mammalian nutrient transporters, have not be documented. However, limited research has been done to elucidate the pathways by which membrane proteins in general are broken down in mammalian cells (Hare, 1990). It is logical to postulate that the degradation of the plasma membrane amino acid transporters may follow one of more of these pathways, described below.

One of the mechanisms by which plasma membrane proteins are

removed and eventually degraded is by a process known as shedding (Hare,









1990; Beaudoin and Grondin, 1991), which means that the proteins are discharged from the cell surface and degraded outside of the cells. Shedding may occur by proteolytic release of the extracellular domains of the membrane proteins or by the release of intact membrane proteins along with membrane lipids in the form of small membrane vesicles. Interestingly, both of these events may be dependent on an initial membrane protein internalization, because lowering the temperature or incubating with lysosomotropic amines prevented the shedding (Johnston and Bystryn, 1984; Teixido et al., 1987). As to the proteolytic cleavage, it was postulated that the internalization of the shed proteins into acidic vesicles would trigger their proteolytic release from the cell surface upon recycling of the vesicles back to the cell surface, probably due to their increased sensitivity to proteases (Hopper et al., 1985; Teixido et al., 1987). For membrane vesicle shedding, the process started with the endocytosis of the clathrin-coated vesicles involving a small domain of the PM containing the condemned proteins, and transfer to multivesicular endosomes. The large vacuoles containing cell surface glycoproteins arising from these vesicles then return their cargo to the extracellular space by exocytosis (Hare, 1990).

The second possible mechanism involves the degradation of membrane glycoproteins in the lysosomes. Endocytosis and lysosomal degradation pathway is well characterized, and abundant evidence shows that it is important for the turnover of many plasma membrane proteins. There are different types of endocytosis in eukaryotic cells, distinguished on the basis of the size of the









endocytic vesicles formed. Among them, the best characterized endocytosis pathway is clathrin-mediated endocytosis, which is also referred to as receptormediated endocytosis (Alberts et al., 1994; Robinson et al., 1996; Steer and Hanover, 1991). The key event during clathrin-mediated endocytosis is the recruitment of soluble clathrin from the cytoplasm onto the intracellular side of the plasma membrane to form a coated-pit. This process is thought to be mediated by the protein complexes called adaptors, which are components of the coat, forming an inner layer and attaching the clathrin to the membrane. One of the hallmarks of clathrin-coated vesicles is their selectivity. Certain membrane proteins, notably receptors for extracellular ligands and some nutrient transporters, such as LDL, Tf, and EGF receptors and GLUT4 transporter, are very efficiently concentrated into clathrin-coated vesicles (Bradbury and Bridges; 1994; Holman et al., 1994). This selectivity of endocytosis would, at least partially, explain why different turnover rates are associated with different proteins, even though all reside at the plasma membrane. In most cases, this selective concentration property of endocytosis has been correlated with the presence of an "internalization signal", for example a Tyr- or Leu-based sorting motif, in the cytoplasmic domain of the membrane protein (Sandoval and Bakke, 1994). There is compelling evidence from in vitro studies showing that the selective sorting is accomplished through the recognition and binding of the internalization signal of the membrane proteins by the adaptors. The clathrin-coated pits then pinch off to form coated vesicles, and this step is regulated by a GTPase, dynamin (Nuoffer and Balch, 1994).









Following the internalization from the cell surface, membrane proteins, lipids, and solutes enter early sorting endosomes. From there, some endocytosed receptors that will be sorted back to the plasma membrane, such as the Tf receptor, also pass through a separation recycling compartment (Steer and Hanover, 1991). This endocytosis-recycling pathway has been best described for several receptor proteins including the LDL and insulin receptors (Alberts et al., 1994; Robinson et al., 1996; Steer and Hanover, 1991). Over the past few years, more and more solute transporters, such as GLUT4, CFTR CI channel, H20 channel, and H+-pump, have been shown to undergo a similar regulated recycling pathway (reviewed by Bradbury and Bridges, 1994).

From early endosomes, internalized molecules can also proceed to

further steps along the endocytic pathway, late endosomes and lysosomes, to be degraded. In vitro assays show that endosomal carrier vesicles, which mediate the vesicle transport between early and late endosomes, can fuse with late endosomes in a microtubule-dependent fashion, but not with early endosomes or with each other (Robinson et al., 1996). It was postulated that microtubles act as tracks along which carrier vesicles can move from one membrane compartment to another. This process is especially important for the transfer of endocytic vesicles from early endosomes to endosomal carrier vesicles and from there to late endosomes. The microtubule depolarizing drug, nocodazole, blocks the transport between early and late endosomes (Gruenberg et al., 1989). Transport from early endosomes to late endosomes, and then to lysosomes, also depends on the acidification of the endosomes and









lysosomes (Steer and Hanover, 1991). The same or similar vacuolar H ATPase is thought to acidify all the endocytic compartments, including early endosomes (pH 6), late endosomes (pH 5.5), and lysosomes (pH 5). This acidic environment plays a crucial part in the function of all these organelles, including the proper vesicle transport, release of bound ligands from their receptors, and proteolysis (Alberts et al., 1994). Most of the hydrolytic enzymes that reside inside lysosomes and late endosomes for protein degradation are most active at low pH. Abrogation of the acidification environment of the endocytic compartments using either bafilomycin A1, a specific inhibitor for the vacuolar H -ATPase, or a weak base, such as NH4Cl or chloroquine, can block the trafficking along the endocytic pathway, as well as inhibit the function of the lysosomal proteases (Steer and Hanover, 1991). Therefore, these inhibitors can be used to test whether the degradation of a particular protein is mediated by the endocytosis/lysosomal pathway.

The third possible pathway responsible for the membrane protein

degradation is the ubiquitination-proteasome pathway (Alberts et al., 1994; Hicke, 1997). Ubiquitin, a small globular protein of 76 amino acids with a protruding carboxyl-terminus, was found to serve as a covalent cofactor for ATP-dependent proteolysis in cytosol (Olson et al., 1992; Hochstrasser, 1996). Ubiquitin is conjugated to the condemned protein via ubiquitin's carboxylterminal glycine residue to form isopeptide bonds with available lysine residues in the target protein. The result is to produce a novel post-translational modification of branched polypeptides. When a targeted protein is poly-









ubiquitinated, a chain of up to 20 ubiquitins can be attached to a single lysine residue of the modified protein. This poly-ubiquitination is accomplished by repeated addition of single ubiqutin units through isopeptide bonds involving lysine-48 of each additional ubiquitin and the carboxyl-terminal glycine residue of each succeeding ubiquitin. Multiple ubiquitination of this type has been shown to serve as a recognition signal for the 26S proteasomal complex, which contains the multicatalytic 20S proteasome as its catalytic core (Alberts et al., 1994; Hochstrasser, 1996)). Ubiquitin is released during protein degradation to be re-used in future rounds of protein catabolism. There are also deubiquitination enzymes that can reverse the ubiquitination reaction and unmodify the target protein (Kalderon, 1996; Wilkinson, 1997). This latter pathway may be part of the quality control system examining each protein continuously and selecting certain protein molecules for degradation while releasing others to continue to function.

The ubiquitination-dependent proteasomal degradation pathway has been best characterized for its pivotal role in regulating the decay of many cytosolic short-lived regulatory proteins, which are involved in a diverse array of regulatory events including cell cycle progression, DNA repair, and transcriptional control. Relevant to the degradation of membrane proteins, Ward and coworkers (1995) first showed that the ubiquitination and proteasome degradation pathway was required for the rapid turnover of the cystic fibrosis transmembrane conductance receptor (CFTR), an integral membrane protein. It was shown that the degradation of both the wild-type and the mutant CFTR was









dramatically reduced by incubation with specific inhibitors for proteasomal proteases. Incubation with these potent proteasome inhibitors caused an accumulation of the ubiquitinated immature CFTR. By expressing a mutant ubiquitin (K48R) and eliminating the formation of the polyubiquitin chain on target proteins, they further confirmed that polyubiquitination is required for the rapid degradation of CFTR.

In contrast to the ubiquitination-proteasome pathway, Strous and

coworkers (1996), using temperature-sensitive mutant CHO cells for ubiquitinactivating enzyme El, showed that ubiquitination of the human growth hormone receptor is required for the ligand-induced endocytosis and degradation through the endosomal/lysosomal pathway. Interestingly, they showed that, ubiquitination of the receptor was dependent on an intact endocytosis pathway, suggesting there was a coupling mechanism between the ubiquitination and endocytosis of this receptor. More recently, Springeal and Andre (1998) reported that, in yeast, ubiquitination of the nutrient permeases is associated with their internalization and degradation. Furthermore, Terrell et al. (1998) published their studies indicating that the mono-ubiquitination of a G proteincoupled receptor can serve as an internalization signal for its endocytosis. This observation suggests that the mechanisms that recognize mono-ubiquitination as an internalization signal for membrane proteins is significantly different from that detecting poly-ubiquitination as a proteasome recognition signal. In another words, the fate of ubiquitinated proteins may be decided by the number and topology of the ubiquitin attached.









Finally, membrane glycoproteins may also be degraded through

individual proteases that are affiliated with the membrane or reside closely to the target proteins. For example, calpain is a calcium-activated protease, which consists of a regulatory subunit and a catalytic subunit (Doherty and Mayer, 1992). The smaller regulatory subunit contains an "EF hand" domain typical for some calcium-binding proteins, and a hydrophobic glycine-rich domain which may associate the enzyme with cell membrane. The other larger subunit also contains an "EF hand" domain and a potent cysteine protease catalytic site. Activation of calpain has been implicated in the degradation of the membrane cytoskeletal protein fodrin (Fukuda et al., 1998; Blomgren et al., 1995; Yokota et al., 1995), actin (Potter et el., 1998), as well as intermediate filament proteins (Resing et al., 1993). Given the previously described association of fodrin with the Na'/K -ATPase (Nelson and Hammerton, 1989) and System A (Handlogten et al., 1996), it is tempting to hypothesize that this or other similar membrane bound or affiliated proteases also specifically act on integral membrane proteins and directly regulate their activity.


C6 Glioma Cell and Its Utilization in SDHA Glutamate/Aspartate Transporter Studies

C6 glioma cells, a rat glial tumor cell line, was originally isolated as a S100 protein-producing clone from rat glial tumors that had been induced by Nnitrosomethylurea. The S-100 protein is a highly acidic protein unique to the vertebrate brain and has been found in numerous brain tumors, both from man and other animals (Benda et al., 1968). This protein was named on the basis of









its solubility in 100 percent saturated ammonium sulfate at neutral pH. Cultured C6 glioma cells have been used extensively to study various aspects of glial biochemistry and physiology. These cells exhibit several biochemical features of normal glial cells, such as expressing S-100 protein (Benda et al., 1968), glial fibrillary acidic protein (GFAP) (Bissel et al., 1974), glutamine synthetase (GS) (Parker et al., 1980), and 2',3'-cyclic nucleotide-3'-phosphohydrolase (CNP) (Zanetta et al., 1972). Brain glial cells can be classified into three subtypes: astrocytes, oligodendrocytes, and microglial (Cooper et al., 1996). Astrocytes are present mainly in regions of axons and dendrites; they tend to surround and closely contact the adventitial surface of blood vessels. Besides the possible insulation and organization roles suggested by its structural characteristics, astrocytes can accumulate glucose, synthesize glycogen, and provide energy substrates to neurons. Whereas, the oligodendrocytes form the myelin sheath along the axons, and the microglial may play a role in signaling for the recruitment of lymphocytes and leukocytes during the repair of the damaged brain tissue (Cooper et al., 1996). GFAP and GS are considered as specific marker proteins for astrocytes, whereas CNP is a marker for oligodendrocytes (Varon, 1978). The presence of both astrocyte and oligodentrocyte properties in C6 cells suggest that they may be most closely comparable to the lessdifferentiated glial stem cells present in the developing brain, which further differentiate into either of the mature glial types (Bhat et al., 1984).

The expression of these unique cell type properties can be regulated within C6 cells by a variety of culture conditions, including cell density, serum









deprivation, and culture passage. Varon (1978) first found that the quantity of GFAP, a specific marker for mature astrocytes, was ten-times greater in stationary phase than in log phase C6 cells. Later, Maltese and Volpe (1979) showed that the expression of an oligodendrocyte marker, CNP, was also significantly up-regulated by growing C6 cells to high density or by culture of non-confluent cells in serum-free medium. When GS and CNP were used to study C6 cells in culture, Parker et al. (1980) reported that at early passages the cells predominantly showed oligodendrocyte-like properties, whereas at later passages they predominantly showed astrocyte-like properties. All of these observations are consistent with the proposal that C6 glioma cells may be analogous to glial stem cells, which can be differentiated to specific mature glial-like cells under different culture conditions.

C6 glioma cells have a high level of SDHA glutamate/aspartate transport activity (Deas and Erecinska, 1989). As to which of the SDHA glutamate transporters are responsible for the high transport activity seen in this cell line, there have been some conflicting reports. Casado and coworkers (1993), using C6 cells, studied the regulation of SDHA glutamate transporter(s) by protein phosphorylation through the PKC pathway. They reported that phosphorylation of the GLT1 transporter protein itself was responsible for the increased transport activity and suggested that the glial-specific transporter GLT1 was expressed in the C6 cells. In contrast, Palos and coworkers surprisingly observed that mRNA for the neuronal specific transporter, EAAC1, but neither of the glial-specific transporters, GLAST or GLT1, could be detected in C6









glioma cells using Northern blotting analyses (Palos et al., 1996). Consistent with these Northern blotting data, Dowd et al. (1996) showed that C6 glioma cells expressed EAAC1 but not GLAST, GLTI, or EAAT4 based on immunoreactivity as detected by Western blotting. The same authors further reported that the EAAC1 -mediated SDHA glutamate transport activities in C6 and in EAAC1 mRNA-injected oocytes shared similar kinetic parameters, but that these kinetics were different from those observed in rat synaptosomes, suggesting that synaptosomes contain a different, or at least a heterogeneous population, of glutamate transporters. The effect of protein phosphorylation on the EAAC1 transporter, directly or indirectly, has been studied using C6 glioma cells as a model system (Dowd and Robinson, 1996; Davis et al., 1998). These investigations showed that uptake activity and cell surface expression of EAAC1 was increased by phorbol ester treatment and with using specific inhibitors, the experiments suggested that both the PKC and PI3K pathways might play roles in regulating EAAC1 transporter function.

The expression of a neuronal-specific SDHA glutamate transporter, that is, EAAC1, in a glial tumor cell line is surprising but not unprecedented. It has been documented that some types of glial tumor cells express glucose transporters, GLUT1 and GLUT3, not usually expressed by normal glial cells (Boado et al., 1994). Similarly, certain PC12 pheochromocytoma cell variants were found to be more sensitive to a toxic aspartate analog, alanosine, than wild-type PC12 cells, suggesting an up-regulation of the aspartate transport activity in tumor cells (Ramachandran et al., 1993).









High-grade astrocytomas, of which C6 may represent a model system, are the most common form of malignant brain tumor in humans, and they are usually resistant to therapy (Bruner, 1994; Lesser and Grossman, 1994). Therefore, understanding the atypical expression of nutrient transporters in these tumor cells may be useful for future clinical diagnosis and chemotherapy.

In this work, I used C6 glioma cells as a model system to study the biosynthesis, degradation, and intracellular targeting of EAAC1 glutamate/aspartate transporter protein. These results obtained from this study may provide us with the basis for studying the role of EAAC1 transporter under diseased states, as well as may serve as a model for studying the life cycle of other mammalian transporter proteins.














CHAPTER 2
MATERIALS AND METHODS


Materials

L-[2,3-3H]-Aspartic acid (1 mCi/ml) (Cat# TRK445), protein A-HRP (Cat# NA9120), HyperfilmTM MP (Cat# RPN1678H), and PromixTM L-[35S]-MethionineCysteine in vivo cell labeling mix (14.3 mCi/ml) (Cat# SJQ0079) were obtained from Amersham LIFE SCIENCE. The RT-PCR reaction reagents were all obtained from Gibco BRL: random hexemers (Cat#48190 - 011), random primers (Cat# Y01212), Superscript II reverse transcriptase (Cat# 18064 - 014), and Tag DNA polymerase (Cat# 18038-042). Goat-anti-rabbit IgG-HRP (Cat# 170-6515), pre-stained protein standards (broad-range) (Cat# 16100318) were obtained from BioRad Laboratories. The EZ-LinkTM Sulfo-NHS-LC-Biotin (Cat# 21335), free D-biotin (Cat# 29129), SuperSignal ULTRA chemiluminescence substrate (Cat# 34075), ImmunoPure immobilized monomeric avidin-sepharose (Cat# 20228) were obtained from PIERCE. L-methionine- and L-cysteine-free DMEM (Cat# 21013) was obtained from Gibco BRL. Protein endoglycosidase F (Cat# 704S) and endoglycosidase H (Cat#702S) were obtained from NEW ENGLAND Biolabs Inc. SuperFect transfection reagent was obtained from Qiagen. Goat-anti-mouse IgG-HRP was obtained from Kirkegaard & Perry Laboratories (Cat# 074-1807). Minimum Essential Medium (MEM) (Cat# M-









0643), protein A sepharose CL-4B (Cat# P3391), D-aspartate (Cat# A-8881), Lmethionine (Cat# M-9625), and L-cysteine (Cat# C-7755), polyoxyethylene 9 lauryl ether (C12E9) (Cat# P-9641), non-immune normal rabbit IgG (Cat# 18140), tunicamycin (T-7765), NH4CI (Cat# A-4514), N-CBZ-Leu-Leu-Leucinal (MG132) (Cat# C-221 1), N-acetyl-leu-Leu-norleucinal (ALLN) (Cat# A-6185), leupeptin (Cat# L-2884), phenylmethyl-sulfonyl fluoride (PMSF) (Cat# P-7626), brefeldin A (BFA) (Cat# B-7651), and nocodazole (NOC) (Cat# M-1404) were all obtained from Sigma Chemical Co. All other chemicals were obtained through either Sigma Chemical Co. or Fisher.



Methods



Cell Culture

C6 glioma cells and BNL CL.2 cells were obtained from American Type Culture Collection (ATCC number for C6 is CCL107; for BNL CL.2 it is TIB 73) and maintained in supplemented Eagle's medium (MEM) containing 10% FBS as monolayer cultures under a humidified atmosphere of 5% CO2/95% air (370C) for a maximum of eight passages in 75 or 175 cm2 flasks. The cultured cells were transferred to 24-well cluster dishes for whole cell transport assays; to plastic 100 - 150 mm culture dishes for metabolic labeling, cell-surface biotinylation, and total cellular protein or membrane protein collection; or to 22 x 22 mm sterilized Corning glass microscope cover slips held in a Falcon six-well cluster trays for immunohistochemistry.











Whole Cell Transport Assay

Amino acid uptake of C6 glioma cells or other adherent cultured cells was measured by the cluster tray method of Gazzola et al. (1981) with modifications by our laboratory (Kilberg et al., 1989). One hundred thousand C6 cells were placed into each well of a 24-well tray and cultured for 20 to 24 hr under normal conditions. To partially deplete the intracellular pool of amino acids to minimize trans-effects on transport (Kilberg, 1989) as well as to remove extracellular Na , cells were incubated at 370C twice for 15 min each (2 x 15 min) in choline-KRP. To initiate transport, 3H-amino acid and the appropriate inhibitor in 250 PIl of either NaKRP or choline-KRP (370C) was added simultaneously to each of the 24 wells in the cluster tray. The Na+-dependent transport is taken as the difference between uptake in NaKRP and choline-KRP. The transport measurement was terminated by discarding the radioactivity and rapidly washing the cells five times with 2 ml ice-cold choline-KRP. The data are expressed as pmolomglprotein.time1' and typically, are presented as the average of 4 independent assays on at least two different batches of cultured cells.



Gel Electrophoresis

Gel electrophoresis was performed essentially following the protocol

originally described by Laemmli (Laemmli, 1970). Protein samples were diluted with at least equal volume of sample dilution buffer (2X SDB) consisting of 2% (w/v) SDS, 5% -mecaptoethanol, 30 pg/ml bromophenol blue, 20% glycerol,









0.125 mM Tris-HCI, pH 6.6 -6.8. The amount of protein loaded per lane will be stated in each figure legend. Vertical 7.5% polyacrylamide slab gels were used and electrophoresis was performed at 30 to 40 mA constant current until the selected pre-stained protein marker band reached the end of the separating gel.



Electrotransfer and Immunoblottinaq

For immunoblotting detection, the fractionated proteins were transferred electrophoretically onto a piece of nitrocellulose membrane in ice-cold transfer buffer containing 25 mM Tris-base, 190 mM glycine, 20% methanol at 299 mA and constant current for 18 to 20 hr. After the transfer, the blot was stained briefly in fast green stain (0.1% fast green, 50% methanol, 10% acetic acid) and de-stained (50% methanol, 10% acetic acid) to check for the efficiency of transfer and the evenness of loading. The blots were blocked in Tris-buffered saline/ 0.1% Tween-20 (TBS-T) (10 mM Tris, pH 7.5, 200 mM NaCI, and 0.1% Tween) containing 1% Carnation non-fat dry milk (NFDM) for 1- 2 hr at room temperature (RT) or overnight at 40C with constant agitation on an orbital shaker. The blots then were incubated in the same blocking solution containing primary antibody (1: 2000 dilution for EAAC1 antibody) for 1 - 2 hr at RT in test tubes with constant "end-to-end" rotation. After extensively washing in TBS-T/1% NFDM to remove the unbound antibodies, the blots were incubated in the same blocking buffer with secondary antibody conjugated to horseradish peroxidase (1:20,000 dilution of goat anti-rabbit IgG-HRP was used for EAAC1 blotting) for 1 hr at RT. The blots were extensively washed with TBS-TI 1% NFDM before









being visualized with SuperSignal Chemiluminescence detection reagents (PIERCE) following the manufacturer's instructions. Light emissions from the blots were captured on Hyperfilm MP (Amersham), and band intensity was quantitated in the linear range of the film on a Visage Bioscan video densitometer.

As shown in Figure 2-1, I typically detect three major bands on EAACI

Western blotting which run with apparent molecular weights of 73, 145, and 219 kDa. All of these three protein bands were detected when either total C6 cellular membrane protein or anti-EAAC1 antibody immunoprecipitated protein was tested by immunoblotting. Pre-incubation of the anti-EAACl antibody with 10 4g/ml of specific EAAC1 -MBP fusion protein completely inhibited the detection of all these bands. These results indicate that all these detected bands are specific EAAC1 protein bands, which may represent the monomer, dimer, and trimer form of EAAC1 protein, as also observed by others (Haugeto, et al., 1996; Davis et al., 1998).



Pulse-Chase Metabolic Labelinq of C6 Glioma Cells

To study the de novo biosynthesis, intracellular targeting, and

degradation of EAAC1 transporter proteins in C6 cells, pulse-chase labeling with L-[wS]-methionine-cysteine (ProMix, Amersham) was used. After placing 6 to 9 x 106 cells onto each 100 mm culture dish or 2.3 x 107 cells onto each 150 mm dish, cell monolayers were cultured for 22 to 24 hr in normal MEM culture medium to permit growth to near confluence. The cells were washed once with









sterile 370C PBS, pH 7.4, and incubated with 15 ml /58 cm2 surface area of methionine- and cysteine-free DMEM medium (Gibco) for 2 x 15 min at 370C to deplete the intracellular pool of free methionine and cysteine. The depletion medium then was aspirated and the cells incubated with 200 pCi/ml of [3S]methionine-cysteine (PreMix cell labeling, Amersham) in methionine- and cysteine-free DMEM medium at 370C for 15 to 30 min. The cells were washed twice with 370C MEM containing 5 mM of non-radioactive methionine and cysteine (chasing medium) and then transferred to fresh chasing medium and chased for 0 to 60 hr followed by immunoprecipitation of the EAAC1 protein. For experiments have chases longer than 24 hr, 1% FBS was added to the medium.



Immunoprecipitation and Fluorography

Immunoprecipitation of the EAAC1 transporter protein was performed

following the procedure outlined by Harlow and Lane(1988)with modifications. After pulse-chase labeling with L-[35S]-methionine and L-["S]-cysteine, cells were washed twice with ice-cold PBS and once with SEB buffer (250 mM sucrose, 2 mM EDTA, 2 mM EGTA, and 10 mM HEPES, pH 7.5), and then frozen in 2.5 ml of SEB buffer containing proteinase inhibitors (1 mM PMSF and 2 pg/ml each of leupeptin, aprotonin, pepstatin, TPCK, and TLCK) at -800C. For analysis, cells were thawed on ice and another 2.5 ml of hypotonic EB buffer (2 mM EDTA, 2 mM EGTA, and 10 mM HEPES, pH 7.5) were added. Cells were scraped from the plates and homogenized on ice with 10 to 15 passes through a pre-chilled steel-block cell homogenizer with a clearance of 0.0025 inches









(Auburn Tool & Dye, Warwick, RI). The cell homogenate was centrifuged at 400 x g for 10 min to get rid of the unbroken cells and nuclei, and the supernatant was centrifuged at 280k x g (65k rpm with Beckman Ti70.1 rotor) for 1- 2 hr at 40C to collect the total membrane pellet, which was then extracted in PES buffer (2% C12E9, 0.1% SDS, 1 mM EDTA in PBS, pH 7.4) for 1 hr on ice with constant magnetic stirring. After centrifugation at 200k x g for 15 min at 40C, the supernatant was recovered and the protein concentration was determined by a modified Lowry method (Kilberg, 1979) or the bicinchoninic acid (BCA) method (Pierce BCA Protein Assay kit), and then an equal amount of starting protein (100 to 500 pg) from each sample was transferred into microcentrifuge tubes, brought up to a volume of 500 p1 with PES buffer, and used for the immunoprecipitation assay. To minimize non-specific interaction between labeled membrane proteins and the immunoglobulins, an unrelated non-immune rabbit serum IgG (5 pg) and a 50% (in PES buffer) suspension of protein Asepharose (50 pl) were added to the extracts and incubated ("pre-cleared") for 13 hr at 40C with constant mixing. Samples were then centrifuged at 100k x g for 15 sec to collect the protein A sepharose - IgG complexes and this "pre-cleared" supernatant was transferred to a new microcentrifuge tube. The specific antibody against EAAC1 (5 pg purified total IgG, # UF91) was added to the precleared extract and incubated at 40C overnight with constant mixing. A 50 pl aliquot of 50% protein A-sepharose beads was then added and incubated for 2 hr to collect the immunoprecipitates. The pellets were washed with PES extraction buffer (3 x 1 min) and with PES containing 0.35 M NaCI (for a total salt









concentration of 0.5 M) (4 x 10 min) at 40C with constant mixing. As shown in Figure 2-2, a series of salt washes were tested, and the data show that the binding between the antibody and the EAAC1 protein can withstand salt washing upto 1M NaCI and 0.5 M salt is sufficient to eliminate basically all of the nonspecific binding. The immunoprecipitated proteins were then eluted from the beads with sample dilution buffer (SDB) containing 6 M urea and 10% ME for 20 min at 370C and separated by SDS-PAGE. The elution condition was determined by the data shown in Figure 2-5 below. Heating at 370C for 30 min was shown to be the optimal elution condition, and was used for all the experiments described in this thesis. For Western blotting, the separated proteins were then electrotransferred onto a nitrocellulose membrane and immunoblotted with a specific anti-EAAC1 antibody as described before. For fluorography, the gels were fixed at RT in 10% TCNA 40% methanol for 30 min, soaked in water for 30 min, and then incubated in 1 M sodium salicylate for 1 hr before drying at 650C under vacuum. The dried gels were exposed to autoradiographic film at -800C with an intensifying screen, and the band intensity was quantitated in the linear range of the film on a Visage Bioscan video densitometer.

To test whether 5 pg of anti-EAAC1 antibody (purified total IgG from the immune rabbit serum) is enough to precipitate the EAAC1 protein from up to 500 Pg starting material, immunoblotting was done to check for the remaining of EAAC1 protein in the immunoprecipitation supernatant. As shown in Figure 2-3,

5 pg of anti-EAAC1 antibody completely depleted the EAAC1 protein from the starting mixture. Also, the amount of EAAC1 protein immunoprecipitated by 5 pg









anti-EAAC1 antibody is proportional to the amount of starting protein up to 500 pg (Figure 2-4). All these data suggest that 5 pg anti-EAAC1 antibody is sufficient to quantitatively immunoprecipitate EAACI protein from up to 500 ag starting material.



Endoqlycosidase Digestions

To test for the endoglycosidase H (Endo H)-sensitivity of the newlysynthesized EAAC1, C6 cells cultured on the 150 mm dishes were pulse-labeled for 15 min with [S]-methionine-cysteine and chased with medium containing an excess of unlabeled methionine and cysteine (5 mM each) for 30 to 240 min as described before. At the end of each chase period, EAAC1 was immunoprecipitated and then the transporter was eluted from the protein Asepharose beads with 10 pl of 5X denaturing solution containing 2.5% SDS, 5% B3-ME in water for 30 min at 370C. After dilution to 1X denaturing solution with 40 pl of water, each of the eluates was collected and then divided into two microcentrifuge tubes (- 20 pl each). A 1/10 volume of 500 mM sodium citrate, pH 5.5 and 500 - 1000 U (1 -2 pl) of Endo H (BioLabs Inc.) was added to each of the tubes. For the non-PNGase F control tube, 1 - 2 pl of enzyme storage buffer (20 mM Tris-HCI, pH 7.5, 50 mM NaCI, and 5 mM Na2EDTA), instead of the enzyme, was added. All samples were incubated at 370C for 1 hr, and then mixed with an equal volume of 2X SDB buffer before the samples were loaded onto SDS-PAGE gel for separation and autoradiographic detection.









To determine the N-linked glycosylation of both the newly synthesized lower molecular weight and the mature forms of the EAAC1 transporter protein, protein endoglycosidase F (PNGase F) digestions were performed. Total C6 cellular membrane protein was collected from the cells that were either pulsechase labeled as described above or not, and then solubilized with PES buffer and immunoprecipitated with specific anti-EAAC1 antibody as described before. The precipitates were then eluted with 10 ipl of 5X denaturing solution (2.5% SDS, 5% 3-ME) for 30 min at 370C, and the collected eluates were diluted to 1X denaturing solution with water as described above for Endo H digestions. For PNGase F digestion groups, 1/10 volume each of 10% NP-40 and a buffer consisting of 10% NP-40, 500 pM sodium citrate, pH 7.5 as well as 500 - 1000 U of PNGase F (BioLabs Inc.) was added. For the non-Endo H controls, the enzyme was replaced with equal volume of enzyme storage buffer only. All the samples were incubated at 370C for 15 to 240 min (a 60 min incubation time was commonly used), then mixed with an equal volume of 2X SDB buffer before they were loaded onto a SDS-PAGE gel for separation and autoradiographic detection.



Cell Surface Protein Biotinylation

Cell membrane impermeable sulfo-NHS-LC-biotin was used to specifically label proteins that are exposed at the exterior surface of the cells. This cellsurface biotinylation method was utilized to study the plasma membrane (PM) localization and half-life of cell-surface exposed EAAC1 protein, as well as to









determine the targeting time and residence time of newly synthesized EAACI in C6 glioma cells. For the determination of the cell-surface localization and the half-life of the PM EAACI protein, C6 glioma cells cultured on 100 mm dishes were washed twice with NaKRP buffer (119 mM NaCl, 5.9 mM KCI, 1.2 mM KHCO3, 5.6 mM glucose, 25 mM Na2HPO4, 0.5 mM CaCI2, 1.2 mM MgSO4, pH

7.5), and incubated with 0.5 - 1 mg/ml of sulfo-NHS-LC-biotin (Pierce Chemical Co.) in NaKRP for 30 min to 1 hr at 4 - 150C. The optimum time and reagent concentration must be determined experimentally, and the specific conditions for each experiment are given in the figure legends. The biotin reagent buffer was aspirated and the cells were rinsed twice with fresh MEM containing 50 mM glycine, and then incubated in fresh 370C MEM + 1% FBS for specific chase times ranging from 0 to 24 hr at 370C. At the end of each chase time period, the cells were washed once with ice-cold NaKRP, another time with ice-cold SEB containing protease inhibitors (described above), and then frozen in 2 ml SEB containing protease inhibitors at -800C. The specificity of the cell surface biotinylation assay was examined, and the data are shown in Figure 2-6. The percentage of the biotinylation for the plasma membrane proteins, Na /K ATPase and EAAC1, was 89% and 69%, respectively; versus below 2% for the intracellular proteins, cytoplasmic AS and ER GRP78. These results indicate that this method reliable for the study of the PM-localization of a protein.

To determine the time it takes for the newly synthesized EAAC1

transporter proteins to arrive at the plasma membrane of C6 glioma cells, cell surface biotinylation was performed after the C6 cells were metabolically labeled









with 35S-methionine and -cysteine for 15 - 30 min and chased in medium containing an excess amount of non-radiolabeled methionine and cysteine for 0 to 36 hr. At the end of each chase time point, these metabolically-labeled C6 cells were washed twice with NaKRP and cell-surface biotinylated with 0.5 - 1 mg/mi of sulfo-NHS-LC-biotin in NaKRP at 150C for 30 - 60 min. After aspirating the biotinylation reagents, the cells were rinsed once in NaKRP and then incubated with NaKRP containing 50 mM glycine for 2 x 15 min at 4 - 150C to quench the remaining free biotin. Then the cells were washed once with ice-cold SEB containing protease inhibitors, and frozen at -800C until all samples were collected. All cell-surface biotinylated C6 cells were thawed on ice and homogenized with a steel-block homogenizer, and total cellular membrane proteins were collected and solubilized with PES buffer as described above. Equal amounts of solubilized proteins was then subjected to two precipitations: first with anti-EAAC1 antibody and protein A-sepharose beads to precipitate all of the EAAC1 protein, and after elution in low pH buffer, the eluates were precipitated again with monomeric avidin-sepharose beads to precipitate only the biotinylated EAAC1. For the immunoprecipitation of total EAAC1 protein, the protein samples were pre-cleared with non-immune rabbit IgG and then immunoprecipitated with specific anti-EAAC1 antibody as described before. The precipitated EAAC1 protein was eluted from the beads with 1 ml of 0.1 M glycine, pH 2.8 containing 0.5% TritonX-100 and 0.2% BSA at RT for 30 min. After centrifugation at 10 Ok x g for 15 seconds to remove the beads, the eluate was removed and neutralized with 50 pl of 0.1M Tris-HCL, pH 9.5. For the









precipitation of biotinylated EAAC1 protein, 25 ~pl of packed monomeric avidinbeads was washed twice with 1 ml of PES buffer and then pre-incubated for 1 - 2 hr with non-biotinylated total C6 cellular membrane proteins to block the nonspecific binding sites on the beads. Then the immunoprecipitated EAAC1 protein collected above was added to the pre-treated monomeric avidin-beads and incubated for 5 - 18 hr with constant mixing at 40C. At the end of the incubation, the beads were washed extensively with 1 ml of PES supplemented with 350 mM NaCI for 40 min at 40C with a total of six changes of buffer. The double precipitated proteins were eluted from the beads with 2X SDB containing 2 mM free D-biotin and separated by gel electrophoresis followed by fluorography as described before. Therefore, only the newly synthesized ("S-labeled) EAAC1 protein that has already arrived at the PM (cell surface-biotinylated) were detected with this double precipitation procedure. Monomeric avidin instead of streptavidin was used because the binding between biotin and monomeric avidin is reversible, whereas, the binding between biotin and streptavidin is nearly irreversible, even in sample dilution buffer. Therefore, when monomeric avidinsepharose is used, the precipitated biotinylated proteins can then be eluted from the beads with an excess amount of free-biotin.



Immunohistochemistry

The immunohistochemistry of human fibroblasts was done as described by Woodard et al (1994). Cells cultured to 50 - 60% confluence on glass cover slips were rinsed in PBS and fixed with 4% paraformaldehyde/PBS for 30 min,









then rinsed 3 times with PBS followed by quenching of free aldehyde by incubating for 30 min in PBS containing 50 mM glycine. Non-specific antibody binding was blocked by incubation in PBS containing 20% normal goat serum (NGS) for 2 hr. For pre-immune or immune serum incubation, cover slips were removed from the six-well trays, inverted and placed on 50 pl drops of antiserum diluted, to the appropriate level for each antibody, in PBS containing 20% NGS with or without 20 pg/ml of competing peptide. After incubation for 2 hr, the cover slips were placed back into the wells and then washed 3 times with PBS. The coverslips were incubated with the secondary antibody (typically 1:200) in 20% NGS/PBS for 1 hr in a similar manner as that used for the primary antibody. After the final wash, the coverslips were mounted onto glass slices with a drop of Fluoromount-G, allowed to dry, and the edges of the cover sip sealed with fingernail polish before being analyzed by fluorescent microscopy.



Expression of EAAC1 Glutamate/Aspartate Transporter in BNL CL.2 Cells

BNL CL.2 cells were both transiently and stably transfected with EAAC1 cDNA using SuperFect transfection reagent following the vendor's protocol (QIAGEN). The cells were cultured to about 40% confluency on 100 mm dishes and washed once with PBS. Twelve pg EAAC1 cDNA in PcDNA3 vector was mixed with 360 pl OptiMEM, and then with 48 pl of SuperFect transfection reagent. The mixture was incubated at RT for 5 - 10 min, and then diluted into 5 ml of OptiMEM immediately transferred onto the cells. After incubation with the transfection reagent for 2 - 3 hr at 370C in a cell culture incubator, the cells were









washed twice and the medium replaced with fresh MEM containing 10% FBS and 4.5 g/L glucose. For stable transfection, the culture medium contained 0.5

1 mg/L of G418 and the expressing cells were selected for at least 2 - 4 weeks with medium changed every 3 days. The cells were passed whenever they reached about 85 - 90% confluency.



RNA Isolation and RT-PCR

Total RNA from 1 x 108 C6 cells was isolated using RNeasy Midi Kits

(QIAGEN) and then poly(A ) mRNA was purified from 500 gg of total RNA using Oligotex mRNA isolation kits (QIAGEN) following the vendor's protocols. The Reverse Transcriptase reactions were carried out following a standard protocol provided by the vendor of the transcriptase using random hexamers to prepare the first strand cDNA (Gibco). PCR primers for each of the glutamate transporters were chosen based on published sequences of the following transporters: rat EAAC1 (5'-GGT GTC GCT GCA CTG GAT TCC AAC G-3', 5'GGC CAT ATA AAG GCC CAA CTT GCG G-3'), rat GLTI (5'-AGG AGC CAA AGC ACC GAA ACC-3', 5'-TCC AGG CCC TTC TTG ATA ACG-3'), rat GLAST1 (5'-TTG GAT TTG CCC TCC GAC CG-3', 5'-GGT GCA TAC CAC ATT ATC ACC GC-3'), and human EAAT4 (5'-TGC GCC CAT ATC AGC TCA CCT ACC3', 5'-TGC CCA GCC TCA TAA TAG CC-3"). Rat brain cDNA was generated from Poly(A)+-selected mRNA by reverse transcriptase reactions and then used as a positive control for all PCR reactions. Thermal cycling using Taq DNA polymerase included 25 cycles at 940C for 1 min, 600C for 1.5 min, and 720C for






47

2 min. The Taq DNA polymerase was added after the heat denaturation at 940C had started. All reactions contained 4 mM MgCI2. The PCR products were visualized by ethidium bromide staining and photographed under UV light.















Figure 2-1 EAAC1 Protein Competition for Immunoblotting A 40 pg aliquot of total C6 glioma cellular proteins (Total) and anti-EAAC1 immunoprecipitates from 50 pg of starting protein (I.P.) were separated on SDS-PAGE. In the absence of competing fusion protein (original antigen) (-EAAC1-MBP), immunoblotting with anti-EAAC1 antibody (1:2000 dilution) as primary antibody and Protein A-HRP (1:10,000 dilution) as secondary antibody detected 3 distinct bands, all of which were competed away by pre-incubating with 10 pg/ml of specific EAAC1 -MBP fusion protein for 3 hr at RToC (+EAAC1 -MBP). The results presented here are representative of at least two independent experiments.










EAAC1 Peptide Competition in Western Blot

" 4 219 kDa o4 145 kDa



g0 -* 73 kDa


I.P. Total (+ EAACI-MBP)


Total I.P. (- EAACI-MBP)















Figure 2-2 Effect of salt on immunoprecipitation of EAAC1 Protein A 500 Pg aliquot of solubilized total C6 membrane protein was subjected to immunoprecipitation with 10 Pg of antiEAAC1 antibody or pre-immune IgG as described in the Methods Section. The precipitates were then divided into 5 groups and each was washed with PES containing a different concentration of NaCI, and then analyzed by SDS-PAGE and Western blotting. These data show that immunoprecipitation by the anti-EAAC1 antibody is very specific; 500 mM NaCI is the salt concentration used for the rest of the immunoprecipitation experiments. The results shown here represent at least two independent experiments.










I.P. of EAAC1 and Washed with Different Salt Concentrations


0"A 4 40


0.15 0.3 0.5 0.8 1.0 0.15 0.3


1 4, 1 5 145kDa I 4 73 kDa


0.5 0.8 1.0 M NaCI


( Pre-immune IgG )


( Anti-EAAC1 Ab)













Figure 2-3 Effectiveness of immunoprecipition of EAAC1 protein A 500 pg aliquot of solubilized total C6 cellular membrane protein was subjected to immunoprecipitation with 5 pg of either anti-EAAC1 antibody (Anti-EAAC1) or pre-immune IgG (P.I.). The supernatant from each precipitation was collected, and 15 pg of protein was loaded onto SDS-PAGE and assayed for EAAC1 Western blotting as described in the text. As a control, 15 pg of starting protein was also tested (Total). The blot shown is representative of several independent experiments using initial protein amount from 500 pg to 1 mg. The complete loss of EAACl immunoreactivity in the supernatant after EAAC1 immunoprecipitation demonstrated antibody specificity and showed that 5 pg of EAAC1 antibody is sufficient for complete transporter precipitation from 500 pg of starting material.











by Immunoprecipitation


4- 145 kDa


b 4- 73 kDa


Total Anti-EAAC1


Depletion of EAAC1


P.I.




















Figure 2-4 EAAC1 Immunoprecipitation Protein Concentration Curve (A) Different amounts of (10 to 500 pg) of solubilized total C6 membrane protein was subjected to immunoprecipitation with 5 pg of anti-EAAC1 antibody (AntiEAAC1) or pre-immune IgG (P.I.), as described in the text. After extensive washes, the precipitates were eluted in 40 pl of 2X SDB, run on SDS-PAGE, and then analyzed by immunoblotting using EAAC1 antibody, using the antibody conditions described in the text. (B) Densitometry analysis of EAAC1 protein from panel A. The densities of the EAAC1 mature monomer bands in each lane was quantified as described in the text and plotted against the initial protein amounts. The results represent a single experiment.



























50 100 250

Anti-EAAC1


4 145 kDa W 473 kDa 500 [ig Starting
I protein


0 100 2o00 300 400 5o00
StartingProtein (vg)


500

P..


I

















e q~- 4 145 kDa


4 73 kDa



Temp (oC): 4 25 37 37 65 95 Time (min): 30 30 10 30 10 10




Figure 2-5 Determination of the Elution Conditions for Immunoprecipitated EAAC1 A 500 pg aliquot of solubilized C6 membrane protein was subjected to immunoprecipitation with 5 ig of anti-EAAC1, as described in the text. After the washing steps, the pellet was then divided into six aliquots, and each was subjected to a different elution condition (elution temperature and time) as labeled in the figure. The eluates were collected, resolved on SDS-PAGE, and analyzed by Western blotting. The results represent a single experiment. Heating at 370C for 30 min was the elution condition used for the rest of the experiments described in this thesis.








Figure 2-6 Specificity of Cell Surface Protein Biotinylation C6 glioma cells on 150 mm dishes were cell-surface biotinylated with or without 1 mg/ml of sulfo-NHS-LC-biotin for 1 hr at 150C, as described in the text. Cells were then harvested and homogenized, and the total cellular protein was solubilized with PES buffer. Equal amounts of solubilized proteins from the biotinylated and non-biotinylated groups were used for monomeric avidin-sepharose precipitation as described in the text. The precipitates were eluted, analyzed with SDS-PAGE, and immunoblotted with specific antibodies. The immunoblots were quantified by densitometry for each of the specific proteins shown (wide hatched bars). The amount of non-specific binding to avidin detected with the non-biotinylated samples was subtracted from the values obtained for each biotinylated protein to give the amount of specific biotinylation. For EAAC1 protein, this non-specific value was 2.7% of the total and was considered insignificant for the remainder of the experiments presented. To demonstrate the total amount of each protein studied, equal amounts of total solubilized total membrane protein were subjected to immunoblotting with each of the antibodies indicated (narrow hatched bars). For the immunoblotting, 1:2000 dilution of anti-EAAC1, 1:1000 dilution of anti-GRP78, and undiluted anti-AS and antiNa'/K' ATPase antibodies were used as primary antibodies, and then 1:20,000 dilution of either goat-anti-rabbit IgG (for EAAC1 and GRP78) or goat-anti-mouse IgG (for AS and Na/K' ATPase) were used as the secondary antibodies. The results represent a single experiment.

















The Specificity of Cell Surface Protein Biotinylation


S100 / Total ET Specific biotinylation E 80
0


C
o






S20


0
S0 1 7
AS GRP 78 EAACl NaIK+ ATP ase















CHAPTER 3
ACTIVITY, PROTEIN CHARACTERISTICS, AND BIOSYNTHESIS OF EAAC1 GLUTAMATE/ASPARTATE TRANSPORTER IN C6 GLIOMA CELLS Introduction

A family of Na+-dependent high affinity (SDHA) glutamate/aspartate

transport systems, previously referred to collectively as System XAG', is essential for the glutamatergic transmission in the central nervous system (CNS), as well as for many other cellular functions. System XAG activity can be determined as Na'-dependent and D-aspartate inhibitable L-glutamate or L-aspartate transport, and five distinct transporter cDNAs encoding this activity have been cloned in the past six years (refer to Chapter 1 for overview). Excitatory amino acid carrier 1 (EAAC1), a member of this SDHA glutamate transporter family, was first identified from rabbit intestine by Kanai and Hediger using oocyte expression cloning in 1992 (Kanai et al., 1992). The rat EAAC1 homolog, which was later cloned by Velaz-Faircloth and coworkers (1995), shares 95% similarity and 90% identity with the rabbit clone. Rat EAAC1 cDNA encoded a protein of 523 amino acids with a predicated non-glycosylated core molecular weight of 56.8 kDa. Sequence analysis predicted an integral membrane protein with ten putative transmembrane domains and intracellular C- and N-termini. There are six putative protein kinase C phosphorylation sites. A recent publication (Davis et al., 1998) indicated that EAAC1 transport activity as well as its cell surface









expression might be regulated by protein phosphorylation, although direct evidence is needed to show the EAAC1 protein is actually phosphorylated. Nevertheless, another member of this SDHA glutamate transporter family, GLT1, was shown to be phosphorylated when the cells were treated with phorbol ester (Casado et al., 1993). EAAC1 also contains four potential N-linked glycosylation sites (Asn-X-Thr/Ser), three of which are localized within the second extracellular loop. It has been documented that EAAC1 protein is N-glycosylated under normal conditions in rat brain and several cell lines (Dowd et al., 1996). Different apparent molecular weights of EAAC1 protein were detected in rat brain, EAAC1 cRNA-injected oocytes, and C6 glioma cells, which show 68 kDa, 80 kDa, and 78 kDa bands, respectively (Dowd et al., 1996; Yang and Kilberg, unpublished data). But after the cleavage of the N-linked oligosaccharide chain(s), only the EAAC1 protein core was detected at ~ 57 kDa (Dowd et al., 1996). This result suggested that the N-glycosylation of EAAC1 might vary according to the tissue/cell type. The possible physiological relevance of this differential Nglycosylation of EAAC1 is not known.

EAACI transporter is the most widely distributed transporter of the anionic amino acid transporter family (reviewed by Malandro and Kilberg, 1996). EAAC1 mRNA or protein have been localized in kidney, small intestine, heart, lung, skeletal muscle, liver, placenta, human fibroblasts, as well as brain. But within the CNS, EAAC1 is believed to be neuron-specific, whereas GLAST and GLT1 are glial-specific transporters as described in Chapter 1. The highest densities of the neuronal transporter EAAC1 are in the cortex and the hippocampus areas









that have high levels of glutamatergic synaptic transmission (Rothstein et al., 1994). Interestingly, these areas are also extremely sensitive to excitatoxic damage caused by stroke, ischemia, and head trauma (Palmer et al., 1993; Nilsson et al., 1996; Greene and Greenamyre, 1996). Therefore, it was postulated that the EAAC1 transporter might play a role in the protection of neurons in these areas from excess glutamate release during normal synaptic activity and pathological conditions. However, little is known about the individual contribution of these SDHA glutamate transporters in the regulation of synaptic excitatory neurotransmitter concentration, because of the lack of specific inhibitors for neurophysiology analysis. In recent years, attempts have been made to decipher the role of each SDHA glutamate transporters by generating mutant mice with individual transporters knocked out (Rothstein et al., 1996; Tanaka et al., 1997; Peghini et al., 1997). These knock-out studies suggested that the glial transporters, but not the neuronal one, play a bigger role in the prevention of excitatoxicity and resultant paralysis and seizures.

Considerable effort has been made in cloning and characterization of amino acid transporters in the past decade, but a major aspect, that is, the biosynthesis, maturation, targeting, and degradation of these transporter proteins, is relatively unexplored. As we have repeatedly learned from the study of many other cellular macromolecules, how they are synthesized and degraded often impacts regulation of their cellular function. Therefore, understanding the life cycle of these amino acid transporter proteins would provide us with the basis for future functional studies in normal and diseased states.









In recent years, a lot of progress has been made in generating the proper tools for the study of several amino acid transporters, including cDNA probes as well as specific antibodies. Our own laboratory successfully generated a polyclonal antibody against the EAAC1 transporter protein, which was raised against the C-terminal 120 amino acid peptide of rat EAAC1 fused with the maltose binding protein (EAAC1 -MBP fusion protein). As I have shown in Chapter 2, this anti-EAAC1 antibody recognizes three distinct protein bands on the Western blots of the C6 glioma cells as well as BNL CL.2 cells transiently transfected with EAAC1 -containing plasmid. Pre-incubation of antibody with EAAC1-MBP fusion protein completely inhibited the detection of all these bands, while pre-incubation with MBP alone failed to compete. All of these observations indicate that this is a reliable antibody, which can be used to study the synthesis and degradation of the EAAC1 transporter protein.

As I stated, the biosynthesis and intracellular trafficking of any individual mammalian amino acid transporter has not been documented. However, given that EAAC1 transporter is predicted to be a N-glycosylated integral membrane protein for which the primary functional site is the plasma membrane (Kanai and Hediger, 1992; Velaz-Faircloth et al., 1995), I hypothesize that the biosynthesis pathway follows the general biosynthetic pathway for plasma membrane glycoproteins. The biosynthesis pathway for integral membrane glycoproteins was reviewed in Chapter 1 and briefly summarized here. Biosynthesis of an integral membrane N-glycoprotein starts with the co-translational translocation of the nascent polypeptide into the rough ER, where it is co-translationally modified









by N-glycosylation. Then oligosaccharide trimming and addition proceed further in the ER and Golgi compartment. Endoglycosidases are a class of enzymes that catalyze the cleavage of oligosaccharide chains at specific sugar residues and can be used for the characterization of glycoproteins. The most commonly used endoglycosidases are protein N-glycosidase F (PNGase F) and endoglycosidase H (Endo H). Whereas PNGase F cleaves between the first GIcNAC residue and the Asn on a N-glycosylated protein (Maley, 1989), Endo H cleaves between the two GIcNAC residues adjacent to the Asn and is specific only to N-glycoproteins that are modified with high mannose and some hybrid type oligosaccharides (Robbins, 1984). Therefore, although the susceptibility to the cleavage by both enzymes suggests that the protein is modified by N-linked glycosylation, Endo H sensitivity can be used to further distinguish the type of the attached oligosaccharide chains. Endo H-resistance plus PNGase F-sensitivity, coupled with metabolic labeling as described below, is routinely used as a detection marker for N-glycosylated proteins that are modified by complex-type oligosaccharide chain(s), that is, they have already passed beyond the medial Golgi.

Metabolic labeling of cells with radiolabeled amino acids, such as 3S-Met or 35S-Cys, is commonly used for the study of the biosynthesis, intracellular trafficking, and degradation of the proteins (Alberts et al., 1994). The radiolabeled amino acid(s) are incorporated into the newly synthesized proteins during the short labeling period, often called the "pulse" period. In order to study the maturation, processing, and degradation of this protein, the labeled protein is









then "chased" by incubating the cells with an excess amount of non-radiolabeled amino acid for a certain period of time. De novo biosynthesis of the interested protein is then studied after immunoprecipitation with specific antibody. The susceptibilities of the radiolabeled protein to the digestion by either PNGase F and/ or Endo H is used to determine whether the newly synthesized protein is Nglycosylated as well as whether it has proceeded beyond the medial Golgi complex.

Membrane proteins that have already arrived at the PM may share the same compositional and structural characteristics as those that have finished their post-translational modification but have not been, or will not be, transferred to the PM. Therefore, additional method(s) have to be designed to determine the PM arrival of the protein. The most commonly used techniques to detect the PM arrival of proteins include plasma membrane isolation and cell surface protein modification by membrane impermeable reagents. PM isolation could be done using differential or gradient centrifugation following pulse-chase labeling. The appearance of the radiolabeled protein in the collected PM fraction could be used to determine the time it took for the newly synthesized protein to be targeted to the PM. However, because it is literally impossible to get absolutely pure PM fractions using currently available membrane fractionation methods, modifying the PM proteins of intact cells with membrane impermeable chemicals has proven to be more effective and reliable approach. In this method, after pulsechase labeling, cell surface proteins are chemically modified by a membraneimpermeable reagent. The appearance of the chemically modified radiolabeled









protein serves as a hallmark for the arrival of the newly synthesized protein at the PM. Many plasma membrane proteins, including the vitronectin receptor (Nesbitt and Horton, 1992), Thy-1 and lymphocyte surface marker proteins (Meier et al., 1992; Altin and Pagler, 1995), insulin receptor (Levy-Toledano et al., 1993), and tumor necrosis factor receptors (Hsu and Chao, 1993) have been successfully studied using this tool. Several membrane impermeable modifiers are now commercially available, with different functional groups to confer different chemical reactivities. Biotin-derivatives, by far, are the most commonly used modifiers. The tight binding between biotin and avidin can then be utilized to detect any proteins that have been modified (Gitlin et al., 1987).

C6 cells are rat glial tumor cells, cloned 30 years ago (Benda et al., 1968), which have high Na'-dependent glutamate/aspartate transport activity. The experiments described in this chapter of my thesis were designed first to test C6 glioma cells as a model system for EAAC1 glutamate/aspartate transporter studies, and then to determine the subcellular localization and other characteristics of the EAAC1 protein in these cells. Finally, the biosynthesis, intracellular targeting, and maturation process of the newly synthesized EAAC1 protein also will be investigated.



Results


Characterization of XAG Activity in C6 Glioma Cells

To determine the effect of substrate deprivation, the Na -dependent high affinity glutamate/aspartate transport activity (System XAG') in C6 cells, the cells









were placed at a density of 1 x 105 cells/ well into 24-well trays and cultured for 20 hr under normal conditions. Then the medium was replaced with either NaKRB containing 10% dialyzed fetal bovine serum (FBS) (for amino acidstarved cells, -AA) or complete MEM containing 10% dialyzed FBS (control) for the rest of the groups. The cells were incubated for another 6 hr at 370C, before being subjected to transport assays for L-[3H]-aspartate in the absence (CholKRP) or presence (NaKRP) of sodium ion. The Na*-dependent transport activity was calculated by subtracting the transport measured in Chol-KRP which, for most of the experiments, was less than 3% of the total uptake activity (NaKRP). As shown in Figure 3-1, the C6 cells incubated in MEM had high endogenous Na*-dependent aspartate/glutamate transport activity, which was almost completely inhibited by the presence of excess amount of D-aspartate as a specific inhibitor for System XAG activity. These data suggested that the Na dependent high-affinity glutamate/aspartate transporter family is the main carrier for the aspartate/glutamate transport activity present in C6 cells.

It was previously reported that amino acid deprivation induced an increase in the Na -dependent L-aspartate transport in renal epithelial (NBL-1) cells without a correspondent increase in either the mRNA (Plakidou-Dymock and McGivan, 1993) or the protein content (Nicholson and McGivan, 1996) of EAACI transporter. To study whether a similar effect would be seen in the C6 glioma cells, cells were treated with amino acid-deprived medium for 6 hr and the transport activity was compared to cells maintained in MEM. As shown in Figure 3-1, amino acid starvation for 6 hr did not cause significant change in the Na'-








dependent L-aspartate uptake in C6 cells. A lack of substrate regulation for EAAC1 in human fibroblasts has also been observed in our laboratory (BarbosaTessmann and Kilberg, unpublished results).

Table 3-1 lists the Na*-dependent L-aspartate transport activity measured by our laboratory or reported by others in a number of different cell lines (Nicholson and McGivan, 1996; Kilberg et al., unpublished data). As shown, C6 glioma cell possessed the highest Na*-dependent transport activity for Laspartate among the cell lines tested. Therefore, this cell line was used as our model system to study Na -dependent high-affinity glutamate/aspartate transporter(s).


Determining Which Glutamate/Aspartate Transporter(S) Are Expressed in C6 Glioma Cells

To determine which members of the Na'-dependent high affinity

glutamate/aspartate transporter family are responsible for the high System XAG activity detected in cultured C6 glioma cells, RT-PCR was performed using the primers specific to each of four glutamate/aspartate likely transporters; EAAT5 has not been found expressed outside of the eye. Four x 106 C6 glioma cells were placed onto each 150 mm dish and cultured for three days under normal culture conditions. The cells were then harvested, total RNA and then Poly(A ) mRNA was purified, and RT-PCR reactions for GLT1, GLAST, EAAC1, and EAAT4 glutamatelaspartate transporters were conducted as described in the Methods Chapter. As shown in the Figure 3-2, only the EAAC1mRNA, but not that for GLAST, GLT1, and EAAT4, was detected from these cells. These results









suggested that EAAC1 transporter was the major, if not the only, SDHA glutamate/aspartate transporter endogenously expressed in the C6 glioma cells, and was responsible for the high XAG" activity observed in this cell line. These results are consistent with previous reports (Palo et al., 1996; Dowd et al., 1996).


Effect of Cell Density/Growth on C6 Glioma Cell Differentiation, XG Activity, and Expression of Glutamate/Aspartate Transporter Proteins

C6 glioma cells are believed to be representative of the less-differentiated stem cells that develop into either astrocytes or oligodendrocytes (Parker et al., 1980). Expression of these cell type properties can be modified by a variety of cultured conditions, including high cell density, which leads to glial differentiation (Varon, 1978; Maltese and Volpe, 1979; Parker et al., 1980). Because the glial fibrillary acidic protein (GFAP) is considered a specific marker protein for differentiated astrocytes (Varon, 1978), the amount of GFAP expressed in C6 cells was measured by immunoblotting under different cell density conditions, as described below. C6 cells (2.4 x 105/well) were placed onto each well of 6-well trays and cultured for up to 12 days under normal conditions with the medium changed every 3 days, except for the cells that would be collected on that same day. As shown in Figure 3-3A, the expression of the GFAP protein is up regulated after C6 cells reach confluency. This result is consistent with the previous report suggesting that high cell density could lead to the glial-like differentiation of C6 cells (Varon, 1978; Maltese and Volpe, 1979), but this experiment alone could not tell whether it was the increased cell density, the slower cell growth rate, or both that caused this glial differentiation.









Given that, under normal culture conditions, C6 glioma cells mainly, if not exclusively, express a neuronal glutamate/aspartate transporter (EAAC1) (Palos et al., 1996; Dowd et al., 1996), it was interesting to investigate whether the type and amount of transporter expressed, would be changed while the cells were undergoing differentiation. Equivalent numbers of C6 cells (0.5 x 10s cells/well) was placed into each well of the 24-well trays and cultured up to 12 days. XAG activity and the amount of EAAC1 and GLT1 proteins in these cells were measured. Protein content for GLAST1 and EAAT4 were below detection. As shown in Figure 3-3B, the XAG activity of C6 cells was high at day 2 and 4, and was reduced at day 6, and then returned back to a certain extent from day 8 to day 10 followed by a drop on day 12. It is not clear why a two-phase response was seen for transport activity under the effect of cell density and growth. However, the idea that EAAC1 might be the major transporter protein in C6 cells was further supported by showing that the amount of EAAC1 protein expressed in these cells correlated very well with the XAG" uptake activity throughout the whole two-phase response. Also shown in Figure 3-3B, as the C6 cells became more glial-differentiated, the expression of the glial-specific GLT1 glutamate/aspartate transporter-like immunoreactivity increased. However, as clearly shown in this figure, the amount of the GLT1 transporter expressed did not correlate with the XAG- activity of the cells. Therefore, GLT1 does not appear to represent a major contributor to the amount of XAG- activity detected in C6 cells.









Therefore, I have characterized the aspartate/glutamate transport

properties of C6 glioma cells, and have determined they have high endogenous Na -dependent high-affinity glutamate/aspartate transport activity mediated by EAAC1. These observations make this cell line an ideal model system for studying the biosynthesis, degradation, and regulation of the EAAC1 transporter protein. These results also demonstrated, for the first time, that the amount of the EAAC1 as well as the total XAG activity of C6 cells are regulated by cell density and growth. The concentration of a protein at a given time is determined by the equilibrium between its biosynthesis rate and its degradation rate. A decrease in EAAC1 transporter protein concentration could be caused either by a reduction of protein synthesis or by an increase of protein degradation. As shown below, my studies of EAAC1 protein synthesis suggest that the transporter expression was reduced by a higher cell density (Figure 3-11), indicating that a reduction of de novo EAAC1 biosynthesis was partially, at least, responsible for the decrease of the steady-state amount of EAAC1, and XAG activity in C6 cells cultured to a higher density.



Subcellular Localization of EAACI in C6 Glioma Cells

As shown above, the neuronal EAAC1 transporter is the major SDHA glutamate transporter endogenously expressed in the C6 glial tumor cells. To determine whether EAAC1 transporter functions normally in C6 cells, it is important to establish the cellular localization of the transporter. Given that amino acid transporters are integral membrane proteins whose main functional site is









the plasma membrane (PM), one would expect that the majority of the transporter protein should be localized at the PM with intracellular pools confined to the biosynthesis and degradation pathways. If, in C6 cells, the subcetlular distribution of EAAC1 protein followed the above predication, the data would suggest that this neuronal transporter is not only expressed in the C6 glioma cells, but also targeted to the right place for its function. Three independent approaches were employed to determine the cell surface and intracellular localization of EAAC1 transporter protein in C6 glioma cells: (1) immunofluorescent-cytochemistry; (2) sucrose gradient fractionation of total cellular membranes followed by immunoblotting; (3) cell surface protein biotinylation followed by avidin-precipitation and immunoblotting.

C6 glioma cells were stained with specific anti-EAAC1 antibody (1:1000 dilution) and FITC-conjugated secondary antibody (1:200) as described in the Methods chapter. As controls for subcellular organelle localization, a specific antibody against the manose-6-phosphate receptor was used as a Golgi marker, and 414 antibody was used as a nuclear envelope marker. 414 antibody was generated against an epitope shared by several proteins of the nuclear pore complex (Davis and Blobel, 1986). As shown in the Figure 3-4, EAAC1 staining analyzed by epifluorescence appears mainly to be PM-like (rather diffused staining throughout the cell with enrichment at the edges) and to a lesser extent, Golgi-like (paranuclear staining) with some intracellular vesicle-like staining (punctuated). The labeling by EAAC1 antibody could be completely inhibited by a









pre-incubation of antibody with 25 pg/ml of EAAC1-MBP fusion protein for 12 hr at 40C (data not shown).

Sucrose gradient fractionation of subcellular organelles was also utilized to determine the subcellular localization of EAAC1 in C6 glioma cells. C6 glioma cells were surface-biotinylated with membrane-impermeable sulfo-NHS-LCbiotin, and then a total cellular membrane fraction was collected and separated using sucrose gradient centrifugation as described in the Methods Chapter. The sucrose concentration and protein content was measured for each of the fourteen fractions collected. Equal amount of protein from each fraction was then separated on SDS-PAGE, and subjected to immunostaining with specific antibodies against EAAC1 and other marker proteins. Streptavidin-HRP was used to test for the distribution of biotinylated cell surface proteins, as described in the Methods Chapter. As shown in Figure 3-5, EAAC1 protein was present in fractions 6 to 9, and peaked in fraction 8 with a sucrose concentration of 36% or a density of 1.16 g/cm3, a value that is consistent with the density of the PM (Griffith, 1986). EAAC1 protein co-migrated with Na*/K' ATPase and with bulk biotinylated cell surface proteins in the sucrose gradient, both considered PM markers, but most of the EAAC1 did not migrate with GRP78, an ER marker protein. The broad distribution profile of GRP78 detected is consistent with the result reported previously, showing this protein was relatively "sticky" and present in most of the fractions (Kitzman, 1995).

To determine the PM localization of EAAC1 transporter in C6 cells, the accessibility of EAAC1 protein from the extracellular space was detected by a









cell surface biotinylation technique using membrane-impermeable sulfo-NHS-LCbiotin reagent. After biotinylation, total C6 cellular membrane proteins were collected, and total biotinylated proteins were isolated using avidin-precipitation and then separated by gel electrophoresis. The biotinylated EAAC1 protein was then detected using immunoblotting with specific anti-EAAC1 antibody. As shown in Figure 3-6, EAAC1 could be biotinylated, demonstrating that a significant portion of the EAAC1 protein was accessible from outside of the cells, that is, it resided at the PM. These results are consistent with those presented in Chapter 2 (Figure 2-6), demonstrating more than 60% of EAAC1 protein was specifically biotinylated by cell-surface biotinylation reaction, compared with only 2% or less of asparagine synthetase (AS) and GRP78. All of these data document that the majority of EAAC1 transporter protein was localized at plasma membrane with a small portion present within either the Golgi compartment, or some other intracellular vesicles.



EAAC1 Protein Oliqomerization

I have observed that specific antibodies against GLT1, GLAST, and

EAAC1 detect high molecular weight protein species, which probably represent the oligomeric forms of the glutamate transporter proteins (Haugeto et al., 1996). Additional experiments showed that reducing reagents (5% 3-ME, or 100 mM DTT), a denaturing reagent (6 M urea), or heating (1000C x 10 min) could not dissociate the oligomerization once it had been formed. Furthermore, the harsher the conditions used to treat the samples or the more handling of the samples, a









greater proportion of the higher MW species was detected (see below). This observation is consistent with the results reported by Haugeto and coworkers showing that GLT1, GLAST, and EAAC1 transporter proteins form homomultimers (Haugeto et al., 1996). Using radiation inactivation analysis, they also suggested that the glutamate transporters might operate as homomultimeric complexes in vivo.

The next set of experiments were designed to determine whether EAAC1 transporter protein from C6 cells forms oligomers in vitro, as well as to demonstrate further the effect of high temperature and solubilization conditions on the formation of the EAAC1 oligomers. Total C6 cellular membrane proteins were either mixed directly with SDB buffer, or solubilized with PES buffer for 1 hr at 40C and then heated at different temperatures for specific lengths of time before mixing with SDB. All samples were then resolved on SDS-PAGE and subjected to immunoblotting analysis with anti-EAAC1 antibody. As shown in Figure 3-7, when untreated, the EAAC1 protein of C6 glioma cells was mainly present as the monomer form. However, after solubilization with PES buffer for 1 hr at 40C, oligomeric EAAC1 was detected. Four different solubilization conditions were tested for their solubilization efficiencies as well as for their effects on EAAC1 olgimerization. These conditions included 0.5% Triton X-100, 1% Triton X-100, 1% Triton X-100 plus 4 M urea, or 2% C12E9 plus 0.1% SDS, all in PBS containing 3 mM EDTA. My results suggest that all four solubilization conditions extract EAAC1 protein from the cellular membrane with comparably high efficiency (> 95%). However, the harsher the solubilization conditions, less









EAAC1 monomer and a greater amount of olgomer was detected by immunoblotting (data not shown). Also shown in Figure 3-7, as the heating temperature increased, the relative amount of EAAC1 monomer form of decreased. Although the corresponding increase of oligomers was not obvious in this latter case, the amount of newly formed oligomer may only represent a small portion of the total, or the maximum capacity of the nitrocellulose membrane transfer may have been reached for the higher MW bands. When membrane proteins were immediately mixed with gel sample dilution buffer and run on SDSPAGE without extra manipulation, most of the EAAC1 protein was detected as a single monomer band with a molecular weight of 73 � 1 kDa (Figure 3-5). Therefore, in apparent conflict with the hypothesis of Haugeto et al. (1996), my data suggest that, in C6 cells, the formation of EAAC1 oligomers may be an artifact only caused by the manipulation of the protein sample, and may not represent functional structures inside the cell.

Over-expression of EAAC1 protein in the BNL CL.2 mouse hepatocyte cell line, which has little or no endogenous glutamate transport activity (0 � 0.4 pmol*mg"1 protein*min' L-aspartate uptake), was also utilized to test for the various forms of EAAC1 protein. Transfection of BNL CL.2 hepatocytes with EAAC1-containing expression plasmids significantly increased the Na dependent aspartate transport activity, and this increase was accompanied by an induction of EAAC1 protein expression detected as both monomer (73 kDa) and oligomers (145 kDa and higher MW species) (Figure 3-8). Thus, formation of homo-oligomers occurs not only to endogenously expressed EAAC1 protein in









C6 cells, but also to the overexpressed EAAC1 protein introduced by transfection of the mouse hepatocytes. Collectively, these results suggest that each of these three bands typically detected by immunoblotting are authentic EAAC1. However, the present data suggest that the oligomerization of EAAC1 protein may not exist in vivo; the high MW forms may only be an artifact caused by sample handling.


N-qlycosylation of EAAC1 Protein


The deduced amino acid sequence for the rat EAAC1 transporter contains four putative N-glycosylation consensus sequences, and three of these predicted N-glycosylation sites are localized within the second extracellular loop (VelazFaircloth et al., 1996). It has been reported that EAAC1 transporter is Nglycosylated, and its glycosylation pattern may vary in different host cell lines (Dowd et al., 1996). To test for the N-glycosylation of EAAC1 protein in C6 cells, protein N-glycosydiase F (PNGase F) digestion was used. As stated in the Introduction section, PNGase F cleaves between the first GIcNAC residue and the asparagine residue on N-glycoproteins (Maley, 1989). As shown in Figure 39, without PNGase F digestion both EAAC1 monomer (73 kDa) and oligomer bands were detected after immunoprecipitation of solubilized total membrane proteins with anti-EAAC1 antibody. The EAAC1 bands were broad, which may be due to the micro-heterogeneity of the glycosylation or other post-translational modifications including phosphorylation. However, a relatively sharp 57 kDa band (EAAC1 core) and a 114 kDa band (core dimer) were detected after incubation of









the immunoprecipitated EAAC1 protein with PNGase F for 15 to 240 min. These results are consistent with the data previously published by Dowd et al. (1996) showing that EAAC1 protein is N-glycosylated in C6 glioma cells, but has a core MW of 57 kDa. However, these data, for the first time, showed that deglycosylation of the EAAC1 dimer does not prevent or reverse its oligomerization.

N-glycosylation has been shown to be important for the functions of many plasma proteins, including amino acid transporters. It has been shown that the N-glycosylation of a glycine amino acid transporter (GLYT1) was important for its transport activity (Olivares et al., 1995). In contrast, abolishment of the Nglycosylation of the GLAST glutamate transporter protein using site-directed mutagenesis had no effect on its transport activity (Conradt et al., 1995). To determine whether N-glycosylation was important for the function of EAAC1l transporter in C6 glioma cells, tunicamycin was used as an inhibitor for the Nglycosylation reaction. Tunicamycin is a uracil-containing nucleoside antibiotic that specifically inhibits the first reaction of the synthesis of dolichol-linked oligosaccharides (Alberts et al., 1994). Incubation with 1 ipg/ml tunicamycin for 8 to 36 hr decreased the aspartate transport activity of EAAC1, and this decrease of transport activity correlated with the reduction of the amount of mature EAAC1 protein in treated C6 glioma cells (Figure 3-10). Furthermore, both effects appeared to be time-dependent up to 36 hr. It is interesting to notice that, as the content of the N-glycosylated mature form EAAC1 decreased, there was no detectable accumulation of core EAAC1. This result may be due to the possibility that core EAAC1 is not as stable as the glycosylated form and is degraded much









faster, as seen with other proteins (Melikian, et al., 1996). These data suggested that the normal N-glycosylation process was required for EAAC1 function in C6 glioma cells. Although it is tempting to postulate that the N-glycosylation of EAAC1 protein itself is important, we have to bear in mind that tunicamycin is a non-selective inhibitor for any proteins that may be N-glycosylated. Therefore, it is also possible that this observed tunicamycin effect on EAAC1 transporter and its activity might be an indirect response to the loss of N-glycosylation of other related protein(s). Further studies using site-directed mutagenesis of the EAACI cDNA must be employed in combination with transfection experiments to differentiate between these two possibilities.

To explore the possibility that EAAC1 protein may associate with other

protein(s) inside C6 glioma cells, protein cross-linking was used with a cleavable cross-linker, dithiobis (succinimidylpropionate) (DSP), followed by immunoprecipitation. C6 cells were first metabolically labeled with 3S-Met-CysH to label EAAC1 protein and other possibly associated proteins, and then treated with DSP to cross-link any proteins that were bound to or in close vicinity with the EAAC1 protein. After immunoprecipitation with anti-EAAC1 antibody, the precipitates were eluted in SDB containing 5% 3ME as reducing reagent to separate potential cross-linked proteins. The samples were resolved on reducing SDS-PAGE, the gel dried under heat and vacuum, and analyzed with fluorography. Any proteins that were cross-linked with EAAC1 protein by DSP would be co-precipitated with EAAC1, and would be revealed as extra bands on the film. As no extra bands were seen in the DSP-treated samples compared









with controls (data not shown), it appeared that there were no major proteins associated with EAAC1 inside C6 glioma cells. This result is also consistent with my finding that, when washed with varied concentrations of salts ranging from the physiologic concentration of 150 mM to almost ten times more (1.2 M), no specific extra protein bands were seen in those samples washed with a low salt concentration after immunoprecipitation (Figure 2-2).


Establish the De Novo Biosynthesis Rate for EAACI Transporter in C6 Glioma Cells

To determine the biosynthesis rate of EAAC1 protein in C6 glioma cells, the cells were metabolically labeled with 200 pCi/ml of 35S-Met-Cys for 15 to 120 min. After the labeling period, the cells were chased in medium containing an excess amount of non-radioactive Met and Cys for 3 hr, then the cells were harvested, total membrane proteins collected, newly synthesized EAAC1 protein was immunoprecipitated, and analyzed by gel electrophoresis followed by fluorography. As shown in Figure 3-11A, as short as 15 min pulse-labeling time is enough for the detection of newly synthesized EAAC1. All three forms (monomer, dimer, oligomer) of EAAC1 appear to be proportional to each other regardless of how long the pulse time. These results suggest that the oligomerization is not part of a maturation process for EAAC1 protein, which is consistent with my hypothesis that the formation of multimers may only be an artifact caused by sample handling. Interestingly, the biosynthesis rate of EAAC1 was decreased when a higher number of cells was plated, that is, EAAC1 biosynthesis is down-regulated by increased cell density (Figure 3-11B). This









observation is consistent with my earlier studies showing the EAACI content is dramatically reduced as cells became more confluent (Figure 3-3). Collectively, these data suggest that higher cell density, or perhaps lower cell growth, decreases the amount of EAAC1 protein, at least partially, through downregulation of its de novo biosynthesis rate. Therefore, in all experiments a fixed number of cells were plated and then cultured for an exact period of time.



Maturation and Targetingq of EAAC1 Transporter in C6 Glioma Cells

To study the maturation process of EAAC1, C6 glioma cells were pulselabeled with 500 pCi/ml of 35S-Met-Cys for 15 min and then chased for 0 to 120 min. When chased for less than 30 min, only the low molecular weight (MW) (~57

- 60 kDa) immature forms of EAAC1 were detected (Figure 3-12). After longer chases, the lower MW form EAAC1 matured into the 73 kDa mature monomer form. In Figure 3-13, it appears that the transition of EAAC1 protein from the lower MW forms to the mature monomer form started after 45 min of chase time and finished after 190 min. Also illustrated in Figure 3-12 is the fact that there were multimer forms detected not only for the mature EAAC1 monomer, but also for the lower MW EAAC1. This observation further supported my hypothesis that the multimer forms may not exist in vivo, because if they were functionally important, we would expect that they should happen more likely, even if not only, to the mature form of the EAAC1 protein.

The next question to address was how long it would take for the newly synthesized EAAC1 protein to be targeted to its destination - the PM. Pulse-









chase labeling of the C6 glioma cells with 35S-Met-Cys followed by cell surface biotinylation was utilized to detect the time at which the radiolabeled (newly synthesized) EAACI protein would also be biotinylated (i.e., arrival at the PM). Cells were metabolically-labeled with 200 pCi/ml of 3S-Met-Cys for 15 min and chased in medium containing an excess amount of non-radiolabeled Met and Cys for 0 to 360 min at 370C. The pulse time was as short as possible (15 min) using an elevated amount of radioactivity to provide a narrow observation "window" for the PM arrival of the newly synthesized EAAC1 protein. At the end of each chase time point, the cells were surface-biotinylated with cell membraneimpermeable sulfo-NHS-LC-biotin to label all the proteins that were accessible from the extracellular space. After the PES-soluble proteins were collected from a total cellular membrane fraction, a double precipitation procedure was done by first precipitating all of the EAAC1 protein using anti-EAAC1 antibody, eluting the transporter from the sepharose beads, and then precipitating only the biotinylated EAAC1 protein using immobilized monomeric avidin-coated beads. Finally, using fluorography analysis, the newly synthesized EAAC1 protein molecules that had arrived at the PM (both 35S- and biotin-labeled) as well as those that had not yet arrived at the PM ("S-labeled and non-biotinylated) were detected. Newly synthesized EAAC1 transporter first arrived at the PM from 45 to 90 min (Figure 3-14A). The densitometry analysis of the data is plotted in Figure 3-14B and show that the targeting of the newly synthesized EAAC1 protein to the PM coincided with its maturation, that is, the transition of EAAC1 from the lower MW forms to the mature monomer.









To study the intracellular targeting of the newly synthesized EAAC1

protein, endoglycosidase digestions were employed. Endoglycosidase H cleaves between the two GIcNAC residues linked to the Asn residue in N-linked glycoproteins, only on high-mannose and some hybrid types of N-linked oligosaccharides. Therefore, resistance to Endo H digestion can be used as a hallmark for the N-glycoproteins that have already proceeded beyond the medial Golgi compartment. C6 glioma cells were pulse-labeled with 200 PCi/ml of 3SMet-CysH for 15 min and then chased in unlabeled medium for 30 to 240 min at 370C. EAAC1 protein was then immunoprecipitated from a solubilized total cellular membrane fraction and subjected to Endo H digestion for 1 hr at 370C. As shown in Figure 3-15, all mature EAAC1 monomers were resistant to Endo H digestion. Recall that the mature EAAC1 monomer form is sensitive to PNGase F digestion (Figure 3-9). This observation is consistent with the results of Figure 314, suggesting that the monomer form EAAC1 has already proceeded beyond the medial Golgi (Endo H-resistant) and has arrived at the PM (accessible to surface biotinylation). In contrast, the immature lower MW forms of EAAC1 protein were sensitive to the Endo H digestion, as shown in the Figure 3-15. These precursor EAAC1 polypeptides have been co-translationally modified by N-glycosylation, but they are still localized in the ER or cis Golgi compartment. Therefore, the oligosaccharide chains attached to them were those of the highmannose type, and thus, still sensitive to Endo H digestion.

As to why only two major bands, instead of a broad smear of multiple bands, were observed for the lower MW forms of EAAC1, we have no definite









answer. Obviously, the antibody would not detect those nascent EAAC1 polypeptides that did not yet contain their C-terminal portion, that is, the last 120 amino acid residues. Judging from the sizes of those two bands after Endo H digestion, I could postulate that the upper one might be the full length EAAC1 protein with a predicted core size around 57 kDa; whereas the lower one might only be a translationally unfinished product, which had already possessed the Cterminal epitome specific for our antibody. Alternatively, one could postulate that the translation of EAAC1 protein may not be a continuous process, but one with distinct pause sites, because if it were a continuous process, a smear of many bands would be expected when the cells were pulse-labeled for a short time without chase. However, proof of this hypothesis will definitely need extensive investigation. Nevertheless, it is clear that both of these two immature forms of newly synthesized EAAC1 protein had been co-translationally modified by Nglycosylation.

In summary, all of the data taken together suggest that the EAAC1 transporter is co-translationally modified by N-glycosylation, and that the transition of the EAAC1 protein from its immature forms to mature form coincides with the alteration of the attached oligosaccharide chains from the high-mannose type to the complex type. Importantly. these results also suggest that once the mature form is fully processed and achieved, it takes so little time as to be undetectable in the experiments, for the EAAC1 monomer to pass through the rest of the Golgi compartment and arrive at the PM.









Discussion

Prior to the present investigation, no mammalian amino acid transporter had been studied with regard to biosynthesis, maturation, and intracellular targeting processes, primarily due to the lack of proper tools. Recent advances in molecular cloning and protein expression have enabled us to obtain specific antibodies to some of the amino acid transporters, and these antibodies serve as valuable tools in addressing these questions. The polyclonal antibody generated by our laboratory against rat EAAC1 specifically recognizes three bands when used to probe C6 glioma cell proteins, all of which could be completely competed away by pre-incubating antibody with the EAAC1 -MBP fusion protein (Figure 21). Consistent with reports by other laboratories who have observed similar results with independent antibodies (Haugeto et al., 1996; Davis et al., 1998), these multiple bands probably represent the mature EAAC1 monomer as well as its multimers, that is, dimer and trimer. However, as described below, the physiologic significance of these multimers is still a subject of debate. Our antiEAAC1 antibody has been used successfully to immunoprecipitate the EAAC1 protein from rat brain plasma membrane protein samples (data not shown), C6 glioma cells (Figure 2-4), EAAC1 -transfected BNL CL.2 cells (Figure 3-8), and Rcho rat placental cells (data not shown). EAAC1 transporter was chosen to investigate what I have termed "the life cycle" of a mammalian amino acid transporter, not only because of the availability of this specific antibody, but also because this transporter is widely distributed among tissues and represents a metabolically-important activity. For example, EAACI knock out mice exhibit









dicarboxylic amino aciduria because of the critical requirement for EAACI in renal resorption (Peghini et al., 1997).

To choose a proper cell line as a model system for studying the

biosynthesis, maturation, targeting, and degradation of the EAAC1 glutamate transporter, several cell lines were tested for their endogenous System XAG activity and EAAC1 protein content. As shown in Table 3-1, C6 glioma cells had the highest XAG activity among the cell lines tested. Therefore, they were the logical choice, regardless of the conflicting reports about which of the five glutamate/aspartate family members are expressed in C6 glioma cells (Casado et al., 1993; Dowd et al., 1996; Palos et al., 1996). My reverse transcriptase PCR results suggested that only EAAC1 mRNA was present when C6 glioma cells were cultured just to the point of confluency. Whereas the EAAC1 protein content decreased as C6 cells approached and passed confluency, GLT1 protein expression was significantly induced after confluency (Figure 3-3). As clearly shown in Figure 3-3, expression of the EAAC1 transporter parallels the observed changes in XAG" activity measured in C6 glioma cells. Therefore, the EAAC1 transporter in C6 glioma cells provided me with an ideal model system for the study of transporter biosynthesis, maturation, targeting, and degradation. To determine if the typically "neuronal" EAACI transporter, when expressed in C6 glioma cells, was localized at the PM, immunocytochemistry and sucrose fractionation techniques were used to determine its subcellular localization. Both approaches gave me consistent results documenting that EAAC1 resides primarily at the PM and, to a lesser extent, in the Golgi complex.









Various nutrient transporters, including amino acid transporters, have been reported to oligomerize as detected by gel electrophoresis and immunoblotting. Haugeto and coworkers reported that GLAST1, GLT1, and EAAC1 transporters all formed homomultimers in rat brain and transfected Hela cells (Haugeto et al, 1996). Furthermore, based on results obtained from radiation inactivation analysis, they postulated that the glutamate transporters operate as homomultimeric complexes in vivo. Similarly, using glioma C6 proteins for immunoblotting, our anti-EAAC1 antibody detected three bands with estimated molecular weights of 73 kDa, 145 kDa, and 200+ kDa, which may represent the EAAC1 monomer, dimer, and trimer, respectively. However, in at least three independent experiments, only the EAAC1 monomer was detected when I collected a total membrane fraction from C6 glioma cells and directly resolved the proteins by gel electrophoresis without additional processing. This result was even true for membrane samples that were fractionated by sucrose gradient centrifugation (Figure 3-5). However, when harsher manipulation conditions were used for the membranes or solubilized proteins, I saw a shift of EAAC1 from its monomer to oligomer forms. In addition, when C6 glioma cells were pulse-chase labeled, the oligomer forms for the newly synthesized immature form EAAC1 (the lower MW forms) were also detected. No change in oligomerization was seen for the newly synthesized EAAC1 mature form as the pulse or chase time was prolonged. Although these results suggest that oligomerization happens before translation is finished, this process seems unlikely to me. If the oligomers were present in the cell and functionally required,









one should never be able to isolate cell membrane containing only monomer, yet it can be. Therefore, it is postulated that these oligomer forms are artifacts caused by sample manipulation. This aggregation may occur because these integral membrane proteins are designed to be surrounded by a lipid bilayer, and have little tolerance for the conditions used to treat the samples. Formation of the multimers may be simply driven by the hydrophobic interactions when these proteins are extracted from the biological membrane into an aqueous surrounding, even with detergent present. However, there is a possibility that the transporter may exist as a homodimer, for example, that is converted to monomer after extraction of the cellular membrane proteins with reducing reagent, that is, -mecaptoethanol.

De novo biosynthesis of a membrane N-glycosylated protein starts from the translocation of the nascent polypeptide into the ER, where it is cotranslationally modified by N-glycosylation. As the protein proceeds through the ER and then Golgi compartments, the oligosaccaride chains attached may be further modified from high mannose- to complex-type. Therefore, when the chase is for a short time, only the immature form of the protein will be detected. Furthermore, these immature forms will be sensitive to endoglycosidase H digestion, only if they have not passed beyond the medial Golgi. This is exactly what I saw for the biosynthesis of EAAC1 transporter. When chased for less than 45 min, only the immature lower MW EAAC1 protein species were detected. These immature forms of EAAC 1 were sensitive to Endo H cleavage, but the mature EAAC1 monomer detected after longer periods of chase was completely




Full Text

PAGE 1

BIOSYNTHESIS, DEGRADATION, AND INTRACELLULAR TARGETING OF EAAC1 GLUTAMATE/ASPARTATE TRANSPORTER IN C6 GLIOMA CELLS By WENBO YANG A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1998

PAGE 2

ACKNOWLEDGMENTS I would like to thank the members of my supervisory committee: Dr. Charles Allen, Dr. Brian Cain, Dr. William Dunn, and Dr. Susan Frost. I wish to extend special thanks to my mentor, Dr. Michael Kiiberg. Without him, this study would not be possible. I would also acknowledge the current and previous members of Dr. Kilberg's laboratory for their friendship and valuable support. Most of all, I would like to thank my husband, Ming, my daughter, Helen, as well as my parents and parents-in-law, whose endless love, support, and personal sacrifice have made my career dream come true. ii

PAGE 3

TABLE OF CONTENTS page ACKNOWLEDGMENTS ii LIST OF FIGURES v ABBREVIATIONS vii ABSTRACT xii CHAPTERS 1 INTRODUCTION 1 Overview of Mammalian Amino Acid Transport 1 Anionic Amino Acid Transport 2 Molecular Cloning of the SDHA Glutamate/ Aspartate Transporters 3 Structural and Functional Characteristcs of the Five SDHA Glutamate/ Aspartate Transporters 7 Tissue and the Cell Distribution of the Five SDHA Glutamate/ Aspartate Transporters 9 Physiological Significance of the SDHA Glutamate/Aspartate Transporters 10 Biosynthesis and Degradation of Membrane Glycoproteins 13 / De Novo Biosynthesis and Intracellular Targeting of the Integral Membrane N-Glycoprotein 15 Degradation of the Plasma Membrane Proteins 19 C6 Glioma Cell and Its Utilization in SDHA Glutamate/Aspartate Transporter Studies 27 2 MATERIALS AND METHODS 32 Materials 32 Methods 33 , , tu

PAGE 4

3 ACTIVITY, PROTEIN CHARACTERISTICS, AND BIOSYNTHESIS OF EAACl GLUT AMATE/ASPART ATE TRANSPORTER IN C6 GLIOMA CELLS 59 Introduction 59 Results 65 Discussion 84 4 DETERMINATION OF THE DEGRADATION RATE, AND THE PM RESIDENCE TIME FOR EAACl TRANSPORTER IN C6 GLIOMA CELLS 121 Introduction 121 Results 126 Discussion 134 5 CONCLUSIONS AND FUTURE DIRECTIONS 159 Conclusions 159 Future directions 171 LIST OF REFERENCES 174 BIOGRAPHICAL SKETCH 184 iv

PAGE 5

LIST OF FIGURES Figure ^ 2-1 EAAC1 Peptide Competition For Immunoblotting 49 2-2 Effect of Salt on Immunoprecipitation of EAAC1 Protein 51 2-3 Effectiveness of Depletion of EAAC1 Protein by Immunoprecipitation 53 2-4 EAAC1 Immunoprecipitation Protein Concentration Curve 55 2-5 Determination of the Elution Conditions for Immunoprecipitated EAAC1 . 56 26 Specificity of Cell Surface Protein Biotinylation 58 31 Na*-Dependent Glutamate/Aspartate Transport Activity of C6 Glioma Cells 91 3-2 RT-PCR Detection of Glutamate/Aspartate Transporters in C6 Glioma Cells 93 3-3 Effect of Cell Density/Growth on the Xag" Activity, EAAC1 Content, and Differentiation of C6 Glioma Cells 95 3-4 Immunocytochemistry Staining of C6 Glioma Cells with Specific EAAC1 , M6PR, and 414 Antibodies 97 3-5 Sucrose Gradient Fractionation of Total Cellular Membranes 99 3-6 Cell Surface Biotinylation of EAAC1 in C6 Glioma Cells 101 3-7 Oligomerization of the EAAC 1 Transporter 1 03 3-8 Over-Expression of EAAC1 Transporter Protein in BNL CL.2 Cells 105 3-9 De-Glycosylation of EAAC1 Protein with N-Glycosidase F 107 3-10 Effect of Tunicamycin on the EAAC1 Transporter Protein Content and Transport Activity 109 3-1 1 De Novo Biosynthesis of EAAC1 Transporter in C6 Glioma Cells 111

PAGE 6

3-12 Maturation of EAAC1 Protein from Lower MW Forms to Its Monomer. .. 113 3-13 Transition of Newly Synthesized EAAC1 from Its Lower MW Form to the Mature Monomer Form in C6 Glioma Cells 115 3-14 Tafficking of Newly Synthesized EAAC1 Protein to the Plasma Membrane in C6 Glioma Cells 117 315 Endoglycosidase H-Sensitivity of Newly Synthesized EAAC1 Protein in C6 Glioma Cells 119 41 Determination of the EAAC1 Protein Degradation Using Pulse-Chase Labeling in C6 Glioma Cells 142 4-2 Degradation Rate of the Cell Surface-Biotinylated EAAC1 Protein 144 4-3 Determining the Plasma Membrane Residence Time of Newly Synthesized EAAC1 Protein 146 4-4 Comparison of the EAAC1 Degradation Rate with the Rate of Its Disapperance from Cell Surface in C6 Glioma Cells 148 4-5 Effect of Various Inhibitors on the Degradation of EAAC1 Protein in C6 Glioma Cells 150 4-6 Effect of ALLN on the Degradation of EAAC1 in C6 Glioma Cells 1 52 4-7 Effect of Leupeptin + NH4CI on the Degradation of EAAC1 in C6 Glioma Cells 154 4-8 Effect of Leupeptin Alone on the Degradation of EAAC1 in C6 Glioma Cells 156 4-9 Effect of NH4CI Alone on the Degradation of EAAC1 1 58 vi

PAGE 7

ABBREVIATIONS uCi mlcrocurie yg microgram ul microliter uM micromolar (BME 3-mercaptoethanol [R]o initial radiolabeled protein concentration [R]t the radiolabeled protein remaining after time t ALLN N-acetyl-leu-leu-norleucinal AS asparagine synthetase ATP adenosine 5' -triphosphate ^ , BFA brefeldin A BSA bovine serum albumin C12E9 polyoxyethylene 9 lauryl ether cDNA complementary deoxyribonucleic acid CFTR cystic fibrosis tramsmembrane regulator protein Chol-KRP sodium-free Krebs-Ringers phosphate buffer CNR 2',3'-cyclic nucleotide-3'-phosphohydrolase CNS central nervous system

PAGE 8

cRNA complementary ribonucleic acid DMEM Dulbecco's modified Eagle medium DTT dithiothreitol EAAC1 rodent excitatory amino acid carrier 1 EAAT1-5 human excitatory amino acid transporters 1-5 EDTA . ethylenediamine tetraacetic acid EGF epidermal growth factor EGTA ethylene glycol-bis((i-aminoethyl ether)-tetraacetic acid Endo H protein endoglycosidase H ER endoplasmic reticulum FBS fetal bovine serum FITC fluorescein isothiocynate GABA y-aminobutyric acid GDP guanosine 5'-diphosphate GFAP glial fibrillary acidic protein GLAST rodent glutamate/aspartate transporter GLT1 rodent glutamate transporter 1 Glut facilitated glucose transporter GLYT1 glycine transporter 1 . » GRP78 glucose regulated protein 78 GS glutamine synthetase GTP guanosine 5' -triphosphate HEPES 4-(2-hydroxyethyl)-1 -piperazineethanesulfonic acid viii

PAGE 9

. ..• , . hr hour HRP horseradish peroxidase IgG immunoglubin G kj, the fractional rate constant for protein degradation kDa kilodalton LDL low density lipoprotein M6PR mannose-6-phosphate receptor mA miliampere MBP maltose binding protein MEM Eagles minimal essential medium min minute mM milimolar mRNA messenger ribonucleic acid MW molecular weight NaKRB sodium-containing Krebs-Ringers bicarbonate buffer NaKRP sodium-containing Krebs-Ringers phosphate buffer NFDM non-fat dry milk NGS normal goat serum NOC nocodazole NSF N-ethylmaleimide-sensitive fusion protein PAGE polyacrylamide gel electrophoresis PBS phosphate buffered saline PGR polymerase chain reaction i.\

PAGE 10

PES buffer detergent-containing protein solubilization buffer PI3K phosphatidylinositol 3-kinase PKA protein kinase A PKC protein kinase C '^^ , . PM plasma membrane . ' • PMSF phenylmethyl-sulfony fluoride PNGase F protein N-glycosidase F RT room temperature RT-PCR reverse transcriptase-PCR SD standard deviation SDB protein gel sample dilution buffer SDHA sodium-dependent high affinity SDS sodium-dodecyl-sulfate SEB buffer sucrose-EDTA-containing Tris-buffer SNAP soluble NSF attachment protein SNARE SNAP receptor SRP signal recognition particle Sulfo-NHS-LC-biotin sulfosuccinimidyl-6-(biotinamido)hexanoate T1/2 half-life TBS-T Tris-buffered saline containing Tween TCA trichloroacetic acid Tf transferrin TLCK Na-p-tosyl-L-lysine chloromethyl ketone X

PAGE 11

TPCK N-tosyl-L-phenylalanine chloromethyl ketone Tris tris(hydroxymethyl)aminomethane xi 3

PAGE 12

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy BIOSYNTHESIS, DEGRADATION, AND INTRACELLULAR TARGETING OF EAAC1 GLUTAMATE/ASPARTATE TRANSPORTER IN 06 GLIOMA CELLS By Wenbo Yang December, 1998 Chairman: Dr. Michael S. Kilberg Major Department: Biochemistry and Molecular Biology Mammalian amino acid transport activities have been extensively studied over the past 30 years by describing their kinetics, substrate specificity, tissue distribution, and regulation. However, the "life cycle", that is, the biosynthesis, processing, and degradation of these transporter proteins, has not yet been documented. This study and others of its kind are beginning to make use of the newly available cDNAs and sequence-specific antibodies as tools to investigate transporter cell biology and function. In the present research, rat C6 glioma cells were used as a model system to study the activity, biosynthesis, intracellular targeting, and degradation of the EAAC1 glutamate/aspartate transporter protein. The EAAC1 transporter is the major sodium-dependent high-affinity glutamate/aspartate transporter expressed endogenously in C6 glioma cells. xii

PAGE 13

Both the glutamate/aspartate uptake activity and the EAAC1 protein content ' were correlatively down-regulated by increased ceil density or decreased cell growth. The EAAC1 transporter protein was co-translationally N-glycosylated with high-mannose oligosaccharide chains, which were processed into complextype sugar chains as the protein matured. The final maturation steps for EAAC1 protein coincided with its PM arrival, which started at about 45 min and finished about 3 hours after initial synthesis. The newly synthesized EAAC1 protein was protected from degradation during the maturation/PM-targeting process, as well as during the first 5 hours after PM arrival. Once started (i.e., 8 hours after the initial synthesis), the degradation of the newly synthesized EAAC1 protein followed first order kinetics with a decay rate of 1 1 .5% per hour yielding a halflife of 6 hours. The EAAC1 transporter protein was degraded, at least partially, through the endocytosis-lysosome protein degradation pathway. Once the EAAC1 protein was endocytosed, it was degraded immediately in lysosomes or late endosomes, no significant intracellular/recycling pool was detected. These results illustrate some interesting and unique features regarding the processing and trafficking of the EAAC1 transporter. . xiii

PAGE 14

CHAPTER 1 INTRODUCTION Overview of Mammalian Amino Acid Transport Amino acids not only sen/e as precursors for protein biosynthesis, but also function as carriers for nitrogen and carbon atoms, as metabolic fuels, as neurotransmitters, as biosynthetic precursors, and as sensitive regulators of cell osmolarity. They can be transferred across the plasma membrane of cells through the mediation of various types of protein carriers. This transferring movement is called "amino acid transport" and these carriers are referred to as . "amino acid transporters". The study of amino acid transport into animal cells actually began in 1913 with the pioneering observations of Van Slyke and Meyer who demonstrated tissue accumulation of amino acids against a . concentration gradient (Van Slyke and Meyer, 1913). In the early 1960s, the Christensen laboratory originated the description of individual transport ; . . ' activities based on the chemical properties, size, and conformation of the amino acid side chain (Oxender and Christensen, 1963). The transport activities were further classified into individual transporter carrier "systems" using substrate competition assays. Therefore, each of the transport systems exhibits distinct substrate specificity for certain classes of amino acids, for example, zwitterionic, anionic, or cationic amino acids. 1

PAGE 15

2 Basically, there are two mechanisms by which the amino acids are transported across the plasma membrane of cells. Some transport systems are secondary active transporters, energized by the Na^-electrochemical gradient, and result in net accumulation of amino acid against a concentration gradient. Others are Na*-independent facilitated transporters and permit flux in either direction depending on the chemical principle of mass action. Anionic Amino Acid Transport The anionic amino acid glutamate is the predominant excitatory neurotransmitter in the mammalian central nervous system (CNS) (Fagg and Foster, 1983; Robinson and Coyle, 1987). It is generally assumed that extracellular concentrations of the excitatory amino acids are regulated primarily by the clearance of these neurotransmitters from the synaptic clefts through ' their specific transporters with little or no extracellular metabolism (Iversen, 1975; Kuhar, 1973). An excessive amount of these excitatory transmitters can be toxic to the neurons that have glutamate receptors (Choi, 1992). Transport of the anionic amino acids across the plasma membrane can be mediated by either high-affinity (Km = 2 50 uM) or low-affinity (Km > 1 00 \iM) Na'dependent transporters, or by a Na*-independent facilitated transporter (Schousboe, 1981; Nicholls and Attwell, 1990; Kanai et al., 1993). However, the Na*-dependent high-affinity (SDHA) glutamate/aspartate transporters, that mediate an activity referred to as System Xag", are believed to play an essential role in glutamatergic transmission in the CNS as well as in the

PAGE 16

glutamate/aspartate metabolism in brain and peripheral tissues (Radian et al., 1990; Danboit et al., 1992; Kanai, 1993). This activity transports both the Land D-steroisomers of aspartate, but only L-glutamate (Christensen and Makowske,1983; Brew and Attwell, 1987), and is very sensitive to membrane potential (Gazzola et al., 1981). The mechanism for anionic amino acid uptake through system Xag' is electrogenic and coupled to the co-transport of three Na* ions and one H*, as well as the counter-transport of one K* ion (Zerangue and Kavanaugh, 1996). This coupling mechanism allows glutamate to be concentrated to high levels in both neurons and epithelia. Even before the cloning of the first SDHA glutamate transporter, the existence of several distinct activities was proposed because transporters in synaptosomes prepared from various brain regions can be differentiated pharmacologically using specific inhibitors (Bridges et al., 1991; Fletcher and Johnston, 1991; Rauen et al., 1992). Over the past decade, five cDNAs, all encoding for SDHA glutamate/aspartate transporter activity, have been cloned. They are designated as EAAT1 -5 for human clones, but for the rodent counterparts, EAAT1-3 were originally named as GLAST, GLT1, and EAAC1, respectively (for review, see Malandro and Kilberg, 1996). The latter rodent nomenclature will pnmahly be used in this report. Molecular C loning of the SDHA Glutamate/Asoartate Transporters In 1992, Storck and coworkers identified the first SDHA glutamate transporter cDNA and named the encoded transporter as GLAST (Storck et al..

PAGE 17

4 1992). During the isolation of a galactosyltransferase from rat brain, they copurified a hydrophobic protein of about 66 kDa. Proteolytic fragments of this protein were sequenced, and then oligonucleotide probes were generated. Using these probes, they identified a cDNA clone from rat brain cDNA library, which encoded a protein of 543 amino acids with a predicated molecular weight of 60 kDa. This clone showed considerable sequence similanty to the previously identified glutamate and monocarboxylate transporters of bacteria (Tolner et al., 1992). The SDHA glutamate/aspartate transport (Xag") activity of GLAST was confirmed by expression in Xenopus oocytes after cRNA injection followed by transport assays using radiolabeled amino acid substrates (Klockner etal., 1993). Pines and coworkers cloned the second member of the SDHA glutamate/aspartate transporter family, GLT1 (Pines et al., 1992). The cloning of GLT1 essentially began several years before the isolation of the cDNA with the purification of the transporter protein to near-homogeneity (Danbolt et al., 1990, 1992). The antibody generated against the partially purified transporter protein was used to immuno-screen a rat cDNA expression library and isolate the GLT1 cDNA, which encoded a 573 amino acid protein with a predicted molecular mass of 63 kDa. The function of this transporter was confirmed by measuring the uptake of radiolabeled substrates following its over-expression in mammalian cells.

PAGE 18

5 Also in 1992, Kanai and Hediger identified the third member of this glutamate transporter family by oocyte expression cloning using fractionated mRNAfrom rabbit intestine and named it EAAC1 (Kanai and Hediger, 1992). This cDNA contained an open reading frame encoding 524 amino acids and a protein of predicted molecular weight of 57 kDa. Expression of the EAAC1 cDNA in Xenopus oocytes resulted in a 1000-fold increase in Na^-dependent uptake of L-glutamate over that of water-injected oocytes. A rat EAAC1 homolog was later isolated, which shares 95% similarity with the rabbit clone (Velaz-Faircloth et al., 1996). A polyclonal antibody against a fusion protein consisting of the C-terminal 120 amino acids of the rat EAAC1 and maltose (Predicted Rabbit EAAC1 Protein Structure)

PAGE 19

6 binding protein was generated by our laboratory, which has been shown to specifically recognize the EAAC1 protein from brain and placenta by immunoblotting (Matthews et al., 1998). • ' H ,,-' ' In 1995, the fourth member of this family, EAAT4, was isolated by Fairman and colleagues using degenerate oligonucleotide primers corresponding to conserved sequences within the other members of the SDHA glutamate transporter family and low-stringency RT-PCR from human cerebellum mRNA (Fairman et al., 1995). Using an oocyte expression system, it was shown that the EAAT4 transporter functioned not only as a SDHA glutamate/aspartate transporter, but also as a chloride channel activated by substrate binding, that is, binding by glutamate/aspartate and sodium. Given this unique property of the EAAT4 transporter, it was postulated to be involved in not only the clearance of glutamate from the synaptic cleft, but also in reestablishment of membrane potential by influencing cellular chlonde permeability. Last year, Arriza and coworkers cloned EAAT5, the fifth member of this family, by screening a human retinal cDNA library with a glutamate transporter cDNA isolated from salamander retina (Arriza et al., 1997). Similarly with EAAT4, EAAT5 was also shown to be a chloride channel as well as a SDHA glutamate/aspartate transporter. The associated chloride conductance activity of EAAT5 was postulated to play a role in visual processing.

PAGE 20

7 Structural and Functional Characteristics of the Five SDHA Glutamate/Aspartate Transporters These five members of the SDHA glutamate/aspartate transporter family exhibit 36 to 55% amino acid identities with each other (51-55% identity between GLAST, GLT1, and EAAC1) (Kanai et al., 1997). The hydropathy profile for each of these transporters predicts a hydrophobic protein with 6 8 putative transmembrane domains at the N-terminus and several shorter hydrophobic regions at the C-terminus, which may compose additional transmembrane domains (Malandro and Kilberg, 1996). There are at least 2-3 consensus sites for N-glycosylation located within the second extracellular loop, as well as several consensus sequences for protein kinase A (PKA) and/or protein kinase C (PKC) phosphorylation within the intracellular loops. Three members of this family, GLAST, GLT1, and EAAC1, have been successfully detected by immunoblotting as well as immunocytochemistry using specific antibodies. It has been reported that they all form homo-multimers. even when detected using immunoblotting under reducing conditions (Haugeto et al., 1996). When detected by immunoblotting, the mature monomers of these transporters run as a broad band between 66 and 74 kDa, depending on the cell type, which is likely due to micro-heterogeneity of their N-glycosylation sites, and possibly other post-translational modifications such as phosphorylation. The same transporter protein prepared from different tissues or cells migrates differently on SDS-PAGE, that is, with different apparent molecular weights (Dowd et al., 1996). This result suggests they may be

PAGE 21

1 modified differently by N-glycosylation and/or phosphorylation according to the . ; individual cell or tissue type. It has been shown that GLAST, GLT1 , and EAAC1 are N-glycosylated in j cultured cells and intact rat brain (Danbolt et al., 1992; Stork et al., 1992; Dowd et al., 1996). N-glycosylation has been shown to be important for the function of many plasma membrane proteins, such as the insulin receptor (Ronnett et al., 1 1984), EOF receptor (Slieker et al., 1986), LDL receptor (Edwards et al., 1989), and glycine amino acid transporter GLYT1 (Olivares et al., 1995). It is ^ postulated that N-glycosylation is important for the correct folding and oligomerization of many membrane proteins and acts as a quality control for the export of newly synthesized proteins from the ER (Montreuil et al., 1995). Nglycosylation may also play a role in stabilizing glycoproteins against proteolytic degradation (Fiedler and Simons, 1 995). More recently, it has been shown that ' : the N-glycan chains may also serve as one form of the selective sorting determinant for the correct intracellular targeting of the glycoproteins (Aridor and Balch, 1996). However, it also has been shown that abolishment of the Nglycosylation of GLAST, using site-directed mutagenesis, had no effect on the function of this transporter (Conradt et al., 1995). At least two members of this transporter family, GLT1 and EAAC1 , have been shown to be regulated by protein phosphorylation through the PKC pathway, and maybe the PI3K pathway as well (Casado et al., 1 993; Dowd and Robinson, 1996; Davis et al., 1998). It was suggested that the increase in transport activity associated with EAAC1 protein phosphorylation correlated

PAGE 22

9 with an increase in cell surface expression of the transporter and could not be attributed to the biosynthesis of new transporter (Davis et al., 1998). Tissue and Cell Distribution of the Five SDHA Glutamate/Aspartate Transporters Among these five members of the SDHA glutamate/aspartate transporter family, the EAAC1 transporter appears to be the most ubiquitously expressed. Although quite abundant in brain, a significant level of expression of this transporter can also be detected outside the nervous system: in small intestine, kidney, heart, skeletal muscle, lung, liver (Kanai and Heidiger, 1992; VelazFaircloth et al., 1996), and placenta (Matthews et al., 1998). Although the expression of the GLT1 , GLAST, and EAAT4 transporters is extremely high in the central nervous system, a low level of the proteins or their correspondent mRNAs are also found in other tissues and cells, including human fibroblasts, placenta, heart, lung, and skeletal muscle (reviewed by Malandro and Kilberg, 1996). EAAT4 is expressed exclusively in cerebellar Purkinje cell soma and dendritic trees (Furuta et al., 1997), whereas EAAT5 is believed to be mainly confined to the retina (Arriza et al., 1997). Within the brain, EAAT4 mRNA is strictly confined to cerebellum (Fairman et al., 1995), whereas EAAC1, GLAST, and GLT1 mRNAs and proteins can be detected throughout the brain (Gegelashvili and Schousboe, 1998). However, when the brain is further dissected into distinct regions, each of these transporters is enriched in a particular area and among different neural cell types. Using in situ hybridization, immunocytochemistry, and

PAGE 23

10 immunohistochemistry techniques, it has been shown that GLT1 and GLAST are exclusively expressed in astrocytes, whereas EAAC1 is expressed only in neurons in the brain (Rothstein et al., 1994; Torp et a!., 1994; Lehre et al., 1995). Moreover, EAAC1 is primarily restricted to dendrosomatic compartment of the glutamatergic and nonglutamatergic neurons, though some immunoreactivity was also found in the presynaptic boutons of GABAergic neurons (Rothstein et al., 1994). This observation suggests that EAAC1 may not be a major contributor for the proper termination of the glutamateric neurotransmission through removal of glutamate from the synaptic clef. In fact, this is consistent with the data obtained from the knockout experiments which will be described below. Phvsioloaical Significance of the SDHA Glutamate/Aspartate Transporters SDHA glutamate/aspartate transporters are key components in the synaptic termination for glutamatergic neurotransmission. Because there is no mechanism for enzymatically degrading the released excitatory neurotransmitters in the synaptic cleft, their immediate removal by neural cells is solely responsible for the termination of the neurotransmission. Both neurons and astrocytes exhibit high capacity for glutamate uptake, although, as described below, astroglial appeared to be the primary site for glutamate clearance (uptake and biotransformation) in brain areas with high glutamatergic activity (Rothstein et al., 1996; Schousboe, 1981; Schousboe and Divac, 1979). It is estimated that the glutamate transporters can concentrate glutamate more

PAGE 24

11 that 10,000-folci across the cell membrane. Consistent with this proposal, the intracellular glutamate concentration in neurons is as high as 10 mM, whereas the synaptic concentration is kept at ~ 1 uM (Hediger, 1994). Therefore, SDHA glutamate transporters are believed to play an important role in protecting neurons from excitotoxicity caused by abnormally high concentrations of extracellular glutamate. In addition, each of these SDHA glutamate transporters has been shown to have a specific and unique distribution, spatially and temporally, during brain development, suggesting that they may also play important roles in brain maturation (Bar-Peled et al., 1997). Selective in vivo and in vitro knockout of the individual glutamate transporters using anti-sense oligonucleotides showed that reduction of GLT1 , and to a much lesser extent GLAST, but not EAAC1, resulted in neurodegeneration (Rothstein et al., 1996). These results provided additional proof that astroglial uptake of glutamate by ' GLT1 and GLAST is the major mechanism in excitotoxicity and synaptic clearance of this neurotransmitter. There is still great uncertainty as to the major function of the neuronal EAAC1 transporter. It is postulated that the EAAC1 transporter may play a role in keeping the neuronal intracellular glutamate at high levels for use as a neurotransmitter, as a precursor for GABA synthesis, or for other metabolic reactions in the brain (Kanai et al., 1995a). In addition, consistent with its ubiquitous expression among different organs and tissues, EAAC1 transporter may function as the major glutamate transporter providing glutamate and aspartate for general metabolism and other intracellular functions. Glutamate has been shown to be important for a variety of cellular

PAGE 25

12 functions including cell differentiation, proliferation, and migration (Pearce et al., 1987; Mattson eta!., 1988). ., As mentioned above, the transport activity of the SDHA glutamate/aspartate transporters is coupled with the co-transport of three Na* and one H* and counter-transport of one K* (Zerangue and Kavanaugh, 1996). As a consequence of this stoichiometry, the functional state of the SDHA glutamate transporter(s) may contribute to the rise of extracellular glutamate to neurotoxic levels in pathological conditions such as anoxia and ischemia after stroke (Kanai, et al., 1995b). In these pathologic conditions, the reduced cellular ATP level might cause a breakdown of electrochemical gradients across the membrane, a rise in extracellular K*, a decrease in extracellular Na*, and depolarization of the membrane (Szatkowski and Attwell, 1994). In this circumstance, the SDHA glutamate transporters are proposed to run in reverse, resulting in a non-vesicular release of glutamate to toxic levels into the synaptic cleft, and subsequently leading to neuron damage (Nicholls and Attwell, 1990; Kanai et al., 1995). Therefore, selective inhibitors of the neuronal glutamate transporter, EAAC1 , may be of therapeutic interest for preventing reversed glutamate transport, without affecting the capacity of glial glutamate transporters to keep the synaptic glutamate concentration at low levels. The GLT1 mRNA and protein are down-regulated in the brain following transient ischemia, while the expression of GLAST and EAAC1 are not altered (Torp et al., 1995). Given that GLT1 appears to be the pnmary mechanism for clearance of glutamate from the synapse (Rothstein et al., 1996), it is possible that the

PAGE 26

13 decrease of GLT1 activity is one of several factors that contribute to the high sensitivity of neurons to post-ischemic damage. Some of the sporadic forms of amyotrophic lateral sclerosis (ALS), a neurodegenerative disease caused by a slow loss of motor neurons, have been associated with a reduction of GLT1 glutamate transporter (Rothstein et al., 1992; Rothstein et al., 1995). It was later shown that aberrant RNA processing of the GLT1 transporter was responsible for the 60 70% reduction in GLT1 protein expression in the motor cortex and spinal cord of ALS patients (Lin et al., 1998). The presence of the aberrant GLT1 mRNA species in cerebrospinal fluid may have diagnosis utility for ALS. Decreased glutamate transporter '" protein in the cortex, especially GLT1 , has also been implicated in the neurodegeneration that occurs in Alzheimer disease (Cowburn et al., 1988; Scott et al., 1995). In contrast, schizophrenia and other psychoses are thought to result, at least partially, from glutamatergic hypofunction, a condition caused by excessive glutamate uptake (Carlsson and Carlsson, 1990). Biosynthesis and Dearadation of Membrane Glycoproteins As stated above, the activity of amino acid transporters has been extensively studied over the past thirty years, but little is known about their biosynthesis, degradation, and intracellular trafficking due to the lack of antibodies. In the past 5 years, the cloning and expression of a number of mammalian amino acid transporters has led to the generation of sequencespecific antibodies from corresponding peptides and fusion proteins. Therefore,

PAGE 27

14 Studies of amino acid transporter "life cycle" can be performed, for the first time, with these specific antibodies. I will use the term "life cycle" to describe the collective process of biosynthesis, targeting, and then degradation which requires many steps of membrane vesicle trafficking carrying the protein from one subcellular compartment to another. In recent years, the intracellular trafficking of many membrane proteins, including hormone receptors, major histocompatibility complex, ion channels, and glucose transporters, has been widely studied in eukaryotic cells (Alberts et al., 1994). Many of the individual steps that contribute to the life cycle of membrane proteins have been documented. Given that all of the amino acid transporter sequences elucidated so far encode for N-glycosylated integral membrane proteins, it is likely that many similanties exist between the life cycle of amino acid transporters and other membrane N-glycoproteins. There are basically two major pathways involved. One is the biosynthesis/exocytosis pathway, which transfers newly synthesized membrane proteins to their different cellular destinations. The other is the endocytosis/recycling pathway, which internalizes plasma membrane proteins and then either stores them in recycling vesicles or targets them for degradation (Alberts et al., 1994). Transfer of proteins from one compartment to another is achieved through the sequential movement of membrane protein cargo between distinct compartments, a process mediated by budding of small membrane vesicles, migration and recognition by the next compartment, and subsequent membrane fusion. The core protein machinery that underlines vesicle transport includes coat proteins, which sculpt a vesicle out of a donor

PAGE 28

15 membrane; the vesicleand target-specific identifiers v-SNAREs and tSNAREs, which bind to each other and thereby dock the vesicles to the acceptor membrane; small GTP-binding proteins (GTPase), which hydrolyze bound GTP to GDP in regulation of vesicle budding and docking; and NSF and SNAP proteins, which bind to the SNARE complex and initiate fusion when NSF hydrolyzes ATP (Nuoffer and Balch, 1994). In most cases, these pathways are highly regulated to meet the needs of the cells. De Novo Biosvnthesis and Intracellular Targeting of the integral Membrane NGlvcoproteins De novo biosynthesis of an integral membrane N-glycoprotein initiates with the co-translational insertion of the nascent polypeptide into the rough ER membrane (Alberts et al., 1994). In most cases, although the nutrient transporters cloned to date are an exception, the ER import step requires the newly synthesized polypeptide to possess an ER signal peptide, which is recognized by a signal recognition particle (SRP). SRP binds both the growing polypeptide chain and the ribosome and directs them to a receptor protein on the cytosolic surface of the rough ER membrane. This docking of the ribosomepolypeptide-SRP-SRP receptor complex on the membrane initiates the translocation process that threads a loop of polypeptide chain across the ER membrane through a hydrophilic pore in a protein translocator (Rapoport, 1991).

PAGE 29

16 All of the amino acid transporter sequences that have been elucidated have predicated N-glycosylation site(s), that is, an Asn-X-Ser/Thr consensus sequence, where X can be any amino acid except possibly proline and asparagine. Having this sequence stretch does not guarantee N-glycosylation of the protein, because other features of the polypeptide also have an influence on whether such potential sites become glycosylated or not. The amino acid transporter proteins containing one or more of these sequences that have been investigated are all N-glycosylated. For the biogenesis of N-glycoproteins, the nascent polypeptide is co-translationally modified, in the ER, with a preformed high-mannose oligosaccharide chain [(GlcNAC)2{Man)9(Glc)3] transferred from dolichol-P-P to Asn in the peptide (Voet and Voet, 1990). This initial Nglycosylation step is identical for ail N-glycoproteins, but the later processing in the ER and Golgi complex by a combination of trimming and addition of specific carbohydrate residues varies widely depending upon the type of the cell, the individual glycoprotein, and possibly the physiological conditions. Processing of the oligosaccharide begins immediately in the ER upon the attachment of the core sugar complex. Two specific glycosidases remove the first two glucose residues in sequence, generating Asn-GlcNAC2Man9 (Alberts et al., 1994). Reglucosylation of the core oligosaccharide in the ER can also happen. It is believed that these glucose residues can interact with the protein chaperones in the ER to retain the newly synthesized N-glycoprotein until it is correctly folded, that is, functioning as quality control.

PAGE 30

17 After the polypeptide is correctly folded, the N-glycoprotein will be transported from the ER to the Golgi compartment by small membrane vesicles budding from the specialized areas of the smooth ER, migrating, then recognizing/fusing with cis-Golgi network (Alberts et al., 1994). While in the Golgi complex, the attached oligosacchande chain(s) will undergo further glycosylation modification, in particular, they are altered from high-mannoseto complex-type in the cis Golgi. This processing pathway is highly ordered and most of the enzymes involved are rather strictly compartmentalized. Therefore, different oligosaccharide residues can serve as hallmarks for individual steps as the protein passes through the ER and Golgi complex. The mature Nglycoprotein will be targeted from the trans Golgi network (TGN) to its functional site, for example, the plasma membrane, through clathrin-coated vesicles (Schmid, 1997). Endoglycosidases are a class of enzymes that can catalyze the cleavage of oligosaccharide chains at specific sugar residues. These enzymes are often useful for characterizing the oligosaccarides on glycoproteins and elucidating the progress through the protein maturation pathway by examining the changes in the sensitivities of the glycoprotein to different enzymes. The best characterized and most widely used enzyme for this purpose is endoglycosidase H (Endo H). When a glycoprotein passes through the medial Golgi compartment, the attached oligosacchande chain will be processed from high-mannoseto a complex-type chain. Endo H cleaves between the two GlcNAC residues adjacent to the Asn residue, but it specifically acts on the

PAGE 31

18 high-mannose and some hybrid types of N-linked oligosaccharides (Robbins, 1984). Therefore, resistance to EndoH digestion can be used as a hallmark for the N-glycosylated proteins that have already proceeded beyond the medial Golgi compartment. In contrast, protein N-glycosidase F (PNGase F) cleaves between the first GlcNAC residue and the Asn residue on all N-glycosylated proteins, and can be used to determine whether a protein is N-glycosylated or not (Maley, 1989). N-glycosylation has been shown to be important for the functions of many plasma membrane proteins, including insulin receptor (Montreeuil et al., 1995; Ronnett et al., 1984), EOF receptor (Slieker et al., 1986), LDL receptor (Edwards et al., 1989), and glycine amino acid transporter GLYT1 (Olivares et al., 1995). N-glycosylation has been shown to be important for the correct folding and oligomerization of many proteins and acts as a quality control for the export from the ER (Ware et al., 1995; Nauseef et al., 1995). Also, it has been suggested that glycosylation may play an important role in stabilizing glycoproteins against proteolytic degradation (Fiedler and Simons, 1995). More recently, distinct N-glycan chains have been shown to comprise one form of a potential sorting signal for the selective transport of N-glycoproteins along their biosynthetic pathway through recognition and binding with specific lectins (Aridor and Balch, 1996). Nevertheless, abolishment of the N-glycosyiation of the SDHA glutamate/aspartate transporter GLAST by site-directed mutagensis had no effect on its plasma membrane arnval and its transport activity (Conradt etal., 1995).

PAGE 32

19 For the amino acid transporter proteins, little is known about how long it might take for an N-glycosylated membrane protein to be synthesized and then targeted to the PM, and even less is known about how long it will stay there. Cariappa and Kilberg (1990) elucidated the biosynthesis and intracellular targeting of an amino acid transporter, System A, using functional assays at a time when no antibodies were available for any amino acid transporter. The de novo biosynthesis of rat hepatic System A was up-regulated by the treatment with glucagon and dexmethasone. Using a Golgi subfractionation technique, they showed that the initial increased transport activity could be detected in the cis Golgi at approximately 45 min following hormone treatment, and in the remaining Golgi fractions and at the cell surface after 60 min. These results were consistent with whole cell transport data showing a 1 hr lag prior to the protein synthesis-dependent increase in plasma membrane System A transport activity following hormone treatment (Christensen and Kilberg, 1987). Studies also have shown that the hormone-induced System A activity has a half-life about 1 ,5 hr after hormone withdrawal (Handlogten and Kilberg, 1984). However no individual integral membrane N-glycosylated amino acid transporter protein has been documented with regard to its de novo biosynthesis and targeting time, as well as its residence time at the PM. Deoradation of Plasma Membrane Proteins After the biosynthesis and targeting of an integral membrane Nglycoprotein to its functional destination has been completed, the part of its life

PAGE 33

20 cycle that is still to come is the degradation. "Birth" and "death" go hand-inhand, this is true for the fate of a protein as much as for any other "living" things. The homeostasis of membrane proteins inside cells is highly regulated through not only the rate of protein synthesis, but also protein degradation. In response to alterations in the environment, proteins that are no longer needed can be eliminated, and their amino acids can be reutilized. The continued turnover of cellular proteins may also prevent the accumulation of a variety of deleterious nonenzymatic modifications such as oxidation, deamination, and glycosylation. In fact, the reduced degradation rates of certain proteins in aged tissue may contribute to the accumulation of aberrant proteins in aging. Furthermore, certain pathways of proteolysis may have evolved to eliminate mistakes in protein biosynthesis and assembly. Rapid and selective degradation of abnormal and mutant proteins is crucial for cell survival (Doherty and Mayer, 1992). To date, the mechanism(s) responsible for the degradation of amino acid transporters, as v^ell as for any other mammalian nutrient transporters, have not be documented. However, limited research has been done to elucidate the pathways by which membrane proteins in general are broken down in mammalian cells (Hare, 1990). It is logical to postulate that the degradation of the plasma membrane amino acid transporters may follow one of more of these pathways, described below. One of the mechanisms by which plasma membrane proteins are removed and eventually degraded is by a process known as shedding (Hare,

PAGE 34

21 1990; Beaudoin and Grondin, 1991), which means that the proteins are discharged from the cell surface and degraded outside of the cells. Shedding may occur by proteolytic release of the extracellular domains of the membrane proteins or by the release of intact membrane proteins along with membrane lipids in the form of small membrane vesicles. Interestingly, both of these events may be dependent on an initial membrane protein internalization, because lowering the temperature or incubating with lysosomotropic amines prevented the shedding (Johnston and Bystryn, 1984; Teixido et al., 1987). As to the proteolytic cleavage, it was postulated that the internalization of the shed proteins into acidic vesicles would trigger their proteolytic release from the cell surface upon recycling of the vesicles back to the cell surface, probably due to their increased sensitivity to proteases (Hopper et al., 1985; Teixido et a!., 1987). For membrane vesicle shedding, the process started with the endocytosis of the clathrin-coated vesicles involving a small domain of the PM containing the condemned proteins, and transfer to multivesicular endosomes. The large vacuoles containing cell surface glycoproteins arising from these vesicles then return their cargo to the extracellular space by exocytosis (Hare, 1990). The second possible mechanism involves the degradation of membrane glycoproteins in the lysosomes. Endocytosis and lysosomal degradation pathway is well characterized, and abundant evidence shows that it is important for the turnover of many plasma membrane proteins. There are different types of endocytosis in eukaryotic cells, distinguished on the basis of the size of the

PAGE 35

22 endocytic vesicles formed. Among them, the best characterized endocytosis pathway is clathrin-mediated endocytosis, which is also referred to as receptormediated endocytosis (Alberts et al., 1994; Robinson et al., 1996; Steer and Hanover, 1991). The key event during clathrin-mediated endocytosis is the recruitment of soluble clathrin from the cytoplasm onto the intracellular side of the plasma membrane to form a coated-pit. This process is thought to be mediated by the protein complexes called adaptors, which are components of the coat, forming an inner layer and attaching the clathrin to the membrane. One of the hallmarks of clathrin-coated vesicles is their selectivity. Certain membrane proteins, notably receptors for extracellular ligands and some nutrient transporters, such as LDL, Tf, and EGF receptors and GLUT4 transporter, are very efficiently concentrated into clathnn-coated vesicles (Bradbury and Bridges; 1994; Holman et al., 1994). This selectivity of endocytosis would, at least partially, explain why different turnover rates are associated with different proteins, even though all reside at the plasma membrane. In most cases, this selective concentration property of endocytosis has been correlated with the presence of an "internalization signal", for example a Tyror Leu-based sorting motif, in the cytoplasmic domain of the membrane protein (Sandoval and Bakke, 1994). There is compelling evidence from in vitro studies showing that the selective sorting is accomplished through the recognition and binding of the internalization signal of the membrane proteins by the adaptors. The clathrin-coated pits then pinch off to form coated vesicles, and this step is regulated by a GTPase, dynamin (Nuoffer and Balch, 1994).

PAGE 36

23 Following the internalization from the cell surface, membrane proteins, lipids, and solutes enter early sorting endosomes. From there, some endocytosed receptors that will be sorted back to the plasma membrane, such as the Tf receptor, also pass through a separation recycling compartment (Steer and Hanover, 1991). This endocytosis-recycling pathway has been best described for several receptor proteins including the LDL and insulin receptors (Alberts et a!., 1994; Robinson et al., 1996; Steer and Hanover, 1991). Over the past few years, more and more solute transporters, such as GLUT4, CFTR CI" channel, H2O channel, and H*-pump, have been shown to undergo a similar regulated recycling pathway (reviewed by Bradbury and Bridges, 1994). From early endosomes, internalized molecules can also proceed to further steps along the endocytic pathway, late endosomes and lysosomes, to be degraded. In vitro assays show that endosomal carrier vesicles, which mediate the vesicle transport between early and late endosomes, can fuse with late endosomes in a microtubule-dependent fashion, but not with early endosomes or with each other (Robinson et al., 1996). It was postulated that microtubles act as tracks along which carrier vesicles can move from one membrane compartment to another. This process is especially important for the transfer of endocytic vesicles from early endosomes to endosomal carrier vesicles and from there to late endosomes. The microtubule depolarizing drug, nocodazole, blocks the transport between early and late endosomes (Gruenberg et al., 1989). Transport from early endosomes to late endosomes, and then to lysosomes, also depends on the acidification of the endosomes and

PAGE 37

24 lysosomes (Steer and Hanover, 1991). The same or similar vacuolar H*ATPase is thought to acidify all the endocytic compartments, including early endosomes (pH 6), late endosomes (pH 5.5), and lysosomes (pH 5). This acidic environment plays a crucial part in the function of all these organelles, including the proper vesicle transport, release of bound ligands from their receptors, and proteolysis (Alberts et al., 1994). Most of the hydrolytic enzymes that reside inside lysosomes and late endosomes for protein degradation are most active at low pH. Abrogation of the acidification environment of the endocytic compartments using either bafilomycin A1 , a specific inhibitor for the vacuolar H*-ATPase, or a weak base, such as NH4CI or chloroquine, can block the trafficking along the endocytic pathway, as well as inhibit the function of the lysosomal proteases (Steer and Hanover, 1991). Therefore, these inhibitors can be used to test whether the degradation of a particular protein is mediated by the endocytosis/lysosomal pathway. The third possible pathway responsible for the membrane protein degradation is the ubiquitination-proteasome pathway (Alberts et al., 1994; Hicke, 1997). Ubiquitin, a small globular protein of 76 amino acids with a protruding carboxyl-terminus, was found to serve as a covalent cofactor for ATP-dependent proteolysis in cytosol (Olson et al., 1992; Hochstrasser, 1996). Ubiquitin is conjugated to the condemned protein via ubiquitin's carboxylterminal glycine residue to form isopeptide bonds with available lysine residues in the target protein. The result is to produce a novel post-translational modification of branched polypeptides. When a targeted protein is poly-

PAGE 38

25 ubiquitinated, a chain of up to 20 ubiquitins can be attached to a single lysine residue of the modified protein. This poly-ubiquitination is accomplished by repeated addition of single ubiqutin units through isopeptide bonds involving lysine-48 of each additional ubiquitin and the carboxyl-terminal glycine residue of each succeeding ubiquitin. Multiple ubiquitination of this type has been shown to serve as a recognition signal for the 26S proteasomal complex, which contains the multicatalytic 20S proteasome as its catalytic core (Alberts et al., 1994; Hochstrasser, 1996)). Ubiquitin is released during protein degradation to be re-used in future rounds of protein catabolism. There are also deubiquitination enzymes that can reverse the ubiquitination reaction and unmodify the target protein (Kalderon, 1996; Wilkinson, 1997). This latter pathway may be part of the quality control system examining each protein continuously and selecting certain protein molecules for degradation while releasing others to continue to function. The ubiquitination-dependent proteasomal degradation pathway has been best characterized for its pivotal role in regulating the decay of many cytosolic short-lived regulatory proteins, which are involved in a diverse array of regulatory events including cell cycle progression, DNA repair, and transcriptional control. Relevant to the degradation of membrane proteins, Ward and coworkers (1995) first showed that the ubiquitination and proteasome degradation pathway was required for the rapid turnover of the cystic fibrosis transmembrane conductance receptor (CFTR), an integral membrane protein. It was shown that the degradation of both the wild-type and the mutant CFTR was

PAGE 39

26 dramatically reduced by incubation with specific inhibitors for proteasomal proteases. Incubation with these potent proteasome inhibitors caused an accumulation of the ubiquitinated immature CFTR. By expressing a mutant ubiquitin (K48R) and eliminating the formation of the polyubiquitin chain on target proteins, they further confirmed that polyubiquitination is required for the rapid degradation of CFTR. In contrast to the ubiquitination-proteasome pathway, Strous and coworkers (1996), using temperature-sensitive mutant CHO cells for ubiquitinactivating enzyme E1, showed that ubiquitination of the human growth hormone receptor is required for the ligand-induced endocytosis and degradation through the endosomal/lysosomal pathway. Interestingly, they showed that, ubiquitination of the receptor was dependent on an intact endocytosis pathway, suggesting there was a coupling mechanism between the ubiquitination and endocytosis of this receptor. More recently, Sphngeal and Andre (1998) reported that, in yeast, ubiquitination of the nutrient permeases is associated with their internalization and degradation. Furthermore, Terrell et al. (1998) published their studies indicating that the mono-ubiquitination of a G proteincoupled receptor can serve as an internalization signal for its endocytosis. This observation suggests that the mechanisms that recognize mono-ubiquitination as an internalization signal for membrane proteins is significantly different from that detecting poly-ubiquitination as a proteasome recognition signal. In another words, the fate of ubiquitinated proteins may be decided by the number and topology of the ubiquitin attached.

PAGE 40

27 Finally, membrane glycoproteins may also be degraded through individual proteases that are affiliated with the membrane or reside closely to the target proteins. For example, calpain is a calcium-activated protease, which consists of a regulatory subunit and a catalytic subunit (Doherty and Mayer, 1992). The smaller regulatory subunit contains an "EF hand" domain typical for some calcium-binding proteins, and a hydrophobic glycine-rich domain which may associate the enzyme with cell membrane. The other larger subunit also contains an "EF hand" domain and a potent cysteine protease catalytic site. Activation of calpain has been implicated in the degradation of the membrane cytoskeletal protein fodrin (Fukuda et al., 1998; Blomgren et al., 1995; Yokota et a!., 1995), actin (Potter et el., 1998), as well as intermediate filament proteins (Resing et al., 1993). Given the previously described association of fodrin with the NaVK*-ATPase (Nelson and Hammerton, 1989) and System A (Handlogten et al., 1996), it is tempting to hypothesize that this or other similar membrane bound or affiliated proteases also specifically act on integral membrane proteins and directly regulate their activity. C6 Glioma Cell and Its Utilization in SDHA Glutamate/Asoartate Transporter Studies C6 glioma cells, a rat glial tumor cell line, was ohginally isolated as a S100 protein-producing clone from rat glial tumors that had been induced by Nnitrosomethylurea. The S-100 protein is a highly acidic protein unique to the vertebrate brain and has been found in numerous brain tumors, both from man and other animals (Benda et al., 1968). This protein was named on the basis of

PAGE 41

28 its solubility in 100 percent saturated ammonium sulfate at neutral pH. Cultured C6 glioma cells have been used extensively to study various aspects of glial biochemistry and physiology. These cells exhibit several biochemical features of normal glial cells, such as expressing S-100 protein (Benda et al., 1968), glial fibrillary acidic protein (GFAP) (Bissel et al., 1974), glutamine synthetase (GS) (Parker et al., 1980), and 2',3'-cyclic nucleotide-3'-phosphohydrolase (CNR) (Zanetta et al., 1972). Brain glial cells can be classified into three subtypes: astrocytes, oligodendrocytes, and microglial (Cooper et al., 1996). Astrocytes are present mainly in regions of axons and dendrites; they tend to surround and closely contact the adventitial surface of blood vessels. Besides the possible insulation and organization roles suggested by its structural characteristics, astrocytes can accumulate glucose, synthesize glycogen, and provide energy substrates to neurons. Whereas, the oligodendrocytes form the myelin sheath along the axons, and the microglial may play a role in signaling for the recruitment of lymphocytes and leukocytes during the repair of the damaged brain tissue (Cooper et al., 1996). GFAP and GS are considered as specific marker proteins for astrocytes, whereas CNP is a marker for oligodendrocytes (Varon, 1978). The presence of both astrocyte and oligodentrocyte properties in C6 cells suggest that they may be most closely comparable to the lessdifferentiated glial stem cells present in the developing brain, which further differentiate into either of the mature glial types (Bhat et al., 1984). The expression of these unique cell type properties can be regulated within C6 cells by a variety of culture conditions, including cell density, serum

PAGE 42

29 deprivation, and culture passage. Varon (1978) first found that the quantity of GFAP, a specific marker for mature astrocytes, was ten-times greater in stationary phase than in log phase C6 cells. Later, Maltese and Voipe (1979) showed that the expression of an oligodendrocyte marker, CNP, was also significantly up-regulated by growing C6 cells to high density or by culture of non-confluent cells in serum-free medium. When GS and CNP were used to study C6 cells in culture, Parker et al. (1980) reported that at early passages the cells predominantly showed oligodendrocyte-like properties, whereas at later passages they predominantly showed astrocyte-like properties. Ail of these observations are consistent with the proposal that C6 glioma cells may be analogous to glial stem cells, which can be differentiated to specific mature glial-like cells under different culture conditions. C6 glioma cells have a high level of SDHA glutamate/aspartate transport activity (Deas and Erecinska, 1989). As to which of the SDHA glutamate transporters are responsible for the high transport activity seen in this cell line, there have been some conflicting reports. Casado and coworkers (1993), using C6 cells, studied the regulation of SDHA glutamate transporter(s) by protein phosphorylation through the PKC pathway. They reported that phosphorylation of the GLT1 transporter protein itself was responsible for the increased transport activity and suggested that the glial-specific transporter GLT1 was expressed in the C6 cells. In contrast, Palos and coworkers surprisingly observed that mRNA for the neuronal specific transporter, EAAC1 , but neither of the glial-specific transporters, GLAST or GLT1 , could be detected in C6

PAGE 43

30 glioma cells using Northern blotting analyses (Palos et al., 1996). Consistent with these Northern blotting data, Dowd et al. (1996) showed that C6 glioma cells expressed EAAC1 but not GLAST, GLT1 , or EAAT4 based on immunoreactivity as detected by Western blotting. The same authors further reported that the EAAC1 -mediated SDHA glutamate transport activities in C6 and in EAAC1 mRNA-injected oocytes shared similar kinetic parameters, but that these kinetics were different from those observed in rat synaptosomes, suggesting that synaptosomes contain a different, or at least a heterogeneous population, of glutamate transporters. The effect of protein phosphorylation on the EAAC1 transporter, directly or indirectly, has been studied using C6 glioma cells as a model system (Dowd and Robinson, 1996; Davis et al., 1998). These investigations showed that uptake activity and cell surface expression of EAAC1 was increased by phorbol ester treatment and with using specific inhibitors, the experiments suggested that both the PKC and PI3K pathways might play roles in regulating EAAC1 transporter function. The expression of a neuronal-specific SDHA glutamate transporter, that is, EAAC1, in a glial tumor cell line is surprising but not unprecedented. It has been documented that some types of glial tumor cells express glucose transporters, GLUT1 and GLUT3, not usually expressed by normal glial cells (Boado et al., 1994). Similarly, certain PC 12 pheochromocytoma cell variants were found to be more sensitive to a toxic aspartate analog, alanosine, than wild-type PC12 cells, suggesting an up-regulation of the aspartate transport activity in tumor cells (Ramachandran et al., 1993).

PAGE 44

31 High-grade astrocytomas, of which C6 may represent a model system, are the most common form of malignant brain tumor in humans, and they are usually resistant to therapy (Bruner, 1994; Lesser and Grossman, 1994). Therefore, understanding the atypical expression of nutrient transporters in these tumor cells may be useful for future clinical diagnosis and chemotherapy. In this work, I used C6 glioma cells as a model system to study the biosynthesis, degradation, and intracellular targeting of EAAC1 glutamate/aspartate transporter protein. These results obtained from this study may provide us with the basis for studying the role of EAAC1 transporter under diseased states, as well as may serve as a model for studying the life cycle of other mammalian transporter proteins.

PAGE 45

CHAPTER 2 MATERIALS AND METHODS Materials L-[2,3-^H]-Aspartic acid (1mCi/ml) (Cat# TRK445), protein A-HRP (Cat# NA9120), Hyperfilm™ MP (Cat# RPN1678H), and Promix™ Lf =S]-MethionineCysteine in vivo cell labeling mix (14.3 mCi/ml) (Cat# SJQ0079) were obtained from Amersham LIFE SCIENCE. The RT-PCR reaction reagents were all obtained from Gibco BRL: random hexemers (Cat#48190 011), random primers (Cat# Y01212), Superscript II reverse transcriptase (Cat# 18064 014), and Tag DNA polymerase (Cat# 18038-042). Goat-anti-rabbit IgG-HRP (Cat# 170-6515), pre-stained protein standards (broad-range) (Cat# 16100318) were obtained from BioRad Laboratories. The EZ-Link™Sulfo-NHS-LC-Biotin (Cat# 21335), free D-biotin (Cat# 29129), SuperSignal ULTRA chemiluminescence substrate (Cat# 34075), ImmunoPure immobilized monomeric avidin-sepharose (Cat# 20228) were obtained from PIERCE. L-methionineand L-cysteine-free DMEM (Cat# 21013) was obtained from Gibco BRL. Protein endoglycosidase F (Cat# 704S) and endoglycosidase H (Cat#702S) were obtained from NEW ENGU\ND Biolabs Inc. SuperFect transfection reagent was obtained from Qiagen. Goat-anti-mouse IgG-HRP was obtained from Kirkegaard & Perry Laboratohes (Cat# 074-1807). Minimum Essential Medium (MEM) (Cat# M32

PAGE 46

33 0643), protein A sepharose CL-4B (Cat# P3391 ), D-aspartate (Cat# A-8881 ), Lmethionine (Cat# M-9625), and L-cysteine (Cat# C-7755), polyoxyethylene 9 lauryl ether (C12E9) (Cat# P-9641), non-immune normal rabbit IgG (Cat# I8140), tunicamycin (T-7765), NH4CI (Cat# A-4514), N-CBZ-Leu-Leu-Leucinal (MG132) (Cat# C-2211), N-acetyl-leu-Leu-norieucinal (ALLN) (Cat# A-6185), leupeptin (Cat# L-2884), phenylmethyl-sulfonyl fluoride (PMSF) (Cat# P-7626), brefeldin A (BFA) (Cat# B-7651), and nocodazole (NOC) (Cat# M-1404) were al obtained from Sigma Chemical Co. All other chemicals were obtained through either Sigma Chemical Co. or Fisher. Methods Cell Culture C6 glioma cells and BNL CL.2 cells were obtained from American Type Culture Collection (ATCC number for C6 is CCL107; for BNL CL.2 it is TIB 73) and maintained in supplemented Eagle's medium (MEM) containing 10% FBS as monolayer cultures under a humidified atmosphere of 5% C02/95% air (37°C) for a maximum of eight passages in 75 or 1 75 cm' flasks. The cultured cells were transferred to 24-well cluster dishes for whole cell transport assays; to plastic 100 150 mm culture dishes for metabolic labeling, cell-surface biotinylation, and total cellular protein or membrane protein collection; or to 22 x 22 mm sterilized Corning glass microscope cover slips held in a Falcon six-well cluster trays for immunohistochemistry.

PAGE 47

34 Whole Cell Transport Ass ay Amino acid uptake of C6 glioma cells or other adherent cultured cells was measured by the cluster tray method of Gazzola et al. (1981) with modifications by our laboratory (Kilberg et al., 1989). One hundred thousand C6 cells were placed into each well of a 24-well tray and cultured for 20 to 24 hr under normal conditions. To partially deplete the intracellular pool of amino acids to minimize trans-effects on transport (Kilberg, 1989) as well as to remove extracellular Na*, cells were incubated at 37°C twice for 15 min each (2x15 min) in choiine-KRP. To initiate transport, ^H-amino acid and the appropriate inhibitor in 250 yl of either NaKRP or choline-KRP (37°C) was added simultaneously to each of the 24 wells in the cluster tray. The Na+-dependent transport is taken as the difference between uptake in NaKRP and choline-KRP. The transport measurement was terminated by discarding the radioactivity and rapidly washing the cells five times with 2 ml ice-cold choline-KRP. The data are expressed as pmol»mg"Vrotein«time"^ and typically, are presented as the average of 4 independent assays on at least two different batches of cultured cells. Gel Electrophoresis Gel electrophoresis was performed essentially following the protocol originally described by Laemmli (Laemmli, 1970). Protein samples were diluted with at least equal volume of sample dilution buffer (2X SDB) consisting of 2% (w/v) SDS, 5% p-mecaptoethanol, 30 pg/ml bromophenol blue, 20% glycerol,

PAGE 48

35 0.125 mM Tris-HCI, pH 6.6 -6.8. The amount of protein loaded per lane will be stated in each figure legend. Vertical 7.5% polyacrylamide slab gels were used and electrophoresis was performed at 30 to 40 mA constant current until the selected pre-stained protein marker band reached the end of the separating gel. Electrotransfer and Immunoblottinq . ^ v,^" ' ' For immunoblotting detection, the fractionated proteins were transferred electrophoretically onto a piece of nitrocellulose membrane In ice-cold transfer buffer containing 25 mM Tris-base, 190 mM glycine, 20% methanol at 299 mA and constant current for 1 8 to 20 hr. After the transfer, the blot was stained briefly in fast green stain (0.1% fast green, 50% methanol, 10% acetic acid) and de-stained (50% methanol, 10% acetic acid) to check for the efficiency of transfer and the evenness of loading. The blots were blocked in Tris-buffered saline/ 0.1% Tween-20 (TBS-T) (10 mM Tris, pH 7.5, 200 mM NaCI, and 0.1% Tween) containing 1% Carnation non-fat dry milk (NFDM) for 12 hr at room temperature (RT) or overnight at 4°C with constant agitation on an orbital shaker. The blots then were incubated in the same blocking solution containing primary antibody (1 : 2000 dilution for EAAC1 antibody) for 1 2 hr at RT in test tubes with constant "end-to-end" rotation. After extensively washing in TBS-T/1% NFDM to remove the unbound antibodies, the blots were incubated in the same blocking buffer with secondary antibody conjugated to horseradish peroxidase (1 :20,000 dilution of goat anti-rabbit IgG-HRP was used for EAAC1 blotting) for 1 hr at RT. The blots were extensively washed with TBS-T/ 1% NFDM before

PAGE 49

36 being visualized with SuperSignal Chemiluminescence detection reagents (PIERCE) following the manufacturer's instructions. Light emissions from the blots were captured on Hyperfilm MP (Amersham), and band intensity was quantitated in the linear range of the film on a Visage Bioscan video densitometer. As shown in Figure 2-1 , I typically detect three major bands on EAAC1 Western blotting which run with apparent molecular weights of 73, 145, and 219 kDa. All of these three protein bands were detected when either total C6 cellular membrane protein or anti-EAACI antibody immunoprecipitated protein was tested by immunoblotting. Pre-incubation of the anti-EAAC1 antibody with 10 yg/ml of specific EAAC1-MBP fusion protein completely inhibited the detection of all these bands. These results indicate that all these detected bands are specific EAAC1 protein bands, which may represent the monomer, dimer, and trimerform of EAAC1 protein, as also observed by others (Haugeto, et al., 1996; Davis et al., 1998). ; ' ) • f' t"'. v*'! Pulse-Chase Metabolic Labelino of C6 Glioma Cells To study the de novo biosynthesis, intracellular targeting, and degradation of EAAC1 transporter proteins in C6 cells, pulse-chase labeling with 35 L-[ S]-methionine-cysteine (ProMix, Amersham) was used. After placing 6 to 9 x 10® cells onto each 100 mm culture dish or 2.3 x 10^ cells onto each 150 mm dish, cell monolayers were cultured for 22 to 24 hr in normal MEM culture medium to permit growth to near confluence. The cells were washed once with

PAGE 50

37 Sterile 37°C PBS, pH 7.4, and incubated with 15 ml / 58 cm^ surface area of methionineand cysteine-free DMEM medium (Gibco) for 2 x 1 5 min at 37°C to deplete the intracellular pool of free methionine and cysteine. The depletion medium then was aspirated and the cells incubated with 200 pCi/ml of [^S]methionine-cysteine (PreMix cell labeling, Amersham) in methionineand cysteine-free DMEM medium at 37°C for 15 to 30 min. The cells were washed twice with 37°C MEM containing 5 mM of non-radioactive methionine and cysteine (chasing medium) and then transferred to fresh chasing medium and chased for 0 to 60 hr followed by immunoprecipitation of the EAAC1 protein. For experiments have chases longer than 24 hr, 1 % FBS was added to the medium. Immunoprecipitation and Fluoroqraphv Immunoprecipitation of the EAAC1 transporter protein was performed following the procedure outlined by Harlow and Lane (1988) with modifications. After pulse-chase labeling with L-[^S]-methionine and L-f^S]-cysteine, cells were washed twice with ice-cold PBS and once with SEB buffer (250 mM sucrose, 2 mM EDTA, 2 mM EGTA, and 10 mM HEPES, pH 7.5), and then frozen in 2.5 ml of SEB buffer containing proteinase inhibitors (1 mM PMSF and 2 Mg/ml each of leupeptin, aprotonin, pepstatin, TPCK, and TLCK) at -80°C. For analysis, cells were thawed on ice and another 2.5 ml of hypotonic EB buffer (2 mM EDTA, 2 mM EGTA, and 10 mM HEPES, pH 7.5) were added. Cells were scraped from the plates and homogenized on ice with 10 to 15 passes through a pre-chilled steel-block cell homogenizer with a clearance of 0.0025 inches

PAGE 51

38 (Auburn Tool & Dye, Warwick, Rl). The cell homogenate was centrifuged at 400 X g for 10 min to get rid of the unbroken cells and nuclei, and the supernatant was centrifuged at 280k x g (65k rpm with Beckman Ti70.1 rotor) for 12 hr at 4°C to collect the total membrane pellet, which was then extracted in PES buffer (2% C12E9, 0.1% SDS, 1 mM EDTA in PBS, pH 7,4) for 1 hr on ice with constant magnetic stirring. After centrifugation at 200k x g for 15 min at 4°C, the supernatant was recovered and the protein concentration was determined by a modified Lowry method (Kilberg, 1979) or the bicinchoninic acid (BCA) method (Pierce BCA Protein Assay kit), and then an equal amount of starting protein (100 to 500 yg) from each sample was transferred into microcentrifuge tubes, brought up to a volume of 500 ul with PES buffer, and used for the immunoprecipitation assay. To minimize non-specific interaction between labeled membrane proteins and the immunoglobulins, an unrelated non-immune rabbit serum IgG (5 ug) and a 50% (in PES buffer) suspension of protein Asepharose (50 pi) were added to the extracts and incubated ("pre-cleared") for 13 hr at 4°C with constant mixing. Samples were then centrifuged at 100k x g for 15 sec to collect the protein A sepharose IgG complexes and this "pre-cleared" supernatant was transferred to a new microcentrifuge tube. The specific antibody against EAAC1 (5 ug purified total IgG, # UF91 ) was added to the precleared extract and incubated at 4°C overnight with constant mixing. A 50 \x\ aliquot of 50% protein A-sepharose beads was then added and incubated for 2 hr to collect the immunoprecipitates. The pellets were washed with PES extraction buffer (3 x 1 min) and with PES containing 0.35 M NaCI (for a total salt

PAGE 52

concentration of 0.5 M) (4 x 10 min) at 4°C with constant mixing. As sliown in Figure 2-2, a series of salt washes were tested, and the data show that the binding between the antibody and the EAAC1 protein can withstand salt washing upto 1M NaCI and 0.5 M salt is sufficient to eliminate basically all of the nonspecific binding. The immunoprecipitated proteins were then eluted from the beads with sample dilution buffer (SDB) containing 6 M urea and 10% ME for 20 min at 37°C and separated by SDS-PAGE. The elution condition was determined by the data shown in Figure 2-5 below. Heating at 37°C for 30 min was shown to be the optimal elution condition, and was used for all the experiments described in this thesis. For Western blotting, the separated proteins were then electrotransferred onto a nitrocellulose membrane and immunoblotted with a specific anti-EAACI antibody as described before. For fluorography, the gels were fixed at RT in 10% TCA/ 40% methanol for 30 min, soaked in water for 30 min, and then incubated in 1 M sodium salicylate for 1 hr before drying at 65°C under vacuum. The dried gels were exposed to autoradiographic film at -80°C with an intensifying screen, and the band intensity was quantitated in the linear range of the film on a Visage Bioscan video densitometer. To test whether 5 ug of anti-EAACI antibody (purified total IgG from the immune rabbit serum) is enough to precipitate the EAAC1 protein from up to 500 yg starting material, immunoblotting was done to check for the remaining of EAAC1 protein in the immunoprecipitation supernatant. As shown in Figure 2-3, 5 ug of anti-EAACI antibody completely depleted the EAAC1 protein from the starting mixture. Also, the amount of EAAC1 protein immunoprecipitated by 5 ug

PAGE 53

40 anti-EAACI antibody is proportional to the amount of starting protein up to 500 ug (Figure 2-4). All these data suggest that 5 ug anti-EAAC1 antibody is sufficient to quantitatively immunoprecipitate EAAC1 protein from up to 500 ug starting material. Endoalvcosidase Digestions To test for the endoglycosidase H (Endo H)-sensitivity of the newlysynthesized EAAC1 , C6 cells cultured on the 150 mm dishes were pulse-labeled for 15 min with [^S]-methionine-cysteine and chased with medium containing an excess of unlabeled methionine and cysteine (5 mM each) for 30 to 240 min as described before. At the end of each chase period, EAAC1 was immunoprecipitated and then the transporter was eluted from the protein Asepharose beads with 10 pi of 5X denaturing solution containing 2.5% SDS, 5% 3-ME in water for 30 min at 37°C. After dilution to IX denaturing solution with 40 vil of water, each of the eluates was collected and then divided into two microcentrifuge tubes (20 yl each). A 1/10 volume of 500 mM sodium citrate, pH 5.5 and 500 1000 U (1 2 ul) of Endo H (BioLabs Inc.) was added to each of the tubes. For the non-PNGase F control tube, 1 2 yl of enzyme storage buffer (20 mM Tris-HCI, pH 7.5, 50 mM NaCI, and 5 mM NazEDTA), instead of the enzyme, was added. All samples were incubated at 37°C for 1 hr, and then mixed with an equal volume of 2X SDB buffer before the samples were loaded onto SDS-PAGE gel for separation and autoradiographic detection.

PAGE 54

41 To determine the N-linked glycosylation of both the newly synthesized lower molecular weight and the mature forms of the EAAC1 transporter protein, protein endoglycosidase F (PNGase F) digestions were performed. Total C6 cellular membrane protein was collected from the cells that were either pulsechase labeled as described above or not, and then solubilized with PES buffer and immunoprecipitated with specific anti-EAAC1 antibody as described before. The precipitates were then eluted with 1 0 ul of 5X denaturing solution (2.5% SDS, 5% 3-ME) for 30 min at 37°C, and the collected eluates were diluted to 1X denaturing solution with water as described above for Endo H digestions. For PNGase F digestion groups, 1/10 volume each of 10% NP-40 and a buffer consisting of 10% NP-40, 500 \iM sodium citrate, pH 7.5 as well as 500 1000 U of PNGase F (BioLabs Inc.) was added. For the non-Endo H controls, the enzyme was replaced with equal volume of enzyme storage buffer only. All the samples were incubated at 37°C for 1 5 to 240 min (a 60 min incubation time was commonly used), then mixed with an equal volume of 2X SDB buffer before they were loaded onto a SDS-PAGE gel for separation and autoradiographic detection. Cell Surface Protein Biotinvlation Cell membrane impermeable sulfo-NHS-LC-biotin was used to specifically label proteins that are exposed at the exterior surface of the cells. This cellsurface biotinylation method was utilized to study the plasma membrane (PM) localization and half-life of cell-surface exposed EAAC1 protein, as well as to

PAGE 55

42 determine the targeting time and residence time of newly synthesized EAAC1 in C6 glioma cells. For the determination of the cell-surface localization and the half-life of the PM EAAC1 protein, C6 glioma cells cultured on 100 mm dishes were washed twice with NaKRP buffer (1 19 mM NaCI, 5.9 mM KCI, 1 .2 mM KHCO3. 5.6 mM glucose, 25 mM Na2HP04, 0.5 mM CaCb, 12 mM MgS04, pH 7.5), and incubated with 0.5 1 mg/ml of sulfo-NHS-LC-biotin (Pierce Chemical Co.) in NaKRP for 30 min to 1 hr at 4 15°C. The optimum time and reagent concentration must be determined experimentally, and the specific conditions for each experiment are given in the figure legends. The biotin reagent buffer was aspirated and the cells were rinsed twice with fresh MEM containing 50 mM glycine, and then incubated in fresh 37°C MEM + 1 % FBS for specific chase times ranging from 0 to 24 hr at 37°C. At the end of each chase time period, the cells were washed once with ice-cold NaKRP, another time with ice-cold SEB containing protease inhibitors (described above), and then frozen in 2 ml SEB containing protease inhibitors at -80°C. The specificity of the cell surface biotinylation assay was examined, and the data are shown in Figure 2-6. The percentage of the biotinylation for the plasma membrane proteins, Na*/K* ATPase and EAAC1 , was 89% and 69%, respectively; versus below 2% for the intracellular proteins, cytoplasmic AS and ER GRP78. These results indicate that this method reliable for the study of the PM-localization of a protein. To determine the time it takes for the newly synthesized EAAC1 transporter proteins to arrive at the plasma membrane of C6 glioma cells, cell surface biotinylation was performed after the C6 cells were metabolically labeled

PAGE 56

43 with ^^S-methionine and -cysteine for 15 30 min and chased in medium containing an excess amount of non-radiolabeled methionine and cysteine for 0 to 36 hr. At the end of each chase time point, these metabolically-labeled C6 cells were washed twice with NaKRP and cell-surface biotinylated with 0.5 1 mg/ml of sulfo-NHS-LC-biotin in NaKRP at 15°C for 30 60 min. After aspirating the biotinylation reagents, the cells were rinsed once in NaKRP and then incubated with NaKRP containing 50 mM glycine for 2 x 15 min at 4 15°C to quench the remaining free biotin. Then the cells were washed once with ice-cold SEB containing protease inhibitors, and frozen at -80°C until all samples were collected. All cell-surface biotinylated C6 cells were thawed on ice and homogenized with a steel-block homogenizer, and total cellular membrane proteins were collected and solubilized with PES buffer as descnbed above. Equal amounts of solubilized proteins was then subjected to two precipitations: first with anti-EAACI antibody and protein A-sepharose beads to precipitate all of the EAAC1 protein, and after elution in low pH buffer, the eluates were precipitated again with monomehc avidin-sepharose beads to precipitate only the biotinylated EAAC1. For the immunoprecipitation of total EAAC1 protein, the protein samples were pre-cleared with non-immune rabbit IgG and then immunoprecipitated with specific anti-EAAC1 antibody as described before. The precipitated EAAC1 protein was eluted from the beads with 1 ml of 0.1 M glycine, pH 2.8 containing 0.5% TritonX-100 and 0.2% BSA at RT for 30 min. After centrifugation at 1 0k x g for 1 5 seconds to remove the beads, the eluate was removed and neutralized with 50 \i\ of 0.1M Tris-HCL, pH 9.5. For the

PAGE 57

44 precipitation of biotinylated EAAC1 protein, 25 yl of packed monomeric avidinbeads was washed twice with 1 ml of PES buffer and then pre-incubated for 1 2 hr with non-biotinylated total C6 cellular membrane proteins to block the nonspecific binding sites on the beads. Then the immunoprecipitated EAAC1 protein collected above was added to the pre-treated monomeric avidin-beads and incubated for 5 18 hr with constant mixing at 4°C. At the end of the incubation, the beads were washed extensively with 1 ml of PES supplemented with 350 mM NaCI for 40 min at 4°C with a total of six changes of buffer. The double precipitated proteins were eluted from the beads with 2X SDB containing 2 mM free D-biotin and separated by gel electrophoresis followed by fluorography as described before. Therefore, only the newly synthesized (^S-labeled) EAAC1 protein that has already arhved at the PM (cell surface-biotinylated) were detected with this double precipitation procedure. Monomeric avidin instead of streptavidin was used because the binding between biotin and monomeric avidin is reversible, whereas, the binding between biotin and streptavidin is nearly irreversible, even in sample dilution buffer. Therefore, when monomeric avidinsepharose is used, the precipitated biotinylated proteins can then be eluted from the beads with an excess amount of free-biotin. Immunohistochemistrv The immunohistochemistry of human fibroblasts was done as described by Woodard et al (1994). Cells cultured to 50 60% confluence on glass cover slips were rinsed in PBS and fixed with 4% paraformaldehyde/PBS for 30 min.

PAGE 58

45 then rinsed 3 times with PBS followed by quenching of free aldehyde by incubating for 30 min in PBS containing 50 mM glycine. Non-specific antibody binding was blocked by incubation in PBS containing 20% normal goat serum (NGS) for 2 hr. For pre-immune or immune serum incubation, cover slips were removed from the six-well trays, inverted and placed on 50 y! drops of antiserum diluted, to the appropriate level for each antibody, in PBS containing 20% NGS with or without 20 yg/ml of competing peptide. After incubation for 2 hr, the cover slips were placed back into the wells and then washed 3 times with PBS. The coverslips were incubated with the secondary antibody (typically 1:200) in 20% NGS/PBS for 1 hr in a similar manner as that used for the primary antibody. After the final wash, the coverslips were mounted onto glass slices with a drop of Fluoromount-G, allowed to dry, and the edges of the cover sip sealed with fingernail polish before being analyzed by fluorescent microscopy. Expression of EAAC1 Glutamate/Aspartate Transporter in BNL CL.2 Cells BNL CL.2 cells were both transiently and stably transfected with EAAC1 cDNA using SuperFect transfection reagent following the vendor's protocol (QIAGEN). The cells were cultured to about 40% confluency on 100 mm dishes and washed once with PBS. Twelve ug EAAC1 cDNA in PcDNA3 vector was mixed with 360 ul OptiMEM, and then with 48 yl of SuperFect transfection reagent. The mixture was incubated at RT for 5 10 min, and then diluted into 5 ml of OptiMEM immediately transferred onto the cells. After incubation with the transfection reagent for 2 3 hr at 37°C in a cell culture incubator, the cells were

PAGE 59

46 washed twice and the medium replaced with fresh MEM containing 10% FBS and 4.5 g/L glucose. For stable transfection, the culture medium contained 0.5 1 mg/L of G418 and the expressing cells were selected for at least 2-4 weeks with medium changed every 3 days. The cells were passed whenever they reached about 85 90% confluency. RNA Isolation and RT-PCR Total RNA from 1 x 10° C6 cells was isolated using RNeasy Midi Kits (QIAGEN) and then poly(A*) mRNA was purified from 500 ug of total RNA using Oligotex mRNA isolation kits (QIAGEN) following the vendor's protocols. The Reverse Transcriptase reactions were carried out following a standard protocol provided by the vendor of the transcriptase using random hexamers to prepare the first strand cDNA (Gibco). PGR primers for each of the glutamate transporters were chosen based on published sequences of the following transporters: rat EAAC1 (5'-GGT GTC GCT GCA CTG GAT TCC AAC G-3', 5'GGC CAT ATA AAG GCC CAA CTT GCG G-3'), rat GLT1 (5'-AGG AGC CAA AGC ACC GAA ACC-3', 5'-TCC AGG CCC TTC TTG ATA ACG-3'), rat GLAST1 (5'-TTG GAT TTG CCC TCC GAC CG-3', 5'-GGT GCA TAC CAC ATT ATC ACC GC-3'), and tiuman EAAT4 (5'-TGC GCC CAT ATC AGC TCA CCT Aces', 5'-TGC CCA GCC TCA TAA TAG CC-3"). Rat brain cDNA was generated from Poly(A)*-selected mRNA by reverse transcriptase reactions and then used as a positive control for all PCR reactions. Thermal cycling using Taq DNA polymerase included 25 cycles at 94°C for 1 min, 60°C for 1 .5 min, and 72°C for

PAGE 60

47 2 min. The Tag DNA polymerase was added after the heat denaturation at 94°C had started. All reactions contained 4 mM MgCb. The PGR products were visualized by ethidium bromide staining and photographed under UV light.

PAGE 61

48

PAGE 62

49

PAGE 63

50

PAGE 65

52

PAGE 66

53

PAGE 67

54 Figure 2-4 EAAC1 Immunoprecipitation Protein Concentration Curve (A) Different amounts of (10 to 500 ug) of solubilized total C6 membrane protein was subjected to immunoprecipitation with 5 ug of anti-EAAC1 antibody (AntiEAAC1 ) or pre-immune IgG (P. I.), as described in the text. After extensive washes, the precipitates were eluted in 40 yl of 2X SDB, run on SDS-PAGE, and then analyzed by immunoblotting using EAAC1 antibody, using the antibody conditions described in the text. (B) Densitometry analysis of EAAC1 protein from panel A. The densities of the EAAC1 mature monomer bands in each lane was quantified as described in the text and plotted against the initial protein amounts. The results represent a single experiment. . . \ . r

PAGE 68

55

PAGE 69

56 m9 l4 Temp(oC): 4 25 37 37 65 95 Time(miii): 30 30 10 30 10 10 ^ 145 kDa ^ 73 kDa Figure 2-5 Determination of the Elution Conditions for Immunoprecipitated EAAC1 A 500 ^ig aliquot of solubilized C6 membrane protein was subjected to immunoprecipitation with 5 \ig of anti-EAAC1, as described in the text. After the washing steps, the pellet was then divided into six aliquots, and each was subjected to a different elution condition (elution temperature and time) as labeled in the figure. The eluates were collected, resolved on SDS-PAGE, and analyzed by Western blotting. The results represent a single experiment. Heating at 37°C for 30 min was the elution condition used for the rest of the experiments described in this thesis.

PAGE 70

57

PAGE 71

58

PAGE 72

CHAPTERS ACTIVITY, PROTEIN CHARACTERISTICS, AND BIOSYNTHESIS OF EAAC1 GLUTAMATE/ASPARTATE TRANSPORTER IN C6 GLIOMA CELLS Introduction A family of Na^-dependent high affinity (SDHA) glutamate/aspartate transport systems, previously referred to collectively as System Xag", is essential for the glutamatergic transmission in the central nervous system (CNS), as well as for many other cellular functions. System Xag" activity can be determined as Na*-dependent and D-aspartate inhibitable L-glutamate or L-aspartate transport, and five distinct transporter cDNAs encoding this activity have been cloned in the past six years (refer to Chapter 1 for overview). Excitatory amino acid carrier 1 (EAAC1 ), a member of this SDHA glutamate transporter family, was first identified from rabbit intestine by Kanai and Hediger using oocyte expression cloning in 1992 (Kanai et al., 1992). The rat EAAC1 homolog, which was later cloned by Velaz-Faircloth and coworkers (1995), shares 95% similarity and 90% identity with the rabbit clone. Rat EAAC1 cDNA encoded a protein of 523 amino acids with a predicated non-giycosylated core molecular weight of 56.8 kDa. Sequence analysis predicted an integral membrane protein with ten putative transmembrane domains and intracellular Cand N-termini. There are six putative protein kinase C phosphorylation sites. A recent publication (Davis et al., 1998) indicated that EAAC1 transport activity as well as its cell surface 59

PAGE 73

60 expression might be regulated by protein phosphorylation, although direct evidence is needed to show the EAAC1 protein is actually phosphorylated. Nevertheless, another member of this SDHA glutamate transporter family, GLT1, was shown to be phosphorylated when the cells were treated with phorbol ester (Casado et al., 1993). EAAC1 also contains four potential N-linked glycosylation sites (Asn-X-Thr/Ser), three of which are localized within the second extracellular loop. It has been documented that EAAC1 protein is N-glycosylated under normal conditions in rat brain and several cell lines (Dowd et al., 1996). Different apparent molecular weights of EAAC1 protein were detected in rat brain, EAAC1 cRNA-injected oocytes, and C6 glioma cells, which show 68 kDa, 80 kDa, and 78 RDa bands, respectively (Dowd et al., 1996; Yang and Kilberg, unpublished data). But after the cleavage of the N-linked oligosaccharide chain(s), only the EAAC1 protein core was detected at ~ 57 kDa (Dowd et al., 1996). This result suggested that the N-glycosylation of EAAC1 might vary according to the tissue/cell type. The possible physiological relevance of this differential Nglycosylation of EAAC1 is not known. EAAC1 transporter is the most widely distributed transporter of the anionic amino acid transporter family (reviewed by Malandro and Kilberg, 1996). EAAC1 mRNA or protein have been localized in kidney, small intestine, heart, lung, skeletal muscle, liver, placenta, human fibroblasts, as well as brain. But within the CNS, EAAC1 is believed to be neuron-specific, whereas GLAST and GLT1 are glial-specific transporters as described in Chapter 1 . The highest densities of the neuronal transporter EAAC1 are in the cortex and the hippocampus areas

PAGE 74

61 that have high levels of giutamatergic synaptic transmission (Rothstein et al., *' 1994). Interestingly, these areas are also extremely sensitive to excitatoxic damage caused by stroke, ischemia, and head trauma (Palmer et al., 1993; Nilsson et al., 1996; Greene and Greenamyre, 1996). Therefore, it was postulated that the EAAC1 transporter might play a role in the protection of neurons in these areas from excess glutamate release during normal synaptic activity and pathological conditions. However, little is known about the individual contribution of these SDHA glutamate transporters in the regulation of synaptic excitatory neurotransmitter concentration, because of the lack of specific inhibitors for neurophysiology analysis. In recent years, attempts have been made to decipher the role of each SDHA glutamate transporters by generating mutant mice with individual transporters knocked out (Rothstein et al., 1996; Tanaka et al., 1997; Peghini et al., 1997). These knock-out studies suggested that the glial transporters, but not the neuronal one, play a bigger role in the prevention of excitatoxicity and resultant paralysis and seizures. Considerable effort has been made in cloning and characterization of amino acid transporters in the past decade, but a major aspect, that is, the biosynthesis, maturation, targeting, and degradation of these transporter proteins, is relatively unexplored. As we have repeatedly learned from the study of many other cellular macromolecules, how they are synthesized and degraded often impacts regulation of their cellular function. Therefore, understanding the life cycle of these amino acid transporter proteins would provide us with the basis for future functional studies in normal and diseased states.

PAGE 75

62 In recent years, a lot of progress has been made in generating the proper tools for the study of several amino acid transporters, including cDNA probes as well as specific antibodies. Our own laboratory successfully generated a polyclonal antibody against the EAAC1 transporter protein, which was raised against the C-terminal 120 amino acid peptide of rat EAAC1 fused with the maltose binding protein (EAAC1-MBP fusion protein). As I have shown in Chapter 2, this anti-EAAC1 antibody recognizes three distinct protein bands on the Western blots of the C6 glioma cells as well as BNL CL.2 cells transiently transfected with EAAC1 -containing plasmid. Pre-incubation of antibody with EAAC1-MBP fusion protein completely inhibited the detection of all these bands, while pre-incubation with MBP alone failed to compete. All of these observations indicate that this is a reliable antibody, which can be used to study the synthesis and degradation of the EAAC1 transporter protein. As I stated, the biosynthesis and intracellular trafficking of any individual mammalian amino acid transporter has not been documented. However, given that EAAC1 transporter is predicted to be a N-glycosylated integral membrane protein for which the primary functional site is the plasma membrane (Kanai and Hediger, 1992; Velaz-Faircloth et al., 1995), I hypothesize that the biosynthesis pathway follows the general biosynthetic pathway for plasma membrane glycoproteins. The biosynthesis pathway for integral membrane glycoproteins was reviewed in Chapter 1 and briefly summarized here. Biosynthesis of an integral membrane N-glycoprotein starts with the co-translational translocation of the nascent polypeptide into the rough ER, where it is co-translationally modified

PAGE 76

63 by N-glycosylation. Then oligosaccharide trimming and addition proceed further in the ER and Golgi compartment. Endogiycosidases are a class of enzymes that catalyze the cleavage of oligosaccharide chains at specific sugar residues and can be used for the characterization of glycoproteins. The most commonly used endogiycosidases are protein N-glycosidase F (PNGase F) and endoglycosidase H (Endo H). Whereas PNGase F cleaves between the first GlcNAC residue and the Asn on a N-glycosylated protein (Maley, 1 989), Endo H cleaves between the two GlcNAC residues adjacent to the Asn and is specific only to N-glycoproteins that are modified with high mannose and some hybrid type oligosaccharides (Robbins, 1984). Therefore, although the susceptibility to the cleavage by both enzymes suggests that the protein is modified by N-linked glycosylation, Endo H sensitivity can be used to further distinguish the type of the attached oligosaccharide chains. Endo H-resistance plus PNGase F-sensitivity, coupled with metabolic labeling as described below, is routinely used as a detection marker for N-glycosylated proteins that are modified by complex-type oligosaccharide chain(s), that is, they have already passed beyond the medial Golgi. Metabolic labeling of cells with radiolabeled amino acids, such as ^^S-Met or ^^S-Cys, is commonly used for the study of the biosynthesis, intracellular trafficking, and degradation of the proteins (Alberts et al., 1994). The radiolabeled amino acid(s) are incorporated into the newly synthesized proteins during the short labeling period, often called the "pulse" period. In order to study the maturation, processing, and degradation of this protein, the labeled protein is

PAGE 77

64 then "chased" by incubating the cells with an excess amount of non-radiolabeled amino acid for a certain period of time. De novo biosynthesis of the interested protein is then studied after immunoprecipitation with specific antibody. The susceptibilities of the radiolabeled protein to the digestion by either PNGase F and/ or Endo H is used to determine whether the newly synthesized protein is Nglycosylated as well as whether it has proceeded beyond the medial Golgi complex. Membrane proteins that have already arrived at the PM may share the same compositional and structural characteristics as those that have finished their post-translational modification but have not been, or will not be, transferred to the PM. Therefore, additional method(s) have to be designed to determine the PM arrival of the protein. The most commonly used techniques to detect the PM arrival of proteins include plasma membrane isolation and cell surface protein modification by membrane impermeable reagents. PM isolation could be done using differential or gradient centrifugation following pulse-chase labeling. The appearance of the radiolabeled protein in the collected PM fraction could be used to determine the time it took for the newly synthesized protein to be targeted to the PM. However, because it is literally impossible to get absolutely pure PM fractions using currently available membrane fractionation methods, modifying the PM proteins of intact cells with membrane impermeable chemicals has proven to be more effective and reliable approach. In this method, after pulsechase labeling, cell surface proteins are chemically modified by a membraneimpermeable reagent. The appearance of the chemically modified radiolabeled

PAGE 78

65 protein serves as a hallmark for the arrival of the newly synthesized protein at the PM. Many plasma membrane proteins, including the vitronectin receptor (Nesbitt and Norton, 1992), Thy-1 and lymphocyte surface marker proteins (Meier et al., 1992; Altin and Pagler, 1995), insulin receptor (Levy-Toledano et al., 1993), and tumor necrosis factor receptors (Hsu and Chao, 1993) have been successfully studied using this tool. Several membrane impermeable modifiers are now commercially available, with different functional groups to confer different chemical reactivities. Biotin-derivatives, by far, are the most commonly used modifiers. The tight binding between biotin and avidin can then be utilized to detect any proteins that have been modified (Gitlin et al., 1987). C6 cells are rat glial tumor cells, cloned 30 years ago (Benda et al., 1968), which have high Na*-dependent glutamate/aspartate transport activity. The experiments described in this chapter of my thesis were designed first to test C6 glioma cells as a model system for EAAC1 glutamate/aspartate transporter studies, and then to determine the subcellular localization and other characteristics of the EAAC1 protein in these cells. Finally, the biosynthesis, intracellular targeting, and maturation process of the newly synthesized EAAC1 protein also will be investigated. Results -jet Characterization of Xar,' Activity in C6 Glioma Cells To determine the effect of substrate deprivation, the Na*-dependent high affinity glutamate/aspartate transport activity (System Xag") in C6 cells, the cells

PAGE 79

66 were placed at a density of 1 x 10^ cells/ well into 24-well trays and cultured for 20 hr under normal conditions. Then the medium was replaced with either NaKRB containing 10% dialyzed fetal bovine serum (FBS) (for amino acidstarved cells, -AA) or complete MEM containing 10% dialyzed FBS (control) for the rest of the groups. The cells were incubated for another 6 hr at 37°C, before being subjected to transport assays for L-[^H]-aspartate in the absence (CholKRP) or presence (NaKRP) of sodium ion. The Na*-dependent transport activity was calculated by subtracting the transport measured in Chol-KRP which, for most of the experiments, was less than 3% of the total uptake activity (NaKRP). As shown in Figure 3-1 , the C6 cells incubated in MEM had high endogenous Na*-dependent aspartate/glutamate transport activity, which was almost completely inhibited by the presence of excess amount of D-aspartate as a specific inhibitor for System Xag' activity. These data suggested that the Na*dependent high-affinity glutamate/aspartate transporter family is the main carrier for the aspartate/glutamate transport activity present in C6 cells. It was previously reported that amino acid deprivation induced an increase in the Na*-dependent L-aspartate transport in renal epithelial (NBL-1) cells without a correspondent increase in either the mRNA (Plakidou-Dymock and McGivan, 1993) or the protein content (Nicholson and McGivan, 1996) of EAAC1 transporter. To study whether a similar effect would be seen in the C6 glioma cells, cells were treated with amino acid-deprived medium for 6 hr and the transport activity was compared to cells maintained in MEM. As shown in Figure 3-1 , amino acid starvation for 6 hr did not cause significant change in the Na*-

PAGE 80

67 dependent L-aspartate uptake in C6 cells. A lack of substrate regulation for EAAC1 in human fibroblasts has also been observed in our laboratory (BarbosaTessmann and Kilberg, unpublished results). ^ Table 3-1 lists the Na*-dependent L-aspartate transport activity measured by our laboratory or reported by others in a number of different cell lines (Nicholson and McGivan, 1996; Kilberg et al., unpublished data). As shown, C6 glioma cell possessed the highest Na*-dependent transport activity for Laspartate among the cell lines tested. Therefore, this cell line was used as our model system to study Na*-dependent high-affinity glutamate/aspartate transporter(s). Determinino Which Glutamate/Aspartate Transporter(S) Are Expressed in C6 Glioma Cells To determine which members of the Na*-dependent high affinity glutamate/aspartate transporter family are responsible for the high System Xag' activity detected in cultured C6 glioma cells, RT-PCR was performed using the primers specific to each of four glutamate/aspartate likely transporters; EAAT5 has not been found expressed outside of the eye. Four x 10^ C6 glioma cells were placed onto each 150 mm dish and cultured for three days under normal culture conditions. The cells were then harvested, total RNA and then Poly(A*) mRNA was purified, and RT-PCR reactions for GLT1 , GLAST, EAAC1 , and EAAT4 glutamate/aspartate transporters were conducted as desaibed in the Methods Chapter. As shown in the Figure 3-2, only the EAACImRNA, but not that for GLAST, GLT1, and EAAT4, was detected from these cells. These results

PAGE 81

68 suggested that EAAC1 transporter was the major, if not the only, SDHA glutamate/aspartate transporter endogenously expressed in the C6 glioma cells, and was responsible for the high Xag" activity observed in this cell line. These results are consistent with previous reports (Palo et al., 1996; Dowd et al., 1996). Effect of Cell Density/Growth on C6 Glioma Cell Differentiation. Xap,' Activity, and Expression of Glutamate/Asoartate Transporter Proteins C6 glioma cells are believed to be representative of the less-differentiated stem cells that develop into either astrocytes or oligodendrocytes (Parker et al., 1980). Expression of these cell type properties can be modified by a variety of cultured conditions, including high cell density, which leads to glial differentiation (Varon, 1978; Maltese and Voipe, 1979; Parker et al., 1980). Because the glial ^ fibrillary acidic protein (GFAP) is considered a specific marker protein for differentiated astrocytes (Varon, 1978), the amount of GFAP expressed in C6 cells was measured by immunoblotting under different cell density conditions, as described below. C6 cells (2.4 x 10^/well) were placed onto each well of 6-well trays and cultured for up to 12 days under normal conditions with the medium changed every 3 days, except for the cells that would be collected on that same day. As shown in Figure 3-3A, the expression of the GFAP protein is up regulated after C6 cells reach confluency. This result is consistent with the previous report suggesting that high cell density could lead to the glial-like differentiation of C6 cells (Varon, 1978; Maltese and VoIpe, 1979), but this experiment alone could not tell whether it was the increased cell density, the slower cell growth rate, or both that caused this glial differentiation. >

PAGE 82

69 Given that, under normal culture conditions, C6 glioma cells mainly, if not exclusively, express a neuronal glutamate/aspartate transporter (EAAC1) (Palos et al., 1996; Dowd et a!., 1996), it was interesting to investigate whether the type and amount of transporter expressed, would be changed while the cells were undergoing differentiation. Equivalent numbers of C6 cells (0.5 x 10^ cells/well) was placed into each well of the 24-well trays and cultured up to 12 days. Xag activity and the amount of EAAC1 and GLT1 proteins in these cells were measured. Protein content for GLAST1 and EAAT4 were below detection. As shown in Figure 3-3B, the Xag" activity of C6 cells was high at day 2 and 4, and was reduced at day 6, and then returned back to a certain extent from day 8 to day 10 followed by a drop on day 12. It is not clear why a two-phase response was seen for transport activity under the effect of cell density and growth. However, the idea that EAAC1 might be the major transporter protein in C6 cells was further supported by showing that the amount of EAAC1 protein expressed in these cells correlated very well with the Xag' uptake activity throughout the whole two-phase response. Also shown in Figure 3-3B, as the C6 cells became more glial-differentiated, the expression of the glial-specific GLT1 glutamate/aspartate transporter-like immunoreactivity increased. However, as clearly shown in this figure, the amount of the GLT1 transporter expressed did not correlate with the Xag' activity of the cells. Therefore, GLT1 does not appear to represent a major contributor to the amount of Xag" activity detected in C6 cells.

PAGE 83

70 Therefore, I have characterized the aspartate/glutamate transport properties of C6 glioma cells, and have determined they have high endogenous Na*-dependent high-affinity glutamate/aspartate transport activity mediated by EAAC1. These observations make this cell line an ideal model system for studying the biosynthesis, degradation, and regulation of the EAAC1 transporter protein. These results also demonstrated, for the first time, that the amount of the EAAC1 as well as the total Xag" activity of C6 cells are regulated by cell density and growth. The concentration of a protein at a given time is determined by the equilibrium between its biosynthesis rate and its degradation rate. A decrease in EAAC1 transporter protein concentration could be caused either by a reduction of protein synthesis or by an increase of protein degradation. As shown below, my studies of EAAC1 protein synthesis suggest that the transporter expression was reduced by a higher cell density (Figure 3-1 1 ), indicating that a reduction of de novo EAAC1 biosynthesis was partially, at least, responsible for the decrease of the steady-state amount of EAAC1 , and Xag" activity in C6 cells cultured to a higher density. Subcellular Localization of EAAC1 in C6 Glioma Cells As shown above, the neuronal EAAC1 transporter is the major SDHA glutamate transporter endogenously expressed in the C6 glial tumor cells. To determine whether EAAC1 transporter functions normally in C6 cells, it is important to establish the cellular localization of the transporter. Given that amino acid transporters are integral membrane proteins whose main functional site is

PAGE 84

71 the plasma membrane (PM), one would expect that the majority of the transporter protein should be localized at the PM with intracellular pools confined to the biosynthesis and degradation pathways. If, in C6 cells, the subcellular distribution of EAAC1 protein followed the above predication, the data would suggest that this neuronal transporter is not only expressed in the C6 glioma cells, but also targeted to the right place for its function. Three independent approaches were employed to determine the cell surface and intracellular localization of EAAC1 transporter protein in C6 glioma cells: (1) immunofluorescent-cytochemistry; (2) sucrose gradient fractionation of total cellular membranes followed by immunoblotting; (3) cell surface protein biotinylation followed by avidin-precipitation and immunoblotting. C6 glioma cells were stained with specific anti-EAAC1 antibody (1 :1000 dilution) and FITC-conjugated secondary antibody (1 :200) as described in the Methods chapter. As controls for subcellular organelle localization, a specific antibody against the manose-6-phosphate receptor was used as a Golgi marker, and 414 antibody was used as a nuclear envelope marker. 414 antibody was generated against an epitope shared by several proteins of the nuclear pore complex (Davis and Blobel, 1986). As shown in the Figure 3-4, EAAC1 staining analyzed by epifluorescence appears mainly to be PM-like (rather diffused staining throughout the cell with enrichment at the edges) and to a lesser extent, Golgi-like (paranuclear staining) with some intracellular vesicle-like staining (punctuated). The labeling by EAAC1 antibody could be completely inhibited by a

PAGE 85

72 pre-incubation of antibody with 25 ug/ml of EAAC1 -MBP fusion protein for 12 hr at 4°C (data not shown). Sucrose gradient fractionation of subcellular organelles was also utilized to determine the subcellular localization of EAAC1 in C6 glioma cells. C6 glioma cells were surface-biotinylated with membrane-impermeable sulfo-NHS-LCbiotin, and then a total cellular membrane fraction was collected and separated using sucrose gradient centrifugation as described in the Methods Chapter. The sucrose concentration and protein content was measured for each of the fourteen fractions collected. Equal amount of protein from each fraction was then separated on SDS-PAGE, and subjected to immunostaining with specific antibodies against EAAC1 and other marker proteins. Streptavidin-HRP was used to test for the distribution of biotinylated cell surface proteins, as described in the Methods Chapter. As shown in Figure 3-5, EAAC1 protein was present in fractions 6 to 9, and peaked in fraction 8 with a sucrose concentration of 36% or a density of 1 .16 g/cm^, a value that is consistent with the density of the PM (Griffith, 1986). EAAC1 protein co-migrated with NaVK* ATPase and with bulk biotinylated cell surface proteins in the sucrose gradient, both considered PM markers, but most of the EAAC1 did not migrate with GRP78, an ER marker protein. The broad distribution profile of GRP78 detected is consistent with the result reported previously, showing this protein was relatively "sticky" and present in most of the fractions (Kitzman, 1995). To determine the PM localization of EAAC1 transporter in C6 cells, the accessibility of EAAC1 protein from the extracellular space was detected by a

PAGE 86

73 cell surface biotinylation technique using membrane-impermeable sulfo-NHS-LCbiotin reagent. After biotinylation, total C6 cellular membrane proteins were collected, and total biotinylated proteins were isolated using avidin-precipitation and then separated by gel electrophoresis. The biotinylated EAAC1 protein was then detected using immunoblotting with specific anti-EAAC1 antibody. As shown in Figure 3-6, EAAC1 could be biotinylated, demonstrating that a significant portion of the EAAC1 protein was accessible from outside of the cells, that is, it resided at the PM. These results are consistent with those presented in Chapter 2 (Figure 2-6), demonstrating more than 60% of EAAC1 protein was specifically biotinylated by cell-surface biotinylation reaction, compared with only 2% or less of asparagine synthetase (AS) and GRP78. All of these data document that the majority of EAAC1 transporter protein was localized at plasma membrane with a small portion present within either the Golgi compartment, or some other intracellular vesicles. EAAC1 Protein Oliqomerization I have observed that specific antibodies against GLT1 , GLAST, and EAAC1 detect high molecular weight protein species, which probably represent the oligomeric forms of the glutamate transporter proteins (Haugeto et al., 1996). Additional experiments showed that reducing reagents (5% (5-ME, or 100 mM DTT), a denaturing reagent (6 M urea), or heating (100°C x 10 min) could not dissociate the oligomerization once it had been formed. Furthermore, the harsher the conditions used to treat the samples or the more handling of the samples, a

PAGE 87

74 greater proportion of the higher MW species was detected (see below). This observation is consistent with the results reported by Haugeto and coworkers showing that GLT1, GLAST, and EAAC1 transporter proteins form homomultimers (Haugeto et al., 1996). Using radiation inactivation analysis, they also suggested that the glutamate transporters might operate as homomultimeric complexes in vivo. The next set of experiments were designed to detemnine whether EAAC1 transporter protein from C6 cells forms oligomers in vitro, as well as to demonstrate further the effect of high temperature and solubilization conditions on the formation of the EAAC1 oligomers. Total C6 cellular membrane proteins were either mixed directly with SDB buffer, or solubilized with PES buffer for 1 hr at 4°C and then heated at different temperatures for specific lengths of time before mixing with SDB. All samples were then resolved on SDS-PAGE and subjected to immunoblotting analysis with anti-EAAC1 antibody. As shown in Figure 3-7, when untreated, the EAAC1 protein of C6 glioma cells was mainly present as the monomer form. However, after solubilization with PES buffer for 1 hr at 4°C, oligomehc EAAC1 was detected. Four different solubilization conditions were tested for their solubilization efficiencies as well as for their effects on EAAC1 olgimerization. These conditions included 0.5% Triton X-100, 1% Triton X-100, 1% Triton X-100 plus 4 M urea, or 2% C12E9 plus 0.1% SDS, all in PBS containing 3 mM EDTA. My results suggest that all four solubilization conditions extract EAAC1 protein from the cellular membrane with comparably high efficiency (> 95%). However, the harsher the solubilization conditions, less

PAGE 88

75 EAAC1 monomer and a greater amount of olgomer was detected by immunoblotting (data not shown). Also shown in Figure 3-7, as the heating temperature increased, the relative amount of EAAC1 monomer form of decreased. Although the corresponding increase of oligomers was not obvious in this latter case, the amount of newly formed oligomer may only represent a small portion of the total, or the maximum capacity of the nitrocellulose membrane transfer may have been reached for the higher MW bands. When membrane proteins were immediately mixed with gel sample dilution buffer and run on SDSPAGE without extra manipulation, most of the EAAC1 protein was detected as a single monomer band with a molecular weight of 73 ± 1 kDa (Figure 3-5). Therefore, in apparent conflict with the hypothesis of Haugeto et al. (1996), my data suggest that, in C6 cells, the formation of EAAC1 oligomers may be an artifact only caused by the manipulation of the protein sample, and may not represent functional structures inside the cell. Over-expression of EAAC1 protein in the BNL CL.2 mouse hepatocyte cell line, which has little or no endogenous glutamate transport activity (0 ± 0.4 pmol»mg'^ protein* min'^ L-aspartate uptake), was also utilized to test for the • various forms of EAAC1 protein. Transfection of BNL CL.2 hepatocytes with EAAC1 -containing expression plasmids significantly increased the Na*dependent aspartate transport activity, and this increase was accompanied by an induction of EAAC1 protein expression detected as both monomer (73 kDa) and oligomers (145 kDa and higher MW species) (Figure 3-8). Thus, formation of homo-oligomers occurs not only to endogenously expressed EAAC1 protein in

PAGE 89

76 C6 cells, but also to the overexpressed EAAC1 protein introduced by transfection of the mouse hepatocytes. Collectively, these results suggest that each of these three bands typically detected by immunoblotting are authentic EAAC1 . However, the present data suggest that the oligomerization of EAAC1 protein may not exist in vivo; the high MW forms may only be an artifact caused by sample handling. N-olvcosvlation of EAAC1 Protein ' The deduced amino acid sequence for the rat EAAC1 transporter contains four putative N-glycosylation consensus sequences, and three of these predicted N-glycosylation sites are localized within the second extracellular loop (VelazFaircloth et al., 1996). It has been reported that EAAC1 transporter is Nglycosylated, and its glycosylation pattern may vary in different host cell lines (Dowd et al., 1996). To test for the N-glycosylation of EAAC1 protein in C6 cells, protein N-glycosydiase F (PNGase F) digestion was used. As stated in the Introduction section, PNGase F cleaves between the first GlcNAC residue and the asparagine residue on N-glycoproteins (Maley, 1989). As shown in Figure 39, without PNGase F digestion both EAAC1 monomer (73 kDa) and oligomer bands were detected after immunoprecipitation of solubilized total membrane proteins with anti-EAACI antibody. The EAAC1 bands were broad, which may be due to the micro-heterogeneity of the glycosylation or other post-translational modifications including phosphorylation. However, a relatively sharp 57 kDa band (EAAC1 core) and a 1 14 kDa band (core dimer) were detected after incubation of

PAGE 90

77 the immunoprecipitated EAAC1 protein with PNGase F for 15 to 240 min. These results are consistent with the data previously published by Dowd et al. (1996) showing that EAAC1 protein is N-glycosylated in C6 glioma cells, but has a core MW of 57 kDa. However, these data, for the first time, showed that deglycosylation of the EAAC1 dimer does not prevent or reverse its oligomerization. N-glycosylation has been shown to be important for the functions of many plasma proteins, including amino acid transporters. It has been shown that the N-glycosylation of a glycine amino acid transporter (GLYT1 ) was important for its transport activity (Olivares et al., 1995). In contrast, abolishment of the Nglycosylation of the GLAST glutamate transporter protein using site-directed mutagenesis had no effect on its transport activity (Conradt et al., 1995). To determine whether N-glycosylation was important for the function of EAAC1 transporter in C6 glioma cells, tunicamycin was used as an inhibitor for the Nglycosylation reaction. Tunicamycin is a uracil-containing nucleoside antibiotic that specifically inhibits the first reaction of the synthesis of dolichol-linked oligosaccharides (Alberts et al., 1994). Incubation with 1 ng/ml tunicamycin for 8 to 36 hr decreased the aspartate transport activity of EAAC1 , and this decrease of transport activity correlated with the reduction of the amount of mature EAAC1 protein in treated C6 glioma cells (Figure 3-10). Furthermore, both effects appeared to be time-dependent up to 36 hr. It is interesting to notice that, as the content of the N-glycosylated mature form EAAC1 decreased, there was no detectable accumulation of core EAAC1 . This result may be due to the possibility that core EAAC1 is not as stable as the glycosylated form and is degraded much

PAGE 91

78 faster, as seen with other proteins (Meliklan, et al., 1996). These data suggested that the normal N-glycosyiation process was required for EAAC1 function in C6 glioma cells. Although it is tempting to postulate that the N-glycosylation of EAAC1 protein itself is important, we have to bear in mind that tunicamycin is a non-selective inhibitor for any proteins that may be N-glycosylated. Therefore, it is also possible that this observed tunicamycin effect on EAAC1 transporter and its activity might be an indirect response to the loss of N-glycosylation of other related protein(s). Further studies using site-directed mutagenesis of the EAAC1 cDNA must be employed in combination with transfection experiments to differentiate between these two possibilities. To explore the possibility that EAAC1 protein may associate with other protein(s) inside C6 glioma cells, protein cross-linking was used with a cleavable cross-linker, dithiobis (succinimidylpropionate) (DSP), followed by immunoprecipitation. C6 cells were first metabolically labeled with "^S-Met-CysH to label EAAC1 protein and other possibly associated proteins, and then treated with DSP to cross-link any proteins that were bound to or in close vicinity with the EAAC1 protein. After immunoprecipitation with anti-EAAC1 antibody, the precipitates were eluted in SDB containing 5% (BME as reducing reagent to separate potential cross-linked proteins. The samples were resolved on reducing SDS-PAGE, the gel dried under heat and vacuum, and analyzed with fluorography. Any proteins that were cross-linked with EAAC1 protein by DSP would be co-precipitated with EAAC1 , and would be revealed as extra bands on the film. As no extra bands were seen in the DSP-treated samples compared

PAGE 92

79 with controls (data not shown), it appeared that there were no major proteins associated with EAAC1 inside C6 glioma cells. This result is also consistent with my finding that, when washed with varied concentrations of salts ranging from the physiologic concentration of 150 mM to almost ten times more (1.2 M), no specific extra protein bands were seen in those samples washed with a low salt concentration after immunoprecipitation (Figure 2-2). Establish the De Novo Biosynthesis Rate for EAAC1 Transporter in C6 Glioma Cells To determine the biosynthesis rate of EAAC1 protein in C6 glioma cells, the cells were metabolically labeled with 200 uCi/ml of ^^S-Met-Cys for 15 to 120 min. After the labeling period, the cells were chased in medium containing an excess amount of non-radioactive Met and Cys for 3 hr, then the cells were harvested, total membrane proteins collected, newly synthesized EAAC1 protein was immunoprecipitated, and analyzed by gel electrophoresis followed by fluorography. As shown in Figure 3-1 1 A, as short as 15 min pulse-labeling time is enough for the detection of newly synthesized EAAC1 . All three forms (monomer, dimer, oligomer) of EAAC1 appear to be proportional to each other regardless of how long the pulse time. These results suggest that the oligomerization is not part of a maturation process for EAAC1 protein, which is consistent with my hypothesis that the formation of multimers may only be an artifact caused by sample handling. Interestingly, the biosynthesis rate of EAAC1 was decreased when a higher number of cells was plated, that is, EAAC1 biosynthesis is down-regulated by increased cell density (Figure 3-1 IB). This

PAGE 93

80 observation is consistent with my earlier studies showing the EAAC1 content is dramatically reduced as calls became more confluent (Figure 3-3). Collectively, these data suggest that higher cell density, or perhaps lower cell growth, decreases the amount of EAAC1 protein, at least partially, through downregulation of its de novo biosynthesis rate. Therefore, in all experiments a fixed number of cells were plated and then cultured for an exact period of time. Maturation and Taroetino of EAAC1 Transporter in C6 Glioma Cells To study the maturation process of EAAC1 , C6 glioma cells were pulselabeled with 500 uCi/ml of ^^S-Met-Cys for 1 5 min and then chased for 0 to 120 min. When chased for less than 30 min, only the low molecular weight (MW) (-57 60 kDa) immature forms of EAAC1 were detected (Figure 3-12). After longer chases, the lower MW form EAAC1 matured into the 73 kDa mature monomer form. In Figure 3-13, it appears that the transition of EAAC1 protein from the lower MW forms to the mature monomer form started after 45 min of chase time and finished after 190 min. Also illustrated in Figure 3-12 is the fact that there were multimer forms detected not only for the mature EAAC1 monomer, but also for the lower MW EAAC1 . This observation further supported my hypothesis that the multimer forms may not exist in vivo, because if they were functionally important, we would expect that they should happen more likely, even if not only, to the mature form of the EAAC1 protein. ; The next question to address was how long it would take for the newly synthesized EAAC1 protein to be targeted to its destination the PM. Pulse-

PAGE 94

81 Chase labeling of the C6 glioma cells with ^^S-Met-Cys followed by cell surface biotinylation was utilized to detect the time at which the radiolabeled (newly synthesized) EAAC1 protein would also be biotinylated (i.e., arrival at the PM). Cells were metaboiically-labeled with 200 uCi/ml of ^S-Met-Cys for 15 min and chased in medium containing an excess amount of non-radiolabeled Met and Cys for 0 to 360 min at 37°C. The pulse time was as short as possible (15 min) using an elevated amount of radioactivity to provide a narrow observation Vindow" for the PM arrival of the newly synthesized EAAC1 protein. At the end of each chase time point, the cells were surface-biotinylated with cell membraneimpermeable sulfo-NHS-LC-biotin to label all the proteins that were accessible from the extracellular space. After the PES-soluble proteins were collected from a total cellular membrane fraction, a double precipitation procedure was done by first precipitating all of the EAAC1 protein using anti-EAACI antibody, eluting the transporter from the sepharose beads, and then precipitating only the biotinylated EAAC1 protein using immobilized monomeric avidin-coated beads. Finally, using fluorography analysis, the newly synthesized EAAC1 protein molecules that had arrived at the PM (both ^^Sand biotin-labeled) as well as those that had not yet arrived at the PM (^^S-labeled and non-biotinylated) were detected. Newly synthesized EAAC1 transporter first arrived at the PM from 45 to 90 min (Figure 3-1 4A). The densitometry analysis of the data is plotted in Figure 3-1 4B and show that the targeting of the newly synthesized EAAC1 protein to the PM coincided with its maturation, that is, the transition of EAAC1 from the lower MW forms to the mature monomer.

PAGE 95

82 To Study the intracellular targeting of the newly synthesized EAAC1 protein, endoglycosidase digestions were employed. Endoglycosidase H cleaves between the two GlcNAC residues linked to the Asn residue in N-linked glycoproteins, only on high-mannose and some hybrid types of N-linked oligosaccharides. Therefore, resistance to Endo H digestion can be used as a hallmark for the N-glycoproteins that have already proceeded beyond the medial Golgi compartment. C6 glioma cells were pulse-labeled with 200 uCi/ml of ^SMet-CysH for 15 min and then chased in unlabeled medium for 30 to 240 min at 37°C. EAAC1 protein was then immunoprecipitated from a solubilized total cellular membrane fraction and subjected to Endo H digestion for 1 hr at 37°C. As shown in Figure 3-15, all mature EAAC1 monomers were resistant to Endo H digestion. Recall that the mature EAAC1 monomer form is sensitive to PNGase F digestion (Figure 3-9). This observation is consistent with the results of Figure 314, suggesting that the monomer form EAAC1 has already proceeded beyond the medial Golgi (Endo H-resistant) and has arrived at the PM (accessible to surface biotinylation). In contrast, the immature lower MW forms of EAAC1 protein were sensitive to the Endo H digestion, as shown in the Figure 3-15. These precursor EAAC1 polypeptides have been co-translationally modified by N-glycosylation, but they are still localized in the ER or cis Golgi compartment. Therefore, the oligosaccharide chains attached to them were those of the highmannose type, and thus, still sensitive to Endo H digestion. As to why only two major bands, instead of a broad smear of multiple bands, were observed for the lower MW forms of EAAC1 , we have no definite

PAGE 96

83 answer. Obviously, the antibody would not detect those nascent EAAC1 polypeptides that did not yet contain their C-terminal portion, that is, the last 120 amino acid residues. Judging from the sizes of those two bands after Endo H digestion, I could postulate that the upper one might be the full length EAAC1 protein with a predicted core size around 57 kDa; whereas the lower one might only be a translationally unfinished product, which had already possessed the Cterminal epitome specific for our antibody. Alternatively, one could postulate that the translation of EAAC1 protein may not be a continuous process, but one with distinct pause sites, because if it were a continuous process, a smear of many bands would be expected when the cells were pulse-labeled for a short time without chase. However, proof of this hypothesis will definitely need extensive investigation. Nevertheless, it is clear that both of these two immature forms of newly synthesized EAAC1 protein had been co-translationally modified by Nglycosylation. In summary, all of the data taken together suggest that the EAAC1 transporter is co-translationally modified by N-glycosylation, and that the transition of the EAAC1 protein from its immature forms to mature form coincides with the alteration of the attached oligosaccharide chains from the high-mannose type to the complex type. Importantly, these results also suggest that once the mature form is fully processed and achieved, it takes so little time as to be undetectable in the experiments, for the EAAC1 monomer to pass through the rest of the Golgi compartment and arrive at the PM.

PAGE 97

84 Discussion Prior to the present investigation, no mammalian amino acid transporter had been studied with regard to biosynthesis, maturation, and intracellular targeting processes, primarily due to the lack of proper tools. Recent advances in molecular cloning and protein expression have enabled us to obtain specific antibodies to some of the amino acid transporters, and these antibodies serve as valuable tools in addressing these questions. The polyclonal antibody generated by our laboratory against rat EAAC1 specifically recognizes three bands when used to probe C6 glioma cell proteins, all of which could be completely competed away by pre-incubating antibody with the EAAC1-MBP fusion protein (Figure 21 ). Consistent with reports by other laboratories who have observed similar results with independent antibodies (Haugeto et al., 1996; Davis et al., 1998), these multiple bands probably represent the mature EAAC1 monomer as well as its multimers, that is, dimer and trimer. However, as described below, the physiologic significance of these multimers is still a subject of debate. Our antiEAAC1 antibody has been used successfully to immunoprecipitate the EAAC1 protein from rat brain plasma membrane protein samples (data not shown), C6 glioma cells (Figure 2-4), EAACI-transfected BNL CL.2 cells (Figure 3-8), and Rcho rat placental cells (data not shown). EAAC1 transporter was chosen to investigate what I have termed "the life cycle" of a mammalian amino acid transporter, not only because of the availability of this specific antibody, but also because this transporter is widely distributed among tissues and represents a metabolically-important activity. For example, EAAC1 knock out mice exhibit

PAGE 98

85 , , dicarboxylic amino aciduria because of the critical requirement for EAAC1 in renal resorption (Peghini et al., 1997). To choose a proper cell line as a model system for studying the biosynthesis, maturation, targeting, and degradation of the EAAC1 glutamate transporter, several cell lines were tested for their endogenous System Xag' activity and EAAC1 protein content. As shown in Table 3-1 , C6 glioma cells had the highest Xag" activity among the cell lines tested. Therefore, they were the logical choice, regardless of the conflicting reports about which of the five glutamate/aspartate family members are expressed in C6 glioma cells (Casado et al., 1993; Dowd et al., 1996; Palos et al., 1996). My reverse transcriptase PGR results suggested that only EAAC1 mRNA was present when C6 glioma cells were cultured just to the point of confluency. Whereas the EAAC1 protein content decreased as C6 cells approached and passed confluency, GLT1 protein expression was significantly induced after confluency (Figure 3-3). As clearly shown in Figure 3-3, expression of the EAAC1 transporter parallels the observed changes in Xag" activity measured in C6 glioma cells. Therefore, the EAAC1 transporter in C6 glioma cells provided me with an ideal model system for the study of transporter biosynthesis, maturation, targeting, and degradation. To determine if the typically "neuronal" EAAC1 transporter, when expressed in C6 glioma cells, was localized at the PM, immunocytochemistry and sucrose fractionation techniques were used to determine its subcellular localization. Both approaches gave me consistent results documenting that EAAC1 resides primahly at the PM and, to a lesser extent, in the Golgi complex.

PAGE 99

86 Various nutrient transporters, including amino acid transporters, have been reported to oligomerize as detected by gel electrophoresis and immunoblotting. Haugeto and coworkers reported that GLAST1 , GLT1 , and EAAC1 transporters all formed homomultimers in rat brain and transfected Hela cells (Haugeto et al, 1996). Furthermore, based on results obtained from radiation inactivation analysis, they postulated that the glutamate transporters operate as homomultimeric complexes in vivo. Similarly, using glioma C6 proteins for immunoblotting, our anti-EAAC1 antibody detected three bands with estimated molecular weights of 73 kDa, 145 kDa, and 200+ kDa, which may represent the EAAC1 monomer, dimer, and trimer, respectively. However, in at least three independent experiments, only the EAAC1 monomer was detected when I collected a total membrane fraction from C6 glioma cells and directly resolved the proteins by gel electrophoresis without additional processing. This result was even true for membrane samples that were fractionated by sucrose gradient centrifugation (Figure 3-5). However, when harsher manipulation conditions were used for the membranes or solubilized proteins, I saw a shift of EAAC1 from its monomer to oligomer forms. In addition, when C6 glioma cells were pulse-chase labeled, the oligomer forms for the newly synthesized immature form EAAC1 (the lower MW forms) were also detected. No change in oligomerization was seen for the newly synthesized EAAC1 mature form as the pulse or chase time was prolonged. Although these results suggest that oligomerization happens before translation is finished, this process seems unlikely to me. If the oligomers were present in the cell and functionally required,

PAGE 100

87 one should never be able to isolate cell membrane containing only monomer, yet it can be. Therefore, it is postulated that these oligomer forms are artifacts caused by sample manipulation. This aggregation may occur because these integral membrane proteins are designed to be surrounded by a lipid bilayer, and have little tolerance for the conditions used to treat the samples. Formation of the multimers may be simply driven by the hydrophobic interactions when these proteins are extracted from the biological membrane into an aqueous surrounding, even with detergent present. However, there is a possibility that the transporter may exist as a homodimer, for example, that is converted to monomer after extraction of the cellular membrane proteins with reducing reagent, that is, p-mecaptoethanol. De novo biosynthesis of a membrane N-glycosylated protein starts from the translocation of the nascent polypeptide into the ER, where it is cotranslationally modified by N-glycosylation. As the protein proceeds through the ER and then Golgi compartments, the oligosaccaride chains attached may be further modified from high mannoseto complex-type. Therefore, when the chase is for a short time, only the immature form of the protein will be detected. Furthermore, these immature forms will be sensitive to endoglycosidase H digestion, only if they have not passed beyond the medial Golgi. This is exactly what I saw for the biosynthesis of EAAC1 transporter. When chased for less than 45 min, only the immature lower MW EAAC1 protein species were detected. These immature forms of EAAC1 were sensitive to Endo H cleavage, but the mature EAAC1 monomer detected after longer periods of chase was completely

PAGE 101

88 resistant to this enzyme digestion. In support of this observation, pulse-chase labeling followed by cell surface biotinylation showed that the arrival of EAAC1 transporter to the PM coincided with the shift of EAAC1 from its lower MW fonns to its mature monomer form. In another words, the maturation process of EAAC1 transporter overlapped with its arrival at the PM, there is no detectable lag time between these two processes. These data indicate that once an EAAC1 protein molecule finishes its maturation process in the Golgi, it is targeted to the PM immediately. Furthermore, the difference between the apparent molecular weights of the EAAC1 mature monomer and its immature forms is ~16 kDa, and no intermediates were seen during its maturation. These results suggest that during the maturation process the trimming and readdition steps for the Nglycosylation modification of each EAAC1 protein molecule is extremely fast. In summary, the EAAC1 biosynthesis and trafficking data presented in this chapter were the first to be documented for any mammalian amino acid transporter. It was shown for the EAAC1 transporter in C6 glioma cells that: (1) the biosynthesis rate needed to maintain steady state was relatively high, suggesting an equally rapid rate of decay; (2) the EAAC1 biosynthesis was regulated by cell density or growth; (3) the protein was co-translationally Nglycosylated with high mannose type of oligosaccaride chains; (4) the N-linked oligosacchande chains shifted from high mannose to complex type as the protein matured; (5) the time it took for this maturation was too fast to permit detection of intermediates; (6) the maturation process coincided with its PM arrival, that is, the time between the intracellular formation of the newly synthesized EAAC1

PAGE 102

89 monomer and its becoming accessible to cell surface biotinylation was too rapid to detect a large intracellular pool; and (7) little or none of the immature lower MW species was targeted to the PM in C6 glioma cells.

PAGE 104

92 (D C O c o Q 0) u. a *o E 03 i_ O) o 1 o E \ E I Q. CO (U (O 03 c 03 q: o a. Q) Q. E CO 03 T3 O > O CO c -t— « sz o 0) < Q o c 03 0) CO s Q) C CD O) CO 0) ~^ x: — o c c > o o > CO O Q. 03 CO 03 o (U CO Cfl 03 CD d O Q. CO c 03 -4— < CD ' 03 c 03 Q. CO TO a5 03 E TO r: UJ H CO < _l O H _i CD o o o CD Q. CO a: E c 03 a: c 03 0) Q. 03 O CO o O (D (D T3 CD O CO (D o O O O o c 03 CO CO CD Q. O 0) CD 0) C3) 0 CO O I 03 O) 03 > 5 ^ S CD < (D FT O (D S" CO CO 03 C g o 03 CD t_ 01 o CL (D o g"CD "5o c CD o c 0) Q. •o C o CD > CO C (D CO 0) k_ Q. CD ^ C£ 0) > CO "a3 O) CD 0) "O E o E g "g T3 0) o 0) 0) o c 0) E CD CL X CD

PAGE 105

93 Vi u U t: o a s « H a :< a O © s Q u & c CQ a S3 < < < «9 S a o o • • • • 00 0\ ^ fJ^* «H 1^ OS rf MO 0^ ^. rf
PAGE 106

94 Figure 3-3 Effect of Cell Density/Growth on the Xap/ Activity, EAAC1 Content, and Differentiation of C6 Glioma Cells An aliquot containing 0.5 x 10^ 2.4 x 10^ or 14.5 X 10^ C6 glioma cells was placed into each well of a 24-well tray, 6-well tray, or a 100 mm dish, respectively. The cells were cultured under normal conditions for 2 to 12 days and the medium changed every three days. The 24-well trays were used to measure the whole cell uptake of 5 pM L-[^H]aspartate at 37°C for 1 min (panel A), the 6-well trays were used to collect a total cellular protein extract for immunoblotting (GFAP, panel A; EAAC1 and GLT1, panel B) and protein quantification (panel A), and the 100 mm dishes were used to establish cell density (panel B). An excess amount of D-Asp (500 uM) was used as an inhibitor for the Na*-dependent aspartate transport activity. The remaining D-Asp-insensitive transport was trivial and remained unchanged over the course of the experiment. The values reported for the transport activity are expressed as the means + standard deviations (SD) for six determinations. Where not shown, the SD bars are contained within the symbols. The data shown in this figure were reproduced in two independent sets of experiments.

PAGE 107

90 Figure 3-1 Na^-dependent High Affinity Glutamate/Aspartate Transport Activity of C6 Glioma Cells C6 glioma cells (1 x 10^/ well) was placed into each well of 24-well trays and cultured for 20 hr under normal conditions. Six hours before the transport assay, the culture medium was replaced with either NaKRB containing 10% dialyzed FBS (the amino acid-starved groups, -AA) or with MEM containing 10% dialyzed FBS for the rest of the groups. The whole cell uptake of 5 \iM of L-[^H]-aspartate was measured for 30 seconds or 1 min at 37°C in the absence ("Control" and "-AA") or presence ("+ D-Asp") of 500 uM of D-aspartate as an inhibitor. Values are reported as the means ± standard deviations (SD) for quadruplicate determinations (n = 4). Where not shown, the SD bars are contained within the symbols.

PAGE 108

95 A. 2 4 6 8 10 12 B. I I 1 1 r Days of Culture

PAGE 109

96 Figure 3-4 Immunocvtochemistrv staining of C6 glioma cells with specific EAAC1, M6PR, and 414 antibodies C6 glioma cells were placed sparsely onto glass coverslips placed inside six-well trays, cultured under normal conditions until the cells reached 50 60% confluency, and then immunostained as described in the Methods Chapter. A 1:1000 dilution of anti-EAAC1, a 1:200 dilution of anti-M6PR (Golgi/late endosomal marker), and a 1:10 dilution of 414 (nuclear marker) antibody was used as primary antibodies, whereas a 1 :200 dilution of goat anti-rabbit IgG-FITC (for EAAC1 and M6PR) or goat-antimouse IgG-FITC (for 414) was used to detect primary antibody bound. The results shown are representative of at least 3 independent experiments. Panel A: Anti-EAACI; Panel B: Anti-414; Panel C: Anti-M6PR.

PAGE 110

97

PAGE 111

98 Figure 3-5 Sucrose gradient fractionation of total cellular membranes C6 glioma cells were surface-biotinylated with 0.5 mg/ml of sulfo-NHS-LCbiotin for 30 min at 4°C, and then a total cellular membrane fraction was collected as described in Methods Chapter. After resuspension in a 62% sucrose solution, the sample was carefully loaded with a long needle onto the bottom of a 1 5 45 % linear sucrose gradient and centrifuged at 35 k rpm for 20 hr at 4°C to separate the membrane organelles. A total of 14 fractions were collected by puncturing a hole through the bottom of the centrifuge tube. Each of the fractions was subjected to protein quantification, refractive index measurement, and Western blot analyses. Panel A: The immunoblot data showing the distnbution of NaVK* ATPase (plasma membrane), GRP78 (ER), EAAC1 (plasma membrane), and total biotinylated cell surface proteins. Panel B: Quantification of results, including densitometry analyses of the Western blot, refractive index of sucrose, and total protein content for each fraction. The results represent a single experiment.

PAGE 112

99 A. Na+/K+ ATPase (-95 kDa) B. I « I GRP78 (78 kDa) EAACl (73 kDa) Biotinylated Protins 60 St » • E C 40 .2 • s n a ti a o I"•5 20 o 2? = g 10 O 0. 12 £ g 6 S a o 0 ± s o o o a c e o w o a. 4 6 8 10 Fraction Numbers 12 14 Sucrose EAACl Na*/K* ATPas* GRP 78 Biotinylated Protein

PAGE 113

100 O O C o 3 O tn o "55 o CD E g a CD O O LU g TO >1 _C g X) o o CD I CO o 1 E in d O o t o (D O ^2 W o •D c CD O c Q) _D i*— C o o CD 03 c CO CD n CO CI T3 E il E c CD (/> =3 = , C 4— « g Id o c o CL CD E CD CO CD sz o o CD (D O D c CD i_ CD o Q) >, C g CD o >, CD O CD o CD xz LU o < Q. I CO Q CO c o o CD O CD E o c o o o Q) C Q) CO (D i_ Q. CD •4— « CD o CD SI \6> c » _g X) o c E E >. XJ "O 0) o CD CD T3 CO CD $ C o CL CO c CD E CD CL X (D C CD o C CD Q. CD o C o O -I I g o "O o LU o 0) >, g CD c CD c CD •o
PAGE 114

101

PAGE 115

102

PAGE 116

103 o o hi i u IT) IT) © 0\ 1-1 + VI o J ] 1 + 1^ O + O + o o a I o .a la "o S C/3 S S H

PAGE 117

104 Figure 3-8 Over-expression of EAAC1 transporter protein in BNL CL.2 mouse hepatocvtes BNL CL.2 fetal mouse hepatocytes were transfected with a vector containing the rat EAAC1 cDNA or vector alone, and the transfected cells were selected for two weeks by culturing in selective medium as described in the Methods Chapter. Panel A. Over-expression of EAAC1 increased the Na*-dependent L-aspartate transport activity of BNL CL.2 cells measured for either 1 or 2 min. Uptake of 5 L-[^H]-aspartate was measured in cells that were transfected with either PcDNA3 vector alone (Vector) or with EAAC1 cDNA insert (EAAC1 ). Values are expressed as means + SD of quadruplicate assays. Panel B. Over-expression of EAAC1 transporter protein in BNL CL.2 cells. For BNL CL.2 cells transfected with either vector alone (lanes 2 and 4) or with EAAC1 inserts (lanes 3 and 5) the total cellular membrane protein was subjected to immunoprecipitation with anti-EAAC1 antibody as described in the Methods Chapter. The samples in lanes 2 and 3 started the entire procedure with 1 mg of solubilized membrane protein, whereas the samples in lanes 4 and 5 were 2.4 mg of protein. Lane 1 is loaded with 10 [ig of PES-solubilized C6 glioma cell total membrane protein. The results represent a single experiment.

PAGE 118

105

PAGE 119

i 106 ;i I E O i: CD O 0) (0 (D g w o o _> a I c o Q O LU C o > (0 o o _> a I 0) Q cn I CO 0) l_ ZD O) Ll T3 O is C o < Q) -9 o w o CO CD CO LU Q. (D i_ Q) q. o — O E _g o > CT 03 C 03 o c 03 o D C 03 o 0) (D o CD C O CO 03 o o co CO c E o CN O in CD to 03 o Z Q_ O CD CO 03 O Z Q. + o B c CD 0) Q. 03 O CO D _ O £ "35 o LU _i_ c 03 I CO CO _>> 03 c 03 C3) C T3 CD O 0) n 3 CO "O c 03 LU O < Q. I (/) Q cyD c o c 3 CD C CD "D C CD CL CD D C o "to 03 0) CD B o o XI o c Z3 E E CD > 03 C CD CO CD Q. CD CD i CU CD Oj uj oi (D $ to Q. E 03 to CD sz XI 0) c CD CO CD i CL to CO CD CD SZ CO I"O O c 03 c CD E q5 X CD

PAGE 120

107 3 3 V* ! I I I + o o + o o in + ^

PAGE 121

108

PAGE 122

109 (o) (^.ujuj. urapjd ^.Biu. loiud) a)|B?dn dsv-Hg ;uapuadep-^BN () (l.0VV3PSJ!unAJBj;!qiV)''^B
PAGE 123

Ill A. i^hMiilll I||||Inmnii^ * 1*1 * § i 145 kDa Pulse: 15 30 60 120 (SxlO^ceUs) B. M 73 kDa 15 30 60 120 min ( 9 X 106 ceUs ) so 40 I 30 & .-8 20 c J 10 • 3x10* cells /dish 0 9x10^ cells /dish 30 60 90 Pulse-labeling Time (min) 120

PAGE 124

112

PAGE 125

113

PAGE 126

no Figure 3-11 De novo biosynthesis of EAAC1 transporter in C6 glioma cells Either 3 x 10^ or 9 x 10^ C6 glioma cells were placed onto 100 mm dishes and cultured for 20 hr before they were metabolically labeled with 200 yCi/ml of ^^S-Met-Cys for 15, 30, 60, or 120 min. After three hours of chase in "MEM + 10% FBS" medium containing 5 mM each of non-radiolabeled methionine and cysteine, a total cellular membrane fraction was PES-solubilized as described in the Methods Chapter. An equal amount of starting protein was subjected to immunoprecipitation with anti-EAACI and then analyzed with SDS-PAGE and fluorography(Panel A). Panel B illustrates the densitometry data obtained from the data in Panel A. The results represent a single experiment. .

PAGE 127

114

PAGE 128

115

PAGE 129

116 Figure 3-14 Trafficking of newly synthesized EAAC1 protein to the plasma membrane in C6 glioma cells C6 glioma cells were pulse-labeled with 200 ^Ci of ^^S-Met-Cys for 15 min and then chased in MEM medium containing 5 mM each of non-radioactive methionine and cysteine for 0 to 360 min. At the end of each chase period, the cells were surface-biotinylated with 0.5 mg/ml of suflo-NHS-CL-biotin in NaKRP buffer at 15°C for 1 hr. After PES solubilization of total cellular proteins, anti-EAACI antibody was utilized to precipitate all of the EAAC1 protein. Precipitated EAAC1 was eluted by incubation in low pH buffer and then immobilized monomeric ayidin-sepharose was used to precipitate the biotinylated EAAC1 (Biotinylated). The non-biotinylated EAAC1 protein (Non-biotinylated) was then collected from the supernatant by a third precipitation using anti-EAAC1 antibody. Finally, all samples were run on SDS-PAGE and then analyzed by fluorography and densitometry. All of these procedures were performed following the protocols described in the Methods Chapter. Panel A. The fluorographic results of a representative gel are shown. Panel B. The plots of the densitometry data obtained from Panel A, showing that the arrival of the newly synthesized EAAC1 protein to the plasma membrane coincides with the transition of newly synthesized EAAC1 protein from its lower MW form to the monomer form. The data reported here were reproducible in four independent experiments.

PAGE 130

117 A. Biotinylated Non-biotinylated 0 30 45 90 120 360 0 30 45 90 120 360 73kDa 57 kDa mm B. 30 s I « o '35 c 20 Q c (0 15 10 '£ 5 < 0 Newly synthesized mature EAAC1 PM arrival of EAAC1 120 180 240 Chase Time (min) 300 360

PAGE 131

118 CO D c CD a c c E CO O I 0) I CO in n o o o CN x: T3 0) Q) I CD Ui Q. 0) $ o CD E to J CD O JO 0) o o CO O o c CD 1CD > o CD g CD I c o c E lO to w CD O X CD c CD O) c 'c 3 c o o E CD E "O CD (O CD x: o CD O t CD Q. (0 CD "D O CD c:l CD CO CD sz o x: o CD (D O "O c CD CD x: c E o CN c g CD CD c CD D c CD c g CD LU C o Q. O CD Q. O C 3 CO c o Q. CD c CD i_ X! E CD E CD O o n CD N O CO I CO LU CL c ro lo X o o c LU 13 o o o o o CO x: •4— » D 0) » ro CD LU CD LU CO Q CO LO CN I 3 Xi O) c c ro c o c c CD o c ro k— CD it= 13 X3 GQ Q CO ro T CO ro 0) X E LO g o ro CD 0) c CD CD H O o CO ro c Q) o ro 0) CD CD > ro c CD CO CD k_ CL CD CD ro CD CD x: c o x: CO ro ro "O CD x: ro Oi o 1— o X3 "O 0) N ro c ro T3 c ro LU CD < CL I CO Q CO JO c CD E i_ CD C2. X CD C CD T3 C CD Q. CD "O c

PAGE 132

119

PAGE 133

120 Table 3-1 Na*-dependent ^H-Asp uptake activity in different cell lines Cell lines BNL CL.2 293 C1 8 Human fibroblasts Rcho cells BNL-1 Hela cells C6 Na"'-dependent L-Asp uptake (pmol>mq'^ protein* mi n'^^ 0 ± 0.4 27.4 ± 1.7 35.2 ± 3.6 48.0 ± 3.2 87.8 ± 0.9 50.2 ± 8.2 241 ± 18 Uptake of 5 uM of L-f^HJ-aspartate was measured for 1 min at 37°C in either Na*or choline-containing KRP buffer as described in the Methods Chapter. The values represent the means (n = 4) ± SD in pmoLmg"^ protein.min-\ BNL CL.2 cells are a fetal mouse hepatocyte cell line, 293 CI 8 cells are a human embryonic kidney cell line stably transfected with Epstein-Barr virus nuclear antigen cDNA, Rcho cells are isolated placental cells, BNL-1 cells are a kidney epithelial cell line, Hela cells are a human colon cancer cell line, and C6 cells are a rat glioma cell line.

PAGE 134

CHAPTER 4 DETERMINATION OF THE DEGRADATION RATE, AND THE PM RESIDENCE TIME FOR EAAC1 TRANSPORTER IN C6 GLIOMA CELLS Introduction Cellular proteins are not only synthesized from amino acids, but also degraded to amino acids continuously. The biochemical mechanisms involved in membrane protein degradation are not weW understood. In fact, no degradation pathway for a mammalian nutrient transporter has been documented fully. However, it is now realized that the regulation and the mechanisms involved in protein degradation are just as important in cellular homeostasis as those associated with biosynthesis. Furthermore, degradation has been identified as one of the primary control points for some biological processes by modulating the concentration of regulatory proteins as well as other regulated short-lived proteins (Alberts et al., 1994). As shown for the zwitterionic amino acid transport system. System A, transporters can be regulated by hormones, substrate availability, cell growth and density, as well as cell differentiation (Shotwell et al., 1983; Kilberg et al., 1985; Fong et al., 1989; Cariappa and Kilberg, 1990; Kilberg et al., 1993). Our laboratory has shown that the regulation of the System A transport activity involves regulation not only of the de novo synthesis, but also of the degradation 121

PAGE 135

122 Of the transporter protein (Handlogten and Kilberg, 1984). In theory, the balance between the rate of the transporter protein synthesis and degradation regulates the steady-state concentration of each amino acid transporter in a cell under normal conditions. Disturbance of this balance by altering either rate will lead to the net change of the transporter protein concentration. A number of approaches have been utilized to determine the rate of plasma membrane protein degradation. The most commonly used method is the pulse-chase labeling technique, which is composed of a radioactive labeling step and a "chase" step. The cells are first metabolically labeled for a short period of time to introduce radiolabeled amino acid tracers (for example, ^H-Leu, ^S-Met, 35 or S-Cys) into the newly synthesized proteins, and then the radiolabeled proteins will be chased into further processing and eventual turnover. Another approach is to label the PM proteins with a chemical reagent (for example, biotin) from the extracellular space and then chase to see the disappearance of the derivatized proteins. Immunoprecipitation techniques can be employed following both of these methods to determine the degradation rate for the specific protein of interest. The obvious difference between these two approaches lies in the two distinct groups of proteins are under examination. First, the pulse-chase-labeling method by itself determines the degradation rate for all newly synthesized membrane proteins, with no differentiation whatsoever of whether they had actually arnved at the PM or not. Whereas the chemical derivatization method studies the degradation rate for only these proteins that are residing at the PM

PAGE 136

123 during the period of chemical modification, regardless of how "old" they are, that is, long ago they were synthesized. Theoretically, plasma membrane purification could be used in conjunction with the pulse-chase labeling method to determine the decay of the newly synthesized proteins that are targeted to the PM, but there is no membrane isolation procedure available yet that yields absolutely pure PM. Protein modification by membrane impermeable reagents restricts the targeted subjects to those proteins already present in the membrane. The chemical derivatization method requires chemical modification of the membrane proteins, whereas the pulse-chase method only replaces non-radioactive atoms in the proteins with radioactive ones. Therefore, when using the protein derivatization method, we have to make the assumption that the chemical modification will not change the studied properties of the protein. By selecting a small molecule for chemical modification, one may be able to minimize possible artifacts that the dehvatization may cause. Therefore, when choosing the methods for studying the degradation of plasma membrane proteins, one has to consider both "PM localization" and "native structure" of the interested proteins. For the EAAC1 transporter degradation studies described in this chapter, both techniques were employed and compared. The breakdown of most proteins is believed to obey first-order kinetics, . which means that the newly synthesized proteins are just as likely to be degraded as the oldest proteins in the same pool, and the reaction rate is only dependent on and limited by the concentration of the proteins (Schimke, 1970; Doherty and Mayer, 1992; Olson et al., 1992). According to this assumption, the

PAGE 137

124 loss of the labeled proteins, either pulse-labeled or chemically modified, should be exponential. This is the observation for the disapperance of a substance with first-order kinetics and is expressed by the equation: d[R]o/dt = -k^R]i where [R]q is the initial radiolabeled protein concentration, [R]t the radiolabeled protein remaining after time t, and the kd is the fractional rate constant for degradation and has the units t'\ When the common logarithm of the remaining radiolabeled protein (logio[R]t) is plotted against time, a straight line is obtained, that is, logio [R]o = logio[R]t (kd/ 2.303) t The slope (x 2.303) of the straight line gives a first-order rate constant for the degradation (kd) of the protein. A convenient way of expressing the first-order rate constant is as a half-life. The half-life is the time taken for loss of half the protein molecules, that is, when [R]o/[R]i is equal to two, and the half-life (fi/2) equals the natural logarithm of two divided by (kd ) (0.693/ kd) (Schimke, 1970; Doherty and Mayer, 1 992). An alternative way of thinking of the degradation rate constant is to convert it into a percentage rate (kdX 100), that is, the percentage of a protein degraded per hour. As mentioned above, there is no plasma membrane transporter protein for which the degradation pathway is known. Several general mechanisms have been proposed for the degradation of plasma membrane proteins (Hare, 1990; Beaudoin and Grondin, 1991; Doherty and Mayer, 1992), including: (1) shedding or exfoliation, (2) lysosomal degradation, (3) ubiquitination-dependent

PAGE 138

125 proteasomal degradation, and (4) in situ degradation at the PM by specific proteases. By shedding, the plasma membrane proteins may be degraded outside of the cells after removal from the cell either by proteolytic release of ectoplasmic domains of the proteins (for example, insulin-like growth factorll/Mannose 6-phosphate receptor) (Scott and Baxter, 1996) or by release of the intact membrane proteins with phospholipid in the form of small membrane vesicles (for example, a-macroglobulin-trypsin receptor) (Kaplan and Keough, 1982). The precise mechanism by which membrane proteins can be selectively released from the cell surface is unclear. The second possible pathway for membrane protein degradation involves the membrane protein endocytosis and lysosomal degradation. Evidence for lysosomal involvement in plasma membrane protein degradation includes studies: (1) showing inhibition by a specific inhibitor of lysosomal proteases (Libby et al., 1980; Christopher and Morgan, 1981; Stoschek, 1984; Hare, 1988), (2) direct electron microscopy showing membrane proteins presented in the lysosomes (Beguinot et al., 1984; Ascoli, 1984; Beguinot et al., 1986; Hemery et al., 1996), as well as (3) biochemical localization of membrane protein degradation products in lysosomes (Russel and Mayer, 1983; Hare and Huston; 1985; Dunn et al., 1986; Draye et al., 1987). Ubiquitination-dependent proteasomal degradation pathway has been considered mainly for the degradation of rapidly regulated cytoplasmic proteins (Olson et al., 1992), but recent evidence suggests that it might also play a role in the elimination of some membrane proteins, especially abnormal membrane proteins (for example, mutant CFTR) (Ward etal., 1995; Hicke, 1997). This

PAGE 139

126 pathway starts with the ubiquitination of the condemned proteins and is followed by the selective degradation in the proteasome, a protein complex containing many proteases (Hicke, 1997). Recently, it has also been shown that, the ubiquitination, especially mono-ubiquitination, of a membrane protein can serve as a signal to trigger for its endocytosis and then decay (Terrell et a!., 1998). Finally, the degradation of some membrane proteins can start from the direct cleavage of the protein by a specific protease that is affiliated with the membrane (Doherty and Mayer, 1992). This part of my research was designed to use C6 glioma cells as a model system to: (1) establish the rate of degradation for newly synthesized EAAC1 proteins after the pulse-chase labeling, (2) establish the rate of decay for the PM resident EAAC1 protein after cell-surface biotinylation, (3) establish the residence time for newly synthesized EAAC1 at the PM, and (4) determine the possible pathway(s) for the EAAC1 protein degradation. Results Determination of the HalfLife and Degradation Rate for the EAAC1 T ransnnrtPr in C6 Glioma Cells Pulse-chase labeling of C6 glioma cells was used to establish the rate of degradation of the newly synthesized EAAC1 protein in C6 glioma cells. The cells were metabolically labeled with ^S-Met-Cys in Metand Cys-free medium for 1 hr and then chased in medium containing an excess amount of nonraiolabeled Met and Cys for 0 to 60 hr. As shown in Figure 4-1, the amount of

PAGE 140

127 radiolabeled EAAC1 protein diminished as the chase time was prolonged. From Figure 4-1 A, it is clear that, all three MW forms of EAAC1 are proportional to each other throughout all chasing time points. These results indicate that the decay rate for the EAAC1 monomer alone is the same as that for all three forms together. Therefore, in the rest of my studies, the densitometry readings for the monomer alone were used for the quantitative plots, unless stated otherwise in the legend. In Figure 4-1 B, the densitometry data from Figure 4-1 A were plotted against the chase time. It took about 14 hr for the decay of one-half of the radiolabeled EAAC1 molecules in C6 glioma cells. When I plotted the data using the common logarithm of the remaining radiolabeled EAAC1 against chase time, I surprisingly observed that there was a noticeable lag time for decay at the beginning of the chase (Figure 4-1 C). This lag time lasted about 8 hr and was observed in ail of my pulse-chase degradation experiments. This lag was further demonstrated in Figure 4-3B, in which more time points were chosen between 0 to 12 hr of chasing. Furthermore, what makes this result even more interesting is that, as presented in Chapter 3, it only takes the newly synthesized EAAC1 protein about 3 hr to arrive at the PM. This means that the newly synthesized EAAC1 protein was protected from the decay, not only during the de novo biosynthesis and targeting process, but also during the initial five hours after the protein reaches the final functional destination the PM. However, once the decay process started, it followed first-order kinetics, that is, expressing the logarithm of the remaining radiolabeled protein against time gives a straight line, as shown in all the degradation experiments presented in this study (Figure 4-

PAGE 141

128 1C). Processing the data presented in Figure 4-1 C gives us the first-order rate constant of degradation (Kd), which is equal to 0. 11 5 (hr'^ ). This means that EAAC1 protein was degraded at a rate of 1 1 .5% per hr, or about 6 hr for half of the total EAAC1 protein to be degraded. Plus, if one adds together the lag period of 8 hr, the total half-time for turnover of a newly synthesized EAAC1 protein would be about 14 hr. As far as I know, the lag time we observed for the EAAC1 ", ,-•_-•>. protein degradation has not yet been documented for any other membrane transporter protein. Degradation of the Cell Surface-Biotinvlated EAAC1 Protein in C6 Glioma Cells As stated in the Introduction, modification of plasma membrane proteins with membrane impermeable chemicals can also be used to study their half-life. The advantage of this chemical method compared with the pulse-chase radiolabeling is that only the PM resident proteins are accessible to the membrane-impermeable chemical modifier. Therefore, it can be used to specifically follow the decay of those proteins that have already arrived at the PM. Many of the chemical modifiers commercially available today are conjugated to biotin, a small molecule that has a tight affinity for avidin. Biotin has many advantages as a molecular tag; it is small and in many cases does not interfere with the biological activity of the modified proteins, and it can be readily detected by binding with avidin. In the previous experiments, I determined the half-life as well as the degradation rate for the newly synthesized EAAC1 protein. In the following studies, a membrane impermeable biotin derivative was used to

PAGE 142

129 determine the half-life of the EAAC1 transporter molecules that have already arrived at the PM. C6 plasma membrane proteins were first chemically modified with sulfo-NHS-LC-biotin by incubating the cells for a short penod of time at 15°C, and then chasing in normal MEM medium lacking serum for 0 to 12 hr at 37°C under normal culture conditions. The remaining biotinylated EAAC1 protein after each period of chasing was detected using precipitation with immobile monomeric avidin-sepharose followed by immunoblotting with specific antiEAAC1 antibody (shown in Figure 4-2). As shown by the logarithm plot, the halflife of the cell surface biotinylated EAAC1 protein was 6 hr with a degradation rate of 1 1 .5% per hr. A result consistent with the pulse-chase labeling described above. It is interesting to notice that there was no detectable lag time for the degradation of the biotinylated EAAC1. In contrast to the pulse-chase studies which determines how long it takes for the newly synthesized EAAC1 to be degraded, cell-surface biotinylation determines the turnover of the cell surface EAAC1 protein as a whole without discriminating with regard to the age of the protein. At any given time point, the percentage of the biotinylation-accessible EAAC1 protein that is still existing within the lag period, that is, from 3 hr (the PM trafficking time) to 8 hr (the decay starting time), may be relatively small. Therefore, the degradation lag period is undetectable when the degradation of the total cell surface EAAC1 protein was under examination, instead of the newly synthesized protein. Considering the results obtained by both approaches, it was clearly shown that, once started, the EAAC1 protein was degraded with a half-life of about 6 hr under normal conditions in C6 glioma cells.

PAGE 143

130 Residence Time of Newly Synthesized EAAC1 at the PM of C6 Glioma Cells From the experiments deschbed aboye, I learned the half-life as well as the degradation rate of the EAAC1 protein in C6 glioma cells. However, I still did not know how long the EAAC1 protein stays at the PM once it reaches there, and if there is a detectable intracellular pool for EAAC1 protein following its arrival at the PM. In Chapter 3, I described the time it took for the newly synthesized EAAC1 protein to arrive at the PM using pulse-chase labeling followed by cell surface biotinylation. The same approach will be used in this part of the study to determine the residence time of the newly synthesized EAAC1 protein at the PM, with the modification that a relatively longer period of chasing time will be used. Figure 4-3A shows the disappearance of the biotinylated radiolabeled EAAC1 protein in C6 cells, for which densitometry analysis is plotted in Figure 4-3B. Consistent with my earlier experiments, the lag of about 8 hr for the disappearance of the newly synthesized EAAC1 from the PM was clearly seen. The loss of the radiolabeled, biotinylated EAAC1 protein in this experiment represents the disappearance of the EAAC1 from the cell surface, which can be interpreted as either protein turnover or protein endocytosis. These data suggest that, if the EAAC1 protein were degraded through an endocytosis pathway, the endocytosis of the newly synthesized EAAC1 protein would not start until 8 hr after its biosynthesis. Similarly, if the EAAC1 protein were degraded in situ at the PM, the protease action would begin only after the protein is 8 hr old. Although, we could not exclude the possibility that more than one mechanism may be

PAGE 144

131 involved in EAAC1 protein degradation, my later studies with specific inhibitors show that the endocytosis-lysosomal pathway is at least partially, if not totally, responsible for the EAAC1 decay in C6 glioma cells. Once started, the PMresident EAAC1 protein disappeared from the cell surface at a rate of 1 1 .5% per hr, which is consistent with the total EAAC1 protein decay rate defined by the pulse-chase studies (Figure 4-3B). To confirm this observation and to show within the same expenment that the disappearance of the newly synthesized, PM-resident EAAC1 protein coincides with the degradation of the total newly synthesized EAAC1 protein in C6 cells, the degradation of newly synthesized (radiolabeled) EAAC1 as a whole and the disappearance of the cell surfacetargeted (radiolabeled & biotinylated) EAAC1 were determined and compared (Figure 4-4). From the densitometry data plotted in Figure 4-4B, it is clear that the disappearance of the newly synthesized EAAC1 from the cell surface correlated well with the degradation of the newly synthesized EAAC1 protein. This result indicates that there is little or no detectable intracellular pool of the endocytosed EAAC1 protein in C6 glioma cells after its PM-residence. However, I cannot exclude the possibility that there may be a small or rapidly equilibrated intracellular recycling pool of EAAC1 protein or that there may be a small, undetectable pool of endocytosed EAAC1 along the endosome-lysosome pathway before it is degraded. In combination with my inhibitor studies described below, these data suggest that, once endocytosed, it takes little time for the EAAC1 protein to be processed through the endocytic pathway and degraded, probably in the lysosomal or late endosomal compartment.

PAGE 145

132 Inhibition of EAAC1 Protein Degradation in C6 Glioma Cells To decipher which pathways are involved in the degradation of the EAAC1 protein in C6 glioma cells, various inhibitors that are specific for different protein degradation pathways were utilized in the following studies. First, to survey for the possible effective inhibitors for the EAAC1 protein degradation, a variety of reagents were tested. In this set of experiments, only one time point of chase (24 hr) after pulse-labeling was used. The remaining radiolabeled EAAC1 protein at the end of the 24 hr-chase was determined following immunoprecipitation, gel electrophoresis, and fluorographic analysis. As shown in Figure 4-5, 1 mM PMSF, an inhibitor for general cellular protein degradation, 20 mM NH4CL, a specific lysosomal pathway inhibitor, as well as 50 yM BFA and 50 uM NOC, inhibitors for cellular vesicle trafficking (Hunziker et al., 1992; Runnegar et al., 1997), all had some degree of inhibition. BFA, an antifungal metabolite, prevents the assembly of the coats of non-clathrin-coated vesicles and causes the fusion of pre-lysosomes with lysosomes, and TGN with late endosomes (Pelham, 1991; Klausner et al., 1992). Whereas, NOC causes the disassembly of microtubules and inhibits microtubule-based vesicle trafficking (Runnegar et al., 1997). By doing so, both BFA and NOC can disturb proper functioning of the endocytosis pathway. In contrast, 250 uM ALLN, a specific proteasomal inhibitor had no effect. ALLN, even at a lower concentration (50 100 uM), has been shown to inhibit the degradation of NF-kappaB (Schow and Joly, 1997) and T-cell antigen receptor alpha chains (Yu et al., 1997). Collectively, these results suggest that

PAGE 146

133 normal intracellular membrane vesicle trafficking as well as lysosomal protease function are required for EAAC1 degradation. To further study the mechanism for EAAC1 degradation, several individual inhibitors or inhibitor combinations were used in experiments that included four to six specific chase time periods to calculate and then compare the decay rates. Consistent with the data shown in Figure 4-5, 50 yM ALLN, a specific inhibitor for the proteasomal proteases (Schow and Joly, 1997; Yu et al., 1997), had no inhibitory effect on the decay rate of the EAAC1 protein (Figure 4-6). In contrast, the combination of 200 ug/ml leupeptin and 1 0 mM NH4CI, used together to test broadly for lysosomal involvement, had an inhibitory effect on the decay of the EAAC1 protein, reducing the decay rate from 1 1 .5% to 8.7% per hr (Figure 4-7). Leupeptin and NH4CI are both considered inhibitors for the lysosomal pathway. NH4CI, a weak base, can accumulate inside acidic compartments like lysosomes and late endosomes and increase the inner pH of these compartments. Maintenance of the low pH is important for the activity of the resident proteases as well as for vesicle trafficking. Therefore, disturbance of the inner acidic environment by NH4CI will lead to an inhibition in the protein degradation by the endocytosis/lysosomal pathway. Leupeptin, a small tri-peptide, strongly inhibits lysosomal serine and cysteine proteases. To dissect whether one or both of these inhibitors was necessary to have the inhibitory effect on EAAC1 protein degradation, the effect of leupeptin and NH4CL were tested separately. As shown in Figure 4-8, 500 yg/ml leupeptin alone could not inhibit the decay of EMC1 protein, although a much lower concentration of leupeptin was shown inhibiting

PAGE 147

134 the degradation of EGF receptor by rat hepatocytes (Dunn and Hubbard, 1984; Dunn et al., 1986). In contrast, 10 to 20 mM NH4CI inhibited the degradation of the EAAC1 protein in C6 glioma cells (Figure 4-9). Also, as shown in the table inside Figure 4-9, the inhibitory effect of the NH4CI seems to be concentrationdependent. All of these data suggest that the function of specific lysosomal serine and cysteine proteases per se may not be involved, but a normal endocytic/lysosomal pathway as a whole is important for the degradation of EAAC1. 1 ^ Discussion Pulse-chase labeling followed by immunoprecipitation was used to establish the half-life and degradation rate for the newly synthesized EAAC1 transporter protein in C6 glioma cells. Surprisingly, an obvious lag period prior to EAAC1 protein decay was detected and, as far as I know, such a lag has not been documented for any other membrane transporter protein. Cell-surface biotinylation and chase followed by avidin-precipitation and anti-EAAC1 immunoblotting was utilized to determine the half-life as well as decay rate for the plasma membrane resident EAAC1 protein. In contrast to the newly synthesized EAAC1 protein decay, there was no delay of degradation for the cell surface EAAC1 that was accessible for biotinylation. To further establish the residence time of newly synthesized EAAC1 protein at the PM, cell-surface biotinylation immediately after the pulse-chase labeling, followed by anti-EAACI immunoprecipitation was employed. When the rate at which the EAAC1 protein

PAGE 148

135 disappeared from the PM was compared with that of its degradation, it was clearly shown that the EAAC1 protein was degraded as soon as it was removed from the cell surface. This result suggests that there is little or no detectable pool of EAAC1 protein within the endocytic compartments. Several specific protease inhibitors were used to characterize the mechanisms by which the EAAC1 protein was degraded. The lysosomal pathway inhibitor, NH4CI, could effectively inhibit the degradation of EAAC1 protein in a dose-dependent manner. In contrast, when the specific proteasomal protease inhibitor, ALLN, was utilized, no inhibitory effect on the decay of the EAAC1 protein was observed. The concentration of any protein at any moment in a given cell is determined by the balance between its biosynthesis and degradation. Although it is understood that a change in transporter protein degradation is as important for the regulation of its function as biosynthesis, little is known about the mechanisms by which plasma membrane transporters are degraded. As far as I can establish, this work will be the first to document the possible mechanism for the decay of a mammalian amino acid transporter. When a pulse-chase labeling approach was used to study the degradation of the radiolabeled EAAC1 protein in C6 cells, it showed that half of the radiolabeled EAAC1 protein would be lost after about 14 hr of chasing. To my surprise, the decay process for the EAAC1 protein did not start as soon as it matured a dogma generally accepted in the protein degradation field (Schimke, 1970; Doherty and Mayer, 1992; Olson et al., 1992). In contrast, there was an obvious delay time between the finish of EAAC1 protein biosynthesis and the

PAGE 149

136 beginning of degradation. This lag time can be divided into two parts: the first three hours occurs while the protein is completing biosynthesis and trafficking; and the following five hours occurs after the protein has already arrived at the PM. As stated in the Introduction section, the theoretical inference from exponential loss of radiolabeled from proteins is that protein degradation happens randomly. A newly synthesized protein is just as likely to be degraded as a protein which has been synthesized some time ago (Schimke, 1970; Doherty and Mayer, 1992; Olson et al., 1992). There is no evidence that proteins show lifetime kinetics (that is, they 'wear out') for most of the protein degradation documented to date, with hemoglobin as one of the few possible exceptions (Doherty and Mayer, 1992). However, what we have described here for the EAAC1 protein degradation can be considered as a hybrid of "first-order" and "lifetime" kinetics. The newly synthesized EAAC1 protein somehow, as a small pool, was protected from the degradation for the first eight hr, but then merged with the rest of the EAAC1 protein pool and was degraded following first-order kinetics. I have no data to explain how these newly arrived EAAC1 protein molecules were protected from the proteolysis within this period of time, but postulate that the endocytosis of the EAAC1 protein may be a selective step during its degradation. Somehow, the newly arrived EAAC1 protein molecules are not susceptible to endocytosis until five hr after they reach the PM, even though they are facing the extracellular surface (biotinylated) and thus, possibly in a functional state. Then they would merge with the general pool of PM EAAC1 protein and start to be endocytosed and degraded following first-order kinetics

PAGE 150

137 with a rate of 1 1 .5% per hr. As to what makes these new arrivals distinguishable from the rest is unclear, but quite a few possibilities exist. These newly synthesized EAAC1 protein molecules could: (1) be sequestered within specific membrane domains that are inaccessible to the endocytosis; (2) lack a stmctural alteration or post-translational modification (for example, phosphorylation or mono-ubiquitination) which are required for the recognition by the adaptor protein for endocytosis; or (3) lack the proper interaction with another regulatory factor or protein for the recognition by the degradation machinery. In the future, it would be interesting to further distinguish between these possibilities as well as to determine whether the function of the EAAC1 protein in these two pools is the same. It has been shown for the human growth hormone receptor, that ubiquitination of the receptor is required as a signal for both its endocytosis and its degradation by the lysosomal pathway (Strous et al., 1996). More recently, Springeal and Andre (1998) reported that, in yeast, ubiquitination of the nutrient permeases is associated with their internalization and degradation. Furthermore, Terrell et al. (1998) published their studies indicating that the mono-ubiquitination of a G protein-coupled receptor can serve as an internalization signal for its endocytosis. . ^ i When the residence time of EAAC1 at the PM was compared with its degradation, I observed the same delay for the disappearance of EAAC1 from the PM as for its degradation. During its 4"^ to 8"^ hr delay, EAAC1 protein was present at the cell surface and accessible to cell surface biotinylation. Once , started, the disappearance of EAAC1 protein from the cell surface also follows

PAGE 151

138 the first-order kinetics, and the rate at which EAAC1 protein disappears from the PM is the same as its degradation rate. There are two possible explanations: (1) EAAC1 protein may be proteolytically cleaved at the PM, thus making the disapperance and the degradation single event; or (2) EAAC1 protein is endocytosed and degraded in the endocytic compartments, such as lysosomes. If the second possibility is true, as my later inhibitor studies suggested, it means that both the endocytosis and the degradation process for EAAC1 protein happen almost simultaneously. The differences observed in the amount of total radiolabeled EAAC1 and the biotinylated-EAACI (as shown in Figure 4-3 & 4-4) may be contributed either to the efficiency of the cell surface biotinylation and the avidin-precipitation procedure, or to the fact that there may be a intracellular pool of mature EAAC1 which are not targeted to the PM. Specific inhibitors have been widely used in studying the mechanisms of protein degradation. It has been shown for the mutant CFTR that the ubiquitination-proteasome pathway was responsible for the degradation of the mutant protein in the ER (Ward et al., 1995). To determine whether the degradation of the EAAC1 protein is dependent on the proteasomal pathway, ALLN, a specific inhibitor for preasomal proteases (Schow and Joly, 1997; Yu et a!., 1997), was used and no inhibitory effect was seen suggesting that the polyubiquitination/proteasome pathway is not involved in EAAC1 degradation. When presented in their neutral state, weak bases such as NH4CI and chloroquine, can rapidly diffuse into any acidic compartment and accumulate there after becoming charged. This accumulation will lead to the neutralization of these compartments

PAGE 152

139 (Steer and Hanover, 1991 ; Pastan and Willingham, 1985). The low inner pH of the lysosomal and other endocytic compartments is important for the proper function of many resident proteases and for vesicle transport after endocytosis. Therefore, weak bases can be used as inhibitors to determine if the degradation of one protein is dependent on the endocytosis-lysosomal pathway. NH4CI was chosen in my studies, because chloroquine (200 400 \iM) caused the C6 glioma cells to detach from the cultured dish. An inhibitory effect of NH4CI on EAAC1 protein degradation was observed, and it was concentration-dependent. These data indicate that endocytosis/lysosomal pathway is, at least partially, responsible for the decay of the EAAC1 transporter in C6 glioma cells. However, my data will not tell exactly where along the endocytosislysosomal pathway, that is, early endosomes, late endosomes, or lysosomes, EAAC1 protein is degraded. Lysosomes are believed to be the main site of degradation in the endocytic pathway (Doherty and Mayer, 1992; Alberts et al., 1994), although endosomal proteolysis has been reported for a number of internalized proteins (Authier et al., 1995 & 1996; Tjelle et al., 1996). As stated above, I could not detect an accumulation of endocytosed EAAC1 after its PM residence. It is possible that EAAC1 might be degraded, or at least the initial cleavages including the chopping off the epitope for the antibody at the Cterminus, within the endosomes as soon as it is internalized. In summary, the EAAC1 degradation data presented in this Chapter are the first to document a degradation study for a mammalian amino acid transporter, or any other solute transporter as far as I am aware. It was shown

PAGE 153

140 that: (1) the newly synthesized EAAC1 protein is protected from the degradation during the first three hr of the biosynthetic trafficking process, as well as during the first five hr of residency at the PM; (2) once the decay of the newly synthesized EAAC1 protein starts, it follows first-order kinetics with a degradation rate of about 1 1 .5% per hr yielding a half-life of about 6 hr; (3) the EAAC1 protein is, at least partially, degraded through the endocytosis-lysosomal protein degradation pathway; and (4) the dispearance of the PM resident EAAC1 protein coincides with the decay of the total cellular EAAC1 , suggesting that there is no detectable accumulation of EAAC1 protein along the endocytic pathway, that is, once it is endocytosed, it is degraded immediately.

PAGE 154

141 Figure 4-1 Determination of the EAAC1 protein degradation using pulsechase labeling in C6 glioma cells (Panel A) C6 glioma cells were pulselabeled with 200 yCi/ml ^^S-Met-Cys in Metand Cys-free medium for 1 hr and then chased in MEM medium containing an excess amount of nonradiolabeled Met and Cys for specific times over 0 to 60 hr. At the end of each chase period, EAAC1 protein was immunoprecipitated, resolved on SDSPAGE, and analyzed with fluorography as described in the Methods Chapter. The data shown here are representative of at least four independent experiments. (Panel B) The plot of the arbitrary densitometry units of the remaining radiolabeled EAAC1 monomer (73 kDa) against the chase time. The densitometric analysis was performed in the linear range of the film. (Panel C) The plot of the logarithm of the densitometry reading of the radioactivity remaining as the EAAC1 protein monomer against the chase time yields an estimate of the half-life by the equation: Ti/2= 0.693/kd.

PAGE 155

142 Chase: 0 6 12 18 24 36 48 60 hr 0 10 20 30 40 50 60 Chase Time (hr) 0 8 16 24 32 40 48 Chase Time (hr)

PAGE 156

143 Figure 4-2 Degradation rate of the cell surface-biotinvlated EAAC1 protein C6 glioma cells were surface-biotinylated with 0.5 mg/ml of membrane impermeable sulfo-NHS-LC-biotin in NaKRP for 1 hr at 1 5°C and then quenched with 50 mM glycine in NaKRP for 15 min at 15°C. After they were transferred to fresh MEM medium containing 1% FBS, the cells were then incubated at 37°C for 0 to 12 hr. At the end of each incubation period, PESsolublilized total membrane proteins were prepared and the remaining biotinylated proteins were precipitated with avidin-sepharose beads. The precipitates were separated on SDS-PAGE, transferred onto a nitrocellulose membrane, and EAAC1 protein detected using immunoblotting. The data shown in panel A are representative of at least two independent experiments. In Panel B, the plot of the densitometry reading of the remaining biotinylated EAAC1 against the chase time is shown.

PAGE 157

144 Degradation of Cell-surface Biotinylated EAACl Chase: < 73kDa 0 8 12 hr B. 10 t5 43 c 3 -T rHalf-life = -6hr Decay rate = ~ 12% -hr-' -1 L. -1 L_ 0 2 4 6 8 10 12 14 Chase Time (hr)

PAGE 158

146 Biotinylated Non-biotinylated 1 r IMI iMJ tf"^ »^ * • ^ 145 kDa IPW ' im ^ »«4 «»f « ' , ^ 73 kDa 0 2 4 6 8 10 12 22 0 2 4 6 8 10 12 22 Chase Time (hr) B. ^ E « c o •a 'E 3 S s (0 D) O 4 8 12 16 Chase Time (hr) 20

PAGE 159

147 Figure 4-4 Comparison of the degradation rate of total EAAC1 protein with the rate of disappearance of cell surface EAAC1 in C6 glioma cells Cells were pulse-chase labeled with 200 uCi/ml ^^S-Met-Cys in Metand Cys-free medium for 30 min and then chased in MEM medium containing 1% FBS and an excess amount of non-radiolabeled Met and Cys for 0 to 36 hr. At the end of each chase period, the cells were surface biotinylated with 0.5 mg/ml sulfoNHS-LC-biotin in NaKRP for 1 hr at 15°C and then quenched for free-biotin with 50 mM glycine in NaKRP for 2 x 15 min at 15°C. As shown in Panel A, after the PES-solublilized total membrane proteins were prepared, total EAAC1 protein was immunoprecipitated with anti-EAAC1 antibody (Total), and then monomeric avidin-sepharose beads were used to precipitate only the biotinylated EAAC1 protein (Biotinylated). The plot of the logarithm of the densitometry reading of the remaining EAAC1 monomer against the chase time is shown in Panel 8. The results shown here were reproducible in at least two independent experiments. The parallel relationship of these two lines can be interpreted to show that the rate of total EAAC1 protein decay was the same as that for the disappearance of EAAC1 from the PM.

PAGE 160

148

PAGE 161

149 Figure 4-5Effect of specific inhibitors on the degradation of EAAC1 protein in C6 glioma cells C6 glioma cells were pulse-labeled with 200 yCi/ml of ^SMet-Cys for Ihr and chased in medium (MEM + 1% FBS + 5 mM Met + 5 mM Cys) with or without specific inhibitors for 24 hr. The PES-solublilized total cellular membrane proteins were prepared and subjected to immunoprecipitation with anti-EAAC1 antibody. After elution, the precipitates were resolved on SDS-PAGE, and then analyzed with fluorography and densitometry. All data are presented as percentages of a control lacking inhibitor. Control (without inhibitor); 1 mM PMSF; 1 mM PMSF + 20 mM NH4CI; 20 mM NH4CI; 50 uM BFA; 50 uM NOC; 250 uM ALLN. The results represent a single experiment.

PAGE 162

150 Control PMSF PMSF + NH4CI NH4C1 BFA NOC ALLN

PAGE 163

151 Figure 4-6 Effect of ALLN on the degradation of EAAC1 in C6 glioma Cells Cells were pulse-labeled with 200 uCi/ml of ^^S-Met-Cys for 1hr and chased in medium (MEM + 1 % FBS + 5 mM Met + 5 mM Cys) with or without 50 uM ALLN for 0 to 36 hr. At the end of each chase period, the PES-solublilized total cellular membrane proteins were collected and subjected to immunoprecipitation with anti-EAAC1 antibody, as described in the Methods Chapter. The precipitates were eluted and resolved on SDS-PAGE before being analyzed by fluorography and densitometry (Panel A). The plots of the densitometry data were expressed as the logarithm of the remaining radioactivity of EAAC1 protein monomer against the chase time (Panel B). The results represent a single experiment. The parallel nature of these two lines demonstrates no change of the EAAC1 decay rate with or without ALLN.

PAGE 164

152 Control A. + 50 ]iM ALLN 145kDa> i»4#^f 73kDa > 57 kDa Chase: 0 6 12 18 24 36 0 6 12 18 24 36 B. Decay rate = 11.5% • hr'^ Half-life = 6 hr Chase Time (hr)

PAGE 165

153 Figure 4-7 Effect of 200 ua/ml leupeptin + 10 mM NH^CI on the degradation of EAAC1 in C6 glioma cells Cells were pulse-labeled with 200 yCi/ml of ^^SMet-Cys for 1 hr and chased in medium (MEM + 1 % FBS + 5 mM Met + 5 mM Cys) with or without 200 ug/ml leupeptin + 10 mM NH4CI for 6 to 36 hr. At the end of each chase period, the PES-solublilized total cellular membrane proteins were collected and subjected to immunoprecipitation with anti-EAACI antibody, as described in the Methods Chapter. The precipitates were eluted and resolved on SDS-PAGE before being analyzed by fluorography and densitometry (Panel A). The plots of the densitometry data were expressed as the logarithm of the remaining radioactivity of EAAC1 protein monomer against the chase time (Panel B). The results represent a single experiment. The altered slope of the data with the inhibitors demonstrated an inhibitory effect of the combination of these two compounds on EAAC1 protein degradation.

PAGE 166

154 145kDa^ I 4 i. -4 * Chase: 6 12 24 36 6 12 24 36 hr (Control) ( + Inhibitors) B. 8 12 16 20 24 Chase Time (hr)

PAGE 167

155 Figure 4-8 Effect of leupeptin on the degradation of EAAC1 in C6 glioma Cells Cells were pulse-labeled with 200 uCi/ml of ^^S-Met-Cys for 1 hr and chased in medium (MEM + 1 % FBS + 5 mM Met + 5 mM Cys) with or without 500 ug/ml of leupeptin for 2 to 24 hr. At the end of each chase period, the PES-solublilized total cellular membrane proteins were collected and subjected to immunoprecipitation with anti-EAAC1 antibody (Panel A), as described in the Methods Chapter. The precipitates were eluted and resolved on SDS-PAGE before analysis by fluorography and densitometry. The plots of the densitometry data were expressed as the logarithm of the remaining radioactivity of EAAC1 protein monomer against the chase time (Panel B). The results represent a single experiment. That these two lines are parallel demonstrates no change of the EAAC1 decay rate with or without leupeptin.

PAGE 168

156 73kDa>|„n||aia|
PAGE 169

145 Figure 4-3 Determining the plasma membrane residence time of newly synthesized EAAC1 protein C6 glioma cells were pulse-chase labeled with 200 uCi/ml ^^S-Met-Cys in Metand Cys-free medium for 30 min and then chased in MEM medium containing 1 % FBS and an excess amount of nonradiolabeled Met and Cys for 0 to 22 hr. At the end of each chase period, the cells were surface biotinylated with 0.5 mg/ml sulfo-NHS-LC-biotin in NaKRP for 1 hr at 15°C and then quenched for free-biotin with 50 mM glycine in NaKRP for 2 X 15 min at 15°C. After the PES-solublilized total membrane proteins were prepared, total EAAC1 protein was first immunoprecipitated with anti-EAAC1 antibody. Monomeric ayidin-sepharose beads were used to precipitate only the biotinylated EAAC1 protein (Panel A). The non-biotinylated EAAC1 protein was immunoprecipitated from the supernatant of the ayidin precipitation through another incubation with anti-EAACI antibody. The logarithm of the densitometry reading of the remaining radiolabed EAAC1 monomer was plotted against the chase time (Panel B). The 8 hr delay of the disappearance of the radiolabeled EAAC1 from the PM is shown in this plot. The data presented here are representative of at least two individual experiments.

PAGE 170

157 Figure 4-9 Effect of 20 mM NH^CI on the degradation of EAAC1 C6 glioma cells were pulse-labeled with 200 yCi/mi of ^^S-Met-Cys for 1 hr and chased in medium (MEM + 1 % FBS + 5 mM Met + 5 mM Cys) with or without 20 mM NH4CI for 12 to 36 hr. At the end of each chase penod, the PES-solublilized total cellular membrane proteins were collected and subjected to immunoprecipitation with anti-EAAC1 antibody, as described in the Methods Chapter. The precipitates were eluted and resolved on SDS-PAGE before analysis by fluorography and densitometry (Panel A). The plots of the densitometry data were expressed as the logarithm of the remaining radioactivity of EAAC1 protein monomer against the chase time. The results represent a single experiment. The reduced slope of the inhibitor line illustrates an inhibitory effect of the 20 mM NH4CI on the EAAC1 protein degradation.

PAGE 171

158

PAGE 172

CHAPTERS CONCLUSIONS AND FUTURE DIRECTIONS Conclusions / Over the past three decades, amino acid transport systems have been extensively studied with regard to their kinetics, substrate specificity, tissue distribution, and regulation, but literally nothing is known about what I have termed the "life cycle" of these transporter proteins. That is, their synthesis, maturation or processing, trafficking to the PM, residence time at the PM, and degradation. In recent years, significant progress has been made in isolating and cloning the amino acid transporter proteins. This study and others of its kind are beginning to make use of the newly available cDNAs and sequence-specific antibodies as tools to investigate the cell biology of the amino acid transporters. By doing so, we definitely get insightful knowledge as to exactly how these transporter proteins are synthesized and degraded in the cells, how they are regulated under different conditions, as well as their roles in both normal and diseased states. What makes this research particularly important is the striking fact that, to date, no study on the biosynthesis of any mammalian amino acid transporter or on the degradation mechanism for any mammalian organic solute transporter has been published. Therefore, my research has not only provided us 159

PAGE 173

160 with new information concerning the biosynthesis and degradation of the individual transporter studied, more importantly, will also serve as a model system for understanding these processes for other nutrient transporters in mammalian cells. In this study, we chose the EAAC1 transporter, a member of the SDHA glutamate/aspartate transport family, as our transporter for study. This choice was mainly due to the fact that EAAC1 is the most ubiquitously distributed transporter within this family, and likely plays an essential role in a variety of cellular functions. In support of this belief, an EAAC1 knockout mouse exhibits amino aciduria and behavior abnormalities (Peghini et al., 1997). These effects presumably result from the importance of EAAC1 in renal resorption, EAAC1 is highly expressed in the kidney (Kanai and Hediger, 1992), and it localized in the central nervous system to neurons (Kanai and Hediger, 1992; Rothstein et al., 1994). Our laboratory has generated a polyclonal antibody specifically against the C-terminus of the EAAC1 transporter, which provides us with a necessary tool for the study of the biosynthesis and degradation of this transporter protein. As to which cell type would be used as my model system, I took the "trial and error" approach of measuring the transport rate and EAAC1 protein content to decide. In most cases, low abundance of the transporter protein of choice would add technical difficulties to the study. Therefore, my first challenge was try to find a cultured cell line that had relatively high SDHA glutamate/aspartate transport activity and an easily detectable amount of EAAC1 transporter protein. After testing many cell lines, I picked the C6 glioma cell line because of its high Na*-

PAGE 174

161 dependent glutamate transport activity (Table 3-1) and selective expression of EAAC1 (Figure 3-2). In the literature, there have been conflicting reports as to which members of the glutamate transporter family are expressed in C6 cells (Casado et al., 1993; Palos et al., 1996; Dowd et al., 1996). Therefore, I decided to test my C6 cell line (obtained from ATCC) by reversed transcriptase and PGR reactions using primers specific for each of the glutamate/aspartate transporters and observed that only EAAC1 mRNA, not GLAST, GLT1, or EAAT4, was detected in C6 cells. Also, when I followed the changes of transport activity and the expression of specific transporter proteins with regard to cell density/growth (Figure 3-3), it was clearly shown that EAAC1 is definitely the major, if not the only, transporter responsible for the Xag" transport activity observed in C6 cells. Furthermore, the results obtained from the immunocytochemistry (Figure 3-4), sucrose fractionation (Figure 3-5), cell-surface biotinylation (Figure 2-6 & 3-6), and PNGase F digestion (Figure 3-9) experiments all suggest that EAAC1 transporter in the C6 glioma cells is synthesized, transferred to the PM, and functions as one would expect. Therefore, I concluded that C6 glioma cell line could be used as a model system for the study of the life cycle of the EAAC1 transporter protein. The expression of a neuronal-specific {in vivo) transporter in a glial tumor cell line is surprising but not unprecedented. It has been shown that some types of glial tumor cells express glucose transporter subtypes that are not usually present in normal glial cells (Boado et al., 1994). Also, although not glial in origin, there are data showing that certain PC 12 pheochromocytoma cells are more

PAGE 175

162 sensitive to a toxic aspartate analog than the wild-type PC12 cells, suggesting an up-regulation of the aspartate transport activity in tumor cells (Ramachandran et al., 1993). Therefore, beyond their usefulness in the laboratory, understanding the atypical expression of EAAC1 transporter in gliomas may be useful for future clinical diagnosis and chemotherapy. Both immunoblotting and fluorography methodology after metabolic labeling of cells detected three EAAC1 bands, with apparent molecular weights of 73, 145, and 200+ kDa, respectively. We and others believe that they represent the monomer and oligomer forms of the EAAC1 protein. A series of experiments were conducted to elucidate the nature of the EAAC1 protein oligomerization. Peptide competition in Western blotting and immunoprecipitation (Figure 2-1 & 2-3) showed that ail three bands were specific to the anti-EAACI antibody. Data obtained from our own laboratory and other research groups using independent antibodies show that GLAST1, GLT1, and EAAC1 transporter proteins all could form oligomers. Haugeto et al. (1996), using radiation inactivation analysis, concluded that the size of the active glutamate transporter was about 160 kDa in rat cerebral cortex. Therefore, they suggested that these glutamate/aspartate transporters may operate as homo-multlmer complexes in the rat brain. However, the results obtained from my research suggests that a significant portion of EAAC1 exists as a monomer in C6 glioma cells. The experiment shown in Figure 3-7 as well as some other studies indicated that the formation of the EAAC1 oligomers could only be detected when the C6 membrane protein sample was solubillzed and subjected to manipulation prior to

PAGE 176

163 analysis. In another words, when freshly isolated C6 membrane proteins were separated by SDS-PAGE immediately after direct extraction with SDB, little or no oligomeric forms of EAAC1 would be detected. However, the longer and the harsher the manipulations are, the greater the proportion of oligomer forms. The harsher conditions include increased heating temperature, prolonged heating time, or solubilization and extended incubation in detergent solutions. Fruitless efforts, by several members of our laboratory and other laboratories, have been made to try to break apart these multimers, including treating the protein samples with reducing reagents, boiling, denaturation, and organic solvents. It seems that, once formed, these multimers are extremely stable. Interestingly, under the same treatment condition, the absolute amount of each EAAC1 form was proportional to each other, independent of the total amount of EAAC1 presented. This was clearly demonstrated with several of the pulse-labeling experiments shown in Chapters 3 & 4. Even the immature low molecular weight forms of EAAC1, detected with shorter chase after the pulse-labeling, could form oligomers. Although the possibility that there might be functional transporter oligomers in vivo, my data suggest that the majority, if not all, of these oligomers are probably artifacts caused by the sample isolation and handling. As presented in Chapter 3, 1 studied the de novo synthesis, Nglycosylation, maturation, and intracellular trafficking of the EAAC1 transporter protein in C6 glioma cells. The de novo biosynthesis of the EAAC1 protein could be easily detected with as short as 15 min pulse-labeling in C6 cells (Figure 31 1), this observation suggested that the EAAC1 transporter has a relatively high

PAGE 177

164 synthesis rate and is probably a relatively short-lived protein. When only a short period of chase time was conducted after pulse-labeling, only the immature lower MWform EAAC1 was detected, which then was transfomned into mature monomer form after longer period of chase. The transition between the immature lower MW form and the mature monomer began at about 45 min and finished around 3 hr after synthesis (Figure 3-13). The EAAC1 protein was cotranslationally N-glycosylated, because the immature lower MW forms, which were immunoprecipitated with our anti-EAACI antibody after short period of chase, were subjected to cleavage by Endo H digestion (Figure 3-15). This biosynthesis process for the EAAC1 protein is similar to the synthesis and maturation process reported for the transforming growth factor-3 type II receptor (TpRII) (Well et al.. 1997; Koli and Arteaga, 1997). The newly synthesized TpRII was detected as a lower MW immature form which was sensitive to Endo H digestion, suggesting it is co-translationally N-glycosylated by a high-mannose oligosaccharide chain. After 30 min chase, the mature form of the receptor appeared which was resistant to Endo H digestion. What makes these studies even more interesting is that, just like I have seen for EAAC1 synthesis and maturation (Figure 3-15), Wells et al. (1997) also reported that the newly synthesized immature T3RII migrated as doublet before and after the Endo H digestion, and that only one broad protein band was detected for the mature T3RII on Western blot. As stated in Chapter 3, I postulate that the translation of the EAAC1 protein may not be a continuous process, but one with distinct pause

PAGE 178

165 Sites. However, the authors studying TpRII synthesis did not propose any specific explanation concerning these two immature TpRII forms (Wells, 1997). To determine how long it would take for the newly synthesized EAAC1 protein to arrive at the PM in C6 cells, both cell surface biotinylation and Endo Hsensitivity assays were employed following pulse-chase labeling. It was shown in Figure 3-14 & 3-15 that the immature lower MWform EAAC1 protein was Endo H-sensitive and inaccessible to cell-surface biotinylation, whereas the mature monomer form EAAC1 was Endo H-resistant and accessible to cell-surface biotinylation. All these data suggested that the immature form EAAC1 existed only before the medial Golgi complex, whereas the EAAC1 mature monomer existed only beyond that compartment along the biosynthesis pathway. There was no detectable lag time between the maturation and cell-surface arrival of the EAAC1 protein. Therefore, I believe that the N-glycosylation modifications needed for transforming the EAAC1 protein from the lower MW form to the mature form in the medial and trans Golgi compartments, together with the transfer of the EAAC1 protein from the Golgi to the PM, were fast. Limited literature has been published on determining the time it takes for newly synthesized membrane proteins to arrive at the PM. In one example, using metabolic labeling in combination with specific cell surface biotinylation, Coupaye-Gerard and Kleyman (1993) determined that the delivery of newly synthesized proteins to the apical ceil surface in the polarized epithelial cell line A6 reached a maximum after 5 min of chase time, and then declined over the

PAGE 179

166 remainder of a 2 hr chase. The bulk flow of newly synthesized proteins to the basolateral membrane domain slowly rose to a maximum after 90 min. In Chapter 4, the half-life and degradation rate of EAAC1 protein, as well as the possible degradation pathway involved in the EAAC1 degradation were examined. Two independent approaches were utilized to determine the half-life of EAAC1 proteins in C6 glioma cells. First, the turnover rate of newly synthesized EAAC1 protein was measured with pulse-chase labeling followed by immunoprecipitation and fluorography analysis. As shown in Figure 4-1, it took about 14 hr for the loss of one-half of the radiolabeled EAAC1 protein molecules in C6 glioma cells. When I plotted the logarithm of the densitometry reading of remaining radioiabed EAAC1 protein against chase time, surprisingly, there was an obvious lag time of about 8 hr at the beginning of the chase. Once passed this lag time, the degradation of EAAC1 protein followed a first order kinetics, i.e., the loss of radiolabeled EAAC1 molecules was exponential, with a degradation rate of 11.5% per hr. These results were reproducible in numerous independent experiments, including the one shown in Figure 4-4, in which the early chase period was expanded and studied in detail. In summary of the data obtained in this series of experiments, it was clearly demonstrated that the degradation of the newly synthesized EAAC1 protein would not start until about 8 hr after its synthesis, and that is about 5 hr after it had been targeted to the PM. This finding was really surprising, because it was contradictory to the general belief that the degradation rate of a protein is dependent and limited by its concentration and that newly synthesized proteins are as likely to be degraded as are old ones. I

PAGE 180

167 am not sure why there is a delay for the onset of the EAAC1 degradation, but a number of possibilities exist: (1) the newly PM targeted EAAC1 protein molecules may be sequestered in special membrane domains, which are inaccessible for endocytosis or protease function; (2) the new arrivals may lack necessary posttranslational or structural modification, like phosphorylation or monoubiquitination, which are required for the recognition by adaptors for initiating endocytosis or specific protease action; or (3) these new comers may not possess the needed association and interaction with other regulatory protein(s) for their degradation by endocytosis and/or specific proteases. As to which of these possibilities is really pertinent to the EAAC1 turnover, further studies need to be done, some of which will be discussed in the Future Directions section of this chapter. In an independent series of experiments, the turnover rate of the general PM resident EAAC1 protein was determined with cell surface biotinylation and chase followed by avidin-precipitation and EAAC1 immunoblotting analysis. As I stated above, using the pulse-chase labeling method, the decay of the newly synthesized EAAC1 protein molecules is studied, without distinguishing whether they have actually arrived at the PM or not. In contrast, the surface biotinylation and chase approach would allow me to estimate the half-life and decay rate of the EAAC1 protein molecules that have already arrived the PM, regardless when they were actually synthesized. As shown in Figure 4-3, the decay of the biotinylated EAAC1 protein followed first order kinetics without a detectable delay of the onset, and the degradation rate was ~ 1 1 .5% per hr. This was exactly what

PAGE 181

168 might be expected, because the majority of the biotinylated EAAC1 molecules have already passed the "lag" period. Another very interesting aspect of the EAAC1 life cycle that I wanted to establish is how long each EAAC1 molecule stays at the PM. This data could not be obtained from either of the degradation studies described above, because before the EAAC1 protein was degraded it might have been gone from the PM for a long time. In other words, both of the previously described approaches could not tell us how long the EAAC1 protein stayed at the PM and whether there was an accumulated pool of endocytosed EAAC1 waiting for degradation. This question was answered directly using pulse-chase labeling followed by cell surface biotinylation and immunoprecipitation, and the data shown in Figure 4-5. The disapperance of the newly synthesized EAAC1 from the PM paralleled the turnover of the protein, including the delay of onset or lag phase as well as the decay period that followed first order kinetics. Although the existence of a . recycling compartment for the EAAC1 protein can not be excluded, that is. one that reaches immediate equilibrium with the PM, these data strongly argue that there was no detectable accumulation of EAAC1 along the endocytic pathway. In contrast, a glucose transporter (GLUT4) has been shown to be endocytosed and stored in both endosomes and specialized recycling vesicles (Martin et al., 1996; ' Lamideet al., 1997). After establishing the half-life and degradation rate for EAAC 1 , and its residence time at the PM, the next question was what is the mechanism responsible for the decay of EAAC1 protein in C6 cells. A number of inhibitors

PAGE 182

169 that were specific for either the endocytosis/lysosomal pathway or proteasomal pathways were utilized. It was shown the decay of EAAC1 protein was inhibited, to varying degrees, by inhibitors for general protein degradation (PMSF) as well as by inhibitors for the endocytosis pathway and vesicle trafficking (NH4CI, BFA, and NOC) (Figure 4-7 & 4-11), but not by an inhibitor for the proteasomal pathway (ALLN) (Figure 4-9). These data suggest that the degradation of EAAC1 protein is, at least partially, dependent on the lysosomal pathway in C6 glioma cells. As I stated above, the specific mechanisms responsible for degradation of mammalian nutrient transporters, regardless of the substrate, have not been documented. However, a yeast general amino acid permease, Gapl , was ' recently shown to be degraded by an ubiquitination-dependent endocytosisvacuole pathway (Springael and Andre, 1 998). On the other hand, the rapid degradation of the mutant CFTR that is retained in the ER has been shown dependent on the function of the proteasomal degradation pathway (Ward et al., 1995; Hicke, 1997). The mechanism for turnover of the PM residing wild type CFTR molecule is unknown. . ' In summary, I propose the following model for the life cycle of EAAC1 protein in C6 glioma cells: (1) the nascent EAAC1 polypeptide is transferred into the ER membrane, where it is co-translationally modified by N-linked highmannose oligosaccharide chain to generate an immature lower MWform; (2) the oligosaccharide chains attached to the newly synthesized EAAC1 protein are further modified into complex-type, presumably in the medial Golgi compartment to generate the mature form; (3) for the maturation of all newly synthesized

PAGE 183

170 Lysosome P.M. cis medial trans (Working Model of the EAAC1 Life Cycle) EAAC1 protein molecules, the shift of the attached oligosaccharide chains from high mannoseto complex-type starts from about 45 min and finishes at about 3 hr after the synthesis, a processing that presumably occurs at a constant rate; (4) the actual duration time for the shift of each EAAC1 protein molecule may be very short, because no intermediates were observed; (5) once a newly synthesized EAAC1 protein molecule is matured, it will arrive at the PM almost immediately; (6) the newly synthesized EAAC1 protein is somehow protected from degradation during the first 3 hr maturation process, as well as during its first 5 hr of residency at the PM; (7) once the decay of the newly synthesized

PAGE 184

171 EAAC1 protein starts, it follows first-order kinetics; (8) EAAC1 protein is, at least partially, degraded through the endocytosis-lysosomal pathway; (9) once endocytosed, the EAAC1 protein is transferred to the lysosomes and degraded there immediately, i.e., there is no detectable accumulation of EAAC1 protein along the endocytic pathway. Future Directions As stated before, through the work described in this research, we have just begun to ask how the amino acid transporters are synthesized and degraded in mammalian cells. Although this study has already helped us understand several aspects of the EAAC1 synthesis and degradation, I believe that what we do not know is still far more than what we do. Therefore, the extension of this study can focus on the following aspects: ; First, it will be of great interest to know whether or not an EAAC1 protein molecule can function as a SDHA glutamate transporter as soon as it reaches the PM. In another words, whether the newly arrived EAAC1 protein molecules that are still in their "lag time" for degradation are functionally active? We detected a separate pool that appeared to be resistant to protein degradation at the early stage of PM residency for the EAAC1 protein, and this pool may correlate with a separate state for its function. A number of approaches can be used to study this problem, for example, monitoring the expression of EAAC1

PAGE 185

172 plasma membrane activity after transfection using a "Tet-on" transient expression system. Secondly, the degradation pathway for EAAC1 protein can be further explored. To determine which of many possibilities are responsible for the delay of EAAC1 degradation, one could compare the phosphorylation and monoubiquitination modification, the PM subdomain localization, and the proteinprotein interactions of the newly synthesized EAAC1 protein molecules that are either in or not in the lag period. Differences in the any one of these aspects may indicate a possible explanation for this delay of EAAC1 protein degradation. Also, a temperature-sensitive ubiquitin conjugating enzyme E1 mutant cell line (Strous et al., 1996) could be used to study whether normal ubiquitination is required for the internalization and/or decay of EAAC1 protein. • " • To further investigate whether the degradation of EAAC1 protein is dependent on the endocytosis pathway, inhibitors specific for clathrin-mediated endocytosis, such as dynamin inhibitor, K*-depletion, hypotonic buffer, or cytosolic acidification, can all be employed. Finally, it will be interesting to find a condition that can be used to prevent or reverse the oligomerization of EAAC1 and other transporter proteins, and then determine whether or not the formation of the oligomers has any physiological relevance. In summary, using C6 glioma cells as a model system, several important facets of the biosynthesis, intracellular processing, degradation, and regulation of EACC1 transporter protein have been elaborated in this work. This research represents the beginning of our understanding concerning the cellular biology of

PAGE 186

173 the mammalian amino acid transporters. Future studies, such as those mentioned above, would definitely provide us with a more complete picture, and perhaps, serve as the basis for future studies focused on clinically relevant problems.

PAGE 187

REFERENCES Alberts, B., Demis, B., Lewis, J., Raff, M., Robert, K., and Watson, J.D. (1994) In Molecular Biology of the Cell New York: Garland Publish, Inc. pp. 500 800 Altin, J.G. and Pagler, E.B. (1995) Anal. Biochem. 224,382-389 Aridor, M., and Balch, W.E. (1996) Trends in Cell Biology 6, 315-320 Arriza, J.L., Eiiasof, S., Kavanaugh, M.P., and Amara, S.G. (1997) Proc. Natl. Acad. Sci. U.S.A. 94, 4255-4160 Ascoli, M. (1984) J. Ce//e/o/. 99, 1242-1250 Authier, F., Mort, J.S., Bell, A.W., Posner, B.I., and Bergeron, J.J. (1995) J. Biol. C/7em. 270, 15798-15807 Authier, F., Posner, B.I., and Bergeron, J.J. (1996) FEBS Lett. 389, 55-60 Bar-Peled, 0., Ben-Hur, H., Biegon, A., Groner, Y., Dewhurst, S., Furuta, A., and Rostein, J.D. (1997) J. Neurochem. 69,2571-2579 Beaudoin, A.R., and Grondin, G. (1991) Biochim. Biophys. Acta 1071, 203-219 Beguinot, L., Lyall, R.M., Willingham, M.C., and Pastan, I. (1984) Proc. Natl. Acd. Sci. U.S.A. 81, 2384-2388 Beguinot, L., Werth, D., Ito, S., Richert, N., Willingham, M.C., and Pastan, I. (1986) J. B/o/. C/7em. 261, 1801-1807 Benda, P., Lightbody, J., Sato, G., Levine, L., and Sweet, W. (1968) Science 161,370 Bhat, N.R., Brunngraber, E.G., and Delpech, B. (1984) J. Neurochem. 44, 1822-1824 Bissell, M.G., Rubinstein, L.J., Bignami, A., and Herman, M.M. (1974) Brain. Res. 82, 77-80 174

PAGE 188

175 Blomgren, K., Kawashima, S., Saido, T.C., Karlsson, J.O., Elmered, A., and Hagberg, H. (1995) Brain Res. 684, 143-149 Boado, R.J., Black, K.L., and Pardridge, W.M. (1994) Mol. Brain. Res. 27, 51-57 Bradbury, N.A., and Bridges, R.J. (1994) Am. J. Physiol. 267, 680-685 Brew, H., and Attwell, D. (1987) Nature 327, 707-709 Bridges, R.J., Stenley, M.S., Andersen, M.W., Cotman, D.W., and Chamberline, A.R. (1991) J. Med. Chem. 34,717-725 Bruner, J.M. (1994) Semin. Oncol. 21, 126-138 Cariappa, R., and Kilberg, M.S. (1990) J. Biol. Chem. 265, 1470-1475 Carlsson, M., and Carlsson, A. (1990) Trends Neurosci. 13, 272-276 Casado, M., Bendahan, A., Francisco, Z., Niels, D.C., Carmen, A., Gimenez, C, and Kanner, B.I. (1993) J. Biol. Chem. 268, 27313-27317 Choi, D.W. (1992) Science 258,241-243 Christensen, H.N., and Kilberg, M.S. (1987) In Physiological Society Study Guides Manchester University Press, Manchester, U.K. pp. 10-46 Christensen, H.N., and Makowske, M. (1983) Life Sci. 33, 2255-2267 Christopher, C.W., and Morgan, R.A. (1981) Proc. Natl. Acd. Sci. U.S.A. 78, 4416-4420 Conradt, M., Storck, T., and Stoffel, W. (1995) Eur. J. Biochem. 229,682-687 Cooper, J.R., Bloom, F.E., and Roth, R.H. (1996) In The Biochemical Basis of Neuropharmacology Oxford University Press, New York, pp. 12-13 Coupaye-Gerard, B., and Kleyman, T.R. (1993) J. Membr Biol. 135, 225-235 Cowburn, R., Hardy, J., Roberts, P., and Briggs, R. (1988) Neurosci. Lett. 86, 109-113 Danbolt, N.C., Pines, G., and Kanner, B.I. (1990) Biochemistry 29, 6734-6740 Danbolt, N.C., Storm-Mathisen, J., and Kanner, B.I. (1992) Neuroscience 51, 295-310

PAGE 189

176 Davis, K.E., Straff, D.J., Weinstein, E.A., Bannerman, P.G., Correale, D.M., Rosthein, J. D., and Robinson, M.B. (1998) J. Neurosci. 18,2475-2485 Davis, LL, and Blobel, G. (1986) Cell 45, 699-709 Deas, J., and Erecinska, M. (1989) Brain Res. 483, 84-90 Doherty, F.J., and Mayer, R.J. (1992) In Intracellular Protein Degradation Oxford; New York: IRL Press Dowd, L.A., Coyle, A.J., Rothstein, J.D., Pritchett, D.B., and Robinson, M.B. (1996) Mol. Ptiarmacol. 49,465-473 Dowd, L.A., and Robinson, M.B. (1996) J. Neurochem. 67, 508-516 Draye, J. P., Quintart, J., Courtoy, P.J., and Baudhuin, P. (1987) 170, 395-403 Dunn, W.A., Connolly, T.P., and Hubbard, A.L. (1986) J. Cell Biol. 102, 24-36 Dunn, W.A., and Hubbard, A.L (1984) J. Cell. Biol. 98, 2148-2159 Edwards, E.H., Sprague, E.A., Kelley, J.L, Kerbacher, J.J., Schwartz, C.J., and Elbein, A.D. (1989) Biochemistry 28,7676-7687 Fagg, G.E., and Foster, A.C. (1983) Neuroscience 9,710-719 Fairman, W.A., Vandenberg, R.J., Arriza, J.L., Kavanaugh, M.P., and Amara, S.G. (1995) Nature 375,599-603 Fiedler, K., and Simons, K. (1995) Ce// 81, 309-312 ' Fletcher, E.J., and Johnston, G.A. (1991) J. Neurochem. 57, 911-914 Fong, A.D., Handlogten, M.E., and Kilberg, M.S. (1989) Biochim. Biophys. Acta 1022, 325-332 ^ Fukuda, S., Harada, K., Kunimatsu, M., Sakabe, T., and Yoshida, K. (1998) J. Neurochem. 70, 2526-2532 Furuta, A., Martin, L.J., Lin, C.-L. G., Dykes-Hoberg, M., and Rothstein, J.D. (1997) Neuroscience 81, 1031-1042 Gazzola, G.C., Dall'Astra, V., Bussolati, 0., Makowske, M., and Christensen H.N. (1981) J. Biol. Chem. 256, 6054-6059

PAGE 190

177 Gegelashvili, G., and Schousboe, A. (1998) Brain Research Bulletin 45, 233238 Gitlin, G., Bayer, E.A., and Wilchek, M. (1987) Biochem. J. 242, 923-926 Greene, J.G., and Greenamyre, J.T. (1996) Prog. Neurobiol. 48, 613-634 Griffith, O.M. (1986) in Techniques of Preparative, Zonal, and Continuous Flow Ultracentrifugation Becl
PAGE 191

178 Hunnziker, W., Whitney, J.A., and Mellman, I. (1992) FEBS Lett. 307, 93-96 Iversen, L.L. (1975) in Handbook of Psychopharmacology Vol. 2, pp. 381-442, New York: Plenum Publishing Corporation Johnston, D., and Bystryn, J. (1984) Expt. Cell. Res. 152, 179-187 Kalderson, D. (1996) Curr. Biol. 6,662-665 Kanai, Y., Bhide, P.G., Difiglia, M., and Hediger, M.A. (1995a) Neuroreport 6 2357-2362 Kanai, Y., and Hediger, M.A. (1992) Nature 360, 467-471 Kanai, Y., Nussberger, S., Romero, M.F., Hebert, S.C., and Hediger, M.A. (1995b) J. Biol. Chem. 270, 16561-16568 Kanai, Y., Smith, CP., and Hediger, M.A. (1993) Trends Neur. Sci. 16, 365-370 Kanai, Y., Trotti, S.N., and Hediger, M.A. (1997) in Neurotransmitter Transporters: Structure, Function, and Regulation Ed. M.E.A. Reith Totowa, NJ: Humana Press Inc. Kaplan, J., and Keough, E.A. (1982) J. Cell Biol. 94, 12-19 Kilberg, M.S. {1989) Meth. Enzymol. 173,564-575 Kilberg, M.S., Stevens, B.R., and Novak, D.A. (1993) Annu. Rev. Nutr 13 137165 Kitzman, H. (1 995) Doctoral Dissertation University of Florida Klausner, R.D., Donaldson, J.D., and Lippincott-Chwartz, J. (1992) J, Cell Biol 116,1071-1080 Klockner, U., Storck, T., Conradt, M., and Stoffel, W. (1993) J. Biol Chem 268 14594-14596 Koli, KM., and Arteaga, C.L. (1997) J. Biol. Chem. 272, 6423-6427 Kuhar, J.M. (1973) Life Sci. 13, 1623-1634 Laemmli, U.K. (1970) Nature 227, C1-C24

PAGE 192

179 Lehre, K.P., Levy, L.M., Ottersen, O.P., Storm-Mathisen, J., and Danbolt, N.C. (1995) J. Neurosci. 15, 1835-1853 Lesser, G.L., and Grossman, S. (19940 Semin. Oncol. 21, 220-235 Levy-Toledano, R., Heleen, L., Caro, P., Hindman, N., and Taylor, S. (1993) Encfocr/no/ogy 133, 1803-1808 . Libby, P., Bursztajn, S., and Goldberg, A.L. (1980) Cell 19, 481-491 Lin, C.L., Bristol, LA., Jin, L., Dykes-Hoberg, M., Crawford, T., Clawson, L, and Rothstein, J. D. (1998) A/etvran 20, 589-602 Malandro, M.S., and Kilberg, M.S. (1996) Annu. Rev. Biochem. 65,305-336 Maley, F. (1989) Anal. Biochem. 180, 195-204 Malide, D., Dwyer, N.K., Blanchette-Mackie, E.J., and Cushman, S.W. (1997) J. Histochem. Cytochem. 45, 1083-1096 Maltese, W.A., and Voipe, J.J. (1979) J. Cell. Physiol. 101, 459-470 Martin, S., Tellam, J., Livingstone, C, Slot, J.W., Gould, G.W., and James, D.E. (1996) J. Cell. Biol. 134, 625-635 Matthews, J.C., Beveridge, M.J., Malandro, M.S., Rothstein, J.D., CampbellThompson, M., Verlander, J.W., Kilberg, M.S., and Novak, D.A. (1998) Am. J. Physiol. 274, C603-C614 Mattson, M.P., Lee, R.E., Adams, M.E., Gutherie, P.B., and Kater, S.B. (1988) Neuron 1, 865-876 Meier, T., Ami, S., Malarkannan, S., Poincelet, M., and Hoessli, D. (1992) AnalBiochem. 204, 220-226 Melikian, H.E., Ramamoorthy, S., Tate, C.G., and Blakely, R.D. (1996) Mol. Pharmacol. 50, 266-276 Montreuil J., Vliegenthart, J.F.G., and Schachter, H. (1995) In Glycoproteins Amsterdam, New York: Elsevier Science B.V. pp. 1-644 Nauseef, W.M., McCormick, S.J., and Clark, R.A. (1995) J. Biol. Chem. 270, 4741^747 Nelson, W.J., and Hammerton, R.W. (1989) J. Cell Biol. 108, 893-902

PAGE 193

iipp III .•^ 180 Nesbitt, S.A., and Horton, M.A. (1992) Anal. Biochem. 206, 267-272 Nicholls, D., and Attwell, D. (1990) Trends Pharmacol. Sci. 11, 462-468 Nicholson, B., and McGivan, J.D. (1996) J. Biol. Chem. 271, 12159-12164 Nilsson, P., Laursen, H., Hillered, L., and Hansen, A.J. (1996) J. Cereb. Blood FlowMetab. 16,262-270 Nuoffer, C, and Balch, W.E. (1994) Annu. Rev. Biochem. 63, 949-990 Olivares, L., Aragon, C, Gimenez, C, and Zafra, F. (1995) J. Biol. Chem. 270, 94379442 Olson, T.S., Terlecky, S.R., and Dice, J.F. (1992) In In Vivo Pathways of Degradation and Strategies for Protein Stabilization, New York: Plenum Press pp. 89-118 Oxender, D.L., and Christensen, H.N. (1963) J. Biol. Chem. 238, 3686-3699 Palmer, A.M., Marion, D.W., Botscheller, M.L., Swedlow, P.E., Styren, S.D., and DeKosky, S.T. (1993) J. Neurochem. 61,2015-2024 Palos, T.P., Ramachandran, B., Boado, R., and Howard, B.D. (1996) Mol Brain Res. 37, 297-303 Parker, K.K., Norenberg, M.D., and Vernadakis, A. (1980) Science 208 179181 Pastan, I., and Willingham, M.C. (1985) \n Endocytosis New York: Plenum Press Pearce, I.A., Cambray-Deakin, M.A., and Burgoyne, R.D. (1987) FEBS Lett 223, 143-147 Peghini, P., Janzen, J., and Stoffel, W. (1997) EMBO 16,3822-3832 Pelham, G.R. (1991) Ce// 67, 449-451 Pines, G., Danbolt, N.C., Bjoras, M., Zhang, Y., Bendahan, A., Eide, L., Koepsell, H., Storm-Mathisen, J., Seeberg, E., and Kanner, B.I. (1992) Nature 360, 464-' Plakidou-Dymock, S., and McGivan, J.D. (1993) Biochem. J. 295, 749-755

PAGE 194

181 Potter, D.A., Tirnauer, J.S., Janssen, R., Croall, D.E., Hughes, C.N., Fiacco, K.A., Mier, J.W., Maki, M., and Herman, I.M. (1998) J. Cell. Biol. 141, 647-662 Radian, R., Ottersen, O.P., Storm-Mathisen, J., Castel, M., and Kanner, B.I. ( 1 990) J. Neurosci. 10,1319-1 330 Ramachandran, B., Houben, K., Rozenberg, Y.Y., Haigh, J.R., Varpetian, A., and Howard, B.D. (1993) J. Biol. Chem. 268,23891-23897 Rapoport, T.A. (1991) FASEBJ. 5,2792-2798 Rauen, T., Jeserich, G., Danbolt, N.C., and Kanner, B.I. (1992) FEBS Lett. 312, 15-20 Resing, K.A., al-Alawi, N., Blomquist, C, Fleckman, P., and Dale, B.A. (1993) J. Biol. Chem. 268, 25139-25145 Robbins, P. (1984) J. Biol. Chem. 259, 7577-7583 Robinson, M.B., and Coyle, J.T. FASEBJ. (1987) 1,446-455 Robinson, M.S., Watts, C, and Zerial, M. (1996) Cell 84, 13-21 Ronnett, G.V., Knutson, V.P. Kohanski, R.A., Simpson, T.L., and Lane, M.D. (1984) J. Biol. Chem. 259, 4566-4575 Rothstein, J.D., Dykes-Hoberg, M., Pardo, C.A., Bristol, L.A., Jin, L., Kunci, R.W., Kannai, Y., Hediger, M.A., Wang, Y., Schieike, J. P., and Welty, D.F. (1996) Neuron 16, 675-686 Rothstein, J.D., Martin, L., and Kunci, R.W. (1992) New. Engl. J. Med. 326, 1464-1468 Rothstein, J.D., Martin, L., Levey, A.I., Dykes-Hoberg, M., Jin, L., Wu, D., Nash, N., and Kund, R.W. (1994) Neuron 13, 713-725 Rothstein, J.D., Van Kammen, M., Levey, A.I., Martin, L., and Kund, R.W. (1995) Ann. Neurol. 38, 73-84 Runnegar, M., Wei, X., Berndt, N., and Hamm-Alvarez, S.F. (1997) Hepatology 26, 176-185 Russel, S.M., and Mayer, R.J. (19830 Biochem. J. 216, 163-175 Sandoval, I.V., and Bakke, 0. (1994) Trends Cell Biol. 4, 292-297

PAGE 195

182 Schimke, R.T. (1970) In Mammalian Protein Metabolism Volume 4, New York: Academic Press, pp. Ml -211 Schmid, S.L (1997) Annu. Rev. Biochem. 66,511-548 Schousboe, A. (1981) Int. Rev. Neurobiol. 22, 1-45 Schousboe, A., and Divac, I. (1979) Brain Res. 177,407-409 Schow, S.R., and Joly, A. (1997) Cell. Immunol. 175, 199-202 Scott, CD., and Baxter, R.C. (1996) Endocrinology 137, 3864-3870 Scott, H.L., Tannenberg, A.E.G., and Dodd, P.R. (1995) J. Neurochem. 64, 2193-2202 Shotwell, M.A., Kilberg, M.S., and Oxender, D.L. (1983) Biochim. Biophys. Acta 737, 267-284 Slieker, L.J., Martensen, T.M., and Lane, M.D. (1986) J. Biol. Chem. 261, 15233-15241 Sprlngael, J.-Y., and Andre, B. (1998) Mol. Biol. Cell 9, 1253-1263 Steer, C.J., and Hanover, J.A. (1991) In Intracellular Trafficking of Proteins Cambridge, England: Cambridge University Press, pp. 3-348 Storck, T., Schutle, S., Hofmann, K., and Stoffel, W. (1992) Proc. Natl. Acad. Asi. U.S.A. 89, 10955-10959 Stoschek, CM. (1984) J. Cell Biol. 98, 1048-1054 Strous, G.J., Kerkhof, P.V., Covers, R., Ciechanover, A., and Schwartz, A.L. (1996) EMBOJ. 15,3806-3812 Szatkowski, M., and Attwell, D. (1994) Trends Neurosci. 17,359-365 Tanaka, K., Watase, K., Manabe, T., Yamada, K., Watanable, M., Takahashi, K., Iwama, H., Nishikawa, T., Ichihara, N., Kikuchi, T., Okuyama, S., Kawashima, N., Hori, S., andTakimoto, M. (1997) Sc/ence 276, 1699-1702 Teixido, J., Gilmore, R., Lee, D.C, and Massague, J. (1987) Nature 326, 883886 Terrell, J., Shih, S., Dunn, R., and Hicke, L. (1998) Mol. Cell 1, 193-202

PAGE 196

183 Tolner, B., Poolman, B., Wallace, B., and Konings, W.N. (1992) J. Bacteriol. 174, 2391-1393 Torp, R., Danbolt, N.C., Babale, E., Bjoras, M., Seeberg, E., Storm-Mathisen, J., and Ottersen, O.P. (1994) Eur. J. Neurosci. 6,936-942 Torp, R., Lekleffre, D., Levey, L.M., Haug, P.M., Danbolt, N.C., Meldrum, B.S., and Ottersen, O P. (1995) Exp. Brain Res. 103, 51-8 Van Slyke, D.D., and Meyer, G.M. (1913) J. Biol. Ctiem. 16, 197 Varon, S. (1978) In Dynamic Propertites of Glial Cells New York: Pergamon Press, pp. 93-103 Velaz-Faircloth, M. (1996) Am. J. Physiol. 270, C67-75 Voet , D., and Voet, J. (1990) In Biochemistry John Wiley and Sons, New York, pp.1 16-1 18 Ward, C.L, Omura, S., and Kopito, R.R. (1995) Cell 83, 121-127 Ware, F.E., Vassllakos, A., Peterson, P. A., Jackson, M.R., Lehrman, M.A., and Williams, D.B. (1995) J. Biol. Chem. 270, 4697-4704 Wells, R.G., Yankelev, H., Lin, H.Y., and Lodish, H.F. (1997) J. Biol. Chem. 272, 11444-11451 Wilkinson, K.D. (1997) FASEB J. 11,1245-1256 Woodard, M.H., Dunn, W.A., Laine, R.O., Malandro, M.S., McMahon, R., Simell, 0., Block, E.R., and Kilberg, M.S. (1994) Am. J. Physiol. 226, 817-824 Yokota, M., Saido, T.C., Tani, E., Kawashima, S., and Suzuki, K. (1995) Stroke 26, 1901-1907 Yu, H., Kaung, G., Kobayashi, S., and Kopito, R.R. (1997) J. Biol. Chem. 272, 20800-20804 Zanetta, J. P., Benda, P., Gombos, G., and Morgan, I.G. (1972) J. Neurochem. 19, 881-884 Zerangue, N., and Kavanaugh, M.P. (1996) Nature 383, 634-637

PAGE 197

BIOGRAPHICAL SKETCH Wenbo Yang was born to Yu-Fang Jiang and Yi Yang on August 9, 1964, in Shandong Province, People's Republic of China. She studied medicine at Nanjing Railway Medical College and was conferred the M.D. degree in 1985. After one year of clinical internship at Nanjing Railway Medical College Affiliated Hospital, she worked as an Assistant Investigator and Lecturer for three years, and later enrolled as a graduate student in the Department of Pathophysiology, Peking Union Medical College. There she graduated with a master's degree in medicine in 1 991 , with a major in pathophysiology. In the following two years, she worked as a postdoctoral fellow in Dr. Edward Block's laboratory in the Department of Medicine at the University of Florida. In January of 1994, Ms. Yang began her Ph.D. study in the Department of Biochemistry and Molecular Biology at the University of Florida and joined the laboratory of Dr. Michael S. Kilberg. She is married to Ming Hu, and they have one daughter, Helen Y. Hu. 184

PAGE 198

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Michael S. Kilberg, urrair Professor of Biochemistry anc Molecular Biology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. )f Doctor of Philosophy. Charles M. Allen Professor of Biochemistry and Molecular Biology I certify that I have read this study and that in my opinion it confonns to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Brian D. Cain Associate Professor of Biochemistry and Molecular Biology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. William A. Dunn Associate Professor of Anatomy and Cell Biology

PAGE 199

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Susan C. Frost Associate Professor of Biochemistry and Molecular Biology This dissertation was submitted to the Graduate Faculty of the College of Medicine and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. December, 1998 Dean, Graduate