* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Identification of the Human Erythrocyte Glucose Transporter (GLUT1
Endogenous retrovirus wikipedia , lookup
Paracrine signalling wikipedia , lookup
Monoclonal antibody wikipedia , lookup
Clinical neurochemistry wikipedia , lookup
Vesicular monoamine transporter wikipedia , lookup
Ribosomally synthesized and post-translationally modified peptides wikipedia , lookup
Biochemical cascade wikipedia , lookup
Proteolysis wikipedia , lookup
Adenosine triphosphate wikipedia , lookup
Ligand binding assay wikipedia , lookup
Polyclonal B cell response wikipedia , lookup
Magnesium in biology wikipedia , lookup
Western blot wikipedia , lookup
Biochemistry wikipedia , lookup
Signal transduction wikipedia , lookup
Oxidative phosphorylation wikipedia , lookup
Two-hybrid screening wikipedia , lookup
Evolution of metal ions in biological systems wikipedia , lookup
University of Massachusetts Medical School eScholarship@UMMS GSBS Dissertations and Theses Graduate School of Biomedical Sciences December 1999 Identification of the Human Erythrocyte Glucose Transporter (GLUT1) ATP Binding Domain: A Dissertation Kara B. Levine University of Massachusetts Medical School Follow this and additional works at: http://escholarship.umassmed.edu/gsbs_diss Part of the Life Sciences Commons, and the Medicine and Health Sciences Commons Recommended Citation Levine, KB. Identification of the Human Erythrocyte Glucose Transporter (GLUT1) ATP Binding Domain: A Dissertation. (1999). University of Massachusetts Medical School. GSBS Dissertations and Theses. Paper 247. http://escholarship.umassmed.edu/gsbs_diss/ 247 This material is brought to you by eScholarship@UMMS. It has been accepted for inclusion in GSBS Dissertations and Theses by an authorized administrator of eScholarship@UMMS. For more information, please contact [email protected]. IDENTIFICA TION OF THE HUMAN ERYTOCYT GLUCOSE TRANSPORTER (GLUTl) ATP BINDING DOMAIN Dissertation Presented Kara B. Levine Submitted to the Faculty of the University of Massachusetts School of Biomedical Sciences, W orces ter in partial fulfilment of the requirements for the degree of Doctor of Philosophy December 1 5 1999 Cellular & Molecular Physiology Copyright Notice (Q Kara B. Levine , 1999 All Rights Reserved Parts of this dissertation Levine, K. B., Cloherty, E. have appeared in: , Fidyk , N. Structural and Physiologic Determinants of Regulation by Adenosine 37(35): 12221- 12232 Sugar Transport Biochemistry. Levine , , and Carruthers , A. (1998) K. B., Hamil S., Cloherty, E. Human Erythrocyte Triphosphate. , and Carruthers, A. (1999) Alanine Scanning Mutagenesis of the Human Erythrocyte Glucose Transport Putative ATP Binding Domain. Submitted. Dr. Cheryl Scheid, Chair of Committee Dr. Thomas Honeyman, Member of Committee Dr. Ann Rittenhouse, Member of Committee Dr. Kendall Knight, Member of Committee Dr. Laurie Goodyear, Member of Committee Dr. Anthony Carruthers, Thesis Advisor Acknowledgements First and foremost I would like part of the learning of the Carruthers which mistakes process. I would also like to thank the members often acted as my personal added bonus of hands in the lab , I don t think this thesis I whom over sounding board my extra would have ever would also Kelley DeZutter and Janice Lloyd , and tricks needed Heard maintaining my sanity. A special thanks go to Erin Cloherty and Stephanie Hamill completed without their help. a are considered a normal laboratory. Especially Dr. Karen the past five years like to for teaching set of been acknowledge Julie me the techniques to do molecular biology. I would like to thank my family for their unending support throughout this experience we like to especially like to thank instiled in Dr. Anthony this thesis. I am grateful that he maintains laboratory environment in which had the my mentor direction and encouragement during Carruthers for his invaluable the creation of to thank call graduate school. my parents Nate and Judy Levine me an intense desire to succeed. Special thanks I would who go to sister Shari , her husband Steve , and my mece " The Chloe " who were always there for advice and to remind me that outside , continues to exist. a real world Abstract The human erythrocyte glucose transport protein (GLUTl) interacts with , and is regulated by, cytosolic ATP. This study asks questions concerning ATP modulation of GLUTI the following mediated sugar transport. 1) Which region(s) of GLUTI form the adenine nucleotide- binding domain? 2) modulation of sugar transport? 3) Is sufficient for sugar transport What factors influence A A TP interaction with GLUTI regulation? The first question was addressed through peptide mappmg, terminal sequencmg, and alanine scannmg mutagenesis usmg (3 2P) -azidoA TP , combination of strategies to a photoacti vatable used a and photolabeling examine how glycolytic intermediates, pH , affect A GLUTI A TP analog. We then transport measurements transporter oligomeric structure of TP regulation and of sugar transport. Finally, GLUTI was reconstituted into proteoliposomes determine whether A TP is function in-vitro. sufficient for the modulation of GLUTI Vll This thesis presents data supporting the hypothesis that residues 332- 335 contribute to the efficiency of adenine nucleotide to GLUTl. In addition , we show that AMP , acidification , and the transporter form antagonize conversion of to its dimeric binding A regulation of sugar transport. Finally, we present results that support the proposal that A for transport modulation. TP interaction with GLUTI is sufficient Vll Table of Contents ACKNOWLEDGMENTS 1V ABSTRACT VIll TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES CHAPTER I INTRODUCTION Sugar Transport Regulation Skeletal Muscle Cardiac Muscle vianlNucleated Erythrocytes ATP Regulation of GLUTl in Human Erythrocytes Strategy CHAPTER II MATERIALS & METHODS CHAPTER III FACTORS INFLUENCING A MODULATION OF ERYTHROCYTE SUGAR TRANSPORT Introduction Results Glycolytic Intermediates Oligomeric Structure Does ATP Modulate GLUTl Mediated Sugar Transport Directly? Discussion 17 CHAPTER IV PEPTIDE MAPPING THE GLUTI A BINDING DOMAIN Introduction Results Peptide Mapping terminal Microsequencing Discussion CHAPTER V ALANINE MUTAGENESIS OF THE GLUTI CONSENSUS A TP BINDING SEQUENCES Introduction Results Purification of Human GLUTl HA His 8 1 Scanning Alanine Mutagenesis of A TP Consensus Binding Sequences glucose Uptake into Cos- 7 cells AzidoATP Labeling of Wild Type & Discussion 101 113 SIGNIFICANCE AND FUTURE DIRECTIONS 120 Mutant hGLUTl CHAPTER VI REFERENCES 136 List of Tables Table Table 2 Table 3 Effects of glycolytic intermediates on erythrocyte 30MG uptake Effect of reductant on A TP-modulation erythrocyte 30MG uptake Characteristics of S. Cerevisiae strain RE700A transfected with GLUT 1 Table 4 Table 5 GLUT 1 alanine mutants Antibody binding With GLUTI to Cos- cells transfected (y- (y- List of Figures Figure Figure 2 putative topography and Consensus A TP binding domains. GLUT Effects of pH on A TP-modulation of sugar transport (A) and GLUT I- labeling by azido- Figure 3 15 P)A TP (B). Effects of pH on the concentration dependence of A TP-modulation of sugar transport (A) and ATP- inhibition of GLUT I- labeling by azidoP)A TP (B). Figure 4 Azido- (y_ 2P)A TP labeling 5 1 of dimeric and tetrameric GLUT Figure 5 A TP stimulates 30MG uptake into reconstituted , purified GLUT proteoliposomes. Figure 6 Peptide mappmg of photolabeled GLUTl. Figure 7 Peptide mappmg of photolabeled GLUTI. Figure 8 Summary of peptide mapping analysis of P)ATP labeled GLUTI (Figures 1 & 2). Figure 9 GLUTl HA His expression in Saccharomyces Cerevisiae. Figure 10 terminal GLUTI antibody western blot of RE700A +/- GLUTI HA His 7 1 Figure 11 Purification of GLUT 1 HA His from Cos- cells. Figure 12 9 1 cells deoxyglucose uptake in Cos- transfected with GLUT 1 HA His Figure 13 Photo labeling of parental GLUT 1 and transfected wild type and GLUT 1 in Cos- 7 cells by the sugar export site ligand (3H) - Figure 14 mutant human 100 CCB. Azido( P)ATP labeling of GLUTI HA His and its corresponding alanine Figure 15 (wild type and mutant). 106 photolabeling affinity purified GLUTI HA His and A TP- inhibition of azidoA TP of 109 GLUT1 332A Figure 17 104 Corrected AzidoA TP labeling GLUTI HA His Figure 16 mutants. AzidoA TP labeling of stripped of Cos- cell membranes peripheral proteins. 112 Xll Abbreviations 2DOG Deoxyglucose 30MG 0-methyl- ATP glucose adenosine 5' triphosphate GLUTI c- terminal antibody CCB cytochalasin- b Cos- 7 simian kidney cell line 8-an ti body tetrameric GLUTI exofacial DTT di thiothrei tol Ghosts sealed erythrocyte antibody membranes lacking cytosol GLUTI His human erythrocyte glucose transport protein histidines hemagglutinin HA- anti- HA antibody GLUTI N- terminal Ni - NT A antibody nickel column media XIV PMSF pheny lmethy lsulfony If! uoride (serine protease inhibitor) PFK phos phofruc tokinase RE700A S. Cerevisiae URA 52 mutant with seven of its glucose transporters (HXT 1- 7) SDS sodium dodecyl sulfate Stripped Ghosts red blood cell ghosts membrane proteins stripped of deleted peripheral Chapter Introduction selectively permeable barrier , all mammalian cells. This barrier and proteins. The hydrophobic is the cell membrane , surrounds composed primarily of lipids nature of this cell membrane allows non- polar molecules, and only the very smallest polar molecules pass through by simple than water such as diffusion. In ions, sugars, metabolites must be transported specialized integral contrast, polar molecules larger ammo acids, across the nucleotides cell membrane by membrane proteins which function or carriers. The channels (selected on the basis of electrochemical gradient are water filled pores that charge and size) , and cell to diffuse as channels allow substances down an across membranes. In contrast, carrier proteins transfer bound membranes by way of conformational changes. solutes across Two types of mediated transport have been described , facilitated and active. Facilitated transport is characterized by rapid, bidirectional transport of solutes across the protein mediated membrane, down electrochemical gradient. Active transport couples a metabolic energy source (i. e. A TP) to the rapid protein mediated movement of solute against a concentration gradient (Carruthers , 1990). Glucose , a hexose monosaccharide , is an essential source energy m mammalian and non-mammalian cells. Cellular metabolic machinery reqUIres D- glucose 1990). A family of for the production of A TP (Darnell facilitative glucose transporters selectively pump glucose across cell membranes areas of high concentration to with a net movement of areas of low concentration (Stein, 1986). Stereoselectivity, saturation kinetics , fungal metabolite sugar from and inhibition by the cytochalasin B (CCB) characterize these facilitative glucose transporters. Five mammalian glucose transporter isoforms have been identified in the facilitative glucose transporter family. These five isoforms are highly homologous (50- 80%) yet exhibit distinct distributions (Gould and Bell , 1990). GLUTl , transporter , is found red blood cells , (Mueckler et in virtually all tissue the erythrocyte glucose mammalian tissues including brain microvessels , kidney, colon, and placenta aI., 1985; Birnbaum et aI. , 1986; Gould and Bell , 1990). GLUT2, the liver type transporter , intestine, and is located in liver , kidney, small -cells (Thorens et aI. , 1988). The presence pancreatic of GLUT3 in fibroblasts and blood vessels suggests a ubiquitous tissue distribution (Kayano et aI. , 1988). GLUT4, the insulin sensitive transporter , is found in skeletal and cardiac muscle as well as adipocytes (Birnbaum , 1989; Charron aI. , 1989; James et aI. , 1989; Kaestner fructose transporter , is found et aI. , 1989). GLUT5, a primarily in the small intestine lower levels of expression identified in subcutaneous fat (Kayano et aI. , The goal of this et aI., 1989; Fukumoto et with kidney, skeletal muscle, and 1990). understand all aspects laboratory is to of faciltative glucose transport. To accomplish this, we have chosen to study GLUT1 , transporter. the human erythrocyte form of the Specifically, our interest GLUTI-mediated sugar transport of GLUTI make it ideal is lies in facilitative glucose understanding how regulated. Certain characteristics for studying transport regulation. First GLUTI comprises 6% of total human red blood cell protein. The ease with which these erythrocytes can be acquired facilitates the purification of large quantities (10- 20 mg/unit of blood) of the transporter for use in experimental protocols. cytosolic environment of red blood In addition , the cells can be readily controlled by reversible hypotonic lysis, thereby permitting analysis of cytosolic determinants of Sugar Transport transport. Regulation Certain cell types , such as skeletal and cardiac adipocytes, and nucleated avian muscle erythrocytes can metabolize sugar at rates that exceed transport capacity. Under these conditions glucose transport becomes the rate- limiting step in sugar utilization and cells become susceptible .to acute transport regulation (Elbrink and Bihler , 1975; Carruthers , 1990). Compounds that regulate glucose carner function can be divided into two groups: hormones and metabolic factors. The primary hormone to which GLUTs respond is insulin , the pancreatic islets in response to (Vander 1990). Cells muscle. In contrast , exercise , which is secreted by changing blood nutrient levels that respond to insulin include adipocytes and metabolic factors such as anOXIa , ischemia and metabolic pOIsons , have been shown to affect skeletal muscle , cardiac muscle, and nucleated aVIan erythrocytes. In either case, exposure to these substances and conditions stimulates glucose uptake (Carruthers, 1990). Two hypotheses exist which explain transport acceleration. following equation: Transport capacity is V max = Kcat (E)t catalytic turnover of an the enzyme , mechanisms governIng defined by the where K cat represents the and (E)t is equal concentration. It is apparent from to total enzyme this relationship that an increase in either transporter catalytic turnover or transporter number would alter transport capacity (V max)' The recruitment suggests that transport stimulation is a redistribution to the hypothesis direct result of transporter ' plasma membrane from a subcellular pool (increased (E)t ) (Cushman and Wardzala , 1980; Suzuki and Kono, 1980). In contrast, the activation hypothesis implies' that acceleration is the result of an increase transporters already present in the transport in the intrinsic activity of membrane (increased k cat (Carruthers, 1990). Individually, carner recruitment or activation could result in sugar transport stimulation. However , mechanisms could also function coordinately in cells. these two Under these conditions, inactivation and reinternalization would result transport inhibition (Carruthers, in 1990). The following sections present four model cell types that exhibit acute regulation of glucose transport , skeletal and cardiac muscle avian erythrocytes, and human red blood cells. Skeletal Muscle Skeletal muscle is the primary site of (Douen et aI. , tissue is 1990; Klip and in-vivo glucose disposal Paquet, 1990). Sugar transport maximally stimulated by insulin , as well as in this exercise, anoxia/hypoxia , ischemia , and inhibition of oxidative phosphorylation , all of which result in an increased demand for A (Clausen , 1975; Elbrink and Bihler, 1975). Subcellular fractionation and cytochalasin B binding experiments have shown that stimulation is characterized by a two- fold increase in plasma membrane and t- tubule carrier levels (Hirshman et Wheeler , 1988; aI. , 1991). Douen et aI. , 1990; Cartee et aI. , aI., 1988; 1991; Goodyear Experimentally, striated muscle has been found to transporter isoforms: GLUT 1 (Mueckler 1986; Gould and Bell , 1990) and et aI., 1985; Birnbaum et aI., GLUT4 (Cushman and Wardzala, 1980; Birnbaum, 1989; Charron et aI. , GLUT1 , contain two 1989; James et aI., 1989). detected primarily at the cell surface, is proposed to be responsible for basal (James et aI. , transport rates 1988). In contrast , in unstimulated skeletal muscle transport stimulation due to anOXIa and exercise appears to result from GLUT4 translocation to the cell membrane from Gao et aI. , an intracellular compartment (Douen et aI. , 1994). More recent evidence rilUscle contains a pool of suggested that skeletal has conformationally masked GLUT4 that becomes exposed during transport stimulation (Wang et aI. , Supplementary experiments have indicated translocation in level of skeletal muscle that 1959; is not sufficient to account for the Morgan et aI. , 1965; Sternlicht et aI., 1988; Wheeler , The implication is that 1961; to anoxia and Morgan et aI. 1988; Klip and transport stimulation in be due to both an Increase 1996). GLUT4 transport stimulation observed in response exercise (Morgan et aI. , 1990; in the turnover Paquet, 1990). skeletal muscle may rate of existing carners total number as well as the - 1988; Klip and of transporters present (Sternlicht et aI. Paquet, 1990). The degree to which recruitment and activation contribute to exercise , and metabolic- depletion induced transport anoxia, stimulation in skeletal muscle has Carruthers suggests that this not yet been determined. question might be answered measunng transport kinetics and carrier by capacity in sarcoplasmic vesicles isolated from basal and stimulated muscle (Carruthers 1990). The structural complexity of skeletal muscle definitive markers for subcellular fractions difficult task that has hindered analysis and the lack of make this a technically (Klip and Paquet , 1990) ;Cardiac Muscle Ii \ Two major glucose 'are expressed in transporter isoforms - cardiac muscle. Levels of "1990; Kraegen et aI. , ;h.eart muscle than in 1993) were found GLUTI and GLUT4 - GLUT 1 (Calderhead et aI. to be 3 to 4 fold greater in skeletal muscle (Doria et aI. , bsence of insulin GLUT4 is 1993). In the localized to tubulovesicular elements tadjacent to the sarcolemma and transverse tubules (Fischer et aI., 1997). Immunohistochemical techniques have detected GLUTI cardiomyocyte plasma membranes and intercalated discs (Doria et aI., 1993). Subcellular fractionation has revealed quantity of GLUTI is also (Fischer et aI. , to undergo present in 1995; Fischer significant intracellular membranes et aI. , 1997). Both carners were found translocation from an intracellular compartment cell surface in response to (Kolter et aI. , 1992; Doria aI., 1994; Zorzano et ischemia , hypoxia or aI., 1997). This plasma membrane (James et aI. , contractile activity is contrary to what is proposed GLUTI is detected only at the 1988). Fischer et al (1997) identified two intracellular pools of carrier in are available for translocation , one of which GLUTl. The evidence suggests is insulin insensitive (Fischer et aI. , whether transport 1997). stimulation in heart cardiomyocytes that contained significant that this pool of surface transporter activation. Wheeler carriers It is not yet certain muscle results from transporter recruitment alone , or occurs in combination transporters in the heart to the et aI. , 1993; Sun et aI. , 1994; Wheeler to occur in skeletal muscle , where quantities of that a with cell found that glucose show a 50% increase in membrane levels in response to anOXIa compared to a 150% (Wheeler, 1988). In contrast , studies undertaken by Haworth et al showed that anOXIa Increases glucose transport cardiomyocytes 10- 20 in skeletal increase in glucose uptake fold (Haworth in rat and Berkoff 1986). In muscle , it appears that translocation heart, as may not account for the degree of transport stimulation seen in response to anoxia ischemia, and exerCIse. This provides additional evidence that carrier activation may constitute a necessary part of the process of transport stimulation. vianlNucleated Erythrocytes In contrast to human erythrocytes , aVIan erythrocytes contain cytosolic structures , primarily mitochondria and a nucleus (Harris and Brown , 1971; Brown , 1975). Sugar transport in avian erythrocytes is subject to regulation by cellular metabolic status. Under basal conditions , protein-mediated net sugar uptake is detectable. Exposure to anoxia or to ' inhibitors of oxidative metabolism results in a robust stimulation of net sugar uptake (Cheung et aI. , 1977; Simons , 1983). Recent studies have barely demonstrated that avian erythrocyte GLUT1 , which is present at the cell surface at all times (Diamond and Carruthers , 1993), functions as an antiporter under normoxic conditions (Cloherty et aI. , This means that sugar metabolic disruption , uptake is coupled to sugar avian red uniport (Cloherty et aI. , absence of exit. Following cell GLUTI can catalyze sugar 1996) allowing net sugar uptake in the intracellular sugar. Antiport function by cellular metabolic status. The mediator switching mechanism is currently unknown. proposed that 1996). appears unaffected of this antiport/uniport Cloherty et al has regulatory factors exist in the avian erythrocyte which allosterically inhibit GLUTI uniport. Cloherty further proposes that AMP dependent protein kinase is involved in activation/deactivation pathway (Cloherty et aI. , this 1996). While avian erythrocyte GLUTI-mediated sugar uniport is controlled more completely by cellular metabolic status than is sugar uniport , the similarity between transport tissues is mediated by and Hinkle , 1985) which striking. Human erythrocyte GLUTI (Mueckler et aI. , is assembled and sugar 1985; Wheeler functions as an allosteric ,' homotetramer (Hebert and Carruthers , 1992; Zottola et aI. , 1995). In . freshly isolated erythrocytes, net uptake of 3- 0-methylglucose (a transported but nonmetabolized sugar) is slower by as much as 150- fold than 30MG exchange transport (Cloherty et aI. , Following hypotonic cell lysis and cellular resealing medium, 30MG net uptake :stimulated by 3 to (but not exchange et aI. , 10- fold (Helgerson erythrocytes are resealed in 1996). in A TP- free transport) is 1989). A TP-containing saline When lysed 30MG net ,uptake and exchange are indistinguishable from transport Jin intact cells (Helgerson et aI., 1989). observed This suggests that as with vian red cell sugar transport, uptake stimulation in human erythrocytes results from accelerated uniport function and that (cytoplasmic A TP may regulate cellular glucose antiport/uniport (Levine et aI. , 1998). ATP Regulation of GLUTl in Human Erythrocytes Human erythrocyte GLUTI is subject ATP (Carruthers to allosteric inhibition by and Helgerson , 1989; Helgerson affinity high capacity net sugar uptake. Maximum rates et aI., 1989) which uptake to high affinity of exchange transport and net " j sugar exit appear unaffected by cellular A TP levels (Helgerson et aI. -- 1989). The net effect of A levels (50 to 100 JLM) TP on 30MG uptake at subsaturating sugar is to stimulate transport by a factor to the increase in the ratio V ma xi K m(app) while levels greater than K m(app) We have demonstrated that the a GLUTI conformational 1989). (saturated transport) is inhibited. isolated human erythrocyte nucleotide binding protein .glucose transport protein is also (Carruthers and Helgerson , transport A TP binding to GLUTI promotes change that leaves the GLUTI carboxyl- terminus less accessible to peptide- directed antibodies. This A TP- dependent conformational change is antagonized by AMP and ADP Which appear to function as competitive inhibitors of A TP binding to GLUTl. Initial peptide mapping studies identified a range of ,residues from 270- 380 as a likely region where A TP attaches (Carruthers and Helgerson , 1989). The primary sequence of GLUTI indicates the presence of three potential consensus A TP binding (Figure 1). Sequence I is a Walker A motif located etween residues 111- 118 , sequence hich spans residues 225- 229, II is a hydrophobic loop motif and sequence III , located at residues i!! (Ii ;'-, !"';. Figure 1. GLUT1 putative topography and consensus ATP binding domains. The three proposed consensus ATP binding sequences located in GLUT1 are highlighted in black. Sequence I is a Walker A motif located between G111 and S117. Sequence II is referred to as a loop motif and stretches from . K225- K229. Sequence III , a Walker B motif , extends from G332- G343. i 1 ':il .'i II.. J:I !i:: 120 ;,(' (j 45 00G G€08G0 (g 126 h.r. 0(00C9co IT \!r. 162 (' 181 (D (9 2070 (0CQ(I 00 (00 (Y 81 (2 -(;63 8 CD Beg : 0 145 M 0 Q0 (! 0l! e 028 0Gf(A 0112 ?i &f 0 (0 :f0 0? Ci 0 000 00c! 188 (! (! 291 (0 0 327 J'Gj 0 272 GJ!J GXXTLXL (Y KXVXK (! fi 307 0338 GOO GXXXXGK (S, 00 C9 0000 (Y 0(y000000 III (9 4220432 Interstitium ._-,- (! (1367 W; 0 GW:(c (0 CItr 008e0 358 (# f# g G) "" 6f 0 !:0 F 4040 450 391 470 (YG00Ci0 I ) XXAG g 000m 06J 0 \V4) (0 (2 0) DI. 0 (9 0 0 4) (Y 0 00 (00CD0 00 I I I II (Y 0235 (Y (Q 492 Cytosol Reg i on Reg i on Reg i on 332- 343, is a Walker B motif which overlaps region of ATP binding on GLUTI (Walker , whether A TP binding with the proposed 1982 #4492). It is unclear to GLUTI is sufficient to mediate sugar transport regulation in human red blood cells molecular species required. are Wheeler has demonstrated that A TP activity of reconstituted human GLUTI protein the (Wheeler , 1989). these results. 1) Required could be mISSIng or dysfunctional in these reconstituted vesicles or 2) the GLUTI used may exhibit an abnormal A TP response sugar (Carruthers and Helgerson however , other does not modulate number of explanations could account for cellular components or whether in these experiments or capacity to 1989). It is transport more probable that the problem lies with Wheeler s use of reduced (DTT- treated) GLUTI which we have shown to exist as a GLUTI dimer (Hebert and Carruthers , 1992). Work in our laboratory has shown ii, that conversion of binding and A tetrameric GLUT 1 TP modulation These results imply that A to its dimeric form ablates of transport (Levine et TP modulation A aI., 1998). of GLUT 1 is dependent upon transporter quaternary structure. In contrast to Wheeler , i work, our studies employ the physiologic , non-reduced form of GLUTI ( a GLUTI tetra mer (Hebert and Carruthers , responds to A 1992)) TP. In the present study we ask: 1) What physiologic factors influence A TP-modulation of sugar transport? 2) Is A TP interaction with GLUTI sufficient for sugar transport regulation? the GLUT 1 nucleotide- binding we use that a combination of 3) Where is domain? To address these questions, transport measurements and azidoA photolabeling strategies to examIne the effects of glycolytic intermediates , pH changes , and transporter oligomeric structure A TP modulation of GLUTl. In addition , purified erythrocyte GLUTI are carried sufficient for modulation of on reconstitution assays using out to determine if A TP GLUTI mediated sugar transport. Finally, peptide mapping, n- terminal sequencing, and alanine scanning mutagenesis are used to locate the GLUTI A TP binding II: If! I:i Our findings suggest that A TP regulation intracellular AMP and acidification of sugar transport , and that A TP- interaction with GLUTI is sufficient for transport modulation. data also support the GLUTl , 364 The hypothesis that tetrameric , not dimeric is regulated by ATP. In addition we show that residues form an integral part of the A TP binding domain on GLUT and that epitope tagged human GLUTI expressed normal with respect to sugar Finally, alanine substitution 301- in Cos- cells is transport and azidoA TP binding. at residues 338- transport, while alanine substitution 340 ablates sugar at residues 225- 229 , and 332- 334 increases (32 P)A TP binding. II! II' (y- , , Chapter n Materials and Methods Azido . Materials. CA. (3H)- P)ATP was purchased from ICN , Costa Mesa, Methylglucose was purchased from Dupont NEN, Wilmington, DE. Rabbit antisera carboxyl- terminal (C- Ab) raised against a synthetic peptide of the rat brain glucose transport protein (residues 480- 492) was obtained from East Acres Biologicals Southbridge MA. Glycoprotein Detection Kits and ECL detection reagents were purchased from Amersham Pharmacia Biotech Piscataway, NJ. Gradient Scientific , Portsmouth SDS PAGE gels were purchased from Owl , NH. Nitrocellulose and Immobilon- P obtained from Fisher Scientific , Medium Pittsburg, PA. Dulbecco s Modified FBS, Penicilin/Streptomycin , Were purchased from Gibco BRL were and G418 (Geneticin) Grand Island , NY. The QuickChange Site- Directed Mutagenesis Kit was acquired from Stratagene , Marco Island , FL. pcDNA3. 1 + vector was supplied Invitrogen , Carlsbad , CA. Fugene 6 Transfection Reagent , ntibody (mouse monoclonal by anti- HA-ab), Endoproteinase Lys- C, and N- - dodecyl-~- D-maltoside were obtained from Roche, Indianapolis , Nickel- NTA (Ni- NTA) was purchased from Qiagen , Valencia , IN. CA. Centricon concentrators were acquired from Millipore, Bedford, MA. materials were purchased from Sigma unless stated otherwise. Tris medium consisted Solutions. of 50 mM Tris- HCI , 5 mM MgCl pH 7.4. Sample buffer (2X) contained 0. 125M Tris- HCI pH 6. 8, 4% SDS , 20% glycerol, and 50 mM DTT. KCI medium consisted of 150 mM KCI , 10 mM Tris- HCI , 5 mM MgCl , 4 mM EGTA, pH 7.4. Stopper consisted of ice cold phloretin , and 10 Tris HCI , 1 JlM KCI medium plus HgCI 10 JlM cytochalasin B , 100 JlM , pH 7.4. Lysis buffer contained 10 mM EGTA , pH 7. 2. Phosphate buffered saline (PBS) :n! I .u contained 171 mM NaCI , 10 mM Na HP0 , 3.4 mM KCI , , pH 7. 3. Buffer A included 300 mM NaCI , 20% Glycerol , and 50 mM NaH A + 0. 1 % N- dodecyl ~- 10 1.84 mM imidazole . Buffer B was composed of Buffer maltoside. Buffer C contained 10 mM Tris- HCI (pH) 7. , 300 mM NaCI , 25 mM imidazole , maltoside.. Elution buffer consisted of 10 and 0. 1 % N- dodecyl mM Tris- HCI (pH 7. 5), .- --- '. -, ,,' .., 500 mM NaCI , 300 mM imidazole , and 0. 1 % N- dodecyl ~- maltoside. Krebs Ringer Phosphate (KRP) contained 4. 5% NaCl t- 75% KCI, 7. 85% CaCI , 19. 1 % MgS0 , and 0. 1 M Na HP0 7,'r:" A peptide corresponding Antisera. to GLUT 1 residues 88- 95 was synthesized by the University of Massachusetts Medical Kchool Peptide Synthesis facility. This peptide was conjugated to keyhole limpet hemocyanin USIng a kit purchased from Pierce. Rabbit antisera (N- Ab) against this GLUTI peptide were obtained i'; ,d ..f from Animal Pharm Services Inc of Healdsburg, CA. Whole IgGs were "'1',' purified from crude (Sogin and Hinkle , ".I. serum by ammonium sulfate precipitation :il. 1980). - 1' I; . Red Cells. Erythrocytes were isolated from whole human blood by "i , I! repeated wash/centrifugation cycles in volume of whole blood was mixed with and centrifuged at 10 000 x g for 5 . Suffy coat were aspirated, and the ice-cold KCI medium. One 3 volumes of KCI medium minutes at 4 o C. Serum and the wash/centrifugation cycle was until the supernatant was clear and the buffy coat was no longer visible. Cells were resuspended in 20 volumes of KCI and, if used in subsequent sugar transport assays , were room temperature for 30 minutes in order to medium incubated at deplete intracellular sugar levels. Erythrocyte Membrane 40 volumes lysis medium , centrifuged at 22 Washed red cells were lysed in Ghosts. incubated on ice for 10 minutes , then 000 x g for 20 was aspirated , and minutes at 4 o C. the lysis/centrifugation/aspiration cycle was repeated. The resulting membranes KCI medium at 37 o C for centrifugation at 22 The supernatant resealed by incubation in 60 minutes and harvested by 000 x g for 20 Glucose Transport were Protein. minutes at 4 o Glucose transport protein was purified from human erythrocytes as described previously and Carruthers is incorporated 1992). The carrier co- purifies with red cell (Hebert lipid and into unsealed proteoliposomes upon removal of by dialysis (Carruthers 1986). Glucose Carrier Labeling with 8- Azidoly32PjATP. 1989). Briefly, a as described previously (Carruthers and Helgerson methanolic solution of resuspended in 500 (azidoA TP)). 8-azido(y- P)A TP is IlL Tris medium pH 8. Purified GLUTI dried under nitrogen , 10 5 (90 /lCi proteoliposomes (500 sedimented by centrifugation at 14 000 x g Labeling was /lg and final /lM protein) are for 20 minutes and combined with the methanol- free 8-azido (y_ 32P)A TP solution. The :11 :111 suspenSIOn is incubated on ice for 30 minutes by which time A : 'PI binding to GLUTI achieves equilbrium. Samples are placed in a plastic weigh boat on Ice and irradiated for 90 seconds at 280 nm a Rayonet Photochemical uv Reactor. Following irradiation (photolabeling), GLUTI is washed three times to remove unbound 8azido (y_ 32P)A TP by centrifugation (14 000 x g for 15 minutes) and resuspended in 500 IlL Tris medium pH 8. 5. Photolabeling of Cos- 'recombinant hGLUTI is identical to that described above except for the following 1) samples eluted from Ni- NTA were combined the methanol- free 8-azido (y_ 32P)A TP at pH 7.4 , rradiation was carried out in separate on Ice. The presence of wells of a and 2) plastic ELISA dish detergent precludes washing the sample to (y- remove unbound azidoA TP. ':; The effects of pH on azidoA to GLUTI were examined by varying the pH of the the range 5. 5 - 8. 5 during A TP binding to the medium over photolabeling. MgCl was present in all experiments. Photolabeling of at 5 GLUTI was carried out at P)A TP that concentrations of 8-azido Tris TP binding were lower than the Km for transporter. In this instance , the amount of bound 8-azido(y- 32P)ATP is gIven as: Bmax . bound -(aZldoATPJ This method makes it possible to detect changes in both the the Kd for 8-azido(y- Glycoprotein P)ATP binding to GLUTl. Detection. carbohydrate was detected Detection Kit. B max and Azido(y- 32P)ATP using the photolabeled GLUT 1 Amersham ECL Glycoprotein Briefly, peptide oligosaccharides are oxidized mM sodium metaperiodate then using 10 labeled using 0. 55 mM biotin , hydrazide. Subsequent detection of biotinylated glycoprotein mploys standard western blotting techniques using streptavidin- peroxidase (1 :5 000) and ECL chemiluminescence. If!, :1' Proteolysis of Labeled Azido(y- 32P)ATP- Carrier. and/or biotin- labeled GLUTI was sedimented by centrifugation at 14, 000 x g for 20 minutes and resuspended in sample chymotrypsin , endoproteinase Lys- (protein/enzyme 20: 1 buffer :t or endoproteinase Glu- by weight). Proteolysis was allowed to proceed for 30 minutes at 37 o and was arrested by addition of phenylmethylsulfonylfluoride (PMSF , 0. 2 fragments were separated by gradient mM). The resulting GLUTI gel electrophoresis (10- 20% polyacrylamide SDS gels) and transferred to nitrocellulose membranes for autoradiography and western blotting. Polyacrylamide Gel Electrophoresis. 10% acrylamide gels where separation of electrophoresis (10- Western Blotting. as described previously Proteins were resolved on (Laemmli proteolytic fragments was required , 1970) gradient gel 20% polyacrylamide gradients) was employed. Proteolytic fragments containing GLUTI carboxyl- terminal residues 480- 492 GLUTI N- terminal residues :!:! 88- and/or a hemagglutinin tag (HA) were detected by western blot analysis. Peptides separated by SDS PAGE were transferred electrophoretic ally to nitrocellulose or immobilon membranes, which were subsequently blocked for 1 hour in PBS containing 15% Carnation TM nonfat dry milk. Following three five-minute washes in PBS, membranes were incubated for (C- Ab and N- Ab) diluted 1:1000 or (HA-ab) 1:400 2% nonfat dry milk. antibody, membranes antibody (protein- 1 hour in pnmary antibody Following three wash cycles to remove primary or goat anti-mouse Horseradish in chemiluminescence using Peroxidase PBS containing 2% nonfat dry milk. Detection of antigen/antibody/protein- was sedimented PBS containing were exposed for 45 minutes to secondary conjugate) diluted 1 :3000 Protein Sequencing. in complexes was by Amersham ECL reagents. Azido(y- 32P)ATP- Iabeled GLUTI (100 Jlg) by centrifugation at 14, 000 x g for 20 minutes then resuspended in 0. 1 % SDS/Tris- HCI pH 7.4 :! Lys- C. Digestion was carried Jlg endoproteinase out for 18- 24 hours at 4 o C. fragments were separated using GLUTI mM thioglycolate pre-equilibrated I.J gradient gels. Peptides were transferred membranes and - Coomassie electrophoretic ally to PVDF following detection by autoradiography and blue staining, bands of interest were excised from the membrane , washed extensively in distilled water and subjected to microsequence analysis by Dr. John Leczyk of the University of Massachusetts Medical School Protein Chemistry facility. Net 3- Methylglucose Uptake. Sugar- depleted cells or ghosts at ice temperature were exposed containing (3H)- to 5 volumes of 0 -methylglucose and variable concentrations unlabeled 3- 0 -methylglucose. of 15 seconds to 1 minute , of stopper ice cold KCI medium Uptake was measured over intervals then 50 volumes (relative to cell volume) solution were added to the cell/ghost suspension. , Cells/ghosts were sedimented by centrifugation (14 000 x g for seconds), washed once in stopper , collected by centrifugation, , extracted in 500 JlL of 3% perchloric acid. The acid '" centrifuged , and duplicate samples of the clear counted. Zero- time uptake of and extract was supernatant were points were prepared by addition of stopper to cells/ghosts prior to addition of medium containing sugar and radiolabel. Radioactivity associated with cells/ghosts at zero., time was subtracted from the activity associated with cells/ghosts following the uptake period. All uptakes were normalized to to sugar equilibrium uptake where cells/ghosts were exposed medium at 37 o C for 60 minutes prior to addition of stopper. Uptake assays were performed using solutions and tubes pre-equilibrated to 4 o Effects of Glycolytic Intermediates Glycolytic intermediates were introduced into human Uptake. erythrocyte ghosts by addition prior to 37 o on 3- Methylglucose to the ice-cold resealing medium membrane resealing. The membranes were warmed to C for 60 min then resealed ghosts were harvested centrifugation (22 in 50 volumes 000 x g for 20 minutes by at 4 O C), and washed once KCI medium to remove extracellular glycolytic intermediates. Ghosts were then uptake measurements processed for 3- 0-methylglucose as described above. Effects of pH on A TP-modulation 3- of Membrane ghosts were resealed uptake. 5 mM MgCl , varying (ATP) and 0. All other KCI medium containing and pH ranging from 5. 5 to stopper) were also same pH. Ghosts were then processed for 0-methylglucose uptake Reconstituted in adjusted to a solutions (uptake media adjusted to the O-methylglucose measurements as described above. sugar transport. GLUTI-mediated EggPC (40 mg in hexane) and cholesterol (10 mg) (:: 80:20 molar ratio) were dissolved in hexane. The organic phase was evaporated stream of N and remaining trace vacuo quantities of under a hexane removed for 3 hr. The resulting lipid film was dissolved in Tris medium containing 5% (by weight) sodium cholate and mixed with 100 /lg GLUTI solubilized in 50 mM sodium cholate. This mixture (4 mL) was dialyzed overnight against 6L detergent- free Tris medium and the resulting suspensIOn of distributed into 0. was frozen in a small unilamellar 5 mL fractions dry ice/isopropyl temperature then subjected to 2 proteoliposomes in microcentrifuge tubes. Each tube alcohol mixture , thawed at room additional rounds of freezelthawing. The resulting large unilammelar proteoliposomes (mean diameter = 2. by oil immersion Jlm phase microscoPY) sedimented readily at 14 000 x g and sugar transport determinations at 24 o C. was measured into at 0, 10, 20 100 JlL = 0. proteoliposome suspension (2 Jlg Tris 10 JlM GLUT!) CCB (a transport Equilibrium uptake was measured JlM). incubation for 2 hr at 24 o C. mL ice cold were used for 30MG uptake (0. 05 mM) , 30 and 60 sec in Tris medium :t inhibitor; K(app) contrast by Uptake was arrested by addition medium containing 10 were sedimented by centrifugation at 14 JlM of 1 CCB. Proteoliposomes 000 x g for 5 min at 4 o the supernatant was aspirated and the liposomes were washed agam m Tris medium containing CCB. Uptake , v = where t is (30MGJ (cpm (cpm time, cpmo, cpm t, and cpm ", associated with proteoliposomes at time v, was computed cpm cpm correspond to radioactivity zero, time t and equili bri um. Tissue Culture. Modified Eagle Cos- 7 cells were maintained in Dulbecco Medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 100 units/mL penicilin, and streptomycin in a 37 Construction C humidified 5% CO 100 ug/mL incubator. Mutant GLUTI- HA His of Wild Type Kbp fragment of GLUTI cDNA was previously A 1. subc10ned into the BamHI cloning site of pGEM3Z (pGEM3Z- GT1) (Zottola, 1995). Site directed mutagenesis via a Stratagene QuickChange Kit was used to introduce a unique Mfel restriction site at residue 1685 GTI. An HA- His in pGEM3Z- epitope tag was inserted into the coding region of ill GLUTI at this location. The presence and orientation of the HA His tag in the GLUTI coding region was confirmed by sequencing. The tagged GLUTI was then further subc10ned into the the pcDNA3. 1 + mammalian expression vector. mutagenesis was undertaken binding domains found in of Scanning alanine through the three conserved A GLUTI. Mutants were again with the Stratagene QuickChange confirmed by sequencing. BamHI site engineered Kit, and their presence was Transient 7 cells. Expression of Wild- type Subconfluent Cos- 7 and Mutant GLUT in Cos- cells were transfected with wild type and mutant GLUTI HA 6His cDNA using the Fugene 6 transfection reagent system. Following transfection , cells were maintained in DMEM for 24 hours, at which time 0. 72 mM G418 (Geneticin) was added to the media and the cells were allowed for an additional 48 hours. For harvesting, to continue growing 6 culture plates of (150 mm) were washed twice with PBS + 5 mM EDT into 20 mL lysis buffer mixed inhibitor cocktail , to Loo & Clark , and PMSF. with 5000 x g for 10 minutes at 4 centrifuged at 100, 000 x fraction. Membranes were 1 mM eukaryotic cell protease 1993). Briefly, cells of Wildtype were with a Fisher Dyna- Mix , then spun C. The supernatant was decanted g for 1 hour L of Buffer A with 1 resuspended in 300 and Mutant and to pellet the crude membrane mM protease inhibitor cocktail and frozen at - Purification A, then scraped Membranes were prepared according 1993 (Loo et aI. , homogenized for five minutes cells C until use. GLUTl. Purification 0 recombinant GLUTI HA His was described in Loo and Clark , 1995 (Loo and Clarke, 1995). Briefly, extracts were solubilized in solubilization buffer (Buffer A + 1 % N- dodecyl ~g for 15 minutes spun at 15, 000 x 1.7 mL maltoside) then to remove insoluble material. The supernatant was allowed to react for 1 hour at C with 2 mls of 4 Ni- NT A column media which was pre-equilibrated with 4 buffer maltoside). The column was B (buffer A + 0. 1 % N- dodecyl ~- washed twice with 8 mL buffer B , and twice with 8 mL buffer Samples were eluted from the column with 1. 5 mL and concentrated in centricon- l0 glucose GLUTI- HA Uptake Cos- were transfected described previously. post transfection. for 2 hours at Cells with Wild with wild- type Mutant Type cells in 12 well culture dishes and mutant GLUTI- HA His glucose uptake was measured at 72 hours Prior to uptake , cells were serum 37 C. Next, incubated at 37 o C elution buffer concentrators. Subconfluent Cos- 7 6His. C. for 15 starved in DME cells were washed twice with KRP then minutes in KRP +/- 10 exposed to 400 j.L (3H)2- deoxyglucose diluted j.M CCB. Cells were 1 :40 in KRP (10 j.Ci) mM unlabeled 2- deoxyglucose. Uptake was measured containing O. at various times , from 5 minutes washing twice with to 30 minutes and stopped by ice cold KRP. Duplicate samples 1 mL were counted following extraction with 450 III 0. 1 % Triton- XI00. zero uptake points well before were prepared by adding ice the addition of Time cold KRP to each medium containing sugar and radiolabel. Radioactivity associated with cell extracts at time zero was subtracted from the activity associated with cell the' uptake period. All uptakes were normalized extracts following ,I for protein. Uptake assays were performed USIng solutions and dishes pre-equilibrated Cos- Binding. well culture dishes as measured 72 cells were grown and transfected in 12 previously described. 8-antibody binding was hours post transfection. a-antisera reacts exclusively with exofacial epitopes of solutions were pre-warmed at 37 , Ii I'; il to 37 o 8-Antibody Ji, ' C with sheep tetrameric GLUTl. Unless noted, all to 37 C. Cells 8-antibody diluted were incubated for 1 hour 1: 1000 in DMEM +10% fetal bovine serum. Following two washes in KRP pH 7.4 , cells were !' incubated for 30 minutes with (125 I)- protein G diluted 1: 1200 in KRP (protein G recognizes and interacts with antibodies raised After two additional washes in KRP , following extraction in 500 ilL of 1 in sheep). duplicate samples were counted % triton in KRP pH 7.4. Protein G cells in the absence of primary antibody was used binding to Cos- to quantitate background radioactivity and was subtracted from each sample. All measurements were normalized to the quantity of protein present in each well. Cytochalasin Membranes from Cos- 7 B Binding. transfected cells with GLUTI HA His and mutant 338A were isolated as described previously. Samples were resuspended in 700 ilL 6. 25 11M (3H)- cytochalasin B (CCB), 100 11M phloretin , and 10 11M cytochalasin D in PBS at pH 7.4 and allowed to equilibrate on ice for 10 minutes. This suspension was placed irradiated for 4 in a plastic weigh boat on ice minutes at 280 nM in a Rayonet Reactor. Following UV irradiation membranes to remove unbound (100 CCB and sedimented by and Photochemical were washed with PBS centrifugation 000 x g for 15 min). Proteins were resolved on 10% SDS PAGE gels and stained for protein or transferred to nitrocellulose for C- and HA- Ab western blotting. Gel slices (0. 2 cm) were extracted from stained gels and incubated containing 300 ilL 30% corrected for C in scintillation fluid Hz O z before being counted. All samples were protein concentration. of Calculation overnight at 50 + and photolabeling experiments In some GLUTI ATP levels. " pH and (A TP) were varied systematically at fixed EGTA (4 mM) and MgCl (5 mM) levels. Exogenous not added , thus total Ca levels CaCl was JI were limited to Ca-contamination , I (approx. 15 tLM; Mgz. ATP , Mg (Helgerson et aI., 1989)). Computation of Mg ATP , ATP , Mg EGTA , Mgz HEGTA and EGTA levels was carried out using the WinMaxC program (Bers et aI. , equilbrium constants cited by Smith Smith , 1974) for H+, Ca and Martell (Martell and , Mg + and K+ interaction with EGTA and ATP adjusted for pH and temperature dissociation constants (- log) (4 O calculated for are: KEGTACa = 2. 91, KEGTAMg = 2. are: KEGTACa = 3. 32, 1994) and KEGTAMg = 3. C). The apparent EGT A and A TP at pH 5. , KATpca = 3. 30, KATPMg 31; at pH 6 , KATpca = 4. 30, KATPMg = 0. 19; at pH 7 1.22; at pH 8 are: GTACa = 3. , KEGTAMg = 3. 89, KATpca = 6. 29, KATPMg are: GTACa = 4. , KEGTAMg = 4. 03, KATpca = 8. 27, KATPMg = 2. 39 and at pH 8. 5 are: K EGTACa = 4. 03, KEGTAMg = 4. 04, Our calculations confirm that Mg pH range 5. 5 KATpca = 9. , KATPMg = 3. 10. + levels decline linearly over the to 8. 5 (from 2.4 to 0. 5 mM) due to increasing by EGT A and variable association with A TP. Mgz A TP plus chelation HA TP levels increase from 2. 85 mM at pH 5. 5 to a peak of 3. 5 mM at 0 and thereafter decline to 3. 2 greatest at pH 5. mM at pH 8. 5. Free ATP levels are (1.4 mM), are lowest at pH 7 (0.4 mM) and (Ji thereafter increase to 0. 8 mM at pH 8. Computation experiments , to change (e. g. of transport transport :HI and labeling constants. rates and GLUTI labeling by A In some TP were shown in a nonlinear (saturable) manner with increasing (A TP) Fig. 3). In these instances, best fit curves were generated by regression using the software package KaleidaGraph 08d (Synergy Software , Reading, P relationships: A) and assuming the following , , Fig 3A pH k2(ATP) kl 8, uptake = (ATP) dATP where kl = basal transport in the absence of A TP, k2 is the increment in saturating A TP levels and KdATP with the transport transport produced is the apparent Kd for k2(ATP) and K dATPI correspond to those k3(ATP) (ATP) K dATPI dATP2 (ATP) parameters for Fig k3 and K dATPz describe an additional interaction where kl , 3A pH 8 with A k2 but where TP that uptake. Fig 3B pH 5. 5 and 8. k5(ATP) , labeling = k4 iATP In the absence of unlabeled A TP produced by saturating A A TP that A TP interaction system. Fig 3A pH 5. , uptake = kl reduces net by , k5 is (ATP) where k4 is labeling the decrease in TP levels and KiATP inhibits labeling half-maximally. labeling is the concentration of Chapter m Factors Influencing A TP Modulation of Erythrocyte Sugar Transport The A TP-sensitivity of suggests that regulation of human erythrocyte sugar transport glycolysis and sugar transport in nucleated erythrocytes (Wood and Morgan, 1969; Whitfield and Morgan , 1983; Diamond and Carruthers , 1993) and cardiac tissue (Mansford , 1968; Haworth and Berkoff , 1986) through a is coordinated common regulatory principle. The major site of glycolytic regulation is phosphofructokinase (PFK) which catalyzes the committed step in glycolysis. PFK is subject to allosteric inhibition by A TP and by citrate and to stimulation by fructosebisphosphate and by AMP and ADP (Carpenter and Hand , Pyruvate kinase , the enzyme controlling the outflow from glycolysis is subject to allosteric inhibition by A fructose- 1986). TP and to stimulation by bisphosphate (Waygood et aI. , 1976; Rosevear et aI. 1987 ). Work in this laboratory has shown that A TP is an allosteric inhibitor of human erythrocyte GLUTI (Carruthers and Helgerson, 1989; Helgerson et aI. , 1989). The net effect of ATP on 30MG uptake II red blood cell ghosts at subsaturating sugar levels (50- 100 to stimulate transport. In contrast , levels ( 100 lM) is inhibited The sugar transport at saturating sugar (Helgerson et aI., 1989). transport experiments that defined this were measured in red M) is A TP effect blood cell ghosts (human red blood cells which have been hypotonically lysed) or inside out red blood cell vesicles (IOV' s) resealed with and without 4 mM ATP at pH 7.4 (Helgerson et aI. , 1989). Under these circumstances, normal parameters such as pH and cellular redox state are artificially controlled. This is not however the case under normal physiologic conditions. For example , sugar transport in cardiac muscle as well as nucleated erythrocytes responds to external stimuli such as muscle contraction , oxidative metabolism (Cheung and skeletal anoxia , and inhibition of et aI., 1977; Simons , 1983; Whitfield and Morgan , 1983; Haworth and Berkoff , 1986; Hansen Lee et aI., 1995). We were lacked the types of et aI., 1995; concerned that our experimental system intercellular communication (i.e. local environmental conditions) that might exist organs. This led us to ask modulation of in intact tissues and 1) What metabolic factors affect A GLUT1? 2) Does the oligomeric structure of GLUTI contribute to the A TP effect? To answer these questions , glycolytic intermediates , oligomeric structure, and changing pH were tested for their ability to alter 30MG uptake into red blood cell ghosts addition , these resealed with 4 factors were examined for their capacity azidoATP photolabeling of purified GLUTl. Our findings that intracellular AMP and acidification antagonize of sugar transport , and that A for transport modulation. that A mM A TP. In A to affect suggest TP regulation TP interaction with GLUTI is sufficient Our findings also support the hypothesis TP regulates only tetrameric GLUT1. Results Glycolytic Intermediates Glycolysis and GLUTI-mediated transport appear coordinately in certain tissues (Cloherty et aI. , 1996). to be regulated We wanted to investigate whether GLUTI-mediated sugar transport might be subject to regulatory feedback mechanisms similar to those found in the glycolytic pathway. Intracellular (but not extracellular; see (Helgerson et aI. , Mg. A TP converts 30MG transport by resealed erythrocyte ghosts from a high capacity, low affinity transport system (V max /lM/min , K m(app) = 12. 3 :t 0. system (V mn = 220 :t net effect of A 1989) =1069 :t 84 8 mM) to a low capacity, high affinity 19 /lM/min , K m(app) = 386 :t 24 /lM; n=4). The TP is to increase 30MG uptake at 0. 1 mM sugar by 2 to 5 fold (see also (Carruthers and Helgerson, 1989)). We measured 30MG uptake at 0. 1 mM sugar to determine how resealing erythrocyte ghosts with or without glycolytic basal and A intermediates affects TP-modulated sugar transport. Table 1 shows modulation of that intracellular AMP (2 mM) inhibits A TP- transport. Extracellular AMP has no measurable effect on basal or A TP-modulated unexpected becaus 30MG transport. previous work in AMP , though unable to mimic competes with A TP This result is not this laboratory has shown that the action of A i for access to the . on transport, transport system (Carruthers and Helgerson , 1989). Resealing ghosts with 2 mM ADP stimulates Table 1: Effects of glycolytic erythrocyte intermediates 30MG uptake 30MG uptake in JLmol/L lAddition cell wa ter/min 0 Mg. A TP 4 mM Mg. A TP 19:t2 49 :t 3 AMP (2 mM) 4:t 1 7:t 1 ADP ( 2 mM) 39 :t 3 no addition NAD (2 mM) no NADH (2 mM) no Pyruvate (4 mM) no Citrate (2 mM) no Lactate (2 mM) Fructose 1 6 bisphosphate (4 mM) Fructose 2 6 bisphosphate (4 mM) no HP0 (4 mM) Diphosphoglycerate (2 mM) no Phosphocreatine (1 mM) no Phosphoenolpyruvate (0. 1 mM) no no Fructose- phosphate (0. 1 mM) no no Glucose- phosphate (0. 1 mM) Glyceraldehyde- phosphate (0. 1 mM) Phosphoglycerate (0. 1 mM) Adenosine (50 JiM) no no no Erythrocyte ghosts were resealed in KCI medium containing 5 mM MgCI plus the addition shown and adjusted to pH 7.4. no indicates that there is no significant difference between the test measurement and the measurement observed with no addition. The number of measurements per condition was 3 or more. Results are shown as mean :t SEM. basal transport. This result was unexpected because ADP , like AMP competes' with A TP for binding to the transporter. We hypothesize that residual adenylate kinase present in red cell ADP to A TP causing transport ghosts converts stimulation (Carru thers and Helgerson , 1989). The remaining glycolytic intermediates and A analogs (including adenosine) are without effect on basal or A modulated transport whether applied extracellular (not shown) surface at the cytoplasmic or of the erythrocyte The experiments describing effects of glycolytic basal and A TP- membrane. intermediates on TP-modulated 30MG uptake were performed at pH 7.4. Under physiologic conditions in muscle and erythrocytes , increased glycolysis elevates cellular lactate levels resulting in acidosis (Cairns et aI. , 1993). We therefore examined the ability of altered pH modulate ATP-regulation of GLUT I-mediated sugar transport. Red blood cell ghosts were resealed 4 mM Mg. ATP adjusted to pH 5. in KCI medium , 6. , 7. 0 containing or lacking or 8. 0. Both extra- intracellular media were adjusted to the same pH to avoid pH and ; :. changes due to transmembrane proton demonstrated that insensitive to flow. Earlier studies have the extracellular sugar uptake site alterations in external pH (Sen is remarkably and Widdas , 1962). KCI both extra- and intracellular media to medium was employed in avoid transmembrane KCI gradients and K+ transport- induced volume changes. Figure 2A shows that as pH decreases from 8. 0 modulation of 100 p,M 30MG uptake to 5. , A is reduced. Subsequent experiments (see Fig 3A) show even stronger , :I inhibition of A TP- modulated transport. Basal transport rates are unaffected. Calculation of free Mg, free A TP and Mg. A TP this inhibition is unrelated to systematic in the ionization state of A of A TP-modulation of levels indicates that (pH- dependent) TP (see Materials and Methods). The loss transport at lower pH could result from altered binding kinetics (affinity or capacity of GLUTI for from the pH dependent alterations A TP) or binding of a regulatory molecule to the ATP- GLUTI complex. To examine these possibilities , was photolabeled at pH 5. , 6. , 7. , and 8. 5 with 10p,M purified GLUTI azidoATP. Previous studies from this laboratory have demonstrated that ;i! Figure 2. Effects of GLUT1- labeling pH on A TP-modulation of sugar transport (A) and by azido-(y_ P)A TP (8). A Erythrocytes were Iyzed and resealed in KCI medium containing (8) or lacking (0) 2 mM exogenous ATP. The pH of the resealing and extracellular media is shown on the abscissa. Ordinate: initial rates of 100 ,uM 30MG uptake (,mol/L cell water/ min). These results summarize triplicate measurements. Lines drawn through the data points were computed by the method of least squares. 8 Purified GLUT1 was incubated in Tris- HCI medium containing minutes at 4 . C. Each sample was azido-(y- P)A TP at the pH indicated for 15 10,uM uv irradiated for 1 min and 30,ug of labeled protein resolved on 10% SOS PAGE minigels. Proteins were transferred to ;J nitrocellulose and labeled bands containing azido(y- P)A TP were identified by autoradiog raphy Figure 2 120 100 5 7. 5 8. GLUTI and azidoA TP associate and that azidoA A TP , ADP , and reversibly to form a binary complex TP binding to GLUTI is inhibited AMP (Carruthers and Helgerson , competitively by 1989). Proteins were resolved on minigels and labeling was quantitated autoradiography. Figure 2B shows that contrary ATP incorporation into GLUTI is by to our expectations as the pH increases from reduced to 8. To better understand this response to A TP- dependence of 30MG uptake in resealed 5 and 8. 0. Figure 3A a saturable however , shows that 100JLM pH , we measured the red cell ghosts at pH 30MG uptake increases in manner with ATP concentration at pH 8. 0. At pH 5. transport is maximally stimula ted inhibited at 2 at 200 JLM mM A TP. Figure 3B summarizes three experiments in which azido(y- out at pH 5. 5 and 8. 0 unlabeled A TP. A A TP , but separate P)ATP labeling of GLUTI was carried in the presence of increasing concentrations of TP inhibition of GLUTI labeling by azidoA TP shows simple competition kinetics at both pH 5. 5 and 8. 0 but maximum labeling by azidoA TP and (app) for A TP inhibition reduced at pH 8. 0. The A TP content of of labeling are erythrocyte ghosts resealed 3. Effects of pHon the concentration- dependence of A TP- Figure modulation of sugar transport (A) and ATP- inhibition of GLUT1- azido-rf P)A TP (B). A. Erythrocytes labeling by were lysed and resealed in KCI medium containing 5 mM MgCI plus increasing concentrations of ATP at pH 8. 0 (8) or at pH 5. 5 (0). Ordinate: min). Abscissa: A initial rates of 100 tlM 30MG uptake (,mol/L cell water/ TP content of the resealing medium in tlM. The results summarize three separate experiments each made in triplicate. Lines drawn through the data points were computed by nonlinear regression (see Materials and Methods). These analyses assume that ATP-modulation of transport at pH 8 is comprised of basal transport (k1 Michaelis- Menten component where seen at 0 ATP) plus a simple , ATP- dependent k2 is the increment in transport produced by saturating A TP levels and K dATP is the apparent Kd transport system. The computed constants are: k1 dATP = 89 tlM for A TP interaction with the = 46 tlM/min = 30 tlM/min; ATP. ATP-modulation of transport at pH 5. 5 is similar to that at pH 8 but is characterized by an additional interaction with A uptake where k2 k3 TP that reduces net is the decrement in transport produced by saturating A TP levels and K dATP2 is the apparent K for A TP inhibition of transport. The computed constants are: tlM/min; K dATP2 k1 = 199 = 48 tlM/min tlM k2 = 259 tlM/min; KdATP1 = 189 tlM ATP k3 = 259 ATP. B. ATP- inhibition of GLUT1- photolabeling by azido-(y- P)ATP at pH 5. 5 (0) and pH 8. 0 (8). Purified GLUT1 was incubated for 15 minutes at 4 o C in Tris- HCI medium containing 5 mM MgC/ . 15 tlM azido- (y_ PJATP at pH 5. 5 or pH 8. 0 in the presence of increasing concentrations of ATP. Each sample was uv irradiated for 1 min and 30 Jig of labeled protein resolved on 10% SDS PAGE minigels. Peptides were transferred to nitrocellulose and labeled bands containing (y_ PJA TP were identified by autoradiography. Ordinate: relative (y_ PJATP incorporation into GLUT1 as judged by densitometric analysis of resulting autoradiograms. Abscissa: (A TPJ present during photolabeling (mM). These data summarize 2 separate experiments. Lines drawn through the points were computed by nonlinear regression (see Materials and Methods). This analysis assumes that k4 is the extent of labeling in the absence of ATP and that ATP competitively inhibits labeling iATP (k5 is the decrease in labeling produced by saturating A TP levels and is the concentration of A TP that inhibits labeling half-maximally). The computed constants are: pH 8. iATP = 0. 5:t iATP = 1. 3 :t 0. 5 mM ATP. 0. 2 mM ATP; pH 5. 5, = (1. 1 :t 0. 1) x 10 k4 k4= (4. 7:t 0. 3) x 10 k5 k5= = (1. 0 :t 0. 1) x 10 (4. 1 :t 0. 5) x 10 -- Figure 3 110 C\ 1 00 I. 1000 2000 (ATPJ 5 1 0 41 3 1 0 CL 2 1 0 :6 pH 5. 1 1 0 pH 8. o 10P (ATPJ mM with 4 mM ATP at pH 5. 5 or 8. 0 is not significantly different (pH 5. (ATP) = 3. 1 mM; pH 8. , (ATP) = 3. 1 mM; ATP content of ghosts resealed in the absence of exogenous A TP: pH 5. 5, (A TP) = 0. 07 mM; pH 8. , (A TP) = 0. 069 mM). Oligomeric Structure Erythrocyte-resident GLUTI exists as a dissociates to form a homodimer upon extracellular reductant (Hebert 1995). In homotetramer but erythrocyte exposure to and Carruthers , 1992; Zottola et aI. the present study we ask: does A TP modulate both dimeric and tetrameric GLUTI-mediated sugar transport? Table 2 shows that erythrocyte ghost exposure to reductant (2 mM DTT) inhibits modulation of 30MG uptake but is without effect on basal A TP- uptake measured in the absence of A TP. Reduced capacity to modulate GLUTI-mediated sugar transport in reductant- treated erythrocyte ghosts could result from reduced ATP binding to GLUTI or from altered interaction of A TP- liganded GLUTI with regulatory cellular components. Figure 4 shows that incorporation of 8-azido(y- 32P)A TP into reduced GLUTI is significantly impaired. :.',' E.:7YJ' Table Effect of reductant of erythrocyte TP-modulation 30MG uptake 30MG uptake Condition Control 2 mM DTT 2 mM A TP 2 mM A TP + 2 mM DTT pmollL cell water/min 41.2 :t 7.4 57. 3 :t 7. 97. 7:t 11.6 48. 1 :t 3.4 ::1 Hypotonically Iyzed erythrocytes were resealed :t 2 mM A TP then incubated :t 2 mM OTT for 30 min at 37 . C. 30MG uptake at 100 JiM 30MG was measured at 4 C. The table summarizes results as mean :t SEM of 3 separate experiments made in triplicate. Figure 4. Azido-(y- P)A TP labeling of dimeric and tetrameric GLUT1. Purified GLUT1 was incubated in Tris- HCI medium + 5 mM MgCI :! 4 mM OTT for 30 minutes at 4 o C. Following labeling with azidoATP 30/lg samples of protein were resolved on 10% SOS PAGE minigels. Peptides were transferred to nitrocellulose and bands containing (y- P)A TP were identified by autoradiography. The presence (+) or omission (-) of OTT treatment is indicated above each lane. , I #&.. Figure 4 105 82 49 33. 28. 19. DTT Does ATP modulate GLUTl mediated sugar transport directly? Reconstitution of purified erythrocyte (EggPC and cholesterol in an lacking 2 GLUTI into liposomes 80:20 molar ratio) containing mM Mg. A TP provides an in-vitro system that solely of GLUT1 , A TP , and the lipid transport into cytochalasin B (10 consists bilayer. Inhibition of sugar GLUTI proteoliposomes JlM and following treatment with CCB) insures that we are measunng protein mediated solute movement across the lipid bilayer as opposed non-specific leakage. Cytochalasin B (CCB) inhibitable 3\, methylglucose uptake by reconstituted GLUTI proteoliposomes is I), robustly stimulated by 2 mM Mg. A TP (Figure 5). Discussion We examined whether A TP regulation of transport in resealed ghosts is subject to modulation by glycolytic intermediates known to regulate glycolysis. Only AMP antagonizes 30MG uptake. Citrate , fructose- bisphosphate (key regulators of the action of A TP on bisphosphate and fructose- PFK or pyruvate kinase) basal or A TP stimulated sugar uptake. Other are without intermediates Figure proteoliposomes. 5. ATP stimulates 30MG uptake into reconstituted , purified GLUT1 GLUT1 (100 Jig) was reconstituted into eggPC:cholesterol proteoliposomes and uptake of 50 JiM 30MG was measured at 20 . C in the presence of 5 mM MgCI :t 2 mM ATP , and in the presence or absence of 20 JiM cytochalasin B (CCB). Ordinate: 30MG uptake expressed as % of the equilibrium 30MG space of proteoliposomes accessed during 15 sec of uptake; Abscissa: (ATPJ of medium during reconstitution and uptake determination (0 or 2 mM ATP). Uptake in the presence of CCB is shown by the solid bars. This figure is representative of three separate experiments. Figure 5 OATP 2 mM ATP were similarly inactive. Adenosine also failed to modulate GLUTI mediated sugar transport. This suggests that A TP is not interacting with the adenosine receptor , nor is the adenine ring responsible for the effects of A TP on GLUT1. ADP , however , mimics the ability of A TP to modulate transport. These effects of AMP transport in erythrocyte ghosts are similar previously (Carruthers ADP inhibits A TP reported to those and Helgerson , 1989). External application of modulation of sugar efflux from inside out erythrocyte membranes vesicles and , like AMP , action of A and ADP on sugar TP which reduces C- Ab antagonizes the binding to purified GLUT 1 (Carruthers and Helgerson , 1989). This suggests that ,adenylate kinase present in erythrocyte ghosts converts " This hypothesis is supported by residual ADP to A TP. A TP assays of ADP- Ioaded ghosts o.' , (Carruthers and Helgerson , 1989). j, regulation of sugar transport We propose , is subject to therefore , that A TP inhibition by AMP and possibly by ADP. The implication is that the glucose epzyme of glycolysis of the cell. transporter - is characteristically sensitive to the energy (y- All experiments intermediates on A performed at fixed investigating the action of glycolytic TP modulation sugar transport of were pH (7.4). Under normal circumstances , glycolytic flux gives rise to elevated lactic acid levels and pH (Elliott et aI. , 1992; Cairns et aI. , 1993; Smith increased reduced et aI. , 1993). therefore examined the effect of altered pH on basal and A regulated sugar transport. To our surprise , we observed that while reduced pH inhibited sugar transport A TP-modulation of azido (y- P)A TP- Iabeling to increase both (app) for of GLUTI was increased. Reduced pH serves maximum photoincorporation A TP- inhibition of photolabeling. transport at pH 8. increases in a concentration. However , A Increases with nucleotide P)A TP and of A TP stimulation of saturable manner with nucleotide TP stimulation of transport at pH 5. concentrations up to 200 11M and thereafter falls with increasing A TP. A TP assays indicate that intracellular A TP is not low pH. Computation of degraded to a free Mg , A significantly greater extent at TP and Mg. A that these observations are not explained dependent changes in TP levels indicates by systematic, pH- these species. We conclude that intracellular A TP exerts antagonistic actions on sugar transport at pH 5. nature of the inhibitory action is presently 5. The unknown. Increased GLUTI labeling by azidoA TP at reduced pH raIses the possibility that a unIque, transport-antagonistic A TP binding acidic pH. Future studies must address this TP results in sugar uptake revealed at possibility. Isolated GLUTI is an A TP binding protein. ghosts with Mg. A site is Resealing erythrocyte stimulation at subsaturating (30MG) and produces uptake inhibition at saturating (30MG) (Carruthers Mg. ATP and Helgerson , 1989). Addition of 2 to reconstituted GLUTI proteoliposomes results in stimulation of 30MG uptake at sub saturating (50 tLM) 30MG levels. ATP modulations of purified GLUTI and erythrocyte resident GLUTI mediated sugar transport are , therefore , fundamentally conclude that ATP , GLUT1 , similar. and the lipid bilayer , are sufficient for A TP modulation of GLUTI function. This conclusion contrasts with that of Wheeler who demonstrated that A TP does not modulate reconstituted human GLUTI protein Wheeler the activity (Wheeler of 1989). However studies employed reduced (DTT- treated) GLUTI which we have shown to exist as a GLUTI dimer (Hebert and Carruthers 1992). Our studies employ the GLUTI (a GLUTI tetramer , physiologic , non-reduced form of see (Hebert and Carruthers , 1992). therefore asked two questions. 1) A TP-regulation? 2) Is reduced GLUT susceptible to Does reduced GLUTI bind A TP? Our studies shown that reductant treatment of erythrocytes converts resident tetrameric GLUTI to dimeric GLUTI (Zottola et aI. , present findings indicate have 1995) and our that reductant also ablates A TP- modulation of 30MG transport. Reductant (DTT) treatment of purified GLUTI results in the loss of azidoA TP labeling. results provide a rational explanation for Wheeler s findings and suggest strongly that dimeric to nuc1eotides. These GLUTI does not present binding sites They further reinforce the view that red cell-resident GLUTI exists as the A TP sensitive , GLUTI tetramer (Zottola et aI. 1995 ). Chapter IV the GLUTl Peptide Mapping ATP Binding Domain Isolated human erythrocyte glucose transporter (GLUT1) functions as an A 1989). TP binding protein (Carruthers and Helgerson Work in this laboratory has shown that the GLUTI carboxyl terminus becomes less accessible to peptide- directed TP exposure. following A This A TP dependent is antagonized by AMP and ADP that appear competitive inhibitors of A out to determine antibodies conformational change to function as TP binding to GLUTl. In this study we set the location of the GLUTI nucleotide binding domain. To answer this question we applied peptide mapping N-terminal sequencing techniques azidoATP , a photoactivatable suggest that GLUTI residues A TP binding domain. to and GLUTI photolabeled with P labeled , ATP analog. Our findings 301- 364 form an integral part of the ' ' Results Peptide Mapping Previous studies from this laboratory have demonstrated that GLUTI and azidoA TP associate and that azidoA reversibly to form a binary complex TP binding to GLUTI is inhibited ATP , ADP and AMP (Carruthers and Helgerson of the azidoA competitively by , 1989). UV TP/GLUTI complex activates the A TP azido irradiation group sulting in covalent attachment of bound nucleotide to GLUTI. In the absence of irradiation , incorporated into the no measurable carrier protein (Carruthers and 1989). Limited proteolysis photoattachment of azidoA 1989). Helgerson studies suggest that the point TP to GLUTI falls witbin the carboxyl- terminal half of the protein (residues 270- Helgerson , 8-azido(y- P)ATP is 456; see (Carruthers We have now undertaken mapping analysis to more precisely define ttachment of nucleotide to and a detailed peptide the site of covalent GLUTl. Peptide mapping studies of integral membrane proteins such as GLUTI are complicated by the typically hydrophobic nature of the resulting peptides and by the presence of N- linked oligosaccharide. For example , human erythrocyte band 7 proteins include the multi-spanning integral membrane glycoproteins aquaporin (28 kDa) and RhD (Mr = 45 kDa) (Smith and Agre , 1991) but are resolved as a 32 kDa protein band upon 10% SDS PAGE (Zottola et aI. , 1995). Intact GLUTI (Mr = 55 kDa) is resolved kDa band (Fig 6A). We therefore decided markers to to use as a 50 GLUTI internal assist in peptide identification. Peptides containing linked oligosaccharide (asn ) were identified by blot analysis streptavidin-conjugated HRP) of GLUTI fragments obtained N(using from glucose transporter biotinylated on carbohydrate residues prior to proteolysis. Pep tides containing GLUTI residues 88- and/or 480- 492 (C-residues) were identified by use antisera (N- Ab and C- Ab peptides corresponding to 95 (N-residues) of rabbit respectively) raised against synthetic these GLUTI domains. Each experiment (see Figure 6A) thus involves biotinylation by unsealed GLUTI proteoliposomes , photolabeling using azido(y- 32P)A TP. of carbohydrate exposed followed by The labeled GLUTI proteoliposomes (y_ Figure 6. Peptide mapping of photolabeled A. Purified GLUT1 (40 j.g GLUT1. biotinylated on carbohydrate residues) was photolabeled using 10j.M azido- fy_ P)A TP then incubated in sample buffer containing 5 j.g chymotrypsin (30 minutes at 37 . C). Peptides were separated on a 10- 20% SOS PAGE gradient gel , transferred to nitrocellulose. Azidofy- P)A TP-containing peptides were detected by autoradiography e p). C- Ab and N- Ab reactive peptides were detected using immunoblot analysis. Biotinylated carbohydrate (Glyc) was detected by streptavidin blot analysis. C- , N- , 32 p and Glyc analyses were performed on a sihgle common lane of electrophoretically separated peptides. Major peptides are assigned a label (a through h) and are indicated by the leftmost solid bars. The mobility and M of molecular weight standards are indicated by the solid bars to the right of the figure. B. EndoLys- C digestion and sequencing of azido-fy- P)ATP- photolabeled GLUT1. Labeled , purified GLUT1 (100 j.g) was incubated in 0. Tris-HCI medium , pH 7.4 containing 5 j.g endoproteinase Lys- j.g endoproteinase Glu- C (EndoGlu- C; separated on a 10- 20% 18- 24 hours at 4 O C). 1 % SOS , 50 mM C (EndoLysC) or 5 Peptides were SOS PAGE gradient gel , which had been pre- equilibrated with 1 mM thioglycolate and subsequently transferred to ImmobilonPVOF for staining with Coomassie blue. Bands containing P)ATP were , identified by autoradiography e p). C- Ab and N- Ab reactive peptides were detected using chemiluminescence immunoblot analysis. Major peptides are assigned a label (i through 0) and are indicated by the leftmost solid bars. The mobility and M of molecular weight standards are indicated by the solid bars to the right of the figure. An intensely labeled 19 kDa band (peptide n) was excised from the membrane , washed 10 times with distiled water and subjected to sequence analysis by the UMMC Microsequencing Facility. A single sequence was obtained as AGVQQPVY A T corresponding to GLUT1 residues 301- 310. Figure 6 Ab N- Ab Glyc 33. 28. 19.4 32p 33. 28. 19.4 EndoLys- EndoGlu- were subjected PAGE , to limited transferred to and/or 32P- Iabelled proteolysis, peptides are resolved by SDS- immobilon or nitrocellulose , and biotinylated peptides are detected by autoradiography. same membrane filters are then probed for the presence of The peptides containing N- and C-residues by chemiluminescent immunoblot analysis. Figure 6A indicates that not all N- or contain 32p or biotinylated carbohydrate. Similarly, containing peptides include (V8) digests labeled GLUTI which were subsequently One of the was subjected not all 32p- N- or C- Ab-reactive sequence. Figure 6B shows EndoLys- C and EndoGlu- C N-residues. C- Ab-reactive peptides of 8-azido(y- 32P)ATP probed for 32p and C- and 32P- Iabeled peptides (figure 6B , peptide n) to N- terminal sequence analysis summarizes additional experiments intact GLUTI was subjected to (see below). Figure 7 in which azidoA TP- Iabeled digestion by chymotrypsin (Figure 7 A) or by V8 protease (Figure 7B). Major peptides of Figures 6 and 7 have an assigned label and are aligned in Figure 8 according to size sequence content specificity. (identified by internal markers) and proteolytic This analysis permits approximate localization of the (y- 7. Peptide mapping of photolabeled GLUT1. A. Purified Figure Jig) was photolabeled using 10 JiM azido-(y- GLUT1 (40 P)A TP then incubated in sample buffer containing 5 Jig chymotrypsin (30 minutes at 37 . C). Peptides were separated on a 10- 20% SOS PAGE gradient gel , transferred to nitrocellulose. P)ATP-containing peptides were detected by autoradiography e p). N- reactive peptides were detected using immunoblot analysis. Major peptides are assigned a label (p through u) and are indicated by the leftmost solid bars. The mobility and M of molecular weight standards are indicated by the solid bars to the right of the figure. B. V8- protease digestion of azido-(y- P)A TPphoto/abeled GLUT1. Labeled , purified GLUT1 (100 Jig) was incubated in 0. 1 % SOS , 50 mM Tris- HCI medium , pH 7.4 containing 8 or 15 Jig endoproteinase Glu- C (EndoGluC; , 18- 24 hours at 4 . C). Peptides were separated on a 10PAGE gradient gel and subsequently transferred to Immobilon- 20% SOS PVOF for staining with Coomassie blue. Bands containing (y_ P)ATP were identified by autoradiography e p). C- Ab reactive peptides were detected using chemiluminescence immunob/ot analysis. Major peptides are assigned a label (v through bb) and are indicated by the leftmost solid bars. The mobility and M molecular weight standards are indicated by the solid bars to the right of the figure. "' - .- Figure 7 p N- 105 f - 33. - 28. - 19. Chymotrypsin 105 -82 33. -28. aabb- va protease (g) 19. , (y_ Figure 8. Summary of peptide mapping analysis of (y- P)ATP- labeled GLUT1 (Figures 6 and 7). The horizontal axis represents GLUT1 residues 492. 1 to labeled and unlabeled peptides are indicated by filled and open bars respectively. Data are from Figures 6 and 7 and are cross-referenced by peptide labels (a - bb). Peptides containing C- Ab reactive sequence overlap with residues 480 - 492. Peptides containing N- GLUT1 residues 88 - GLUT1 Ab reactive sequence overlap with 95. Peptides containing biotinylated carbohydrate also contain GLUT1 residue asn . When two peptides are linked by a single vertical bar , this indicates that localization is ambiguous and the range of possibilities is shown for the linked peptides. The length of each peptide was computed by assuming that electrophoretic mobility accurately represents peptide molecular weight. The shaded vertical box spanning GLUT1 residues 291- 364 represents that region of GLUT1 that most likely contains covalently attached (y_ The striped bar represents a 19 kDa (peptide n) of GLUT1 whose N- terminal PJATP. PJA TP- Iabeled , endoLysC fragment residue (A 301 ) was confirmed by sequencing. Peptides e and y should (but do not) contain (y_ pJA TP and are shaded differently for comparison. Presumed specificities of proteolysis are: Endo- LysC , K; V8 protease , 0 , E; chymotrypsin ,F //,/ ;,, / / / ".- , ;,.. Figure 8 45 88 95 291 364 nv / 480 492 / //1 :J;,, .-J ':1 YI 45 88 95 291 364 480 492 (y_ site of covalent attachment 364. Recognizing of P)A TP to residues 291 through that the masses of hydrophobic and glycosylated proteins obtained by SDS- PAGE analysis of nucleotide attachment should be are approximate , this regIOn viewed with some caution. Two peptides have properties inconsistent with this conclusion. Peptide e (Figure 6A) and peptide y (Figure 7B) cross react with C- , are characterized by Mrapp 19. 9 no and 22.4 kDa respectively but contain p. Three observations suggest that these peptides are detectable dysfunctional GLUTI fragments present in GLUTI proteoliposomes pnor to labeling and digestion. 1) These C- Ab-reactive peptides observed occasionally in GLUTI preparations proteolytic enzymes (Hebert and Carruthers , Carruthers , unpublished correspond to any Peptide n are that are untreated by 1992); Cloherty and observations). 2) Peptide does not possible V8-specific GLUTI cleavage product. 3) overlaps with peptides e and y, contains 32p and identified definitively by N-terminal sequencing (see below). (y- N- Terminal Microsequencing To further refine definition of the site of nucleotide attachment azido (y- P)A TP labeled GLUTI was subjected to digestion. Endoproteinase Lys- Lys- ester , endoproteinase specifically hydrolyzes amide and peptide bonds at the carboxylic side of lysine (Jekel et aI. 1983). Assuming the conditions of proteolysis reveal each cleavage site and also favor EndoLys- produce a 16 kDa 456. Purified , GLUTI fragment azido- A TP labeled for 18- 24 hours at 4 o 10- 20% specificity, this enzyme is expected to corresponding to GLUT 1 was exposed to endoLys- C and the resulting peptides were resolved the blots were exposed to stained with Coomassie blue. Bands membrane , and of film for autoradiography and interest were excised from the subjected to N- terminal microsequence analysis. Figure 6B indicates that a 19 kDa band (peptide n) is heavily labeled P)ATP but lacks recognition sites for both N- terminal sequence analysis of and C- Ab. N- this peptide yielded 10 pmol of a single sequence of AGVQQPVY A 301- 310. on gradient SDS- PAGE gels. Following transfer to Immobilon- membranes , by residues 301 - T corresponding to GLUTI residues Discussion The results of our 291- 364 as peptide mappIng studies point to GLUTI residues the smallest GLUTI cassette attached azido (y32PJ- A TP. This that contains covalently result is consistent with previous indirect studies suggesting that GLUTI residues involved in nucleotide binding (Carruthers 270- 380 and Helgerson , 1989). The present study offers significant improvement analyses because we have employed internal approximate origin of assumed residues), GLUTI residues linked carbohydrate (asn 480- markers to indicate the assigned on residues 84- 96 (N- 492 (C-residues) and asparagIne- ). Even USIng these markers , however GLUTI peptide mapping studies the frequently anomalous mobility of is complicated by GLUTI fragments upon SDS :PAGE. For example , an azido(yP)- ATP , fragment with intact carboxylterminus Jragment. EndoLys- earlier proteolytic enzyme specificity alone. These markers are: GLUTI interpretation of over our proteolytic fragments previously the basis of peptide mass and may labeled endoLysC GLUTI is resolved as a 23 kDa specifically hydrolyzes amide , ester , and peptide bonds at the carboxylic side of in proteolysis (an assumption lysine. Assuming specificity supported by sequence analysis smaller endoLysC GLUTI fragment), this peptide could cleavage at lysine 300 or , peptides (21 of a derive from lysine 256. The predicted masses of these kDa and 25. 8 kDa respectively) differ significantly from the observed mass on SDS PAGE. The smaller azido(yfragment (19. 1 kDa) was P)- ATP labeled endoLysC GLUTI subjected to N- terminal sequence analysis. I: This confirmed the origin of the azido(y P)- ATP fragment as labeled GLUTI ala 01 with hydrolysis occurring at lys 30o specificity and precise places the site of i: . Assuming molecular weight determination , this result azido(y P)- ATP incorporation within a GLUTI cassette comprised of residues weight = 19223 D). 301- 477 (theoretical molecular Our peptide mapping studies suggest that azido(y P)- A TP incorporation occurs between and 364. This permits further refinement GLUTI residues of the azidoA TP- incorporation domain to residues 301 - 364. This GLUTI contains primary sequence that shares 291 domain 50% identity with a region of adenylate kinase previously shown to form a hydrophobic strand of parallel B- pleated sheet flanking the adenine-ribose and triphosphate chain of MgA TP (Fry et aI. , 1986). This sequence adenyl ate kinase GLUT GQPTLLL YVDAG Izl GRRTLHLIGLAG 33z 343 IS also found in the adenine nucleotide binding proteins phosphorylase (Newgard et aI. , 1988), glycogen RAD3 (Reynolds et aI. , the ATP/ADP translocase (PIano and Winkler , 1991), the brown fat uncoupling protein (Casteilla et aI. , 1989; Cassard 1990), phosphofructokinase (Valdez et aI. , subunits of Fj ATPase (Ohta and Kagawa , GLUTI domain to the glucose while strengthened of a of this remaInS detailed GLUTI structure. It is contains the consensus which is found in a may be exposed Walker number of A TP and GTP binding proteins (Fry et aI. , 1986). In GLUTl , (residues 111- 118) et aI. the (X- & 1986). Assignment , mappIng studies , interesting to note that GLUTI also motif GXXXXGKS/T , human transport nucleotide binding pocket by our peptide inferential in the absence 1989) and 1985), this sequence to the exterior of the cell (Hresko et aI. , 1994) suggesting that it intracellular nucleotide binding. assumes no role in Chapter Alanine mutagenesis of the GLUTl consensus ATP binding sequences Peptide mappIng and n- terminal sequencIng studies laboratory have suggested that residues 301- the GLUTI ATP binding domain. pnmary sequence that is This in this 364 comprise part of region of GLUTI contains 50% identical with a region of adenylate kinase previously shown to form a hydrophobic strand of a parallel pleated sheet flanking the adenine ribose and triphosphate of Mg. ATP (Fry et aI. , sequences have been 1986). Two additional consensus ATP binding identified in GLUTI (Fry et aI. , to delineate further the location of the A we have the data also show that transporter function. three proposed regions suggest that residues 332- the efficiency of adenine nucleotide 1986). In order TP binding site on GLUTI mutated all of the residues in these to alanine. Our findings chain binding to residues 338- 340 335 contribute to GLUTl. In addition are necessary for normal Results Purification of Human GLUTl HA His GLUTI is a glycosylated membrane protein that contains linked oligosaccharide a single at asn . The presence of this carbohydrate makes it difficult to synthesize GLUTI in a bacterial system. We felt that either a yeast system fulfill our needs. or a transient mammalian system would Our data suggest that we were fairly successful in expressing hGLUTI in a strain of Saccharomyces Cerevisiae (RE700A) in which seven (hxt 1- 7) of 20 putative glucose transporter genes are disrupted of yeast is unable to Figure 9B). Following or deleted. Unmodified , this strain survive on standard glucose media (Table transfection with human GLUT1 , clones are able to grow on source (Figure 9A & 9B). and selected transport glucose as a sole In addition , separation of 3 carbon whole cell extracts for western blotting reveals a 55 Kd protein that reacts with GLUTI C- Ab (Figure 10). The data suggest that we have successfully, heterologously, expressed human GLUTI in yeast. Unfortunately, this mutant yeast grows too sl wly and produces insufficient GLUTI for our purposes. Table 3: Characteristics of S. Cerevisiae strain RE700A transfected with Yeast Carbon Source Wild Type Glucose Wild Type Galactose Wild Type Maltose RE700A Glucose RE700A Galactose RE700A Maltose RE700A + GLUT1 Glucose RE700A + GLUT1 Galactose RE700A + GLUT1 Maltose GLUT1 Glucose Uptake Growth S. Cerevisiae 2 RE700A is aURA 52 mutant with 7 of its glucose transporters (HXT 1deleted or disrupted. RE700A transfected with HA His epitope tagged human erythrocyte glucose transporter (GLUT1). Figure 9. GLUT1 HA His expression in Saccharomyces Cerevisiae. A. H)0- glucose uptake measured in RE700A +/- tagged GLUT1. Yeast were grown to an 600 = 0. 6 for uptake. 1- 2 mL of cells were pelleted and washed in PBS prior to addition of uptake media. All solutions except stop were preheated to 37 C. Uptake of 100 /lM O- glucose was terminated at 5, 30, and 60 seconds by washing with ice cold PBS pH 7.4. Following lysis , duplicate samples of supernatant were counted in a Beckman scintillation counter. B. RE700A +/tagged GLUT1 grown on YP-maltose or minimal media + glucose. Figure 9 + GLUT1 HA HIS - GLUT1 HA HIS :: 350 -- 250 150 Null Null YP- Maltose Minimal Media + Glc +/- Figure 10. C-terminal GLUT1 antibody western blot of RE700A GLUT1 HA His Cell extracts were prepared from RE700A +/- tagged GLUT1 grown to an OD 600 of 0. 6. Prior to extraction , the yeast were washed twice with ice cold 50mM Tris- HCI pH 7. 5, 10 mM NaN . Cells were combined with 0. 2 mM glass beads and 100 JlI of protein sample buffer containing 1 mM PMSF. The mixture was vortexed briefly and 50 JlI of each sample was resolved on 10% SDS PAGE gels that were transferred to nitrocellulose for western blotting with GLUT1 terminal antibody. ;: .: Figure t(5 q, e (r 105 33. 28. 19. ,-, We therefore mammalian transient decided to develop a eXpreSSIOn system for GLUTl. Cos- cells are a simian kidney cell GLUTl. line that is capable of correctly processIng 1985; Schurmann et aI. , (HA), 6 histidine (His 1997). We introduced (Mueckler et aI., a hemagglutinin )' tag to facilitate the purification of the expressed human GLUTI (GLUTI HA His ). To accomplish this, a unique Mfel restriction site was introduced at residue Kbp fragment of GLUTI cDNA previously subc10ned into the site of pGEM3Z (pGEM3Z- GT1) (Zottola et aI. , insertion of an HA His further subcloned expression vector. the His into the epitope tag at this BamHI 1685 in a 1.7 BamHl 1995). Following the site, the GLUTI was site of a pcDNA3. 1+ mammalian The benefits of attaching this tag are twofold; region binds Nickel- NT A allowing affinity purification recombinant protein and , 2) HA can be detected with standard western blotting techniques , allowing us parental Cos- 7 of to differentiate between GLUTI and tagged human GLUTI (GLUTI HA His To purify hGLUT1 , Cos- 7 cells were transfected with pcDNA3. 1 GLUT 1 via the Fugene 6 transfection Cells were harvested and homogenized reagent system. 72 hours post transfection and membranes were isolated by high-speed centrifugation in Beckman TI50. 2 rotor at 100 000 x g (Loo and Clarke , 1994). the hydrophobic nature of hGLUT1 , detergent , n- dodecyl-~ D maltoside , purification because ionic to solubilize the it is necessary applied to the Ni- NTA. membranes before they are detergents of the protein to bind to the A non- ionic was used to prepare samples for may interfere with the ability column (Loo and Clarke , solubilized protein was allowed Due to 1995). The to bind to the nickel matrix for one hour with gentle agitation. The Ni- NTA column was washed with increasing concentrations were eluted with a of imidazole (10- 25 mM), and samples solution of 500 mM NaCI and 300 mM imidazole. Fractions were resolved on 10% SDS- PAGE gels and transferred to nitrocellulose membranes for immunoblotting. Standard western blotting techniques were used GLUTI present on these membranes. Gels stained suggest that GLUTI HA His pure. Rabbit antisera corresponding to to identify the with Pro- blue containing eluate is approximately 50% raised against a synthetic peptide hGLUTI residues 480- 492 (C- Ab) was used to locate and identify intact GLUTI in various purification fractions. Human GLUTI and Cos- 7 GLUTI show significant homology, thus C- terminal Ab binding is indicative of total GLUTI concentration. To quantitate hGLUT1 , we used anti- HA antisera which binds selectively to epitope tagged hGLUTI. Figures llA & B show both the resulting terminal and HA western blots from a standard GLUTI HA His purification protocol using Cos- cell membranes. A single band with an approximate molecular weight of 55 large KDa is visible in the final elution fraction (eluate (E) on both blots. The presence of both an intact terminus and an HA tag suggests that we have C- successfully affinity- purified His tagged hGLUTI protein. Alanine Scanning Mutagenesis of Fry et al (Fry et aI. , potential consensus A ATP Consensus 1986) showed TP binding Binding Sequences that GLUTI contains sequences (Figure 1). three Current GLUTI topology would place consensus site I (Gresidues 111- 117), within transmembrane II (K- X- V - domain 3. amino acids 225- 229) lies within the loop between domains 6 & 7. S/T Finally site III (G- Consensus site large cytoplasmic L (Y , I) X- Figure 11. Purification of GLUT1 HA His from Cos- 7 cells. HA- Ab (A) and G- Ab (8) immunoblot analysis of samples of GLUT1 HA His transfected cell membranes obtained during GLUT1 affnity purification on Ni- NT A. Isolated Gos- 7 cell membranes were solubilized in n- dodecyl-~- O-maltoside (lane 2) and were applied to aNi- NT A column. Following flowthrough of non bound GLUT1 (lane 3), columns were washed with 300 mM NaGI plus 10 mM imidazole (lane 4), 300 mM NaGI plus 30 mM imidazole (lane 5) and 500 mM NaGI plus 300 mM imidazole (lane 6). Fractions were collected and proteins were separated on 10% SOS- PAGE gels and transferred to nitrocellulose for immunoblot analysis. A. The presence of GLUT1 in each of the samples was detected with G- Ab. Purified human erythrocyte GLUT1 was placed in the far left lane as a control. GLUT1 runs as a diffuse band with an approximate molecular weight of 55 kOa. Lower molecular weight bands are proteolytic fragments of GLUT1. B. GLUT1 HA His was identified with a mouse anti- HA antibody. Fractions correspond to and are aligned with those in panel A. The mobility of molecular weight standards are indicated by the solid bars to the left of the panels A and 8. ..- .. ..::. .. .:: := ,- Figure 11 I- Q) 0 or E 'C Q) u. T" (' 'C .. en ca a. C/ .. :s Ct 9750352922- 143 97503529- 143 1-- residues 332- 343) occupIes both a portion of the cytoplasmic loop between helices 8 & 9 and transmembrane mapping studies from this laboratory (Levine et aI. , that the point of covalent Arg 301 - A TP attachment to GLUT 1 Met 364 . This proposed sequence III. Once a was in place , site- Peptide domain 9. region of binding 1998) suggest is between overlaps with procedure for producing and purifying GLUT directed mutagenesis could be undertaken. To test the involvement of the three domains in changed the residues in each of the GLUTI A TP binding, we three proposed sites (Table 4) and determined whether transporter to alanine function or azidoA TP binding was altered. D- glucose Uptake into Cos- Cells To assess whether wild type and mutant hGLUTI are functional we measured 2- deoxy- glucose (2DOG) uptake into Cos- cells transfected with each of our pcDNA3. 1+ GLUTI constructs. 2DOG is used in this instance because it undergoes phosphorylation and remams trapped in the intracellular space. Cells expressing epitope tagged hGLUTI show a two- fold increase in cytochalasin B (CCB) Table Mutant Name 4 - GLUT1 Alanine Mutants Original Sequence 111A 111 225A 225 GFSK 114 KSVLK229 332A 332 335A 335 338A 338 341A 341 334 Mutation 111 225 AAAA 114 AAAAA 229 332 AAA 334 335 AAA 337 LIG340 338 AAA LAG 343 341 AAN GRR TLH337 340 inhibitable sugar transport above basal levels (i.e. measured in untransfected cos- 7 cells , transport or in cells transfected with pcDNA3. 1 + lacking GLUT1 HA His cDNA) (Figure 12). Transfection with alanine mutants: lIlA , 225A , 332A , 335A, and results in a 2- fold stimulation of hypothesized that this transport (Figure 12). We increase was a result presence at the plasma membrane. measuring 8-antibody binding to Cos- is an approximate of greater by cells +/- hGLUTl. Because epitope of measurement of expression at the membrane. Cos- GLUT1 We tested this possibility this antibody recognizes an exofacial its binding 341A also tetrameric hGLUT1 transporter cells transfected with tagged GLUT 1 exhibit a 2- fold increase in 8-antibody binding above Cos- cell levels (Table 5). This confirms our original assumption that transport stimulation is a direct result of enhanced GLUTI presence at the plasma membrane. In contrast, mutant 338A shows only basal transport activity and is considered later , retains non- functional. its ability to bind This mutant, as we wil discuss azidoA TP. This result implies that Figure 12. 2- deoxyglucose uptake in Cos- 7 cells transfected with GLUT1 Comparison of 200G (0. 1 mM) uptake at 37 o C by Cos7 cells , Cos- HA Hisij. cells transfected with vector (pcONA3. 1) lacking cells expressing GLUT1 HA His and GLUT1 HA GLUT1 His sequence and by Cos7 alanine mutants. Ordinate: CCB- inhibitable 200G accumulation (mol per min per Jig total cell protein). Abscissa: cells in which transport was measured. Results are shown as mean :tSEM of triplicate measurements. Cos- His6 Cos- 7 pcDNA3. GLUT1 HA GLUT1 GLUT1 225A GLUT1 332A GLUT1 GLUT1 338A GLUT1 341A 1 1 0- 21 3 10- 2DODG uptake 4 10- mol per min per Jlg cell protein 51 Table 5: o- Antibody binding to Cos- 7 Cells transfected with GLUT1 Antibody Binding 125 cpm per Cos- 7 Cells Cos- 7 Cells + GLUT1 HA His Cos- 7 Cells + 338A 1 N= /lg protein G SEM total protein 2658 704 5595 427 2199 experiments carried out in duplicate alanine substitution at residues 338- 340 does not grossly affect transporter quaternary structure (i.e. GLUTI is folded correctly respect to A TP binding). To test this hypothesis, CCB (cytochalasin B) photoincorporation into GLUTI was measured isolated from unaltered Cos- 7 cells , as well as with in membranes cells that had been transfected with GLUT1 HA His and 338A. Membranes containing tagged wild type GLUTI bound 2- fold more CCB than Cos- 7 cells alone (Figure 13A). In contrast CCB binding was equivalent to that found in basal 13B). The absence of 338A membranes Cos- 7 membranes CCB binding and the binding imply that some aspect in retention of of the mutant (Figure azidoATP transporter structure disrupted. It is also possible that membrane is disrupted. To better understand these results , antibody binding Cos- cells transfected with mutant and 2- fold (Table 5). 8- studies as previously described were applied to exhibited the same level cells , GLUTI trafficking to the plasma 338A. Interestingly, 338A of antibody binding as untransfected Cos- less binding than cells containing wild type hGLUTI Figure 13. Photolabeling of parental GLUT1 and transfected wild type and mutant human GLUT1 in Cos- 7 cells by the sugar export site ligand (3 HJ- CCB. Cos- 7 cell membranes were photolabeled in the presence and absence of 1 00 M phloretin. Membranes were washed to remove unincorporated (3 H)- CCB and membrane proteins were separated by SOS-PAGE. Following Coomassie staining, each gel lane was sliced into 2 mm sections , dissolved and counted for associated (3 H)- CCB. Ordinate: phloretin inhibitable CCBincorporation (cpm). Abscissa: relative mobility (gel slice number). A. CCB incorporation in Cos- cells , and cells transfected with GLUT1 HA His . B. CCB incorporation in Coscells transfected with GLUT1 HA His and alanine mutant 338A. 100 Figure 13 600 500 .. Cos- 0 WT GLUT1 400 300 200 100 Slice# 2000 o WT GLUT1 . 338A -g 1500 .. 1000 t. 500 Slice# 101 AzidoATP Labeling of Wild Mutant hGLUTl Type To test whether alanine substitution affects A TP binding transporter , we measured azidoA TP incorporation (mutant & wild type) purified from Cos- cells. in GLUTI HA His Wild type mutant hGLUTI were affinity- purified from Cos- to the and cells as described previously. Proteins eluted from the Ni- NTA column were concentrated using an hour in a Amicon centricon- 10 spun at 5000 x g for Beckman JA- 17 rotor. Samples labeled with azidoA TP were equilibrated for 30 minutes in 10- 15 IJM azidoATP irradiated for 90 seconds at 280 then UV nM in a Rayonet Photochemical reactor. Fractions were simultaneously resolved on three 10% SDS- P AGE gels and transferred to nitrocellulose. Two of the membranes were used for C- Ab and HA- Ab western blotting. The third blot, which contained photolabeled hGLUT1 , autoradiographic film for 5- 24 hours. was exposed Densitometry was used to quantitate epitope tagged GLUTI and azidoA TP incorporation. protein concentration for each of the a Pierce BCA kit , applied to all and a uniform three gels. Total constructs was measured with quantity of each mutant protein was 102 Western blots and azidoA TP autoradiographs for each alanine mutant are shown in Figure 14. Lane 3 (4 mM ATP) shows purified epitope tagged GLUTI (from presence of 4 mM cold A TP. The absence suggests that A TP HA His lane 4) labeled with azidoA of radiolabel incorporation and 8-azidoA TP bind to the same site on GLUTI . This same effect has been seen previously, using purified human erythrocyte GLUTI (Carruthers et aI. , TP in the and Helgerson , 1989; Levine 1998). AzidoA TP binding in each incorporation in wild type (azidoA TP) to alanine mutant was compared to GLUTI HA His . The ratio HA antibody was used as a measure attachment to tagged transporter. Mutants lilA , 341A bound quantities of azidoA TP that found in GLUTI HA His (Figure 15). of photolabel of A TP 335A , 338A, and were equivalent to that In contrast , mutants 225A and 332A exhibited a 1. to 2. incorporation (Figure 15). Three possible explanations for this increased A fold Increase in photolabel TP incorporation are: 1) The small sidechain reduces steric interactions that normally limit transporter- binding site. 2) By exchanging A TP access polar of alanine to the residues with 103 Figure 14. Azido( p)ATP labeling of GLUT1 HA His and its corresponding alanine mutants. A. Transporter was purified by Ni- NT A affinity chromatography of membranes isolated from Cos- 7 cells transfected with wild- type or mutant GLUT1 HA His (see Figure 11). Purified GLUT1 was photolabeled using 10 pM azido( P)ATP , resolved on three separate 10% SOS PAGE gels (25pg total protein per lane per gel) and transferred to nitrocellulose for western blotting and autoradiography. For each group of mutants an internal control of GLUT1 HA His was also processed. Control protein is visible at the far left of each group of blots. GLUT1 constructs were identified by C- Ab and HA- Ab immunoblot analysis. Incorporated affnity label (azidoATP) was detected by autoradiography. Antibody and ATP binding were quantitated by densitometry using NIH Image. The ratio of azidoATP labeling to HA- Ab reactivity is shown below the blots. Lane 3 (4 mM ATP) shows that Ni- NTA purified GLUT1 HA His photolabeling byazidoATP (lane 4) is inhibited in the presence of 4 mM unlabeled ATP. ,, .. .. 104 Figure 14 GLUT- 111A 4 mM ATP GLUT1- 225A 332A GLUT1- a. 335A 338A 8, 341A 105 Figure 15. Corrected AzidoATP labeling of GLUT1 HA His (wild type and mutant). AzidoA TP labeling of tagged GLUT1 (wild type & mutant) is presented as a ratio of A TP/ HA antibody binding. This distinguishes azidoA TP labeling of recombinant human GLUT1 HA His from parental GLUT1. 106 Figure 15 111A 225A 332A 335A 338A 341A 107 uncharged alanines , A TP binding are structure of the sequence increased. 3) hydrophobicity and subsequently Charge alterations loosen the binding pocket. GLUTI HA His and 332A were also analysis of A TP inhibition of subjected to competition azidoA TP photoincorporation. Affinity- purified GLUT1 HA His and 332A were photolabeled with azidoA in the presence of increasing concentrations of unlabeled A TP. Membranes were then subjected nitrocellulose and quantitated to SDS- PAGE , transferred to for azidoA TP incorporation autoradiography and densitometry. As with of isolated purified (Levine et aI. , azidoA TP photolabeling human erythrocyte glucose transport 1998), photoincorporation of by protein azidoA TP into affinity purified GLUT1 HA His and 332A is inhibited half-maximally by 50150 JLM ATP (Figure 16). With GLUTI 332A , however, photoincorporation of azidoA TP in the fold greater. This suggests that increased photoincorporation of azidoA TP into absence of A TP is some 10- mutant 332A results from increased efficiency of mutant GLUTI labeling. HA- Ab immunoblot efficiency is unchanged by azidoA TP photolabeling of GLUTl. 108 Figure 16. ATP- inhibition of azidoATP photolabeling of affinity purified GLUT1 HA His6 (0) and GLUT1332 (8). Purified , epitope- taggedGLUT1 was photolabeled as in Figure 14 in the presence of increasing unlabeled (A TP). Proteins were resolved by SDS- PAGE and label incorporation and epitopeP-autoradiograms and Ha- tagged GLUT1 were quantitated by densitometry of Ab immunoblots. Ordinate: the ratio (32 P) incorporation: HA- Ab reactivity. Abscissa: (ATP) present during photolabeling in JiM. Curves drawn through the points were computed by nonlinear regression assuming that inhibition of photolabeling shows simple saturation kinetics and is characterized by the . P(ATP) expression: labelincorporation = N - where N IS label Kj(app) + (ATP) incorporation in the absence of unlabeled ATP , P is the amount by which saturating unlabeled A TP inhibits label incorporation and Kj(app) is that concentration of unlabeled ATP which reduces labeling by P/2. The following constants were obtained: GLUT1 HA His , N= 0. 66 :! 0. 01; P=0.41 :! 0. 54:! 17 JiM ATP. ATP; GLUT1 332 3A' N= 11. 2:! 0. 7; P=11. 3:! 1. 0, Kj(app) = , Kj(app) = 127:! 49 JiM 109 Figure 16 000 1500 500 (ATP) 2000 110 Parental Cos- 7 be able to form heterotetramers GLUTI may with GLUTI HA His . This could make it difficult to visualize the effects of GLUTI mutagenesis on azidoA TP parental GLUTI and 1 binding. For example , if 3 units of unit of tagged alterations in labeling caused by GLUT1 form a tetramer alanine substitution in the single subunit may be partially camouflaged. To examIne this possibility, Cos- 7 membranes +/- GLUTI HA His were stripped membrane proteins with an alkaline wash azidoA TP. and labeled with Samples were resolved on 10% SDS- PAGE photolabel incorporation was measured. protein was found to bind 3- fold transfected with GLUTI HA His 17). This suggests that gels and The same quantity of total more azidoA TP from cells than non- transfected cells (Figure parental Cos- 7 GLUT1 , present in preparations of nickel purified , epitope tagged hGLUT , obscure the peripheral of should not affects of alanine mutagenesis on azidoA TP binding. III Figure 17. AzidoATP labeling of Cos- 7 cell membranes stripped of peripheral proteins. Isolated Cos- 7 cell membranes from control cells or from cells transfected with GLUT1 HA His were labeled with azidoA TP following 3':=HE" '''I'"r.", removal of peripheral proteins by alkaline wash. Samples were resolved on 10% SDS PAGE gels and transferred to nitrocellulose. Protein contents were determined using a Pierce SCA assay kit , and 35 g of total protein were loaded in each lane. Label incorporation was detected by exposure of membranes to autoradiographic film. AzidoATP binding was quantitated by densitometry using NIH Image. Membranes from GLUT1 HA His transfected cells contain 3 fold more azidoA TP label than membranes isolated from nontransfected cells. :: :', .; 112 Figure 17 'i 143 Kd 97Kd 50Kd 35Kd 29 Kd 22Kd ,,1 113 Discussion In this study we ask the the three proposed question; Does alanine substitution at GLUT 1 consensus A TP binding sequences affect sugar transport and azidoATP binding? As a expressed HA His mutagenesis studies , we have erythrocyte GLUTI in: CosGLUTI HA His result of our alanine cells. Intact Cos- tagged human cells expressing exhibit normal sugar transport under basal conditions. In addition , purification of the protein on a Ni- NT A column provides a transporter that is normal with respect to A binding capacity. We compared 2- deoxyglucose (2DOG) uptake in CosGLUTI HA His as a measure of and/or cell surface epitope tagged GLUT 1 above basal Cos- function expression. Transfection with GLUT 1 HA His (wild type) and 5 of the six mutant 335A , and 341A) cells +/- constructs (111A , 225A , 332A resulted in a two- fold stimulation of transport cell levels , and levels observed in cells transfected with empty pcDNA3. 1 + vector. This increase in transport is probably the result of a 2- fold increase in cell surface most transporter 114 expreSSIOn. We tested this hypothesis by binding to Cos- 7 measurIng 8-antibody cells +/- GLUTI HA His . We observe that cells expressmg GLUTI HA His present 2- fold more 8-antibody reactive cell surface sites than untransfected Cos- our initial hypothesis cells. This result that transport stimulation in confirms transfected cells IS a direct result of greater transporter presence at the plasma membrane. In comparison , cells transfected with unchanged sugar transport later , retains rates. This construct its ability to bind oligomeric structure of mutant 338A display azidoA , as we wil discuss TP. This suggests that the 338A is not affected by alanine substitution at least with respect to A TP binding. This hypothesis was examined by measuring (3H)- CCB binding to membranes isolated from Coscells transfected with GLUTI HA His , and 338A. Tagged GLUTI (wild type) binds 2- fold more CCB than 338A or Cos- cells. These results suggest that some structural component of mutant 338A is altered (although ATP is still able to bind). Most probably, regIOn of GLUTI that contains it is the the closely related sugar efflux and CCB binding sites (Sultzman and Carruthers 1999). Alterations in 115 the sugar- binding site might also explain the inability of 338A to transport sugar. Two other possibilities could explain the inability of 338A to i:- - transport sugar. 1) trafficking to the plasma Transporter The affinity of 338A for sugar may be altered. may be disrupted. 2) 8-antibody binding studies were used expressed at the plasma binding in membrane. 338A transfected due to the absence of test whether 338A is The data show that 8-antibody cells is the same as that observed in alanine substitution at 338- 340 of a epitope of tetrameric GLUT1 may disrupt GLUT1 oligomeric configuration insensitive GLUTI exofacial- in this protocol may resolve question. this Schurmann et al. observed that transport was lost in when conserved residues 333R/334R were converted (Schurmann et aI. , may binding 338A at the plasma membrane or, because 8- antibody recognizes an exofacial structure. The use to cells. The lack of 8-antibody untransfected Cos- membrane 1997). These alanine results contrast with our work that shows normal transport activity in difference may be explained to GLUT4 GLUTI mutant 332A. This by their reconstitution of GLUT4 into 116 vesicles to measure uptake. carried out in intact Cos- cells. Substitution of alanine results in a loss of charged responsible for In their at this residues. These amino acids retaining transporter topography in the absence , exposed to In companson , our experiments were 333R/334R may site might be membrane. lose topographical stability when detergent during reconstitution. We assumed that the strategy of alanine mutagenesis would decrease, or ablate , azidoA TP binding to GLUTI. However this was not the case; in fact two constructs (225A , 1.5 to 10- and 332A) incorporate fold more azidoATP than GLUTI HA His mutants. This increase in does not reflect altered or the other labeling was somewhat unexpected affinity for A TP but rather , and altered efficiency of photolabeling. The overall and 332A might increase in be a azidoA TP labeling observed in 225A result of alanine substitution reducing the steric hindrance that normally limits ATP access to the GLUTI ATP binding pocket. Alternatively, the makes consensus sequences II , their similarity to the proposed presence of additional alanines and III more hydrophobic. Increasing consensus sequences may allow ... -' 117 more azidoA TP to bind transporter charge may to result in the loss interactions (Gallivan and I/. the transporter. Dougherty, binding. Assuming that increased a secondary A of alterations in stabilizing cation- 1999), which could affect A TP incorporation does not reflect action in which mutations introduced at a nonspecifically alter ligand binding within pocket, this Finally, distal site the nucleotide- binding effect might be consistent with a direct perturbation of the nucleotide- binding domain. ,,:1 large range appears to exist in the maXImum amount of azidoA TP that can be incorporated into mutant 332A. Densitometry of photolabeled 332A shows a 1.5- labeling (Figure15), as opposed inhibition of to 3 fold increase in azidoA competition analysis (ATP azidoA TP incorporation) which suggests azidoA TP labeling (Figure 16). This discrepancy autoradiogram overexposure in which could alter the The effects 10- fold higher may be a result the photolabeling experiments 32p detection sensitivity of the of mutating residues 332- 335 on of technique. GLUT1 A binding are consistent with N- terminal sequence analysis of peptides obtained upon proteolysis of azido A TP- photolabeled GLUTI. These 118 studies indicate that the GLUTI A TP binding domain lies between GLUTI residues 301 and 364 (Levine et aI. , 1998). Since GLUTI sequences 307 to 327 and 338 to 358 are strongly hydrophobic and thus putative membrane spanning a- helices (Mueckler et aI., they are unlikely to be GLUTI residues accessible to cytosolic A TP. In 332- 335 fall in the intervening and the remainder of contrast the Walker B domain (cytosolic) sequence between these membrane spanning regions and may be well- positioned A TP and modulation 1985), for interaction with of the sugar translocation pathway. AzidoA TP and A TP binding to purified human erythrocyte GLUTI are mutually exclusive , suggesting that they bind at site on the protein (Carruthers and Helgerson , 1989). To the same determine that this is also the case in epitope tagged GLUT1 , we labeled purified GLUT1 HA His with azidoATP in the presence of 4 mM ATP. The absence of 32 P incorporation confirms the stereospecificity of azidoA TP binding to tagged GLUTl. It is probable oligomerize to form that parental GLUTI and tagged GLUTI tetrameric GLUT1. However , membranes transfected with GLUTI HA His Cos- cell bind 3- fold more 119 azidoA TP than a comparable amount of protein from untransfected membranes. This suggests that the presence of samples of parental GLUT1 in affinity purified epitope tagged GLUTI does changes in azidoA mutagenesis. TP binding that are a direct not obscure result of alanine '-- 120 Chapter Significance & Future Directions This thesis examInes several hypotheses. First , we tested capacity of various physiologic factors TP modulation to influence A of GLUTI mediated sugar transport. Specifically, we looked effects of glycolytic intermediates, oligomeric structure. Next we asked , is the ATP interaction with GLUTI nucleotide binding domain erythrocyte glucose transporter. at intracellular pH , and transporter sufficient for sugar transport modulation? Third , we location of the the investigated the on the human Finally, having narrowed the possible location of the ATP binding site on GLUT1 , we analyzed whether alanine mutagenesis altered sugar transport and A binding. We have concluded that only AMP , ATP regulation of GLUTl. and possibly ADP , inhibit Previous work in this laboratory has shown that AMP , and ADP , though unable to mimic the effects of A TP compete with A TP for access to the transport system (Carruthers and Helgerson , 1989). This suggests that the adenosine ;;.. 121 ri; binding, while phosphate chain moiety is responsible for dictates the ability of a compound to length modulate GLUT 1 activity. The r. - F;t:: present study shows that modulation of adenosine alone is insufficient for human GLUTI function. An increase in glycolysis can lead to an elevation in lactate levels and subsequent acidosis (Cairns et aI. , encouraged us to examine how alterations in cellular 1993). This intracellular pH affect A TP regulation of GLUTI. From these studies, we observed that as pH decreases from 8. A TP modulation of GLUT1 is inhibited. 0 to We propose that this loss of A TP regulation results from altered binding kinetics (affinity or capacity pH dependent of GLUTI for A TP) or from the binding of a regulatory molecule to the A TP- GLUTI complex. The data indicate that pH affects A TP- Surprisingly, we found that low pH increases the (app) for A TP- inhibition et al. (Merino et aI. , (32P)A of photolabeling is 1997) has shown structure of the nucleotide- binding Specifically, low pH was GLUTI binding kinetics. TP binding while unchanged. Merino that pH also modulates the domain on A TPase. found to increase access to the A TP binding 122 site. A similar mechanism could explain the increase binding that we have observed in in azidoA GLUTI as pH is lowered from 8. to 5. To better understand how low pH inhibits A TP regulation of GLUT1 , we examined the ATP dependence of 30MG uptake cell ghosts at pH 5. 5 and 8. 0. Uptake was found to increase saturable manner at pH 8. 0. into red in a In contrast, 30MG uptake at pH 5. 5 was maximally stimulated at 200 JlM A TP and inhibited at 2 mM A TP. Degradation of A TP at low pH was ruled out as a cause of this unexpected result. We have concluded that A TP acts in, an inhibitory fashion sugar transport at low pH. incorporation at pH 5. 5, on In view of the increase seen In photolabel transport inhibition was unexpected. This result alluded to the possibility that low pH unmasks a unIque transport-antagonistic ATP binding site on GLUTl. The exposure of such a site would account the inhibition of for both the increase in A TP binding and transport seen under acidic conditions. Transport inhibition could also be the result of GLUTI interacting with a dependent inhibitory protein. 123 We have shown that GLUT1 reconstituted into proteoliposomes containing 2 stimulation at sub that GLUT1 , exhibits transport saturating 30MG. We have therefore the lipid bilayer required for A mM A TP , and ATP are the concluded only components TP modulation of GLUTI function. This same system could be used to rule out or confirm the presence of a regulatory protein. If transport inhibition is seen in pH inducible this system at pH 5. 5, we can say with some certainty that a regulatory protein not the agent arresting transport at low pH. Wheeler (1989) was unable to recreate the A TP effect in a reconstituted system. His ' experimental design used the reduced dimeric form of physiologic or Carruthers the transporter. In contrast , our tetrameric form of the transporter (Hebert and 1992). This laboratory has observed that exposure to reductant (dithiothreitol- DTT) inhibits uptake. studies employ the A TP modulation of 30MG Specifically, conversion of tetrameric GLUTI to its form results in loss between the two regulated by A of azidoA TP labeling. This explains data sets. We conclude TP. that the dimeric disparity tetrameric GLUTI is ,t' f, 124 Although GLUT1 functions as an A TP binding protein , location of its nucleotide binding domain is experiments have suggested ;Jmj .n that to GLUTI falls within the range the currently unknown. Past covalent incorporation of azidoA TP of C- terminal residues (270- 456) (Carruthers and Helgerson , 1989). The peptide mapping studies carried out in this thesis identify residues 291- 364 as a likely component of the GLUTI azidoA TP-attachment site. The strong hydrophobicity of GLUTI presents molecular weight of this uncertainty, carbohydrate , a problem for estimating the GLUTI fragments by gel mobility. To balance we used internal markers (C- terminus , and azidoA N- terminus TP), and proteolytic enzyme specificity to identify peptide fragments more definitively. More recent work in our laboratory has shown that A TP increases GLUTI resistance to proteolytic digestion (Cloherty and Carruthers, observations). This may offer an explanation unpublished for the encountered obtaining complete proteolytic digests difficulty of azidoA labeled GLUTl. Digestion of purified erythrocyte GLUTI with endoproteinase 125 Lys- peptide fragment whose first ten produces a identified by N'-terminal sequencIng as GLUT1 This sequence and the peptide mapping data amInO acids were residues 301- 310. allow us to further refine the proposed azidoA TP-attachment site on GLUTI to cassette that contains residues 301- 364. a Importantly, this range of residues overlaps with a proposed Walker B motif in GLUT1. We have shown that A TP competitively inhibits azidoA TP binding to purified GLUT1 (Carruthers and Helgerson , that azidoA TP and A TP bind to the same site on the transporter. The 1989). This proposed pocket structure of a nucleotide- binding implies domain (Fry et aI., 1986) presents certain problems when using 8-azidoA TP for peptide mappIng. Specifically the azido moiety that of adenine may crosslink site. For this reason 8-azidoA TP 8 to a peptide adjacent to the actual binding GLUTI peptide mapping experiments 2-azidoA TP should also be performed in GLUTI segments sits on carbon contributing to the 2-azidoA TP crosslink of 2-azido labeled GLUTI might pattern of labeled peptides. using an effort to identify other nucleotide binding pocket. If to adjacent peptides , digestion produce a significantly different ;::!. 126 To further delineate the location of the A constructed alanine mutants for each of consensus sequences on GLUT 1 TP binding site, we the three proposed and measured azidoA TP labeling of and sugar transport by, each mutant GLUTl. This required development of a the GLUTI expression system. Unfortunately, the yeast strain we created (RE700A + GLUTI HA His ) grew too slowly and produced insufficient GLUTI for our purposes. clones are being screened to and human GLUT select one with a high Currently, more rate of growth expression. This problem led us to develop expressIOn system for a transient mammalian human GLUTI usmg Cos- cells. Purification of HA His tagged GLUTI on a nickel column produced large quantities of a protein which run as a diffuse band with approximate molecular weight of 55 kDa. GLUTI C- , and HA- , identified this as erythrocyte glucose transporter. found to be normal Western blotting with with respect transporter was normal with epitope tagged human In addition , in-vivo transporter was to sugar transport , and respect to azidoA the isolated TP binding. B:. 127 Transfection of Cos- 7 and five of the cells with epitope tagged wild type GLUT1 SIX alanine mutants, resulted in a 2- fold increase in DOG uptake above basal Cos- 7 cell levels. 8- Antibody binding studies have shown that cells transfected with GLUT1 HA His present a 2- fold excess of GLUTI at the plasma membrane relative to untransfected cells. This suggests that the transport stimulation we observed cells expressing epitope tagged GLUT1 , transporter expression at the cannot rule out the stimulate transport, Cos- in Cos- is the result of increased cell membrane. As a cautionary note , one possibility that these newly expressed proteins in part , by having a higher rate of catalytic turnover than parental GLUT1. However , quantitatively the change increased transport parallels in cell surface binding of a antiserum (8- Ab) directed against exofacial epitopes of GLUT I-mediated 2DODG transport 1986). The measurements in GLUTl. Km(app) for human red cells is 15 mM (Stein reported in the present study were made at 0. mM 2DOG. Transport is, therefore, given by uptake polyclonal max f2DODG) m(app) ::-\%'; 128 where V max cellular 2DOG transport capacity) = kcat (GLUTILell surface (the where kcat is the intrinsic catalytic turnover of GLUTI. Since bo th transport and (GLUTl)ceii surface are doubled by Cos- GLUT1 HA His this suggests that the k / Km(app) HA His cell transfection with ratio for expressed GLUTI cells is not significantly different from that of Cos- 7 cell in Cos- endogenous GLUT1. 338A was the only mutant whose catalytic activity may have been affected by alanine substitution. 30MG uptake by this construct was not significantly different from uptake in untransfected Cos- cells. This was suprising because these residues (338- 340) are located in a proposed transmembrane domain of GLUTI and are , like alanine , uncharged. The fact that this retains its ability to bind azidoA TP suggests structure is unchanged with In contrast , phloretin 55 kDa protein mutant that its oligomeric respect to the nucleotide binding site. inhibitable photoincorporation of CCB into a in membranes expressing mutant 338A is not significantly different from that observed with membranes isolated from non- transfected cells. a portion of mutant The absence 338A structure of CCB binding implies that is disrupted. Because this 129 mutant retains azidoA TP binding, the disrupted region in the CCB labeling data suggest that question encompasses the proposed GLUT CCB binding and sugar efflux sites , which are closely related (Sultzman and Carruthers 1999). The lack of 2- deoxyglucose uptake by the 338 mutant could result from disabled sugar transport, a result loss of consistent with the CCB binding. Alternatively, the transporter may not be expressed correctly at the plasma membrane. GLUT 1 that is structurally intact , could studies because total solubilized to the Cos- Incorrectly trafficked still bind azidoA TP in these cell membranes are applied nickel column for epitope tagged GLUT 1 purification. To test this possibilty, cell surface 8:-antibody binding was measured in Cos- cells transfected with mutant 338A. The results show that mutant 338A expressing antibody as Cos- cells bind the cells alone , but 2- expressing wild type same amount of fold fewer IgGs than cells epitope tagged GLUTl. The lack of o-an ti body binding may be due to the absence of 338A at the plasma membrane or alternatively, alanine substitution at 338- 340 may convert tetrameric GLUT1 to dimeric GLUT1. In this instance 8- 130 antibody, which recognizes an exofacial epitope of GLUT1, would be unable to recogmze the protein. If this is the case the affinity of mutant 338A for 2DOG or its ability 2DOG may be altered. Immunohistochemistry question. of the following: 1) dysfunctional GLUTI trafficking to the plasma structure in a small region of membrane , 2) abnormal oligomeric the transporter, and/or 3) an altered structure. Originally; we hypothesized amIno acid to alanine would knock out , were found to bind 1.5 to 2. essential that converSIOn of an or reduce , azidoA to the transporter. Contrary to this expectation , His this configuration conclude that the lack of sugar transport in mutant 338A is a result of one or more quaternary to translocate using a insensitive exofacial GLUTI antibody may resolve We therefore tetrameric fold more azidoA TP binding 225A and TP than 332A GLUTI HA or the other alanine mutants. It is possible that the charge alterations in these mutants loosen the structure resulting in increased access of the N of the transporter group of azidoA TP to the A binding pocket. Another explanation is based on the small side chain found on alanine. In both 225A and 332A , alanine replaces ~~~~ 131 polar amino acids that contain large side chains. Removal of barriers could reduce steric hindrances controlling access nucleotide binding pocket. Another possibility is related This stretch hydrophobic nature of Sequence III. GLUTI was found to contain 50% is highly hydrophobic to the to the amino acids in homology to a Walker B box which (Fry et aI. , alanines for polar residues in of these 1986). Substitution of GLUTI (i.e. 332- 334 the hydrophobic nature of this regIOn. uncharged GRR), Increases By making GLUTI sequence III more like a true Walker B motif, we may have enhanced azidoA TP binding to the transporter. In the absence of a change in the k A TP inhibition 332A , of azidoA TP labeling in GLUT 1 HA His for and mutant it is doubtful that the increased label incorporation that we have observed in mutants 225A , transporter affinity for and 332A is due ATP. The reactive N to altered of azidoATP may simply be in closer proximity to (i. e. have increased reacting probability of with) side chains in the pocket. Another obstacle that concerned us was the potential for parental GLUTI oligomerization with tagged GLUTI (both and wild type). GLUT1 HA His containing Cos- mutant cell membranes 132 bound 3- fold more azidoA TP than membranes from cells tagged GLUT1. We propose that the presence of lacking parental GLUTI in these preparations should not obscure the affects of alanine mutagenesis. We were particularly interested to observe that azidoA TP incorporation into increases in a purified human erythrocyte GLUTI at low pH manner similar to that found with epitope GLUTI mutant 332A purified from Cos- cells. Ki(app) inhibition of azidoA TP GLUTI HA His indicating that transporter affinity for A mutation of ATP- inhibition ATP- labeling is not reduced TP is not these residues. AzidoA TP labeling of increased 5 to 10- fold by lowering for tagged pH from 8. 5 increased by red to 5. cell GLUTI is 5 yet Ki(app) for of labeling is unchanged (Levine et aI., 1998). Histidine is the best candidate as pH is lowered from 8. 5 amino acid to 5. 5. for side-chain protonation GLUT1 contains 5 histidine residues of which four (H 160, 239, 337 and 484) are located within cytoplasmic domains. One of these , GLUT 1 Walker B H 337 , is found in the putative domain although direct involvement in pH- dependent modulation of azidoA TP binding efficiency wil require 133 confirmation by substitution mutagenesis. Why would His protonation mImIC (alanineh 332 GRR? substitution for One clue may tg' be obtained from the hydrophobic GLUT1 sequence preceding the putative Walker B mqtif. This sequence contains a phenylalanine that could pair with Arg333 or Arg334 by cation-n (Gallivan and Dougherty, 1999). These interactions, which attain approaching - 4 kcal/mol , can structure (Gallivan and Dougherty, interactions significantly stabilize protein 1999). If the neighboring His ::iJ. ':(.\:1 iL' 'J. . were protonated at reduced pH , interactions by charge by mutagenesis it might disrupt Phe- Arg repulsion of Arg or protonation of His structure permitting cation-n . Removal of Arg 333 or Arg 334 ? could relax local GLUTI uv closer interaction of the activated azidoA nitrene with GLUTI side chains or backbone. This structural relaxation could increase the efficiency of label without changing the affinity of the transporter While speculative , photoincorporation for nucleotide. this hypothesis warrants experimental scrutiny. Fluorescence resonance energy transfer (FRET) is another technique that can be used to further GLUT 1 nucleotide binding domain. delineate the location of the We have collaborated with a 134 visiting orgamc synthetic chemist (Dr. Boris Yagen , University, Jerusalem) to synthesize a fluorescent A Hebrew TP derivative anthraniloyl- A TP (AN- A TP). AN- A TP has an absorbance maximum at 343 nm and AN- an emISSIOn maXImum at approximately 450 nm. ATP binding to 2 with k tLM GLUTI is half constants of 0. and koff 3 per tLM maximal at 7 tLM AN- ATP per min and 2. 1 per min respectively (Cloherty, Yagen and Carruthers, unpublished). AN- ATP also mimics the abilty of transport in red cells. ATP to modulate GLUT I-mediated Using the intrinsic tryptophan fluorescence of GLUT1 (excitation at 298 nm , emission max 332 nm) to excite AN- A TP bound GLUT1 , we have observed fluorescence resonance energy transfer (FRET) between GLUT1 tryptophan residues and bound AN- A TP. AN- A TP fluorescence requires GLUTI tryptophan fluorescence because in the absence of this energy extremely low fluorescence. location of the source , AN- A Using this TP shows technique, we can map the GLUTI nucleotide- binding domain as tryptophan residues within the it relates to transporter. By utilizing column purification of epitope tagged human Ni- NT GLUT 1 , we can A 135 synthesize and isolate GLUT1 mutants in which tryptophan 1l- converted to glycine or leucine , thereby identifying tryptophan residues in close proximity to i , has been bound AN- A TP ;;? 136 References Bers , D. M., (1994). " C. W. Patton and R. Nuccitelli to the preparation of Ca2+ buffers. practical guide 40: 3- Methods Cell Biol 29. Birnbaum , M. J. (1989). " Identification of a novel gene encoding an insulin-responsive glucose transporter protein. Cell 5 7: 3 05- 315. Birnbaum, M. J. , H. C. Haspel and O. M. Rosen (1986). " Cloning characterization of a cDNA encoding transporter protein. and the rat brain glucose- Proc. Natl. Acad. Sci. USA 83: 5784- 5788. Brown , J. N. (1975). " The avian electron microscopy. Cairns, S. P. , H. Westerblad erythrocyte: a study of fixation for J. Microsc. 104: 293- 305. and D. G. Allen (1993). " Changes in myoplasmic pH and calcium concentration during exposure to lactate in isolated rat ventricular myocytes. 464: 561- 574. J Physiol (Lond) 137 Calderhead , D. M. , K. Kitagawa, G. E. Lienhard and G. W. Gould (1990). " Translocation of the brain- type Carpenter , myocytes. transporter insulin stimulation of glucose transport in largely accounts for BC3H- l glucose 269(3): 597- 601. Biochemical Journal J. F. and S. C. Hand (1986). " Comparison dependent allostery and dissociation of pH- for phosphofructokinases from Artemia embryos and rabbit muscle: nature of the enzymes acylated with diethylpyrocarbonate. Biophys 248(1): 1- Carruthers, A. (1986). " ATP regulation transporter. J. Biol. Chem. Carruthers, A. (1990). " Facilitated Rev. 70(4): 1135- Carruthers, A. and of the human red cell sugar 261(24): 11028- 11037. diffusion of glucose. Physiol. 1176. A. L. sugar transporter Biochemistry Arch Biochem Helgerson (1989). " The human erythrocyte is also a nucleotide 28(21): 8337- 8346. binding protein. ;; ,;..",'" 138 Cartee , G. D. , A. G. Douen , (1991). " Stimulation hypoxia. T. Ramlal, A. Klip and J. O. of glucose transport in Thomas and D. skeletal muscle by 70(4): 1593- 600. Journal of Applied Physiology Cassard , A. M. , F. Bouilaud , M. Holloszy G. Mattei , E. Hentz, S. Raimbault , M. Ricquier (1990). " Human uncoupling protein gene: structure , comparison with rat gene , and assignment .(i the long arm of chromosome 4. J Cell Biochem 43(3): 255- 64. Casteilla , L. , F. Bouilaud , C. Forest and D. Ricquier (1989). Nucleotide sequence of a cDNA encoding bovine brown fat uncoupling protein. Homology with ADP binding site of ADP/ATP carrier. Nucleic Acids Res 1 7(5): 2131. Charron , M. J. , F. C. Brosius III , S. L. Alper and h. F. Lodish (1989). glucose transport protein expressed predominately in insulin-responsive tissues. 2535- 2539. P roc. Natl. Acad. Sci. U.S. A. 86: 139 Cheung, 1. Y. , D. M. Regen , Morgan (1977). " T. (1975). " muscle cells. C. F. Whitfield and of Anaerobic stimulation avian erythrocytes. Clausen , M. E. Schworer , Biochim. Biophys. Acta H. E. sugar transport in 470: 212- 229. The effects of insulin on glucose transport in 6: 169- 226. Curro Top. Memb. Transp. Cloherty, E. K. , D. L. Diamond , K. S. Heard and A. Carruthers (1996). GLUTI-mediated sugar transport by Regulation of Biochemistry antiport/uniport switch mechanism. 3 5: 13231- 13239. Cloherty, E. K. , K. S. Heard and A. Carruthers (1996). " Human erythrocyte sugar transport is incompatible with available Cushman , S. W. and 35: 10411- 10421. Biochemistry carrier models. L. J. Wardzala (1980). " Potential mechanism of insulin action on glucose transport cel1." Darnell, J. , J. Biol. Chem. in the isolated rat adipose 255: 4755- 4762. Lodish , H. , and Baltimore , D. (1990). Biology . New York , Scientific Molecular Cell American Books, Inc. pp Pages 140 Diamond , D. and A. Carruthers (1993). " Metabolic control transport' by derepression of cell surface glucose of sugar transporters: an insulin- independent, recruitmen t- i ndependen t mechanism J. Biol. of regulation. 268: 6437 - 6 444. Chem. Doria , M. C., D. D. Lund , A. Pasley, A. Sandra and W. 1. Sivitz (1993). Immunolocalization of GLUT - 1 glucose skeletal muscle and in American Journal of Douen , A. G. , T. Ramlal , normal and hypoxic cardiac tissue. 265: E454- E464. Physiology S. Rastogi , Vranic , J. O. Holloszy and recruitment of Evidence for P. J. Bilan, G. D. Cartee , A. Klip (1990). " Exercise M. induces the " insulin-responsive glucose transporter distinct intracellular insulin- and exercise- recruitable transporter pools in skeletal Biological Chemistry Elbrink , transporter in rat Journal of 265(23): 13427- 30. J. and 1. Bihler (1975). " Membrane cellular metabolic rates. Ellott, A. C. , G. L. Smith muscle. Science 188: transport. Its relation to 1177- 1184. , D. A. Eisner and D. G. Allen (1992). Metabolic changes during ischaemia and their role 141 contractile failure in isolated ferret hearts. 467- 90. 454: Fischer , J Physiol (Lond) Y., J. Thomas, P. Rosen and H. Kammermeier (1995). " Action of metformin on glucose transport and glucose transporter GLUTI and GLUT4 in heart muscle cells from healthy diabetic rats. Endocrinology 136(2): Fischer , Y., J. Thomas, L. Sevila, P. Munoz , Kozka , M. Palacin , X. Testar , 412- 20. C. Becker , H. Kammermeier (1997). " Insulin- induced recruitment G. Holman , of glucose transporter 4 Evidence existence of different intracellular GLUT4 vesicle populations. Fry, D. C. 1. J. and A. Zorzano (GLUT4) and GLUTI in isolated rat cardiac myocytes. of the and J Biol Chem 272(11): 7085- 92. , S. A. Kuby and A. S. Mildvan (1986). " ATP- binding site of adenylate kinase: Mechanistic implications of its with ras-encoded p21 , FI- A TPase , and binding proteins. Proc. Natl. Acad. Sci. other nucleotideU.S. A. Fukumoto , H., T. Kayano , J. B. Buse , Y. Edwards , and S. Seino homology 83: 907- 911. P. F. Pilch , (1989). " Cloning and characterization G. 1. Bell of the ,,,',' 142 major insulin-responsive glucose transporter expressed in human skeletal muscle and other insulin-responsive tissues. 264: 7776- 7779. Biol. Chem. Gallvan, J. P. and D. A. Dougherty (1999). " Cation- pi interactions in structural biology. Proc Natl Acad 96(17): 9459- 64. Sci USA Gao , J., J. Ren , E. A. Gulve and J. O. Holloszy (1994). " Additive effect of contractions sarcolemma. Goodyear , L. J. , and insulin on GLUT - translocation into the 77(4): 1597- 601. J Appl Physiol M. F. Hirshman , R. J. Smith and E. S. Horton (1991). Glucose transporter number, activity, and isoform content in plasma membranes of red and white skeletal muscle. American Journal of Physiology Gould , G. W. and G. 1. Bell (1990). " Facilitative an expanding family. Hansen , P. (1995). " Kinetics muscle: effects of Physiology 1 5(1): 18- 23. Trends Biochem Sci , E. Gulve , 1. Gao, J. Schluter , glucose transporters: M. Mueckler and J. Holloszy of 2- deoxyglucose transport in of insulin and contractions. 268(1 Pt 1): C30- skeletal American Journal 143 . V' ,i Harris , J. R. and J. N. Brown (1971). " Fractionation erythrocyte: an ultrastructural study. of the aVIan 36: J. Ultrastruct. Res. 23. Haworth , R. A. and H. A. Berkoff (1986). " The uptake by metabolic demand 5 8(1): 1 57 - control of sugar in isolated heart cells. Circ Res 65. Hebert , D. N. and A. Carruthers (1992). " Glucose transporter oligomeric structure determines transporter function. Reversible redox- dependent interconversions of and dimeric GLUT 1. " tetrameric Journal of Biological Chemistry 267(33): 23829- 38. Helgerson , A. L. , D. N. Hebert, S. N aderi Characterization of and A. Carruthers (1989). two independent modes of action on human erythrocyte sugar transport." of A Biochemistry 2 8(15 ) : 6410- 6417. Hirshman , M. F. , H. Wallberg- Henriksson and E. S. Horton , L. J. Wardzala, E. D. Horton (1988). " Acute exercise increases the number 144 of plasma membrane glucosetransporters in muscle. rat skeletal 238: 235- 239. Febs. Lett. Hresko , R. c., M. Kruse, M. Strube and M. Mueckler (1994). Topology of the Glut 1 glucose transporter deduced from glycosylation scannIng mutagenesis. J Biol hem 269(32): 20482James , D. E. , R. Brown , J. Navarro and P. F. Pilch (1988). " Insulin- regulatable tissues express a unique insulin-sensitive transport protein. James , D. E. Nature glucose 3 3 3: 1 83 - 18 5. , M. Strube and M. Mueckler (1989). " Molecular cloning and characterization transporter. Nature of an insulin-regulatable 3 3 8: 8 3 glucose -87. Jekel , P. A. , W. J. Weijer and J. J. Beintema (1983). " Use of endoproteinase Lyssequence analysis. from Lysobacter enzymogenes Anal Biochem in protein 134(2): 347- 54. Kaestner , K. H., R. J. Christy, 1. C. Mclenithan, L. T. Braiterman , Cornelius , P. H. Pekala and D. Lane (1989). " Sequence , distribution and differential expression of P. tissue mRNA for a putative 145 insulin-responsive glucose transporter adipocytes. Proc. Natl. Acad. Sci. U. gf' mouse 3T386: 3150- 3154. A. Kayano , T., C. F. Burant , H. Fukumoto , G. W. Eddy, M. G. Byers , in Gould, Y. S. Fan , R. L. T. B. Shows, S. Seino and G. 1. Bell (1990). Human facilitative glucose transporters. Isolation , functional ;J' characterization , and gene localization of cDNAs encoding an isoform (GLUT5) expressed in small intestine, kidney, muscle and adipose tissue and an pseudogene- like T. sequence (GLUT6). glucose transporter Journal of Biological 265(22): 13276- 82. Chemistry Kayano , unusual , H. Fukumoto , R. L. Eddy, Y. S. Fan , M. G. Byers, T. B. Shows and G. 1. Bell (1988). " Evidence for a family of human glucose transporter- like proteins. J. Biol. Chem. 263: 15245- 15248. Klip, A. and M. R. Paquet transporters in (Review). (1990). " Glucose transport and glucose muscle and their metabolic regulation. Diabetes Care 13(3): 228- 43. 146 Kolter , T. , 1. Uphues, A. Wichelhaus, H. Reinauer and J. Eckel (1992). Contraction- induced translocation of the glucose transporter Biochemical & Glut4 in isolated ventricular cardiomyocytes. Biophysical Research Communications 189(2): 1207- 14. Kraegen, E. W., J. A. Sowden , M. B. Halstead , P. W. Clark , K. J. Rodnick , D. J. Chisholm and transporters and in vivo D. E. James (1993). " Glucose glucose uptake in skeletal and cardiac muscle: fasting, insulin stimulation and immunoisolation studies of Laemmli , U. K, Lee, A. D. , GLUTI and GLUT4. Nature (1970). Biochemical Journal 270: 680- 685. P. A. Hansen and J. O. Holloszy (1995). " Wortmannin inhibits insulin-stimulated but not contraction-stimulated glucose transport activity in skeletal muscle. Febs Lett 361(1): 51- Levine , K. B., E. K. Cloherty, N. J. Fidyk and A. Carruthers (1998). Structural and physiologic determinants of human erythrocyte sugar transport regulation by adenosine triphosphate. Biochemistry 37(35): 12221- 32. 147 Loo , D. D. , A. Hazama , S. Supplisson , (1993). " Relaxation kinetics of E. Turk and E. M. Wright the N a+/glucose cotransporter. Proceedings of the National Academy of Sciences of the United 90(12): 5767- 71. States of America Lon , T. W. and D. M. Clarke (1994). " Prolonged association temperature-sensitive mutants of human Pcalnexin during biogenesis. Loo , of glycoprotein with 269(46): 28683- J Biol Chem T. W. and D. M. Clarke (1995). " Rapid purification of human Pglycoprotein mutants expressed transiently in HEK 293 cells by nickel chelate chromatography and characterization of their drug-stimulated ATPase activities. J Biol Chem 270(37): 21449- 52. Mansford , K. R. (1968). " Effects glycolysis. of heart uptake and 61(8): 816- Proc R Soc Med Martell , A. E. and R. M. Smith work on glucose (1974). Critical stability constants New York " Plenum Press pp Pages Merino, J. M. , F. Henao and C. Gutierrez- Merino (1997). " Structural changes of the sarcoplasmic reticulum Ca(II)- A TPase , ' 148 nucleotide binding domain by pH and La(III). Arch Biochem 348(1): 152- Biophys Morgan , H. E. , M. J. Henderson Regulation of glucose , D. M. Regen and C. R. Park (1961). transport in muscle: the effects of insulin and anoxia on glucose transport and phosphorylation in the isolated perfused heart of normal rats. J. Biol. Chern. 236: 253- 261. Morgan , H. E., J. R. Neely, R. D. Wood , C. Liebecq, H. Liebermeister and C. R. Park (1965). " Factors affecting glucose transport in heart muscle and erythrocytes. Fed. Proc. 24: 1040- 1045. " A \1! 1iir tit 'il Morgan , H. E. , P. J. Randle and D. M. Regen (1959). Biochem. J. 73: 573- 579. Mueckler, M. , C. Caruso , S. A. Baldwin , M. Panico , 1. Blench , H. R. Morris, W. J. Allard , G. E. Lienhard and H. F. Lodish (1985). Sequence and structure of a human glucose transporter. :tJfr ,lei Science 2 2 9: 941- 945. Newgard , C. B. , D. R. Littman , C. van genderen, M. Smith and R. J. Fletterick (1988). " Human brain glycogen phosphorylase. ..:. 149 analysis, chromosomal mappmg, tissue Cloning, sequence expreSSIOn and comparison with J. Biol. isozymes. the liver and muscle 263: 3850- 3857. Chem. Ohta, S. and Y. Kagawa (1986). " Human FI- ATPase: molecular cloning of cDNA for the beta subunit. J Biochem (Tokyo) 99(1): 135- 41. Piano, G. V. and H. H. Winkler (1991). " Identification and initial topological analysis translocase. Reynolds, P. , of the Rickettsia prowazekii A 173(11): 3389- 96. JBacteriol D. R. Higgins , L. Prakash and S. Prakash (1985). " The nucleotide sequence of cerevISlae: a the RAD3 gene of Saccharomyces potential adenine nucleotide binding amino acid sequence and a Nucleic Acids TP/ ADP Res. nonessential acidic carboxyl terminal region. 13: 2357- 2372. Rosevear , P. R. , T. L. Fox and A. S. Mildvan (1987). " Nuclear Overhauser effect studies bound to kinase. of the conformations of MgA TP the active and secondary sites of muscle pyruvate Biochemistry 26(12): 3487- 93. 150 Schurmann , A., H. Doege , H. Ohnimus , Joost (1997). " Role V. Monser , A. Buchs and H. G. conserved arginine and glutamate of residues on the cytosolic surface of glucose transporters for transporter function. 3 6(42): 12897- 902. Biochemistry Sen , A. K. and W. F. Widdas (1962). " Determination of the temperature and pH dependence of glucose transfer human erythrocyte membrane measured T. J. B. (1983). " Characterization of sugar transport in the pIgeon red blood cell. Simons , T. J. glucose exit." 160: 392- 403. Physiol. Simons , by across the B. (1983). " 338: 477 - 5 00. J. Physiol. The role of calcium in the sugar transport in the pigeon red blood cell. regulation of J. Physiol. 338: 501- 526. Smith , B. L. and P. Agre (1991). " Erythrocyte Mr 28, 000 transmembrane protein exists as a multi subunit oligomer similar to channel proteins. J Biol Chem 266(10): 6407- 15. Smith , G. L., P. Donoso , C. J. Bauer and D. A. Eisner (1993). Relationship between intracellular pH and metabolite 151 I r: in isolated ferret Sogin , D. C. and P. C. Hinkle (1980). " Immunological identification concentrations during metabolic inhibition heart." Acad. Sci. Stein , W. D. (1986). New York Proc. Natl. erythrocyte glucose transporter. of the human Sternlicht, E. 472: 11- 22. J Physiol (Lond) U.S. A. 77: 5725- 5729. Transport and diffusion across cell membranes Academic Press pp Pages , R. J. Barnard and G. K. Grimditch (1988). " Mechanism of insulin action on glucose transport Am. J. Physiol. in rat skeletal muscle. 254: E633- E638. Sultzman , L. A. and A. Carruthers (1999). " Stop- flow analysis cooperative interactions between GLUTI sugar export sites. Biochemistry import and 3 8(20): 6640- 50. Sun, D. , N. Nguyen , T. R. DeGrado , M. Schwaiger and F. C. r. Brosius (1994). " Ischemia induces translocation responsIve glucose of of the transporter GLUT4 to the membrane of cardiac myocytes. insulin- plasma Circulation 89(2): 793- 152 Suzuki , K. and T. Kono (1980). " Evidence that insulin causes translocation of glucose transport activity to the membrane from an intracellular storage site. Sci. U.S. A. Pro plasma c. Natl. Acad. 77: 2542- 2545. Thorens, B., H. K. SarJ.ar , H. R. Kaback Cloning and functional and H. F. Lodish expression in bacteria of a novel glucosetransporter present in liver , beta- pancreaticislet cells. (1988). Cell intestine , kidney, and 5 5: 2 81- 2 90. Valdez , B. C. , Z. Chen , M. G. Sosa , E. S. Younathan and S. H. Chang (1989). " Human 6- phosphofructo- kinase gene has an additional intron upstream of start codon. Gene Vander , A. J. , Sherman , James H. , Luciano , Dorothy Human Physiology. New York , 7 6(1): 167- S. (1990). McGraw- Hill Publishing Company pp Pages Wang, W. , P. A. Hansen , B. A. Marshall , J. O. Holloszy and M. Mueckler (1996). " Insulin unmasks a COOH- terminal Glut4 153 epitope and increases glucose transport across T - tubules in skeletal muscle. 135(2): 415- 30. J Cell Biol Waygood , E. B. , J. S. Mort and B. D. Sanwal (1976). " The pyruvate kinase of control of Escherichia coli. Binding of substrate allosteric effectors to the enzyme activated by fructose bisphosphate. Wheeler , Biochemistry of glucose J. Biol. heart." 1 15(2): 277- 82. T. J. (1988). " Translocation response to anoxia in and Chem. transporters in 263: 19447- 19454. Wheeler , T. J. (1989). " ATP does glucose transporter. not regulate the Biochemistry reconstituted 28: 3413- 3420. Wheeler , T. J. , R. D. Fell and M. A. Hauck (1994). " Translocation of two glucose transporters in heart: effects of rotenone uncouplers, workload , palmitate , insulin and anoxia. Biophys Acta Biochim 1196(2): 191- 200. Wheeler , T. J. and P. C. Hinkle (1985). " The glucose transporter mammalian cells. Ann. Rev. Physiol. 47: 503- 518. of 154 Whitfield , C. F. and J. E. Morgan (1983). " Effect transport in avian erythrocytes. Biochim. of anoxia on sugar 307: Biophys. Acta 181- 196. Wood, R. E. and H. E. Morgan (1969). " Regulation of sugar transport erythrocytes. II aVIan J. Biol. Chem. 244: 145 Zorzano , A. , L. Sevila , M. Camps, C. Becker , J. Meyer , Kammermeier , P. Munoz , A. Guma , X. Testar , Blasi and Y. Fischer (1997). " Regulation 1 - 1460. H. M. Palacin , of glucose J. transport, and glucose transporters expressIOn and trafficking in the heart: studies in cardiac myocytes. Am J Cardiol 8 0(3A): 65A- 76A. Zottola , R. J. , E. K. Cloherty, P. E. Coderre , A. Hansen , and A. D. N. Hebert Carruthers (1995). " Glucose Transporter Function Is Controled By Transporter Oligomeric Structure. A Single Intramolecular Disulfide Promotes Glutl Tetramerization. Biochemistry 34: 9734- 9747.