Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Mechanistic studies of rabbit muscle glycogen phosphorylase Stirtan, William G. 1993

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1993_spring_phd_stirtan_william.pdf [ 8.25MB ]
JSON: 831-1.0061783.json
JSON-LD: 831-1.0061783-ld.json
RDF/XML (Pretty): 831-1.0061783-rdf.xml
RDF/JSON: 831-1.0061783-rdf.json
Turtle: 831-1.0061783-turtle.txt
N-Triples: 831-1.0061783-rdf-ntriples.txt
Original Record: 831-1.0061783-source.json
Full Text

Full Text

MECHANISTIC STUDIES OF RABBIT MUSCLEGLYCOGEN PHOSPHORYLASEByWILLIAM G. STIRTANB. Sc. McMaster University, 1988A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIES(Department of Chemistry)We accept this thesis as conformingTHE UNIVERSITY OF BRITISH COLUMBIAJanuary 1993© William G. Stirtan, 1993In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)Department of  C L.9 XV) VS The University of British ColumbiaVancouver, CanadaDateDE-6 (2/88)ABSTRACTGlycogen phosphorylase catalyzes the reversible phosphorolysis of glycogen. Theenzyme contains a molecule of pyridoxal phosphate (PLP), of which the phosphate moietyis essential for catalysis. To investigate the role of the cofactor phosphate moiety twophosphonate analogues of PLP, 5'-deoxypyridoxal-5'-methylenephosphonic acid (9) and5'-deoxypyridoxal-5'-difluoromethylene phosphonic acid (13), were prepared andreconstituted into apoglycogen phosphorylase b. Both such reconstituted enzymes hadactivities of approximately 25% - 30% of the native enzyme, and moreover, despite theconsiderable difference in cofactor pKa, the pH-dependences of Vmax , Km and Vmax/Kmfor the two enzyme systems were quite similar. These results suggest that PLP does notfunction as an essential acid catalyst in glycogen phosphorylase; rather they suggest that thecofactor phosphate remains dianionic throughout catalysis. Evidence concerning theprotonation state of the PLP phosphate moiety within the R- and T-state enzymeconformations was obtained from solid-state magic angle spinning 31P NMR experimentsof the native enzyme. The cofactor phosphate moiety, in both allosteric forms of theenzyme, exhibited axially symmetric shielding parameters. These results suggest that thePLP phosphate moiety exists as a dianion in both enzyme conformations, and therefore,that a change in protonation state does not accompany the allosteric transition.The compound, 1-nitro-D-glucal, was found to irreversibly inactivate both the R-and T-state forms of glycogen phosphorylase, behaving as an active site-directed affinitylabel. The pH-dependence of the inactivation was consistent with 1-nitro-D-glucalundergoing a conjugate addition reaction at the C-2 position with an active site nucleophile,possibly Tyr-573. X-ray structural studies of the inactivated T-state enzyme identified Tyr-573 as a likely candidate for the inactivation, and further, identified a surface amino acid(His-73) covalently bonded to the C-2 position of a second molecule of the affinity label,the product of a conjugate addition reaction.The cofactor-substrate analogues, pyridoxa1-5'-pyrophospho-1-a-D-glucose(PLPP-Glu) (21) and pyridoxa1-5'-pyrophospho-1-(2-deoxy-2-fluoro)-a-D-glucose(PLPP-2FG1u) (23), were prepared and reconstituted into apoglycogen phosphorylase b.Results from kinetic, 19F NMR and cofactor abstraction studies demonstrated that PLPP-2FG1u, unlike PLPP-Glu, is completely resistant to decomposition at the enzyme activesite, and furthermore, does not transfer its glucosyl moiety to glycogen. Glycogenphosphorylase reconstituted with PLPP-2FG1u has allowed crystallization and initialstructural analysis of the activated enzyme / substrate complex.The solution binding interaction between glycogen and phosphorylase has beenprobed by 19F NMR using a glycogen analogue in which all the non-reducing terminalglucosyl residues have been replaced by 4-deoxy-4-fluoro-glucosyl moieties (4-F-glycogen). Results from titrating the fluorinated glycogen analogue with phosphorylasesuggest that 60 phosphorylase dimers bind per glycogen particle. 4-F-glycogen has alsobeen investigated as an "incompetent" substrate analogue in an attempt to provide evidencefor a glucosyl-enzyme intermediate in phosphorylase. Results from kinetic andradiolabelling studies demonstrated that the fluorinated glycogen analogue used was notcompletely inert to glucosyl transfer, but rather, that it possessed approximately 1% of theactivity observed with normal glycogen, most likely due to incomplete derivatization of theglycogen.TABLE OF CONTENTSABSTRACT^  iiTABLE OF CONTENTS^  ivLIST OF FIGURES  viiiLIST OF TABLES  xLIST OF SCHEMES^  xiLIST OF ABBREVIATIONS AND DEFINITIONS^  xiiACKNOWLEDGMENT  xivCHAPTER IGLYCOGEN PHOSPHORYLASE: INTRODUCTION1. General Introduction^  21.1. Function and Role  21.2.^Allostery  31.3.^Structural Information^  32. The Catalytic Mechanism  72.1. General Information  72.2. Evidence for Oxocarbonium ion-like Transition States^ 102.3. Evidence for Acid-Base Catalysis^  122.4. The Active Site; Covalent Enzyme Intermediate or Stabilized Ion Pair?^ 143. Pyridoxal Phosphate: The Essential Cofactor  163.1. General Role and Use of Cofactor Analogues^  163.2. The "Interacting Phosphates" Hypothesis  193.3. 31 P NMR Studies^  223.4. The Role of PLP in Catalysis?^  264. The Aims of This Study  31CHAPTER IITHE ROLE OF PYRIDOXAL PHOSPHATE IN GLYCOGEN PHOSPHORYLASE:RESULTS AND DISCUSSION1. Synthesis^  341.1. PLP Phosphonate Analogues^  341.2. 5'-Deoxy-5'-difluoromethylpyridoxal  362. Kinetic Studies with Reconstituted Glycogen Phosphorylase b^ 372.1. Kinetic Constants for 5-CH2PLP- and 5-CF2PLP-phosphorylase b^ 37iv2.2. Phosphonic Acid Ionizations and pH/Rate Profiles^ 392.2.1. Phosphonate pK2 Values for 5-CH2PLP and 5-CF2PLP^ 392.2.2. pH-Dependence of Phosphorylase b Reconstituted with5- CH2PLP and 5-CF2PLP^  413. NMR Experiments of Phosphorylase b Reconstituted with 5-CF2PLP^ 453.1^31 P NMR Experiments^  453.2. 19 F NMR Experiments  494. 31P MASNMR Spectroscopy of R- and T-state Glycogen Phosphorylase b^ 515. Discussion^  57CHAPTER IIIINACTIVATION OF GLYCOGEN PHOSPHORYLASE BY A NOVEL AFFINITYLABEL: RESULTS AND DISCUSSION1. Inactivation of Glycogen Phosphorylase b with Nitroglucal^ 721.1. Background and Synthesis^  721.2. Preliminary Inactivation Studies  731.3. Inactivation Kinetics for R- and T-state Glycogen Phosphorylase b^ 751.4. Active Site-Directed Inactivation^  781.5. pH-Dependence of the Inactivation Kinetic Constants (ki and Ki)^ 811.6. Structural Analysis of the Nitroglucal-Glycogen Phosphorylase b Complex^ 851.7. Mass Spectral Analysis of Nitroglucal Inactivated GlycogenPhosphorylase b^  882. Discussion^  89CHAPTER IVA NOVEL COFACTOR-SUBSTRATE ANALOGUE FOR GLYCOGENPHOSPHORYLASE: RESULTS AND DISCUSSION1. Synthesis^  992. Rates of Reactivation of Glycogen Phosphorylase b Reconstituted with PLPP-Gluand PLPP-2FG1u^  1013. 19F NMR and Cofactor Abstraction Studies of Glycogen Phosphorylase bReconstituted with PLPP-2FGlu^  1033.1. 19F NMR Experiments  1033.2. Cofactor Abstraction Study  1054. Structural Analysis of PLPP-2FG1u-Phosphorylase b^  1075. Discussion^  108vCHAPTER VBINDING INTERACTIONS AND SUBSTRATE ACTIVITY OF 4-FLUORO-GLYCOGEN: RESULTS AND DISCUSSION1. 4-F-Glycogen Binding Study^  1141.1. Titration of 4-F-Glycogen with Glycogen Phosphorylase b^ 1142. 4-F-Glycogen as an "Incompetent" Dead-End Substrate Analogue?  1182.1. 31 P NMR of PLPP-Glu-phosphorylase b in the Presence of 4-F-Glycogen^  1182.2. Mass Spectral Studies of PLPP-Glu-Phosphorylase b in the Presence of 4-F-Glycogen  1222.3. Transfer of Radiolabelled Glucose From PLPP-Glu-Phosphorylase b to 4-F-Glycogen^  1242.4. Kinetic Evaluation of 4-F-Glycogen with Native GlycogenPhosphorylase b^  1273. Discussion^  127CHAPTER VIMATERIALS AND METHODS1. Synthesis^  1341.1. General Procedures and Materials^  1342. Enzymology  1502.1. General Procedures^  1502.2. Protein Purification  1502.3. Resolution and Reconstitution of Glycogen Phosphorylase b^ 1512.4. Enzymic Synthesis of 4-F-Glycogen^  1512.5. Kinetic Experiments^  1532.5.1. General Procedures ^1532.5.2. Kinetic Evaluation of Glycogen Phosphorylase b Reconstituted with5 -CH2PLP and 5-CF2PLP^  1532.5.3. pH-Dependence of the Kinetic Parameters for GlycogenPhosphorylase b Reconstituted with 5-CH2PLP and 5-CF2PLP^ 1542.5.4. Reconstitution and Reactivation of ApoglycogenPhosphorylase b with PL-CF2H^  1552.5.5. Reactivation Kinetics for PLPP-Glu- and PLPP-2FG1u-Phosphorylase b^  155vi2.5.6. Incubation of the R- and T-State forms of GlycogenPhosphorylase b with Glucal Analogues^  1562.5.7. The Determination of ki and Ki for the Nitroglucal Inactivation ofGlycogen Phosphorylase b^  1562.5.8. Glucosyl Transfer from PLPP-[ 14C]-Glu-Phosphorylase b toGlycogen and 4-F-Glycogen  1573. NMR Experiments^  1593.1. General Procedures^  1593.2. 31 P NMR Titration of 5-CH2PLP and 5-CF2PLP^  1603.3. 31P NMR of Native Glycogen Phosphorylase b and of5-CF2PLP-Phosphorylase b^  1603.4. 19F NMR of 5-CF2PLP- and PL-CF2H-Phosphorylase b^ 1603.5. Solid State 31 P MASNMR Spectroscopy of free PLP and ofR- and T-state Glycogen Phosphorylase b^  1613.6. 19F NMR of PLPP-2FGlu-Phosphorylase b  1623.7. 19F NMR Titration of 4-F-Glycogen with Glycogen Phosphorylase b^ 1633.8. 31P NMR of PLPP-Glu-Phosphorylase b in the Presence of4-F-Glycogen^  1634. Analysis of Pyridoxal Cofactors from PLPP-2FG1u-Phosphorylase b^ 1645. Ion-Spray Mass Spectral Studies^  165APPENDIX IKINETIC DERIVATIONS AND pH-DEPENDENCE OF ENZYME REACTIONS1. Kinetic Derivations^  1672. pH-Dependence of Enzyme Reactions^  169REFERENCES^  173viiLIST OF FIGURESFigure 1-1.Figure 1-2.Figure 2-1.Figure 2-2.Figure 2-3.Figure 2-4.Figure 2-5.Figure 2-6.Figure 2-7.Figure 2-8.Figure 2-9.Figure 2-10.Figure 3-1.Figure 3-2.Figure 3-3.Figure 3-4.Figure 3-5.Figure 3-6.Figure 3-7.Figure 4-1.Figure 4-2.Figure 4-3.Figure 5-1.Figure 5-2.A schematic ribbon diagram of the T-state glycogen phosphorylase bdimer^  4Inhibition of phosphorylase by transition state analogues^ 12Determination of the kinetic parameters for 5-CH2PLP-phosphorylase b and5-CF2PLP phosphorylase b^  3731p NMR titration of phosphonate cofactor analogues^ 40The pH-dependence of Vmax , Km , and Vmax/Km for5-CH2PLP-phosphorylase b and 5-CF2PLP-phosphorylase b^ 43The structure of a G 1 CP^  4531P NMR spectra of glycogen phosphorylase b reconstituted with5-CF2PLP and in its native form^  4619F NMR spectra of glycogen phosphorylase b reconstituted with5-CF2PLP^ 4931P MASNMR spectra of free PLP and the R- and T-state formsof glycogen phosphorylase b^  54An earlier assignment of the changes in cofactor ionization statethought to accompany the allosteric transition in glycogen phosphorylase ^ 62Ground state destabilization^  66The 'under the ring' conformation for aG1P^  68Inactivation of glycogen phosphorylase b with glucal analogues^ 74Nitroglucal inactivation kinetics for T-state glycogen phosphorylase b^ 76Nitroglucal inactivation kinetics for R-state glycogen phosphorylase b^ 77Protection against nitroglucal inactivation by a competitive ligand^ 79Nitroglucal inactivation kinetics over a range of pH^ 82The pH-dependence of nitroglucal inactivation  83X-ray crystallographic analysis of T-state glycogen phosphorylase b inthe presence of nitroglucal^  87Time course for reactivation of PLPP-Glu- andPLPP-2FG1u-phosphorylase b^  10219F NMR spectra of glycogen phosphorylase b reconstitutedwith PLPP-2FGlu^  104Analysis of pyridoxal compounds from PLPP-2FG1u-phosphorylase.b^ 10619F NMR titration of 4-F-glycogen with glycogen phosphorylase b^ 115Plot of peak heights of 19F signals versus enzyme concentration^ 117viiiFigure 5-3. 31P NMR spectra of glycogen phosphorylase b reconstituted withPLPP-Glu^  120i xLIST OF TABLESTable 1-1.^Reconstitution of apophosphorylase b with various analogues of PLP^ 18Table 2-1.^Kinetic parameters for glycogen phosphorylase b reconstitutedwith 5-CH2PLP and 5-CF2PLP^ 38Table 2-2.^Principal components of the 31P shielding tensors for free PLP andfor the cofactor phosphate in R- and T-state glycogen phosphorylase.b^ 55Table 5-1.^Mass spectral data collected for glycogen phosphorylase b^ 123Table 5-2.^Radiolabel incorporation from PLPP-[ 14C]-Glu-phosphorylase b^ 125XLIST OF SCHEMESScheme 1-1. The reaction catalyzed by glycogen phosphorylase^ 2Scheme 1-2. A general mechanism for glycogen phosphorylase  8Scheme 1-3. The kinetic mechanism for glycogen phosphorylase^ 9Scheme 1-4. Utilization of glucal and heptenitol in phosphorylase  13Scheme 1-5. Positional isotope exchange between bridging and non-bridgingoxygens via an enzyme intermediate^  15Scheme 1-6. The reaction catalyzed by PLPP-Glu-phosphorylase^ 21Scheme 1-7. The acid catalyzed mechanism proposed by Palm et al., 1990^ 27Scheme 1-8. The proposed electrophilic mechanism for glycogen phosphorylase^ 30Scheme 2-1. The synthetic route for preparation of 5-CH2PLP and 5-CF2PLP^ 35Scheme 2-2. The synthetic route for preparation of PL-CF2H^ 36Scheme 2-3. A model for the decarboxylation reaction catalyzed bypyruvate decarboxylase^  67Scheme 2-4. A possible catalytic role for PLP in glycogen phosphorylase^ 70Scheme 3-1. Analogues of D-glucal  73Scheme 3-2. A kinetic model for the irreversible inactivation of glycogenphosphorylase b by nitroglucal^  80Scheme 4-1. Two alternate reactions for PLPP-Glu-phosphorylase b^ 98Scheme 4-2. The synthetic route for the preparation of PLPP-Glu and PLPP-2FG1u^ 100Scheme 5-1. 4-F-Glycogen as an "incompetent" substrate analogue^ 113xiLIST OF ABBREVIATIONS AND DEFINITIONSAbbreviationsADP^Adenosine-5'-diphosphateAMP^Adenosine-5'-monophosphateAMPS^Adenosine-5'-thiomonophosphateAsn^AsparagineAsp^Aspartic acidATP Adenosine-5'-triphosphateBES^2-[bis(2-hydroxyethypamino] ethanesulphonic acid5-CF2PLP^5'-Deoxypyridoxal-5'-difluoromethylenephosphonic acid5-CH2PLP^5'-Deoxypyridoxal-5'-methylenephosphonic acidddH2O^Double deionized waterD.C.I.^Desorption Chemical IonizationDTT DithiothreitolEDTA^Ethylenediaminetetraacetic acidF.A.B.^Fast Atom Bombardment4-F-glycogen 4-Deoxy-4-fluoro-glycogen2F-aG1P^2-Deoxy-2-fluoro-a-D-Glucose-l-phosphate (2-Deoxy-2-fluoro-a-D-glucopyranosyl phosphate)aG1CPaG1PGluHEPESHisKIELCLC-MSLysMESm.p.NADPNitroglucalNMRPi(1-Deoxy-a-D-glucopyranosyl) methylphosphonatea-D-Glucose- 1-phosphate (a-D-glucopyranosyl phosphate)Glutamic acid(N-2-Hydroxyethyl piperazine N'-2-ethanesulfonic acid)HistidineKinetic isotope effectLiquid chromatographyLiquid chromatography - mass spectrometryLysine(2[N-Morpholino] ethane sulfonic acid)Melting pointNicotinamide adenine dinucleotide phosphate1-Nitro-D-glucalNuclear magnetic resonanceOrthophosphatePLPL-CF2HPLPPLPPPLPP-2FG1uPLPP-GluTCATFATyrPyridoxal5'-DeoxydifluoromethylpyridoxalPyridoxal-5'-phosphatePyridoxal-5'-pyrophosphatePyridoxal-5'-pyrophospho-a-1-(2-deoxy-2-fluoro)-D-glucosePyridoxal-5'-pyrophospho-l-a-D-glucoseTrichloroacetic acidTrifluoroacetic acidTyrosineKinetic ConstantsKi^Dissociation constant for the enzyme-inhibitor complexki First-order rate constant of inactivationKm^Michaelis-Menten constantkcal First-order rate constant for catalysisVmax^Maximal rate of an enzyme-catalyzed reactionACKNOWI .FDGMENTI would like to express my gratitude to my supervisor, Professor Stephen Withers,for his advice and encouragement over the years. Thanks also to Professor C. A.McDowell, Dr. Robin Challoner, Professors Neil Madsen, Steve Sprang, and LouiseJohnson for their collaborative efforts, without which this work would not have beenpossible. I also thank my coworkers for their suggestions and friendship, especially KarenRupitz for her technical support and advice. Thanks also to the technical staff in theDepartment of Chemistry for their assistance, and especially Dr. K. Wong in theDepartment of Biochemistry for his assistance.Special thanks to Doug "Spanker" Spraklin, for his friendship over the years.Finally, I must dedicate this thesis to my parents, my sister Julie-Bird, my wifeKim, and baby Jesse, for their love, constant support and patience.x i vCHAPTER 1Glycogen PhosphorylaseIntroduction11. General Introduction1.1. Function and RoleGlycogen phosphorylase (a-1,4-glucan-orthophosphate glucosyl transferase, EC2.4.1.1) is a well characterized enzyme, found widely distributed in nature. The primaryfunction of phosphorylase is to meet cellular demands to convert glycogen andorthophosphate (Pi) to a-D-glucopyranosyl phosphate (aG1P). 0I IHO—P-0"OOHHOBGLYCOGENHOHOOH^Glycogen PhosphorylaseOH^ HOO HO •HOHO^00-Scheme 1 -1. The reaction catalyzed by glycogen phosphorylase.The equilibrium constant for the phosphorolysis of glycogen is 0.28 at pH 6.8 (Cori andCori, 1940) and therefore lies slightly in favor of glycogen synthesis. However, in -vivo,where the concentration of phosphate is considerably greater 100 fold) than aG1P, thereaction is driven in the direction of glycogen degradation (Lanier et al., 1960), thusproviding the cell with a regulated supply of phosphorylated glucose units.HOGLYCOGEN21.2. AllosteryGlycogen phosphorylase is a complex allosteric enzyme which makes use of avariety of allosteric mechanisms for control of its activity. The enzyme assumes at leasttwo conformations, one of which is active. Equilibrium among these conformations isregulated by "effector" molecules which interact at specific sites on the protein and promotea catalytically active R-conformation or alternatively, an inactive T-conformation. Theenzyme, first purified from rabbit skeletal muscle and termed phosphorylase b, was foundto be dependent on adenosine-5'-monophosphate (AMP) for activity (Cori and Cori,1936), but could also be inhibited by several negative effectors such as ATP, ADP,glucose-6-phosphate and glucose. Thus, phosphorylase b normally exists in the inactiveT-state unless activated by AMP. In 1941 a second form of phosphorylase was isolatedand found to be active in the absence of AMP (Green et al., 1942). This form of theenzyme, termed phosphorylase a, is predominantly found in the active R-state and isactivated by the phosphorylation of phosphorylase b at the hydroxyl group of a serineresidue at position 14 on the protein (Fischer and Krebs, 1955; Fischer et al., 1959). Thephosphorylation reaction that converts phosphorylase b into active phosphorylase a iseffected by a specific protein kinase (Fischer and Krebs, 1955), which itself is subject tohormonal and nervous control mechanisms (Madsen and Withers, 1986).1.3. Structural InformationThe phosphorylase monomer consists of 842 amino acids with a molecular weightof 97,444 (Johnson, 1992). In solution, phosphorylase can exist as a dimer or tetramercomposed of identical subunits (Madsen and Cori, 1956). The equilibrium between thetwo forms of the enzyme is dependent on the extent of enzymic activation (Madsen andCori, 1957). In vitro, in the absence of glycogen and any other effectors, the covalentlyactivated phosphorylase a molecule is present as a tetramer, while phosphorylase b ispresent as an inactive dimer (Dombradi et al., 1981) (Figure 1-1).3Cotelytsc *rte .Glycogenetoesge sots'Alioetericeffector site'Copill^ ^Cetlectot e^teGlycogenelvers s■teCotoiyticFigure 1-1. A schematic ribbon diagram of the T-state glycogen phosphorylase b dimer.Johnson et al., 1990.4Once activated by AMP, however, the phosphorylase b dimers associate into tetramers, anassociation which can be reversed by negative allosteric effectors such as glucose andcaffeine (Withers et al., 1979). Each tetrameric species, phosphorylase a or b, once in thepresence of glycogen dissociates into the fully active R-state dimers found in -vivo (Wangand Graves, 1964; Wang et al., 1965).X-ray crystallographic studies have been carried out on both phosphorylase a andb, providing extensive structural information. The two structures have been solved to aresolution of 2.1 A for phosphorylase a (Sprang et al., 1979; Sprang et al., 1982) and 1.9A for phosphorylase b (Johnson et al., 1987; Acharya et al., 1991). Each monomer iscomposed of two domains, the N- (1-489) and C- (490-842) terminal domains, bothconsisting of a-helical segments (45%), and 0-sheet regions (25%). Arguably the mostvaluable benefit of the past crystallographic studies was the identification of the primaryligand binding sites located throughout the enzyme (Figure 1-1). Thus, the AMP bindingsite was located in the N-terminal domain, in close proximity to the serine-14-phosphateresidue where covalent activation occurs. Also found in the N-terminal domain was theglycogen "storage" site where oligosaccharides such as maltoheptaose are found to bind inthe crystal structure (Kasvinsky et al., 1978). Kinetic studies have shown that glycogenbinds to the storage site with higher affinity than does its individual oligosaccharide chainsto the active site (Kasvinsky et al., 1978). Thus, the storage site is thought to serve as ananchor point which firmly binds the enzyme to the glycogen particle thereby providing ahigh local concentration of oligosaccharide chains in the region of the active site. Theremainder of the ligand binding sites are located at the juncture of the N- and C-terminaldomains. An inhibitor site, or I site, located at the entrance to the active site binds aromaticmolecules such as caffeine, thus blocking catalysis. The exact physiological role of the Isite is not yet fully understood. The catalytic site binds both glucose (a negative effector)and the glucose portion of the substrate (aG1P) in the same high specificity pocket, withthe phosphate portion of aG1P binding to a subsite directly adjacent. The remainder of the5catalytic site is occupied by an essential cofactor molecule, pyridoxal phosphate (PLP),covalently bound to the enzyme through a Schiff base to lysine-680 (Lys-680). Thepyridoxal ring of PLP is located in a hydrophobic pocket, in contrast the phosphate moietyinteracts with the solvent and with several basic residues and is oriented directly toward thesubstrates phosphate binding site (Madsen and Withers, 1986).Initially, all structural information originating from X-ray crystallographic studiescame from crystal forms in the inactive T-state conformation since the active R-state formswere not easily crystallized. T-State phosphorylase a was crystallized in the presence ofglucose, which stabilizes the inactive conformation, whereas phosphorylase b crystals weregrown using a very weak activator of the enzyme, inosine-5'-monophosphate, and thesecrystals were also considered to be predominantly T-state. It was hoped that comparison ofthe two T-state structures would identify the conformational changes that result from thephosphorylation at serine-14 in phosphorylase a. Unfortunately, any uniqueconformational changes communicated to the active sites of these two protein structureswere considerably restricted due to the crystal lattice packing forces (Sprang et al., 1988).In fact, diffusing substrate (an R-state effector) into the active site of T-state phosphorylasea crystals results in crystal cracking and disintegration of the crystal lattice (Sprang et al.,1982), an indication of gross conformational changes which accompany activation of theenzyme.Recently however, under very different conditions, two partially activated R-stateforms of phosphorylase have been crystallized and the structures for both crystal formshave been solved (Barford and Johnson, 1989; Barford et al., 1990; Sprang et al., 1991).The first of these R-state crystals, solved to 2.9 A, was grown in the presence of 1.0 - 1.2M ammonium sulphate. Sulphate acts as a mimic of phosphate (an R state effector) andbinds at the AMP site and the serine-14 phosphate site, resulting in the T- to R-statetransition. It had been shown previously that tetrameric crystals of phosphorylase a or bcould be obtained in the presence of ammonium sulphate, however, the crystal structure,6until recently, had not been solved. The second R-state crystal form, solved to 3 Aresolution, was crystallized through the use of a cofactor analogue, pyridoxa1-5'-pyrophosphate (PLPP), and requires AMP to induce the active R-state conformation(Sprang et al., 1991). In this case, the second phosphate of the cofactor analogue mimicsthe substrate phosphate and binds in its absence. Even though both sets of crystals weregrown under entirely different conditions, in general, the quaternary structure and AMPbinding interactions for the two R-state crystal forms are quite similar (Johnson, 1992).Unfortunately, because both R-state crystals contain tetrameric enzyme, oligosaccharide isoccluded from binding to the active site, and therefore, it is unlikely that they represent thefully active R-state dimers found in -vivo. Thus, a fully activated form of phosphorylase,wherein oligosaccharide is bound at the active site of the enzyme has yet to be observed inthe crystalline state.2. The Catalytic Mechanism 2.1. General InformationThe catalytic mechanism for glycogen phosphorylase in the direction of glycogensynthesis, consists of a glucosyl transfer from phosphate to the non-reducing terminalsugar of glycogen. The reaction is thought to involve a double displacement, possiblyemploying a covalent enzyme intermediate, thereby accounting for the observed retention ofconfiguration at the anomeric center and cleavage of the anomeric carbon-oxygen bond.7HO A- GLYCOGEN0HO—P--O-oI -^HO HO ^HO OH1^OHHO?-H^HO GLYCOGENScheme 1-2. A general mechanism for glycogen phosphorylase.The main features of the catalytic mechanism are thought to include general acid catalysis tolabilize the glycosidic linkage, stabilization of the incipient glucosyl cation by covalent orelectrostatic means, base catalysis to assist the incoming nucleophile, and finally bondformation between the glucosyl anomeric center and the oligosaccharide 4-hydroxyl group(or phosphate oxygen in the reverse direction).The kinetic mechanism for glycogen phosphorylase has been derived from initialvelocity measurements, inhibition studies, and equilibrium isotopic exchange reactions(Engers et al., 1969, 1970a, 1970b). The enzyme can best be described as having a rapidequilibrium random bi-bi kinetic mechanism whereby isomerization of the enzyme ternary8complex is slow compared to dissociation of the enzyme substrate complex (Scheme 1-3).When one of the substrates is present at saturating concentrations (e.g. glycogen) the modelsimplifies to that of a single substrate system.E-Gn^E-Gn+1/ \\N / \E^E-GnGP —.----='',- E-Gn+i P^EE-GP/^\ E -13 /Scheme 1-3. The kinetic mechanism for glycogen phosphorylase.E-GP is the enzyme-aG1P complex, E-G, is the enzyme glycogen complex, E-P is the enzyme phosphatecomplex, and E-GnGP and E-Gn+ iP are the ternary enzyme substrate complexes.From the above kinetic scheme it can be seen that glycogen phosphorylase requires anactive ternary complex, wherein all substrates (aG1P or Pi, and glycogen) are present atthe active site, before product formation can occur (Engers et al., 1969). Indeed, earlyexperiments (Cohn and Cori, 1948) have shown that no isotopic exchange between aG1Pand 32P-[Pi] occurs in the absence of the second substrate (e.g. glycogen), thus supportingthe proposal that no bond making or breaking can occur in phosphorylase catalysis until anactive ternary complex is formed.92.2. Evidence for Oxocarbonium ion-like Transition StatesOxocarbonium ion-like transition states are thought to precede and follow theformation of a glucosyl enzyme intermediate (or stabilized ion pair) in the phosphorylasemechanism. Good evidence for the involvement of a charged transition state comes fromthe work of Street et al., (1989) where a series of deoxy- and deoxyfluoro-aG 1Panalogues were synthesized and tested as substrates in the phosphorylase reaction. Allsubstrates were found to be utilized by the enzyme, however, at considerably reduced ratesas a consequence of electronic and binding effects at the transition state for the rate limitingstep. The substituent effects on the enzymic reaction were compared with those sameeffects on the model non-enzymic reaction, namely acid catalyzed hydrolysis of the sameseries of compounds. A logarithmic plot of the rate constants for the enzymic reactionversus those for acid catalyzed hydrolysis yielded a linear free energy relationship with acorrelation coefficient p = 0.90, indicating considerable similarities in the electronic natureof these two transition states. Since the acid catalyzed hydrolysis of aG1P is thought toproceed via a mechanism involving an oxocarbonium ion intermediate or at least a transitionstate with substantial oxocarbonium ion character (Bunton and Humeres, 1969), this linearfree energy relationship suggests that the phosphorylase reaction proceeds through a similaroxocarbonium ion-like transition state.Another line of evidence often used to detect oxocarbonium ion-like transition statesis the study of kinetic isotope effects (KIE). Two such studies with phosphorylase,investigating the a-secondary KIE with modified substrates have shown contrastingresults. Tu et al., (1971) separately measured the phosphorylase reaction rate for C-1protio and C-1 deuterio aG1P and observed an a-secondary deuterium KLE of 1.10(kH/kD = 1.10). This result is consistent with a mechanism involving considerable sp 2character on forming the transition state at the rate limiting step. Later, however, Firsov etal., (1974) improved on the experimental procedure by synthesizing doubly labelled aG1Psubstrates. This allowed the kinetic experiments to be performed on a single solution10where a direct competition between substrates could be established. Firsov et al., (1974)were able to demonstrate that the deuterium and tritium KIE, for both the forward andreverse reaction was zero (kH/kD = kH/kT = 1.0). This finding was more in keeping withan SN2-like transition state, and the methodology employed circumvents the problem ofphosphorylase being sensitive to numerous inhibitors. However, the results obtained fromKIE studies with phosphorylase are only valid if the rate limiting step in catalysis isformation or breakdown of the putative glucosyl-enzyme intermediate (or stabilized ionpair).The use of tight binding transition state analogues which mimic the transition statefor the natural substrate have also been used to obtain support for an oxocarbonium ion-liketransition state in phosphorylase. Since these studies measure only binding, the problem ofdetermining which catalytic step is rate limiting is avoided. D-Glucono-(1,5)-lactone(Figure 1-2) possesses a half chair conformation with possibly some positive character onthe endocyclic oxygen and negative character on the exocyclic oxygen, similar to thatproposed for the intermediate oxocarbonium ion in the acid catalyzed hydrolysis of aG1P.Thus, D-glucono-(1,5)-lactone may act as a transition state analogue for the substrateportion of the phosphorylase reaction if indeed phosphorylase proceeds through a similartransition state. D-Glucono-(1,5)-lactone was found to bind to the enzyme-glycogencomplex, in direct competition for the glucose binding site of aG1P (Tu et al.,1971).However, the binding was found to be quite weak for a transition state analogue (K1= 1mM), when compared to the dissociation constant for glucose (K1= 2 mM). A short timelater it was found (Gold et al., 1971), that D-glucono-(1,5)-lactone bound much moretightly to the enzyme-glycogen-Pi complex (Ki = 0.025 mM), an affinity two orders ofmagnitude greater than the substrate (Km = 2 - 3 mM), and more in line with that expectedfor a transition state analogue. Thus it would seem that to truly mimic the transition statefor the phosphorylase reaction, phosphate must be present to bind at the aG1P phosphatebinding site such that the enzyme can fully adopt an activated conformation. More recently,11another transition state analogue locked in the half chair conformation, norjirimycintetrazole (Figure 1-2), has been shown to bind the enzyme-glycogen-Pi complex with highaffinity (Ki = 0.014 mM) (Withers and Vasella, Unpublished results). That these glucosylanalogues bind phosphorylase with such high affinity in the presence of phosphateprovides further support for the enzyme passing through transition states with substantialoxocarbonium ion character.OH/,OH^ OH0   =NHO^8+^HOHON'N /HO^NOxocarbonium ion-like^Glucono-(1,5)-lactone^Nojirimycin tetrazoletransition stateFigure 1-2. Inhibition of phosphorylase by transition state analogues.2.3. Evidence for Acid-Base CatalysisGlycosylic substrates, such as glucal or hepenitol, have been shown previously tobe efficient substrates for phosphorylase provided the necessary co-substrates (Pi andoligosaccharide) are present (Klein et al., 1982; Klein et al., 1986). The common catalyticfeature between these substrates is the reactive enolic double bond which can be protonatedby general acid catalysis and subsequently react with phosphate, probably in a concertedfashion (Scheme 1-4).HO HOOR8 _HO^ HO^ HO12OHHOOHI-0—P-0"II0HOHOOHHOHO ^0"IHO—P-0-II0Glucal + phosphate Heptenitol + phosphateScheme 1-4. Utilization of glucal and heptenitol in phosphorylase.Klein et al., (1982) were able to demonstrate that the products formed from incubatingphosphorylase with D-glucal, Pi and oligosaccharide were 2-deoxyglucosylated-polysaccharide and 2-deoxy-aG1P, the latter likely being derived from the enzymicdegradation of the modified oligosaccharide. NMR studies employing deuteriumincorporation from D20 were used to show that, indeed, D-glucal was protonatedstereospecifically into the equatorial position of C-2, thus from below the C-2 position ofthe pyranose ring.The other glycosylic substrate, hepenitol, possesses a reactive exocyclic doublebond. In the presence of phosphate and heptenitol, phosphorylase catalyzes the formationof heptulose-2-phosphate, a potent inhibitor of phosphorylase (Klein et al., 1986). In thiscase, deuterium incorporation studies showed that deuterium uptake in the enzymecatalyzed reaction occurred at the methylene position, producing a deuteromethyl (CH2D)group. Heptenitol is a unique substrate in that it is the first known substrate forphosphorylase which does not require a primer (oligosaccharide) (Klein et al., 1986),suggesting that heptenitol is used exclusively in the degradative pathway where it competeswith oligosaccharide substrates. That both glucal and heptenitol are utilized by13phosphorylase is considered supportive evidence for general acid-base catalysis in thenormal phosphorylase reaction (Klein et al., 1984).2.4. The Active Site; Covalent Enzyme Intermediate or Stabilized Ion Pair?Catalysis by glycogen phosphorylase proceeds with overall retention ofconfiguration at the anomeric center, the C-0 bond being the scissile bond. One keyquestion has always been whether the anomeric center is involved in a doubledisplacement, with the production of a stable enzyme intermediate, as seen with sucrosephosphorylase (Voet and Abeles, 1970), or if the leaving group departs in a more SN1-likefashion whereby a stabilized ion pair is formed and then broken down by an incomingnucleophile. In some cases, a fully concerted (SNi) mechanism excluding any intermediatehas been proposed whereby direct front-side attack of the nucleophile occurs as the leavinggroup departs (Kokesh and Kakuda, 1977; Palm et al., 1990).In the case of sucrose phosphorylase, the involvement of an intermediate has beenclearly demonstrated with the exchange of phosphate groups between aG1P and inorganicphosphate in the absence of an acceptor molecule (Doudoroff et al., 1947). Moreover,Voet and Abeles, (1970) were able to capture the glucosyl-enzyme intermediate. Thekinetic mechanism for sucrose phosphorylase can be described as 'ping-pong', wherebyaG1P can bind to the active site of the enzyme and transfer the glucosyl moiety to asuitable enzymic nucleophile before the acceptor molecule (fructose for example) binds tothe enzyme and completes the glucosyl transfer reaction. Similar evidence has not beenobtained with potato or rabbit muscle phosphorylase since the kinetic mechanism for theseenzymes (as discussed earlier) requires that all substrates be bound at the active site beforeany reaction can occur. Consequently, any enzyme intermediate formed in these enzymeswould never accumulate due to rapid turnover. Thus, in order to use similar methodologyon potato or rabbit muscle phosphorylase, experiments must be developed which allow theformation of the enzyme ternary complex but prevent enzymic turnover of any enzyme14A1intermediate formed. This has been demonstrated in potato phosphorylase through the useof a-cyclodextrin, a cyclic oligosaccharide of glucose which can mimic the binding ofstarch and therefore activate potato phosphorylase, but which can not act as a glucosylacceptor since it lacks the necessary free 4-hydroxyl group (Kokesh and Kakuda, 1977).Kokesh and Kakuda (1977) were thus able to provide the first evidence consistent with aglucosyl-enzyme intermediate in potato phosphorylase when they demonstrated that theenzyme could catalyze the positional isotope exchange of bridging phosphate oxygens in180-labelled aG1P provided a-cyclodextrin was present.%■Nuc)HOHO 0OHHO0(.*.P —OHoI-OH)■HOH    HOHO00—P —OHI*0-Scheme 1-5. Positional isotope exchange between bridging and non-bridgimg oxygens viaan enzyme intermediate.HOHO0-0—P —OHoI *15Support for this work was later found when Klein et al., (1981) incubated potatophosphorylase with radioactive glucal, a-cyclodextrin and arsenate (an analogue ofphosphate) and observed incorporation of up to 1 mol of radioactive glucal per mol ofenzyme subunit. Unfortunately however, since that time, Klein and coworkers havequestioned these results and no longer consider that a glucosyl-enzyme intermediate is partof the potato or rabbit muscle phosphorylase mechanism (Helmreich, 1992).The amino acid sequence of potato phosphorylase is highly homologous to therabbit muscle enzyme, showing 51% identity of amino acid residues. Further, the highlyconserved regions are found at the active site suggesting that the catalytic machinery andpossibly the reaction mechanisms for the two enzymes are very similar. Indeed, Withersand Rupitz, (1990) have used a series of deoxy- and deoxyfluoro-aG1P substrates in alinear free energy study to demonstrate that the electronic structure of the transition stateand the hydrogen bonding interactions between the enzyme and substrate at the transitionstate for the two enzymic reactions are extremely similar. Thus, it might be expected thatrabbit muscle glycogen phosphorylase proceeds through the same putative enzymicintermediate as potato phosphorylase.3. Pyridoxal Phosphate: The Essential Cofactor 3.1. General Role and Use of Cofactor AnaloguesIn 1957 glycogen phosphorylase was first found to contain one molecule ofpyridoxal phosphate (PLP) per enzyme subunit (Baranowski et al., 1957). A short timelater the apoenzyme (PLP removed) was shown to be inactive, but could regain activityupon reconstitution with PLP (Illingworth et al., 1958). The PLP coenzyme is covalentlylinked to phosphorylase via a neutral tautomer of the Schiff base to Lys-680, where it actsin a structural role maintaining quaternary structure (i.e. subunit interactions), and acatalytic role, through its catalytically essential phosphate moiety. In fact, the role of PLPin phosphorylase is unique amoung PLP-containing enzymes in that the imine linkage16which covalently bonds the enzyme and cofactor can be reduced, with the enzyme retainingmost of its catalytic activity (Fischer et al., 1958). The stuctural and catalytic role of thecofactor was first delineated with the observation that pyridoxal (PL) or 5'-deoxypyridoxal,both of which are lacking the 5'-phosphate moiety, could be used to reconstitute theapoenzyme, restoring quaternary structure and allosteric properties but no catalytic activity(Kastenschmidt et al., 1968). Later, it was demonstrated that monomers of the enzymereconstitued with PLP analogues that did not restore activity, such as pyridoxal, could behybridized with monomers of the native enzyme to form active phosphorylase dimershaving 50% of the normal specific activity (Feldman et al., 1976). Since monomers of thenative enzyme are inactive under normal conditions, this study suggested that the"incompetent" monomers were able to confer the correct conformation on to the"competent" monomers. Thus, the structural role of the cofactor is effected through itsability to bind at Lys-680, whereas the catalytic function resides in the phosphate moiety.Over the years, in an effort to test every part of the PLP molecule for its role in thephosphorylase mechanism, several derivatives have been synthesized and tested for theirability to bind to (i.e. restore quaternary structure) and reactivate the apoenzyme. Aselection of these analogues is presented in Table 1-1. The reconstitution studies shown inTable 1-1 demonstrate that portions of the PLP molecule are clearly not required forcatalytic activity. The activity of the 3-0-methyl derivative effectively excludes the 3-OHgroup in any catalytic role. In addition, the nitrogen of the pyridine ring is unlikely to beinvolved in catalysis since 6-fluoro-PLP is active, yet the fluorine reduces the pyridine pKaby over 6 units. Conversely, the essential regions of the cofactor include the 4-positionwhich must possess an aldehyde functionality for cofactor binding, and also the 5'-positionwhich, in order to restore catalytic activity, must contain a phosphate derivative capable offorming a dianion.17Table 1-1 . Reconstitution of apophosphorylase b with analogues of PLPCHOCHT-OP032-Pyridoxal Analogue ModifiedPositionModification %ReactivationEvidenceforBindingNone — — 0 —Pyridoxal 5'-phosphate — — 100 +N-methyl PLP 1 -CH3 0 -N-Oxide PLP 1 —N-0 25* +2-Nor PLP 2 —H 65 +3-0-methyl PLP 3 —OCH3 25 +3-0-methyl N-oxide PLP 3& 1 -OCH3 & N-0 0 +Pyridoxamine 5'-phosphatephosphate44-4-deoxypyridoxine-CH2NH2—CH300 -Lys-679 linkage reduced 4 -CH2—NH2 60 -4-vinyl PLP 4 -C=CH2 7 +Pyridoxal 5 —CH2OH 0 +Pyridoxal + Phosphite 5 HPO3 24 +Pyridoxal + Fluorophosphate 5 HFPO3 12 +Deoxypyridoxalyl methanephosphonic acid5 —CH2PO3 25 +Pyridoxal 5'-sulphate 5 -CH2OSO2 0 +Pyridoxal 5'-fluorophosphate 5 -CH2OPFO2 0 +Pyridoxal methyl ester 5 -CH2PO3OMe 0 +6-Methyl PLP 6 -CH3 8 +6-Fluoro PLP 6 —F 28 +Table from Madsen and Withers (1986). * Some of the bound derivative reverts to the natural coenzyme.18Thus the methyl ester of PLP and pyridoxal-5'-fluorophosphate are both inactivederivatives. However the phosphonic acid derivative, which is able to form a dianion,restores the apoenzyme to 25% of the native activity. Furthermore, if apophosphorylase isreconstituted with pyridoxal (PL-phosphorylase), the enzyme can be reactivated by theaddition of phosphate, phosphite or fluorophosphate. It is thought that phosphate and itsanalogues activate the enzyme by binding at the vacant cofactor phosphate binding siteadjacent to the pyridoxal ring, acting in its absence. Indeed, this proposal has beensupported by crystallographic evidence which suggests that phosphite binds to PL-phosphorylase in a site that is only 1 A removed from the normal PLP phosphate bindingsite (Oikonomakos et al., 1987). Studies such as these have served to focus attention onthe coenzyme phosphate as the principal component involved in catalysis. For reviews ofthis work see (Madsen and Withers 1986; Parrish et al., 1977; Graves and Wang, 1972).3.2. The "Interacting Phosphates" HypothesisThe PLP replacement studies described above suggest that the coenzyme phosphatehas an essential catalytic role at the active site of glycogen phosphorylase. There is furtherevidence from X-ray crystallographic and solution studies indicating that the cofactorphosphate and substrate phosphate interact directly in some way. Using X-raycrystallography, Withers et al., (1982c) were able to show an inter-phosphate separation ofsome 6.8 A between the cofactor phosphate in phosphorylase a and the bound substrateanalogue glucose-1,2-cyclic phosphate. This distance was reduced to approximately 4.8 Ain both the phosphorylase b-heptulose-2-phosphate complex (McLaughlin et al., 1984) andthe nojirimycin tetrazole-phosphate complex (E. Mitchell, personal communication),wherein both analogues are thought to mimic the transition state for the glucosyl transferreaction. Moreover, as these crystal structures are those of predominantly inactive T-stateforms, it cannot be ruled out that the inter-phosphate distance between the cofactor andsubstrate is further reduced on going to the fully active R-state enzyme.19Several solution studies have substantiated the hypothesis that the cofactor andsubstrate phosphates are indeed located in adjacent binding sites, close enough to possiblyinteract. Parrish et al., (1977) were able to demonstrate that pyrophosphate couldcompetitively inhibit the binding of both substrate (aG1P) and activator (phosphite) to PL-phosphorylase (remembering that phosphite binds at the position normally occupied by thecofactor phosphate). In addition, it was shown that one pyrophosphate molecule boundper enzyme monomer, suggesting that pyrophosphate could simultaneously bind at bothphosphate sites and that the two sites were directly adjacent. This approach was laterextended by Withers et al., (1982a) who demonstrated that phosphite activation of PL-phosphorylase b could be inhibited by a series (methylene, ethylene and propylene) ofalkane diphosphonate analogues, the most efficient inhibitor being the methylene analoguewhich possesses an inter-phosphorus distance of only 3A.Substrate-cofactor analogues possessing a pyrophosphate linkage have alsoprovided insight. Glucose- 1-pyrophosphate can be utilized as a substrate in PL-phosphorylase as it can transfer its glucosyl unit to glycogen with concomitant productionof pyrophosphate (Klein et al., 1984). This result suggests that the pyrophosphate moietybinds at both the substrate and cofactor phosphate binding sites, the latter of which must bebound for enzymic activation. In addition, it has been shown by Withers et al., (1982b)that glycogen phosphorylase b reconstituted with pyridoxal-5'-pyrophosphate (PLPP-phosphorylase), although unreactive, exists locked in the activated R-state conformationsince the enzyme possesses high affinity for R-state effectors such as AMP, and lowaffinity for T-state effectors like glucose and caffeine. That PLPP-phosphorylase is lockedin an activated conformation in the presence of AMP was later confirmed by X-raycrystallographic analysis (Sprang et al., 1991).The cofactor-substrate analogue pyridoxal-5'-pyrophospho-a-D-glucose (PLPP-Glu) is composed of the PLP cofactor and the aG1P substrate covalently linked through a20pyrophosphate linkage, and can be efficiently reconstituted into the apoenzyme (Takagi etal., 1982; Tagaya and Fukui, 1984; Withers et al., 1981a).Scheme 1-6. The reaction catalyzed by PLPP-Glu phosphorylase.In the presence of an oligosaccharide acceptor (e.g. glycogen) the glucosyl moiety ofPLPP-Glu-phosphorylase is rapidly transferred, forming a new a-1,4-glucosidic linkage,along with production of the catalytically inactive PLPP-enzyme (Scheme 1-6). Thus,phosphorylase reconstituted with PLPP-Glu is capable of catalyzing only a single glucosyltransfer, yet this is accomplished with the correct stereochemistry and is thought to veryclosely mimic the normal catalytic reaction (Tagaya and Fukui, 1984). There is little doubtthat phosphorylase positions its cofactor phosphate directly adjacent to the substratephosphate, however the nature of the interaction between the phosphates is a topic ofcontinuing controversy.213.3. 31 P NMR StudiesIt has been shown that glycogen phosphorylase is only active if the coenzymephosphate is capable of forming a dianion and that the cofactor phosphate is injuxtaposition with the substrate phosphate. It would seem, therefore, that 31P NMR wouldbe an ideal technique for further investigation of the role of the coenzyme phosphate since31P NMR chemical shifts of phosphates are known to be sensitive to both ionization stateand environment. Indeed, direct observation of the 31 P NMR resonance of the cofactor inthe enzyme in the presence of various substrates and effectors has provided furtherinformation, though rather inconclusive, concerning the mechanistic role of PLP. Early31P NMR experiments with phosphorylase were difficult to interpret as the resonances dueto PLP, AMP, and any other substrates and effectors such as aG1P all resonate at similarfrequencies (Feldmann and Helmreich, 1976). A short time later, however, Feldmann andHull, (1977) showed that thiophosphate analogues, which resonate approximately 40 ppmdownfield of the phosphate region, could be used to resolve the PLP resonance. Thus,phosphorylase could be activated by adenosine-5'-thiomonophosphate (AMPS) (Murrayand Atkinson, 1968) and the 31P NMR spectrum before and after AMPS addition waseasily interpreted. In the absence of AMPS the coenzyme phosphate resonates at a shift (80.5 ppm) equivalent to that of a mono-protonated phosphomonoester, termed 'form 1'.Upon addition of AMPS, the NMR spectrum shows a downfield (8 40 ppm) exchange-averaged signal representing free and bound AMPS. In addition, two signals representingPLP were observed approximately 3 ppm apart, the upfield peak (8 0.5 ppm) being form1 and a new downfield peak (8 --- 3.5 ppm), termed 'form 3', which was thought to be thecofactor phosphate in its deprotonated form. Upon addition of arsenate, an analogue ofphosphate which activates the enzyme and tightens AMPS binding, the 31P NMR spectrumwas seen to change once again. The AMPS signal was observed as two distinctresonances, representing free and bound AMPS. Further, the PLP resonance wascompletely converted to form 3, thought to represent the deprotonated form of the cofactor.22Thus, activation of the enzyme appeared to be accompanied by deprotonation of thecofactor. It was soon shown (Withers et al., 1979) that T-state effectors such as glucoseand caffeine could force the PLP phosphate resonance upfield from form 3 to form 1,possibly through a protonation event. Further, the resonances representing the free andbound AMPS species are observed to collapse into a single exchange-averaged signal underthese T-state conditions, consistent with the expected weakening of nucleotide binding(Withers et al., 1979).The results from previous 31 P NMR studies have been interpreted in terms of aPLP deprotonation which accompanies the T- to R-state allosteric transition, and indeed,have contributed to the proposal that PLP acts as an acid-base catalyst in the phosphorylasemechanism (Klein et al., 1981; Palm et al., 1990; Helmreich, 1992). It should be realized,however, that in any interpretation of 31 P NMR studies the understanding of factors whichdetermine chemical shifts of phosphate moieties is very limited. The problem becomesparticularly apparent when one considers the following evidence. As was discussedearlier, addition of glucose to nucleotide activated phosphorylase b results in an upfieldshift of the PLP resonance, interpreted as a protonation. If the same experiment isperformed with phosphorylase a reconstituted with the phosphonate analogue of PLP, 5'-deoxypyridoxa1-5'-methylenephosphonic acid, the result is a similar upfield shift (Hoerl etal., 1979). However, the 31 P NMR response to titration of this phosphonate analogue freein solution is known to be opposite to that of PLP (Schnackerz and Feldmann, 1980), theresonance shifting downfield upon protonation. The same reversal of effects is observedwith phosphorylase b reconstituted with the phosphonic acid analogue, since addition ofAMPS to the unliganded T-state enzyme results in a downfield shift in the 31P NMRsignal, the same response observed for PLP. It was suggested (Klein et al., 1984) thatthese reversed effects indicated that the ionization states are inverted in the reconstitutedenzyme. It seems unlikely, however, that the apparent inversion of ionization states for thephosphonic acid analogue of PLP is simply due to the small change in cofactor pKa pKa231.2 units). In addition, if PLP were involved in some catalytically essential protontransfer and the ionization state of the cofactor in the active R-state enzyme was important,it seems unlikely that the enzyme would tolerate such a reversal of ionization states with aPLP analogue that confers one of the highest apoenzyme reactivations to date. Theseobservations, therefore, cast doubt on the previous interpretation of the 31 P NMR chemicalshifts used to describe the allosteric transition in phosphorylase (Madsen and Withers,1986).One alternative explanation for the observed chemical shifts involves the cofactorphosphate undergoing a distortion of O-P-O bond angle due to the changing conformationof nearby positively charged amino acids known to interact with the cofactor and substratephosphate moieties. Gorenstein (1975) has demonstrated that the 31P NMR chemical shiftof phosphate esters correlates well with the smallest O-P-O bond angle. For example, thefirst ionization of an acyclic phosphate monoester produces no change in O-P-O bond angleand very little change in 31 P chemical shift (Gorenstein, 1975). Removal of the secondproton results in a change in chemical shift of 3 - 4 ppm downfield and a correspondingreduction in O-P-O bond angle from 104°) to 102° (Gorenstein, 1975). There is supportfor the proposal that changes in O-P-O bond angle, or phosphate conformation could beresponsible in some way for the shifts observed in the 31P NMR spectra of glycogenphosphorylase. Schinzel and coworkers (Schinzel et al, 1992) have recently used PLP as a31P NMR reporter nucleus for the functional changes which occur at the active site of E.coli maltodextrin phosphorylase with various active site mutations. Mutations made in theactive site region resulted in unusual dependences of PLP 31 P NMR chemical shifts onsolution pH, changes which could not solely be explained by simply invoking changes inionization state. Thus, the effects of the mutations on the PLP environment were discussedin terms of changes in the O-P-O bond angle. This explanation has also been used torationalize the 31P NMR chemical shifts in aspartate amino transferase, another enzymewhich carries an essential PLP cofactor and produces 31 P NMR data not easily interpreted24on the basis of ionizations alone (Schnackerz et al., 1989). Without a more preciseunderstanding of the factors affecting isotropic solution state 31P NMR chemical shifts, theinformation deduced from such experiments will remain rather speculative.31P NMR studies of phosphorylase have not solely focussed on the allosterictransition between T- and R-state enzyme conformations, but also on the catalytic role ofthe cofactor phosphate when a fully activated ternary enzyme complex is formed. To thisend, glucose-1,2-cyclic phosphate (Withers et al., 1981b; Klein et al., 1984) andheptulose-2-phosphate (Klein et al., 1984) were used as appropriate aG1P analogues sinceboth molecules are effective R-state inhibitors of the enzyme (Hu and Gold, 1978; Klein etal., 1984), and possess 31P NMR chemical shifts which are considerably displaced fromthat of PLP. Addition of glucose-1,2-cyclic phosphate to AMPS-activated phosphorylase bcauses an upfield shift of the PLP resonance from the normal R-state shift (form 3) to anew position slightly upfield of form 1. Extensive line-broadening of the PLP resonancewas also observed. Two possible interpretations of these results were given (Withers etal., 1981b), the first being that the PLP phosphate could become involved in a protonexchange, shifting to a more protonated state upon binding of the aG1P analogue. Thesecond interpretation requires that the PLP phosphate, be it a monoanion or distorteddianion, becomes tightly coordinated by enzymic groups, the large linewidth then resultingfrom a reduction in mobility. The suggestion that the PLP phosphate could be a distorteddianion in the activated form of phosphorylase has led to the proposal that PLP could beacting as a Lewis acid or electrophile in the catalytic mechanism rather than a BrOnsted acid(Withers et al., 1981b; Madsen and Withers, 1986).Similarly, when 31P NMR was used to monitor heptulose-2-phosphate binding topotato phosphorylase, the PLP resonance was observed to undergo a significant upfieldshift with considerable line-broadening (Klein et al., 1984). In addition, the 31P resonancefor heptulose-2-phosphate undergoes considerable broadening. It was suggested (Klein etal., 1984) that the change in linewidth could represent a protonation-deprotonation25equilibrium between the two phosphate moieties, an event which is unlikely to occur withbound glucose-1,2-cyclic phosphate due to its low pKa (Withers et al., 1981b; Klein et al.,1984).3.4. The Role of PLP in Catalysis?The most popular mechanism used to describe the role of PLP in thephosphorylase reaction involves the cofactor as a BrOnsted acid catalyst (Klein et al., 1982;Klein et al., 1984). Alternatively, it has been proposed that the cofactor phosphate remainsdianionic throughout the reaction, possibly acting in an electrophilic role (Withers et al.,1982a; Takagi et al., 1982) or as an essential dianionic phosphate anchor which functionsas an indispensable structural element (Chang et al., 1987).Largely on the basis of 31P NMR data coupled with evidence from the "interactingphosphates" hypothesis, and studies with glycosylic substrates, it has been suggested thatthe PLP phosphate group within glycogen phosphorylase functions as a proton donor-acceptor shuttle in a general acid-base catalyzed reaction (Helmreich, 1992; Palm et al.,1990; Klein et al., 1982; 1984). The postulate involves a direct proton transfer between thecoenzyme phosphate and that of the substrate as shown in Scheme 1-7. Thus, in thedirection of glycogen synthesis, the cofactor phosphate protonates aG1P, labilizing theglucosidic bond and effecting acid catalysis. In the direction of glycogen degradation thecofactor phosphate protonates the bound phosphate substrate which then protonates theglucosidic linkage at the terminal glucosyl residues of the oligosaccharide chain. Theglucosyl bond is broken and the incoming nucleophile (oligosaccharide or phosphate)attacks in a general base catalyzed process. The most recent version of this mechanisminvolves an oxocarbonium ion-intermediate stabilized by the substrate phosphate, and isreviewed by Helmreich, (1992).26HOHOHOHOHO^0If IIO—P —0"I,-' 0"OH, Ho^OH1I^HO^0—PL-0—P=0^HOO"HOGLYCOGENHHOHO0")I0=P —0"IOH k,GLYCOGEN^' --(31IPL-0—P —0"IIO/0 HOII^HOHO—P-0"OH0HO'Cy HO HO0II—PL-0—P —OHO -GLYCOGENScheme 1-7. The acid catalyzed mechanism proposed by Palm et al, 1990.Arguably the strongest evidence for the involvement of the cofactor phosphate as anacid-base catalyst comes from the work with glycosylic substrates, such as glucal, where ithas been shown that proton uptake occurs from below the plane of the sugar ring. Theutilization of glucal does indeed suggest that a proton source below the plane of thesubstrate, possibly phosphate, is available for activation of the glucal double bond.27However, this does not rule out the possibility that other protein functional groups couldfunction as the acid-base catalyst in the normal phosphorylase reaction.The proposed Br6nsted mechanism requires that a proton be relayed between thecoenzyme phosphate and the substrate phosphate. Experiments designed to probe thisproton transfer have centered on the use of PLP analogues. For example, reconstitution ofapophosphorylase b with pyridoxal-5'-fluorophosphate was found to result in an inactiveenzyme (Klein et al., 1982; Withers et al.,1982a). Since this PLP analogue does notpossess a second phosphate ionization capable of acting in a proton transfer, and alsocannot act as an essential dianion, this outcome is consistent with each of the proposalsdescibed earlier. Evidence against the involvement of PLY as a BrOnsted catalyst has comefrom studies with phosphorylase reconstituted with pyridoxal (PL-phosphorylase), anenzyme which is only active in the presence of activator anions which bind in place of themissing coenzyme phosphate moiety (Parrish et al., 1977; Withers et al., 1982c).Originally, Parrish et al., (1977) were able to demonstrate that bothfluorophosphate (pK2 = 4.8) and phosphite (pK2 = 6.6) were equally good activators ofPL-phosphorylase. The relatively low second pKa value for fluorophosphate essentiallyensures that the activator is in a dianionic state, likely excluding this phosphate analoguefrom participation in a proton shuttle mechanism. Indeed, Chang et al., (1983) haveprovided 19F NMR evidence suggesting that fluorophosphate binds to the PLP phosphatesite as a dianion and remains dianionic throughout catalysis. Further, a pH-activity profilewas perfomed for both the fluorophosphate- and phosphite-activated PL-enzymederivatives (Withers et al., 1982c). If these phosphate analogues were involved in anessential proton transfer then one might expect large differences in the observed pH-profiles, reflecting the different pKa values. However, no significant differences betweenthe enzyme systems were observed. These results are thought to suggest that thephosphate analogues which activate PL-phosphorylase, and therefore the native PLPcofactor, bind as dianions and are unlikely to be involved in an essential catalytic step28involving a proton transfer (Withers et al., 1982c). Still, doubts remain as to how well thePL-phosphorylase system, activated by non-covalently bound phosphate analogues,reflects the behaviour of the native PLP cofactor which possesses a covalent bond betweenpyridoxal and the 5'-phosphate moiety. Klein et al., (1984) have argued that slightdifferences in binding due to the lack of the PL-P covalent bond make it unlikely that anyactivating anion, regardless of its pKa , can replace the 5'-phosphate of the cofactor as aproton donor-acceptor group. Alternatively, and without justification, Klein andcoworkers suggest that these phosphate mimics force PL-phosphorylase to follow aslightly different mechanism using alternate amino acid side chains as general acid-basecatalysts (Klein et al., 1984).31P NMR results have not really provided satisfactory answers concerning the roleof the cofactor phosphate. However, Withers et al., (1985) have used 31P NMR T1 and T2relaxation measurements to provide evidence suggesting that the cofactor phosphate in theR-state enzyme is indeed a more tightly constrained, possibly dianionic, species than that inthe T-state enzyme. On this basis, in addition to the earlier evidence from fluorophosphateactivation studies with pyridoxal phosphorylase, and reconstitution studies with thecofactor-substrate analogue PLPP-Glu, the PLP cofactor in phosphorylase has beenpostulated to act in an electrophilic role (Madsen and Withers, 1986; 1984). It wasproposed that the coenzyme phosphate in the active ternary complex is constrained bypositive charges towards a trigonal bipyramiclal conformation with an empty apical positionoriented towards the substrate phosphate (Scheme 1-8). In this way, it was thought thatPLP could act as an electrophile interacting with the substrate phosphate and labilizing theglycosidic bond with the abortive formation of a quasi-pyrophosphate linkage. Thismechanism is somewhat analogous to the normal phosphoryl-transfer reactions catalyzedby kinase and phosphatase enzymes. Interestingly, uridine diphosphoglucose, the naturalsubstrate for glycogen synthase in the synthesis of glycogen is also very similar to theactive site complex proposed for the electrophilic mechanism, perhaps suggesting a29HOHOHOOHHO—P-0- HO1Og-8-0- it- 08-APL-0---3 P/i8+‘■HOGLYCOGENcommon structural requirement in the catalysis of glucosyl transfer to and from phosphates(Madsen and Withers, 1986).NucHO^00^HA11HO^s■ 11O—P —O-1O-A8- 0- — 0 8-PL—O-- P 8+I08_HONuOHHOH-TA\roHO'HOA-H^HOGLYCOGENScheme 1 -8. The proposed electrophilic mechanism for glycogen phosphorylase.In an attempt to probe the possible distortion of the coenzyme phosphate duringcatalysis, various oxyanions of the early transition metals (molybdate, tungstate andvanadate) have been tested with PL-phosphorylase as potential transition state analogues(Chang et al., 1983; Madsen and Withers, 1984). Since these analogues can adopt atrigonal bipyramidal configuration, it was thought that if a distorted phosphate is indeedOH30formed in the phosphorylase reaction then these metals may efficiently inhibit PL-phosphorylase activity by competing for the cofactor phosphate binding site. In one suchstudy, molybdate was found to bind pyridoxal phosphorylase some 13 times more tightlythan phosphate itself. Binding of this magnitude is considered good, but cannot beconsidered as conclusive evidence that molybdate binds as a transition state analogue insupport of the electrophilic mechanism. Unfortunately, it is difficult to test the electrophilicmechanism directly, and the exact role of the PLP phosphate moiety in phosphorylasecatalysis remains uncertain.4. The Aims of This Study While there is currently little disagreement as to the general features of the catalyticmechanism of glycogen phosphorylase, the exact role of the cofactor phosphate and theidentity of the acid-base catalytic group(s), in addition to the source of the stabilization ofthe intermediate glucosyl moiety remain unclear.Two approaches are proposed to answer the question of the role of the PLPcofactor in catalysis. One involves the synthesis of two phosphonate analogues of PLPwhich differ considerably in their phosphonic acid pK a values, thus in their ability to effectproton transfer reactions. Kinetic and NMR studies of glycogen phosphorylasereconstitued with such analogues should provide valuable insights. The second involvesuse of magic-angle spinning 31 P NMR to investigate the ionization state of the nativecofactor phosphate moiety in both the R- and T-state conformations through determinationof the shielding tensor components describing the phosphorus nucleus in each case.A second aim concerns the identity of active site amino acid residues, and threeapproaches are proposed. The first will involve the development and testing of novelcovalent inhibitors of glycogen phosphorylase. This will include kinetic andcrystallographic analysis of a novel sugar analogue, 1-nitro-D-glucal, which may act as aneffective enzyme inactivator, functioning through a conjugate addition reaction. The31second approach involves the synthesis of a stable derivative of PLPP-Glu, in which thehydroxyl group at the C-2 position of the glucose ring is replaced by a fluorine atom(PLPP-2FG1u). Once prepared and reconstituted into apophosphorylase b, kinetic, 19FNMR, and cofactor abstraction studies will be completed to characterize the PLPP-2FG1u-enzyme derivative. Such an analogue should not decompose once bound to the active siteof phosphorylase, and further, should not transfer its glucosyl moiety to acceptoroligosaccharide. PLPP-2FG1u should, therefore, allow crystallization and structuralanalysis of the activated enzyme / substrate complex. The third approach involves thetesting of a glycogen analogue in which all of the non-reducing terminal glucosyl residueshave been replaced with 4-deoxy-4-fluoro-glucosyl moieties (4-F-glycogen). 4-F-Glycogen will be investigated as an "incompetent" substrate analogue in an attempt to trapor provide evidence for a glucosyl-enzyme intermediate in phosphorylase catalysis.Finally, since 4-F-glycogen possesses an NMR active nucleus ( 19F), 19F NMRstudies will also be used to probe the solution binding interaction between glycogen andphosphorylase. Such studies should provide insight into the mode of binding betweenphosphorylase and glycogen, and also may allow an estimate of the stoichiometry ofbinding.For clarity, the aims relevant to each chapter are reviewed in a brief Introduction atthe beginning of Chapters 3, 4, and 5.32CHAPTER 2The Role of Pyridoxal Phosphate in Glycogen PhosphorylaseResults and Discussion33Results 1. Synthesis The synthesis of two PLP analogues and a derivative of pyridoxal was largelycarried out according to general procedures established by Korytnyk and Ikawa, (1970),and Hullar, (1969) as described below. For a review of the synthesis of Vitamin B6 (PLP)and its analogues, refer to Korytnyk, (1986).1.1. PLP Phosphonate AnaloguesThe synthesis of both 5'-deoxypyridoxal-5'-methylenephosphonic acid (5-CH2PLP) (9) and 5'-deoxypyridoxal-5'-difluoromethylenephosphonic acid (5-CF2PLP)(13) utilized the same basic strategy, as shown in Scheme 2-1. In each case, theisopropylidene derivative of 5'-deoxy-5'-chloropyridoxol (4) was first prepared frompyridoxol hydrochloride (Korytnyk and Ikawa, 1970) in high yield (80 - 90%). Thefollowing step required carbon-carbon bond formation between (4) and the protectedphosphonate moiety. While inefficient, the coupling could be achieved by reacting (4) withthe lithio anions of dimethyl methylphosphonate (Corey and Volante, 1976) and diethyl1,1-difluoromethylphosphonate (Obayashi et al., 1982; Bigge et al., 1989) in a reactionthat proceeded with very low yield (< 15%) in both cases. The highly unstable nature ofthe lithio difluoromethylphosphonate carbanion has been reported previously (Yang andBurton, 1991; Blackburn et al., 1987), yet all attempts to react [(diethoxyphosphoryl)difluoromethyl] zinc bromide (Burton et al., 1982; 1989), a stable alternate reagent, with(4) were unsuccessful. Once the fully protected phosphonate analogues were obtained,however, the subsequent deprotections and oxidation proceeded with the expectedmoderate to very good yields, as reported for the previous synthesis of 5-CH2PLP (Hullar,1969).340I IP-0" Li+IOH0I IP-0 - Li+IOHCHOHON(9)CHO(13)CH2OHCH2OH.,,..N../H+Cl -1)HC1, Acetone2) SOC12HOCH2C1^( 4 )1)Li+ -CH2P0(0Me)2 , THF2)HCOOH3) Mn02 / \4) HC11)Li+ -CF2P0(0E02, THF2)HCOOH3)Mn024) HC1Scheme 2-1. The synthetic route for preparation of 5-CH2PLP and 5-CF2PLP.35CH2OHCH2OHHO1.2. 5'-Deoxy-5'-difluoromethylpyridoxalThe synthesis of 5'-deoxy-5'-difluoromethylpyridoxal (PL-CF2H) (17) wascarried out predominantly according to well established procedures for pyridoxal chemistry(Korytnyk, 1986). Thus, the isopropylidene derivative of isopyridoxal (14) was preparedfrom pyridoxol hydrochloride with good overall yield (78%). The aldehyde functionalitywas fluorinated with diethylaminosulfur trifluoride (DAST) according to Middleton, (1975)to produce the protected 5'-difluoromethyl analogue of pyridoxol in good yield (72%).The subsequent deprotection and oxidation steps proceeded with good yield (64 and 79%,respectively) to afford PL-CF2H, as shown in Scheme 2-2.CHO1)Acetone, HCl2) Mn02^N(14)1)DAST (15)2) HC1^(16)3) Mn02CHOHO----_)/c— CF2HU(17)Scheme 2-2. The synthetic route for the preparation of PL-CF2H.36-1.5 1 -0.5^01/[G1 P] (1/mM)0.50.152. Kinetic Studies with Reconstituted Glycogen Phosphorylase b 2.1. Kinetic Constants for 5-CH2PLP- and 5-CF2PLP-Phosphorylase bGlycogen phosphorylase b was reconstituted with 5-CH2PLP and 5-CF2PLP byincubating apoenzyme with a 5 - 25 fold excess of each cofactor analogue as described inthe Materials and Methods section. The reconstituted enzymes were assayed for catalyticactivity in the direction of glycogen synthesis by measuring the initial reaction rates usingthe standard Fiske-Subbarow phosphate analysis, as described in Engers et al., (1970a, b).The kinetic parameters (V max and Km ) describing substrate (aG1P) utilization, for eachreconstituted enzyme, were determined by fitting the initial rate data to the non-linear formof the Michaelis-Menten equation. The initial rate data for each reconstituted enzyme,presented in Lineweaver-Burk form, are shown in Figure 2-1, whereas the calculatedvalues of V max and K m are presented in Table 2-1 along with that of the native enzyme.Figure 2-1. Determination of the kinetic parameters for 5-CH2PLP-phosphorylase b (0)and 5-CF2PLP-phosphorylase b (0).Reaction conditions are given in Table 2-1. The following substrate concentrations were used (m/4): 1.28,1.92, 2.56, 3.84, 5.12, 19.20.37Table 2-1. Kinetic parameters for glycogen phosphorylase b reconstituted with 5-CH2PLPand 5-CF2PLP.aEnzyme Km (mM) Vmax (pmolimiiilmg) % ReactivationNative 2.0 ± 0.1 62.4 ± 1.3 1005-CH2PLP-enzyme 0.70 ± 0.03 16.4 ± 0.1 265-CF2PLP-enzyme 2.1 ± 0.2 20.0 ± 0.5 32a Reaction conducted at pH 6.8 and 30° C in 100 mM KC1, 50 mM triethanolamine hydrochloride, 1 mMEDTA, 1 mM DTT. The reaction mixtures each contained 1 mM AMP, 1% glycogen, and reaction timeswere 5 minutes. Enzyme concentrations were 1.94, 8.02 and 6.24 ug mL -1 for the native, 5-CH2PLP-, and5-CF2PLP-phosphorylase enzymes, respectively. Apoenzyme activity was 0.5 1..unol min -1 mg-1 .Kinetic studies with reconstituted glycogen phosphorylase b revealed that Vmaxvalues for 5-CH2PLP-phosphorylase are approximately 26% of that observed in the nativeenzyme, and futhermore, 5-CH2PLP-phosphorylase was found to bind its substrate(aG1P) with an affinity slightly higher than that of the native enzyme. These results are ingood agreement with that reported previously for the same enzyme derivative (Vidgoff etal., 1974). Reconstitution of apophosphorylase b with 5-CF2PLP reactivated theapoenzyme to a level consistently higher than 5-CH2PLP, exhibiting approximately 32% ofthe activity measured with the native enzyme. Moreover, 5-CF2PLP-phosphorylase wasobserved to bind aG1P with an affinity about equal to that of the native enzyme.Interestingly, since apophosphorylase b is reactivated to similar levels with bothphosphonate analogues, these results suggest that differences in ionization behaviourbetween the two phosphonic acid cofactors, once bound at the active site of the enzyme,may not be important in catalysis. Indeed, since the 5-CF2PLP cofactor is expected toremain dianionic at this pH (6.8), these results bring into question the postulated role of the38cofactor phosphate (phosphonate in this case) as an essential proton shuttle in the catalyticmechanism.2.2. Phosphonic Acid Ionizations and pH/Rate ProfilesTo further investigate the potential involvement of the cofactor analogues in anessential proton transfer, a study of the effects of pH on the kinetics of 5-CH2PLP- and 5-CF2PLP-phosphorylase b was undertaken. The pK2 values for both 5-CH2PLP and 5-CF2PLP, free in solution, were first determined by 31 P NMR. This was followed by acomplete determination of the pH-dependence of V max , Km and Vmax/K m for eachreconstituted enzyme. By directly comparing the pH-dependences of the two reconstitutedenzymes, in parallel studies, many of the difficulties inherent in the interpretations ofpH/rate profiles, as previously discussed by Knowles, (1976) should be eliminated. Adiscussion of the interpretation of pH-profiles, necessarily based on a number ofassumptions about the enzyme-substrate system, is included in Appendix A.2.2.1. Phosphonate pK2 Values for 5-CH2PLP and 5-CF2PLPThe catalytically relevant second pKa value for each phosphonic acid analogue, freein solution, was determined in an effort to estimate the difference in pKa values between thetwo cofactor analogues once bound at the active site of the enzyme. Thus, the freephosphonate analogues of PLP were dissolved in a solution of 50% D20 and 100 mM KC1to final concentrations of 28 mM (5-CH2PLP) and 23 mM (5-CF2PLP), and were titratedover a wide range of pH by the sequential addition of small volumes of concentrated acid orbase. The titrations were followed by 31 P NMR spectroscopy as previously reported forthe pK2 determination of the phosphate moiety within PLP and several of its derivatives(Schnackerz, 1986; Schnackerz and Feldmann, 1980). The titration curves obtained in thismanner are shown below in Figure 2-2.396pH0^2^4 8^100^2^4^6^8pH b)a)Figure 2-2. 31P NMR titration of phosphonate cofactor analogues. (a) 5-CH2PLP (s); (b)S-CF2PLP (0).The titration was carried out in a 50% 1)20 solution in 100 mM KC1. The concentrations of PLPanalogues were 28 mM (5-CH2PLP) and 23 mM (5-CF2PLP). Adjustments in pH were made by theaddition of small volumes of 2 N NaOH or 2 N HCI.40Interestingly, the response of chemical shift to pH was in opposite directions in thetwo cases. For the titration of 5-CH2PLP, two ionizations were observed; pKi = 1.3 ±0.1 and pK2 = 7.2 ± 0.1, the second phosphonate ionization of which is in good agreementwith that determined previously (pKa = 7.3) for 5-CH2PLP (Schnakerz and Feldmann,1980; Vidgoff et al., 1974). Two ionizations were also observed for the titration of 5-CF2PLP; pKi = 4.2 ± 0.1 and pK2 = 8.2 ± 0.2. However in this case, since the first pKavalue for difluoromethylenephosphonic acids is known to be below the limit of this pHtitration, the observed pKa of 4.2 is assigned as the pK2 value for the phosphonic acidmoiety of 5-CF2PLP, in agreement with earlier studies of similar phosphonic acidspossessing the same difluoromethylene functionality (Blackburn et al., 1987; Bigge et al.,1989). Thus, as expected, the pK2 value for the fluorinated phosphonate is well below(ApK2 .----. 3 units) that of the non-fluorinated analogue. Interestingly, the second ionizationobserved for 5-CF2PLP (pKa = 8.2 ± 0.2), which produced only a small change in the 31PNMR chemical shift, closely matches the expected pK a for the pyridoxal phenolic moietywhen the ring nitrogen is deprotonated (Schnackerz, 1986; Kallen et al., 1985; Vidgoff eta1.,1974). Thus, it appears that the phosphonate moiety is sensitive to the ionization of thephenolic group on the pyridoxal ring, possibly through some preferred solutionconformation which places the two in juxtaposition, but only became visible in the 5-CF2PLP titration because of the significantly lower phosphonic acid pKa values.2.2.2. pH-Dependence of Phosphorylase b Reconstituted with 5-CH2P L Pand 5-CF2PLPWith a significant difference in pK2 values (ApK2 ...-- 3 units), between the freecofactor phosphonate analogues, a parallel study of the pH-dependence of glycogenphosphorylase reconstituted with 5-CH2PLP and 5-CF2PLP could provide valuableinformation concerning the role of the cofactor phosphate (or phosphonate in this case) inany essential catalytic steps. Thus, one might expect large differences in the pH-profiles41for the two enzyme systems, reflecting the expected differences in cofactor pK a values atthe active site of the enzyme, if indeed the cofactor analogues were involved in acatalytically essential proton transfer. To this end, the pH-dependence of the kineticparameters (Vmax , Km and Vmax/Km) for glycogen phosphorylase b reconstituted with 5-CH2PLP and 5-CF2PLP was determined by calculating the Vmax and Km values from aplot of initial reaction velocity versus aG1P concentration, at every pH studied. Reactionswere assayed in the direction of glycogen synthesis at 30 °C using the triethanolaminebuffer system previously described by Withers et al., (1982c). The range of pH values atwhich each enzyme is stable was investigated by incubating the reconstituted enzymes atvarying pH and subsequently measuring the activity at pH 6.8. The final reaction pH wasmeasured with all substrates, effectors and enzyme present. The pH-dependence of thekinetic parameters for each enzyme system, obtained in this way, is presented in Figure 2-3as the logarithm of the kinetic parameter versus pH. Apparent pKa values were calculatedwith the aid of the GraFit computer program (Leatherbarrow, 1990).The pH-profiles shown in Figure 2-3 are extremely similar for the two enzymesystems and certainly do not reflect the large change in cofactor pK a measured in solution.The pH-dependence of log Vmax/Km for both enzyme systems was found to be nearlysuperimposable. Indeed, the apparent pKa values for the acidic limbs (5-CH2PLP = 6.5;5-CF2PLP = 6.3) and the basic limbs (5-CH2PLP = 7.1; 5-CF2PLP = 6.9) are the samewithin experimental error. The pH-dependence of Vmax/Km represents the ionizations ofthe free enzyme and/or free substrate (Knowles, 1976), and thus it would seem that theionizations monitored here are quite possibly those of identical residues, for the most partunaffected by the differing cofactor pK a values. The plots describing the pH-dependenceof pKm for the two enzyme systems were also found to be very similar, the apparent pK avalues for the acidic limbs (5-CH2PLP = 6.2; 5-CF2PLP = 6.2) and basic limbs (5-CH2PLP = 7.0; 5-CF2PLP = 7.1) also being the same within experimental error.42b)E0.0.20-0.2-0.4-0.6-0.8- 1 _ I^i^11^I^i^1^1^1^i^I^1^_I^I^I^1^I^1^I^15.6 5.8^6^6.2 6.4 6.6 6.8^7^7.2 7.4 7.6xei3>a)o1. 0.6YE->a)0.40.20-0.25.6 5.8^6^6.2 6.4 6.6 6.8^7^7.2 7.4 7.65.6 5.8^6^6.2 6.4 6.6 6.8^7^7.2 7.4 7.6pHFigure 2-3. The pH-dependence of Vmax , Km and Vmax/Km for 5-CH2P LP -phosphorylase b (0) and 5-CF2PLP-phosphorylase b (0).At each pH value the initial reaction velocity was measured at six different aG1P concentrations rangingfrom 1.5 mM to 22 mM. All re-actions were carried out over 5 minutes at 30 °C and the indicated pH inbuffer containing 50 mM triethanolamine hydrochloride, 100 mM KC1, 1 mM EDTA, 1 mM DTT, in thepresence of 1 mM AMP and 1% glycogen. The enzyme concentration in each reaction mix was 9.15 and7.01 ug m1.7 1 for the 5-CH2PLP- and 5-CF2PLP-phosphorylase enzymes, respectively. Apparent P}Cavalues were calculated from the V max , Km , and Vmax /Km versus pH data, with the aid of the GraFitcomputer program. The error associated with the calculated pK values was low (5 ± 0.3 pK units) unlessotherwise noted. The curves shown in the logarithmic plots in Figure 2-3 are not meant to represent anyspecific pH function and are for presentation purposes only.43The pH-dependence of Km monitors ionizations within the free enzyme, free substrate andenzyme-substrate complex which are essential for binding. It would seem that theionizations monitored here, as with the log V max/Km plot, are unaffected by the differentcofactor pKa values. Furthermore, it was found that the pH-dependences of log Vmax forthe two enzyme systems show very similar apparent pK a values for the basic limb region ofthe plot, those being 7.9 and 7.6 for the 5-CH2PLP- and 5-CF2PLP-phosphorylaseenzymes, respectively. However, the calculated pK a of 7.9 for the 5-CH2PLP-enzyme isquite approximate since the pH-optimum (pH 7.0), and the limited pH-range in whichphosphorylase is active, restricts data collection in the basic limb region for this enzymederivative. Still, the pH-profile observed for 5-CH2PLP-phosphorylase is in goodagreement with that reported previously by Vidgoff et al., (1974). The acidic limb regionof the log Vmax plot is the only one in which significant differences are observed betweenpH-profiles for the two reconstituted enzyme species. Thus, the apparent pKa values forthe two enzyme systems differ by approximately 0.8 pH units, the pK a for 5-CH2PLP-phosphorylase being 6.4, the pK a for 5-CF2PLP-phosphorylase being 5.6. Thisdifference in pKa is also reflected in a difference in pH-optima (ApH-opt. .-  0.5 units)between the two enzyme systems, an optimal pH of 7.0 and 6.5 being observed for the 5-CH2PLP- and 5-CF2PLP-enzymes, respectively. Thus, it would appear that 5-CF2PLP-phosphorylase is most active at slightly lower pH, and furthermore, since the pH-dependence of Vmax monitors the ionizations of the enzyme-substrate complex immediatelypreceeding the rate-determining step (Knowles, 1976), it would seem that the enzyme-substrate ternary complex in 5-CF2PLP-phosphorylase exhibits a slightly lower pKa thanthe same complex in 5-CH2PLP-phosphorylase.44HOHO0IICH2----P-OH0-3. NMR Experiments of Phosphorylase b Reconstituted with 5-CF2pLP 3.1. 31 P NMR Experiments31P NMR spectra of glycogen phosphorylase b reconstituted with 5-CF2PLP wererecorded at 28 °C, in a 10 mm NMR tube at high enzyme concentration (0.9 - 1.0 mM).Adenosine-5'-thiomonophosphate (AMPS) was used as a nucleotide activator, in place ofAMP, since its resonance is well downfield of the region expected for the cofactorphosphonate. The 31P NMR spectra of 5-CF2PLP-phosphorylase b in the presence ofAMPS and (1-deoxy-a-D-glucopyranosyl) methylphosphonate (aG 1 CP), a substrateanalogue (Figure 2-4), are shown in Figure 2-5 in addition to a 31P NMR spectrum ofnative glycogen phosphorylase b recorded under similar AMPS activating conditions.Figure 2-4. The structure of aG1CP.4560^40^30^20^10^0ppmFigure 2-5. 31P NMR spectra of glycogen phosphorylase b reconstituted with 5-CF2PLPand in its native form.NMR samples contained 50-60% D20, 50 mM triethanolamine hydrochloride, 100 mM KC1, 1 mMEDTA, 1 mM DTT, pH 6.8, and spectra were recorded at ZS °C. a) Signal averaged over 33675 transients,reaction mixture contained 5-CF2PLP-phosphorylase b (0.92 mM), AMPS (1.8 mM). b) Signal averagedover 21600 transients, reaction mixture contained 5-CF2PLP-phosphorylase b (0.90 mM), AMPS (1.7mM), aG1CP (3.4 mM). c) Signal averaged over 4760 transients, reaction mixture contained nativeglycogen phosphorylase b (1.0 mM) , AMPS (2.4 mM).4 6The 31 P NMR spectrum of glycogen phosphorylase b reconstituted with 5-CF2PLP, in the presence of activating AMPS, is shown in Figure 2-5a. The largeresonance observed at 43.8 ppm is that of free AMPS and is present as an exchange-broadened signal (A1)1/2 = 99 Hz) with enzyme-bound AMPS, appearing as a small broadshoulder peak located at 8 ---. 41.1 ppm. Extensive line-broadening is also seen in the signalrepresenting the cofactor phosphonate, which is centered at 8 --- 2.2 ppm with a linewidthof 380 Hz. A portion of the cofactor broadening undoubtedly arises as a result of theAMPS exchange process, however, significant line-broadening is also expected as a resultof the scalar coupling from the difluoromethylene functionality directly adjacent to thephosphorus nucleus. Free in solution, the 31P NMR spectra of 5-CF2PLP shows a tripletat 5.5 ppm with a Jp_F coupling on the order of 100 Hz. Immobilized at the active site ofthe protein, in a chiral environment, the two fluorine nuclei will be inequivalent, eachcoupling to the phosphorus nucleus separately, the result being considerable signal-broadening. The sharp resonance observed at 1.9 ppm is assigned as contaminatingphosphate, derived from contaminated AMPS or possibly released from the protein itself(Withers et al., 1979).On addition of aG1CP, an inactive substrate analogue, to the nucleotide-activatedenzyme (Figure 2-5b), the free AMPS resonance is observed to shift slightly downfield to44.0 ppm with a significant reduction in linewidth (Apia = 84 Hz). Correspondingly, thebound AMPS signal has shifted upfield slightly, to 40.9 ppm, providing slightly betterresolution of the free and bound forms of the nucleotide, and together these changessuggest that a decrease in the AMPS exchange rate has accompanied aG1CP binding.Further, the cofactor resonance for 5-CF2PLP-phosphorylase b has shifted slightlydownfield to 2.4 ppm and also undergone a reduction in linewidth (O14/2 = 365 Hz). Thisresult is consistent with that seen previously for the native enzyme (Feldmann and Hull,1977; Hoerl et al., 1979; Withers et al., 1981b) and for 5-CH2PLP-phosphorylase b (Kleinet al., 1984). The resonance for contaminating phosphate has shifted slightly upfield to 1.847ppm, with an increase in signal intensity and a considerable reduction in linewidth, asestimated by re-Fourier transforming the data using a 10 Hz line-broadening factor. Theeffects on contaminating phosphate are consistent with its initial involvement in anexchange process at the active site of the enzyme, binding to the vacant phosphate sitenormally occupied by aG1P, but then being displaced from the active site upon aG1CPaddition. The resonance due to aG1CP is observed as a single exchange-averaged signal,in which both the chemical shift (21.4 ppm) and the linewidth (Avii2 = 40 Hz) represent aweighted average of the free and enzyme-bound species. Thus, in the presence of aG1CP,5-CF2PLP-phosphorylase b appears to shift to a more activated state, reflected in theresonances for both AMPS and the phosphonate cofactor.The 31 P NMR spectrum of native glycogen phosphorylase, collected under similarconditions of AMPS activation, is shown in Figure 2-5c. The signal from free AMPS isobserved at 43.9 ppm with a linewidth of 63 Hz, whereas the enzyme-bound AMPS signalis clearly resolved at 41.1 ppm (.6:01/2 = 143 Hz), illustrating a considerable difference inthe AMPS exchange rate between the native enzyme and 5-CF2PLP-phosphorylase b. Theresonance from the PLP cofactor is also well defined at 8 4 ppm, with a small resonanceat 0.8 ppm, representing the two forms of the enzyme, previously defined in theIntroduction. The sharp resonance at 3.6 ppm is commonly observed in spectra of thenative enzyme (Withers et al., 1979), and assigned as AMP. Thus, it seems that nativeglycogen phosphorylase b exists in a more activated R-state than the 5-CF2PLP-enzyme, asjudged from the variation in AMPS exchange rates and thus the apparent AMPS affinity forthe two enzyme species.48L-20rf-30 ppm -40•-50a)b)3.2. 19F NMR Experiments19F NMR spectra of glycogen phosphorylase b reconstituted with 5-CF2PLP wererecorded at room temperature in a 5 mm NMR tube at high enzyme concentration. The 19FNMR spectra of 5-CF2PLP-phosphorylase b, obtained in this way, in the absence andpresence of activating AMP are shown in Figure 2-6.Figure 2-6. 19F NMR spectra of glycogen phosphorylase b reconstituted with 5-CF2PLP.The NMR sample contained 50% D20, 50 mM triethanolamine hydrochloride, 100 mM KC1, 1 mMEDTA, 1mM DTT, pH 6.8. a) Signal averaged over 55824 scans, sample contained 5-CF2PLP-phosphorylase (0.84 mM). b) Signal averaged over 54399 scans, sample contained 5-CF2PLP-phosphorylase (0.80 mM), AMP (1.8 mM). Chemical shifts are referenced to TFA.In the absence of activating nucleotide (Figure 2-6a) the 19F NMR spectrum for 5-CF2PLP-phosphorylase b shows an extremely broad resonance (AD1/2 710 Hz) centeredat approximately -32 ppm. That the observed chemical shift for the enzyme-bound cofactoris shifted slightly downfield (AS ---- 3 ppm) from that observed for the cofactor free in49solution is consistent with shifts observed in other systems (Gerig, 1989; Percival andWithers, 1992), and thought to reflect a change to a relatively hydrophobic environment(Sykes and Hull, 1978). The observed line-broadening could result from chemical shiftheterogeneity, or alternatively, from two broad overlapping 19F signals possessing slightlydifferent chemical shifts. Indeed, immobilized at the active site of the enzyme, eachfluorine nucleus within 5-CF2PLP would be distinct, geminally coupled to both fluorineand phosphorus. Similar large linewidths have been observed in previous 19F NMRstudies of phosphorylase reconstituted with 6-fluoropyridoxal phosphate (Chang et al.,1986), a system with a single fluorine bonded to the cofactor, and lacking any large J-coupling. Upon addition of AMP to the sample of 5-CF2PLP-phosphorylase b (Figure 2-6b), the 19F signal appeared as two broad components together shifted slightly upfield to 5-34 ppm, with a considerable reduction in total linewidth (M)112 610 Hz). Thereduction in linewidth upon nucleotide activation might be due to an increase in thehomogeneity of the population of enzyme species, or alternatively, a change in mobility forthe two distinct fluorine nuclei.Extensive 19F signal-broadening was also observed for the difluoromethylenefunctionality in preliminary studies with glycogen phosphorylase b reconstituted with 5'-deoxy-5'-difluoromethylpyridoxal (PL-CF2H) (Scheme 2-2). PL-CF2H is an analogue ofpyridoxal that was originally designed to probe possible fluorine-phosphorus interactions atthe active site of glycogen phosphorylase, previously reported with phosphoglucomutase(Percival and Withers, 1992). While PL-CF2H was found to reconstitute and reactivateglycogen phosphorylase b to a similar extent to that reported previously for other 5'-deoxyanalogues of pyridoxal (Chang et al., 1987), the 19F NMR signal observed for PL-CF2Hreconstituted phosphorylase b was essentially broadened completely into the baseline.While it is possible that the extreme broadening in this case may be due, in part, to proteinprecipitation resulting from the instability of PL-CF2H reconstituted phosphorylase, thuslimiting data acquisition, it seems that other factors associated with the difluoromethylene50moiety contribute to the extensive signal-broadening observed. One potential source ofadditional signal-broadening for the PL-CF2H analogue is exchange between magneticenvironments in the active site. This is reasonable since PL-CF2H lacks the 5'-phosphategroup which would be expected to aid PLP positioning.4. 31 P MASNMR Spectroscopy of R- and T -State Glycogen Phosphorylase The previous studies of glycogen phosphorylase b reconstituted with 5-CH2PLPand 5-CF2PLP have focussed on the role of PLP in the native enzyme. In particular, thestudies have tried to answer the specific question of whether or not the cofactor phosphate(or phosphonate in this case) is involved in an essential proton transfer step duringcatalysis. The answer to this question may or may not be directly related to the ionizationstate of PLP in the T- and R-state allosteric forms of glycogen phosphorylase. However,in an attempt to address the question of possible changes in cofactor ionization state thatoccur upon enzymic activation, many solution 31 P NMR studies (Feldmann and Hull,1977; Withers et al., 1981b, 1985; Hoerl et al., 1979; Klein et al., 1984) have beenperformed and the results have suggested that changes in the isotropic chemical shift of thecofactor phosphate upon activation (AS 3 ppm) may reflect a cofactor deprotonationevent. However, it is also possible that geometrical changes occurring as a result of theallosteric transition influence 31P chemical shift.In an effort to more clearly define the ionization state of the cofactor phosphate inthe R- and T-state allosteric forms of glycogen phosphorylase, high resolution 31P magic-angle spinning NMR (MASNMR) spectroscopy was employed in a collaboration with Dr.Robin Challoner and Professor C. A. McDowell. 31P MASNMR was used to obtain theshielding tensor components for the phosphate moiety of free (solid) PLP in various stagesof ionization as well as the shielding tensor components for the PLP phosphate moietywithin the R- and T-state forms of crystalline native glycogen phosphorylase b.Knowledge of the shielding tensor components (a11, 622, and a33) describing the51shielding tensor interaction allow a comparison of the symmetry of the electron densitydistribution at the cofactor phosphate nucleus of the model compounds with that in both theR- and T-state forms of the enzyme (Challoner et al., 1992). The MASNMR techniqueyields a centerband at the isotropic frequency in addition to a spinning sideband manifold,which after analysis provides information concerning the anisotropy of the chemical shift.Since dianionic phosphorus nuclei of phosphate monoesters possess three equivalentoxygen nuclei and one P-0(R) bond, the principal shielding components will exhibit axiallysymmetric shielding in which the least shielded component (033) of the shielding tensor isparallel to the unique P-0(R) bond and 022 and 011 will be equivalent in magnitude and liein a perpendicular plane with no preferred orientation. In the case of the monoanionicforms of phosphate monoesters, the phosphorus nucleus is not axially symmetric and willpossess shielding tensor components of lower symmetry than in the dianionic case(Challoner et al., 1992). Thus, if the cofactor phosphate moiety in glycogenphosphorylase undergoes a change in protonation state upon activation, this will bereflected in the principal shielding components, whereas if the allosteric transition does notinvolve a protonation event, then the symmetry of the principal shielding components mayremain relatively unperturbed (Challoner et al., 1992). Indeed, the recent study of Un andKlein, (1989) in which O-P-O bond angles and P-0 bond lengths were related to principalshielding tensor elements suggests that very small changes in such parameters couldaccount for a 3 ppm shift in isotropic resonance.The 31P MASNMR spectra for the model compounds of PLP, and for the R- andT-state forms of glycogen phosphorylase b were collected and analyzed by Dr. R.Challoner, and are shown in Figure 2-7. The shielding tensor components for variousionizations of PLP and for the cofactor phosphate in the R- and T-state conformations ofglycogen phosphorylase are shown in Table 2-2.52Figure 2 - 7. Figure legend on following page.53100 0PPMIII-100d),^I^,50,-50100 50 0PPP.1^1^,^,^'-50 -100C)IFFigure 2-7. 31P MASNMR spectra of free PLP and the R- and T-state forms of glycogenphosphorylase b.Spectra arising from the application of single-pulse 31 P MASNMR experiments to the model PLPcompounds and to glycogen phosphorylase b. (a) PLP-disodium salt; (b) PLP-monosodium salt. Themicrocrystalline samples of glycogen phosphorylase were prepared as outlined in the Materials and Methodssection. (c) R-state glycogen phosphorylase b (contains AMPS), 29,128 transients; (d) T-state glycogenphosphorylase b, 11,548 transients. The isotropic resonances are indicated (*) on the spectra of enzyme.54Table 2-2. Principal components of the 31P shielding tensors for free PLP and for thecofactor phosphate in R- and T-state glycogen phosphorylase b.aCompound 633 022 an aaniso 0 6solnPLP-disodium salt 79.5 -14.6 -43.6 108.6 7.1PLP-monosodium salt 68.1 9.5 -73.0 99.9 1.5PLP-free acid 67.2 5.5 -78.7 103.8 -2.0R-state phosphorylase b 70.8 -30.0 -30.0 100.8 3.6 3.8T-state phosphorylase b * 66.4 -33.9 -33.9 100.3 --0.5, 2.5 0.6a Table was taken from Challoner et al., (1992). All chemical shift values are quoted in ppm. The shiftswere referenced to 85% H3PO4 with signals occurring downfield being positive. The anisotropy is definedas 033 - 1/2(022 + ail), and the shift (a) is equal to 1/3 (a33 + 022 + all).* The shielding tensor components correspond to the peak at -0.5 ppm. Note, however, that a somewhatgreater error is introduced into the measured components for this species with respect to the others listed inthe table, given the assumptions made when deconvoluting the centerband.The shift in isotropic resonance for the model compounds, as a function ofprotonation state, is in agreement with that observed in solution. Further, although thedianionic form of PLP deviates somewhat from axial symmetry (Challoner et al., 1992),the principal shielding tensor components describing the phosphorus environment in eachof the ionization states for the model compounds have shifted in the expected directionsrelative to one another, and are in good agreement with values reported previously (R.Challoner, personal communication; Un and Klein, 1989). The free acid is included inTable 2-2 but is not considered to be a likely form of the PLP cofactor located at the activesite of glycogen phosphorylase.The isotropic shifts observed for the PLP phosphate moiety in the R- and T-statecrystalline forms of glycogen phosphorylase b were also in good agreement with thoseobserved previously in solution studies (Feldmann and Hull, 1977; Withers et al., 1981b;55Hoerl et al., 1979; Klein et al., 1984), wherein a downfield shift was observed uponAMPS activation. Further, for the R-state enzyme, considerable shielding anisotropy, andtherefore good sideband intensities, was observed for the PLP phosphate moiety and forthe thiophosphate moiety of AMPS. Direct simulation of the spectra for the R-state enzymereveals axially symmetric shielding parameters, typical of those previously determined fordianionic phosphate monoesters (Challoner et al., 1992). Motional influences on the PLPphosphate moiety, which in a different ionization state could lead to the observed axialsymmetry, are considered unlikely (Challoner et al., 1992). For the T-state form of theenzyme, two resonances were observed in the 31P MASNMR spectrum, suggestingheterogeneity in the enzyme sample. The major signal (8 --= -0.5 ppm) has a spinningsideband manifold similar to that observed for the R-state allosteric form of phosphorylase,but with the isotropic peak shifted slightly upfield. Direct simulation of such a sidebandmanifold leads to axially symmetric parameters with a nearly identical shielding to thatobserved in the R-state enzyme, therefore suggesting that the cofactor phosphate presentwithin the inactive T-state enzyme is also dianionic. In this case, however, greater error isintroduced by the overlap of the spectral intensity of several components in addition to theuncertainty in the isotropic shift as a result of the uncertainty in the phasing of theoverlapping centerbands (Challoner et al., 1992). The symmetry of the shielding tensor forthe minor (downfield) signal in the T-state spectrum could not be evaluated due to its lowintensity, and it is thought that this component of the T-state spectrum could correspond toa more mobile phosphate moiety. It is clear, however, that the shielding tensor symmetryfor the major component in the T-state spectrum is very similar to the correspondingphosphate moiety in the R-state spectrum.565. Discussion PLP and CatalysisEarly studies with glycogen phosphorylase identified separate structural andcatalytic roles for PLP, since cofactor analogues such as pyridoxal, pyridoxal-5'-phosphatemonomethylester, and pyridoxal-5'-fluorophosphate were all found to reconstitute aholoprotein with intact quaternary structure and allosteric properties, but with no catalyticactivity (Kastenschmidt et al., 1968; Pfeuffer et al., 1972; Feldmann et al., 1976; Klein etal., 1982; Sprang et al., 1982). Not until the cofactor contains a phosphate derivativecapable of forming a dianion will enzyme activity be restored, as illustrated with nativePLP, phosphonic acid derivatives of PLP, or even pyridoxal-phosphorylase in the presenceof phosphite or fluorophosphate. This in hand, two catalytic roles for PLP have beenproposed, both consistent with the need for a cofactor phosphate capable of forming adianion. The first involves the PLP phosphate moiety as an essential acid catalyst (Klein etal., 1982), the second as a distorted dianion capable of acting as an electrophilic catalyst(Withers et al., 1982a; Takagi et al., 1982).In this study, two phosphonic acid analogues of PLP were reconstituted intoglycogen phosphorylase b in an effort to address the specific question of whether or notPLP acts as an essential acid catalyst. For a review of the proposed mechanism involvingPLP as an essential acid catalyst please refer to the Introduction, section 3.4. Thedeoxymethylene (5-CH2PLP) and deoxydifluoromethylene (5-CF2PLP) analogues of PLPwere synthesized since a large difference in phosphonic acid pK a (ipK2 ,--- 3) could beobtained with only minimal steric considerations. Measurement of 31 P and 19F NMRspectra, in addition to the pH-dependence of activity for glycogen phosphorylasereconstituted with each analogue should provide insights into the cofactor ionization state.Undoubtedly, once the two cofactors are bound to the enzyme, the active site environmentcould well perturb the individual in-situ pKa values, however, what is important for thisstudy is the intrinsic difference in cofactor pK a , and the fact that the reconstituted enzyme57systems, otherwise, are internally consistent. Sterically, the two cofactors are very similar,the C-F bond length (1.39 A) and van der Waals radius (1.35 A) being only slightly greaterthan that of the C-H bond (1.09 A and 1.20 A, respectively) (Withers et al., 1988). Thus,large differences in enzymic activation between the two cofactors is unlikely to arise as aresult of these small steric considerations. However, unlike hydrogen, fluorine can formweak hydrogen bonds, possibly mimicking the phosphate ester oxygen in the native PLPcofactor. Thus, it is possible that favourable interactions between the fluorinated cofactorand residues at the active site of the enzyme may, in part, be responsible for smalldifferences in catalytic efficiency between the reconstituted enzyme derivatives.Reconstitution of glycogen phosphorylase b with 5-CH2PLP (Table 2-1) wasfound to reactivate the apoenzyme to approximately 26% of the activity found with thenative enzyme. Further, this reconstituted enzyme was found to bind its substrate (aG1P)with slightly higher affinity than the native enzyme in complete agreement with previousstudies (Vidgoff et al., 1974). Interestingly, however, 5-CF2PLP-phosphorylase b alsosignificantly reactivated the apoenzyme (32%), indeed, to a level slightly greater than thatof 5-CH2PLP-phosphorylase. The higher activity with 5-CF2PLP might be due tofortuitous interactions between the cofactors polar difluoromethylene group and active siteresidues, resulting in slightly better placement of the cofactor phosphonate for catalysis.Indeed, 19F NMR experiments of 5-CF2PLP-phosphorylase suggest that significantchanges in the environment surrounding the difluoromethylene functionality occur uponnucleotide activation of the enzyme. Substrate (aG1P) was found to bind to 5-CF2PLP-phosphorylase b with an affinity (2.0 mM) only slightly less than that for the 5-CH2PLP-enzyme (0.70 mM). Indeed, the Km values calculated for the two enzymes are within therange commonly observed for the native enzyme from one enzyme preparation to the next.Given the similiarity in apoenzyme reactivation for the two cofactor analogues, and thetendency for 5-CF2PLP to remain dianionic at pH 6.8, it seems likely that both analoguesbind to the enzyme as dianions.58In an effort to fully investigate the possibility that the cofactor analogues could beacting as essential acid catalysts, a complete determination of the pH-dependence of thekinetic parameters (Vmax, Km and Vmax/Km) for glycogen phosphorylase b reconstitutedwith 5-CH2PLP and 5-CF2PLP was completed. Although enzymic pH-profiles invariablysuffer from an uncertainty of approximately 0.3 pH units (Kasvinsky and Meyer, 1977)and are often over-interpreted (Knowles, 1976), it seems likely that given the extremedifferences in solution pK2 (ApK2 --. 3) for the two cofactors, if the phosphonic acids wereindeed involved in a catalytically essential proton transfer, then the large differences incofactor pK2 would necessarily be reflected in the pH-profiles for the two enzymic species.In this study, it was found that the ionizations of the free enzyme and free substratederived from the pH-dependence of log Vmax/Km and pKm , are extremely similar for bothreconstituted enzyme species. In fact, the apparent pKa values for each reconstitutedenzyme, in both plots, were approximately 6.3 and 7.0 for the acidic and basic limbs,respectively. These ionizations are very similar to the pKa values observed in previous pH-studies with the native enzyme (Withers et al., 1982c; Kasvinsky and Meyer, 1977). It isunlikely, therefore, that changes in ionization state of the phosphonate cofactor are beingreflected in these plots. Similarly, the alkaline limbs in the plot describing the pH-dependence of log Vmax for 5-CH2PLP- and 5-CF2PLP-phosphorylase b are quite similarwith pKa values of 7.9 and 7.6, respectively. If the cofactor phosphonate analogues wereinvolved as essential acid catalysts then this alkaline limb should be quite different for thetwo reconstituted enzymes, showing a much lower pKa value for the 5-CF2PLP-enzymederivative. However, no such differences were observed, thus it seems highly unlikelythat the phosphonic acid analogues of PLP could be functioning as essential acid catalystsin glycogen phosphorylase. Rather, the results suggest that the cofactor phosphonates arebound to the enzyme as dianions, and possibly remain as such throughout catalysis.The only significant difference observed in any of the pH-profiles for the tworeconstituted enzyme derivatives was in the acidic limb of the log Vmax plot, where59apparent pKa values of 6.4 and 5.6 were observed for the 5-CH2PLP- and 5-CF2PLP-enzymes, respectively. The acidic limb of the log Vmax plot reflects an ionization of agroup within the enzyme-substrate complex which must remain deprotonated for effectivecatalysis. Since this difference in pKa values was not observed in the free enzyme (logVm/Km plot), it must arise from ionizations which are perturbed upon substrate bindingand formation of the ternary enzyme complex. This difference would seem to be too smallto arise from the cofactors themselves, and therefore may have its origin in the way the twocofactors interact with residues located at the active site, resulting in a small change in thepKa of the ionizing group normally responsible for the acidic limb of the log Vmax profile.That cofactor analogues with significantly different pK2 values can reactivatephosphorylase b and show similar pH-dependent profiles has been reported previouslywith studies of pyridoxal (PL) reconstituted phosphorylase (Parrish et al., 1977; Chang etal., 1983; Withers et al., 1982c). In these studies it was shown that PL-phosphorylasecould be reactivated to very similar levels with phosphite (pK2 = 6.6) and fluorophosphate(pK2 = 4.8) activator anions which bind to the vacant cofactor phosphate site and act in itsabsence. It has been argued (Klein et al., 1984), but with little justification, that studieswith PL-phosphorylase and activating anions do not reflect the reaction with the naturalcofactor, and further that PL-phosphorylase follows a different mechanism employing adifferent acid catalyst. The kinetic studies shown here, with 5-CH2PLP and 5-CF2PLPreconstituted phosphorylase b, do not suffer from the same difficulties, and therefore,serve to fully support earlier suggestions (Withers et al., 1982c) that PLP is not theessential acid catalyst in glycogen phosphorylase.PLP and the Allosteric Transition While previous studies have focussed on the catalytic role of PLP as a potential acidcatalyst, much effort has also centered on the allosteric transition from the inactive T-stateto the active R-state. Of particular interest is the mechanism by which the enzyme achieves60activation and is regulated in vivo. Activation of glycogen phosphorylase b, throughnucleotide (AMP) binding or covalent activation by phosphorylation, is known to beaccompanied by many structural changes throughout the enzyme molecule, several ofwhich are localized at the active site (Johnson, 1992). Solution state 31 P NMRspectroscopy has often been used to observe these changes in the PLP environment. Early31P NMR studies demonstrated that the resonance for the PLP cofactor phosphate shiftedapproximately 3 ppm downfield upon activation with AMPS (Feldmann and Hull, 1977),analogous to the shift which accompanies the deprotonation of a free phosphate monoester.Thus, activation of the enzyme has often been thought to involve deprotonation of thecofactor phosphate (Klein et al., 1981; Helmreich 1992) (Figure 2-8).Results from high resolution solid state 31P MASNMR experiments with the R- andT-state forms of the enzyme, however, cast doubts on the interpretation of the isotropicchemical shifts observed in solution, and suggest that the cofactor phosphate is a dianionicspecies in both forms of the enzyme. The isotropic chemical shifts observed for the R- andT-state crystalline samples of phosphorylase b (Table 2-2) are in good agreement withthose observed in solution. However, the spinning side-band manifolds for the two formsof the enzyme suggest that axially symmetric shielding parameters describe the phosphorusenvironment in each case (Challoner et al., 1992), therefore characterizing it as beingdianionic in both cases.61T-State R-State• •ATP, Glucose-6-phosphateGlucose, Caffeine0Py ridoxal 0—P-0"OH 0Pyridoxal 0—P-0"44v^I20^10^0^–10^420^10^0^—10PPm PPInFigure 2-8. An earlier assignment of the changes in cofactor ionization state thought toaccompany the allosteric transition in glycogen phosphorylase.Feldmann and Hull, (1977); Helmreich, (1992)Thus, it would seem that differences arising between the isotropic shifts of the PLPphosphate moiety in the R - and T-state forms of phosphorylase are not a consequence ofdifferences in protonation state (Challoner et al., 1992). Rather, given the structuralchanges known to occur throughout the enzyme upon activation, the observed change inchemical shift (= 3 ppm) for the PLP phosphate moiety probably occurs as a result ofchanges in phosphate geometry. Differences in O-P-O bond angle have been shownpreviously to correlate with the varying 31 P chemical shifts of phosphate monoesters62(Gorenstein, 1975). Furthermore, changing phosphate geometry has been suggested as thesource of the different shifts in the phosphorus resonance for PLP bound at the active siteof a series of mutant E.coli maltodextrin phosphorylases (Schinzel et al., 1992). Thesuggestion that the PLP phosphate moiety is present as a dianion in both the T- and R-stateconformations contrasts with earlier proposals concerning its ionization state. However, itis quite consistent with X-ray crystallographic studies which show that the amino acidcontacts between phosphorylase and the cofactor phosphate are essentially unchanged inthe R- and T-state forms of the enzyme (Oikonomakos, 1991). Furthermore, kineticstudies with 5-CH2PLP- and 5-CF2PLP-phosphorylase b also support the proposal that thecofactor analogues are bound as dianionic species, and in fact suggest that the cofactorphosphate may remain dianionic throughout catalysis. Thus, the PLP phosphate, oncebound to the active site of the enzyme, may indeed, exist exclusively as a fully ionizedspecies, with no changes in ionization state.While evidence gathered from 31 P MASNMR points to the PLP phosphate as adianion in both the R- and T-state conformations of native glycogen phosphorylase b, therole of the cofactor in the allosteric transition is still unclear. Allostery in phosphorylasehas been described in terms of conformational changes which serve to eitherelectrostatically create or destroy phosphate recognition sites within the enzyme (aG1P,AMP, Serine-P), and in doing so affect subunit-subunit contacts (Oikonomakos, et al.,1991). It has been suggested, on the basis of kinetic and 19F NMR studies, that thecofactor phosphate may have a structural role in the positioning of catalytic groups in thecorrect orientation for catalysis (Chang et al., 1987), thus it is tempting to also assign asimilar allosteric role. Indeed, 31 P NMR relaxation studies have provided evidence that thecofactor phosphate becomes more rigidly fixed upon enzymic activation (Withers et al.,1985), perhaps suggesting that the strengths of the interactions between the enzyme and thecofactor phosphate change with activation. It is therefore somewhat surprising thatphosphorylase reconstituted with pyridoxal, possessing no phosphate moiety whatsoever,63is known to possess allosteric properties, and indeed, this enzyme derivative exists in asignificantly more R-state conformation than the native enzyme (Kastenschmidt et al.,1968; Withers et al., 1982b). Thus, the presence of a dianionic cofactor phosphate reducesthe extent of enzymic activation. Consistent with this observation, is the suggestion, basedon 31P NMR studies of 5-CH2PLP-phosphorylase b (Klein et al., 1984) and 5-CF2PLP-phosphorylase b, that when phosphorylase is reconstituted with phosphonic acid analoguesthe enzyme exists in an activated state similar to, or indeed less activated than, the nativeenzyme. While it is still quite unclear as to why the PLP phosphate moiety would reduceenzymic activation, it must be remembered that phosphorylase is a highly regulated enzymewhich uses a variety of allosteric mechanisms to meet the changing needs of the hostorganism. Thus, it seems possible that the role of the cofactor phosphate in enzymicactivation may be somewhat regulatory, the dianionic phosphate moiety acting to down-regulate or dampen activation such that the enzyme can always be maximally responsive tochanges in the concentration of the various allosteric affectors.A Mechanistic Alternative Enzymes are capable catalysts because they are able to reduce activation energy,either by stabilization of the transition state or by destabilization of the ground state, orboth. Glycogen phosphorylase is known to contain an essential molecule of PLP whosephosphate moiety must be capable of forming a dianion if the enzyme is to be active. Inthis study, all evidence points to a cofactor phosphate which binds to the enzyme as adianion in both the R- and T-state conformations, with no role as an essential acid catalyst.Thus, alternatives for the role of the cofactor phosphate moiety in catalysis must beconsidered. While the results from this study contradict earlier proposals suggesting thatthe cofactor phosphate moiety behaves as an essential BrOnsted acid catalyst (Helmreich,1992; Klein et al., 1982; 1984), they are in agreement with proposals suggesting that thecofactor phosphate possibly acts in an electrophilic role (Withers et al., 1982a; Takagi et64al., 1982), or as a phosphate anchor that functions to orient active site catalytic residues(Chang et al., 1987). Indeed, the dianionic cofactor phosphate may function to coordinatethe orientation of different domains within the protein, thereby mediating the structuralchanges which are essential to catalysis and the allosteric transition. Alternatively, adianionic cofactor phosphate could also function to destabilize the ground state enzyme-substrate complex through electrostatic interactions. From an evolutionary viewpoint,phosphate would be a unique candidate for just such a role. Of the 20 common aminoacids none are capable of forming a dianion, yet like amino acids, the phosphate moiety(covalently bound to PLP) can be held in juxtaposition to reacting groups within the activesite complex.While enzymes utilize ground state binding interactions to stabilize the enzyme-substrate complex, some of this energy can be used to destabilize reacting groups in theground state relative to the transition state (Jencks, 1987). Thus, in order to be catalytic thedestabilization must be relieved on reaching the transition state. For example (Figure 2-9),if the binding interactions between component groups of the substrate and the active site ofan enzyme provide a favorable standard free energy change for substrate binding, LG'ES,in the absence of any ground state destabilization or transition state stabilization, theactivation energy for the bound substrate, AG4Es, will be the same as that for the freesubstrate (AGt). However, if the enzyme can make use of binding interactions to force thesubstrate into a structural or electrostatic interaction that destabilizes its ground state by anamount AGD, but has no effect on the transition state, then the activation energy for thebound substrate will be reduced by AGD.Ground state destabilization has been suggested previously to contribute to thecatalytic effectiveness of the bound metal ion (Zn+2) of carboxypeptidase A (Lipscomb etal., 1969), an enzyme that cleaves the carboxy-terminal amino acid from polypeptidesubstrates. The replacement of solvating water by the substrate was postulated to reduce65AstS°G $^AES^:.1GJES ES ^t '1GD#^#ES ^ ES^GEst^1A (° GRs-L GD)the dielectric constant surrounding the metal atom and thereby increase its activity inpolarizing the acyl group for nucleophilic attack.AGFigure 2-9. Ground state destabilizationSimilarly, pyruvate decarboxylase is thought to utilize ground state destabilization as ameans of rate enhancement (Crosby et al., 1970). Decarboxylation of pyruvate is thoughtto occur through a zwitterionic intermediate formed from pyruvate and thiaminepyrophosphate. Analogues of the pyruvate adduct (Scheme 2-3) undergo decarboxylation104 - 105 times more rapidly in ethanol than water and still faster in aprotic solvents. Therate increase in this non-catalyzed reaction is thought to occur because of a decrease incharge localization at the transition state, thus in a poorly ion-solvating environment theground state is destabilized relative to the transition state. It becomes apparent then, that ifthe enzyme forces the carboxylate of pyruvate and the cationic nitrogen of thiaminepyrophosphate into an environment which is less polar than water, a large rate enhancementmay be expected. The favorable binding interactions between substrates and the enzyme66CH3provide the binding energy to initially hold the charged portions of the substrates in anunfavorable environment.CH3N8+ CH38-0.^OH^,.,. CH3,.otCH3N (CH3^0^HOII^CH3IIC0Scheme 2-3. A model for the decarboxylation reaction catalyzed by pyruvatedecarboxylase.A somewhat different role for ground state destabilization of the substrate (aG1P)in phosphorylase catalysis, has been suggested previously on the basis of recent X-raycrystallographic studies of phosphorylase b in the presence of various aG1P derivatives(Martin et al., 1990, Johnson et al., 1990). These studies show that the bound substratephosphate occupies a position under the C-2 hydroxyl group of the sugar ring distinct fromthe more stable rotamer in which the phosphate group is oriented away from the sugar ring,trans to C-2 (O'Connor et al., 1979). Steric restrictions imposed on the conformationalfreedom of the glucosyl phosphate, in addition to the formation of an internal hydrogen67OH0HOHOH,-, 0IIObond between phosphate and the hydroxyl group at C-2, are both thought to be somewhatresponsible for the positioning of the glucosyl phosphate moiety (Martin et al., 1990).Mechanistically, it was suggested (Martin et al, 1990) that the 'under the ring'conformation prevents exo-anomeric stabilization of the sugar-phosphate bond, and therebyinduces a ground state destabilization of the aG1P substrate (Figure 2-10).Figure 2-10. The 'under the ring' conformation for aG1P.Since the 'under the ring' conformation places the substrate phosphate in juxtaposition tothe PLP phosphate, it was proposed that the PLP phosphate moiety is better able to act asan essential acid catalyst (Martin et al., 1990). However, results from the present studypreclude the PLP phosphate moiety as an essential acid catalyst, and therefore, while the'under the ring' conformation may serve to destabilize the ground state substrate complex itmust serve to preferentially orient the substrate and cofactor phosphates for some reasonother than one of essential proton transfer. Indeed, this preferred substrate conformation isconsistent with the electrophilic mechanism previously described (Withers et al., 1982a;Takagi et al., 1982), or alternatively a ground state electrostatic destabilization.Thus, mechanistically, glycogen phosphorylase may proceed via an initial bindingstep which provides the necessary binding energy to bring the substrate phosphate into68Nuc1-0IIPL-0—P-0-GLYCOGENo-close contact with the PLP phosphate and provide the proposed electrostatic destabilization,possibly employing the 'under the ring' conformation (Scheme 2-4).0 HO^HNuc II^HO ^ Nzuc)HOHOHO^...i.-}^HO—P —o-^HO'---70 (-))fI Ii S../AM/P\AH "0 a oI^I IPL-0—P-0 -1—0"O0 H OHO ^0HHO—P--0"^HOB JHOA-H HO OI—PL-O—P —0"GLYCOGENScheme 2-4. A possible catalytic role for PLP in glycogen phosphorylaseTo assist in bond cleavage, acid catalysis from an active site amino acid (not PLP) likelyprotonates the phosphate leaving group. Destabilization of the ground state enzyme-nHOHOA" HOOH69substrate complex could be somewhat relieved at the transition state by any movement ofthe phosphate leaving group away from the cofactor phosphate. Indeed, very recent X-raystructural studies have found a second substrate phosphate binding site directly adjacent tothat which binds the phosphate within aG1P (Sprang et al., 1992). The formation of aglucosyl-enzyme covalent intermediate or a stabilized ion pair is thought to be the initialproduct of aG1P bond cleavage. The enzyme intermediate or ion pair can then be attackedby the 4-hydroxyl group of the oligosaccharide acceptor molecule, in a general basecatalyzed process such that the anomeric configuration is retained.While the results from the present study suggest that the PLP phosphate moietybound at the active site of glycogen phosphorylase is not involved in an essential protontransfer process, and may indeed remain a fully ionized species throughout the allosterictransition, the exact catalytic role of the cofactor phosphate is not known and willundoubtedly remain a topic of continuing controversy for years to come. However, thisnew role in ground state destabilization is another alternative, consistent with all theevidence, which has not been considered previously.70CHAPTER 3Inactivation of Glycogen Phosphorylase by a Novel Affinity LabelResults and Discussion71Introduction The mechanistic details behind the action of glycogen phosphorylase have eludedinvestigators for decades predominantly because the enzyme is extremely substrate specific,providing catalysis to only a select few substrate analogues. While X-ray crystallographicstudies can be immensely useful to mechanistic investigations, it may prove very difficult(as with glycogen phosphorylase) to prepare crystalline samples of the enzyme in aconformation which presents the active site residues in catalytically relevant positions.Previous studies have addressed this issue with reagents that inactivate glycogenphosphorylase through the chemical modification of amino acids essential to the catalyticmechanism (Avramovic-Zikic et al., 1974; Battell et al., 1968; Takagi et al., 1989; Dreyfuset al., 1980). While these studies have provided valuable clues regarding which aminoacids might be essential to catalysis, or to the allosteric transition, they can suffer from alack of specificity since the reagents employed do not utilize active site binding interactionsdesigned specifically for the substrate.Results 1. Inactivation of Glycogen Phosphorylase b with Nitroglucal 1.1. Background and SynthesisPrevious experiments in this laboratory have shown that 1-nitro-D-glucal(nitroglucal) (Scheme 3-1) irreversibly inactivates the enzyme p-glucosidase in a time-dependent fashion, likely behaving as an affinity label via a conjugate addition reaction atthe activated C-2 position. Since glycogen phosphorylase is known to bind D-glucal, albeitweakly (KD = 80 mM) (Sprang et al.,1982), nitroglucal, in addition to 1-carboxylic acid-D-glucal and 1-carboxylic acid methyl ester-D-glucal were tested as potential inactivators ofglycogen phosphorylase b. Nitroglucal was a generous gift from Dr. A. Vasella (Beer etal., 1986), whereas the acid and methyl ester derivatives of D-glucal were prepared fromglucal, and were available from previous synthetic work (Scheme 3-1). The synthesis of72the acid and ester analogues of glucal involves a C-1 vinylic deprotonation of protected D-glucal (Boeckman and Bruza, 1981; Lesimple et al., 1986), the lithio anion of which canthen be reacted with carbon dioxide to afford the carboxylic acid derivative. However, itshould be noted that complications in this reaction can arise from the use of tert-butyldimethylsilyl hydroxyl protecting groups, due to the formation of a-silyl carbanions(Friesen et al., 1991), and hence, the use of tert-butyldiphenylsilyl protecting groups isadvised. The methyl ester derivative is simply prepared from the carboxylic acid analogueby treatment with diazomethane. The silylated acid and methyl ester analogues of D-glucalwere deprotected using fluoride ion, and subsequently purified using ion-exchange andsilica-gel chromatography, respectively.R = NO2, C00 - , or COOMeScheme 3-1. Analogues of D-glucal.1.2. Preliminary Inactivation StudiesSince both the T- and R-state conformations of phosphorylase bind glucosylanalogues and are important for the allosteric transition, all three derivatives of glucal (1-nitro, acid, and methyl ester) were tested as potential inactivators of glycogenphosphorylase b under conditions promoting both the T-state (Figure 3-1a) and R-state(Figure 3-lb) enzyme conformation.73200^300Time (min)100 400 500400^800Time (min)1200 16000.0-2.0-3.00.0-1.0"."6> -2.0c-3.0-4.0Figure 3-1. Inactivation of glycogen phosphorylase b with glucal analoguesThe inactivations were conducted at pH 6.8 and 30 °C in 100 mM KC1, 50 mM triethanolaminehydrochloride, 1 mM EDTA and 1 mM DTT. Activity measurements were carried out in the same buffercontaining 16 mM aGIP, 1 mM AMP and 1% glycogen. a) Inactivation of the T-state enzyme (no R-stateeffectors present): (II) Nitroglucal (40.6 mM), kobs = 2.76 x 10-2 min -1 ; (0) 1-Carboxylic acid-glucal(38.4 mM), kobs = 3.88 x 10 -4 min -1 ; (A) 1-Carboxylic acid methyl ester-glucal (31.0 mM), k obs = 1.12x 10-4 min -1 ; T-state enzyme control (0). b) Inactivation of the R-state enzyme (1 mM AMP and 1%glycogen present): (•) Nitroglucal (26.6 mM), kobs = 4.61x 10-3 min-1 ; (0) 1-Carboxylic acid-glucal(38.4 mM), kobs = 1.15 x 10-4 min -1 ; (A) 1-Carboxylic acid methyl ester-glucal (31.0 mM), kobs = 1.80x 10-5 min -1 ; R-state enzyme control (0).74As can be seen in Figure 3-1, at the concentrations employed, the methyl ester andcarboxylic acid derivatives of glucal had no significant effect as time-dependentinactivators. Activity loss was not significantly greater than that in controls containing noinactivator. Conversely, when either allosteric conformer of glycogen phosphorylase bwas incubated with nitroglucal, enzyme activity was clearly observed to decrease in a time-dependent fashion. The inactivation was shown to be irreversible since the enzyme activitycould not be restored after excess nitroglucal was removed by dialysis. The time-dependent decrease in enzyme activity (V/V0) was approximately first order in both cases.1.3. Inactivation Kinetics For R- and T-state Glycogen Phosphorylase bTo further characterize the inactivation due to nitroglucal, the constants, ki andfor the reaction of nitroglucal with both the R- and T-state forms of glycogenphosphorylase b were determined. Residual activity was assayed at time intervals using thecontinuous phosphoglucomutase / glucose-6-phosphate dehydrogenase coupled assay sincethis assay was far more convenient over the course of long inactivation times than thestopped phosphate analysis previously employed. Reducing agents such asmercaptoethanol and dithiothreitol were omitted from the incubation mixtures since separateexperiments confirmed that these reagents underwent a conjugate addition reaction withnitroglucal. While the enzyme possesses slightly less activity under non-reducingconditions the loss of activity is fully reversible. The first order rate constant (kobs)obtained at each concentration of nitroglucal was fitted to the non-linear form of theMichaelis-Menten equation and the results obtained in this way are presented inLineweaver-Burk form in Figure 3-2 and Figure 3-3.750.0 -0.5 --1.0 -o-1.5-2.0 --2.5 -• 5.04 mM• 10.1 mM• 20.2 mM• 57.3 mM-3.00^ 100Time (min)200b)t0.20300 —200 -00.00^0.05^0.10^0.151/[Nitroglucal] (1/mM)Figure 3-2. Nitroglucal inactivation kinetics for T-state glycogen phosphorylase ba) A semilogarithmic plot of residual activity versus time for the inactivation of T-state glycogenphosphorylase at 30 °C and pH 6.8 in buffer containing 10 mM MES, 10 mM HEPES, 10 mMtriethanolamine hydrochloride, and 100 mM NaCl. The following nitroglucal concentrations were used andthe kobs values obtained are listed in parentheses: 5.04, 10.1, 20.2, 57.3 mM; (4.35 ± 0.31, 8.34 ± 0.78,12.4 ± 0.5, 16.2 ± 0.7) x 10-3^b) A double reciprocal plot of the k obs values obtained at theirrespective nitroglucal concentrations. Activity measurements were performed using the coupled assaydescribed previously by Engers et al., (1969).760.0-0.50,_.-1.0-1.5-2.00 6.14 mM• 24.6 mMO 36.8 mM0^100 ^200^300-;-.300 -200 -Time (min)b). ff..00100  -0 I I I0.00^0.05^0.10^0.15^0.201/[1\litroglucal] (1/mM)Figure 3-3. Nitroglucal inactivation kinetics for R-state glycogen phosphorylase b.a) A semilogarithmic plot of residual activity versus time for the inactivation of R-state glycogenphosphorylase at 30 °C and pH 6.8 in buffer containing 10 mM MES, 10 mM HEPES, 10 mMtriethanolamine hydrochloride, 100 mM NaCI, and 1 mM AMP. The following nitroglucal concentrationswere used and the kobs values obtained are listed in parentheses: 6.14, 12.3, 24.6, 36.8, 49.1 mM; (3.66 ±0.23, 4.44 ± 0.31, 4.69 ± 0.31, 5.70 ± 0.35, 5.49 ± 0.33) x 10 -3 min - I. For clarity, only three of thefive lines are shown. b) A double reciprocal plot of the kobs values obtained at their respective nitroglucalconcentrations. Activity measurements were performed using the coupled assay described previously byEngers et al., (1969).77As observed in Figures 3-2a and 3-3a, the inactivation data for the R-state enzyme,unlike that for the T-state enzyme, appears to deviate somewhat from first order behaviour,possibly due to labelling of an alternate amino acid which can also affect enzyme activity.However, reasonable fits of the earlier time points could be obtained allowing estimates ofthe inactivation rate and binding constant (ki and Ki) to be made. The rate constant forinactivation (ki) of glycogen phosphorylase b was found to be almost four fold greater inthe T-state enzyme ((2.1 ± 0.2) x 10 -2 min -1 ) than in the R-state conformation ((0.60 ±0.03) x 10 -2 min-1 ). If the mechanism of inactivation is the same for both allosteric formsof the enzyme, it seems reasonable that the active site residue responsible for theinactivation could be slightly better positioned in one conformation than in the other.Interestingly, the binding constant for nitroglucal in the R-state enzyme (4.1 ± 1.2 mM)suggests a four fold greater affinity than that observed in the T-state enzyme (15.9 ± 2.9mM). The second order inactivation constant ki/Ki is however approximately the same inthe two cases.1.4. Active Site-Directed InactivationTo test for the active site-directed nature of the nitroglucal inactivation, glycogenphosphorylase b was incubated with nitroglucal in the presence of competitive ligandswhich are known to reversibly bind to, or block, the enzyme active site. Two suitableligands for glycogen phosphorylase include glucose and caffeine. Glucose binds tophosphorylase (KD ti 2 - 4 mM) utilizing the same active site interactions which bind thesubstrate, while alternatively, caffeine binds to the enzyme (KD 0.1 - 0.2 mM) at aninhibitor site approximately 10 A from the active site, and once bound, effectively blocksaccess to the active site pocket. Thus, glycogen phosphorylase b was incubated withnitroglucal in the presence of glucose and / or caffeine at 30 °C and pH 6.8 and residualactivity was assayed at time intervals using the continuous phosphoglucomutase / glucose-6-phosphate dehydrogenase coupled assay as previously described. The time course for78nitroglucal inactivation, obtained in this way, in the presence of competitive ligands ispresented in Figure 3-4.-2 150•^i100Time (min).^115010 200Figure 3-4. Protection against nitroglucal inactivation by a competitive ligand.Inactivation of glycogen phosphorylase b at 30 °C and pH 6.8 in buffer containing 10 mM MES, 10 mMHEPES, 10 mM triethanolamine hydrochloride, and 100 mM NaCl. The inactivation was performed at aconstant concentration of nitroglucal (24.6 mM) with the following competitive ligands: ■ No additionalligands (kobs = (7.98 ± 0.18) x 10-3 min-1 ); 0 glucose (7.5 mM) (kobs = (6.46 ± 0.22) x 10-3 min-1); •caffeine (0.75 mM) (kobs = (3.78 ± 0.41) x 10 -3 min-1 ); 0 glucose (7.5 mM) and caffeine (0.75 mM)(kobs = (2.88 ± 0.26) x 10-3 min-1 ). Activity measurements were performed using the coupled assaydescribed previously by Engers et al., (1969).Results from the protection experiment show that glucose and caffeine both affordprotection against inactivation by nitroglucal. In particular, at the concentrations employed,glucose (= 2 x KD) provided approximately 20% protection, and caffeine (= 5 x KD)reduced by more than half the rate of inactivation observed in the absence of competitiveligands. Together, acting as non-exclusive competitive inhibitors, glucose and caffeineprovided the expected level of protection against inactivation. That protection againstnitroglucal inactivation was afforded by these competitive ligands is in full agreement withthe proposal that the inactivation observed with glycogen phosphorylase b is an active site79event. However, this is not to suggest that nitroglucal cannot react at other sites on theprotein, only that reaction at the active site is responsible for the observed enzymeinactivation. A simple kinetic model for the proposed inactivation is shown below inScheme 3-2.K1^kiI + E^EI --> E-IScheme 3-2. A kinetic model for the irreversible inactivation of glycogen phosphorylaseby nitroglucal.In this model, nitroglucal (I = inactivator) is involved in an initial binding step (Ki) and anirreversible, rate limiting bond forming step (k 1). If the concentration of I is much largerthan [E] o , such that [I] - [E]o [I], then the concentration of I is essentially constant overthe course of the reaction and the kinetics are pseudo first order with respect to enzymeconcentration. The Michaelis-Menten equation for this inactivation can be written asfollows:vi = ki[E]o[I]^(1)Ki + [I]where ki is the rate constant for inactivation and Ki is the dissociation constant for theinactivator. Since [I] is assumed to be constant, Equation 1 can be rewritten as:vi = kobs[E]t^(2)where^ kobs =  ki [I]^(3)K 1 + [I]80The value of kobs at each inactivator concentration can be determined by fitting the velocitydata to a standard first order decay function. The values of ki and Ki are determined byfitting the observed rate constants (kobs) to Equation 3.1.5. pH-Dependence of the Inactivation Kinetic Constants (ki and KOThe pH-dependence of the kinetic constants (ki and Ki) for the inactivation of T-state glycogen phosphorylase b was investigated in an effort to monitor changes in theinactivation process with varying pH values. The inactivation was carried out in a singlebuffer system at 30 °C over a range of pH. Residual enzyme activity was measured in thedirection of glycogen degradation using the phosphoglucomutase / glucose-6-phosphatedehydrogenase coupled assay, described previously by Engers et al., (1969). The firstorder rate constants (kobs) obtained for the various concentrations of nitroglucal were fittedto the non-linear form of the Michaelis-Menten equation to provide the inactivationconstants (ki and Ki) at each pH studied (Figure 3-5). The pH-dependences of the kineticparameters (ki, Ki and ki/Ki) for the inactivation of T-state glycogen phosphorylase bobtained in this way are shown in Figure 3-6. Apparent pK a values were calculated withthe aid of the GraFit computer program (Leatherbarrow, 1990).81O pH 8.5• pH 9.0500400SEen.ra3002001000300250Cs 200Een 1500Y■,— 100500140120S 100Ecn80.00 60E402000^0.2^0.414Nitroglucal] (1/mM)o pH 6.8• pH 5.9o pH 7.9• pH 7.20.60^0.2^0.4^0.6^0.81/Nitroglucal) (1/mM)0^2^4^6^8^101/[Nitroglucal] (1/mM)Figure 3 -5. Nitroglucal inactivation kinetics over a range of pH.Double reciprocal plots of kob s versus nitroglucal concentration, determined over a range of pH.Inactivation reactions were carried out at 30 °C in a buffer containing 10 mM MES, 10 mM HEPES, 10mM triethanolamine hydrochloride, and 100 mM NaCI at the pH values listed. Residual activitymeasurements were performed using the coupled assay described previously by Engers et al., (1969).82•a)-1.2-1.4-1.6-1.8-2-2.25 6 7 8 9 10b)0.40-0.4-0.8-1.2-1.65 6 7 8 9 105 6 7^8pH9 1 0-1-2.5-3Figure 3 -6. The pH -dependence of nitroglucal inactivation.a) A plot of log ki versus pH. The following values of ki were obtained at the pH values indicated inparentheses: 0.064 (5.90), 0.021 (6.80), 0.010 (7.20), 0.019 (7.90), 0.032 (8.47), and 0.060 min -1 (9.00).b) A plot of pKi versus pH. The following values of Ki were obtained at the pH indicated in parentheses:32.7 (5.90), 15.9 (6.80), 5.6 (7.20), 4.3 (7.90), 1.1 (8.47), and 0.4 mM (9.00). c) A plot of log ki/Kiversus pH. Apparent pKa values were calculated with the aid of the GraFit computer program. The curvesin these plots are not intended to represent any pH function and are for presentation purposes only.83The pH-profiles shown in Figure 3-6 demonstrate a considerable variation for boththe inactivation rate constant (k,) and the nitroglucal binding constant (Ki), over the rangeof pH studied. The pH-dependence of log which reflects ionizations in the EI complexjust prior to the rate limiting bond formation step, (Figure 3-6a) shows an increase at lowand high pH, the apparent pK a values for the acidic and basic limbs being 6.1 ± 0.1 and8.8 ± 0.1, respectively. Since there is good evidence that the inactivation is active site-directed, and furthermore because nitroglucal is not expected to change ionization state inthe pH-range studied, the observed increase in k, at low and high pH must arise fromchanges in protonation originating from active site residues in the EI complex. Theobserved pH-dependence of lc, is consistent with an inactivation reaction that is enhancedby acid and base catalysis, however, it cannot be ruled out that changes in the active siteconformation (i.e. positioning of potential active site nucleophiles) also contribute to theobserved variation in k1.The nitroglucal binding constant (KO (Figure 3-6b) was observed to decreasesignificantly with increasing pH, the apparent pK a for the variation being 6.5 ± 0.1. ThepH-dependence of K, depends on the ionizations of the free and bound enzyme in additionto the ionizations of the free inhibitor (Knowles, 1976). In this case only the enzyme isexpected to change ionization state over the pH-range studied, and thus it appears that theground state interactions between the enzyme and nitroglucal considerably increase with thedeprotonation of active site amino acid residues.The pH-dependence of WK,, the second order rate constant for inactivation, whichreflects ionizations in the free enzyme and free inhibitor which affect the bond formationprocess, is shown in Figure 3-6c. This describes an ionization of pK a = 9.7 ± 0.3 and canlikely be assigned to the deprotonation of the putative nucleophile in the free enzyme.841.6. Structural Analysis of the Nitroglucal-Glycogen Phosphorylase bComplexTo further investigate the interactions between nitroglucal and glycogenphosphorylase b, X-ray crystallographic studies of nitroglucal bound to the T-state form ofthe enzyme were carried out in collaboration with E. P. Mitchell and Professor L. N.Johnson in the Laboratory of Molecular Biophysics at Oxford University. Glycogenphosphorylase b was crystallized in the presence of IMP (a weak allosteric activator) andthen soaked (Martin et al., 1990) with nitroglucal for an extended period of time (-...  2 days)before acquiring a difference Fourier electron density map of the enzyme in the presence ofinactivator. Results from the structural analysis revealed that nitroglucal was bound at twosites within the enzyme, the active site and also a secondary site near the surface of theenzyme (Figure 3-7). At the active site, nitroglucal was observed bound to the enzyme inthe same high specificity pocket used to bind the glucosyl portion of the substrate (aG1P)and also the allosteric effector glucose, although no covalent bonds to the enzyme wereobserved. While disappointing, this is quite reasonable since it is possible that, in thecrystalline lattice, motional flexibility is considerably restricted, thereby preventing covalentbond formation at the active site. Nonetheless, all hydrogen bonding interactions normallyobserved at the active site of glycogen phosphorylase b were in place with nitroglucal,excluding of course those hydrogen bonds normally made to the C-2 hydroxy group.Those amino acids which normally hydrogen bond to the C-2 hydroxyl group include Tyr-573, Glu-672 and Asn-284. However, in the nitroglucal structure Asn-284 is observed todonate a hydrogen bond to the nearest oxygen atom within the nitro group. Residues inclose proximity to the activated C-2 position of nitroglucal include Tyr 573 (4.5 A) and Glu672 (3.4 A). Attempts at longer soaking periods with nitroglucal (.,--. 2 weeks) resulted incrystals that no longer diffracted (E. Mitchell, personal communication).858 6Figure 3-7. X-ray crystallographic analysis of T-state glycogen phosphorylase b in thepresence of nitroglucal.a) Nitroglucal bound at the active site, shown with the Fourier difference map in blue. b) Nitroglucalbound at the active site with the five closest amino acids to C-2 shown. c) Nitroglucal covalently bound toHis-73, shown with the Fourier difference map in blue.87While no covalent bonds between the inactivator and enzyme were observed at theactive site of the enzyme, a covalent bond was found between His-73 and the sugar moietyof a second molecule of what was nitroglucal (Figure 3-7). The difference Fourier electrondensity map clearly shows the sugar moiety in the chair conformation with a covalent bondextending axially from the C-2 position of the sugar to the E-2 nitrogen (NE2) atom of thehistidine moiety. Not surprisingly, since no binding site for nitroglucal exists at the surfaceof glycogen phosphorylase, the nitroglucal moiety once bound to His-73 does not makeany strong binding interactions to nearby amino acids. In fact there is only one hydrogenbond formed between the protein and the nitroglucal-His-73 adduct, and that is between theC-6 hydroxyl and the OH of Tyr-74. Interestingly, His-73 is located in close proximity tothe AMP binding site, possibly suggesting a secondary labelling site which might beimportant to the R-state inactivation process observed in solution. X-ray crystallographicstudies of nitroglucal bound to the R-state form of glycogen phosphorylase are currently inprogress at Oxford University.1.7. Mass Spectral Analysis of Nitroglucal Inactivated GlycogenPhosphorylase bThe results from kinetic and X-ray crystallographic studies appear to confirm therole of nitroglucal as an active site-directed affinity label which, not surprisingly, is alsoable to covalently bond to residues beyond those at the active site. Preliminary data fromion-spray mass spectral studies of the inactivated enzyme support the suggestion thatmultiple labelling indeed occurs. In collaboration with Dr. J. Gebler and Dr. R. Aebersold,mass spectral data on the native enzyme and the inactivated enzyme were collected andcompared. Mass spectral analysis of the native protein, untreated with inactivator,produced a molecular weight of 97,221 ± 10 Da. The monomer molecular weight ofglycogen phosphorylase is reported to be 97,444 Da (Johnson, 1992), however, the PLPcofactor is extracted under the conditions employed for mass spectral analysis, and88therefore the target mass is reduced to 97,213 Da, thus corresponding well with theobserved mass. Mass spectral analysis of the inactivated enzyme (> 90% inactivated)suggests that the sample was composed of many species of various molecular weights.Indeed, the results suggested that glycogen phosphorylase was labelled with several (up to10) nitroglucal moieties, and that accurate molecular weights of these adducts could not becalculated from the mass spectral data due to the heterogeneity of the protein sample.Furthermore, based on low molecular weight species also present within the proteinsample, these results suggest that glycogen phosphorylase may have undergone somedegradation under the conditions of extensive labelling. This latter result suggests that thefully labelled form of the enzyme may be less stable than the native enzyme, and that theconditions used for mass spectral analysis may enhance peptide cleavage reactions. Proteinprecipitation observed at high concentrations of the inactivated enzyme further suggests thatextensive labelling of glycogen phosphorylase may, in addition, affect enzyme solubilityproperties.2. Discussion Over the years considerable effort has been extended to the study of various affinitylabels for glycogen phosphorylase (Avramovic-Zikic et al., 1974; Battell et al., 1968;Takagi et al., 1989; Dreyfus et al., 1980). The focus behind the effort is to identify aminoacid residues essential to catalysis and / or the enzyme allosteric transition. One prominentexample of an active site amino acid studied by chemical modification is Arg-569, whichcan be modified by butanedione, but only in the activated R-state conformation.Subsequent X-ray crystallographic studies revealed that upon enzyme activation, Arg-569swings into the active site, displacing Asp-283, and creates the phosphate binding site towhich substrate can bind. This is clearly consistent with the butanedione labelling studies.While chemical modification studies can be extremely useful, especially when they arecoupled to X-ray structure analysis, they often lack specificity since the reagents used do89not utilize active site interactions, thus they are not placed in juxtaposition to importantcatalytic residues.Preliminary studies with nitroglucal suggested that both the R- and T-stateconformations of glycogen phosphorylase b were inactivated by the nitro derivative, andfuthermore, that the carboxylic acid and methyl ester analogues of glucal were essentiallyunreactive. Since the a, 0-unsaturated methyl ester is expected to be considerably lessactivated than the nitro analogue (Shenhav et al., 1970), and further, since the carboxylicacid derivative should be essentially unreactive unless it is protonated upon binding to theenzyme, these results are therefore consistent with an inactivation involving a conjugateaddition reaction. Nitroglucal inactivation of glycogen phosphorylase b was shown to beactive site-directed since protection against inactivation could be afforded by ligands whichare competitive for, or sterically block the active site. This is not unreasonable given thatnitroglucal is structurally similar to D-glucal, an analogue that is known to bind to, and inthe presence of phosphate, undergo phosphorylation at the active site of glycogenphosphorylase (Klein et al., 1982). These results strongly support the proposal thatnitroglucal indeed binds to the active site of glycogen phosphorylase utilizing active sitebinding interactions, and inactivates the enzyme through derivatization of an active siteamino acid.The nitroglucal bound at the active site was positioned with the activated glucosylC-2 position in close proximity to the side chain residues of both Tyr-573 (4.5 A) and Glu-672 (3.4 A). While either of these amino acid side chains might be expected to act in aconjugate addition reaction, no covalent bonds to nitroglucal were observed at the activesite. However, additional evidence supporting a nucleophilic role for Tyr-573 in theinactivation was obtained from the pH-dependences of ki/Ki and ki, wherein, apparent pKavalues calculated for the basic limb of the plots were 9.7 and 8.8 respectively, certainlywithin range of a tyrosine residue located at the active site of the free enzyme (ki/Ki) andenzyme-inhibitor (ki) complex. Indeed, based on the pH-dependence (pK a = 8.3) of the in90situ degradation of a cofactor-substrate analogue, previous studies (Horinishi et al., 1988)have implicated Tyr-573 in a catalytic role, possibly functioning as a base to deprotonatethe C-2 hydroxyl moiety and thereby catalyze the degradation process. Thus, thenitroglucal inactivation of glycogen phosphorylase b might well involve derivatization of anactive site residue such as Tyr-573. Unfortunately since a covalent bond was not observedat the active site of the T-state enzyme, the involvement of either Tyr-573 or Glu-672remains rather speculative. X-ray structural studies of the R-state enzyme, in the presenceof nitroglucal, are currently in progress and thus a structural comparison of the twoinactivated enzymes will be possible.Structural differences between the R- and T-state enzyme conformations are likelyto be responsible for the differences observed in nitroglucal affinity, the R-state enzymebinding nitroglucal with approximately four fold greater affinity than the T-state enzyme.While this difference in affinity between the two enzyme conformations is not large it mightstill be rationalized on the basis of the half chair conformation adopted by nitroglucal, oralternatively, by differing interactions localized to the nitro group. Catalysis by glycogenphosphorylase is thought to proceed through oxocarbonium ion-like transition states (Streetet al., 1989), and thus, half chair analogues mimicking the transition state such asgluconolactone or norjirimycin tetrazole tend to bind phosphorylase with increased affinity(Tu et al., 1971; Withers, Unpublished results). Some of the same R-state interactions thatprovide increased affinity for these analogues might also be somewhat responsible for themoderate increase in nitroglucal affinity observed here in the R-state enzyme. It is alsoknown that one of the main changes accompanying the T- to R-state allosteric transition isthe creation of a positively charged active site pocket which can electrostatically stabilizenegatively charged substrates. It seems possible that the increase in nitroglucal affinity forthe R-state enzyme could arise from favorable interactions localized to the nitrofunctionality, interactions which are less available in the T-state enzyme. X-ray structureanalysis of the T-state enzyme suggests that Asn-284 is the only amino acid to interact with91the polar nitro group. Asn-284 normally donates a hydrogen bond to the C-2 hydroxylgroup in glucose and since nitroglucal lacks this hydroxyl moiety, Asn-284 donates itshydrogen bond to the nearest oxygen atom within the nitro group. Once completed, thestructural analysis of nitroglucal bound to the R-state enzyme will aid in understanding theobserved difference in affinity for the two enzyme conformations.Further information related to the binding interactions between nitroglucal and theT-state enzyme conformation comes from the pH-dependence of Ki. The binding constant(Ki) for nitroglucal was observed to decrease by approximately 80 fold (an increase inaffinity) as the pH was increased from 5.8 to 9.0, the apparent pK a for the variation being6.5. Previous studies on the hydrogen bonding interactions around the glucose ring haveindicated that the most important interactions between glucose and the enzyme are those atthe 3, 4, and 6 hydroxyl positions (Street et al., 1986). Since these are the only hydroxylsavailable in nitroglucal, any change in the network of hydrogen bonding interactions nearthese positions could significantly affect the nitroglucal affinity for the T-state enzyme.Thus, while the source of the increased affinity at higher pH is not fully understood, it ispossible that the hydrogen bonds at the 3, 4 or 6 positions become better oriented upondeprotonation of a nearby amino acid residue. Alternatively, the creation of one or morefavorable interactions between the nitro group and the enzyme could also be responsible forthe observed increase in affinity with increasing pH. Previous studies with glycogenphosphorylase (Street et al., 1986) have indeed shown that deletion of a charge-neutralhydrogen bond at the active site of the enzyme can easily result in a similar decrease inaffinity, thus it is not unreasonable to suggest that the formation of one or more favorableinteractions between nitroglucal and the enzyme as the pH increases could be responsiblefor the increase in affinity observed here.While nitroglucal was observed to inactivate the T-state enzyme according to a firstorder process, the inactivation kinetics for the R-state enzyme appear to deviate somewhatfrom first order. The most likely explanation for this result involves a slowly reacting92secondary labelling site which serves to complicate the inactivation process for the R-stateenzyme, but which is less likely to affect the more rapidly inactivating T-state enzyme. X-ray structural analysis of the T-state enzyme, previously soaked with nitroglucal, revealsone such secondary binding site located near the surface of the enzyme in close proximityto the AMP binding site. Thus, a covalent bond was observed between the C-2 position ofnitroglucal and the E-2 nitrogen atom (NE2) of His-73, indicating that conjugate additionhad occured from the top face of the nitroglucal ring, and furthermore, that protonation atCl had occurred from below, the source of the proton probably being solvent water. Oneof the predominant interactions involved in the binding of AMP to glycogen phosphorylaseb is a hydrogen bond between the hydroxyl group of Tyr-75 and the phosphate moiety ofAMP (Oikonomakos et al., 1991). If covalent bond formation at His-73 interferes in anyway with the interactions that stabilize AMP binding, this secondary labelling site couldcertainly contribute to the deviation from first order inactivation kinetics observed with theR-state enzyme. Of course, secondary labelling at His-73 could also be occurringthroughout the T-state inactivation process, however, since the T-state kinetics do notdeviate significantly from first order decay it seems unlikely that derivatization of His-73occurs at a comparable rate. Further, since structural differences exist between the R- andT-state conformations at the AMP binding site (Oikonomakos et al., 1991), the relativeaccessibility of His-73 in solution, and therefore the rate of labelling in the two conformers,could be quite different.While covalent bond formation was not observed at the active site of the enzyme inthe structural analysis, it seems highly likely that a conjugate addition reaction isresponsible for the inactivation observed in the R- and T-state conformations of glycogenphosphorylase b. The tendency for a, 13-unsaturated nitro compounds to undergoconjugate addition reactions with a wide variety of nucleophiles is well known (Sidgwick,1966). Moreover, given the clear evidence that His-73 reacts at the C-2' position ofnitroglucal in the same way, the proposal that a similar conjugate addition reaction at the93active site of glycogen phosphorylase is responsible for inactivation of the enzyme is notunreasonable. Indeed, the pH-dependence of ki also supports this proposal since the rateconstant for inactivation was observed to considerably increase at low and high pH,consistent with general acid and base catalytic involvement in the conjugate additionreaction. Conjugate addition reactions are normally carried out under base catalyzedconditions (Bergmann et al., 1959), the base of which deprotonates the nucleophile foraddition to the activated double bond. General acid catalysis might also be envisioned toparticipate in the conjugate addition reaction, if indeed, at the active site of the enzyme, thenitro group receives some degree of proton donation from a neighboring amino acid sidechain, or alternatively, if a low pH-induced conformational change places the nucleophilecloser to the C-2 position.Conclusion The results obtained for the inactivation of glycogen phosphorylase b withnitroglucal suggest that the inactivation process is active site-directed and that derivatizationof nitroglucal occurs through a conjugate addition reaction. Further, the results suggestthat one active site residue is predominantly responsible for the inactivation, however, theydo not rule out the possibility of secondary labelling sites. Indeed, X-ray crystallographicstudies confirm at least one other derivatization site, His-73, which is available in thecrystalline T-state conformation. Crystal lattice packing constraints likely preclude theaccessibilty of many other potential derivatization sites on the surface of the enzyme. Thatseveral secondary labelling sites could well be derivatized in solution was supported in amass spectral study of the fully inactivated enzyme. This was not unexpected given thelarge number of potential nucleophiles available on the surface of glycogen phosphorylase.Structural studies designed to locate nitroglucal bound to the R-state enzyme are currentlyin progress. Once completed, future studies with radiolabelled nitroglucal might serve toidentify or confirm the active site residues within the T- and R-state enzyme labelled by the94inactivator. Finally, since nitroglucal possesses no anomeric configuration, and indeed,has now been shown to inactivate an a-glucosyl transferase (glycogen phosphorylase) aswell as a 1i-glucosyl transferase (Ellen Lai, personal communication), it may be possible toextend the use of nitroglucal or similar affinity labels to other more general glucosylbinding proteins.95CHAPTER 4A Novel Cofactor-Substrate Analogue for Glycogen PhosphorylaseResults and Discussion96IntroductionApophosphorylase b can be efficiently reconstituted with pyridoxa1-5'-pyrophospho-a-D-glucose (PLPP-Glu), a cofactor-substrate analogue in which PLP andaG 1P are covalently linked through a pyrophosphate bond (Takagi et al., 1982; Withers etal., 1981a; Tagaya and Fukui, 1984). In the presence of glycogen, PLPP-Glu-phosphorylase quickly transfers the glucosyl moiety to the acceptor oligosaccharide, thusforming a new a-1,4-glucosidic linkage along with the production of the catalyticallyinactive cofactor, pyridoxa1-5s-pyrophosphate (PLPP), which remains bound at the enzymeactive site (Scheme 4-la). Previous studies have clearly demonstrated the mechanisticsimilarities of this glucosyl transfer reaction with that catalyzed by the native enzyme(Tagaya and Fukui, 1984). Cofactor-substrate analogues of this type, wherein thesubstrate is covalently linked to the PLP cofactor, are an attractive means of getting a fullequivalent of aG1P to the active site of the enzyme for mechanistic or structural studieswithout having to present the enzyme with a saturating concentration of substrate.However, it is well known that PLPP-Glu-phosphorylase b, in the absence of the secondsubstrate (glycogen), spontaneously breaks down to form native PLP-enzyme and glucose-1,2-cyclic phosphate (Scheme 4-1b), a process thought to involve deprotonation of thehydroxyl group at the C-2 position by a nearby active site amino acid residue (Withers etal., 1981a; Horinishi et al., 1988). The hydroxyl group at C-2, once deprotonated, isthought to break the pyrophosphate linkage by attacking the glucosyl phosphate moiety(Horinishi et al., 1988). Horinishi and co-workers (1988), have implicated Tyr-573 as apossible candidate for the deprotonation event at C-2, a suggestion based on X-raystructural data in addition to the pH-dependence of reactivation. To date, this reactivationprocess has precluded X-ray crystallographic investigation of PLPP-Glu-phosphorylasesince the sample decomposes before crystallization and data collection.97a) Normal Reaction in the Presence of Glycogenb) Reactivation in the Absence of GlycogenScheme 4-1. Two alternate reactions for PLPP-Glu-phosphorylase b.98The aim of this study was to synthesize an analogue of PLPP-Glu which will notundergo the reactivation and which will also not undergo the glucosyl transfer reaction at asignificant rate. To this end, the 2-deoxy-2-fluoro analogue of PLPP-Glu wassynthesized, reconstituted into glycogen phosphorylase, and tested for both glucosyltransfer and reactivation in a series of kinetic, 19F NMR, and cofactor abstraction studies.Future X-ray structural studies of the reconstituted enzyme derivative should providevaluable information about the R-state enzyme conformation, and in addition, may identifyan active site nucleophile for the proposed double displacement enzyme mechanism.Results 1. Synthesis The synthesis of PLPP-Glu and pyridoxa1-5'-pyrophospho-1-(2-deoxy-2-fluoro)-a-D-glucose (PLPP-2FG1u) was carried out using the same basic strategy as outlined forthe previous synthesis of PLPP-Glu (Shimomura and Fukui, 1978; Takagi et al., 1982)(Scheme 4-2). In each case, PLP was converted to PLP-diphenyl pyrophosphate withdiphenyl phosphochloridate under basic conditions. This was then reacted with theappropriate phosphorylated glucose moiety. Purification of the pyrophosphate productswas achieved using ion-exchange chromatography (AG-1X8, Cl - form), to yield PLPP-Glu and PLPP-2FG1u in an overall yield of99Scheme 4-2. The synthetic route for the preparation of PLPP-Glu and PLPP-2FG1u.1 0 02. Rates of Reactivation of Glycogen Phosphorylase b Reconstituted with FLPP-Glu and PLPP-2FGlu Glycogen phosphorylase b was reconstituted with PLPP-Glu and PLPP-2FG1u inplace of the native cofactor (PLP) by incubating separate samples of the apoenzyme withapproximately one equivalent of each cofactor-substrate analogue. Evidence for thesuccessful reconstitution of the enzyme with each cofactor analogue came from routineU.V. / Vis. studies which identify the formation of the imine linkage between the cofactorand Lys-679 (for details refer to the Materials and Methods section). The reconstitutedenzymes were then incubated at room temperature, at the indicated pH, in the presence orabsence of glycogen and reactivation was monitored over a period of five days bymeasuring the increase in enzymic activity due to the formation of native PLP-enzyme. Theactivity was measured in the direction of glycogen synthesis, under standard assayconditions (Engers et al., 1970a, b). The time courses for reactivation of the tworeconstitued enzymes under various conditions of pH and effector concentration are shownin Figure 4-1.As shown in Figure 4-1, all reconstituted enzyme derivatives initially showed verylow levels of enzyme activity, this activity presumably being due to trace amounts of PLPpresent in the PLPP-Glu and PLPP-2FGlu preparations. However, when PLPP-Glu-phosphorylase was incubated at pH 6.8 in the absence of glycogen, the enzyme activityincreased significantly with an observed first older rate constant of approximately 0.02 ±0.01 hr 1 (Figure 4-la). Furthermore, the reactivation process could be accelerated at pH7.9 (kobs = 0.04 ± 0.02 hr 1 ) (Figure 4-1b), in full agreement with earlier reactivationstudies on the same enzyme derivative (Withers et al., 1981a; Horinishi et al., 1988).1010^20^40^60^80 100 120 140Time (hrs)0^20^40^60^80^100 120 140Time (hrs)Figure 4-1. Time course for reactivation of PLPP-Glu- and PLPP-2FG1u-phosphorylase b.Activity measurements were conducted at pH 6.8 and 30 °C in 100 mM KC1, 50 mM triethanolaminehydrochloride, 1 mM EDTA and 1 mM DTT. The reaction mixtures each contained 18 mM aG1P, 1mMAMP, 1% glycogen, enzyme (30 gg m1.7 1 ), and reaction times were 5 minutes. a) Reactivation timecourses at pH 6.8. b) Reactivation time courses at pH 7.9. Samples contained: (0) PLPP-Glu-enzyme;(A) PLPP-Glu-enzyme plus glycogen; (U) PLPP-2FG1u-enzyme; (A) PLPP-2FG1u-enzyme plus glycogen.The enzyme and glycogen concentrations in each incubated sample were 15.2 gM and 0.24%, respectively.Apoenzyme activity was 0.5 p.mol min -1 mg-1 . The activity of control samples of reconstituted PLP-enzyme remained essentially unchanged over the span of the experiment (= 50 gmol min-1 mg-1 ).102In the presence of glycogen, enzyme activity from the PLPP-Glu samples, regardless ofpH, remained essentially unchanged throughout the experiment. This is the expected resultif, in the presence of glycogen, the rate of glucosyl transfer to polysaccharide greatlyexceeds the rate of the reactivation reaction since the resultant PLPP-enzyme cannotreactivate. Glycogen phosphorylase reconstituted with PLPP-2FGlu, however, remainedessentially inactive regardless of the solution pH or the presence or absence of glycogen.This was expected since the fluorine at the glucosyl C-2 position cannot engage in anintramolecular attack on the phosphate moiety releasing PLP. Thus, glycogenphosphorylase b reconstituted with PLPP-2FGlu should be sufficiently stable forcrystallization studies.3. 19F NMR and Cofactor Abstraction Studies of Glycogen Phosphorylase b reconstituted with PLPP-2FG1u 3.1. 19F NMR Experiments19F NMR studies of glycogen phosphorylase b reconstituted with PLPP-2FG1uwere carried out in an effort to prove the presence of the cofactor, to monitor possiblechanges in the fluorine environment upon activation of the enzyme and possibly to detectwhether glycosyl transfer occurs in the presence of glycogen. The latter point could also beconfirmed by removal of the cofactor and subsequent identification. 19F NMR spectra ofglycogen phosphorylase b reconstituted with PLPP-2FG1u were recorded at roomtemperature, in a 5 mm NMR tube at high enzyme concentration (0.7 - 1.0 mM). The 19FNMR spectra of PLPP-2FG1u-phosphorylase in the presence of various effectors andsubstrates are shown below in Figure 4-2.103c)qtrLIrAf■-"vA"'AC)-110^-120^-130^-140PP=a)VVIAI\VJ^AN\tvIVAIVFigure 4-2. 19F NMR spectra of glycogen phosphorylase b reconstituted with PLPP-2FG1u.The NMR sample contained 55 - 60% D20, 50 mM triethanolamine hydrochloride, 100 mM KC1, 1 mMEDTA, 1 mM DTT, pH 6.8. a) Signal averaged over 20954 scans, sample contained PLPP-2FG1u-phosphorylase b (1.03 mM). b) Signal averaged over 23000 scans, sample contained PLPP-2FG1u-phosphorylase b (0.8 mM), AMP (2.4 mM). c) Signal averaged over 37200 scans, sample containedPLPP-2FG1u-phosphorylase b (0.73 mM), AMP (2.2 mM), glycogen (0.75%). Chemical shifts arereferenced to TFA, and 2-fluoro-glucal (-92.2 ppm) was used as an internal standard (0.3 - 0.8 mM) in eachexperimentThe 19F NMR spectrum of glycogen phosphorylase b reconstituted with PLPP-2FGIu (Figure 4-2a) shows a broad resonance (tIv1/2 = 480 Hz) located at -121.8 ppm, ingood agreement with the chemical shift observed for the cofactor free in solution. Further,the linewidth observed for the fluorine nucleus within PLPP-2FGIu-phosphorylase isconsistent with that observed previously for fluorinated analogues of PLP immobilized atthe active site of the enzyme (Chang et al., 1986). Upon addition of activating AMP104(Figure 4-2b), the 19F resonance for PLPP-2FG1u shifts slightly downfield (8 = -121.2ppm) with a moderate reduction in linewidth (AD112 = 450 Hz). The source of thenarrowing for the nucleotide activated signal is not fully understood, although it could bedue to a change in the local mobility at the C-2 position of the sugar moiety upon AMPaddition. Further, it cannot be ruled out that the population of PLPP-2FG1u-enzymespecies becomes increasingly homogeneous with AMP activation. With the addition of thesecond substrate, glycogen (Figure 4-2c), the 19F signal shifted upfield to approximately -122.0 ppm with a significant increase in linewidth (AD112 = 510 Hz). The signal-broadening observed upon glycogen addition is not unexpected given the expected increasein correlation time for the enzyme-glycogen complex, in addition to any exchange-broadening between the free and bound forms of the enzyme.The absence of any large chemical shift changes suggests that only small structuralchanges occur at the C-2 position of the glucose binding site upon nucleotide activation andglycogen addition. This is consistent with earlier studies (Street et al., 1989) whichsuggest that the glucose binding pocket in the T-state enzyme remains essentially intactduring the T- to R-state transition. The 19F NMR results also suggest that no glucosyltransfer occurs from phosphorylase upon glycogen addition since such turnover wouldvery likely result in considerable narrowing of the 19F signal upon transfer of the 2-fluoro-glucosyl moiety from its immobilized position at the active site of the enzyme, to glycogen,where the local mobility of the sugar residue would likely be greater. However, in order toconfirm the suggestion that turnover in PLPP-2FG1u-phosphorylase is considerablyretarded, the pyridoxal cofactor was removed from the sample of PLPP-2FGlu-phosphorylase b used in the NMR experiment and identified as described below.3.2. Cofactor Abstraction StudyThe analysis of the pyridoxal compounds bound to PLPP-2FGlu-phosphorylase bwas carried out essentially according to Takagi et al., (1982) and Tagaya and Fukui,105(1984). Thus, the NMR sample of PLPP-2FG1u-phosphorylase (see above, sample (c)),in the presence of AMP and glycogen was treated with trichloroacetic acid (TCA) toprecipitate the protein and release all protein bound cofactors. The pyridoxal compoundswere then identified using ion-exchange chromatography, and the elution profiles arepresented in Figure 4-3.Figure 4-3. Analysis of pyridoxal compounds from PLPP-2FG1u-phosphorylase b.The elution profiles represent two standard runs in addition to that for the pyridoxal compounds extractedfrom the NMR sample of PLPP-2FGlu-enzyme. PLP was added as an internal standard in each case. (A)PLPP-2FG1u standard solution; (^) PLPP-2FG1u plus PLPP standard solution; (•) PLPP-2FG1u-phosphorylase b.PLP was included as an internal standard in each of the samples examined. As canbe seen in the elution profiles shown in Figure 4-3, PLP elutes early in the profile and iscentered about fraction 15. Three peaks are seen in the profile for a standard solutioncontaining PLP, PLPP-2FG1u and PLPP (0). PLP eluted early and the two diphosphocofactors eluted late, between tubes 50 and 68. Earlier studies with PLPP-Glu and PLPPhave demonstrated that PLPP elutes at a higher salt concentration than does PLPP-Glu106(Takagi et al., 1982), as expected since PLPP possesses one extra ionizable oxygen atom.This order of elution was confirmed in the present study with a standard solution of PLPP-2FG1u (♦ ), which was observed to elute between fractions 50 and 60. However, asevidenced from the small shoulder peak observed in the elution profile for the standardsample of PLPP-2FG1u, this sample also appears to contain a small amount of PLPP,probably derived from the hydrolysis of the parent molecule during the TCA work-up ofthe sample. The elution profile for the cofactors extracted from PLPP-2FG1u-phosphorylase b (II) shows that the major pyridoxal compound bound to the enzyme, evenafter several weeks at 4 °C, is PLPP-2FG1u. Again, small quantities of the PLPP cofactormay also be present in the sample, the most likely origin being hydrolysis during samplepreparation, though some enzymic turnover cannot be ruled out. Nonetheless, these resultsindeed show that PLPP-2FG1u-phosphorylase b is quite resistant to turnover, unlikePLPP-Glu-phosphorylase which quickly Ow --=, 10 sec.) transfers its glucosyl moiety toglycogen (Tagaya and Fukui, 1984).4. Structural Analysis of PLPP -2FG1u -Phosphorylase b In collaboration with Dr. N. B. Madsen (University of Alberta) and Dr. S. R.Sprang (University of Texas), X-ray structural studies of crystalline PLPP-2FGlu-phosphorylase b have been possible and are currently ongoing. At the University ofAlberta, crystals of PLPP-2FG1u-phosphorylase b of sufficient size and diffraction qualitywere grown from Zeppezauer tubes by vapor diffusion using polyethylene glycol asdescribed elsewhere (Sprang et al., 1991). X-ray crystallographic analysis of thereconstituted enzyme is presently being carried out at the University of Texas. Initialstructural studies of the enzyme grown in the presence of oligosaccharide have indicatedthat after several weeks at ambient temperature, only PLPP-phosphorylase b, the productof turnover, is observed (S. R. Sprang, personal communication). Apparently, after anextended period of time, crystalline PLPP-2FG1u-phosphorylase b grown in the presence107of oligosaccharide acceptor, indeed transfers the 2-deoxy-2-fluoro-glucosyl moiety to theoligosaccharide acceptor.5. Discussion While earlier studies have clearly shown that glycogen phosphorylase breconstituted with PLPP-Glu transfers the enzyme bound glucosyl moiety to glycogenforming a new a-1,4-glucosidic linkage, unfortunately the reactivation process has limitedits use in X-ray crystallographic studies. In this study, PLPP-2FG1u was synthesized andreconstituted into glycogen phosphorylase b, producing an enzyme derivative completelyresistant to reactivation, even at elevated pH. These results are in full agreement with theproposed mechanism for PLPP-Glu reactivation (Horinishi et al., 1988). By substitutingthe glucosyl C-2 hydroxyl group for a fluorine nucleus the reactivation process is preventedsince there is no longer a nucleophile at C-2 to cleave the pyrophosphate bond linking thecofactor and substrate moiety. Similar supportive evidence for the reactivation mechanismhas been reported previously in a study that separately reconstituted glycogenphosphorylase b with PLPP-Glu and pyridoxal-5'-pyrophospho-a-D-mannose, an isomerof PLPP-Glu in which the C-2 hydroxyl moiety is axial rather than equatorial thus cannotundergo the reactivation process (Horinishi et al., 1988). This study compared thereactivation rate for the two enzyme species and found that enzyme reconstituted with themannose derivative, just like PLPP-2FG1u-phosphorylase, was indeed completely resistantto reactivation.In the presence of an acceptor oligosaccharide (e.g. glycogen), PLPP-Glu-phosphorylase b transfers its glucosyl moiety to the acceptor, PLPP-2FG1u-enzyme washowever shown to be essentially inactive under these conditions. 19F NMR studies withPLPP-2FG1u-phosphorylase b suggest that the 2-fluoro glucosyl moiety remainsimmobilized at the active site of the enzyme throughout the addition of AMP and glycogen.Any turnover would transfer the 2-fluoro glucosyl moiety to glycogen, most likely108resulting in a considerable increase in local mobility, which would be reflected in anarrowing of the signal linewidth and possibly a change in chemical shift. Since only asmall variation in the 19F signal was observed throughout the study, the cofactor-substrateanalogue is thought to remain intact, even in the presence of glycogen. This was laterconfirmed in a cofactor extraction study, in which it was shown that PLPP-2FG1u was theonly cofactor bound to the enzyme after incubation with AMP and glycogen. The fact thatonly moderate changes to the 19F signal were observed upon AMP and glycogen additionalso supports recent kinetic and structural studies suggesting that the glucosyl binding siteremains essentially intact throughout the allosteric transition (Street et al., 1989).It is not unreasonable that glycogen phosphorylase b reconstituted with PLPP-2FG1u is essentially inactive, given what is known to occur upon substitution of the C-2hydroxyl group for a fluorine atom in aG1P. The substrate analogue 2-deoxy-2-fluoro-aG1P has a Vmax value in the normal catalytic reaction about five orders of magnitudelower than that of the parent substrate (aG1P) (Street et al., 1989). The large ratereduction is thought to arise from a combination of electronic and binding effects whichserve to destabilize an already electron-deficient oxocarbonium ion-like transition state(Street et al., 1989). Since PLPP-Glu is thought to mimic the native enzyme reaction(Tagaya and Fukui, 1984; Withers et al., 1981a), it is not unreasonable to suggest thatthese same electronic and binding effects would considerably reduce the turnover rate ofPLPP-2FGlu-phosphorylase b.Preliminary structural studies have been carried out on glycogen phosphorylase breconstituted with PLPP-2FG1u, in collaboration with Professor N. B. Madsen at theUniversity of Alberta and Professor S. R. Sprang at the University of Texas. Initial resultssuggest that protein crystals of sufficient size and diffraction quality can be prepared understandard conditions (N. B. Madsen, personal communication). Furthermore, initialstuctural results suggest that the enzyme crystallizes as an R-state tetramer withoutoligosaccharide bound at the enzyme active site. Interestingly however, in the crystalline109state, it appears that PLPP-2FGlu-phosphorylase b is somewhat active since pyridoxa1-5'-pyrophosphate (PLPP) was the only cofactor observed at the active site of the enzyme afterseveral weeks of standing at ambient temperature. Since the results from solution studiessuggest that PLPP-2FG1u-phosphorylase is essentially inactive, or at least extremely slowto turnover, while the early structural studies suggest that turnover in the crystalline stateindeed occurs, it might be possible to acquire X-ray crystallographic data before, duringand after glucosyl transfer, thereby gathering structural information throughout the transferprocess. Presently, structural studies are ongoing with crystals of PLPP-2FG1u-phosphorylase b grown in the absence of oligosaccharide, such that structural informationwith the cofactor intact can be obtained. It should also be feasible to investigate crystalswhich have noi been stored for any significant period of time.In the presence of AMP, native glycogen phosphorylase b exists as tetramers.However, upon addition of glycogen the tetramers dissociate into the fully activated R-statedimers found in vivo . As yet, only the tetrameric form of the crystalline R-state enzymehas been prepared (Barford and Johnson, 1989; Sprang et al., 1991), thus structuralinformation concerning the fully active dimer conformation with oligosaccharide bound atthe active site is highly desirable. Initial structural analysis of PLPP-2FG1u-phosphorylaseb suggests that this enzyme crystallizes as the R-state tetramer and not the fully activedimer. Future structural studies, however, may well provide new information concerningthe interactions between the R-state enzyme and the aG1P portion of the cofactor-substrateanalogue. This may allow identification of an active site nucleophile.110CHAPTER 5Binding Interactions and Substrate Activity of 4-Fluoro-GlycogenResults and Discussion111Introduction "Incompetent" substrate analogues have proved valuable in studying enzymemechanisms, and this is especially true with glycogen phosphorylase since this enzymerequires a ternary complex, wherein all substrates are bound, before the enzyme canassume a catalytically active conformation. Previous kinetic and NMR studies with"incompetent" analogues of aG1P, such as aG1CP (Withers and Street, Unpublishedresults) and glucose-1,2-cyclic phosphate (Withers et al., 1981b) have provided usefulinformation concerning the activated conformation since these analogues can bind toglycogen phosphorylase and produce a ternary enzyme complex, yet are unable to react.While cyclodextrin, a cyclic oligosaccharide, has been shown to mimic oligosaccharidebinding in potato phosphorylase, and indeed, has been used to provide evidence for aglucosyl-enzyme intermediate in that enzyme (Kokesh and Kakuda, 1977), no goodanalogues of glycogen have been available for testing with glycogen phosphorylase. Ananalogue such as 4-deoxy-4-fluoro-glycogen (4-F-glycogen), in which all terminal sugarspossess a fluorine at the 4-position, would be a useful analogue since it should bind toglycogen phosphorylase, allow the formation of a ternary complex, yet be unable to act as aglucosyl acceptor since the hydroxyl group at the 4-position (the nucleophile in the normalreaction) has been replaced by a fluorine (Scheme 5-1). Thus, 4-F-glycogen should be avaluable analogue in the search for the putative glucosyl-enzyme intermediate in glycogenphosphorylase since it would promote the initial bond breaking and bond making catalyticsteps, yet be inert as a substrate in the direction of glycogen synthesis. In addition, since4-F-glycogen possesses an NMR active nucleus ( 19F), 19F NMR studies could provideuseful insight into the binding interaction between glycogen and phosphorylase.112Scheme 5-1. 4-F-Glycogen as an "incompetent" substrate analogue.4-Deoxy-4-fluoro-glycogen has been enzymatically synthesized previously usingglycogen phosphorylase and the substrate analogue 4-deoxy-4-fluoro-aG1P (Withers,1990). To synthesize 4-F-glycogen the phosphorylase reaction is run in the direction ofglycogen synthesis and the "capped" 4-deoxy-4-fluoro-glucosyl residues are transferredfrom phosphate to the 4-hydroxyl moiety of normal glycogen, just as in the normalreaction. However, once transferred to the glycogen termini, 4-deoxy-4-fluoro-glucosylresidues are no longer capable of accepting another glucosyl residue since the fluorine at the4-position cannot act as a nucleophile. While the equilibrium constant for glycogenformation is only 3.5, the reaction can be brought essentially to completion by repeatedremoval of the product phosphate by dialysis and subsequent re-addition of the 4-deoxy-4-fluoro-aG1P substrate.Interestingly, 4-F-glycogen binds to phosphorylase some 100 fold more tightlythan does normal glycogen (Withers, 1990) though the basis of this improved binding isnot yet understood. In addition, 4-F-glycogen has been shown (Withers, 1990) to beessentially inactive as a substrate for glycogen phosphorylase in the direction of glycogensynthesis (i.e. in the presence of aG1P) (Withers, 1990), as expected. In this study, the113binding interaction between 4-F-glycogen and phosphorylase was investigated by 19FNMR spectroscopy and an estimate of the stoichiometry of binding was determined. Inaddition, 4-F-glycogen was tested as an "incompetent" dead-end substrate analogue inexperiments designed to trap or provide evidence for a glucosyl-enzyme intermediate inglycogen phosphorylase.Results 1. 4-F-Glycogen Binding Study 1.1. Titration of 4-F-Glycogen with Glycogen Phosphorylase bThe presence of an NMR active nucleus ( 19F) in 4-F-glycogen has been used tomonitor the binding interaction between 4-F-glycogen and glycogen phosphorylase b, thusallowing estimates to be made of the stoichiometry of binding and providing insights intothe mode of binding. 4-F-Glycogen was enzymatically synthesized according to Withers,(1990) by incubating glycogen phosphorylase b with normal glycogen and 4-deoxy-4-fluoro-aG1P. The titration of 4-F-glycogen with glycogen phosphorylase b was achievedby sequential addition of aliquots of enzyme to a sample of the modified glycogen in a 5mm NMR tube and acquisition of 19F NMR spectra after each addition. The experimentwas performed in the presence of AMP such that 4-F-glycogen bound to activated R-stateenzyme. However it is important to note that the enzyme does not catalyze any reactionunder these conditions since the second substrate (aG1P or phosphate) is not present.This titration was performed at two different constant concentrations of 4-F-glycogen(0.6% and 1.2%) and representative spectra from one such titration are shown in Figure 5-1.114ba-120.00^-12:1 00PPM .128.00C-118.00dFigure 5-1. 19F NMR titration of 4-F-glycogen with glycogen phosphorylase b.The NMR sample contained 75% D20, 50 mM triethanolamine hydrochloride, 100 mM KC1, 1 mMEDTA, 1 mM DTI', 1.2 % 4-F-glycogen, 2.1 mM AMP (initially), 2.3 mM 2-fluoro-D-glucal (initially),pH 6.8. Only 5 of the original 10 spectra are shown and these contained the following concentrations ofglycogen phosphorylase; a) 0 mM; b) 0.054 mM; c) 0.10 mM; d) 0.15 mM; and e) 0.22 mM. 2-Fluoro-D-glucal was used as an internal standard in all experiments.115As shown in Figure 5-1, no significant change in the chemical shift (8 = -121.6ppm) was observed as the enzyme concentration was increased, nor did any newresonances appear. Further, the total peak area, as determined by integration, remainedapproximately constant throughout the series. There was, however, a progressive decreasein the peak height with increasing enzyme concentration up to a limiting point beyondwhich no further decrease was observed. Some broadening of the line was also observedas the enzyme concentration increased, linewidths changing from Av1/2 = 88 Hz in theabsence of enzyme to Av12 = 135 Hz in the presence of saturating concentrations. Theseobservations are consistent with progressive binding of the enzyme to 4-F-glycogenparticles until the surface of the particles is completely covered, when no further bindingcan occur.Since the increase in linewidth upon enzyme binding would appear to indicate thatthe spin-spin relaxation time (T2) for this resonance was decreased by the binding ofphosphorylase (though this could also be due to chemical shift heterogeneity), the spin-lattice relaxation time (T1) of the resonance observed in the presence of saturating enzymewas also determined in order to assess possible changes in mobility of the fluorosugarsgiving rise to this resonance. A value of T1 = 0.24 s was measured for this signal from asample of 4-F-glycogen (1.2%) containing sufficient phosphorylase (0.22 mM) tocompletely coat the glycogen particle. This value is slightly smaller than that previouslyobserved for free 4-F-glycogen (T1 = 0.36 s) (Withers, 1990).As noted previously, the peak heights of the 19F resonance for 4-F-glycogen wereobserved to decrease progressively with increasing concentration of glycogenphosphorylase up to a point beyond which no further decrease was observed. Figure 5-2ashows a plot of peak height versus enzyme concentration (expressed as monomers ofmolecular weight 97,400) using a 4-F-glycogen concentration of 1.2%. Peak heightclearly decreases monotonically down to a phosphorylase concentration of 0.15 mM withno further decrease beyond that point. Such behaviour is suggestive of very tight binding116of 4-F-glycogen to phosphorylase (KD for 4-F-glycogen binding to phosphorylase«1.2%) , a result which is consistent with the tight binding observed in previous kineticstudies (Withers, 1990). A similar set of NMR spectra was accumulated using a 4-F-glycogen concentration of 0.6%, and the peak heights for these 19F signals are shown inFigure 5-2b. In this case the peak height decreased to a limiting phosphorylaseconcentration of approximately 0.07 mM, with no further change beyond that point.9080706050400.05^0.1^0.15^0.2[Glycogen Phosphorylase) mM60SO-.40—.30— 20 ^ 1^1^1^1^1^1^1^1^1^1^1 1^1^1^1^10^0.05^0.1^0.15[Glycogen Phosphorylase] mMFigure 5-2. Plot of peak heights of 19F NMR signals versus enzyme concentration.Conditions for the NMR experiments are described in Figure 5-1 and also in the Experimental section. a)1.2% 4-F-glycogen, 2.1 mM AMP. b) 0.6% 4-F-glycogen, 2.1 mM AMP.117The results presented here provide insight into the solution binding interactionbetween glycogen phosphorylase and 4-F-glycogen (as will be discussed later). Moreover,the 19F NMR titration of 4-F-glycogen with phosphorylase has provided a direct methodfor estimating the number of enzyme molecules bound to one glycogen particle.2. 4-F-Glycogen as an "Incompetent" Dead-End Substrate Analogue? 2.1. 31 P NMR of PLPP-Glu-Phosphorylase b in the Presence of 4-F-GlycogenAs previously outlined, glycogen phosphorylase requires a ternary complex beforecatalytic steps of bond making and bond breaking can occur. Thus, any attempt toaccumulate and trap an enzyme intermediate in glycogen phosphorylase requires an"incompetent" glycogen analogue which activates the enzyme, yet prevents turnover. Anideal enzyme system to observe catalytic steps of bond cleavage within the enzyme ternarycomplex might be glycogen phosphorylase reconstituted with pyridoxa1-5'-diphospho-a-D-glucose (PLPP-Glu). It is well known that PLPP-Glu-phosphorylase mimics the nativeenzyme reaction by catalyzing the glucosyl transfer from the enzyme-bound PLPP-Glucofactor to glycogen with retention of anomeric configuration (Takagi et al., 1982; Tagayaand Fukui, 1984) (for a review please refer to Chapter 4 Introduction). However, unlikethe reaction with aG1P, the glucosyl-transfer from PLPP-Glu-phosphorylase can beconveniently followed by 31P NMR spectroscopy since the PLPP-enzyme (the product ofglucosyl transfer from PLPP-Glu) has a significantly different spectrum from that ofPLPP-Glu-phosphorylase (Withers et al., 1981a). Since 4-F-glycogen is expected toactivate the enzyme and promote the initial steps of bond cleavage, yet be inert to glucosyltransfer, PLPP-Glu-phosphorylase should serve as a good system in which to test theability of 4-F-glycogen to activate the enzyme and promote PLPP-Glu bond cleavage,possibly forming the putative glucosyl-enzyme intermediate.118Thus, glycogen phosphorylase b was reconstituted with PLPP-Glu and the 31 PNMR spectra of this enzyme derivative in the absence and presence of 4-F-glycogen wererecorded at 30 °C in a 10 mm NMR tube at high enzyme concentration (0.8 - 0.9 mM).Adenosine-5'-thiomonophosphate (AMPS) was used as a nucleotide activator in place ofAMP since its resonance is located well downfield of the signals of interest. The 31 P NMRspectra of PLPP-Glu-phosphorylase b obtained in this way are presented in Figure 5-3.The 31 P NMR spectrum for PLPP-Glu-phosphorylase b in the presence ofactivating AMPS (Figure 5-3a) was completely consistent with that observed previously forthe same enzyme system (Withers et al. 1981a). The resonance observed at 44.5 ppm isthat of free AMPS and is present as an exchange-averaged signal (Au in = 38 Hz) withenzyme-bound AMPS, appearing at 41.0 ppm (O1)1/2 = 110 Hz). The well resolvedsignals representing free and bound AMPS reflect an exchange process occurring in theslow exchange regime. The broad signal centered at 8 - 11 ppm (Av in = 210 Hz)represents the pyrophosphate moiety within PLPP-Glu bound at the active site of theenzyme. In solution the PLPP-Glu phosphate moieties proximal to pyridoxal and glucoseare observed at -11.0 and -12.6 ppm respectively, whereas bound at the active site of theenzyme the two signals merge to form an asymmetric broad resonance as observed here.As shown in Figure 5-3b, after one hour of incubating 4-F-glycogen with PLPP-Glu-phosphorylase the 31 P NMR spectrum had considerably changed to that of PLPP-phosphorylase b. A new resonance was observed at -5.0 ppm (6,1)112 = 125 Hz) which, onthe basis of previous studies with PLPP-phosphorylase (Withers et al., 1981a), can beassigned to the terminal phosphate moiety of the PLPP-enzyme. The other upfield signal,now located at -10.5 ppm (O1)1/2 = 210 Hz), has shifted slightly downfield from itsoriginal position as a result of the glucosyl transfer.119abSO^40^s0^so^10^0^-10PPmFigure 5-3. 31P NMR spectra of glycogen phosphorylase b reconstituted with PLPP-Glu.NMR samples contained 60% D20, 50 mM triethanolarnine hydrochloride, 100 mM KCI, 1 mM EDTA, 1mM DTT, pH 6.8, and spectra were recorded at 30 °C. a) Signal averaged over 8400 transients, reactionmixture contained PLPP-Glu-phosphorylase b (0.89 mM), AMPS (1.9 mM). b) Signal averaged over 2400transients (first hour of acquisition only), reaction contained PLPP-Glu-phosphorylase b (0.84 mM),AMPS (1.76 mM), 4-F-glycogen (0.59%). c) Signal averaged over 2400 transients (second hour ofacquisition only), reaction mixture was the same as listed for (b). d) Signal averaged over 26200transients, reaction mixture was same as for (b).120The 31P NMR spectrum shown in Figure 5-3c is that of the same sample after a secondhour of incubation and acquisition time. In this spectrum, the signal representing thephosphate moiety proximal to pyridoxal has shifted slightly downfield to -10.3 ppm (Avi/2= 110 Hz) while the terminal phosphate moiety is still observed at -5.0 ppm (Av1/2 = 90Hz). Moreover, both resonances have undergone a considerable reduction in signallinewidth which is especially clear in the resonance at -10.3 ppm and is consistent with anincrease in the homogeneity of the NMR sample (only PLPP-enzyme present). Thus, itappears that essentially complete conversion of PLPP-Glu to PLPP occurred within thefirst hour of incubation. Finally, after the second hour of data acquisition was completedspectral data were accumulated over an extended period and the resulting spectrum isshown in Figure 5-3d. The signals observed at -5.0 ppm (Av1/2 = 107 Hz) and -10.3 ppm(Au1/2 = 130 Hz) are completely consistent with earlier 31 P NMR studies of glycogenphosphorylase reconstituted with PLPP (Withers et al., 1981a). The broad resonanceobserved at 8 3 ppm results from a small quantity of contaminating native enzyme.These results clearly show that upon Addition of 4-F-glycogen, cleavage of theglucosyl moiety from PLPP-Glu-phosphorylase indeed occurs. While this would suggestthat the glucosyl moiety has been transferred to the enzyme as the putative covalentintermediate, it cannot be ruled out that transfer to a small number of free 4-hydroxylmoieties within the 4-F-glycogen derivative has occurred. It is also possible that anyglucosyl-enzyme intermediate formed in the presence of 4-F-glycogen is susceptible tohydrolysis. Since this 31P NMR experiment only monitors the cleavage of the glucosyl-phosphate bond within PLPP-Glu, the results shown here do not distinguish between thepossible fates of the glucosyl moiety outlined above.1212.2. Mass Spectral Studies of PLPP-Glu-Phosphorylase b in the Presenceof 4-F-GlycogenThe 31P NMR results outlined above show that in the presence of 4-F-glycogen theglucosyl moiety is cleaved from PLPP-Glu-phosphorylase, and further, that it may form acovalent glucosyl-enzyme intermediate. To investigate this possibility ion-spray massspectrometry in the liquid chromatography-mass spectrometric (LC-MS) mode wasemployed with technical assistance from Dr. J. Gebler and Dr. D. Hess in Dr. R.Aebersold's laboratory in an attempt to identify glycogen phosphorylase enzyme modifiedwith the glucosyl moiety. This technique has been used previously in this laboratory toidentify covalent glucosyl-enzyme intermediates in other glucosyl transferring enzymesknown to proceed through a covalent intermediate (Withers, Unpublished results). Theresults from the mass spectral analysis of PLPP-Glu-phosphorylase b in the presence ofdifferent concentrations of 4-F-glycogen, in addition to the mass analysis of the nativeenzyme and the apoenzyme are summarized below in Table 5-1.As shown in Table 5-1, the mass values determined for all enzyme species wereapproximately 97,200 Da. The reported mass for the monomeric form of glycogenphosphorylase b (Johnson, 1992), based on its amino acid composition is 97,444 Da.Since the masses of the native enzyme and the apoenzyme were found to be approximatelyequal, it appears that the cofactor is removed under the acidic conditions required for theanalysis. Further support for this suggestion comes from the untreated sample of PLPP-Glu-phosphorylase b which was also found to have approximately the same mass as eitherthe native or apophosphorylase b enzymes. The preparation of enzyme samples for massanalysis involves liquid chromatography (LC) containing trifluoroacetic acid (0.05%) suchthat the proteins are in a protonated state when they enter the mass spectrometer. It is theseconditions which are most likely responsible for the removal of the cofactor, and as willbecome apparent, this in fact simplifies the analysis.122Table 5-1. Mass spectral data collected for glycogen phosphorylase b.aEnzyme MassNative phosphorylase 97,206Apophosphorylase 97,221PLPP-Glu-phosphorylase(untreated) 97,196PLPP-GIu-phosphorylase(low [4-F-glycogeri])b 97,200PLPP-Glu-phosphorylase(high [4-F-glycogen])c 97,202a The apoenzyme and PLPP-Glu-phosphorylase were prepared as previously described. The protocol forsample preparation is outlined in the Materials and Methods section. The error within each massdetermination for all experiments was 5 ± 10 Da.PLPP-Glu-phosphorylase (67 gM) was preincubated with AMP (1.7 mM) and different concentrations of 4-F-glycogen at room temperature and pH 6.8 in the standard triethanolamine buffer system. b Reactionincluded (0.07%) 4-F-Glycogen. The same mass was observed after 5 minutes and 45 minutes ofpreincubation with 4-F-glycogen. c Reaction included (0.35%) 4-F-glycogen. The incubation time was 90minutes.The results from the mass spectral analysis of PLPP-Glu-phosphorylase b,preincubated with two different concentrations of 4-F-glycogen, indicate that no significantconcentration of covalently derivatized enzyme species is formed even after an extendedincubation period. Thus, if indeed a glucosyl-enzyme intermediate is formed in thepresence of 4-F-glycogen, it is either transferred to free non-fluorinated terminal sugarresidues within the 4-F-glycogen derivative, or alternatively, is hydrolyzed from theenzyme under the conditions of the experiment.1232.3. Transfer of Radiolabelled Glucose from PLPP-GIu-Phosphorylase bto 4-F-GlycogenThe previous mass spectral study suggests that the ultimate fate of the glucosylmoiety within PLPP-Glu-phosphorylase b, upon incubation with 4-F-glycogen, is not aglucosyl-enzyme intermediate, but rather, that it is hydrolyzed from the enzyme ortransferred to chain ends within 4-F-glycogen which are not capped with 4-deoxy-4-fluoro-glucose residues. In an attempt to determine the fate of the glucosyl moiety within PLPP-Glu-phosphorylase b under these conditions, a radiolabelled analogue in which the sugarwas radioactive was prepared and reconstituted into glycogen phosphorylase. Previousstudies have used this technique to study the reaction of PLPP-Glu-phosphorylase b in thepresence of normal glycogen (Takagi et al., 1982; Tagaya and Fukui, 1984).The synthesis of radiolabelled PLPP-[ 14C]-Glu was carried out as previouslydescribed by Takagi et al., (1982). Glycogen phosphorylase b was reconstituted withpLpp_ [ 14C] _ Glu and the fates of the radioactive glucosyl moiety, in the presence of 4-F-glycogen and normal glycogen were determined as follows. The glucosyl transferreactions were initiated by the addition of PLPP-[ 14 C]-GIu-phosphorylase b to reactionmixes containing either normal glycogen or 4-F-glycogen, in the presence of AMP.Trichloroacetic acid (TCA) was used to stop the reaction, to precipitate the protein and alsoto liberate any remaining protein-bound cofactors. After removing the protein fraction bycentrifugation, the soluble glycogen was precipitated from the supernatant with coldethanol. The percentage incorporation of the radiolabelled glucosyl moiety into the variousfractions is summarized in Table 5-2.1 24Table 5-2. Radiolabel incorporation from PLPP-[ 14C]-Glu-phosphorylase b.aSupernatant Protein Fraction Glycogen FractionNormal Glycogenb4-F-Glycogens55%60%5%16%40%24%a Reactions were carried out at pH 6.7 and room temperature in 100 mM KC1, 50 mM triethanolaminehydrochloride, 1mM EDTA, 1 mM DTT. Incubation time was 60 minutes.b Incubation mixture contained 5.6 p.M PLPP-0 4Q-Glu-phosphorylase b (2.5 x 106 cpm gmol-1 ), 2.3mM AMP, 0.98% normal glycogen.c Incubation mixture contained 18.7 1.1M PLPP-[ 14C]-Glu-phosphorylase b (2.5 x 106 cpm gmol-1 ), 2.3mM AMP, 1% 4-F-glycogen.As shown in Table 5-2, 40% of the radiolabelled glucosyl moiety within PLPP-[14C] -Glu-phosphorylase b was transferred to normal glycogen after an extendedincubation period. The radioactivity remaining associated with the protein fraction 5%)probably represents a small fraction of unreacted cofactor in addition to small amounts oflabelled glycogen which became trapped in the protein upon TCA precipitation. Indeed,earlier studies investigating the glucosyl transfer from PLPP-[ 14C1-Glu-phosphorylase b tonormal glycogen (Takagi et al., 1982) have also found that significant amounts 10%) ofradiolabel are trapped within the precipitating protein. The radiolabel observed within thesupernatant fraction 55%), while higher than expected, is also consistent with earlierstudies (Takagi et al., 1982) in which considerable amounts of the radiolabel wereobserved within the supernatant fraction (20 - 30%) after an extended reaction period.While no explanation for the radioactivity within the supernatant was given, it probablyarises from a combination of incomplete glycogen precipitation, cofactor breakdown toform glucose-1,2-cyclic phosphate, and possibly hydrolysis of the putative glucosyl-enzyme intermediate.125The distribution of radiolabel was found to be considerably different for the reactionof PLPP-[ 14 C]-Glu-phosphorylase b with 4-F-glycogen. After the same period ofincubation, 4-F-glycogen was labelled with 24% of the label, about half of that observedwith normal glycogen. This result suggests that indeed 4-F-glycogen can act as an acceptoroligosaccharide at high enzyme concentrations, but that transfer of the glucosyl moiety to 4-F-glycogen occurs much more slowly in the presence of the blocked glycogen derivative.The protein fraction was found to contain approximately three fold more radiolabel 16%)than was observed in the reaction with normal glycogen, possibly indicating that aglucosyl-enzyme intermediate is formed, yet is ultimately transferred to free 4-hydroxyltermini within the 4-F-glycogen molecule and therefore does not appreciably accumulate.Finally, the supernatant fraction was also observed to contain slightly more radiolabel,possibly supporting the proposal that in the presence of 4-F-glycogen, the glucosyl-enzymeintermediate is somewhat more susceptible to hydrolysis.While the results clearly indicate that both normal glycogen and 4-F-glycogenbecome labelled upon incubation with PLPP-[ 14C]-Glu-phosphorylase b, it is necessary toshow that the radiolabel incorporated into the glycogens occurs as a result of normalphosphorylase catalysis and therefore that the labelled glucosyl moiety was indeedtransferred to a free 4-hydroxyl residue. To this end, the labelled glycogens were isolated,dissolved in standard triethanolarnine buffer containing phosphate (25 mM) (the substratefor the reverse reaction), and treated with a small amount of native glycogenphosphorylase. After an incubation period of 30 minutes the glycogen was re-precipitatedand counted for radioactivity, as was the supernatant. The results showed that 100% of theradiolabel previously associated with the glycogens was transferred back to thesupernatant, in full agreement with the reversibility of the normal enzyme catalyzedreaction. These results demonstrate that the glucosyl transfer from PLPP-[ 14C]-Glu-phosphorylase b to both normal glycogen and 4-F-glycogen is indeed a competent glucosyltransfer reaction.1262.4.^Kinetic Evaluation of 4-F-Glycogen with Native GlycogenPhosphorylase bThe previous experiments have shown that while a glucosyl-enzyme intermediatemay form upon incubation of PLPP-Glu-phosphorylase b with 4-F-glycogen, the blockedglycogen derivative is not completely inert towards addition of further sugar residues.Rather, it possesses free 4-hydroxyl residues which can act as acceptor oligosaccharides.However, the relative rates of glucosyl transfer between phosphorylase and 4-F-glycogencompared to that with normal glycogen are still unknown. In an attempt to estimate areaction velocity for the enzyme catalyzed glucosyl transfer to 4-F-glycogen, enzymereactions were carried out with native glycogen phosphorylase and aG1P under standardassay conditions in the direction of glycogen synthesis, but at high enzyme concentration.The reaction rate for the glucosyl transfer to 4-F-glycogen was compared to that in thepresence of normal glycogen at the same concentration (0.5%). The results showed thatnative glycogen phosphorylase catalyzes the glucosyl transfer from aG1P to 4-F-glycogen(--= 0.5 gmol min -1 mg-1 ) approximately 100 fold more slowly than to normal glycogen (g50 gmol min -1 mg -1 ). Thus, it appears that of the approximately 5000 glucose end groupsfound within one glycogen particle (Madsen and Cori, 1958) about 1% of these remainunblocked in the 4-F-glycogen analogue, and that these free ends are indeed able to act ascompetent acceptor oligosaccharide moieties.3. Discussion 4-F-Glycogen Binding Study While much is known about the branching structure of glycogen, the overallstructure can only be inferred (Goldsmith et al., 1982). Despite X-ray crystallographicstudies that have shown how glycogen phosphorylase binds oligosaccharides such asmaltoheptaose (Goldsmith et al., 1982; Oikonomakos et al., 1991), very little is knownabout the same interaction with the natural substrate. The synthesis of 4-F-glycogen127(Withers, 1990) has allowed a further look into the solution-state interaction betweenphosphorylase and glycogen.The binding interaction between 4-F-glycogen and glycogen phosphorylase b wasinvestigated by 19F NMR spectroscopy. Titration of 4-F-glycogen with phosphorylase bresulted in a progressive decrease in the 19F NMR peak height up to a limiting point andthen no further. The decrease in peak height with increasing enzyme concentrationpresumably reflects binding of the enzyme until saturation is achieved. Interestinglyhowever, the total signal area (reflected in the integration) remained essentially constantthroughout the series. Thus, one explanation for the change in peak height is that as theenzyme concentration is increased, a new broadened component grows into the spectrum ata similar chemical shift. Such a broadened component might well be expected if binding ofthe enzyme to 4-F-glycogen reduces the local mobility of the terminal 4-fluorosugar suchthat it now acquires the effective correlation time of the total 4-F-glycogen/enzyme complexrather than the much shorter correlation time of a locally mobile sugar residue attached tothe glycogen particle.If the former assumption is valid then the 4-F-glycogen 19F NMR signal obtainedin the presence of enzyme results from a combination of "free" and "bound" species. The"bound" population represents 4-fluoro-glucosyl residues which are either directly boundto the protein or are very close in space to bound protein. These likely give rise to a broadcomponent in the 19F NMR signal. The "free" 4-fluoro-glucosyl residues, present withinthe ((phosphorylase)n/ glycogen) macromolecule, are those whose mobility is affected verylittle by the binding of the protein, but which are sterically occluded from interaction withother phosphorylase molecules and these make up the major component of the 19F signal.The T1 value for the 19F signal (T1 = 0.24 sec) was found to be slightly less than thatobserved previously for 4-F-glycogen in the absence of enzyme (T1 = 0.36 sec) (Withers,1990). This result suggests that while a small decrease in mobility has occurred there isstill considerable local motional freedom. Indeed, such a difference might well be expected128even in the absence of any changes in local motion given that the correlation time for the((phosphorylase)n/4-F-glycogen) complex would be considerably larger (i.e. lower rate ofreorientation) than that for 4-F-glycogen itself.Gratifyingly, both sets of titration data revealed an equivalence point ofapproximately 0.12 mM phosphorylase for titrating a 1% 4-F glycogen solution. From thisit is possible to estimate the number of glycogen phosphorylase molecules bound per 4-F-glycogen particle. Assuming a molecular weight of 1 x 10 7 Da for rabbit liver glycogen(Manners, 1957), then a 1% solution of glycogen corresponds to 1 p.M glycogen particles.Since a 1% solution of 4-F-glycogen is titrated by 120 p.M glycogen phosphorylase(expressed as monomers), then an average of 60 glycogen phosphorylase dimers (theoligomeric form of the enzyme known to associate with the enzyme) must be bound perglycogen particle. A similar value for the number of glycogen phosphorylase moleculesbound per glycogen particle has been determined previously in ultracentrifugal studies withhigh molecular weight corn phytoglycogen (Madsen and Cori, 1958). Determination of theamount of phosphorylase co-sedimented at different glycogen/phosphorylase ratios allowedthe number of equivalents bound to be determined. A value of 33 phosphorylase moleculesof molecular weight 500,000 (as thought at that time) per corn phytoglycogen particle (2x107 Da) was determined. Recalculation of this earlier data in light of the known monomermolecular weight of 97,400 would give a total of 85 dimers per phytoglycogen particle.This value is close to that of 60 dimers per particle determined in this work when thedifference in molecular weight, thus of surface areas, of phytoglycogen and mammalianglycogen particles is considered.Interestingly, a glycogen particle of molecular weight 1 x 107 Da will have some5000 glucose end groups (Madsen and Cori, 1958), of which essentially all should bear afluorine label. However, only 60 dimers will be bound to this glycogen particle, providingat best only 240 binding sites (one storage site and one active site per enzyme monomer).Thus, a maximum of only 4.8% of the available glycogen end groups will be directly 129bound to phosphorylase molecules, yet clearly the 19F peak height decreases by a muchgreater percentage than this. This indicates that the mobilities of many more glycogen chainends are decreased significantly upon binding of phosphorylase, even though they are notdirectly bound to the protein. This is quite reasonable given the highly branched structureof glycogen and the physical interference with motion likely afforded by binding of thisrelatively large enzyme.4-F-Glycogen as a Potential Dead-End Substrate? The reaction catalyzed by glycogen phosphorylase proceeds with retention ofconfiguration at the anomeric center. However, one important question concerning themechanism has been whether or not this involves a double displacement reaction yia acovalent glucosyl-enzyme intermediate. Evidence for just such an intermediate has beenobtained previously in both sucrose phosphorylase (Voet and Abeles, 1970) and potatophosphorylase (Kokesh and Kakuda, 1977). However, no such evidence has beenobtained for muscle glycogen phosphorylase. The recent synthesis of 4-F-glycogen(Withers, 1990) has allowed the first attempt at producing an enzyme ternary complex inglycogen phosphorylase with an analogue of glycogen which could allow the formation ofa stable glucosyl-enzyme intermediate, yet prevent its turnover.Initial 31P NMR studies with PLPP-Glu-phosphorylase b were designed to showwhether or not 4-F-glycogen was able to confonnationally activate the enzyme and allowformation of the putative glucosyl-enzyme intermediate. In the presence of 4-F-glycogen,PLPP-Glu-phosphorylase b was converted to the PLPP-enzyme in a reaction that wasessentially complete within the first hour of incubation. These results are consistent withthe glycogen derivative binding to the enzyme and allowing the formation of an enzymeternary complex, and indeed, suggest that a glucosyl-enzyme intermediate may form.However, the ultimate fate of the glucosyl moiety cleaved from PLPP-Glu could not bedetermined by this technique.130Ion-spray mass spectrometric studies of PLPP-Glu-phosphorylase b preincubatedwith 4-F-glycogen did not detect a covalent glucosyl-enzyme intermediate, rather only theapoenzyme mass was observed. Glycogen phosphorylase is only conformationallyactivated when all substrates are bound to the enzyme and thereby has evolved to avoidhydrolysis reactions. Since the LC-MS technique dissociates all 4-F-glycogen from theenzyme, it is not unreasonable to suggest that any glucosyl-enzyme intermediate formedmight be very sensitive to hydrolysis once the 4-F-glycogen derivative is dissociated.Further, due to the high monomer molecular weight of glycogen phosphorylase, it is alsopossible that small amounts (<20%) of the modified enzyme might go undetected in thepresence of predominantly unmodified enzyme (Dr. D. Hess, personal communication).Thus, while ion-spray mass spectrometry has been utilized previously in this laboratory toclearly identify glucosyl-enzyme intermediates in other enzyme systems (Withers,Unpublished results), and indeed, in this case was able to confirm the molecular weight ofglycogen phosphorylase b, it was unable to determine the fate of the glucosyl moietytransferred from PLPP-Glu-phosphorylase in the presence of 4-F-glycogen.The results from radiolabelling studies with glycogen phosphorylase b reconstitutedwith PLPP-[ 14C]-Glu clearly demonstrated that 4-F-glycogen can indeed accept theglucosyl label upon incubation with the enzyme. While increased levels of radiolabel foundin the protein fraction are encouraging and suggest that a glucosyl-enzyme intermediate mayform, the label is clearly being transferred to free 4-hydroxyl groups within the 4-F-glycogen derivative and thus does not accumulate. A small percentage of free 4-hydroxylgroups within the 4-F-glycogen derivative could well be responsible for the observedglucosyl transfer reaction since many glucosyl residues can be successively transfered toone free glycogen chain end. While glucosyl transfer to 4-F-glycogen is not observedunder normal assay conditions, kinetic studies with aG1P at high concentrations of thenative enzyme suggest that 1% of the 4-F-glycogen terminal sugar moieties may act ascompetent acceptors. This value of 1%, however, represents an upper limit since the131kinetic studies do not take into account the potential competing hydrolysis reaction in whichthe glucosyl moiety is transfered to a molecule of water rather than a 4-hydroxyl moietywithin 4-F-glycogen. The source of the free ends within the 4-F-glycogen derivative mostlikely arises from the presence of terminal oligosaccharide residues which are not cappedduring the synthesis. In addition, small numbers of free 4-hydroxyl residues could alsoarise from low level hydrolysis reactions of the glycogen molecule, be they spontaneous orenzyme-catalyzed, from trace glycosidase contaminants. These results have demonstratedthat further attempts to trap or provide supportive evidence for a glucosyl-enzymeintermediate in glycogen phosphorylase, using 4-F-glycogen as a dead-end acceptor, arelikely to be complicated with low level glucosyl transfer reactions to the glycogenderivative, thus alternatives should be sought.Future studies in this area, extending on the 4-F-glycogen theme, include thesynthesis of purified 4-deoxy- or 4-deoxy-4-fluoro-oligosaccharides which are still able tobind glycogen phosphorylase and act as potential dead-end substrates. With the productionof oligosaccharides such as the hexamer, difficulties associated with the preparation andpurification of a homogeneous sample are minimized.Conclusion This study has demonstrated that 4-F-glycogen is not completely inert to glucosyltransfer from phosphate, and indeed an upper limit of 1% was estimated for the number offree 4-hydroxyl residues available for reaction within the 4-F-glycogen derivative. While4-F-glycogen is not perfectly suited to act as an "incompetent" substrate in phosphorylasecatalysis, it has been effectively used in conjunction with 19F NMR to study the solutionbinding interaction between phosphorylase and glycogen. Moreover, 19F NMR hasprovided a direct method for titrating glycogen binding proteins and determining thestoichiometry of such interactions.132CHAPTER 6Materials and Methods1331. Synthesis1.1. General Procedures and MaterialsMelting points were determined using a Laboratory Devices Mel-temp II melting-point apparatus, and are uncorrected.1 H Nuclear magnetic resonance (NMR) spectroscopy was performed on either aVarian XL-300 spectrometer operating at 300 MHz or a Bruker WH-400 operating at 400MHz. Chemical shifts are given in the 8 scale, and are referenced to internaltetramethylsilane (8 = 0.00 ppm) for samples run in CDC13 whereas samples run in D20are referenced to external 2,2-dimethyl-2-silapentane-5-sulphonate (8 = 0.00 ppm). 19Fand 31 P NMR spectra were recorded on a Bruker AC-200E (quad. nuclei probe (Q.N.P.))spectrometer operating at 188 MHz and 81 MHz, respectively. 19F Signal positions aregiven in the 8 scale and are referenced against external trifluoroacetic acid (8 = 0.00 ppm),with signals occurring downfield of the reference being assigned positive S values. 31P-NMR resonances are also given in the S scale and are referenced to external 85%phosphoric acid (8 = 0.00 ppm), with signals occurring downfield of this position beingassigned positive S values. Both 19F and 31 P NMR spectra were collected with broadbandproton decoupling.Low resolution mass spectra (electron ionization) were recorded on a Kratos MS 50mass spectrometer operating at 70 eV. Desorption Chemical Ionization (D.C.I.) massspectra were recorded on a Delsi Nermag R 10-10 C mass spectrometer with NH3 as the CIgas. Fast Atom Bombardment (F.A.B.) mass spectra were recorded on a AEI MS 9 massspectrometer with Xenon as the F.A.B. gun, operating at 7 - 8 KV, 1 mAmp current.Micro-analyses were performed by Mr. P. Borda, Micro-analytical laboratory,University of British Columbia, Vancouver.Solvents and reagents used were either reagent grade, certified or spectral grade.Dry solvents were prepared as follows; methylene chloride was washed several times withconcentrated sulphuric acid, followed by several washings with water and a saturated1 34solution of sodium bicarbonate. The solvent was then pre-dried with sodium sulphate anddistilled from calcium hydride. Pyridine was distilled from calcium hydride.Tetrahydrofuran (THF) was distilled from sodium and benzophenone.Activated manganese dioxide was prepared by heating 454 g of powderedmanganous carbonate in a Pyrex glass dish at 220 - 280 °C for about 18 hours. The blackpowder was stirred with a 1 litre solution of 15% concentrated nitric acid in distilled water.The slurry was filtered with suction, the remaining solid being washed with distilled wateruntil the washings were about pH 5 and then dried at 220 - 250 °C. The resultingmanganese dioxide retained its oxidizing ability for about 2 months kept in a looselystoppered container F. Fukui, personal communication).Diazomethane was prepared with assistance from Anna Dora Gudmundsdottir inProf. Scheffer's laboratory. Ethanol (95%, 2.5 mL) was added to a solution of potassiumhydroxide (0.5 g) in water (0.8 mL) in a 50 mL distilling flask fitted with a droppingfunnel and a condenser (no ground glass joints were used). A receiver flask was attachedto the condenser and cooled in an ice bath throughout the distillation. The flask containingthe alkali solution was heated in a water bath to --,- 60 °C and a solution of Diazald (AldrichChemical Co.) (2.15 g) in diethyl ether (20 mL) was added dropwise through the droppingfunnel, followed by an additional aliquot of diethyl ether (--- 10 mL) to rinse the droppingfunnel. The rate of addition approximately equaled the rate of distillation. The distillationwas continued until the the distilling ether was colorless. The combined etheral distillatecontained approximately 0.3 g of diazomethane (Black, 1983).Thin-layer chromatography (t.l.c.) was performed on aluminum-backed MerckKieselgel 60 F-254 analytical plates. Compounds were detected visually with U.V. light orwith 10% sulphuric acid in methanol. Column chromatography was carried out accordingto Clark-Still et al., (1977) using Kieselgel 60 (180-230 mesh) silica gel.Pyridoxol hydrochloride (pyridoxine hydrochloride) and a-D-glucopyranosylphosphate (bis potassium salt) were purchased from Sigma Chemical Co. [ 14C]-a-D-1 3 5Glucose- 1-phosphate (bis potassium salt) (335 .tCi gmo1 -1 ) was purchased fromAmersham Chemicals. The compounds 3, 4, 6-tri-O-(tert-butyldimethylsilyl)-D-glucal and3, 4, 6-tri-0-(tert-butyldiphenylsily1)-D-glucal were synthesized and kindly provided byEllen Lai and Eric Lam, respectively. In addition, 2-deoxy-2-fluoro-a-D-glucopyranosylphosphate (bis cyclohexylammonium salt), 4-deoxy-4-fluoro-a-D-glucopyranosylphosphate (bis cyclohexylammonium salt) and (1-deoxy-a-D-glucopyranosyl)methylphosphonate (bis cyclohexylammonium salt) were all previously prepared by Dr. IanStreet. 1-Nitro-D-glucal was generously provided by Prof. A. Vassella. The dipotassiumsalts of a-D-glucopyranosyl phosphate and 2-deoxy-2-fluoro-a-D-glucopyranosylphosphate were converted into their bis-(tri-n-butylammonium) salts by passage down acolumn of Dowex 50W-X8 (H+ form) ion-exchange resin into a stirred solution of tri-n-butylamine followed by lyophilization.a4, 3-0—Isopropylidene pyridoxol hydrochloride (1)Compound 1 was prepared according to Korytnyk and Ikawa, (1970). Pyridoxolhydrochloride (10.0 g, 48.6 mmol) was stirred with reagent grade acetone (200 mL) at 0 °Cfor 30 minutes. Gaseous HC1 was bubbled through the suspension for 45 minutes, or untilthe solution was saturated, at which time the solution was allowed to warm to roomtemperature while stirring for an additional 75 minutes. Crystallization was initiated bycooling at -20 °C in the freezer overnight, then diethyl ether (200 mL) was added and thecrystals were collected using a sintered glass funnel and washed with diethyl ether to yield1 as a white solid (10.6 g, 93%). 1 H NMR data (CDC13, 300 MHz): 8 8.30 (s, 1 H, H-6), 4.99 (s, 2 H, C4'-CH2) 4.71 (s, 2 H, C5'-CH2), 2.68 (s, 3 H, C2'-CH3), 1.57 (s, 6H, Isopropylidene).Compound 1 could be quantitatively converted to its free base, 2 , by stirring with anexcess of saturated sodium bicarbonate solution. The mixture is filtered and washed withwater to yield 2 as white solid.136a4, 3-0-Isopropylidene-5'-deoxy-5'-chloropyridoxol hydrochloride (3)Compound 3 was prepared according to Korytnyk and Ikawa, (1970). Compound 1 (6.1g, 24.9 mmol) was crushed to a fine powder and suspended in reagent grade benzene (150mL) with stirring. Reagent grade thionyl chloride (10 mL, 0.14 mole) was addeddropwise, the suspension was heated to reflux and then immediately cooled. Thin-layerchromatography was used to monitor the reaction progress. A further 6 mL (0.08 mole) ofthionyl chloride was added and the suspension was heated to reflux and subsequentlycooled once more. Once complete, the suspended material was collected by suctionfiltration and washed with cold diethyl ether yielding 3 as a white solid (6.2 g, 95%). 1 HNMR data (CDC13, 300 MHz) : 8 8.30 (s, 1 H, H-6), 5.13 (s, 2 H, C4'-CH2), 4.64 (s, 2H, C5 --CH2), 2.82 (s, 3 H, C2'-CH3), 1.64 (s, 6 H, Isopropylidene).Compound 3 could be quantitatively converted to its free base, 4, by stirring with anexcess of saturated sodium bicarbonate solution. After all reaction had ceased the productwas extracted into methylene chloride followed by drying over magnesium sulfate andevaporation of solvent under reduced pressure to yield a green oil which slowlydecomposes with time.Dimethyl 2-(a4, 3-0--isopropylidene-2-methyl-5-pyridyl) ethylphosphonate (5)To dimethyl methylphosphonate (2 mL, 19.0 mmol) in dry THF (20 mL) under nitrogenwith stirring at -78 °C was added n-butyl lithium (12 mL, 19.2 mmol). After 45 minutes atthis temperature the generated dimethyl lithiomethylphosphonate carbanion was canulatedinto a stirred suspension of 4 (1.4 g, 6.16 mmol) and a catalytic amount of tetra n-butylammonium iodide in THF (20 mL) at -78 °C. The mixture was stirred at -78 °C for 30minutes and then allowed to warm to room temperature. The THF was evaporated invacuo followed by an extractive workup (CH2C12-H20), drying over magnesium sulfateand evaporation of solvent under reduced pressure to yield a crude oil. Compound 5 couldbe separated from the main pyridoxol-containing side product, 6 (dimer of 5), only after137extensive silica-gel chromatography which produced 5 and 6 in a ratio of 3 : 2 respectively,with very low overall yield. Data for compound 5. 1 H NMR data (CDC13, 300 MHz) : 87.88 (s, 1 H, H-6), 4.82 (s, 2 H, C4'-CH2), 3.77 (d, 6 H, Lp 10 Hz, -OCH3), 2.72 (m,2 H, C5 '-CH2), 2.39 (s, 3 H, C2'-CH3), 2.01 (m, 2 H, CH2-P), 1.55 (s, 6 H,Isopropylidene); Mass spectral data m/z: 315; Expected mass: 315.Purification of compound 5 from its 'dimer' side product, 6, was not optimized at thisstage, rather the mixture was carried on until the next step where purification wasconsiderably easier.Data for compound 6, the 'dimer' of 5.1 H NMR data (CDC13, 300 MHz): 8 7.81 (s, 2 H, H-6), 4.78 (d, 2 H, JHa,Hb 16 Hz, C4'-CH a), 4.68 (d, 2 H, JHb,Ha 16 Hz, C4'-CHb), 3.58 (d, 6 H, J}Lp 10 Hz, -OCH3), 2.90(ddd, 2 H, J 7, 11, 15 Hz, C5'-Clic), 2.56 (ddd, 2 H, J 7, 15, 19 Hz, C5'-CHd), 2.38 (s,6 H, C2'-CH3), 2.21 (m, 1 H, CH-P), 1.54 (s, 6 H, Isopropylidene), 1.53 (s, 6 H,Isopropylidene); Mass spectral data miz: 506; Expected mass: 506.Dimethyl 243 -hydro.xy-4-hydroxymethy1-2 -methyl-5 -pyridyl) ethylphosphonate (7)Compound 5, present as a crude oil, was dissolved in 10% formic acid (10 mL) andrefluxed for 1 hour then cooled and neutralized (pH 8) with saturated sodium bicarbonate.The water was then evaporated under reduced pressure to give a dark brown solid. Thesolid was suspended in methanol, filtered to remove sodium salts, and reconcentrated toyield an oil. The crude product was purified by flash chromatography (chloroform-1 3 8methanol 9:1) to give 7 as a light yellow oil which solidified upon storage at 4 °C overnight(0.16 g, 11% from 4). 1 H NMR data (CDC13, 400 MHz) 8 7.82 (s, 1 H, H-6), 4.89 (s, 2H, C4'-CH2), 3.72 (d, 6 H, .1Hy 11 Hz, -OCH3), 2.86 (dt (5 lines), 2 H, JH,H 8.0, 8.0,41,P 14 Hz, C5'-CH2), 2.48 (s, 3 H, C2'-CH3), 1.98 (dt, 2 H, JH,H5' 8.0, 8.0, kp 16Hz, CH2-P).Dimethy12 -(4 -formy1-3 -hydroxy -2 -methy1-5-pyridyl) ethylphosphonate (8)The oxidation of 7 was carried out according to Hullar (1969). Compound 7 (0.25 g, 0.9mmol) was dissolved in chloroform (20 mL) and stirred with activated manganese dioxide(1.5 g) in the dark at room temperature for two hours at which time thin-layerchromatography indicated quantitative conversion to a compound that tested positive withdinitrophenylhydrazine spray, an indicator of the aldehyde functionality. The reactionmixture was filtered through Celite and the solvent evaporated in vacuo to give 8 as ayellow oil (0.24 g, 100%). 1 H NMR data (CDC13, 300 MHz) : 6 10.42 (s, 1 H, C4'-CHO), 8.03 (s, 1 H, H-6), 3.78 (d, 6 H, .14-,p 10 Hz, -OCH3), 3.25 (m, 2 H, C5'-CH2),2.55 (s, 3 H, C2'-CH3), 2.09 (m, 2 H, CH2-P).2 -(4 -Formy1-3 -hydroxy -2 -methyl-5 -pyridyl) ethylphosphonic acid (9)The final deprotection of 8 was carried out according to Hullar (1969). Compound 8 (0.24g, 0.90 mmol) was dissolved in 5.7 N HCl (5 mL) and refluxed for 10 hours in the dark.The resultant orange solution was concentrated in vacuo and then diluted with deionizedwater (2 L) until the pH was 3.5. One half of this solution was applied to a column of AG-1X8 resin (Cl - form, 2 cm x 10 cm) and eluted with a linear gradient (2 x 500 mL) of H2O/ 0.01 N HC1, 0.02 N LiCl. Fractions were analyzed by removing aliquots (50 .tL),diluting them into 0.1 N HC1 (1 mL), and measuring the absorbance at 295 nm.Appropriate fractions were collected and lyophilized to yield 9 as a yellow solid mixed withlithium chloride. The yield was estimated from the absorbance at 295 nm knowing the139extinction coefficient to be 6050 M -1 cm -1 (0.1 N HC1) (Hullar, 1969) (80 mg, 35%).Thin-layer chromatography: Rf= 0.5, 13:3:3 (butanol, formic acid, water),dinitrophenylhydrazine positive. 1 H NMR data (D20, 300 MHz) (Solution pH .-- 7):aldehyde species 8 10.48 (s, 1 H, C4'-CHO), 8.17 (s, 1 H, H-6), 3.34 (m, 2 H, C5'-CH2), 2.66 (s, 3 H, C2'-CH3), 2.01 (m, 2 H, CH2-P); hemi-acetal: 8 8.02 (s, 1 H, H-6),6.52 (s, 1 H, H-4'), 3.05 (m, 2 H, C5 '-CH2), 2.58 (s, 3 H, C2'-CH3), 2.01 (m, 2 H,CH2-P); 31P NMR data (D20, 122.5 MHz): 8 24.4 (s, broad); Mass spectral data, F.A.B.(para-nitrobenzylalcohol matrix) (m+1): 252; Expected mass for the mono lithium salt of 9(m+1): 252.The analytical sample was prepared by passing 9 down a small column of Bio-Rex 70 (H+form, 0.7 cm x 12 cm) ion-exchange resin which after lyophilization yielded the free acidform of 9 as a yellow solid. Elemental analysis required C9H12NO5P + 1.2 H2O; C,40.51; H, 5.40; N, 5.25; Found: C, 40.35; H, 5.60; N, 4.94.Diethyl 1 ,1 -thfluoro -2 -( a4 , 3-0—is opropylidene-2 -methy1-5 -pyridyl) ethylphosphonate(10)Butylithium (10 mmol) in dry THF (10 mL) and diisopropylamine (10.3 mmol) was stirredat -20 °C for 20 minutes and then cooled to -78 °C at which time diethyldifluoromethylphosphonate (Bergstrom and Shum, 1988) (10.3 mmol) in dry THF (10mL) was added dropwise. After stirring for 30 minutes at -78 °C to allow generation of thelithio carbanion of diethyl difluoromethylphosphonate, compound 4 (1.56 g 6.9 mmol)suspended in a minimal volume of dry THF ( ---. 5 mL) was added with stirring. Thereaction mixture was allowed to warm to room temperature over approximately 1 hour atwhich time the THF was evaporated in vacuo to leave a brown oil. The crude oil wasredissolved in methylene chloride, washed with saturated sodium bicarbonate, dried overmagnesium sulphate, filtered, and the solvent was evaporated under reduced pressure. Theproduct mixture was purified by flash chromatography (ethyl acetate) to yield 10 as a140colorless oil (120 mg, 5%). 1H NMR data (CDC13, 400 MHz): 6 7.95 (s, 1 H, H-6), 4.83(s, 2 H, C4'-CH2), 4.25 (m, 4 H, -OCH,CH3), 3.25 (dt, 2 H, JH,F 18, 18, J}Lp 5.0 Hz,C5'-CH2), 2.42 (s, 3 H C2'-CH3), 1.52 (s, 6 H, Isopropylidene), 1.32 (t, 6 H, J 7.0 Hz,OCH2CH3); 31P NMR data (CDC13, 81 MHz): 6 6.20 (t, Jp,F. 105 Hz); 19F NMR data(CDC13 188 MHz): 6 -34.50 (d, JF,p 105 Hz, CF2 -P).Diethyl 1 ,1 -difluoro-2 -(3- hydroxy-4 -hydroxymethy1-2 -methy1-5-pyridyl)ethylphosphonate (11)Compound 10 (0.19 g, 0.50 mmol) was dissolved in 10% formic acid/methanol (3 mL,2:1 mixture) and stirred at 60 °C for 5 hours at which time excess chloroform was addedand the reaction mixture was washed with saturated sodium bicarbonate, dried withmagnesium sulphate and filtered. The solvent was evaporated under reduced pressure andfinal purification was achieved by column chromatography (CH2C12-Me0H 25:1) yielding11 as an oil that solidified on standing overnight at 4 °C (0.13 g, 76%). 1 H NMR data(CDC13, 400 MHz): 6 7.89 (s, 1 H, H-6), 4.88 (s, 2 H, C4'-CH2), 4.18 (m, 4 H,OCHzCH3), 3.34 (dt, 2 H, JH,F 18, 18, Jii,p 7.0 Hz, C5'-CH2), 2.46 (s, 3 H, C2'-CH3),1.30 (t, 6 H, J 6.0 Hz, -OCH2CH2); Mass spectral data m/z: 339; Expected mass: 339.Diethyl 1 ,1-difluoro-2-(4-formy1-3-hydroxy-2-methy1-5-pyridyl) ethylphosphonate (12)The oxidation of compound 11 was carried out according to Hullar (1969). Compound 11(0.13 g, 0.38 mmol) was dissolved in chloroform (6 mL) and stirred with 1.5 g ofactivated manganese dioxide in the dark at room temperature for 2 hours at which timecomplete conversion (by t.l.c.) to the desired aldehyde had occurred, and was confirmedby a positive dinitrophenylhydrazine test. The reaction mixture was filtered through Celiteand the solvent evaporated in vacuo to give 12 as a yellow oil (0.10 g, 78%). 1H NMRdata (CDC13, 400 MHz): 6 11.55 (s, 1 H, C3-OH), 10.38 (s, 1 H, C4 --CH0), 8.06 (s, 1H, H-6), 4.25 (m, 4 H, -OCH2CH3), 3.66 (dt, 2 H, JH,F 19, 19, hu, 4.0 Hz, C5'-CH2),1412.52 (s, 3 H, C2'-CH3), 1.36 (t, 6 H, J 7.0 Hz, -OCH2CHa); 31P NMR data (CDC13, 81MHz): 8 5.95 (t, Ju 105 Hz); 19F NMR data (CDC13, 188 MHz): 8 -34.85 (d, Jp,p 105Hz, CF2-P).1,1-Difluoro-2-(4-formy1-3-hydroxy-2-methy1-5-pyridyl) ethylphosphonic acid (13)Compound 12 (50 mg, 0.16 mmol) was dissolved in 5 N HC1 (1.5 mL) and heated to amild reflux in the dark for 16 hours. The solution was cooled and diluted with deionizedwater (2 - 3 L) until the pH was 3.5. The solution was then applied to a column of AG-1X8 (Cl - form, 2 cm x 9 cm) ion-exchange resin and eluted with a linear gradient of H2O /0.01 N HC1, 0.02 N LiC1 (2 x 500 mL). Fractions were analyzed by removing aliquots(100 gL), diluting them into 0.1 N HC1 (1 mL), and measuring the absorbance at 295 nm.Appropriate fractions were collected and lyophilized to yield 13 as a yellow solid mixedwith lithium chloride. The yield was estimated from the absorbance at 295 nm (20 mg,40%). Thin-layer chromatography: Rf= 0.45, 13:3:3 (butanol, formic acid, water),dinitrophenylhydrazine positive. 1 H NMR data (D20, 400 MHz) (Solution pH -..-- 10):aldehyde species 8 10.30 (s, 1 H, C4'-CHO), 7.45 (s, 1 H, H-6), 3.65 (t (br.), 2 H, JH,F20 Hz, C5'-CH2), 2.36 (s, 3 H, C2'-CH3); 31P NMR data (D20, 81 MHz): 8 6.00 (t, Ju86 Hz); 19F NMR data (D20, 188 MHz): 8 -34.70 (d, JF,p 86 Hz, CF2-P).The analytical sample was prepared by passing 13 down a small column of Bio-Rex 70(H+ form, 0.7 cm x 12 cm) ion-exchange resin which after lyophilization yielded themono-lithium salt of 13 as a yellow solid. This sample was used for mass spectralanalysis, U.V./Vis. spectroscopy and elemental analysis. Mass spectral data, F.A.B.(glycerol matrix): (m+1) = 288; Expected mass for the mono lithium salt of 13 (m+1): 288;U.V/Vis. data: Amax 295 nm, E295 = 5700 ± 500 M -1 cm -1 (0.1 N HC1). Elementalanalysis required C9F2H9LiNO5P + LiCI + 3H20; C, 28.20; H, 3.92; N, 3.65; Found: C,28.23; H, 3.55; N, 3.26.142a4, 3-0-Isopropylidene isopyridoxal (14)Compound 2 (1.23 g, 5.88 mmol) was dissolved in CH2C12 (100 mL) and stirred withactivated Mn02 (11 g) at room temperature for 3 hours. At this time t.l.c. indicatedcomplete conversion, and dinitrophenylhydrazine spray gave a positive test for thealdehyde functionality. The Mn02 solids were suction filtered through Celite and thesolvent was evaporated in -vacuo to yield a slightly yellow oil which spontaneouslycrystallized to give 14 as pale yellow crystals (1.00 g 82%). 1H NMR data (CDC13, 200MHz): 8 10.03 (s, 1 H, C5'-CHO), 8.47 (s, 1 H, H-6), 5.18 (s, 2 H, C4'-CH2), 2.50 (s,3 H, C2'-CH3), 1.55 (s, 6 H, Isopropylidene).a'1, 3 -0 -lsopropylidene -5 '-deoxy -5'-chfluoromethylpyridoxol (15)Compound 14 (0.40 g, 1.93 mmol) was dissolved in dry CH2C12 (2 mL) and then addeddropwise to a stirred solution of diethylaminosulfur trifluoride (DAST) (4.6 mmol) in dryCH2Cl2 (1.5 mL). The solution was stirred under nitrogen at room temperature overnightafter which excess CH2C12 was added and the solution was washed once with saturatedsodium bicarbonate and water. The solvent was evaporated in vacuo and the residue waspurified by flash chromatography (ethyl acetate-methylene chloride 1:20) to yield 15 as acolorless oil (0.32 g, 72%). 1 H NMR data (CDC13, 200 MHz): 6 8.05 (s, 1 H, H-6), 6.60(t, 1 H, JH,F 55, 55 Hz, C5'-CH2), 4.99 (s, 2 H, C4'-CH2), 2.45 (s, 3 H, C2' -CH3),1.58 (s, 6 H, Isopropylidene); 19F NMR data (CDC13, 188 MHz): 8 -34.80 (s, C5'-CF2H).5'-Deoxy -5'-difluoromethylpyridoxine (16)Compound 15 (0.72 g, 3.14 mmol) was dissolved in a solution of 0.2 N HCl / Me0H(1:1, 2 mL) and refluxed for 2 hours. Two more aliquots of 0.2 N HC1 (2 mL each) wereadded over an additional hour at which time the the solution was cooled to -20 °C and leftovernight. A portion of compound 16 precipitated out of solution at -20 °C and was143collected by suction filtration. After evaporating the solvent from the filtrate the remainderof 16 could be crystallized from Me0H-ethyl ether (0.47 g, 79%). 1 H NMR data (D20,200 MHz): 8 8.40 (s, 1 H, H-6), 7.20 (t, 1 H, JH,F 55, 55 Hz, C5'-CF2H), 5.10 (s, 2 H,C4'-CH2), 2.70 (s, 3 H, C2'-CH3); 19F NMR data (D20, 188 MHz): 8 -38.60 (s, C5'-CF2H).5 ' -D eoxy-.5 '-difluoromethylpyridoxal (17)Compound 16 (0.30 g, 1.59 mmol) was dissolved in CHC13 (50 mL) and stirred withactivated Mn02 (4 g) at room temperature in the dark for 2 hours. Total conversion to thealdehyde was evidenced by t.l.c. and confirmed by dinitrophenylhydrazine spray. TheMn02 solids were removed by filtration through Celite, then the bulk of the chloroformwas removed by evaporation in vacuo. The volatile product was crystallized upon coolingon ice under a constant stream of nitrogen to yield 17 as yellow crystals (0.19 g, 64%). 1HNMR data (CDC13, 400 MHz): 5 11.75 (s (br.), 1 H, C3-OH), 10.50 (s, 1 H, C4'-CHO),8.20 (s, 1 H, H-6), 6.80 (t, 1 H, JH,F 54, 54 Hz, C5'-CF2H), 2.60 (s, 3 H, C2'-CH3);19F NMR data (CDCI3, 188 MHz): 8 -24.90 (s, C5'-CF2H); Mass spectral data m/z: 187;Expected mass: 187. Elemental analysis required for C8F2H7NO2 + 0.25 H2O; C, 50.13;H, 3.92; N, 7.31; Found: C, 50.05; H, 3.74; N, 7.64.2,6-Anhydro-3-deoxy-D-arabino-hept-2-enoic acid (sodium salt) (18)(1-(Sodium carboxylate)-D-glucal )3, 4, 6-Tri-0-(tert-butyldimethylsily1)-D-glucal (0.53 g, 1.08 mmol) was dissolved in dryTHE (0.5 mL) under nitrogen and cooled to -78 °C with stirring. Tert-butyllithium (1.5mL, 2.5 mmol) was added dropwise and the reaction was stirred at -78 °C (10 minutes) andthen warmed to 0 °C for 30 minutes after which time a small volume of dry THE (0.5 mL)was added. The reaction was then re-cooled to -78 °C and excess CO2, previously passedthrough Drierite, was bubbled into the solution. The solution was allowed to warm to1 44room temperature and then the reaction was quenched with a small volume of water. Afterevaporation of solvents in vacuo, the crude oil obtained was redissolved in CHC13, washedsuccessively with HCl (0.6 N) and saturated NaC1 solution, dried over MgSO4, filteredand solvent evaporated under reduced pressure. Deprotection was achieved by dissolvingthe crude product (-- 0.2 mmol) in THE / 1 M tetra-n-butylammonium fluoride (2.6 mL)with a few drops of triethylamine, stirring at room temperature overnight then evaporationof the solvent under reduced pressure. The crude deprotected compound 18 was dissolvedin Me0H with an excess of Dowex X-2 (Li+ form) ion-exchange resin and the suspensionwas stirred for 5 hours to remove excess tetra-n-butylammonium ions. The resin wasfiltered and the solvent removed by evaporation in vacuo. The residue obtained was re-dissolved in a minimal volume of methanol and filtered through a layer of Celite to removeinsoluble LiF, then the solvent was evaporated in vacuo to yield the deprotected compound18. Final purification was achieved by ion-exchange chromatography using AG-1X8 ion-exchange resin (HC00 - form, 2 cm x 10 cm). The deprotected sugar was loaded on to thecolumn in a small volume of deionized water (:,-- 2 mL) and eluted with a linear gradient offormic acid (0 - 0.5 M) yielding 18 in the protonated form. Bio-Rex 70 ion-exchange resin(Na+ form) was used to convert the protonated carboxylic acid into the sodium salt, whichafter lyophilization yielded 18 as a white powder (0.02 g, 46 %). 1H NMR data (D20,400 MHz): 8 5.63 (d, 1 H, J2,3 2.5 Hz, H-2), 4.33 (dd, 1 H, J3,2 2.5, J3,4 7.4 Hz, H-3),3.97 (m, 1 H, H-5), 3.94 (dd, 1 H, J6,5 2.5, J6,6 , 13 Hz, H-6), 3.86 (dd, 1 H, J6',5 6.1,J6',6 13 Hz, H-6'), 3.60 (dd, 1 H, J4,3 7.5, J4,5 9 Hz, H-4).The analytical sample was prepared by ethanol precipitation of 18 from a minimal volumeof water. Mass spectral data, F.A.B. (3-nitrobenzyl alcohol-5% HCl matrix): (m+1) =213; Expected mass for the sodium salt of 18 (m+1): 213. Elemental analysis required forC7O6H9Na + 1.5 H2O; C, 35.15; H, 5.02; Found: C, 35.37; H, 5.09.145(3, 4, 6-Tri-0-(tert-butyldiphenylsily1)-methyl-2 ,6-anhydro-3-deoxy-D-arabino-hept-2-enoate (19)3, 4, 6-Tri-O-(tert-butyldiphenylsilyl)-D-glucal (1.72 g, 2.00 mmol) was dissolved in dryTHF (1 mL) under nitrogen and cooled to -78 °C with stirring. Tert-butyllithium (4.5 mL,7.2 mmol) was added dropwise and the reaction was stirred at -78 °C (10 minutes) and thenwarmed to 0 °C (0.75 hours) at which time a small volume of dry 'THF was added (1.5mL). The reaction was then re-cooled to -78 °C and excess CO2, previously passedthrough Drierite, was bubbled into the solution. The solution was allowed to warm toroom temperature, the reaction was quenched with a small volume of water and solventswere removed by evaporation in vacuo. The crude oil obtained was redissolved in CHC13,washed successively with HCl (0.6 N) and saturated NaCl solution, dried over MgSO4,filtered and concentrated by rotary evaporation. The crude carboxylic acid was dissolved indiethyl ether (10 mL) and treated with diazomethane. Once esterification was complete theether was evaporated in vacuo and the crude product was purified by columnchromatography (ethyl acetate-hexanes 1:30) to yield 19 as an oil (0.37 g, 20%). 1 HNMR data (CDC13, 400 MHz): 5 7.60 - 7.25 (m, 30 H, aromatics), 5.67 (dd, 1 H, J2,35.4, J2,4 1.5 Hz, H-2), 4.50 (m, 1 H, H-3), 4.10 (m, 1 H, H-4), 4.02 (dd, 1 H, J6,6'11.4, J6,5 7.9 Hz, H-6), 3.89 (m, 1 H, H-5), 3.82 (s, 3 H, 0-CH3), 3.76 (dd, 1H, J6',611.4, J6',5 4.6 Hz, H-6'), 1.00 (s, 9 H, 3 x CH3), 0.91 (s, 9 H, 3 x CH3), 0.73 (s, 9 H, 3x CH3); Mass spectral data m/z: 918; Expected mass: 918.Methyl-2,6-anhydro-3-deoxy-D-arabino-hept-2-enoate (20)(1-(Carboxylic acid methyl ester)-D-glucal )Compound 19 (0.31 g, 0.34 mmol) was dissolved in a solution of THF / 1M tetra-n-butylammonium fluoride (1.70 mL) and stirred at room temperature for 3 hours. Thesolvent was evaporated in vacuo and the crude deprotected sugar was dissolved inmethanol (10 - 20 mL) and stirred with an excess of Dowex X-2 (Li+ form) ion-exchange146resin. The suspension was allowed to stir overnight to remove excess tetra-n-butylammonium ions and was then filtered and concentrated by rotary evaporation. Theresulting oil was redissolved in a minimal volume of ethyl acetate/methanol (60:40) andfiltered through silica to remove insoluble LiF. After evaporation of the solvent in vacuothe residue was purified by column chromatography (ethyl acetate-methanol 10:1) andcrystallized from methanol-ethyl acetate-ether to yield white crystals of 20 (0.024 g, 35%).M.p. 119-121 °C; 1 H NMR data (D20, 400 MHz): 6 6.00 (d, 1 H, J2,3 2.7 Hz, H-2), 4.37(dd, 1 H, J3,2 2.7, J3,4 7.4 Hz, H-3), 4.02 (m, 1 H, H-5), 3.92 (AB multiplet, 2 H, J6,6'12.6, J6,5 5.0, J6',5 2.3 Hz, H-6, H-6'), 3.80 (s, 3 H, OCH3), 3.72 (dd, 1 H, J4,5 9.6,J4,3 7.4 Hz, H-4); Mass spectral data (D.C.I., NH3): (m+NH4+) = 222; Expected mass(m+NH4+): 222. Elemental analysis required for C8H1206 + 0.3 H2O: C, 45.71; H, 6.03;Found: C, 45.66; H, 6.01.Pyridoxal-5 1-pyrophospho-l-a-D-glucose (21)Compound 21 was prepared according to Takagi et al., (1982). Pyridoxal phosphate(0.26 g, 0.97 mmol) was dissolved in dry CHC13 (9 mL) and triethylamine (0.5 mL), thensolvents were removed by evaporation in vacuo and the resultant gum pumped dry. Theresidue was redissolved in CHC13 (8 mL) and pyridine (1 mL) and once again concentratedby rotary evaporation and pumped dry. The residue obtained was redissolved in CHC13 (7mL) and triethylamine (0.4 mL) and stirred with diphenyl phosphochloridate (0.27 mL, 1.3mmol) at room temperature under nitrogen for 3 hours. The solution was then concentratedby rotary evaporation to yield a yellow syrup which was extracted (2x) with diethyl ether(= 3 mL) to remove excess diphenyl phosphochloridate, and the resulting gum was dried invacuo for 45 minutes. To the gum was added bis-tri-n-butylammonium a - D-glucopyranosyl phosphate (1.28 mmol) in dry pyridine (4 mL) and the reaction was stirredat room temperature for 15 hours, then solvent removed by rotary evaporation and theresultant gum pumped dry. This was then dissolved in double-deionized H2O (ddH2O)147(20 mL) and extracted with CHC13 followed by diethyl ether (30 mL each). The aqueousfraction was further diluted with ddH2O (0.5 L) and loaded onto a column of AG-1X8 (C1 -form, 2 cm x 18 cm) ion-exchange resin. The column was washed with ddH2O (150 mL)followed by a solution of 10 mM HC1, 5 mM LiC1 (400 mL). A linear gradient of 10 mMHC1, 5 mM LiC1 / 10 mM HC1, 20 mM LiC1 (500 mL each) was used to elute compound21. Fractions were collected and the absorbance at 390 nm was measured by takingaliquots (0.1 mL) and diluting this into 0.1 N NaOH (3 mL). The elution profile (A390), inaddition to t.l.c (butanol, water, formic acid 13:3:3), was used to identify compound 21.Appropriate fractions were combined and neutralized with LiOH (2 N) to pH 7 and thenlyophilized to yield a yellow powder. Final purification was achieved by precipitating 21from a minimal volume of ddH2O with a combination of methanol/acetone (1:4) yielding21 as a yellow-orange solid (0.056 g, 12%). 1H NMR data (D20, 300 MHz): 6 10.40 (s,1 H, C4' -CHO (pyridyl)), 7.79 (s, 1 H, H-6 (pyridyl)), 5.60 (dd, 1 H, J1,2 3.1, hi,p 7.0Hz, H-1 (glucosyl)), 5.32 (d, 2 H, Jg,p 7.5 Hz, C5'-CH2 (pyridyl)), 3.83-3.45 (m, 6 H,H-2, H-3, H-4, H-5, H-6, 6' (glucosyl)), 2.50 (s, 3 H, C2'-CH3 (pyridyl)); 31 P NMRdata (D20, 81 MHz): 8 -11.00 (d, Jp,p 21 Hz, pyridyl-P), -12.61 (d, Jp,p 21 Hz,glucosyl-P).Pyridoxal pyrophospho-a-D-[ 14C1-glucose, 22, was prepared from PLP and[ 14q-a-D -glucose - 1-phosphate according to the above procedure but on a 1/250 scale andwith the following changes. [ 14q-a-D-Glucose-1-phosphate (0.131=01, 3351.1.Ci pmol-1) was diluted with unlabelled material (5 ilmol) before the reaction with PLP. UnlabelledPLPP-Glu (3 grnol) was added to the crude PLPP-[ 14C]G1u before purification by AG-1X8 (Cl - form, 0.7 cm x 15 cm) ion-exchange resin. Finally, since only a very smallamount of the material was synthesized, the product was not precipitated, but rather wasstored as a lyophilized powder. Thin-layer chromatography (6:4:3); butanol, pyridine,water) of this material showed predominantly a single spot with the same Rf as the148unlabelled material when detected by U.V. and radioactivity. The specific activity of thefinal product was 8 x 106 cpm pmo1 -1 .Pyridoxal-5'-pyrophospho-1-(2-deoxy-2-fluoro)-a-D-glucose (23)Pyridoxal phosphate (0.09 g, 0.33 mmol) was dissolved in dry CHC13 (8 mL) andtriethylamine (0.3 mL), then solvents were removed by evaporation under reducedpressure. The residue was redissolved in CHC13 (5 mL) and pyridine (1 mL) and onceagain concentrated by rotary evaporation and the gum pumped dry. The residue obtainedwas redissolved in CHC13 (5 mL) and triethylamine (0.15 mL) and stirred with diphenylphosphochloridate (0.10 mL, 0.48 mmol) at room temperature under nitrogen for 3.5hours. Solvents were then removed by rotary evaporation to yield a yellow syrup whichwas extracted (2x) with diethyl ether 5 mL) to remove excess diphenylphosphochloridate and the remaining gum was dried in vacuo for 30 minutes. To the gumwas added bis-tri-n-butylammonium 2-deoxy-2-fluoro-a-D-glucopyranosyl phosphate(0.35 mmol) in dry pyridine (5 mL) and the reaction was stirred at room temperature for15.5 hours. After removal of solvents by rotary evaporation and extensive pumping undervacuum the resulting gum was dissolved in ddH2O (15 mL) and extracted twice withCHC13 and once with diethyl ether (30 mL each). The aqueous fraction was diluted to 150mL with ddH2O and then loaded onto a column of AG 1X8 (Cl - form, 1 cm x 20 cm) ion-exchange resin at 4 °C. The column was washed with ddH2O (75 mL) followed by 10 mMHC1 , 5 mM LiC1 (75 mL). A linear gradient of 10 mM HC1, 5 mM LiC1/ 10 mM HC1, 20mM LiC1 (75 mL each), followed by a further single solution of 10 mM HC1, 20 mM LiC1(100 mL) was used to fully elute the desired compound. Fractions were collected and theabsorbance at 390 nm was measured by taking aliquots (0.2 mL) and diluting them into 0.1N NaOH (2 mL). The elution profile (A390), in addition to t.l.c (butanol, water, formicacid 13:3:3), was used to identify compound 23. Appropriate fractions were combinedand neutralized with LiOH (2 N) to pH 7 and then lyophilized to yield a yellow powder.149Final purification was achieved by precipitating 23 from a minimal volume of ddH2O witha combination of methanol/acetone (1:4) which yielded 23 as a yellow solid (0.018 g,11%). 1 H NMR data (D20, 400 MHz): 5 10.42 (s, 1 H, C4'-CHO (pyridyl)), 7.78 (s, 1H, H-6 (pyridyl)), 5.73 (dd, 1 H, J1,2 4.0, JH,p 7.5 Hz, H-1 (glucosyl)), 5.30 (d, 2 H,J13,13 7.0 Hz, C5'-CH2 (pyridyl)), 4.39 (ddd, 1 H, J2,F 49, J2,3 7.5, J2,1 3.5 Hz, H-2(glucosyl)), 3.96 (m, 1 H, H-3 (glucosyl)), 3.85 (m, 1 H, H-5 (glucosyl)), 3.77 (m, 2 H,H-6, 6' (glucosyl)), 3.49 (dd, 1 H, J4,5 9.5, J4,3 9.5 Hz, H-4 (glucosyl)), 2.46 (s, 3 H,C2'-CH3 (pyridyl)); 31P NMR data (D20, 81 MHz): 5 -11.10 (d, Jp,p 20 Hz, pyridyl-P), -13.00 (d, Jp,p 20 Hz, glucosyl-P); 19F NMR data (D20, 188 MHz): 5 -123.60 (s, C2-F(glucosyl)).2. Enzymology,2.1. General ProceduresAbsorbance measurements were carried out on either a Pye Unicam PU-8800 orPU-8600 UV-Visible spectrophotometer. Rabbit muscle was obtained from Pel-FreezBiologicals. T-state glycogen phosphorylase b seed crystals were kindly donated by Prof.N. Madsen. Rabbit liver glycogen (type III) was purchased from Sigma Chemical Co. andwas purified with AG-1X8 (MO - 400 mesh, Cl - form) ion-exchange resin. Glycogen andits analogue 4-F-glycogen were assayed by the method of Dische (Ashwell, 1957).Ammonium sulphate (ultra pure grade) was obtained from Schwarz / Mann Biotech. AMP,AMPS and caffeine were purchased from Sigma Chemical Co. Potassium phosphite wasobtained from ICN Pharmaceuticals.2.2. Protein PurificationGlycogen phosphorylase b (E.C. was prepared from rabbit muscle by themethod of Fischer and Krebs (1962) using DTT instead of cysteine and recrystallized atleast three times before use. Protein concentrations were determined from absorbance150measurements at 280 nm using an absorbance index E2800'1% 1.32 mL mg -1 cm-1 (Bucand Buc, 1968).2.3. Resolution and Reconstitution of Glycogen Phosphorylase bApophosphorylase b was prepared by an adaptation of the method described byShaltiel et al., (1966). To a solution of glycogen phosphorylase b (freed of AMP) wasadded an equal volume of resolution buffer (0.4 M imidazole and 0.1 M cysteine-citrate,pH 6.2) and the mixture was incubated at room temperature for 1.5 hours. The proteinsolution was then passed down a Sephadex G-25 column (2 cm x 30 cm) equilibrated withhalf-strength resolution buffer to complete the resolution, then desalted by passage down asecond Sephadex G-25 column equilibrated with buffer containing 50 mMglycerophosphate / 50 mM mercaptoethanol, pH 7.0. Assays of this apoenzyme routinelyshowed less than 0.5% activity and could be reconstituted with PLP to normal activity.Apoenzyme was reconstituted with various PLP analogues using a molar excess (pyridoxalderivative / apoenzyme) of anywhere from 1 (pyridoxal-5'-pyrophospho- 1 -a-D-glucoseanalogues) to 25 (pyridoxal-5'-phosphonate analogues) fold. Incubation of the apoenzymewith the PLP analogues was carried out at 37 °C in the dark over a period of 45 - 90minutes, the 90 minute incubation time being for those analogues present in only slightexcess over the apoenzyme. The reconstituted enzyme was then precipitated with an equalvolume of saturated ammonium sulphate (pH 6.8), pelleted by centrifugation at 4 °C,redissolved and dialyzed against the appropriate buffer at 4 °C. The absorbance at 333 nmin the UV - Vis spectrum of reconstituted enzyme samples was routinely used as positiveevidence for cofactor binding (Chang et al., 1987; Feldmann and Helmreich, 1976). Anyalterations to this general procedure are outlined within the individual experiments.2.4. Enzymic Synthesis of 4 -F-GlycogenThe synthesis of 4-F-glycogen was carried out as follows, according to Withers,(1990). The reaction mixture (4.0 mL) contained glycogen phosphorylase (5.5 mg mL -1 ),151AMP (1 mM), glycogen (5%) and 4-deoxy-4-fluoro-a-D-glucopyranosyl phosphate (10mM) in buffer containing KC1 (100 mM), sodium glycerophosphate (20 mM), DTT (5mM), and EDTA (1 mM), pH 6.8. After 50 hours at room temperature the reaction mixturewas dialysed (30,000 M.W. cutoff) against the above buffer (500 mL, two changes) over-night, then was replaced into a reaction tube with the same concentration of AMP, 4-deoxy-4-fluoro-a-D-glucopyranosyl phosphate and an additional 1 mg of glycogen phosphorylaseto replace activity lost through denaturation. The reaction mixture was then incubated for afurther 50 hours, and this process was repeated twice more, followed by a final dialysisagainst buffer containing sodium glycerophosphate (2 mM), EDTA (1 mM), and DTT (5mM), pH 7.0. Removal of protein from the glycogen analogue was achieved by passingone half of the the sample down a column (1 cm x 25 cm) of DE53 cellulose equilibratedwith 10 mM Tris buffer, pH 8.3. The column was eluted with the same buffer andfractions (3 mL) were collected and assayed for the presence of glycogen. The processwas repeated for the second half of the sample. The fractions containing glycogen (first 3-4 tubes) from both runs were combined and lyophilized. The glycogen was thenredissolved in a small volume of deionized water (4 mL) and passed down a secondcolumn of DE53 cellulose (1 cm x 25 cm) equilibrated with the same buffer. Fractionswere assayed for contaminating phosphorylase activity and for the presence of glycogen.Appropriate fractions were pooled and lyophilized. In order to desalt the material, theglycogen derivative was dissolved in ddH2O (2 mL) and applied to a BioGel P2 column(1.6 cm x 66 cm) pre-equilibrated with deionized water. The column was eluted withU.1120 and the effluent was monitored with a U.V. detector, the glycogen being eluted inthe void volume. Fractions containing the desalted glycogen derivative were combined andlyophilized to yield 75 mg. Kinetic parameters for the reaction of this 4-F-glycogenderivative with glycogen phosphorylase were identical to those of samples previouslysynthesized in this laboratory by Karen Rupitz. The glycogen derivative bound very tightly152to the enzyme and was unable to act as a substrate in the direction of glycogen synthesisunder normal assay conditions.2.5. Kinetic Experiments2.5.1. General ProceduresGlycogen phosphorylase was primarily assayed in the direction of glycogensynthesis with initial reaction rates determined by the Fiske-Subbarow phosphate analysisas described in Engers et al., (1970a, b). The buffer employed for all kinetic experiments,unless otherwise stated, contained 50 mM triethanolamine hydrochloride, 100 mM KC1, 1mM EDTA, 1 mM DTT, pH 6.8. Reaction mixtures were 0.5 mL containing 1 mM AMPand 0.5% - 1% glycogen, and studies were performed at pH 6.8, 30 °C with 5 minutereaction times. Generally, the enzyme was preincubated with AMP and glycogen and thenadded to a solution containing substrate. The concentration of enzyme in the reactionmixture depended on the enzyme system employed. When glycogen phosphorylase wasassayed in the direction of glycogen degradation, the phosphoglucomutase / glucose-6-phosphate dehydrogenase coupled assay described previously by Engers et al., (1969) wasemployed. All values of Vmax and Km were calculated by fitting the rate data to the non-linear form of the Michaelis-Menten equation with the aid of the GraFit computer program(Leatherbarrow, 1990). The data were, however, plotted according to Lineweaver andBurke, (1934) for presentation purposes.2.5.2. Kinetic Evaluation of Glycogen Phosphorylase b Reconstituted with5-CH2PLP and 5-CF2PLPApophosphorylase b was reconstituted under standard conditions with a 25 foldexcess of 5-CH2PLP and 5-CF2PLP. Initial reaction rates for the reconstituted enzymeswere measured in the direction of glycogen synthesis with a varying concentration of a-D-glucopyranosyl phosphate (1 - 19 mM), under standard assay conditions. Reaction times1 53were 5 minutes, conducted at pH 6.8 and 30 °C in standard triethanolamine buffer, andcontained 1 mM AMP and 1% glycogen. The enzyme concentrations used in the reactionswere the following: native glycogen phosphorylase (1.94 p.g inL -1 ), 5-CH2PLP-phosphorylase b (8.02 pg mL-1 ), 5-CF2PLP-phosphorylase b (6.24 p.g tnL-1 ).2.5.3. pH-Dependence of the Kinetic Parameters for GlycogenPhosphorylase b Reconstituted with 5-CH2PLP and 5-CF2PLPInitial reaction rates for the reconstituted enzymes at various pH values weremeasured in the direction of glycogen synthesis with varying a-D-glucopyranosylphosphate (1.5 - 22 mM) concentration. All reaction times were 5 minutes, conducted at30 °C in buffer containing 1 mM AMP and 1% glycogen. Reactions were conducted at pHvalues ranging from 5.78 to 7.55 for 5-CH2PLP-phosphorylase and 5.55 to 7.55 for 5-CF2PLP-phosphorylase. The standard triethanolamine buffer system was used over theentire pH-range, the buffer pH being adjusted with concentrated HCl or NaOH (2 N). Thefinal pH values of 5.55, 5.78, 6.06, 6.27, 6.61, 6.84, 7.28 and 7.55 were recorded forreaction mixtures containing all substrates, effectors and enzyme. The followingequilibrium constants (Xe), (Street, 1988) were used in the calculation of initial rates at thevarious pH values.pH Xe5.55 0.8885.78 0.8766.06 0.8566.27 0.8366.61 0.7966.84 0.7667.22 0.7197.55 0.698Control experiments were completed with each enzyme system to ensure that irreversiblepH-dependent inactivation did not occur at the pH values studied. The enzyme1 54concentration employed in each reaction mix was 9.15 p.g mL -1 and 7.01 p.g mL-1 for the5-CH2PLP- and 5-CF2PLP-phosphorylase enzymes, respectively.2.5.4. Reconstitution and Reactivation of Apoglycogen Phosphorylase bwith PL-CF2HApophosphorylase b was reconstituted with PL-CF2H according to methodsdescribed previously for 5'-deoxypyridoxal analogues (Chang et al., 1987). Initial reactionrates for PL-CF2H-phosphorylase were measured in the direction of glycogen synthesiswith varying a-D-glucopyranosyl phosphate concentration (8 mM - 60 mM) at a fixedphosphite (6 mM) concentration. All reaction times were 5 minutes, conducted at pH 6.8and 30 °C in standard triethanolamine buffer, and contained 1 mM AMP and 1% glycogen.The reaction mixture contained an enzyme concentration of 62.6 ug rnL -1 . Values of Vm(1-2 grnol min -1 mg-1 ) and Km mM) were determined from a Lineweaver-Burkeanalysis.2.5.5.^Reactivation Kinetics for PLPP-GIu- and PLPP-2FG1u-Phosphorylase bApophosphorylase b was reconstituted with PLPP-Glu and PLPP-2FGluessentially as previously described for PLPP-Glu by Withers et al., (1981a). Thus theapoenzyme (15.2 p.M) was mixed with one equivalent of PLPP-Glu and PLPP-2FG1u in abuffer containing 25 mM sodium glycerophosphate, 25 mM mercaptoethanol, 50 mM KC1,25 mM triethanolamine hydrochloride, 0.5 mM EDTA, and 0.5 mM DTT. The incubationswere carried out either in the presence or absence of 0.24% glycogen at the indicated pH,and at room temperature for a period of 5 days. The recovery of native enzyme activitywas followed by removing a small aliquot (5 - 101.1.L) of the reconstituted enzyme sampleand assaying it in the direction of glycogen synthesis under standard conditions. Thus,reactions were carried out over 5 minutes at 30 °C in mixtures containing a - D-1 5 5glucopyranosyl phosphate (18 mM), AMP (1 mM), glycogen (1 %) and enzyme (30 ggmL-1 ), all in triethanolamine buffer (pH 6.8).2.5.6 Incubation of the R- and T-state forms of Glycogen Phosphorylase bwith Glucal AnaloguesR-state glycogen phosphorylase b was incubated in triethanolamine buffercontaining AMP (1 mM) and glycogen (1%) at pH 6.8 and 30 °C in the presence of thevarious glucal analogues 30 - 40 mM, refer to the legend in Figure 3-1 for details). TheT-state enzyme was incubated under the same conditions but without AMP and glycogen.Residual activity was measured by removing small aliquots of the inactivation mixture (5 -10 p.L) (2 - 4 gg enzyme) and diluting this into a reaction mixture (0.5 mL) containing asaturating concentration of a-D-glucopyranosyl phosphate (...16 mM), AMP (1 mM) andglycogen (1%). The initial reaction rates were assayed in the direction of glycogensynthesis at 30 °C and pH 6.8 in triethanolamine buffer under standard assay conditons(Engers et al., 1970a, b). The pseudo first order rate constants (k obs) for the decrease inenzyme activity were determined with the aid of the GraFit computer program(Leatherbarrow, 1990).2.5.7. The Determination of lc; and K1 for the Nitroglucal Inactivation ofGlycogen Phosphorylase bFor the R-state enzyme, solutions of glycogen phosphorylase b were incubated at30 °C and pH 6.8 in buffer containing 10 mM MES, 10 mM HEPES, 10 mMtriethanolamine hydrochloride, 100 mM NaC1, and 1 mM AMP in the presence of a varyingconcentration of nitroglucal (5 - 57 mM). The T-state enzyme was incubated at 30 °C andpH 6.8 in the same way, but in the absence of AMP. In addition to the inactivationexperiment described above, the T-state enzyme was also inactivated in the presence ofcompetitive ligands (glucose and/or caffeine) as outlined in Figure 3-4, and also in another156experiment, was inactivated over a range of pH using the same buffer system describedabove. The range of pH values covered in this latter experiment included 5.90, 6.80, 7.20,7.90, 8.47, 9.00. In all experiments the inactivation time course was followed byremoving a small aliquot from the inactivation mixture (5 gL) and diluting this into areaction mix (0.6 mL) containing 40 mM imidazole, 10 mM magnesium acetate, 0.4 mMEDTA, 1 JIM glucose-1,6-diphosphate, 0.5% glycogen, 1 mM AMP, 21 mM phosphate, 1mM NADP, 15 units mL-1 phosphoglucomutase, 3.4 units mL-1 glucose-6-phosphatedehydrogenase (pH 6.8). The reaction rate at each time interval was measured in thedirection of glycogen degradation by following the change in absorbance at 340 nm.Pseudo first order rate constants (kobs) were determined for the time-dependent decrease inenzyme activity at each concentration of nitroglucal with the aid of the GraFit computerprogram (Leatherbarrow, 1990). Values of ki and Ki were determined by fitting the k obsvalues to the non-linear form of the Michaelis-Menten equation. Values of pK a quoted forthe pH-dependence of the inactivation due to nitroglucal were calculated by fitting the datato standard pH-functions (Appendix A) with the aid of the GraFit computer program.Note: thiol reducing agents (mercaptoethanol and DTT) were excluded from the buffersystem used in these experiments because of their reaction with nitroglucal. Technicalassistance with the nitroglucal inactivation experiments was obtained from Karen Rupitz.2.5.8. Glucosyl Transfer from PLPP-[ 14 Q-Glu-Phosphorylase b toGlycogen and 4-F-Glycogen.Glycogen phosphorylase b was reconstitued with PLPP-[ 14C]-Glu using thestandard conditions for reconstitution previously described except for the followingchanges. The apoenzyme was concentrated with a Centri-prep concentrator (30,000 M.W.cutoff) at 4 °C before reconstitution such that ammonium sulphate precipitation ofradiolabelled protein could be avoided. The apoenzyme was reconstituted with a 1.9 foldexcess of PLPP-[ 14C]-Glu and the incubation period was extended to 90 minutes. Dialysis157of the reconstituted enzyme was done in the standard triethanolamine buffer system exceptthat the pH was 6.7 such that spontaneous regeneration of native enzyme would beminimized. The specific activity of the reconstituted enzyme was 2.5 x 10 6 cpm grno1 -1 .Reactions with PLPP-[ 14Q-Glu-phosphorylase b were initiated by addition of theenzyme to reaction mixtures containing the standard triethanolamine buffer (pH 6.7), AMP(2.3 mM), normal glycogen (0.98 %) or 4-F-glycogen (1 %) in a final volume of 200 p.L atroom temperature. Reactions were terminated by the addition of cold 5% TCA (400 j.t1.)and the protein was then spun down with a microfuge at 4 °C. The protein pellet waswashed at least three times with 2.5% TCA (100 )11.) and then dissolved in 0.1 N NaOH(100A) and counted for radioactivity. Glycogen was precipitated from the supernatantwith cold ethanol (800 1,1L) and then spun down with a microfuge at 4 °C. The glycogenpellet was washed at least three times with cold ethanol. The glycogen pellet was thenredissolved in water and counted for radioactivity. A portion of the supernatant was alsocounted for radioactivity.The same procedures were used to isolate labelled glycogen and test the reversibilityof the labelling reaction with native glycogen phosphorylase. In this case once the labelledglycogen was precipitated, a portion was redissolved in the standard triethanolamine buffer(pH 6.8). To the glycogen solution was added phosphate (25 mM) and native glycogenphosphorylase 10 ug). The reaction was allowed to proceed for 30 minutes at roomtemperature at which time the reaction was stopped with trichloroacetic acid. Unlabelledglycogen was added to the reaction mixtures to enhance precipitation of glycogen, afterwhich the samples were analyzed as described above.The kinetic evaluation of 4-F-glycogen was carried out under standard experimentalconditions. Reaction was monitored in the direction of glycogen synthesis at 30 °C and pH6.8 in the standard triethanolamine buffer. The reaction contained a saturatingconcentration of a-D-glucopyranosyl phosphate (18 mM), activating AMP (1 mM) and 4-158F-glycogen (0.5%). High concentrations of glycogen phosphorylase (--= 100 .tg triL -1 )were required to detect appreciable turnover.3. NMR Experiments 3.1. General ProceduresGenerally, all protein 19F and 31P NMR experiments were carried out under thefollowing conditions unless otherwise noted. Signals for the 19F and 31 P NMRexperiments are reported in the 8 scale and were referenced against trifluoroacetic acid or85% phosphoric acid, signals occurring downfield of the reference being assigned positive8 values. Experiments were generally conducted without proton decoupling.Protein NMR experiments were done at high enzyme concentration (0.7 - 1.0 mM)in buffer containing 50 mM triethanolamine hydrochloride, 100 mM KCI, 1 mM EDTA and1mM DTT (pH 6.8). Protein was concentrated using a Millipore CL filter (30,000 M.W.cutoff) and then dialyzed against a small volume (= 10 mL) of D20-triethanolamine buffer(pH 6.8) such that the D20 concentration was 50 - 60%. All D20 used in these NMRstudies was previously treated with Chelex to remove any paramagnetic impurities. TheD20 present in the buffer was used for field/frequency lock. Solutions of effectors indeionized water or triethanolamine buffer (pH 6.8) were added from stock solutions. Theexact conditions for each NMR experiment are reported in the legend to the appropriateFigure.All solution-state 31P NMR experiments with glycogen phosphorylase and itsderivatives were collected in the Department of Biochemistry in Prof. P. Cullis' laboratorywith the very generous technical assistance of Dr. K. Wong.1593.2. 31P NMR Titration of 5-CH2PLP and 5-CF2PLP31P NMR spectra were recorded on a Varian XL-300 spectrometer operating at 121MHz at room temperature using the standard parameters for non-protein samples asdescribed in Section 1.1. (General Procedures and Materials). The 5-CH2PLP (28 mM)and 5-CF2PLP (23 mM) phosphonic acids were present in a solution of 50% D20 and 100mM KC1. The pH of the solution was adjusted with a small volume (1-5 W.) ofconcentrated HC1 or NaOH (2 N), after which the spectra were accumulated (384 scans).Generally a 4 Hz line-broadening factor was used in data processing. Values of pKa werecalculated with the aid of the GraFit computer program (Leatherbarrow, 1990).3.3. 31 P NMR of Native Glycogen Phosphorylase b and of 5-CF2PLP-Phosphorylase b31P NMR spectra were recorded at 28 °C on a Bruker MSL 200 operating at 81MHz with a 10 mm probe. A spectral width of 10,000 Hz was employed, with a 20 - 30°pulse angle and a repetition time of 1.5 seconds. Exponential line-broadening (25 Hz) wasused prior to Fourier transformation, and all linewidth data have been corrected for this.Sample size was 1.3 - 1.5 mL in a 10 mm NMR tube, with enzyme concentrations between0.9 and 1.0 mM calculated for the phosphorylase monomer.Native glycogen phosphorylase was dialysed at 4 °C for 48 hours against twochanges of triethanolamine buffer (pH 6.8) to remove AMP and then was further dialysedagainst the same buffer containing 75 % D20. 5-CF2PLP-phosphorylase b was preparedas previously described and both enzymes were concentrated using the conditionsdescribed above in General Procedures.3.4. 19 F NMR of 5 -CF2PLP- and PL -CF2H -Phosphorylase bThe 19F NMR spectra of glycogen phosphorylase b reconstituted with 5-CF2PLPand PL-CF2H were recorded on a Bruker AC-200E (quad. nuclei probe (Q.N.P.))1 60spectrometer operating at 188 MHz with a 5 mm probe. A spectral width of 30,000 Hzwas employed with a 35 - 40° pulse angle and a repetition time of 0.8 - 1.1 seconds.Exponential line-broadening (75 Hz) and data 'left shifts' were used prior to Fouriertransformation and all linewidth data have been corrected for the line-broadening factor.Sample size was 0.4 - 0.5 mL, with enzyme concentrations between 0.6 and 0.8 mMcalculated for the phosphorylase monomer.Enzyme, present in 50 mM triethanolamine hydrochloride, 100 mM KC1, linMEDTA, 1 mM DTT (pH 6.8) was concentrated using a Millipore CL filter (30,000 M.W.cutoff). An equal volume of D20-triethanolamine buffer (pH 6.8) was added such that thefinal concentration of D20 was 50%. Further concentration with the Millipore filter wascompleted until the desired enzyme concentration was reached.3.5. Solid State 31P MASNMR Spectroscopy of free PLP and of R- and T-State Glycogen Phosphorylase bThe PLP monosodium and disodium model compounds were obtained from thefree acid of PLP by titrating solutions with sodium hydroxide to pH 4.0 and 8.0respectively, followed by lyophilization. AMP was removed from glycogen phosphorylaseby dialysis against large volumes of triethanolamine buffer (pH 6.8). Crystalline samplesof the T-state enzyme were obtained by seeding of a solution of phosphorylase b (15 - 30mg mL-1 ) in buffer containing 1 mM EDTA, 5 mM DTT, 10 mM 2-fbis(2-hydroxyethypaminolethanesulphonic acid (BES) (pH 6.7) containing glucose (50 mM)with a highly diluted suspension of seeds obtained by grinding tetragonal crystals ofphosphorylase b (Johnson et al., 1974; Kasvinsky and Madsen, 1976). MicrocrystallineR-state phosphorylase b was obtained from solutions of phosphorylase b (16 mg mL -1 ) inbuffer containing 100 mM KC1, 50 mM triethanolamine hydrochloride, 1 mM DTT, 10mM magnesium acetate, 0.4 mM adenosine-5'-thiomonophosphate, pH 6.8. The proteinmicrocrystals were packed into the rotor using an Eppendorf centrifuge followed by161spinning the rotor in the probe at several kHz for fifteen minutes and removing the residualbuffer by pipette. This process was repeated several times if necessary to fill the rotor.Experiments on both the T- and R-state enzymes were initially non-reproducible since the31P-MASNMR spectrum sometimes consisted of only isotropic peaks. However when theabove protocol was followed and the samples were spun for prolonged periods before theremoval of excess buffer by pipette, a sideband manifold was always observed.NMR experiments were performed and analyzed in the laboratory of Professor C.A. McDowell by Dr. R. Challoner on a Bruker MSL 200 pulse spectrometer with anoperating frequency of 81.05 MHz for 31 P. A spectral width of 50,000 Hz was employedand experiments were proton decoupled. An exponential line-broadening factor of 100 Hzwas applied for the spectral analysis, and spinning frequencies in the 2.0 - 2.2 kHz rangewith a recycle time of 20 s were typically used. 31P-MASNMR spectra were simulated byDr. R Challoner as described in Challoner et al., (1992). The convention a33>a22>ai Iwas used for the assignment of the principal tensor components obtained from the spectralsimulations of the spinning sideband intensities.3.6. 19F NMR of PLPP -2FG1u -Phosphorylase bThe 19F NMR spectra of glycogen phosphorylase b reconstituted with PLPP-2FG1u were recorded on a Bruker AC-200E (quad. nuclei probe (Q.N.P.)) spectrometeroperating at 188 MHz with a 5 mm probe. A spectral width of 20,000 Hz was employedwith a 40 - 60° pulse angle and a repetition time of 1.2 - 2.0 seconds. Exponential line-broadening (75 Hz) and data 'left shifts' were used prior to Fourier transformation and alllinewidth data were corrected for the line-broadening factor. Sample size was 0.4 - 0.5mL, with enzyme concentrations between 0.7 and 1.0 mM calculated for the phosphorylasemonomer. 2-Fluoro-D-glucal (-92.2 ppm) was used as an internal standard in allexperiments. PLPP-2FG1u-phosphorylase b was prepared as previously described andconcentrated using the conditions described above in the General Procedures.1 623.7. 19F NMR Titration of 4-F-Glycogen with Glycogen Phosphorylase b19F NMR spectra were recorded at room temperature on a Bruker AC-200E (quad.nuclei probe (Q.N.P.)) spectrometer operating at 188 MHz with a 5 mm probe. A spectralwidth of 20,000 Hz was employed with a 35 - 40° pulse angle and a repetition time of 1.2seconds. Exponential line-broadening (25 Hz) was used prior to Fourier transformationand all linewidth data have been corrected for the line-broadening factor. Measurements of(Ti) values were performed using the progressive saturation method (Freeman and Hill,1971).All chemical samples were dissolved in the standard triethanolamine buffer (pH6.8) containing 75% D20. Glycogen phosphorylase was dialysed at 4 °C for 48 hoursagainst two changes of triethanolamine buffer (pH 6.8) to remove AMP and then wasfurther dialysed against the same buffer containing 75% D20. 2-Fluoro-D-glucal (8 = -92.3 ppm) was used as an internal standard in all experiments.Titrations were performed by the sequential addition of glycogen phosphorylase(0.97 mM) to a sample (0.52 mL) of 4-F-glycogen (1.2 or 0.6%), 2.1 mM AMP, 2.3 mM2-fluoro-D-glucal in 75% D20 buffer. The 4-F-glycogen concentration was kept constantas the volume increased by the addition of small aliquots (3 - 8 41.) of concentrated(5.75%) 4-F-glycogen. Such constancy in concentration made subsequent data analysisconsiderably more simple. 19F NMR spectra (256 scans) were recorded after each additionof enzyme and compared after setting the absolute intensity of the first free induction decayto 1.3.8. 31 P NMR of PLPP-Glu-Phosphorylase b in the Presence of 4-F-Glycogen.31P NMR spectra were recorded at 30 °C on a Bruker MSL 200 operating at 81MHz with a 10 mm probe. A spectral width of 10,000 Hz was employed with a 20 - 30 °pulse angle and repetition time of 1.5 seconds. Exponential line-broadening of 20 Hz was163used prior to Fourier transformation, and all linewidth data have been corrected for this.Sample size was 1.7 - 1.8 mL in a 10 mm NMR tube, with enzyme concentrations between0.8 and 0.9 mM calculated for the phosphorylase monomer. PLPP-Glu-phosphorylase bwas prepared as previously described and concentrated using the conditions describedabove in the General Procedures.4. Analysis of Pyridoxal Cofactors from PLPP -2FG1u -phosphorylaseThe analysis of pyridoxal compounds was carried out essentially according toTakagi et al., (1982), and Tagaya and Fukui, (1984). Both the control and enzymesamples were treated in the following way. The samples (0.06 - 0.15 gmole) in ddH2O orstandard triethanolamine buffer (pH 6.8) 0.1 mL) were treated with 1 mL of cold 5%trichloroacetic acid (TCA) and chilled on ice for 5 minutes, then the protein was spun downand the supernatant removed. This procedure was repeated once more, combiningsupernatants. The same procedure was repeated two more times with 2.5 % TCA. Thecombined TCA supernatants were extracted several times (7 - 8) with ether, and then theaqueous fraction was rotary evaporated under low vacuum conditions such that residualether was removed. At this time the PLP standard 0.06 gmole) was added to thecofactor solution and the volume was adjusted to 5 mL with ddH2O. The 5 mL samplewas loaded onto a column of AG-1X8 ion-exchange resin (Cl - form, 0.5 cm x 10 cm),followed by a 5 mL wash with ddH2O. The column was then eluted with a linear gradientof 2 mM HC1, 10 mM NaC1 / 2 mM HCI, 70 mM NaC1 (50 mL each), followed by asolution of 2 mM HC1, 70 mM NaCl (20 mL), and finally a solution of 2 mM HCI, 85 mMNaC1 (20 mL). Fractions (--= 2 mL) were collected and analyzed by the method of Wadaand Snell (1961).1 645. Ion-Spray Mass Spectral StudiesAll ion-spray mass spectral studies were completed at the Biomedical ResearchCenter at the University of British Columbia in Dr. R. Aebersolds laboratory on a triplequadrupole mass spectrometer from Sciex. The spectra were collected in the LC-MS mode(single MS) by Dr. J. Gebler and Dr. D. Hess. All protein samples were treated in thesame way prior to mass spectral analysis. The protein samples were centrifuged to removeany insoluble material and then separated from low molecular weight compounds by HPLCusing a 2.1 mm narrow bore C4 column. The following gradient was applied to elute theprotein; 20% buffer B in buffer A to 90% buffer B in buffer A over 10 minutes. Buffer A:0.05% TFA (aq.), 2% acetonitrile (aq.); Buffer B: 0.045% TFA (aq), 80% acetonitrile(aq)•The experimental conditions used for the preparation of the nitroglucal inactivatedsample of glycogen phosphorylase are the same as those used in the previous inactivationexperiments (see Section 2.5.7. of the Experimental).For the reactions of PLPP-Glu-phosphorylase b, incubation mixtures containedenzyme (67 gM), AMP (1.7 mM) and 4-F-glycogen at either 0.07% or 0.35%.Incubations were carried out at room temperature and were between 5 and 90 minutes inlength as described in the legend to Table 5-1.165APPENDIX AKinetic Derivations and pH-Dependence of Enzyme Reactions1 661. Kinetic DerivationsThe expressions used to analyze the data in this work were derived by applicationof the steady state approximation to the following general kinetic expression:k 1^k2E + S i.--b ES -4 E + Pk.1in which k2 is the rate limiting step of the reaction. Thus the reaction rate will be equal to:1) = k2 [ES]where k2 (or kcal in this case) is simply the first order rate constant for the chemicalconversion of the enzyme-substrate complex (ES) to the enzyme-product complex. Sincethe initial reaction rate is approximately constant over the time interval concerned the steadystate approximation applies to the concentration of ES where:d[ES] = 0dtAssuming that [ET] = [E] + [ES], and that [So] >> [E0], then the velocity equationcan be expressed in the following way:v = d[P] = k2 [ES] = k2 [Eol [S] dT^ Km + [S]where Km is the Michaelis constant and is defined as:^k2 + k.1 k1The Michaelis constant is equal to Kd (i.e. k-1 / k1), the dissociation constant of theenzyme-substrate complex, when k2 << k_1 (rapid equilibrium assumption).The specificity constant (kcatfi(m) is a second order constant which relates thereaction rate to the concentration of free, rather than total enzyme. This becomes apparent1 67at low substrate concentrations where the enzyme is largely unbound and [E] .. [E 0]. Atsuch concentrations the reaction rate is given by:1) = [E] [S] kcatKmGlycogen phosphorylase is a two-substrate enzyme. However, when one of thesubstrates is present at saturating concentrations (e.g. glycogen), the kinetic modelsimplifies to that of a single substrate system as shown above. The rate limiting step in thephosphorylase mechanism is the isomerization of the central enzyme ternary complex.Thus, isomerization is slow compared to the dissociation of the enzyme-substrate complex(k2 << k.i) and values of Km will be equal to Kd, the dissociation constant for the enzyme-substrate complex.In the presence of a competitive inhibitor, the kinetic scheme is modified as shownbelow:ki^k2E + S „=-2 ES -4 E + PKi j t^k-1EIand gives rise to the following rate equation:u =^rEol IS1 k2 Km (1 + [I]/Ki) + [S]1682. pH-Dependence of Enzyme ReactionsInvestigating the pH-dependence of the kinetic parameters from the Michaelis-Menten equation is a common practice and the interpretation of such pH-profiles isgenerally based on equations derived by Alberty et al., (1963). In their derivations,Alberty and coworkers made two important assumptions which have been disregarded inthe past (Knowles, 1976). In summary, the authors stated that if there is only one activesite ionization state capable of catalyzing the interconversion of substrate and product, andif all prototropic equilibria involving the ionizing groups are fast with respect to all othercatalytic steps in the reaction, then a number of conclusions can be made from the variationof kcat/Km and of kcat with pH. Plots of kcat/Km versus pH will yield pKa values forionizing groups within the free enzyme and / or free substrate, and plots of kcat versus pHwill yield pKa values for the enzyme-substrate complex whose decomposition is ratelimiting (Knowles, 1976). The pH-dependence of Km follows ionizations which affectbinding of the substrate to the enzyme, thus can yield information on either the free enzymeand free substrate or the enzyme-substrate complex.The active site of an enzyme may contain several ionizing groups which may beinvolved in substrate binding or catalysis. The following model illustrates simple diproticequilibria in which the free enzyme and the enzyme-substrate complex can each be presentin three different ionizations.EH2 + S^EH2S^EH2Kb t,^Kb'EH + S^EHS k2EH + PE + S^ 1" ES169In this scheme, reaction can only occur through the monoprotonated form of the enzyme-substrate complex and it is assumed that all prototropic equilibria (vertical processes) arefast compared to the rate limiting interconversion step. Thus, the pH-dependence of thereaction velocity and the various kinetic parameters can be written as follows:1) = ^k2 [Eo] [S]Kd (1 + [Hi- ]/Kb + Ka/[H+]) + [S] (1 + [H +]/K'b + Wai[11 4 ])k2 (pH) = ^k2 (1 + [H+]/K'b + Wa/[1 .1 -4 ])Kd (pH) = Kd x (1 + [11 4- ]/Kb + Ka/[H +]) (1 + EH -11/K% + K'ai[H+])k2/Kd (pH) =  k2  x ^1 Kd^1 + [H -f]/Kb + Ka/[11+]where k2 (kcat) and Kd (Km) are pH-independent.At low pH where [H -1 >> Ka these equations can be simplified. For example:k2 (pH) = ^kz. (1 + [H -I- ]/1C•b)and similarly at high pH where [H4-] is << Kb,k2 (pH) = ^kz (1 + K' a i[H -F])170Previous studies have employed these equations in an effort to aid in the interpretation ofpH-profiles collected from glycogen phosphorylase (Kasvinsky and Meyer, 1977).In terms of the second assumption stated by Alberty and coworkers and how itrelates to glycogen phosphorylase, it seems very likely that all protonation-deprotonationequilibria are rapid relative to the interconversion of the enzyme ternary complex. Forexample, the transfer of a proton from a stronger acid to the base of a weaker acid isgenerally thought to be a diffusion-controlled process with a bi-molecluar rate constant onthe order of 10 10 M -1 sec -1 in water at 25 °C (Alberty and Bloomfield, 1963). The bi-molecular rate constant for the transfer of a proton from a weaker acid to the basic form of astronger acid may be estimated in the following way (Alberty and Bloomfield, 1963):kiHA + B - ,--z-1 HB + A -k. 1K = (HB) (A -)  = ki/k.1 = KHA/KH B(HA) (B-)KHA and KHB are the acid dissociation constants for HA and HB. If the rate constant forthe proton transfer from HA to B - is taken to be 10 10 M -1 sec -1 and the difference in pKabetween HA and HB is 4 pK units, then the rate constant for proton transfer from theweaker acid BH to A - will be 106 M-1 sec -1 . Since the rate constant for proton transfer isvery large and the concentration of buffer species is also large, it is reasonable to assume,given the rate constant for turnover in glycogen phosphorylase is known to be on the orderof kcat = 100 sec -1 , that all protonation-deprotonation equilibra are rapid compared to theinterconversion of the enzyme ternary complex, and therefore that proton transfer reactionsin glycogen phosphorylase remain at equilibrium.The first assumption by Alberty and coworkers states that there can be nodiversionary pathways in the course of the interconversion of substrate and product. Thisassumption requires that there be only one catalytically active ionization state in the active171site. While very little is known about the nature of the amino acids involved phosphorylasecatalysis, it has been proposed (Klein et al., 1982, 1984; Helmreich, 1992) that essentialacid catalysis is required for bond cleavage of the substrate, however the source of the acidcatalysis is still unclear, and possible diversionary pathways with varying pH cannot becompletely ruled out.1 72REFERENCESAcharya, K. R., Stuart, D. I., Varvill, K. M., and Johnson, L. N. (1991), GlycogenPhosphorylase b, World Scientific Press, Singapore.Alberty, R. A., and Bloomfield, V. (1963), J. Biol. Chem. 233., 2804.Ashwell, G. (1957), Methods Enzymol. 1, 73Avramovic-Zikic, 0., Breidenbach, W. C., and Madsen, N. B. (1974), Can. J. Biochem.52, 146.Baranowski, T., Illingworth, B., Brown, P. H., and Cori, C. F. (1957), Biochim.Biophys. Acta 25_, 16.Barford, D., and Johnson, L. N. (1989), Nature (London) 340, 609.Barford. D., Hu, S-H., and Johnson, L. N. (1990), J. Mol. Biol. 2.11 233.Battell, M. L., Zarkadas, C. G., Smillie, L. B., and Madsen, N. B. (1968), J. Biol.Chem. 243, 6202.Beer, D., Bieri, J. H., Macher, I., Prewo, R., and Vasella, A. (1986), Helvetica ChimicaActa §.2, 1172.Bergmann, E. D., Ginsburg, D., and Pappo, R. (1959), Organic reactions IQ, 179.Bergstrom, D.E., and Shum, P. W. (1988), J. Org . Chem. 51, 3953.Bigge, C. F., Drummond, J. T., and Johnson, G. (1989), Tet. Lett. 30, 7013.Black, T. H. (1983), Aldrichimica Acta 16, 3.Blackburn, G. M., Brown, D. Martin, S. J., and Parratt, M. J. (1987), J. Chem. Soc.Perkin Trans. I, 181.Boeckman, R. K., and Bruza, K. J. (1981), Tetrahedron 31, 3997.Buc, M. H., and Buc, H. (1968), Regul. Enzyme Act. Allosteric Interact., Proc. MeetFed. Eur. Chem. Soc., 4th, 1967, 109.Bunton, C. A., and Humeres, E. (1969), J. Org . Chem. 14, 572.Burton, D. J., Ishihara, T., and Maruta, M. (1982), Chemistry Letters 755.Burton, D. J., and Sprague, L. G. (1989), J. Org . Chem. 54 613.Challoner, R., McDowell, C. A., Stirtan, W., and Withers, S. G. (1993), "A Solid-State31P-NMR Investigation of the Allosteric Transition in Glycogen Phosphorylase."Biophys.J., (In Press).Chang, Y. -C., McCalmont, T., and Graves, D. J. (1983), Biochemistry 22, 4987.173Chang, Y. -C., Scott, R. D., and Graves, D. J. (1986), Biochemistry 25., 1932.Chang, Y. -C., Scott, R. D., and Graves, D. J. (1987), Biochemistry 2., 360.Clark-Still, W., Kahn, M., and Mitra, M. (1977), J. Org . Chem. 42, 2923.Cohn, M., and Cori, G. T. (1948), J. Biol. Chem. 175, 89.Corey, E. J., and Volante, R. P. (1976), J. Amer. Chem. Soc. 9_a, 1291.Cori, C. F., and Cori, G. T. (1936), Proc. Soc. Exp. Biol. Med. 24., 702.Cori, G. T., and Cori, C. F. (1940), J. Biol. Chem. .12„1, 733.Crosby, J., Stone, R., and Leinhard, G. E. (1970), J. Amer. Chem. Soc. 92, 2891.Dreyfus, M., Vandenbunder, B., and Buc, H. (1980), Biochemistry 12, 3634.Dombradi, V. (1981), Mt. J. Biochem.1,1 125.Doudoroff, M., Barker, H. A., and Hassid, W. Z. (1947), J. Biol. Chem. 168, 725.Engers, H. D., Bridger, W. A., and Madsen, N. B. (1969), J. Biol. Chem. 244, 5936.Engers, H. D., Shechosky, S., and Madsen, N. B. (1970a), Can. J. Biochem. la, 746.Engers, H. D., Bridger, W. A., and Madsen, N. B. (1970b), Can. J. Biochem. 4$_, 755.Feldmann, K., and Helmreich, E. J. M. (1976), Biochemistry^2394.Feldmann, K., Zeisel, H., and Helmreich, E. J. M. (1976), Eur. J. Biochem. a, 285.Feldmann, K., and Hull, W. E. (1977), Proc. Natl. Acad. Sci. (USA) 14., 856.Firsov, L. M., Bogachena, T.I., and Bressler, S. E. (1974), Eur. J. Biochem. 41, 605.Fischer, E. H., and Krebs, E. G. (1955), J. Biol. Chem. 216, 121.Fischer, E. H., and Krebs, E. G. (1962), Methods Enzymol. 1, 369.Fischer, E. H., Kent, A. B., Snyder, E., and Krebs, E. G. (1958), J. Amer. Chem. Soc.IQ, 2906.Fischer, E. H., Graves, D. J., Crittenden, E. R. S., and Krebs, E. G. (1959), J. Biol.Chem. 2.3_4, 1698.Friesen, R. W., Sturino, C. F., Daljeet, A. K., and Kolaczewska, A. (1991), J. Org .Chem. 5.2, 1944.Gerig, J. T. (1989), Methods Enzymo1.111, 3.Gold, A. M., Legrand, E., and Snachez, G. R. (1971), J. Biol. Chem. 246, 5700.1 74Goldsmith, E., Sprang, S., and Fletterick, R. (1982), J. Mol. Biol. 156, 411.Gorenstein, D. (1975), J. Amer. Chem. Soc. 2/, 898.Graves, D. J., and Wang, J. H. (1972), Enzymes, (Third Edition) 1 435.Green, A. A., Cori, G. T., and Cori, C. F. (1942), J. Biol. Chem. .142, 447.Helmreich, E. J. M. (1992), BioFactors 3., 159.Hoerl, M., Feldmann, K., Schnackerz, K. D., and Helreich, E. J. M. (1979),Biochemistry la, 2457.Horinishi, N., Tagaya, M., and Fukui, T. (1988), FEBS Len. 221, 426.Hu, H. Y., and Gold, A. M. (1978), Biochim. Biophys. Acta 525., 55.Hullar, T. L. (1969), J. Med. Chem. .12, 58.Illingworth, B. Jansz, H. S., Brown, P. H., and Cori, C. F. (1958), Proc. Natl. Acad.Sci. USA 41, 1180.Jencks, W. P. (1987), Catalysis in Chemistry and Enzymology, Dover Publications Inc.,New York.Johnson, L. N., Madsen, N. B., Mosely, J., and Wilson, R. S. (1974), J. Mol. Biol. 2a,703.Johnson, L. N., Hadju, J., Acharya, K. R., Stuart, D. I., McLaughlin, P. J.,Oikonomakos, N. G., and Barford, D. (1987), in Allosteric Proteins (Herve, G., Ed.),CRC Press.Johnson, L. N., Acharya, K. R., Jordan, M. D., and McLaughlin, P. J. (1990), J. Mol.Biol. 211, 645.Johnson, L. N. (1992), FASEB Journal 2274.Kallen, R. G., Korpela, T., Martell, A.E., Matsushima, Y., Metzler, C. M., Morozov, Y.V., Ralston, I. M., Savin, F. A., Torchinsky, Y. M., and Ueno, H. (1985),Transaminases (Christen, P., and Metzler, D. E., Eds.), p. 37, John Whey and Sons,New York.Kastenschmidt, L. L., Kastenschmidt, J, and Helmreich, E. J. M. (1968), Biochemistry/.3590.Kasvinsky, P. J., and Madsen, N. B. (1976), J. Biol. Chem. 251, 6852.Kasvinsky, P. J., and Meyer, W. L. (1977), Arch. Biochem. Biophys. all 616.Kasvinsky, P. J., Madsen, N. B., Fletterick, R. J., and Sygusch, J. (1978), J. Biol.Chem. 253, 1290.Klein, H. W., Schlitz, E., and Helmreich, E. J. M. (1981), Cold Spring HarborConference on Cell Proliferation 305.1 75Klein, H. W., Palm, D., and Helmreich, E. J. M. (1982), Biochemistry 22, 6675.Klein, H. W., Im, M. -J., Palm, D., and Helmreich, E. J. M. (1984), Biochemistry a5853.Klein, H. W., Im, M. -J., and Palm, D. (1986), Eur. J. Biochem.51, 107.Knowles, J. R. (1976), CRC Crit. Rev. Biochem. 165.Kokesh, F. C., and Kakuda, Y. (1977), Biochemistry LE, 2467.Korytnyk, W., and Ikawa, M. (1970), Methods Enzymol. 18A, 524.Korytnyk, W. (1986), in Vitamin B6 Pyridoxal Phosphate: Chemical, Biochemical, andMedical Aspects, Part A, (Dolphin, D., Poulson, R., Avramovic, 0., Eds.) p. 335, JohnWiley and Sons, Inc. USA.Lamer, J., Villar-Palasi, C., and Rechman, D. J. (1960), Arch. Biochem. Biophys. ,56.Leatherbarrow, R. J. (1990), GraF it Version 2 .0, Erithacus Software Ltd., Staines, U. K.Lesimple, P. Beau, J-M., Jaurand, G., and Sinay, P. (1986), Tet. Lett. 22, 6201.Lineweaver, M., and Burke, D. (1934), J. Amer. Chem. Soc. hf, 1965.Lipscomb, W. N., Hartsuck, J. A., Quiocho, F. A., and Reeke, G. N. (1969), Proc. Nat.Acad. Sci. USA 64, 28.Madsen, N. B., and Cori, C. F. (1956), J. Biol. Chem. 223, 1055.Madsen, N. B., and Cori, C. F. (1957), J. Biol. Chem. 224, 889.Madsen, N. B., and Cori, C. F. (1958), J. Biol. Chem. 233, 1251.Madsen, N. B., and Withers, S. G. (1984), in Chemical and Biological Aspects of VitaminB6 Catalysis: Part A. p. 117, A. R. Liss, Inc. New York.Madsen, N. B., and Withers, S. G. (1986), in Vitamin B6 Pyridoxal Phosphate: Chemical,Biochemical, and Medical Aspects, Part B, (Dolphin, D., Poulson, R., Avramovic, 0.,Eds.) p. 355, John Wiley and Sons, Inc. USA.Manners, D. J. (1957), Adv. Carb. Chem. f2, 261.Martin, J. L., Johnson, L. N., and Withers, S. G. (1990), Biochemistry 22, 10745.McLaughlin, P. J., Stuart, D. I., Klein, H. W., Oikonomakos, N. G., and Johnson, L.N. (1984), Biochemistry 23, 5862.Middleton, W. J. (1975), J. Org . Chem. 411, 574,Murray, A. W., and Atkinson, M. R. (1968), Biochemistry 2, 4023.176O'Connor, J. V., Nunez, H. A., and Barker, R. (1979), Biochemistry .13,, 500.Oikonomakos, N. G., Johnson, L. N., Acharya, K. R., Stuart, D. I., Barford, D., Hadju,J., Varvill, K. M., Melpidou, A. E., Papageorgiou, T., Graves, D. J., and Palm, D.(1987), Biochemistry 2.6, 8381.Oikonomakos, N. G., Acharya, K. R., and Johnson, L. N. (1991), in Post-TranslationalModification of Proteins (Crabbe, J., and Harding, J., Eds.) p. 81, CRC Press, BocaRaton, Florida.Obayashi, M., Ito, E., Matsui, K., and Kondo, K. (1982), Tet. Lett. 22, 2323.Palm, D., Klein, H. W., Schinzel, R., Buehner, M., and Helmreich, E. J. M. (1990),Biochemistry 22, 1099.Parrish, R. F., Uring, R. J., and Graves, D. J. (1977), Biochemistry .1.k, 4824.Percival, M. D., and Withers, S. G. (1992), Biochemistry a, 505.Pfeuffer, T., Ehrlich, J. and Helmreich, E. (1972), Biochemistry 11, 2125.Schinzel, R., Palm, D., and Schnackerz, K. D. (1992), Biochemistry 21., 4128.Schnackerz, K. D., and Feldmann, K. (1980), Biochem. Biophys. Res. Comm. 2,5, 1832.Schnackerz, K. D. (1986), in Vitamin B6 Pyridoxal Phosphate: Chemical, Biochemical,and Medical Aspects, Part A, (Dolphin, D., Poulson, R., Avramovic, 0., Eds.) p. 245,John Wiley and Sons, Inc. USA.Schnackerz, K. D., Wahler, G., Vincent, M. G., and Jansonius, J. N. (1989), Eur. J.Biochem. 185, 525.Shaltiel, S., Hedrick, J. L., and Fischer, E. H. (1966), Biochemistry 5, 2108.Shenhav, H., Rappoport, Z., and Patai, S. (1970), J. Chem. Soc. B 469.Shimomura, S., and Fukui, T. (1978), Biochemistry 11, 5359.Sidgwick, F. R. S., (1966), The Organic Chemistry of Nitrogen (Millar, I. T., Springall,H. D., Eds.), Clarendon Press, Oxford.Sprang, S. R., and Fletterick, R. J. (1979), J. Mol. Biol. 131, 523.Sprang, S. R., Goldsmith, E. J., Fletterick, R. J., Withers, S. G., and Madsen, N. B.(1982), Biochemistry 21, 5364.Sprang, S. R., Acharya, K. R., Goldsmith, E. J., Stuart, D. I., Varvill, K. M., Fletterick,R. J., Madsen, N. B., and Johnson, L. N. (1988), Nature (London) Ilk, 215.Sprang, S.R., Withers, S.G., Goldsmith, E. J., Fletterick, R. J., and Madsen, N. B.(1991), Science 254, 1367.Sprang, S. R., Madsen, N. B., and Withers, S. G. (1992), Protein Science I, 1100.177Street, I. P., Armstrong, C. R., and Withers, S. G. (1986), Biochemistry 21, 6021.Street, I. P. (1988), Ph. D. Thesis, University of British Columbia.Street, I. P., Rupitz, K., and Withers, S. G. (1989), Biochemistry za, 1581.Sykes, B. D., and Hull, W. E. (1978), Methods Enzymol. 42, 270.Tagaya, M., and Fukui, T. (1984), J. Biol. Chem. 252, 4860.Takagi, M. Fukui, T., and Shimomura, S. (1982), Proc. Natl. Acad. Sci. USA. 22, 3716.Takagi, S. Kobayashi, M., and Matsuda, K. (1989), J. Biochem. 105, 933.Tu, J. I., Jacobson, G. R., and Graves, D. J. (1971), Biochemistry _a 1229.Un. S., and Klein, M. P. (1989), J. Amer. Chem. Soc. 111, 5119.Voet, J. G., and Abeles, R. H. (1970), J. Biol. Chem. 241, 1020.Vidgoff, J. M., Pocker, A., Hullar, T. L. and Fischer, E. H. (1974), Biochem. Biophys.Res. Comm. 51, 1166.Wada, H., and Snell, E. E. (1961), J. Biol. Chem. 216., 2089.Wang, J. H., and Graves, D. J.(1964), Biochemistry 2, 1437.Wang, J. H., Shonka, M. L., and Graves, D. J.(1965), Biochemistry 4, 2296.Withers, S. G., Sykes, B. D., Madsen, N. B., and Kasvinsky, P. J. (1979),Biochemistry la, 5342.Withers, S. G., Madsen, N. B., Sykes, B. D., Takagi, M., Shimomura, S., and Fukui,T. (1981a), J. Biol. Chem. 256, 10759.Withers, S. G., Madsen, N. B., and Sykes, B. D. (1981b), Biochemistry 2_Q, 1748.Withers, S. G., Madsen, N. B., Sprang, S. R., and Fletterick, R. J. (1982a),Biochemistry 21, 5372.Withers, S. G., Madsen, N. B., and Sykes, B. D. (1982b), Biochemistry 21, 6716.Withers, S. G., Sheckosky, S., and Madsen, N. B. (1982c), Biochem. Biophys. Res.Comm. la, 322.Withers, S. G., Madsen, N. B., and Sykes, B. D. (1985), Biophys. J. 4$., 1019.Withers, S. G., Street, I. P., and Percival, M. D. (1988), in Fluorinated Carbohydrates:Chemical and Biochemical Aspects (Taylor, N. F. Ed.) p. 59, American Chemical Society,Washington, D. C.Withers, S. G., and Rupitz, K. (1990), Biochemistry 22, 6405.Withers, S. G. (1990), Carbohydr. Res. 197, 61.178Yang, Z. -Y., and Burton, D. J. (1991), Tet. Lett. 32, 1019.


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items