MECHANISM OF THE REACTION CATALYZED BY THE OXALATE DECARBOXYLASE FROM Bacillus subtilis By DRAÂŽENKA SVEDRUÂŽI A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2005
Copyright 2005 by DraÂženka SvedruÂži
This thesis is dedicated to my br other ÂŽeljko, my parents and Chris. Ova je teza posve ena mome bratu ÂŽeljku, mojim roditeljima i Krisu.
iv ACKNOWLEDGMENTS This study was supported by grants from the National Institutes of Health (DK61666 and DK53556) and by the University of Florida Chemistry Department. Partial funding was also receiv ed from Dr. Ammon B. Peck. Thanks go to my doctoral dissertation committee: Dr. Steven A. Benner, Dr. Ammon B. Peck, Dr. Michael J. Scott, Dr. J on D. Stewart, and especially my advisor, Dr. Nigel G. J. Richards, for making this re search project a reality, but also for his constant support and guidance. Dr. Laurie A. Renhardt, Yang Liu a nd Dr. Wallace W. Cleland I thank for fruitful collaboration on hea vy atom isotope effects resear ch. Also, I thank Dr. Wallace W. Cleland for hospitality during my stay in Madison, Wisconsin. Thanks go to my EPR collaborators Dr . Lee Walker, Dr. Andrzej Ozarowski and Dr. Alexander Angerhofer. I am grateful to my coworkers and fri ends in the Richards research group, especially Dr. Christopher H. Chang for proofreading, countless discussions, valuable insights, guidance and support. Sp ecial thanks go to Stefan Jonsson for all his help, Sue Abbatiello for her mass spectrometry efforts and Lukas Koroniak for help with NMR experiments. Also thanks go to all members of Richards group, esp ecially Mihai, Jemy and Cory, for providing a pleasant and supporting environment in and out of lab. I thank Dr. James A. Deyrup, Graduate Coordinator, and Lori Clark in the Graduate Student Office for their help and advice.
v Finally, I thank my mother, father and my brother for all their support and their sacrifice that enabled me to get high educat ion. Also, thanks go to my brother ÂŽeljko for guidance and inexhaustible inspiration.
vi TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES...........................................................................................................viii LIST OF FIGURES...........................................................................................................ix ABSTRACT......................................................................................................................x ii CHAPTER 1 INTRODUCTION........................................................................................................1 Biological Significance of Oxalic Acid........................................................................1 Oxalate-Degrading Enzymes........................................................................................3 Oxalate Decarboxylase.................................................................................................4 Biological Distribution..........................................................................................4 Induction and Biological Roles.............................................................................5 Enzyme Structure: OXDCase Belongs to Bicupin Structural Family...................6 Enzyme Â–Manganese Interaction..........................................................................9 Enzyme-Oxygen Interaction and Production of Peroxide...................................12 Chemical Degradation of Oxalate..............................................................................14 Research Objective.....................................................................................................15 2 EXPRESSION AND PURIFICATION OF RECOMBINANT B. subtilis OXALATE DECARBOXYLASE..............................................................................16 Introduction.................................................................................................................16 Results and Discussion...............................................................................................17 Experimental Section..................................................................................................21 3 STEADY STATE CHARACTERISATI ON OF THE REAC TION CATALYZED BY B. subtilis OXALATE DECARBOXYLASE......................................................26 Introduction.................................................................................................................26 Results and Disscusion...............................................................................................27 Progress Curve and Michaelis-Menten Equation................................................27 Activation, Inhibition and Substrate Specificity.................................................30 The pH Dependence on OxdC Catalysis.............................................................33
vii Formate Production: Irreversible or Reversible Process.....................................36 Experimental Section..................................................................................................37 4 HEAVY ATOM ISOTOPE EFFECTS ON THE REACTION CATALYZED BY B. subtilis OXALATE DECARBOXYLASE.............................................................39 Introduction.................................................................................................................39 Results and Discussion...............................................................................................41 13C and 18O Isotope Effects.................................................................................41 Kinetic Mechanism for the Enzymatic Reaction of OxdC..................................43 Fractionation Factors...........................................................................................44 Molecular Mechanism of the Reaction Catalyzed by OxdC...............................50 Experimental Section..................................................................................................53 5 ROLES OF THE ACTIVE SITE ARGI NINE AND GLUTAMATE RESIDUES....57 Introduction.................................................................................................................57 Results and Discussion...............................................................................................61 Mutational Effects on OxdC Structure................................................................61 Steady-State Characterization.............................................................................62 Oxalate Oxidase and Dye Oxidation Activities...................................................65 13C and 18O Kinetic Isotope Effects.....................................................................68 Experimental Section..................................................................................................72 6 IDENTIFICATION AND STUDIES OF TYROSYL RADICAL FORMED DURING TURNOVER..............................................................................................74 Introduction.................................................................................................................74 Results and Discussion...............................................................................................76 Variation of Radical Intens ity as a Function of Time.........................................76 Effect of Temperature and Micr owave Power on Radical Signal.......................79 Identification of a Tyrosyl Radical......................................................................83 In Search for Tyrosine Residues Forming the Radical........................................83 Spin-trapping experiments...................................................................................86 Potential Functional Implicati ons of the Observed Radical................................87 Experimental Section..................................................................................................89 APPENDIX A PROTEIN PURIFICATION CHROMATOGRAMS.................................................94 B DERIVATION OF EQUATIONS FOR HEAVY ATOM ISOTOPE EFFECT........97 C EPR SPIN-TRAPPING OF A RADICAL DURING OXALATE DECARBOXYLASE TURNOVER.........................................................................100 BIOGRAPHICAL SKETCH...........................................................................................109
viii LIST OF TABLES Table page 2-1 Purification table for a typical OxdC preparation....................................................19 3-1 Specific activities (SA) a nd Michaelis-Menten constants ( KM) for oxalate decarboxylation cataly zed by native and recombinant OXDCase...........................29 3-2 Specific activities and kinetic constants: kcat, KM and kcat/ KM for oxalate decarboxylation catalyzed by recombinant OxdC...................................................30 3-3 Results of the deuterium exchange experiments......................................................37 4-1 13C and 18O isotope effects (IE) on the reaction catalyzed by Bacillus subtilis .OxdC............................................................................................................41 5-2 Molecular weights (MW) of OxdC in solution and manganese incorporation per momomer.................................................................................................................62 5-3 Steady state characterizati on of Arg and Glu mutants.............................................64 5-4 Dye oxidation and oxalate oxidase ac tivities for the wild type and OxdC mutants.....................................................................................................................66 5-5 13C and 18O isotope effects on the reaction cat alyzed by OxdC wild type and mutants.....................................................................................................................69 5-6 Kinetic constants for reactions catalyzed by OxdC mutants derived based on 13C and 18O isotope effects.............................................................................................70 5-7 Comparison of the ( V/K )/[ E ]0k1 calculated based on IE data and (V/K)exp determined based on activity assays.........................................................................71 6-1 Results of kinetic and EPR characteriza tion of some of the Tyr and Trp mutants .86
ix LIST OF FIGURES Figure page 1-1 Classes of enzyme that catalyze the de gradation of oxalate. Found in (A) mostly plants, (B) mostly fungi and so il bacteria, and (C) bacteria.......................................3 1-2 Sequence alignment of fungal and bacter ial oxalate decarboxylases and barley oxalate oxidase showing diagnostic motifs for cupin superfamily HXHX3,4EX6G and GX5PXGX2HX3N................................................................................................7 1-3 High-resolution cr ystal structure of Hordeum vulgare oxalate oxidase monomer (A) and hexamer (B)..................................................................................................8 1-4 High-resolution cr ystal structure of Bacillus subtilis oxalate decarboxylase (OxdC) monomer (A) and trimer (B).........................................................................9 1-5 Comparison of metal binding sites in the three Mn-dependent enzymes. (A) B. subtilis oxalate decarboxylase, (B) barley oxalate oxidase and (C) human Mndependent superoxide dismutase..............................................................................10 1-6 EPR spectrum of the resting OxdC..........................................................................13 1-7 Chemical degradation of oxalate..............................................................................14 2-1 Sodium dodecyl sulfate-polyacrylamide gel (12% polyacrylamide) showing the expression of OxdC in E. coli and purified OxdC with the specific activity of 60 U/mg.........................................................................................................................18 2-2 Multimerization state determination by size-exclusion chromatography................21 3-1 Progress curve for formate production.....................................................................28 3-2 Michaelis-Menten curve for OxdC catalyzed decarboxylation................................29 3-3 Effect of o -PDA on Vmax and Vmax/ KM for oxalate decarboxylation by OxdC.........32 3-4 Inhibition of OxdC catalyzed oxa late decarboxylation by malonic acid.................33 3-5 pH effect on Vmax......................................................................................................35 3-6 OxdC deactivation by oxalate at lower pH..............................................................35
x 3-7 pH deffect on Vmax/ KM..............................................................................................36 3-8 Hypothetical exchange of hydrogen and deuterium when formate is incubated with D2O...................................................................................................................36 4-1 Minimal kinetic mechanism up to first irreversible step for OxdC catalyzed reaction consistent with 13C and 18O isotope effects................................................43 4-2 Fractionation factors of oxalate car bons and oxygens as a function of C-O bond order.........................................................................................................................4 5 4-3 Structure of the putative oxalyl radica l intermediate in the OxdC catalyzed reaction.....................................................................................................................48 4-4 Mechanism of the OxdC-catalyzed C-C bond cleavage consistent with IE and EPR experiments......................................................................................................52 4-5 Schematic diagram of the 13C and 18O isotope effect experiments..........................55 5-1 Comparison of the manganese-binding sites in oxalate oxidase and oxalate decarboxylase...........................................................................................................58 5-2 Effect of pH on log( Vmax) for the wild type and E162Q...........................................65 5-3 Coupled assay used to measure OXOXase activities...............................................66 5-4 Progress curve following oxalate de carboxylase activity and oxalate oxidase activity .....................................................................................................................6 8 6-1 EPR spectrum of O xdC frozen 30 seconds af ter oxalate addition...........................75 6-2 The steady-state OxdC radical inte nsity as a function of reaction time...................77 6-3 Kinetic model for the steady-state ra dical observed during OxdC catalysis............78 6-4 Effects of temperature and isotopic editing on the observed radical spectrum in OxdC........................................................................................................................80 6-5 Effects of temperature on the observed radical spectrum in OxdC..........................81 6-6 Power saturation of the g = 2.0 signal formed by OxdC in the presence of oxalate and dioxygen................................................................................................82 6-7 Optical density indicating growth of E. coli cells in time........................................84 6-8 OxdC monomer showing di stribution of tyrosine a nd tryptophan residues.............85 A-1 DEAE-Sepharose Fast Flow chromatograms...........................................................94
xi A-2 phenyl-Sepharose Hi-Performance chromatograms.................................................95 A-3 Q-Sepharose Hi-Performance chromatogr am. Arrow points out to the peak that showed OxdC activity and was co llected and purified further................................96 C-2 EPR spectra showing that trapped radical is not tyrosyl radical............................100 C-1 EPR spectra of the spin-traped ra dical formed during OxdC turnower.................101 C-3 Rate of formate production versus co ncentration of DMPO present during OxdC catalysed oxalate deacrboxylation..........................................................................102
xii Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy MECHANISM OF THE REACTION CATALYZED BY THE OXALATE DECARBOXYLASE FROM Bacillus subtilis By DraÂženka SvedruÂži December 2005 Chair: Nigel G. J. Richards Major Department: Chemistry The oxalate decarboxylase from Bacillus subtilis (OxdC) catalyzes the mechanistically intriguing reac tion in which the C-C bond in oxalate is cleaved to give carbon dioxide and formate. Crystallographic structures have shown that OxdC is a member of the Â“bicupinÂ” structural family with a mononuclear manganese center in each cupin domain, and amino acid liga nds that are structurally similar to those from oxalate oxidase, another oxalate-degrading enzyme. While oxalate oxidase catalyzes a redox reaction in which oxalate is oxidized and mo lecular oxygen is reduced to peroxide, OxdC catalyzes a non-redox reaction that is uniquely dependent on molecular oxygen, which is not consumed in the reaction. In this study OxdC was overexpressed from E. coli with full manganese occupancy and a specific activity of 79 I.U./mg, equal to that obtained for the native enzyme. Oxalate 13C and 18O heavy atom isotope effects on Vmax/KM suggest the presence of a partially rate-determining step prior to the irreversible C-C bond cleavage, which was
xiii proposed to be proton-coupled electron transfer. From the pH studies it was determined that monoprotonated oxalate is the true substrate for OxdC. Mutagenesis studies of ac tive site arginine (R92 a nd R270) and glutamate (E162 and E333) residues and heavy atom isotope effects on the reaction catalyzed by the mutants suggested that (1) the N-terminal ma nganese binding site hos ts the catalytically active site, (2) Glu162 is the key residue determ ining reaction specificity and has a role of acid/base catalyst in the O xdC mechanism and (3) the role of Arg92 is to orient the oxalate molecule upon binding, polarize a C-O bo nd in oxalate and th erefore facilitate decarboxylation. A narrow EPR signal observed in the presen ce of enzyme, oxalate and dioxygen is identified as a protein-based tyrosyl radical. Kinetic studies have s hown that the observed radical is not kinetically competent. Furthe r characterization of the radical EPR signal and mutagenesis of conserved tyrosyl residue s in the proximity of the Mn-binding sites suggest that the tyrosyl radica l is formed more than 10 Ã… from the Mn-binding site and in a position inaccessible to solvent.
1 CHAPTER 1 INTRODUCTION Biological Significance of Oxalic Acid Oxalic acid and oxalate salts are widely di stributed in nature and are involved in many geochemical and biochemical processes in soils. Organisms can produce oxalate on primary or secondary metabolic pathways. General reviews on oxalate metabolism have been written previously by Li bert and Franceschi (1987), Fr anceschi and Horner (1980) and Hodgkinson (1977). A heterogeneous bact erial taxonomic group capable of using oxalate as a sole carbon and energy source, known as Â“oxalotropic,Â” has been reviewed recently by Sahin (2003). Many fungi belonging to Ascomycetes , Basidiomycetes and Zygomycetes produce calcium oxalate crys tals in some phase of th eir life cycle or excrete oxalic acid under specific growth conditions. Oxalate secretion by fungi provides advantages for their growth and colonization of substrates and is associated with plant pathogenesis (Dutton and Evans, 1996). Brown -rot wood decay fungi use oxalic acid as a proton source in enzymatic and non-enzyma tic hydrolysis of carbohydrates and as a metal chelator. In white-rot wood decay f ungi oxalic acid has multiple roles in redox chemistry: it is a source of formate radica ls that reduce dioxygen and ferric ion to superoxide anion radical and ferrous ion, respectively; a nd, it is directly oxidized in a process that stimulates lignin peroxidase (LiP) and manganese peroxidase (MnP) (Shimada et al., 1997). Although it is often a ssociated with plant pathogens (the most common being Sclerotinia sclerotiorum ) (Loewus et al., 1995), oxalic acid is also produced by many plants themselves. In lignin synthesis oxalate is a substrate of a Mn-
2 dependent enzyme, oxalate oxidase (see text below), which provides H2O2 required for cell wall cross-linking (Requena and Born emann, 1999). Interestingly, oxalate and manganese ions are involved in both lignin synthesis in plants a nd lignin degradation by fungi. Humans and animals are exposed to oxalate through diet by consuming oxalate directly, or through compounds that can be c onverted to oxalate in the body such as ascorbate, xylitol, or ethylene glycol (usually accidentally). Infection by Aspergillus species (aspergillosis) can also result in an excessive accumulation of oxalate in the body, a condition referred to as oxa losis (Dunwell et al., 2000). Dietary oxalate can be catabolized in mammalian organisms only through the action of oxalotropic bacter ia. A prominent example is Oxalobacter formigenes , which was found to colonize the intestines of both rats and humans (Allison et al., 1985). Hyperoxaluria in humans and animals, which might be due to absence of Oxalobacter formigenes from gut flora, causes several pathological conditions (Williams and Wandzilak, 1989), the most common one in humans being urolithiasis (formation of stones in the kidneys, which are predominantly composed of calcium oxalate). In addition to this relatively common condition caused by peripheral oxalate, there is a less common but much more severe genetic condition known as primary hyperoxaluria, which leads to deposition of insoluble oxalate salt s throughout the body (oxalosis). Primary hyperoxaluria appears in two forms: type 1 is caused by misfunction of the liver enzymes alanine:glyoxylate aminotransferase and serine pyruvate aminotransferase; and type 2 is caused by a defect of glyoxalate reductase and D-glycerate dehydrogenase, enzymes from hepatocytes and leukocytes (De Pauw and Toussinet, 1996).
3 Oxalate-Degrading Enzymes Three known types of enzymes have evol ved in nature that degrade oxalate: oxalate oxidase (OXOXase), characteristic most ly for plants (Kotsira and Clonis, 1997); oxalate decarboxylase (OXDCase), characteri stic primarily for fungi (Shimazono, 1955), but also found in soil bacteria (Tanner a nd Bornemann, 2000); and bacterial oxalyl-CoA decarboxylase (Quayle, 1963) (Figure 1-1). In the case of oxalyl-CoA decarboxylase, oxalate is derivatized to the coenzyme Athioester prior to de carboxylation, and it is decarboxylated employing thiamin pyrophosphate-dependent chemistry. Although the main focus of the research presented in th is dissertation concer ns OXDCase, different aspects of both OXOXase and OXDCase will be compared and discussed in more detail throughout the text, since they share many stru ctural and mechanistic features. As the oxalyl-CoA decarboxylase utilizes a very diffe rent and largely unrel ated mechanism, it will not be discussed further. O (H)O O O (A) O (H)O O O (B) O H O O -O SCoA O (C) O H SCoA H+ Oxalate Oxidase O2, H+2CO2 + H2O2Oxalate Decarboxylase "cat. O2" CO2 + Oxalyl-CoA Decarboxylase CO2 + Figure 1-1. Classes of enzyme that cataly ze the degradation of oxalate. Found in (A) mostly plants, (B) mostly fungi a nd soil bacteria, and (C) bacteria. Oxalate degrading enzymes have either actual or potential commercial application in many areas of medicine, ag riculture and industry (Dunwel l et al., 2000). Related to
4 human and animal health, these enzymes show great promise in the prevention, diagnosis and treatment of different types of oxalosis and hyperoxaluria. Oxalate Decarboxylase In 1955, Shimazono was the first to defi ne Â“oxalic acid-decarboxylaseÂ” as an enzyme that catalyzes non-oxidative decarboxy lation of oxalic acid to formic acid and CO2. OXDCase employs manganese and is in active under anaerobic conditions even though decarboxylation is a redox -neutral reaction. During tu rnover molecular oxygen is reduced to peroxide only s ub-stoichiometrically, which suggests a potential role of reduced dioxygen species in the decarboxylation. Biological Distribution OXDCase was first isolated from the basidiomycete fungi Collyvia velutipes ( Flammulina velutipes ) and Coriolus hersutus (Shimazono, 1955) and has subsequently been identified in other white-rot fungi: Sclerotinia sclerotiorum (Magro et al., 1988), Coriolus ( Trametes ) versicolor (Dutton et al., 1994), Dichomitus squalens , Phanerochaete sanguinea and Trametes ochracea (MÃ¤kelÃ¤ et al., 2002). The filamentous fungi Myrothecium verrucari (Lilehoj and Smith, 1965), certain strains of deuteromycete Aspergillus niger (Emiliani and Bekes, 1964; Emiliani and Riera, 1968) and the common button mushroom Agaricus bisporus (Kathiara et al ., 2000) are also known to host this enzyme. In spite of the common belief that OXD Case activity is related only to the whiterot wood decaying fungi, activity was also discovered in the liquid cultures and mycelial extract of Postia placenta , a brown-rot fungus (Micales , 1997). While fungal enzymes have been known in the scientific literature for a relatively long time, bacterial OXDCases were discovered only recentl y. It was proposed, based on sequence information and alignments with fungal OXDCases, that genes from microorganisms
5 such as Bacillus subtilis, Streptococcus mutans and Synechocystis might encode an enzyme with oxalate decarboxylase activity (Dunwell and Gane, 1998; Dunwell et al., 2000). Subsequently, it was conf irmed experimentally that B. subtilis expresses cytosolic OXDCase (Tanner and Bornemann, 2000). Next, B . subtilis enzymes encoded by genes yvrK and yoaN (referred to in newer literature by their more functional designations oxdC and oxdD , respectively) were over-expressed in E. coli and the activity of the recombinant enzymes confirmed (Tanner et al., 2001). More recently, a structural genomics study of proteins encoded by putative open reading frames in Thermatoga maritima has identified an oxalate-binding monoc upin as a putative bacterial oxalate decarboxylase (Schwarzenbacher et al., 2004). To date, OXDCase from C. velutipes, A. niger and B. subtilis are the best characterized. Induction and Biological Roles Expression of OXDCase can be induced by addition of substrate or other carboxylic acids (glycolic, malonic or citric) to the culture medium and/or by lowering culture medium pH (Shimazono and Haya ishy, 1957; Lilehoj and Smith, 1965; Mehta and Datta, 1991; Micalles, 1997, Tanner and Bo rnemann, 2000; Azam et al, 2002). It was shown that OXDCase from A. bisporus (Kathiara et al., 2000) wa s not inducible by either pH or oxalic acid. OXDCase from P. placenta (brown-rot fungus) requires 40 times higher concentration of oxalate (100-200 mM) for induction than the OXDCase from C. versicolor (white-rot fungus) (Micales, 1997). High oxalate requirements for the induction may explain why ju st one OXDCase-expressing br own-rot fungus has been discovered so far. It was proposed that di fferential regulation of OXDCase expression might imply different roles this enzyme has in different organisms (Tanner and Bornemann, 2000; Azam et al., 2001). Acid-i nduced OXDCases might have a role in
6 proton consumption, while oxalate-induced ones might protect an organism from harmful metabolic effects of oxalate. Interestingly, ev en though oxalate stimulat es activities of the lignin-degrading enzymes LiP and MnP, at higher concentrations it acts as a noncompetitive inhibitor (Shimada et al., 1997) . In this prospect the role of OXDCase expressed by wood-rotting fungi might also be to control oxalat e concentration and therefore indirectly pa rticipate in the wood degradation process. Fungal OXDCase can be both an intracellular enzyme or excreted outside the hypha l cells (Dutton et al., 1994; Micales, 1997; MÃ¤kelÃ¤ et al., 2002). A proteomic-based study showed that both OXDCases from B. subtilis (OxdC and OxdD) are spore-asso ciated proteins (Kuwan et al., 2002). OxdC is a cytosolic protein, indu ced in acidic conditions (optimum pH = 5) but not by oxalate, which might imply its role in proton consumption within the cytoplasm (Tanner and Bornemann, 2000). OxdD is specifically associated with the B. subtillis spore coat structure (Costa et al., 2004). Enzyme Structure: OXDCase Belong s to Bicupin Structural Family Sequence alignment (Dunwell and Gane , 1998; Dunwell et al., 2000), homology modeling (Chakraborty et al., 2002 ) and crystal structures (Ana nd et al., 2002; Just et al., 2004) established common structural featur es of fungal and bacterial OXDCases. All studied OXDCases are found to be bicupi ns (Figure 1-4, A), while OXOXases are monocupins (Figure 1-3, A), which are both su b-classes of the cupin structural family. Only recently the first bicupin OXOXase wa s discovered from dikaryotic white rot fungus Ceriporiopsis subvermis pora (Escutia et al., 2005). The structural scaffold that characterizes a member of the cupin superfamily of proteins is that of a conserved -barrel fold ( cupa is a Latin term for a small barrel) containing two sequence motifs: GX5HXHX3,4EX6G and GX5PXGX2HX3N (Dunwell et
7 al., 2000) (Figure 1-2). The cupins are one of the most functionally diverse protein superfamilies due to the ancient nature of this protein fold (Anantharaman et al., 2003). A characteristic feature of cupins is their high thermal stabilit y and resistance to proteases owing to a tightly packed structure, la rge number of subunit contacts, hydrogen bonds, hydrophobic interactions and shor t loops (Dunwell et al., 2000) . The active site of an enzymatic cupin is usually lo cated in the center of the -barrel with a metal ion coordinated by the residues from both diagnosti c motifs. Absence or presence of different metals such as Ni, Fe, Zn, Mn, or Cu contri butes to a functional diversity of the cupin superfamily (Dunwell et al , 2004, Gopal et al., 2005). Domain 1 Motif 1 Motif 2 A.niger OXDCase -G VIREL H W H REA E WAYVLA G--------G DLWYF P S G HP H SLQG-C.velutipes OXDCase -G AIREL H W H KNA E WAYVLK G -------G DLWYF P P G IP H SLQA-B.subtilis OXDCase -G AIREL H W H KEA E WAYMIY G -------G DLWYF P S G LP H SLQG-(H.vulgare OXOXase -G GTNPP H I H PRAT E IGMVMKG------G ETFVI P R G LM H FQFN -) Domain 2 Motif 1 Motif 2 A.niger OXDCase -G AIREM H W H PNAD E WSYFKR G-------G DVGIV P RNMG HFIE N-C.velutipes OXDCase -G ALREL H W H PTED E WTFFIS G ------G DIAYV P ASMG H YVE N-B.subtilis OXDCase -G AMREL H W H PNTH E WQYYIS G ------G DVGYV P FAMG H YVE N-(H.vulgare OXOXase -G GTNPP H I H PRAT E IGMVMK G ------G ETFVI P RGLM H FQF N-) Figure 1-2. Sequence alignment of fungal and ba cterial oxalate decarboxylases and barley oxalate oxidase showing diagnosti c motifs for cupin superfamily HXHX3,4EX6G and GX5PXGX2HX3N. Bold letters denote residues conserved in a motif and blue manganese bindin g residues. OXOXase sequence is in brackets to notify that it has just on e domain, which is aligned with both OXDCase domains. B. subtilis OXDCase is OxdC (not OxdD). It was proposed that bicupins most like ly have evolved from the duplication and then fusion of a single domain ancestor, or in some cases by fusion of two different monocupin precursors (Dunwell et al., 2000, 2001). Bicupins were found to be both microbial (from Archaea, bacteria, fungi) a nd plant-associated proteins. Apart from
8 OXDCase other bicupins known in literature include several dioxygenases, seed storage globulins and sucrose binding prot eins (Dunwell et al., 2004). To illustrate characteristic cupin motifs and conserved metal binding ligands, part of the sequence alignment of the fungal and ba cterial OXDCases and barley OXOXase is given in Figure 1-2. From sequence alignments it is evident that the OXOXase is more similar to OXDCase domain 2 (Dunwell et al ., 2000). Both enzymes have manganese ion bound in the middle of each cupin domain. Crystal structures of OXOXase (Woo et al., 2000) and OxdC (Anand et al., 2002; Just et al., 2004) showed that both enzymes crystallize as a hexamers. Hordeum vulgare OXOXase monomers (Figure 1-3A) appear to form dimers with an overall quaternary structure best described as a Â“tri mer of dimersÂ” (Figure 1-3B). The B. subtilis OxdC protein crystallizes as a hexamer made up of two trimeric layers packed face-to-face with D3 point pseudosymmetry (Figure 1-4B). Figure 1-3. High-resolution crystal structure of Hordeum vulgare oxalate oxidase monomer (A) and hexamer (B). Ribbon re presentation of the monomers is showing the secondary structural elemen ts and the location of the Mn-binding site. Helices and strands are colored yellow and red, respectively; manganese is represented as a silver sphere. Mo nomers in a quaternary structure are shown in different colors in a ribbon representation. Images were prepared using the CAChe Worksystem Pro V 6.0 software package (CAChe Group, Fujitsu, Beaverton OR) B A
9 Enzyme Â–Manganese Interaction Metal analysis studies (T anner et al., 2001; Chakra borty et al., 2002), EPR and UV-visible spectroscopy (Tanner et al., 2001; Chang et al., 20 04) and crystal structure determination (Anand et al., 2002; Just et al., 2004) have amply demonstrated that OXDCase is a manganese-containing enzyme. Crystal structure of OxdC and homology modeling of OXDCase from C. veltipes (Chakraborty et al., 2002) showed that each cupin domain has mononuclear manganese coordinated by three histidine and one glutamate residues (Figure 1-5A). Overall ge ometry around the metal ion is octahedral with two exchangeable coordination sites assumed to be filled with aquo or hydroxo ligands in the enzymeÂ’s catalytic resting state. Therefore, in the crystal structure the total number of Mn ions bound to an OxdC hexamer is 12. Figure 1-4. High-resolution crystal structure of Bacillus subtilis oxalate decarboxylase (OxdC) monomer (A) and trimer (B). (O xdC quaternary structure is actually hexameric, only trimer was shown for cl arity). Color-coding as in Figure 1-3. Images were prepared using the CA Che Worksystem Pro V6.0 software package (CAChe Group, Fujitsu, Beaverton OR) A B
10 The manganese-binding sites in each domain are almost identical to each other and remarkably similar to Mn-binding site s from OXOXase (Woo et al., 2000) (Figure 15B) and manganese-dependent superoxide dismutase (MnSODase) (Edwards et al., 1998). Chakraborty and coworkers (2002) show ed experimentally for the fungal OXDCase that each of the 6 metal-binding his tidine residues is essential, since His-toAla single variants did not have any dete ctable OXDCase activity. This study does not differentiate the cases in which mutated resi dues and absence of Mn have an affect on enzyme structure, or on substr ate recognition and catalysis. A B C Figure 1-5. Comparison of metal binding sites in the three Mn-dependent enzymes. (A) B. subtilis oxalate decarboxylase, (B) barley oxalate oxidase and (C) human Mn-dependent superoxide dismutase. Bound waters are shown in their protonated form for simplic ity, although it is likely th at these are replaced by other ligands during catalytic turnov er. [Carbon Â– black; nitrogen Â– blue; oxygen Â– red; manganese silver. Hydrogens attached to carbon atoms have been omitted for clarity]. Images were prepared using the CAChe Worksystem Pro V6.0 software pack age (CAChe Group, Fujitsu, Beaverton OR) Together with the knowledge that OXDCase is in fact a manganoenzyme, results reported for fungal OXDCase (Lilehoj a nd Smith, 1965 and Emiliani and Riera, 1968) and bacterial OXDCase (Tanner et al., 2001) show that Mn is tightly bound since extensive dialysis with buffers that were de void of manganese salts did not lead to a detectable decrease in the activity. Shimazono and Hayaishi (1957) also showed that
11 activity of the native OXDCase from C. velutipes was not affected by chelating reagents such as EDTA, , Â’-diphyridyl and p -chloromercuribenzoate. Tanner and coworkers (2001) observed that OxdC and OxdD overexpressed in E. coli are expressed as soluble and active only if cells were exposed to heat treatment (42 ÂºC, 2 min) and Mn2+ salts prior to induction. When other metal cations such as Fe2+, Cu2+, Ni2+, Co2+, Mg2+or Zn2+ were added to a bacterial culture, soluble but cat alytically inactive enzyme was overexpressed. Similar manganese-binding properties, incl uding tight binding, thermally triggered metalation, non-specific metal uptake, and an absolute dependence on Mn incorporation for activity are, also observed experimentally for MnSODase (Whittaker and Whittaker, 1999, 2000; Mizuno et al., 2004). Calorimetric st udies of MnSODase showed that the binding constant for Mn(II) in Mn SODase is surprisingly low (KMn(II)=3.2 x 108 M-1) in light of the essentially irreversible metal binding. This indicates that metal binding and release are dominated by kinetic rather th an thermodynamic constraints (Mizuno et al., 2004). The kinetic stability of the metal-prot ein complex is due to stability of the elements of protein structure and is indepe ndent of the presence or absence of metal, which is reflected in thermally trigged metala tion characteristic of MnSODase proteins. Considering the structurally similar Mn-b inding site, and similar properties observed experimentally, this model obtained for Mn SODase could be hypothesized for OXDCase and OXOXase metal centers in the absence of detailed study. It was proposed for OxdC that the Mn oxi dation state alternates during catalysis, between Mn(II) and Mn(III) (Taner et al. 2001) or Mn(III) and Mn (IV) (Anand et al., 2002). However, only a Mn(II) state was detect ed experimentally by either standard perpendicular-mode (sensitive to half-integer-spin KramersÂ’ systems) or parallel-mode
12 (sensitive to integer-spin non-KramersÂ’ sy stems) EPR spectroscopy for the resting enzyme (Tanner et al., 2001; Chang et al., 2004) or during turnover (Chang et al., 2004). The X-band spectrum of the resting enzyme in perpendicular mode is shown in Figure 16A, B. The characteristic sextet pattern centered at g = 2.003 comes fr om the hyperfinesplit MS = -1/2 +1/2 central-field transition of Mn(II) (Chang et al ., 2004). Hyperfine splitting (93 G) is consistent with octahedral coordination of Mn, previously observed in the crystal structures. Not surprisingly, the EPR spectrum obtained for barley OXOXase (Whittaker and Whittaker, 2002) was similar to one for OxdC. In both studies the Mn(II) signal was perturbed upon an aerobic oxalate addition, ch anging the nuclear hyperfine splittings in the g = 2 region and increasi ng the modulation dept h of the hyperfine multiplet features. In order to detect presence of Mn in the higher oxidation states during the turnover Chang and co-workers (2004) us ed X-band parallel mode EPR. In this experiments no spectroscopic signature for either Mn(III) or Mn(IV) was observed. Enzyme-Oxygen Interaction and Production of Peroxide It was shown experimentally that fungal (Shimazono and Hayaishi 1957; Emiliani and Bakes, 1964 and Lilehoj and Smith, 1965) and bacterial (Ta nner et al., 2001) OXDCase are inactive under strictly anaer obic conditions and the activity can be recovered by the addition of mol ecular oxygen. The requirements for O2 are specific. Shimazono and Hayaishi (1957) demonstrated that O2 can not be replaced with H2O2 or reducing reagents such as paraquinone, 2-methyl-1,4-napthoquinone, flavin adenine dinucleotide or cytochrome c . Emiliani and Riera (1968) showed that the optimal partial pressure of oxygen (pO2) for A. niger OXDCase activity was around 0.1 atm. They also reported that presence of reagents such as o -phenylenediamine ( o -PDA) shift the optimal pO2 towards higher values. OXDCase in an atmosphere of pure O2 was irreversibly
13 inactivated during turnover. It was found that A. niger OXDCase was sensitive to oxygen only during turnoverÂ—bubbling pure O2 through the enzyme solution for up to an hour prior to oxalate expos ure did not affect CO2 evolution. These results are consistent with the observation that the EPR signal for Mn(I I) has the same profile if obtained in anaerobic or aerobic conditions, and changes onl y by addition of oxalate (Chang et al., 2004). This suggests that if O2 binds directly to the metal, which takes place only after oxalate binding and therefore only enzyme unde rgoing turnover might be sensitive to an excess of O2. Oxygen effects on A. niger OXDCase differ from the effects reported for M. verrucaria OXDCase (Lilehoj and Smith, 1965.) In th is study activity in air was equal to that in a pure oxygen atmosphere. It was stated earlier th at both OXDCase and OXOXase ha ve oxalate as substrate and very similar active sites; it is therefore not surpri sing that OXDCase shows minor OXOXase activity. For OxdC it was determin ed to be 0.2 % of decarboxylase activity (Tanner et al., 2001). Figure 1-6. EPR spectrum of the resting O xdC. (A) Narrow scan continuous-wave Xband spectrum of the central-field mangane se transition in resting-state, aspurified OxdC. (B) Overlay of the full experimental spectrum shown partially in (A) (light line) with a simulation (d ark line) of the broad features arising from a small zero-field splitting. Parameters used were giso = 2.0034, = 9.437713 GHz, D = 0.023 cm-1, and a 100 G Gaussian linewidth; hyperfine interactions were not included in the spin Hamiltonian. (Chang et al., 2004).
14 Chemical Degradation of Oxalate Chemical precedent for direct homolytic C-C bond cleavage in the breakdown of carboxylic acids is provided by the Kolbe reac tion. In such an oxidative decarboxylation, oxalate is broken into two carbon dioxide mol ecules through stepwise radical mechanism (Figure 1-7, iv ). In the first step oxalate can be oxidized electrochemically (Andrieux et al., 2001), photochemically (Buche r et al., 2001), or by high-vale nt metal cations (Pelle et al., 2004) to form an oxalate radical anion (C2O4 . -) (Figure 1-7, i ), which subsequently decarboxylates to give formate radical anion (CO2 . -) (Figure 1-7, ii ). It was proposed based on experimental data that the first electron transfer (ET) and C-C bond cleavage occur by a successive rather than by concerted mechanism and therefore C2O4 . forms as a true intermediate (Kanouf i and Bard, 1999). Finally, a second oxidation gives the second CO2 molecule from CO2 . (Figure 1-7, iii ). Kanoufi and Bard also showed that the rate of the overall reaction is controlled by the first ET. Reaction scheme, standard potentials of individual steps and free en ergy for the fragmentation reaction calculated based on the standard potentia ls are given in Figure 1-7. ( i ) C2O4 2C2O4. -+ eE0 C2O4 . -/ C2O4 2= 1.41 V (Kanoufi and Bard, 1999) ( ii ) C2O4. CO2. + CO2 G0 fragmentation= -0.61 eV (Kanoufi and Bard, 1999) ( iii ) CO2. CO2 + eE0 CO2/ CO2 . -= -1.9 V (Butler and Henglein, 1980) ( iv ) C2O4 22CO2 + 2eE0 2CO2/ C2O4 2= -0.55 V (Galus, 1985) Figure 1-7. Chemical degradation of oxalate. Reaction scheme for oxidative decarboxylation of oxalate and corres ponding standard oxidation potentials vs. NHE or Gibbs free energy: ( i ) first ET, ( ii ) fragmentation reaction, ( iii ) second ET and ( iv ) overall reaction.
15 Research Objective Oxalate degradation is a challenging chemical reaction, which involves highly reactive radical species. It can be expected that enzymes able to catalyze C-C bond fragmentation utilize a redox-act ive cofactor. No signature of an organic cofactor was observed experimentally in the OXDCase or OXOXase structures. Even though the role of manganese and dioxygen in OXDCase is poorly understood, it can be suggested that Mn(II) and O2 act together as an el ectron sink during the decarboxylation. In the case of OXOXase this proposal is evident since mo lecular oxygen is reduced in the overall redox reaction to H2O2. On the other hand, in the OXDCase-catalyzed reaction dioxygen is not consumed, although reaction is dioxygen-depe ndent and presumably involves reduced dioxygen species as intermediate s. Furthermore, density func tional theory ca lculations on a Mn-binding site from OXDCase suggest that the intrinsic reactivity of the Mn-complex is to oxidize oxalate to two CO2 (reaction characteristic for OXOXase), which implies that the protein environment is essentia l to modulate activity of the OXDCase metallocenter (Chang and Richards, 2005). To summarize, the central mechanistic question for OXDCase is how the protein en vironment controls the high reactivity of putative radical intermediates, and therefore the outcome of the reaction. Understanding the OXDCase mechanism would greatly contri bute to our understanding of radical enzymology in general. The main goals of the presented research were: (1) overexpression of high quality B. subtilis OxdC; (2) detailed steady-stat e kinetic characterization of the enzyme reaction; (3) identification of radical species produced during the turnover and determination of their contribution to the kinetic and chem ical mechanism; and, (4) to understand the role(s) of selected residues in the Nand C-terminal manganese-binding sites.
16 CHAPTER 2 EXPRESSION AND PURIFICATION OF RECOMBINANT B. subtilis OXALATE DECARBOXYLASE Introduction The OXDCase gene was cloned from C. velutipes (Kesarwani et al., 2000), A. phoenices (Scelonge and Bidney, 1998) and B. subtilis (Tanner et al., 2001) and the corresponding enzymes from C . velutipes (Azam et al., 2001) and B. subtilis (Tanner et al., 2001, Anand et al., 2002) were subsequently overexpresse d. Specific activities of overexpressed/purified OXDCases differ si gnificantly depending on the expression system and conditions. For example, C. veultipes OXDCase was overexpressed catalytically active only from yeast S. pombe , while the enzyme expressed from another yeast S. cerevisiae (Azam et al., 2001) or E. coli (Azam et al., 2002) was functionally inactive. Bacterial enzymes O xdC and OxdD were overexpressed in an active form in E. coli only if cells were heat-treated (42 C, 2 min) and Mn2+ salts were added to the growth medium prior to induction (Tanner et al., 2001). In general, overexpression of metalloproteins is often a challenging tas k. Within the cell, successful synthesis and assembly of a metalloprotein require specifi c metal delivery, metal recognition and in some cases expression of a metallochaperone, a protein with the role to assist metal incorporation and protein folding. Overexpression of catalyti cally active metalloprotein also depends on the availability of the required metal in a host cell. Concentrations of Mn2+ in bacterial cells might vary over several loga rithms of concentration (up to 1 mM) even when environmental Mn2+ is scarce. A general review on manganese transport in
17 bacteria was written by Kehres and Maguire (2003). Although B. subtilis OxdC overexpressed in E. coli is found to be catalytically activ e only if cells were heat-treated prior to induction, at present there is no clear evid ence of a role for chaperone proteins (expressed upon heat shock in E. coli cells) that are able to assist OxdC folding and manganese incorporation. Initial studies of fungal OXDCase were pe rformed with partially purified enzyme (Shimazono and Hayaishi, 1957; Emiliani and Riera, 1968) or even crude extracts (Lillehoj and Smith, 1965). Detailed mechan istic studies require high amounts of pure OxdC. To simplify the purification procedure, overexpression of B. subtilis OxdC in E. coli was reported with a polyhistidine tag engi neered on Cor N-terminus (Anand et al., 2002; Just et al., 2004). Such enzymes showed lower specific activitie s than native OxdC (Tanner et al., 2001), especially if the His-tag was construc ted on the C-terminus. Lower activity of His-tagged OXDCases might be due to instability of tertia ry structure caused by additional His residues or lo wer incorporation of Mn in the active site as imidazole rings from histidines ligate Mn ions and th erefore reduce concentration of available Mn for binding to the active site(s). Results and Discussion OxdC was overexpressed from E. coli following the procedure from Tanner and coworkers (2001) with a few changes. To fac ilitate expression of specific manganese transporting proteins in E. coli and subsequently higher in tracellular concentration of Mn2+ , the expression protocol was modified such that cells were grown at 30 C to slow growth rate, and overexpressi on induced with isopropyl-D-thiogalactoside (IPTG) close to the stationary growth phase. Approximately half of the OxdC, overexpressed in
18 97.4 66.2 45.0 31.0 21.0 kDa E. coli , was expressed as insoluble inclusion bodies (see Figure 2-1). To obtain higher yields, soluble OxdC was partially recovere d from inclusion bodies by extraction with buffer containing 0.5 M sodium chloride, wh ich was found to enhance OxdC solubility. This is assumed to be a consequence of ge neral ionic strength, rather than a specific effect of sodium or chloride ions on the protein structure. Figure 2-1. Sodium dodecyl sulf ate-polyacrylamide gel (12% polyacrylamide) showing the expression of OxdC in E. coli and purified OxdC with the specific activity of 60 U/mg. Gel was stained by Coomassi e Blue dye. Line 1 shows molecular weight markers, lane 3 crude extract, line 4 soluble proteins extracted from lysis pellets, lines 2 and 4 purified OxdC in concen trations 0.3 and 0.6 mg/ml loaded on a gel respectively. Soluble proteins from crude extract and ex tracted proteins from lysis pellets were combined and purified in a four-step pur ification procedure including two anionexchange and one hydrophobic in teraction chromatography st eps and ammonium sulfate protein precipitation. A representative purific ation table (Table 2-1) is given below. Fractions pooled after the DEAE column (A ppendix A, Figure A-1) were combined and mixed with ammonium sulfate (AS) up to 1.7 M salt concentration. Pr ecipitated proteins are removed by centrifugation, while those rema ining soluble were applied to a phenyl 1. 2. 3.4. 5.
19 Sepharose (PS) column. OxdC activity was found in three distinct peaks that eluted from PS column (see Appendix A, Figure A-2). A lthough all three peaks containing OxdC showed some activity (relati ve specific activities 1, 0.5 and 0.1 respectively), only the fraction with the highest specific activity and the least affinity for hydrophobic resin was pooled and purified further. The other two peaks may comprise OxdC partially folded and/or in a lower multimerisation state (mono mer or trimer), since according to the chromatogram they bind to a hydrophobic resi n more tightly. The AS precipitation and PS chromatography fractionation steps cause some loss of total OxDc, but both steps were necessary to obtain high quality prot ein and high specific activities of purified OxdC. Although OxdC was purified at room te mperature, overall OxdC activity did not decrease significantly from the beginning to th e end of purification (a ctivity of lost OxdC in AS and PS steps included). As a bicupi n OxdC has a high thermal stability and is resistant to proteases, features characteristic of a member of the cupin superfamily. This stability has been attributed to a struct ure with a large number of subunit contacts, hydrophobic interactions, hydrogen bonding, effici ent packing, short loops and fewer cavities (Dunwell et al., 2001). Table 2-1. Purification table for a typical OxdC preparation. Proteins from crude extract and proteins extracted from pellets we re combined and purified further. Activity is defined as a mol of formate produced per minute (I.U.) per mg of total protein. Total mass Total Activity Specific activity mg/L culture I.U I.U./mg Crude extract 392 5 2540 3 6.5 Lysis pallets extract 279 9 1198 2 4.3 DEAE 233 4 3728 1 16 AS precipitation 70 7 2590 2 37 phenyl Sepharose 39 4 2027 1 52 Q-Sepharose 34 4 1972 1 58 Final 31 2 1922 1 62
20 Specific activity of the purified OxdC for different protein preparations varied from 40-79 I.U./mg. As determined with ICP-MS OxdC with the specific activity of 55 I.U./mg has 1.3 Mn atoms incorporated per mono mer. If it is assumed that activity varies linearly with concentrati on of bound manganese, 2 Mn per monomer corresponds to approximately 80 I.U./mg, which is close to th e highest specific activity obtained in our laboratory or in the literature (Tanner et al., 2001). During dialysis against the storage buffer (up to 8 h at 4 C in 20 mM hexamethylenetetra mine hydrochloride, pH 6, 0.5 M NaCl) specific activity did not decrease. Th is is surprising considering that the manganese binding site is accessible to solven t and in principle Mn ion could dissociate from the protein. Also, incubation of O xdC, expressed with lower than full Mn occupancy, in a buffer containing Mn(II) salts did not revive activit y. This experimental data are consistent with the data obtain ed for Mn-dependent superoxide dismutase (MnSODase), an enzyme with a similar meta l binding site as in OXDCase (Figure 1-3). In the case of MnSODase metal is bound essent ially irreversibly. Mizuno and co-authors (2004) showed that the binding constant for Mn(II) in MnSODase is surprisingly low (K=3.2 108 M-1), implying essentially irreversib le metal binding characteristic of MnSODase protein and indicates that metal binding and release are dominated by kinetic rather than thermodynamic constraints. They proposed that the kinetic stability of the metalloprotein complex is due to stability of the protein structure and independent of the presence or absence of metal. The crystallographic structure of OxdC s hows that the enzyme crystallizes as a hexamer. Multimerization state in a solution of overexpressed and purified OxdC was determined by size exclusion chromatography. E xperiments showed that even OxdC with
21 the highest specific activity (79 I.U./mg) el utes both as a hexamer and as a high molecular weight aggregate (Figure 2-2). Monomers and trimers were not observed in solution. OxdC aggregates were also presen t when single cysteine residue found in the sequence was mutated to serine (C383S), whic h implies that formation of aggregates is not due to formation of disulfide bonds betw een monomers. The highe st concentration of OxdC obtained in the conditions of stor age buffer was 12 mg/ml. It was observed experimentally that solubility of the OxdC decreases by de creasing pH and ionic strength of solution. Figure 2-2. Multimerization state determina tion by size-exclusion chromatography. Logarithm of protein molecular weight and elution volume (Ve) from size exclusion chromatography. Ve is nor malized with a void volume (Vo). Calibration curve was obtained using pr otein standards (black squares). Red squares represent theoretical elution volume for OxdC monomer (M), trimer (T) and hexamer (H) calculated using best linear fit from standard curve. Blue squares are elution volumes obtai ned experimentally for OxdC. Experimental Section Expression and purification. The oxdC gene from B. subtilis was cloned into the pET-9a plasmid that was subsequently us ed to transform the expression strain BL21(DE3) (plasmid construct prepared by Chang, C.H), and the expression of the oxdC 1 1.5 2 2.5 3 0.9220.127.116.11.18.104.22.168 log(MW)= 5.4219 2.72099(Ve/Vo) R2= 0.95464 log (MW)Ve/VoM T H
22 encoded protein was accomplished by inoculating 0.5 L of LBK with oxdC :pET9a/BL21(DE3). The cells were grown at 30ÂºC un til the cultures reached an optical density (OD600) value of 1.7, at which time the 0.5 L of media were heat -shocked at 42ÂºC for 16 min prior to addition of IPTG and MnCl2 to final concentrations of 1 and 5 mM, respectively. The cells were further shak en at 37ÂºC and harvested after 4 h by centrifugation (5000 rpm, 15 min, 4ÂºC). The pe llets were resuspended in 50 mL lysis buffer (50 mM TrisÂ·HCl , pH 7, containing 10 M MnCl2) before being sonicated for 30 s. After sonication, lysis pellets were separate d from the crude extract by centrifugation (8000 rpm, 20 min, 4ÂºC) and resuspended in 50 mL of extraction buffer containing 1 M sodium chloride, 0.1% Triton X-100, and 10 mM 2-mercaptoethanol. The mixture was stirred overnight at room temperature. Ce ll debris was removed by centrifugation and the supernatant was combined with the crude ex tract. This solution (100 mL) was diluted 10fold before it was applied to a 2.5 Ã— 30 cm DEAE-Sepharose Fast Flow (Sigma) column. The column was washed with imidazole hydrochloride buffer (50 mM, pH 7.0, containing 10 M MnCl2) and developed with a 500 mL NaCl gradient (0 1 M). Fractions were collected and assayed for their ability to oxidize o -phenylenediamine ( o PDA) in a side reaction cata lyzed by OxdC, and those fractio ns exhibiting activity were pooled before the addition of solid (NH4)2SO4 to a final concentration of 1.7 M. The precipitate was removed by centr ifugation (8000 rpm, 30 min, 4 C), and the supernatant was loaded onto a phenyl-Sepharose Hi-P erformance (Amersham Pharmacia Biotech) column. The column was washed with imid azole hydrochloride buffer (50 mM, pH 7.0, containing 10 M MnCl2) and developed with a 500 mL (NH4)2SO4 gradient (1.7 0 M). The fractions were pooled as for the DEAE co lumn and diluted 15-fold before they were
23 loaded onto a Q-Sepharose Hi-Performance (Amersham Pharmacia Biotech) column. The protein was eluted with an imidazole hydroc hloride buffer (50 mM, pH 7.0, containing 10 M MnCl2) and a 500 mL NaCl gradient (0 1 M) as for the DEAE column. The purified protein was then concentrated by ultr afiltration (Amicon) to a volume of 10 mL and dialyzed for 5 h against 1 L of stor age buffer (20 mM hexamethylenetetramine hydrochloride, pH 6, containing 0.5 M NaCl) or applied on G-25 -Sepharose Fast Flow (Sigma) column and eluted with storage buffer . The resulting solution was again concentrated, and aliquots were flash-frozen in liquid N2 and stored at -80ÂºC. Protein concentration . Protein concentrations were determined using Lowry assay (Lowry et al., 1951) with bovine seru m albumin solutions as standards. Qualitative o-PDA assay . A reaction mixture used to detect active OxdC in fractions during purification consisted of 50 mM acetate buffer pH 4.2, 0.2% Triton X100, 0.5 mM o -PDA, 50 mM potassium oxalate and 20 L enzyme solution (100 L total volume). If active OxdC is pres ent, the solution turns yellow in a few minutes depending on enzyme concentration and activity. Formate assay . For measuring OxdC specific activity assay mixtures were composed of 50 mM acetate buffer pH 4.2, 0.2% Triton X-100, 0.5 mM o -PDA, 50 mM potassium oxalate and 5 g of OxdC (100 L total volume). Turnover was initiated by addition of substrate. Mixtures were incubated at ambient te mperature (21-23 C) for 1 or 2 min, and then the reaction was quenched by addition of 10 L of 1 N NaOH. The amount of formate produced in an aliquot (50 L) taken from the OxdC-catalyzed reaction mixture was determined by a format e dehydrogenase (FDH) assay (Schute et al., 1976) consisting of 50 mM potassi um phosphate, pH 7.8, 0.9 mM NAD+, and 0.4 IU of
24 FDH (1 mL final volume). The absorbance at 340 nm was measured after overnight incubation at 37 C, and formate was quantitated by comparison to a standard curve generated by spiking OxdC assay mixtures with added NaOH (as for quench) before OxdC with known amounts of sodium formate. Measurements at specific activity were performed in triplicate. Metal analysis . Samples for metal analysis we re pretreated to remove any adventitious metals. OxdC pur ified to homogeneity in con centrations of 6-10 mg/ml and total volume of 0.2 ml was mixed with 2 L 100 mM EDTA, pH 11, and the mixture incubated for 15 min. Next it was applied to a column loaded with 2-3 ml G-25 resin resuspended in a metal-free water (resistivity 18.3 ). The resin was previously pretreated with 0.5 ml 1 M EDTA, pH 11 to remove any adventitious metals. Enzyme was eluted with metal-free water an d fractions pooled based on quick o -PDA assay. Protein solution was diluted up to 10 ml with metal-free 1% HNO3 and sent for metal analysis by inductively coupled plasma-ma ss spectrometry (ICP-MS) at the Soil and Plant Analysis Laboratory (Madison, WI). Concentration of metals in the enzyme samples was subtracted by concentration of metals of sample containing only storage buffer pretreated the same as enzyme samples. Size exclusion chromatography . A BIOSEP SEC-S2000 column (300 x 7.8 mm with 75 x 7.8 mm guard column) (Phenomenex, Torrance, CA) was equilibrated with 100 mM potassium phosphate at pH 7 with 100 mM KCl an ran at 1.0 mL/min flow rate. The column was calibrated with molecular weight standards: lysozyme (14.4 kDa), carbonic anhydrase (29.0 kDa), peroxidase (44.0 kDa), bovine serum albumin (66.0 kDa), alcohol dehydrogenase (150 kDa), amylase (200 kDa), apoferritin (443.0 kDa) and
25 thyroglobuline (669.0 kDa). The void volume (Vo) of the column was measured by injecting blue dextran. To get calibrati on curve log(MW) was plotted vs. Ve/Vo. Between 10 and 50 g OxdC were loaded on the column. After obtaining the elution volume, the molecular weight of OxdC was determin ed by interpolation using the plot obtained for standards (Figure 2-2). Site-directed mutagenesis . Mutagenic primers were designed using Gene Runner v. 3.05 (Hastings Software, Inc.) and obtaine d from Integrated DNA Technologies, Inc (Coralville, IA). The p ET-9a plasmid with the oxdC gene insert was used as a template for PCR with the mutagenic primers using the QuikChangeÂ® Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA). The mutated pET-9a plasmids were isolated from XL-1 Blue supercompetent cells (Stratagene, La Jolla, CA) by standard alkaline lysis, chloroform extraction, PEG precipitation. Th e desired mutations were verified by DNA sequencing of the oxdC gene insert using the services of the University of Florida Interdisciplinary Center for Biotechnology Research. Mutated plasmid was subsequently transformed into BL21(DE3) competent cells (Novagen, Madison, WI) and further expressed and purified as a wild type.
26 CHAPTER 3 STEADY STATE CHARACTERISATION OF THE REACTION CATALYZED BY B. subtilis OXALATE DECARBOXYLASE Introduction The first step in studying the kinetic and chemical mechanism of an enzymecatalyzed reaction is to determine enzyme catalytic properties under conditions of steady state and multiple turnover. The steady state is defined as conditions where reaction rate does not change as a function of time as the co ncentrations of the reaction intermediates reached a constant level. Advantages of steady state measurements include (1) experiment set up is simple and theref ore cheap and readily accessible and (2) experiments do not require large quantities of enzyme. The basic weakness of the steady state experiment is that it does not give dir ect evidence about catalytic mechanism since it reflects the overall reaction ra ther than individual steps. Nevertheless, steady state experiments can help rule out some of po ssible mechanisms. Typically, in steady state conditions concentration of enzyme is much lo wer than concentration of substrate(s) and rate is measured before substrate is apprecia bly depleted (initial rate ). In such conditions initial rates are described in terms of concen trations of substrate by the Michaelis-Menten equation given below (Micha elis and Menten, 1913): v = ] [ ] [ ] [0S K k S EM cat= ] [ ] [maxS K S VM (3-1) Rate of reaction (v) at concentrations of substrate [ S ] that are significantly higher than KM is equal to maximal rate Vmax. Apparent first order rate constant kcat is Vmax
27 multiplied by the concentration of enzyme [E]0. KM is defined as concentration of substrate at which rate is equa l half the maximum rate (v = Â½ Vmax). When [ S ] is significantly lower than KM then: v = M catK k[ E ]0[ S ] (3-2) While KM is only an experimental parameter, kcat and kcat/KM have actual physical interpretations: kcat, also known as a Â“turnover number,Â” is the apparent first order rate constant for the reaction and is a mathematical function of all indivi dual first order rate constants from substrate binding to product release while kcat/ KM, a so called Â“specificity constant,Â” is an apparent second order rate constant and a function of all individual rate constants up to and including the first irre versible step. The exact expression and more precise definition for both kcat and kcat/ KM will depend on the specific kinetic mechanism. For a mechanism that describes oxalate deca rboxylation by OxdC kinetic parameters will be derived and discussed in more details in the next chapter. The main goal of the research presented in this chapter was to de velop an enzyme assay, determine steady state parameters and range of conditions optimal fo r OxdC activity and identify activators and inhibitors. Besides being a necessary first step in the study of catalytic mechanism, these results will also be important information for potential clinical application of OxdC. Results and Disscusion Progress Curve and Michaelis-Menten Equation To make sure that assays are done under th e conditions of initial rate, production of formate was measured at different time points. In the literature it is often stated that initial rate conditions can be only satisfied if subs trate conversion is less than 10 %. This is generally true for reversible reactions. In the case of oxalate decarboxylation, and most
28 decarboxylation reactions in gene ral, the reaction is irreve rsible as a gaseous product (CO2) is formed, and the entropic barrier fo r back reaction is extremely high (low probability of trace product CO2 returning in a reac tive conformation). Th erefore, the rate of formate production will start decreasing only when the concen tration of residual oxalate drops close to the KM value (see text below), which is about 80 % oxalate conversion (Figure 3-1). As a consequence re actions can be run up to oxalate conversion that give enough formate to measure easily experimentally. Figure 3-1. Progress curve for formate production. In itial oxalate concentration was 50 mM, concentration of enzyme 0.06 g/ L in 50 mM acetate buffer pH 4.2. Initial formate production was fitted to linear equation y=ax Prior reports for fungal and bacterial OXD Cases showed that oxalate dependence of the activity followed Michaelis-Menten kinetics. The only exception is recombinant OXDCase from C. velutipes (Chakraborty et al., 2002) that showed sigmoidal dependence with the Hill coefficient >1, whic h might suggest that th ere are at least two substrate binding sites and bindi ng of oxalate to this enzyme is a cooperative process. 0 10 20 30 40 50 051015202530t / minc ( f o r m a t e ) / m M
29 Michaelis -Menten kinetic parameters KM and specific activities from literature are summarized below in Table 3-1. Table 3-1. Specific activities (SA) and Michaelis-Me nten constants ( KM) for oxalate decarboxylation cataly zed by native and recombinant OXDCases. Organism SA / I.U./mg KM / mM Reference C. velutipes n /a 2.05 Shimazono and Hayaishi, 1957 166 4.5 Mehta and Datta, 1991 n /a 0.49 Azam et al., 2001 A. niger n /a 2 Emiliani and Bekes, 1964 M. verrucaria n /a 1.7 Lillehoj and Smith, 1965 B. subtilis (OxdC) 75 15 Tanner et al., 2001 The Michaelis-Menten curve for OxdC catalyzed decarboxylation is shown in Figure 3-2. Kinetic parameters kcat, KM and kcat/ KM measured for OxdC are summarized in Table 3-2. Figure 3-2. Michaelis-Mente n curve for OxdC cataly zed decarboxylation. Assay conditions: 1-70 mM oxalate, acetate buffer pH=4.2, 0.2 % Triton X-100, 0.5 mM o -PDA, 0.052 g/ L OxdC. Initial reaction rate (v) is defined as mM of produced formate per min. Data were fitted to hyperbolic equation y=ax/(b+x). Specific activities varied considerably for different enzyme preparations (see Chapter 2). kcat was calculated based on specific activity obtained for the enzyme preparations that yielded OxdC with highest purity and two Mn ions bound per monomer. 0 0.5 1 1.5 2 2.5 3 01020304050607080v / mM/minc0(oxalate) / mM
30 KM is an average value acquired from a series of experiments done with different enzyme preparations. Experimental error for KM is relatively large due to its dependence on the precision of rate measurements at low oxalate concentrations, which is close to detection limits in the conditions of the assay. Table 3-2. Specific activiti es and kinetic constants: kcat, KM and kcat/ KM for oxalate decarboxylation catalyzed by recombinant OxdC. Kinetic parameter Value Units Specific activity 79 Â± 3 I.U./mg kcat 56 s-1 KM 5 Â± 2 mM kcat/ KM 11 mM-1 s-1 Activation, Inhibition and Substrate Specificity Inhibition/activation studies with the OXDCase from C. velutipes and A. niger demonstrated that weak reducing agents (with E =0.7-0.8 V, such as ascorbic acid, o and p -phenols, o and p -aromatic amines and ferrocyanide) activate fungal OXDCase activity while stronger reducing agen ts (dithionite, sulf ite, hydroxylamine) inhibit the enzyme. Since those reagents were effective only on enzyme during turnover it was proposed that they interfere with or destroy the EoxalateO2 complex (Shimazono, 1955; Emiliani and Riera, 1968). The most intriguing activating and inhibiting effect on OXDCase activity is that of molecular oxygen. For OXDCase isolated from A. niger Emiliani and Riera (1968) found that enzyme was re versibly inactivated in the atmosphere of pure nitrogen, activity reached optimum at oxygen pressure about 0.1 atm and enzyme was irreversibly inactivated in the atmosphere of pure oxygen. In the case where pure O2 was bubbled through enzyme solution in the absence of substrate for 1 h, no inactivation was observed (Emiliani and Bekes, 1964). This is consistent with the proposal in literature for another manganese dependent enzyme homopr otocatechuate 2,3-dioxygenase, where O2 binds to
31 the metal center only after substrate is bound already, which might be necessary to prevent adventitious oxida tion of the metal center by O2 (Vetting et al., 2004). For A. niger OXDCase Emilian and Bekes (1964) also dem onstrated that weak reducing reagent o -phenylenediamine ( o -PDA) increased OXDCase activity and also increased its tolerance to molecular oxygen. B. subtilis OxdC was also activated by o -PDA in concentrations up to 5 mM while it was inhibited at higher o -PDA concentrations 9Figure 3-3). It was observed that OxdC catalyzes oxidation of o -PDA in a side reaction in which a pale yellow color appeared. Since o -PDA oxidation is a consequence of the turnover the appearance of yellow color can be used as a quick and convenient qualitative assay for OXDCase activity. It was s hown experimentally that o -PDA (concentrations 0.05-5 mM) increases both Vmax and Vmax/ KM comparing to values obtained from the assays without o PDA added (Figure 3-3). From the available experimental data, the mechanism of O2 deactivation or o -PDA activation is not clear. Howeve r, a hypothesis cons istent with the experimental observation is following: O2 binds to enzyme-oxalate complex and subsequently during the reacti on becomes reduced to superoxi de. If the superoxide is further reduced in a side reacti on peroxide is formed. As a c onsequence, the active site is left electron deficient since pe roxide is a more stable form of reduced dioxygen than superoxide and therefore a w eaker reducing reagent. This hypothesis is supported by the observation that deactivation by O2 is accompanied by peroxide production (Emiliani and Riera, 1968). In this case, the role of o-PDA might be to prevent superoxide reduction or to provide electrons to the el ectron-deficient active site or reaction intermediates. Both cases are consistent with effect o -PDA has on both Vmax and Vmax/ KM (Figure 3-3).
32 Apart from the redox active reagents, other reagents are also found to affect OxdC activity. Non-ionic detergen t Triton X-100 was observed to enhance OxdC activity. OxdC activity is inhibited with 50 mM fluorid e and iodide salts (30 and 15 % of OxdC activity in salt-free solution respectively), while either chloride or bromide salts did not have significant effect. Similar but more pronounced effect was observed for barley oxalate oxidase (Whittaker and Whittaker, 2002) where it was explained by competition of halides and oxalate for the Mn binding. Th e fact that chloride does not inhibit OxdC up to 50 mM anion concentration is important fo r its clinical applications for analysis of oxalate in biological fluids. Figure 3-3. Effect of o -PDA on Vmax and Vmax/ KM for oxalate decarboxylation by OxdC. Assay conditions: 1-60 mM oxalate, acetate buffer pH=4.2, 0.2 % Triton X100, 0.063 g/ L OxdC. Reaction rate (v) defined as mM of produced formate per min. Carboxylic acids such as pyruvic, oxaloace tic, 2-ketoglutaric, malonic, succinic, glutaric, fumaric, maleic, citric , phthalic and oxamic were test ed as alternative substrates for OxdC. Only oxalate, in th is group of compounds, was shown to be substrate for 0 0.5 1 1.5 2 2.5 -0.4-0.200.20.40.60.811.2 0 0.05 0.1 0.25 1 1/v = 0.77336 + 1.63591/s R2= 0.99935 1/v = 0.40221 + 1.77921/s R2= 0.99885 1/v = 0.35674 + 1.60711/s R2= 0.99868 1/v= 0.32025 + 1.18941/s R2= 0.99479 1/v= 0.28936 + 1.10361/s R2= 0.99622 1/v / min/mM1/c0(oxalate) / mM-1c( o -PDA)/mM
33 OxdC, which was also found for fungal OXDCases (Shimazono, 1955; Emiliani and Riera, 1968). Substrate specificity is not surprising considering that oxalate decarboxylation is a redox reaction in which red ox potential of all par ticipants has to be finely tuned with redox properties of the enzy me. The same sets of compounds were also tested as inhibitors. The highest inhibitory effect was achieved with malonic acid, which is smallest of all tested dicarboxylic acids . This illustrates th at presence of both carboxylic groups is important for interaction wi th the enzyme but also steric hindrance between enzyme and inhibitor. Inhibition pr ofile shown on Figure 34 was closest to one for competitive inhibition with Ki = 8 mM. In competitive inhibition both substrate and inhibitor compete for the same active site. Figure 3-4. Inhibition of OxdC catalyzed oxa late decarboxylation by malonic acid. Assay conditions: 2.5-50 mM oxalate, aceta te buffer pH=4.2, 0.2 % Triton X-100, 0.071 g/ L OxdC. Reaction rate (v) defined as mM of produced formate per min. The pH Dependence on OxdC Catalysis Although enzymes contain many ionizing gr oups, it is usually found that plots of rate of enzyme catalyzed reaction against pH take the form of si mple single or double 0 1 2 3 4 5 6 7 -0.0500.050.10.150.22.214.171.124 0 25 50 1/v= 0.44+ 1.22 1/s R2= 0.99332 1/v= 0.54+ 7.90 1/s R2= 0.99761 1/v = 0.65 + 17.00 1/s R2= 0.99904 1/v / min/mM1/c0(oxalate) / mM-1c(malonate) / mM
34 ionization curves, since the onl y ionizations of importance are those of groups in the active site participating in catalysis, or groups elsewhere that maintain active site conformation. pH variation st udies are used to discover and identify those ionizing groups and their role in catalysis. Detailed re views illustrating how to analyze pH effects on enzyme activity, such as ones written by Cleland (1979) and Tipton and coworkers (1983), are numerous in literat ure. pH studies in the condi tions of steady state usually focus on effects pH has on kcat and kcat/ KM. pH dependence on kcat (or Vmax) is due to the ionization constant of enzyme-substr ate complex while pH dependence of kcat/ KM comes from the ionization of free enzyme or free substr ate. In theory pH studies can be used to identify ionizing groups that participate in catalysis, while in practice this information can be obscured if pH change also affects ch emical properties and stability of enzyme, substrates or co-factors. For the fungal OXDCases maximal decarboxy lation rate was observed at acidic pH: 2.5-4.0 for C. velutipes (Shimazono and Hayaishi, 1957) and 4.7 for A. niger (Emiliani and Bekes, 1964). When C. velutipes OXDCase was heated at 78 C for 10 min in solutions with different pH, cooled to room temperature and tested for activity, it lost 5 % activity at pH 4.0 but almost 90 % at pH 2.0 and 7.0. pH effect on B. subtilis OxdC activity was tested for pH range 2.8-8.2. Th e optimal activity was observed at pH 4.2 (Figure 3-5). Intriguingly at lower pH OxdC is inactivated by high oxalate concentrations (Figure 3-6). Both Vmax/ KM (Figure 3-7) and Vmax (Figure 3-5) decreased as the pH of the reaction buffer was raised above 4. When the observed Vma x/ KM values were multiplied by (1 + KA/H), where KA is the second acid dissociation co nstant of oxalic acid and is equal to 4.2 (Speakman, 1940) and H is the hydrogen ion concentr ation, the resulting
35 Vmax/ KM values were essentially invariant with pH, suggesting that monoprotonated oxalate is the actual substrate for OxdC. Figure 3-5. pH effect on Vmax. Assay conditions: oxalate concentrations 50 mM, 0.064 g/ L OxdC, serial of buffers were used in pH range 3.2-7.8 (see Experimental section). Figure 3-6. OxdC deactivation by oxalate at lower pH. OxdC activity vs. oxalate concentrations at pH 2.8 and pH 3.2. v / mM/min pH 2.8 pH 3.2 0 0.5 1 1.5 2 01020304050c0(oxalate) / mMpH pHV m a x / m M / m i n 0 0.5 1 1.5 2 2.5 3 3.5 345678 -1.5 -1 -0.5 0 0.5 34567logVmax
36 Figure 3-7. pH deffect on Vmax/ KM. The experimental data were fitted to log y = log [C/(1 + K/H)] (Cleland, 1979) which gave a pKa value of 4.2 Â± 0.1. Formate Production: Irreversible or Reversible Process During the oxalate decarboxylation ca talyzed by OxdC a carbon-carbon bond is broken, and a new carbon-hydrogen bond is formed to generate formate. While overall reaction is clearly irreversible, as discussed at the beginning of this chapter, it is not clear if formate production is revers ible or irreversible proce ss. To answer this question formate was incubated in th e presence of OxdC in D2O. The reaction of interest is shown in Figure 3-8. Figure 3-8. Hypothetical exchange of hydrogen and deuterium when formate is incubated with D2O. Exchange of proton and deuterium can be easily measured by proton NMR. If OxdC catalyzes proton exchange, the inte nsity of the formate hydrogen signal will decrease more comparing to the analogous signal from formate that was not incubated with OxdC. Incubation of formate in D2O in the presence of resting OxdC or OxdC under H O O D O O + + D2O HD O l o g ( Vm a x/ KM)pH
37 turnover (oxalate present) did not effect the rate of deuterium exchange (Table 3-3), which leads to a conclusion that formate formation from its hypothetical deprotonated precursor is irreversible in OxdC-catalyzed decarboxylation. Table 3-3. Results of the deuterium exchange experiments. All experiments were done in D2O solutions: 50 mM acetate buffer (pD ~4.2), 25 mM formate and oxalate, 1.5 M OxdC; overnight at room temperat ure. 25 mM dioxane were used as internal standard and it was added prior to 1H NMR experiment. Assay Composition 1H NMR Relative Peak Intensity Formate Dioxane 1. Formate 0.44 0.02 8 2. Formate, OxdC, O2 0.46 0.03 8 3. Formate, oxalate, OxdC, O2 0.52 0.02 8 Initial ratio 1 8 Experimental Section Assay conditions. OxdC activity was measured as a concentration of produced formate versus time. A detailed protocol is given in the Experimental section of Chapter 2. Effect of o -PDA (Sigma) was tested in a conc entration range of 0-10 mM. The carboxylic acids: pyruvic, oxaloacetic, 2-ke toglutaric, malonic, succinic, glutaric, fumaric, maleic, citric, phthalic and oxamic were tested as inhibitors and alternative substrates in 25 and 50 mM concentrations. In the inhibition experiments enzyme was preincubated for 5 min with potential inhib itor and reactions were started by addition of oxalate. The pH of all stock solutions was adjusted to pH of the assay mixture (4.2 usually). The pH dependence for oxalate de gradation catalyzed by OxdC was measured in the following series of buffers: pH 2.8 and 3.2, glycine; pH 3.8, 4.2, 4.7, and 5.2, acetate; pH 5.7, 6.2, and 6.7, MES; pH 7, BES; and pH 7.7 and 8.2, EPPS. Stock solution for all buffers were 200 mM, so that pH did not change significantly by final dilution to 50 mM in assay mixtures.
38 Curve fitting. All curve fitting were done using KaleidaGraphTM Version 3.5 (Synergy Software) curve fitting program. Michaelis-Menten curves were fitted to the hyperbolic equation y = ax/(b+x). Vmax/ KM values for experiments done at pH 3.2 and 3.8 are estimated by linear regression on 1/v vs. 1/[oxalate] plot while high oxalate concentrations were omitted. Deuterium exchange experiments. 1H NMR experiments were done using a Varian Gemini 300 spectrometer. All experiments were done in D2O solutions: 50 mM acetate buffer (pD ~4.2), 25 mM formate and oxalate, 1.5 M OxdC; incubated overnight at room temperature. 25 mM dioxane were us ed as internal standa rd and it was added prior to 1H NMR experiment.
39 CHAPTER 4 HEAVY ATOM ISOTOPE EFFECTS ON THE REACTION CATALYZED BY B. subtilis OXALATE DECARBOXYLASE Introduction Isotope effects (IE) are a power ful tool that have been used for the last 30 years to study kinetic and chemical mechanism of en zyme-catalyzed reac tions. Methodology was originally developed by the Dexter Nort hrop (Northrop, 1975), Wallace W. Cleland (Cleland, 1982) and Paul Cook (Cook and Cl eland, 1981). There are many reviews in literature written by enzymologi sts mentioned above and othe rs that describe theory behind isotope effects and ways they can be employed in mechanistic studies. A review that describes the most closely methodology a pplied in the research described in this chapter was written by OÂ’Leary (1980). Difference in the reaction rate resulting fr om isotopic substitution for atoms other than hydrogen are usually very small. Fo r carbon, nitrogen and oxygen (Â“heavy atomsÂ”) those rate differences are commonly in th e range 1-3% and seldom exceed 10%. Since difference in rates of enzyme catalyzed reac tion with heavy and light isomer is usually less than experimental error, heavy atom isotope cannot be measured by the standard kinetic techniques. To improve sensitivity of the method, isotope effects are determined using a competitive method, in which change s in isotopic distribut ion in the starting material or product were followed over the course of reaction. Is otopic ratio of the required precision can be obt ained only by isotope-ratio mass spectrometer. Such an instrument is specifically designed to meas ure isotope ratios in simple and volatile
40 molecules such as CO2, N2 and CH3Cl. For carbon, oxygen and nitrogen it is possible to use the small natural abundances of the h eavier isotope and ther efore synthesis of isotopically labeled substrates are not neces sary, which greatly simplify the method. In steady-state measurements, since substrate is a mixture of isotopic isomers that compete for the enzyme active site, observed isotope effects will affect only Vmax/ KM (simplified V/K), sometimes referred as Â“specificity cons tant,Â” which is an apparent second-order rate constant that include all steps from substrate binding up to first irreversible step. An advantage of the fact that IE affects only V/K is that it is easier to interpret results since overall effect is effect on fewer steps. C ontrary, the main disadva ntage is that such experiments cannot provide information about st eps after the first ir reversible step. In general, the magnitude of an isotope eff ect on collective rate parameters such as Vmax/ KM, is usually lower than the magnitude of an intrinsic isotope effects unless the isotope sensitive step is only rate-limiting or ra te-contributing step. This phenomenon is an important tool in elucidating kinetic mechan isms. An observed isotope effect is usually compared with the maximal intrinsic isotope effect (usually from non catalyzed chemical reaction) and if significantly lower it implies that other non-isotope sensitive steps are relatively slow comparing to isotope-sensitive one. Observed IE can be sometimes increased by decreasing relative rates of individual kinetic st eps, which can be achieved by changing reaction conditions, most ofte n by running a reaction at nonoptimal pH. Prior to the work described in this ch apter Tanner and co -workers (2001) and Anand and co-workers (2002) proposed a mechanism for oxalate decarboxylation catalyzed by OxdC that involves free radica l chemistry. Such proposals are based solely on the observation that OxdC is a manga nese and dioxygen dependent enzyme. The
41 mechanism proposed in this chapter is based on experime ntal results: primarily on 13C and 18O kinetic isotope effects and also EPR spectroscopy and steady-state characterization (Chapter 3). The work described in this chapter was published (Reinhardt and et al., 2003). Results and Discussion 13C and 18O Isotope Effects The 13C and 18O isotope effects (IEs) were meas ured on the reaction catalyzed by OxdC at pH 4.2 and pH 5.7 (Table 4-1). Table 4-1. 13C and 18O isotope effects (IE) on the reaction catalyzed by Bacillus subtilis . Results presented in this table show: (1) there is observable 13C and 18O IEs, (2) 13C IEs are smaller that 3-5%, (3) IEs for CO2 differ than for HCO2 -, (4) CO2 is enriched with heavy isotopes comp aring to the initial oxalate. For more detailed analysis see text below. 13(V/K), % 18(V/K), % pH (buffer) CO2 HCO2 CO2 HCO2 4.2 (BHEP) 0.5 0.1 1.5 0.1 -0.2 0.2 1.1 0.2 5.7 (piperazine) 0.8 0.1 1.9 0.1 -0.7 0.1 1.0 0.1 For rate-limiting C-C cleavage reactions, normal 13C IEs of 3-5% are typical. However, for the OxdC-catalyzed reaction the 13C IE values for the CO2 product (normal 0.5-0.8%) were lower than expected, indicati ng that C-C bond cleavage is not the major rate-limiting step. The observation that the 13C IE for formate production (normal 1.51.9%) is 3-fold larger than that for CO2 is consistent with the pr esence of a different step that is slower than C-C bond cleavage. Moreover, the normal 13C IEs observed are most likely due to a decrease in C-O bond order in this step. Given that the formation of CO2 is most likely irreversible, the slower, revers ible step precedes C-C bond cleavage to give CO2, a gaseous product. At pH 5.7, the chemical steps are somewhat more rate-limiting
42 as compared to those at pH 4.2, as shown by the increase in the observed 13C IEs. This difference is presumably the result of an exte rnal commitment at the higher solution pH. As observed for the 13C IE, the 18O IE is also normal (1.0-1.1%) for formation of formate (Table 4-1) . In contrast, the 18O IE for CO2 production is inverse (-0.2% to 0.7%), consistent with the fact that the bond order for these oxygens increases from 1.5 in the substrate to 2 in the product. Similar measurements on decarboxylases, for which the chemistry step is rate-limiting, have given 18O IE values of -0.4% to -0.5% (Headley and OÂ’Leary, 1990; Waldrop et al., 1994). In a ddition to the effects of changes in bond orders, 18O IEs can also be influenced by changes in protonation state. For example, the equilibrium 18O IE for protonation of a carboxyl group has been shown to be 0.98 (Tanaka and Araki, 1985). Since the V/K prof ile (Figure 3-7) for OxdC suggests that monoprotonated oxalate is the su bstrate, the effects of prot onation must be taken into account in deducing the chemistry of the en zyme-catalyzed reaction from the observed 18O IE values. A further complicatio n in interpreting the experimental 18O IEs values is the possibility that water attacks an inte rmediate in the OxdC -catalyzed reaction, resulting in the presence of oxyge ns in the residual substrate or products derived from the solvent. It was therefore examined whether exchange of the solvent oxygens into formate or oxalate took place under the reaction conditi ons by incubating OxdC with substrate in a solution containing 5 % H2 18O. Only 0.2 % of 18O incorporation into both formate and oxalate was observed under these conditions , and no exchange wa s found in control reactions performed under identic al conditions, with the excep tion that formate instead of OxdC was added at the star t of the incubation. While 18O exchange is therefore
43 dependent on the presence of active OxdC, it occurs to a minimal extent, confirming the validity of 18O IE measurements. Kinetic Mechanism for the Enzymatic Reaction of OxdC The observed 13C and 18O isotope effects rule out any mechanism for the OxdCcatalyzed breakdown of oxalate that does not incl ude an isotope-sensitive, reversible step prior to decarboxylation. Th erefore a minimal kinetic mechanism up to the first irreversible step for the OxdC catalyzed re action can be written as shown in Figure 4-1. Figure 4-1. Minimal kinetic mechanism up to fi rst irreversible step for OxdC catalyzed reaction consistent with 13C and 18O isotope effects. The first step ( k1) is oxalate binding to the OxdC, second step ( k2) formation of oxalyl radical intermediate and third step ( k5) is decarboxylation. The first reversible step is oxalate bind ing, than formation of a reactive oxalate intermediate in a slow and reversible step and subseque ntly decarboxylation. IE results cannot provide evidences whether oxygen binding occurs before or after oxalate binding. Based on EPR studies of OxdC catalyzed re action (Chang et al, 2004) and analogy drown from the studies on other oxygen binding meta loenzymes in literature (Vetting et al., 2004), it can be proposed that oxygen binds after oxalate is already bound to the manganese and that Mn(II)-O2 complex is an active part aker in redox chemistry (see detailed discussion in Chapter 1). For the proposed minimal kinetic mechanism (Scheme 4-1) Vmax/KM is derived using Â“net rate constantÂ” method (Cleland, 1977) : E + oxalate k1 k2 E-oxalate E-oxalate. E-formate. + CO2 k3 k4 k5
44 Vmax/ KM = [E]0 5 3 5 4 2 5 3 1) ( k k k k k k k k (4-1) The equation for the effects of isotopic substitution in the minimal mechanism can be derived from equation 4-1 as ratio of Vmax/ KM for reaction with light isotope (12C or 16O) and Vmax/ KM for reaction with heavy isotope (13C or 18O). For simplicity, x(V/K) will stand for the ratio of Vmax/ KM for the light and heavy isotope (i.e. isotope effects) in which x = 13 or 18 for 13C and 18O IE respectively. If one assumes that isotope sensitive steps are k3, k4, and k5 then equation 4-2 can be derived as: x(V/K) = 2 3 4 5 4 2 5 3 4 5 3 5 31 1 k k k k k k k k k k k k Kx x eq x (4-2) Fractionation Factors The BEBOVIB program (Sims and Lewis, 1984; Berti, 1999) was used (done by N.G.J. Richards) to estimate the 13C and 18O fractionation factors of oxalate relative to CO2 gas, using a force field that closely reproduced the experimental IR and Raman frequencies of oxalate (Ito and Berstei n, 1956). Assuming C-O bond orders of 1.5 in oxalate, the 13C and 18O fractionation factors versus CO2 were computed to be 0.9838 and 0.9674, respectively, which are reasonable numbers. Then C-O bond order was systematically decreased in one end of oxalate to a value of 1.0, while maintaining the CO bond order at 1.5 in the other carboxylate group. These calculations showed that reducing the C-O bond order lowered the 13C and 18O fractionation factors in a linear fashion to 0.9537 and 0.9465, respectively (Figure 4-2). In a similar manner, it can be demonstrated that the 18O fractionation factor in the ot her carboxylate end was almost
45 unchanged and the 13C value decreased slightly to 0.9828. These calculated values were subsequently employed in estimating the actu al bond orders in the oxalate radical (see below). Analysis of the Observed 13C and 18O Isotope Effects Since, on the basis of th e pH dependence of the Vmax/ KM (Figure 3-7), the monoprotonated form of oxalate is the true substrate (see detailed discussion in Chapter 3), 13C and 18O isotope effects analysis is complicated by whether ( i ) the protonated carboxylic acid gives rise to CO2 or formate in the OxdC-catalyzed reaction and ( ii ) removal of the proton occurs in the first ( k3) or second ( k5) step of the minimal mechanism. Figure 4-2. Fractionation factors of oxalate carb ons and oxygens as a function of C-O bond order. The sloping lines are for the carboxyl whose C-O bond order is varied, while the nearly horizontal li nes are for the one where the bond order remains 1.5. Â— 13C, Â·Â·Â·Â·Â· 18O. Figure was prepared by N.G.J. Richards. C-O Bond Order Fractionation Factor Relative to CO2(g)
46 At pH 5.7, where oxalate exists almost co mpletely as a dianion, protonation of the substrate must take place to yield the monoanion. The simplest assumption is that formation of monoprotonated oxalate occurs pr ior to binding, although this has not been experimentally establ ished. In this case, 18O will be enriched in the protonated carboxyl group of the substrate by 2%, so that 18(V/K) IE at the protonate d end of the molecule must be multiplied by 0.98. This then raises the question of whet her the carboxylic acid moiety becomes formate or CO2 when the substrate is bound w ithin the OxdC active site. In all cases, given the minimal kinetic m echanism shown above (Figure 4-1) and if assumed ( i ) that the oxalate carboxyl group giving ri se to formate was initially protonated and ( ii ) a hydrogen atom was remove d prior to decarboxylation ( k3), a reasonable fit of the experimental data to equation 4-2 for all of the four observed heavy-atom IEs was not obtained. In a similar manner, assuming that the protonated substr ate carboxyl ending up as CO2 was deprotonated in the same step as decarboxylation ( k5) does not give a consistent model at pH 5.7. In contrast, th e assumption that the proton was removed from this group in a step ( k3) prior to C-C bond cleavage ( k5) gave a model consistent with experimental observations (vide infra). Thes e two conclusions were also necessary to model the isotope effects at pH 4.2, although reproducing the data at this pH required the introduction of an external commitment. Gi ven these conclusions, analysis starts by assuming that the 13C IE on decarboxylation to give CO2 is 1.04, which is an average value for such reactions (OÂ’Leary, 1989) and that k3/k2 is small at pH 5.7. Under these conditions, equation 4-2 can be rewritten as 1.008 = ) / ( 1 ) / ( 04 . 14 5 4 5k k k k (4-3)
47 In obtaining this equa tion, the values of 13Keq3 and 13k3 were set to unity because proton removal from the carboxylic acid proceeds with a negligible 13C isotope effect (Bayles et al., 1974). Solving equation 4-3 gives k5/ k4 = 4, indicating that decarboxylation is 4-fold more likely to occur from the oxalate radical than the back reaction to regenerate the substrate. Turning to the 18O isotope effect at the carboxyl that becomes CO2, 18(V/K) must be first multiplied by 0.98 to correct for the pr otonated carboxyl being the initial form of the substrate. 1 8Keq3 will then be 1.02, as a result of proton removal in the first step. From fractionation factor calculat ions an estimate of 0.967 for 18Keq5 gives 18k5 = 0.983, when set midway between unity and 18Keq5. Assuming that k5/ k4 is 4 and k3/ k2 is small, as before, allows equation 4-2 to be rewritten as 0.993 = 4 1 4 983 . 0 02 . 1 98 . 03 18 k (4-4) From equation 4-4, 18k3 = 1.0159, which is a reasonable value for the deprotonation step. For the isotope effects at the carboxylate group of oxalate that becomes formate, equation 4-2 can be employed to calculate 13Keq3, assuming that ( i ) 13k5 is 1.03, a typical value for decarboxylation and ( ii ) 13k3 is midway between 13Keq3 and unity. With this value, equation 4-2 becomes: 1.019 = 4 1 2 / ) 1 ( 4 03 . 13 13 3 13 eq eqK K (4-5) Solving equation 4-5, it was found that 13Keq3 = 1.0215, which when divided into the fractionation factor of oxalate give s 0.963, corresponding to a C-O bond order of the oxalate radical of 1.16 (Figure 4-1). Assuming 18k5 = 1.003 to model the loss of C-C-O
48 bending modes in the formate radical anion allo ws to perform a similar calculation for the 18O IE at this end of the molecule, since equation 4-2 takes the form 1.01 = 4 1 2 / ) 1 ( 4 003 . 13 18 3 18 eq eqK K (4-6) From equation 4-6, 18Keq3 = 1.0157, which when divide d into the fractionation factor of oxalate gives 0.9524, correspondi ng to a bond order of 1.14 for the C-O bonds at this end of the oxalate radical, in g ood agreement with the value of 1.16 computed on the basis of the 13C IE. Average value for the C-O bond order 1.15 implies that the structure of the oxalate radical is a hybrid of two resonan ce forms, in a ratio of 70:30 (Figure 4-3). The partial positive charge on th e carbon going to formate likely facilitates the decarboxylation of the radi cal (Figure 4-3). Hydrogen tran sfer to the formate radical anion, formed by loss of CO2, then yields formate, which is subsequently released from the enzyme. Figure 4-3. Structure of the putative oxalyl ra dical intermediate in the OxdC catalyzed reaction. The radical deca rboxylates to give CO2 from the left carbon and formate radical anion from the right carbon. Relative ratio of the two resonance structure was determined us ing isotope effects and fractionation factors (see text). At pH 4.2, oxalate is half protonated to the active monoanion form, and the small heavy-atom IEs observed at this pH (Table 4-1) suggest that k3/ k2 is no longer negligible. Thus calculations can be repeat ed in attempt to determine k3/ k2 at pH 4.2. At the carboxyl end of oxalate that becomes CO2. Substitution of the observed 13C IE into equation 4-2 gives: 70% : 30%
49 1.005 = 2 3 2 31 4 1 1 4 04 . 1 k k k k (4-7) Solving equation 4-7 gives k3/ k2 = 0.75, and hence oxalate more often dissociates from the enzyme than proceeds to products. For the 18O IE the equation is multiplied by 0.99, rather than by 0.98, since half of th e substrate is monoprotonated, to give 0.998 = 2 3 2 31 4 1 4 ) 0159 . 1 ( 4 ) 983 . 0 )( 02 . 1 ( 99 . 0 k k k k (4-8) This equation then gives k3/ k2 = 0.81, a value similar to that computed on the basis of the 13C IE above. Since the 13C IE is better determined than the 18O IE at pH 4.2, in subsequent calculations value k3/ k2 = 0.75 will be used. Using the values computed for 13Keq3, 13k3, 13k5, k3/ k2, and k5/ k4 from our observed isotope effects for the minimal mechan ism (Figure 4-1), we could calculate 13(V/K) at pH 4.2 for the carboxyl group at the end of oxala te that becomes formate. Substitution into equation 4-2 gives 3 4 1 3 ) 0108 . 1 ( 4 ) 0215 . 1 )( 03 . 1 ( = 1.012 = 13(V/K) (4-9) that is in good agreemen t with the experimental value of 1.015. For the 18O isotope effect, equation 4-2 by 1.01 to correct for the half-pro tonation of the substrate, and so equation 4-2 becomes 3 4 1 ] 3 ) 0078 . 1 ( 4 ) 0157 . 1 )( 003 . 1 [( 01 . 1 = 1.016 = 18(V/K) (4-10)
50 that compares with the experimental value of 1.011. While the agreement between theory and experiments is not perfect in this cas e, the kinetic model for OxdC-catalyzed decarboxylation does fit the experimental data in a reasonable fashion. Molecular Mechanism of the Re action Catalyzed by OxdC The catalytic mechanism of OxdC proposed here invokes the pa rticipation of a substrate-based radical in orde r to facilitate cleavage of the C-C bond. This hypothesis is supported by (i) the dependence of OxdC activity on catalytic levels of dioxygen (ii) the presence of manganese, which can abstract an electron from bound substrate, in the active site(s) of the enzyme (iii) the absence of other cofactors in the OxdC structure and (iv) the lack of a net redox change betw een substrate and products. Observed 13C and 18O isotope effects for OxdC catal yzed reaction rule out any me chanism that does not include an isotope-sensitive, reversible step prio r to decarboxylation. The normal IEs found for both carbon and oxygen also limit the possible number of mechanisms employed by the enzyme. Proposed mechanism that is consistent with the observed 13C or 18O IEs and with the crystal structures of the OxdC active site (Anand et al., 2002; Just et al., 2004) is given on Figure 4-4. Based on IE data analysis CO2 is formed from the protonated end of bound oxalate. On the basis of si mple electrostatic considerations, it can be proposed that the carboxylate anion of the substrate, rather than the carboxylic ac id moiety, would be directly attached to the meta l ion. In proposed mechanism, the reversible slow step required by the IE results presumably corres ponds to the oxidative transfer of a single electron from oxalate to a Mn(II)-oxygen complex, accompanied by proton transfer from the protonated substrate to a nearby residue. On the basis of the OxdC crystal structures, the side chain of Glu-162 in N-terminal dom ain and/or Glu-333 in C-terminal domain could function as the general base in the depr otonation reaction due to their proximity to
51 the Mn-binding site (Figure 5-1). From the avai lable data it is not cl ear which of the two metal-binding sites (Nor Cterminal) or both are catalytically active. The proton transfer at this step is sup ported by not only the IE studies bu t also the fact that oxalate dianion is easier to oxidize (~0.7 eV) than the mononanion (Isse et al., 1999). The IE measured here for formation of formate appears to be largely equilibrium one. Thus, the oxalate radical formally has an empty p orbi tal on carbon and a singl e odd electron in a p orbital on oxygen, which may delocalize direc tly onto the p orbital of the nearby oxygen, forming a three-electron bond (Figur e 4-4) (Jonsson et al., 1996). Removing an electron therefore reduces the bo nd order between carbon and oxygen, causing the normal 13C and 18O IEs observed for formation of formate. Arginine residues Arg92 and/or Arg-270 in Nand C-terminus domain, resp ectivelly (Figure 5-1) might pl ay a role in polarizing the C-O bond in the Mn-bound, radical intermedia te through an electrostatic interaction. Decarboxylation and C-C bond cleavage in a second step, facilitated by the partial positive charge on the carbon that will become formate, then results in formation of a formate radical anion having a negative charge at the carbon atom (Figure 4-4). This step, which is faster than reversal of the first one, is consistent with the observation of the small normal 13C IE that is measured for CO2. IE experiments do not provide information about the steps after the first ir reversible step but it can be pr oposed that in the final steps of the OxdC-catalyzed reaction, the formate radical anion acquires a proton, possibly back from Glu-162 and/or Glu-333 residue s, and an electron from Mn(II)-reduced oxygen complex to yield the product fo rmate and regenerated dioxygen. A key conclusion of presented here IE studi es is that the radical used to initiate decarboxylation is not located on the car boxyl group that eventually forms CO2. As well,
52 the large drop in bond order determined fo r the substrate C-O bonds of the carboxylate that forms formate in the OxdC-catalyzed r eaction, is clear evid ence that the radical initiating decarboxylation cannot be present on the carboxylate that is converted to CO2. Figure 4-4. Mechanism of the OxdC-catalyzed C-C bond cleavage consistent with IE and EPR experiments. The mechanism also s hows possible roles of active site Arg and Glu residues. (A) Monoprotonated oxalate and oxygen bind to the Mn metal center. The oxidation state of Mn is proposed to be 2+ (Chang et al. 2004). (B) Glu-162 acts as a general base to remove a proton while electron transfer takes place to generate oxaly l radical. Arg-92 polarizes the C-O bond of the substrate, building up positive charge on the carbon atom, thereby facilitating heterolytic cleavag e of the C-C bond to give CO2. (C) Representation of the product of decarboxyl ation that is consistent with the model for the C-O bond orders derived from the measured KIEs. This species is a resonance form of the formate radical anion. Coupl ed proton-electron transfer, possibly involving Glu-162, then yields formate. It is interesting to compare proposed m echanism of OxdC catal yzed reaction with the chemical mechanism of oxalate decarboxy lation (see discussion in Chapter 1). In each case oxalate has to be first oxidized , which is followed by C-C bond cleavage by a Mn His His Glu O His O O O OH O Mn His His Glu O His OO O OO PCET -O Glu162 O H2N NH2 + N H Arg92 H2N NH2 + N H Arg92 HO Glu162 O Mn His His Glu O His OO O H2N NH2 + N H Arg92 O Glu162 O H CO2 Mn His His Glu O His O O O H2N NH2 + N H Arg92 -O Glu162 O H II II II II (A)(B) (C) (D)
53 successive rather than by concerted mech anism. In a chemical mechanism produced radical intermediate is furthe r oxidized to form second CO2 molecule, while in the OxdC catalyzed reaction the radical intermediate is reduced which results with formate as a final reaction product. Interestingly, produc ts of chemical oxala te decarboxylation are more similar to the products of reaction cat alyzed by oxalate oxidase. Here, it can be proposed that the function of OxdC is not only to catalyze reaction but also control its outcome. Experimental Section Isotope effect nomenclature. The nomenclature used in this work is that due to Northrop (1977) in which 13(V/K) represents the ratio of Vmax/KM for the 12C-containing species relative to the 13C-containing species, and 18(V/K) is a similar ratio for 16O and 18O. 13C and 18O isotope effect determinations. The primary 13C and secondary 18O isotope effects for the OxdC-catalyzed decarboxylation reaction were measured by isotope ratio mass spectro metric analysis of CO2(g) (OÂ’Leary, 1980) by the internal competition method, and hence these isotope effects represent the effects of isotopic substitution on Vmax/KM. Their values are therefore asso ciated only with steps up to, and including, the first irrevers ible step in the mechanism. Natural-abundance levels of isotopic label in the carbon a nd oxygen atoms of oxalate were used in these experiments. The analyses to determine the isotopic ratios (R values) were performed on CO2(g), which was either isolated directly from O xdC-catalyzed partial conversion of oxalate (RCO2) or produced by oxidation reacti ons run to completion with I2 in anhydrous DMSO (Hermes et al., 1984) of initial oxalate (R0), produced formate (Rformate), and residual oxalate (Rs) in reaction run to completion with I2 in anhydrous DMSO. In these isotope
54 effect determinations, the enzymatic reacti ons were performed by incubation of OxdC with oxalate (27-43 mM) and o-phenylenediamine (0.5 mM) at 22 C in either 0.1 M 1,4bis-(2-hydroxyethyl)piperazine, pH 4.2, or piperazine, pH 5.7 (920 L total volume). All gases were passed over Ascarite to remove CO2(g) prior to use. Buffer solutions and aqueous solutions of oxalate/o-phenylenediamine were sparged beforehand with O2(g) and N2(g) (which had been passed through 1 N aqueous H2SO4 in addition to being treated with Ascarite), resp ectively. The presence of oxygen is required for OxdC activity, precluding the sparging of both solutions with only N2(g). After addition of buffer to the substrate solution, sufficient OxdC was added to yi eld reaction times of between 7 min and 2 h. OxdC-catalyzed reac tions were quenched by raising the solution pH to 7.5 with 1 N Tris-H2SO4, pH 7.5. The fraction of conversion, f, was determined by use of formate dehidrogenase coupled assay (F DH) (Chapter 2) and an oxalate detection kit (Trinity Biotechnology) to measure form ate and remaining oxalate concentrations, respectively. CO2(g) was collected during the OxdCcatalyzed reaction and purified as previously described (OÂ’Leary, 1980). Oxal ate and formate were separated on Bio-Rad AG-1 ion-exchange resin (1 Ã— 4 cm) and eluted with aqueous H2SO4 (pH 2.5-2.7). The column fractions were monitored by the F DH and oxalate assays described above, and the desired fractions were pooled. HEPES (1 M, pH 7.0, 0.5 mL) was added to the fractions containing formate, while oxalate samples were neutralized with NaOH. After rotoevaporation to remove most of the H2O, samples were again sparged with N2(g) to remove any dissolved, atmospheric CO2(g). After removal of residual water under vacuum (<1 Torr), these samples were placed on a vacuum line (<1 Torr) and heated at 70 C overnight. An anhydrous solution of I2 (0.15-0.2 g) dissolved in DMSO (2 mL)
55 was then syringed into the dried samples, which were residing on the vacuum line, and the CO2(g) evolved in the oxidation reaction was distilled through two pentane/N2(l) traps before being collected in a N2(l) trap. CO2(g) was further distilled from a pentane/N2(l) trap into a collection flask cooled with N2(l). Whereas formate oxidation produces 1 equivalent of CO2(g), each molecule of oxalate gives 2 equivalents of CO2(g). Control samples in which the reaction was quenched be fore the addition of OxdC did not contain detectable amounts of CO2(g). Experiments with 18O-enriched water (Isotec) were performed in a manner similar to that descri bed above except that the reaction solutions contained 5% H2 18O. Figure 4-5. Schematic diagram of the 13C and 18O isotope effect experiments. Blue color represents components of OxdC reaction mixture, while red represents oxalate used as a standard with isotope ratio R0. Isotope ratios: R0, RCO2, RS and Rformate were obtained by mass spectrometar (MS). MS gives isotope ratios only for CO2 molecules, which is obtained directly as a product of the OxdC reaction or from chemical conversion of initial oxalate, residual oxalate and formate to CO2.
56 The isotopic ratios of each sample of CO2(g) were measured on a Finnegan Mat spectrometer, and the following equation was used to determine the apparent isotope effect (appÂ·IE): appÂ·IE = 01 ln 1 ln R R f fs= 01 ln 1 ln R R f fP (4-11) where Rs is the mass ratio of the isotope (13C or 18O) in residual oxalate at the fraction of reaction f, and R0 is the value for the oxalate substrate before reaction. Similarly, RCO2 and Rformate are the cognate mass ratios in CO2(g) or formate at f, respectively, and Rp is their average. The 13C or 18O isotope effects on the formation of CO2 and formate were then calculated by use of eqs 4-12 and 4-13: Isotope effect for formation of CO2 = x x 2 1 appÂ·IE (4-12) Isotope effect for formation of formate = 2 1 x appÂ·IE (4-13) where x is the RCO2/ Rformate ratio. The 18O mass ratio in the CO2 product was determined by comparing the original R0 value with that for Rs and Rformate. A complete derivation of eqs 4-12 and 4-13 is given in the Appendix B . Schematic diagram of the IE experiments were given in Figure 4-5.
57 CHAPTER 5 ROLES OF THE ACTIVE SITE ARGI NINE AND GLUTAMATE RESIDUES Introduction Oxalate decarboxylase and oxalate oxidas e are good model systems to study how protein environment controls the outcome of a chemical reaction since both enzymes have almost identical Mn-bindi ng site(s) (Figure 5-1), both use oxalate as a substrate, and both employ molecular oxygen in the reaction. While OXOXase catalyzes a formal redox reaction in which dioxygen is reduced, OXDC ase catalyzes a redox-neutral reaction in which dioxygen has a role of a cofactor (F igure 1-1). Interestin gly, recent density functional theory (DFT) calculations have suggested that the intrinsic gas-phase reactivity of the B. subtilis oxalate decarboxylase (OxdC ) active center is to oxidize oxalate such as in OXOXase catalyzed reactions (Chang and Richards, 2005). These DFT results imply that one role for the OxdC protein envi ronment is modulation of the intrinsic metallocenter reactivity, presumably by affecti ng electronic distributi on at the Mn center during catalysis. Another interesting feature of the OxdC is that each cupin domain has an almost identical metal binding site to the other, extending to the residues in the second coordination sphere (Figure 5-1). The presen ce of two Mn-binding sites opens up several issues, including: (1) whether both Mn bindi ng sites are catalytically active, (2) assuming they are, whether they work independently, (3) if only one site is catalytic, which of the two metal binding sites is the one, and (4) if the Mn-binding sites have different functions how these functions are controlled. As in the case of OXOXase versus OXDCase
58 activity, it can be proposed that protein e nvironment controls f unctions of each Mnbinding site. Proposals in the literature that address those questions will be discussed in more detail below. Figure 5-1. Comparison of the manganese-binding sites in oxalate oxidase and oxalate decarboxylase. (A) H. vulgare OXOXase Mn-binding site (Woo et al., 2000), (B) OxdC C-terminal Mn-binding si te (Anand et al., 2002), (C) OxdC Nterminal Mn-binding sites in Â“openÂ” conformation (Ana nd et al., 2002) and (D) OxdC N-terminal Mn-binding sites in Â“closedÂ” conformation (Just et al., 2004). Bound water molecules are shown in their protonated form given that the optimum catalytic activity for both enzymes is reported to be approximately pH 4.2. Atoms are colored as follows: C, grey; H, white; N, blue; O, red; Mn, silver. Hydrogens attached to carbon atoms have been omitted for clarity. Images were render ed using the CAChe Worksystem Pro V6.0 software package (CAChe Group, Fujitsu, Beaverton OR). For OxdC two crystal struct ures are available (Anand et al., 2002; Just et al., 2004). Both crystal structures show ed that non-ligand residues cl osest to the Mn are Arg92 in A B C D
59 the N-terminal domain and Arg270 and Glu333 in the C-terminal. The main difference between the two crystal structures is in th e N-terminal metal binding site. In the first crystal structure (Anand et al ., 2002) this site is solventexposed (Â”openÂ”) and lacks a residue that can participate in the general aci d/base catalysis, which is required for OxdC reactivity as proposed in the Figure 4-4. Cons idering that the C-terminal Mn-binding site has such a residue (E333), Anand and cowo rkers (2002) suggested that it is the Cterminal domain that hosts the catalytically active site. The authors also showed that mutations E333A, R270E and Y340F in the C-term inus led to proteins with 4, 8 and 5 % of the wild type activity, respectively. A s econd crystal structure (J ust et al., 2004) shows a loop containing a 310 helix near the N-terminal metal binding site in an alternative conformation. The loop forms a Â“closedÂ” conformation forming a lid covering the entrance to the Mn-binding site. This conforma tional change brings Glu-162 close to the metal, which makes it a candidate for general acid/base catalysis. Consequently, Just and coauthors (2004) have proposed that the N-te rminal Mn-binding site is the sole or dominant catalytically active site while th e C-terminal domain has a predominantly structural role. The second crystal struct ure also showed that in the Â“closedÂ” conformation the Mn-binding site is not solvent exposed. Seve ral mutations of the active site Arg (R92 and R270) and Glu (E162 and E 333) residues when made individually have caused activity to drop significantly below 50 % (Just et al., 2004), i ndicating that neither of the two domains can carry out catalysis independen tly, although the nature of interaction between the two sites was not clar ified. In general, the role an individual residue has in catalysis cannot be deduced based merely on decrease of specific activity, if it is not established what effects this mutation has on the enzyme structure.
60 From the isotope effects measurements pr esented and analyzed in Chapter 4 it was calculated that the C-O bond in the oxalyl ra dical anion is polarized according to Scheme 4-2. Due to proximity to the metal center it was suggested th at the active site arginine residue (Arg92 or 270) is liable for polarizat ion of C-O bond in the oxalyl radical anion. The active site glutamate (Glu162 or 333) was proposed to act as a general base in the proton-coupled electron transfer. The goals of the research presented in this chapter are (1) to find experimental evidence for the role s of active site Arg a nd Glu residues in the OxdC-catalyzed reaction, and (2) to find out if the mutations affect reaction specificity and can give a hint as to which of the two Mn-binding sites hosts the catalytically active site. Several active site Arg and Glu mutants were already characteri zed in the literature (Anand et al., 2002; Just and al., 2004). As allude d to earlier, a drawback of these reports was that conclusions were made based solely on specific activities of the mutants without analyzing how those mutations affect the en zyme structure. Also, OxdC mutants were engineered with the polyhi stidine tag on the C-terminus (Anand et al., 2002) or Nterminus (Just et al., 2004), which have si gnificantly reduced enzyme specific activity. For example, activity of the His-tagged w ild type was 21 I.U./mg (Just et al., 2004) comparing to the 75 I.U./mg reported previ ously by the same group for the non-tagged wild type (Tanner et al., 2001). Mutants of the OxdC prepared in presented study are summarized in Table 5-1. Table 5-1. Prepared active site argi nine and glutamate OxdC mutants. N-terminal C-terminal Double mutants R92K* R270K* R92K/R270K E162D E333D E162Q E333Q E162Q/E333Q * Prepared by Ewa Wroclawska
61 Results and Discussion Mutational Effects on OxdC Structure Purification of the wild type OxdC that yields protein with high Mn incorporation (1.3-2.0 Mn per bicupin monomer) includes tw o ion-exchange and one phenyl-Sepharose (PS) steps (Chapter 2). During the PS purification step OxdC is fractionated into three distinct peaks (Figure A-2). The least hydr ophobic peak (one that elutes first) was collected and further purified. All of the mutants in the N-terminal domain behave as wild type during the purification, while Cterminal domain mutants and double mutants did not show the same chromatogram as w ild type when run through PS column. For all of the C-terminal mutants, the most hydrophi lic peak was missing and the other two were more pronounced (Figure A-2). All C-terminal and double mutants we re purified just by ion exchange chromatography, since addi ng PS step did not significantly increase specific activity for these mutants. When th e wild type enzyme wa s purified without the PS step, the average specific activity was 30 I. U./mg a value used to determine how much the specific activity of C-terminal and double mutants cha nged relative to the wild type. The fact that interaction with hydrophobic resin changed for some of the mutants indicates that those mutations may have an effect on the overall enzyme structure, resulting in a more exposed hydrophobic surface. All N-terminal mutants behaved as did wild type on the PS column suggesting that global structure for these mutants was preserved. Effect of mutations on the quaternary st ructure was assessed by size-exclusion chromatography (SEC). Wild type with th e maximum specific activity of 79 I.U./mg eluted in two fractions: as a hexamer, as e xpected from the crysta l structure, and as a higher molecular weight aggregate (Chapter 2) . Results of SEC for different mutants are
62 summarized in Table 5-2. As during the purification, most of the N-terminal mutants have the same structural characteristics as the wild type while the C-terminal mutants showed lower multimerization states. Interpreta tion of SEC is inconclusive in the cases when only aggregates eluted from the column, since even the fully active wild type also showed aggregate formation. Table 5-2. Molecular weights (MW) of OxdC in solution and manganese incorporation per momomer. Calculated Multimerisation Mn MW/kD per monomer wt (79 I.U./mg) 430, 280 aggregate, 6 2.0 R92K 540 aggregate 1.7 E162D 470, 270 aggregate, 6 tbd E162Q 480, 280 aggregate, 6 0.5 R270K 370, 140, 45 aggregate, 3, 1 R333D 420, 120 aggregate, 3 Finally, metal incorporation for the mutants purified to homogeneity was determined by ICP-MS. The results in Tabl e 5-2 illustrate that mutation of Arg 92 (R92K) did not affect Mn binding significan tly, whereas mutation of Glu162 (E162Q) did. Considering that the E162Q mutant behaves on the PS column as the wild type and forms hexamers in solution, but binds fewe r Mn ions might suggest that the overall structure for this mutant is unperturbed, but may have changed locally in a way that affects manganese binding. In conclusion, for the R92K mutant, the decrease in the specific activity can be attributed to the e ffect on chemistry, while for all other mutants the activity decrease is at least partially a re sult of an effect on the enzyme structure. Steady-State Characterization As an initial step in the steady-state characterization, pr oduct formation (here formate) was followed as a function of time. For the wild type, si nce decarboxylation is irreversible and OxdC is stable in the c onditions of the assay, the reaction goes to
63 completion and formate production is linear up to the point when concentrations of residual oxalate reaches Mi chaelis-Menten constant (KM) values. The same was observed for the E162D, E333D and R270K mutants, while progress curves for E162Q, E333Q, R92K and double mutants were not linear and did not go to completion. For the R92K it was established experimentally that the activ ity loss was not due to product inhibition but enzyme instability under the assay conditions. Nonlinearity of the formate production for the E162Q mutant will be discussed below. Fo r all the mutants, even the ones with the lowest specific activity, concentration of produced formate was well above the detection limit and activities could be measured with c onfidence, since mutants were stable long enough to produce detectable product concen trations while the rate of chemical decarboxylation was insignificant. Specific activity (SA) and KM for the mutants were determined for oxalate concentrations up to 80 mM. Results are summarized in Table 5-3. For the mutants purified to homogenity turnover number ( kcat) was also calculated, although as in the case of the E162Q this is just apparent kcat since metal analysis has showed that this mutant has only 25% of the full metal occupancy. KM obtained for the mutant s varies in the range 0.1 to 2 KM of that of the wild type (Table 5-3). It was proposed in the Chapter 4 that the active site Glu residues participate in acid/base catalysis, first as acceptor of a pr oton from the monoprotonated oxalate in the first proton-coupled electron transfer and s econd as a proton donor to the formate radical anion intermediate. If the hypothesis is correct, it can be expected that mutation of Glu to Asp would affect specific activity less than the mutation Glu to Gln, since the Gln side chain lacks an ionizable group at the assay pH. As shown in Table 5-3, E162D has lost
64 only about 50% of the activity of wild t ype, while for E162Q activity dropped around 200-fold. Activity of the E333D mutant decreas ed around 10-fold compared to wild type, while E333Q had about the same activity as E162Q. Higher decrease in specific activity of the E333D compared to E162D is consiste nt with what was obser ved previously, that residues in the C-terminal domain have more effect on the overall OxdC structure than residues in N-terminal domain. Both double mu tants have about the same or even higher activities comparing to the single mutants (Table 5-3), which suggests that overall activity of each double mutant is determined by the activity of the slower site. This result supports the idea that Nand C-terminal Mn-binding site communicate, although the nature of this communi cation is not clear. Table 5-3. Steady state characterization of Ar g and Glu mutants. Specific activities (SA) were defined as a maximum rate of formate production. Turnover number (kcat) was calculated for the OxdC puri fied to homogeneity. OxdC purified omitting phenyl-Sepharose step is marked w ith * (percentage of SA relative to wild type for these mutants was calculated based on 30 I.U./mg). OxdC SA kcat KM SA %wt I.U./mg s-1 mM wt 79 2 57 5 1 100 wt* 30 3 5 1 100* R92K 1.3 0.4 0.9 10 2 1.6 E162D 40 3 29 3.1 0.7 51 E162Q 0.4 0.1 0.3 10 2 0.5 R270K* 2.1 0.4 2.6 0.6 7* E333D* 4 1 7 1 13* E333Q* 0.4 0.1 1.7 0.2 1.3* R92K/R270K* 0.4 0.1 tbd 1.3* E162Q/E333Q* 0.7 0.1 0.6 0.2 2.3* To investigate further the role of the Glu162 residue, the effect of pH on Vmax was determined and compared to the profile for th e wild type. The differe nt slope seen in the pH profile for E162Q versus wild type, as shown in Figur e 5-2, indicates that the
65 mutation has changed either the rate-limiting step, or the pKa of the group that controls the rate-limiting step. Figure 5-2. Effect of pH on log( Vmax) for the wild type and E162Q. This graph clearly illustrates that the slope of the pH prof ile change for the mutant. All assays were done in 50 mM potassium citrate buffer. Oxalate Oxidase and Dye Oxidation Activities The wild type OxdC is known to posse ss low oxalate oxidase and dye oxidation activities during turnover, which is 0.2 and 0.5 % relative to the oxalate decarboxylase activity, respectively (Tanner et al., 2001). For all prepared Arg and Glu mutants presented in this work, dye oxidation and oxala te oxidase activities were measured in a peroxidase-coupled assay (Figure 53) as previously described in the literature (Just et al., 2004). Direct oxalate-dependent dye oxidation by OxdC wild type and mutants were measured when peroxidase was not added to a reaction mixture. To get OXOXase activity, the rate of dye oxidat ion in the assay without peroxi dase was subtracted from the rate of dye oxidation in the assay with the peroxidase present. The obtained activities, -4 -3 -2 -1 0 1 2 44.555.566.5 E162Q wtlog(Vmax)pH
66 given in the Table 5-4, are the same or hi gher than it previously reported by Just and coworkers (2004). Figure 5-3. Coupled assay used to measure OXOXase activities. Table 5-4. Dye oxidation and oxalate oxidase activities for the wild type and OxdC mutants. SA / I.U./mg OxdC Dye oxidationOxalate oxidase activity wt 0.07Â±0.01 0.05Â±0.02 R92K 0.04Â±0.01 0.04Â±0.02 E162D 0.04Â±0.01 0.04Â±0.02 E162Q 0.6Â±0.1 0.56Â±0.03 R270K 0.02Â±0.01 0.01Â±0.01 E333D 0.04Â±0.01 0.04Â±0.01 E333Q 0.07Â±0.01 0.05Â±0.01 R92K/R270K 0.04Â±0.03 0.01Â±0.01 E162Q/E333Q0.04Â±0.03 0.01Â±0.01 The discrepancies between data in literature (Just et al., 2004) a nd data presented in Table 5-4 can be explained if it is consider ed that the peroxidase -coupled dye oxidation assay was not linear in time: the rate drops below its initial value after less than 20 s, making it is easy to miss the initial rate phase of the time course. The most striking observation was the 10-fold difference in activ ity of the E162Q measured here and that published previously. Results in Table 5-4 also showed that this E162Q mutant is an equally efficient oxalate oxidase (0.6 I.U./m g) as it is an oxalate decarboxylase (0.4 I.U./mg). To check these results for the E162Q mutant under the same experimental conditions for both OXDCase and OXOXase activ ities, it was necessary to design an alternative assay in which OXOXase activity was established by measuring concentration OXOX a se HC2O4 + O2 + H+ H2O2 + Dyered 2CO2 + H2O2 2H2O + Dyeox Peroxidase
67 of produced formate and residual oxalate and subtracting it from the initial oxalate concentration. Using this procedure, as s hown in Figure 5-4 OXOXase activity shows an initial burst, then drops to zero after 50 min. Beside s having 10-fold higher OXOXase activity than the wild type a nd other mutants, the E162Q va riant also has approximately 10-fold higher dye oxidation activity. This was observed with both 2,2Â´-azinobis-3ethylbenzthiazoline-6-su lfonic acid (ABTS, the dye used in OXOXase assays herein) and o -phenylenediamine (o-PDA, a reagent usually added to OxdC assays because of its beneficial effect on the oxalate decarboxylat ion activity). This enhanced dye oxidation activity of E162Q is intriguing, especially the nature of the ox idant. If the oxidant is an intermediate formed on the main catalytic pa thway or in the side reaction, the E162Q mutation may increase its lifetime or decrea se its binding. Alternatively, if E162Q mutation causes the Â“lidÂ” closing the N-te rminal Mn-binding site in the Â“closedÂ” conformation not to fit properl y during the turnover, a dye mo lecule might be able to interact more directly with the active site a nd reaction intermediates than it is possible for the unaltered protein. Lid opening can also explain the high oxalate oxidase activity of the E162Q mutant from th e perspective of the O2 reaction participant. Assuming the protonation state of this moiety is at least one determinant of the oxidation:decarboxylation branching ratio, as has been suggested from density functional studies (Chang and Richards, 2005) , if the active site is e xposed to the solvent during turnover and superoxide is not shielded from exogenous protons, it can be protonated and further reduced to hydrogen peroxide. Because of the significantly lower pKa of superoxide compared to the carbene of a fo rmate radical anion intermediate, it can be proposed that protonation of superoxide will lead to oxa late oxidase activity and not
68 absence of proton for formate radical inte rmediate. Observation that the analogous E333Q C-terminal variant does not show significantly higher OXOXase activity comparing to the wild type is consistent with the proposal from the literature that Cterminal Mn-binding site has only a structural ro le (Just et al., 2004, Escutia et al., 2005). Figure 5-4. Progress curve following oxalate decarboxylase activity () and oxalate oxidase activity (). Initial oxalate concentration was 20 mM. 13C and 18O Kinetic Isotope Effects Heavy atom 13C and 18O isotope effects (IE) were determined for the oxalate decarboxylase reaction catalyzed by R92K , R270K, E162Q and E333D mutants, following the methodology applied for the wild t ype (Chapter 4). In the case of R92K mutant it can be expected that changes in IE will reflect the effect this mutation has on chemical catalysis since metal analysis has suggested that enzyme structure was preserved. In the case of the C-terminal domain mutants it was established that mutations have an effect on the overall enzyme structur e. Considering that IE results are dependent only on active enzyme forms, IE results can hi nt if residues in the C-terminal can also 0 2 4 6 8 10 050100150200250300350c(product) / mMt / min
69 effect the catalysis and not only percentage of the active enzyme. Obtained IE results are given in Table 5-5. Since IE for the mutants did not change radi cally from those seen for the wild type, it can be assumed that mutations affected re lative rates of the kine tic steps but not the overall kinetic model (Figure 4-1). Therefore IE s are analyzed in the same way as for the wild type (Chapter 4). Results of analysis are summarized in Table 5-6. For the E162Q variant, IE results could not be analyzed si nce it is unknown what per centage of produced CO2 is coming from oxalate oxidase activity, and what from the oxalate decarboxylase activity in a conditions of IE experiments. The observation of a si gnificant new activity with a necessarily different mechanism that generates a common product to the reference reaction requires a different analytical pr ocedure. Unfortunately, insufficient data regarding either the natural oxalate oxidase m echanism or that of th e OxdC side reaction exists to develop such a procedure. Table 5-5. 13C and 18O isotope effects on the reaction ca talyzed by OxdC wild type and mutants. pH 13(V/K) / % 18(V/K) / % CO2 formate CO2 formate Wild type 4.2 5.7 0.5 0.1 0.8 0.1 1.5 0.1 1.9 0.1 -0.2 0.2 -0.7 0.1 1.1 0.2 1.0 0.1 R92K 4.2 5.7 0.4 0.1 1.1 0.1 1.0 0.1 1.9 0.1 -0.5 0.1 -0.4 0.2 0.7 0.1 0.8 0.2 R270K 4.2 5.7 0.5 0.1 1.5 0.1 1.4 0.1 2.0 0.1 -0.002 0.2 -0.5 0.2 0.4 0.2 -0.03 0.2 E333D 4.2 5.7 0.8 0.1 1.3 0.1 1.5 0.1 2.0 0.1 -0.8 0.1 -0.8 0.1 0.7 0.1 1.5 0.1 E162Q 4.2 0.9 0.1 1.5 0.1 -1.4 0.2 0.9 0.1 Results for R92K are consistent with the role proposed for th e R92 residue, which is to polarize the oxalate C-O bond and theref ore facilitate decarboxyl ation. For the R92K
70 mutant contribution of the nonpolarized resonanc e structure is higher (Figure 4-3) and as a result relative rate of decarboxylation ( k5/ k4) decreased versus the wild type. On the other hand, the relative rate of th e proton-coupled el ectron transfer ( k3/ k2) increased slightly for the R92K mutant versus the wild type. This can be interpreted as a destabilizing effect that po larization of the C-O bond has on the transition state in the kinetic step forming oxalyl radical ( k3). Table 5-6. Kinetic constant s for reactions catalyzed by O xdC mutants derived based on 13C and 18O isotope effects. IE data were analyzed using the same kinetic model as for the wild type. FF fract ionation factor, BO C-O bond order in oxalate radical anion, a:b ratio of two resonance stru ctures of oxalate radical anion where a is percentage of resonance with bond order 1 and b for bond order 1.5 (Scheme 4-2), a-calculated from 13C IE, b-calculated based on 18O IE, c-experimental and calculated value of 13C IE for formate at pH 4.2, data in brackets are one that are not consistent with the model pr obably due to large experimental error. The most interesting parameters are in bold. wt R92K R270K E333D k5/ k4 4.00 2.64 2.08 2.08 18k3 1.016 1.021 1.021 1.017 13Keq3 1.021 1.017 1.015 1.015 FFa 0.963 0.968 0.969 0.969 BOa 1.16 1.22 1.25 1.25 a : ba 70:30 55:45 50:50 50:50 18Keq3 1.016 1.011 (0.998) (1.021) FFb 0.952 0.956 BOb 1.14 1.22 k3/ k2 0.75 2.41 2.36 0.92 calc. IEc 1.2 0.7 (0.8) 1.2 exp. IEc 1.5 1.0 (1.4) 1.5 For R270K and E333D, IE results cannot be interpreted in as straightforward manner as for R92K because of effects thes e mutations have on the enzyme structure. The changes in the isotope effects relative to the wild type showed that drop in specific activity for these mutants is not due only to enzyme instability and lower concentrations of active enzyme but also to an effect on cat alytic mechanism. Both C-terminal mutants R270K and E333D have similar effects on k5/ k4 and the C-O bond polarization. If E333
71 residue participates in acid/base catalysis it can be expected that mutation to Asp, a residue with the same functi onal group, would not effect IEs significantly. IE results suggest that C-terminal mutant s have a structural effect on the N-terminal active site in the way that active site residues participating directly in catalysis are not properly aligned during the turnover. If the relative rates k3/ k2 and k5/ k4, calculated based on isotope effects, are plugged into the equation 4-1, (V/K)/[E]0k1 can be calculated (see equation 6-1). 4 5 2 3 4 5 4 5 2 3 0 1/ / ) / ( 1 / / /k k k k k k k k k k E k K V (6-1) The Â“( V/K )/[ E ]0k1Â” value on its own lacks much physical meaning; however, if combined with experimental Vmax/ KM values, it can be used to estimate [ E ]0k1 for each mutant. In Table 5-7 [ E ]0k1 has been expressed as a percentage of the [ E ]0k1 for the wild type. For the R92K this number reflects drop in the rate of substrate binding ( k1) relative to the k1 of the wild type, which sugge sts that role of R92 in the OxdC catalysis is also to ensure productive binding of oxa late. For other two mutants k1 is less affected or not affected at all, and decrease in the [ E ]0k1 value relative to wild type reflects lower percentage of the active OxdC forms relativ e to the total protein concentrations. Table 5-7. Comparison of the ( V/K )/[ E ]0k1 calculated based on IE data and (V/K)exp determined based on activity assays (T able 5-3). Comparison of two values gives a product of active enzyme con centration and oxalate binding constant (k1) relative to the wild type. ( V / K )/[ E ]0k1 (V/K)exp [ E ]0k1% wt wt 0.375 16 100 R92K 0.636 0.35 1.3 R270K 0.615 2.45 9.3 E333D 0.383 0.65 3.9
72 Experimental Section Most of the experimental procedures app lied in the research described in this chapter are already described in the previous chapters: site-directed mutagenesis, enzyme expression and purification, size-exclusion chromatography, metal analysis, protein determination and formate assay in Chap ter 2 and heavy atom isotope effects experimental set up and data analysis in Chapter 4. Only the specific changes in the experimental procedure will be mentioned he re. R92K and R270K mutants were cloned, transformed, expressed and purified by Ewa Wroclawska. Purification. All N-terminal mutants were purified as previously described for the wild type. In the case of Cterminal and double mutants puri fication on phenyl-Sepharose column did not result in an increase in speci fic activity and was omitted. Wild type OxdC purified only by ion-excange chromatography yi elded specific activity of 30 I.U./mg. Formate assays. Oxalate decarboxylase activity for all the mutants were measured by formate assay. In the cas e of E162Q mutant 0.5 mM o -PDA was not added to the reaction mixture, since when added it did not improve OxdC activity as for the wild type and other mutants. pH profiles for the wild type and E162Q were done in the 50 mM potassium citrate buffers with a pH 4.2-6. 7. Because of higher buffer capacity citrate buffer has relative to other buffers, 100 l reaction mixture was quenched with 20 l 1M NaOH. Oxalate oxidase and dye oxidation assays. Oxalate oxidase activity for wild type and all mutants were determined spectrophoto metrically using a continuous assay in which the production of hydrogen peroxide was coupled to the oxidation of 2,2Â´azinobis-3-ethylbenzthiazoline-6sulfonic acid (ABTS) usi ng horseradish peroxidase
73 previously described by Requena and Bornem ann (1999). For the E162Q mutant oxalate oxidase activity was also measured with an alternative assay. Composition of this assay mixture was the same as for the standard oxalate decarboxylase assay: 50 mM potassium acetate pH 4.2, 0.2 % Triton X-100, 20 mM pot assium oxalate pH 4.2, 0.2 mg/ml E162Q. Reaction was started by addition of substrate. 50 l aliquots were quenched with 10 l 1M NaOH after 5 to 300 min. Concentration of the formate produced in the aliquots was measured by end-point formate dehydrogenase assay described previously and residual oxalate was measured using the Oxalate detection kit from Trinity Biotechnology. 13C and 18O isotope effects. Isotope effects for the mutants were determined as previously described in Chapter 4. Consideri ng that the specific activities of the mutants were lower than for the wild type, reactions had to be run longer to reach 50 and 100 % oxalate conversion. Also, portions of fresh en zyme had to be added occasionally during the reaction.
74 CHAPTER 6 IDENTIFICATION AND STUDIES OF TYROSYL RADICAL FORMED DURING TURNOVER Introduction Kinetic and chemical mechanism of OxdC catalyzed oxalate decarboxylation consistent with the observed 13C and 18O isotope effects are described in detail in Chapters 4 and 5. Proposed mechanism incl udes several radical species. It can be expected that those radicals are transient and do not build up to the steady state levels. Progress curves also showed that OxdC is stable during the tu rnover, implying that chemistry of highly reactive radical intermed iates is well controll ed. It was therefore surprising when a sharp g = 2.0 EPR signal with a peak-to-trough width of ~23 G was observed upon exposure of enzyme to oxalate , as shown in Figure 6-1 (Chang et al., 2004). The signal width is too narrow to arise from a transition meta l-associated signal, implying that an organic radical is formed during oxalate turnover. Generation of this species required the simulta neous presence of enzyme, oxa late, and oxygen, and was not observed in any of the isolated reacti on components. The requirement for the simultaneous presence of O2, enzyme, and oxalate for both activity and radical formation strongly suggests that the obser ved species is related to en zymatic catalysis, although not necessarily on the main catalytic pathway. Experiments done in our laboratory by C.H. Chang showed that oxalate isotopically enriched with 13C showed no detectable change in the radical signal therefore eliminating a meta-stable oxalyl or even formyl intermediate as a possibility for the paramagnetic site . Enrichment of cofactor dioxygen with 17O also
75 produced no change in spectral lineshape. Ha ving eliminated known reaction participants, the most likely alternative identity for the radical host is a protein residue. The OxdC radical observed manifests additional structur e at higher temperatures that might give easy clues to its identity. Unfortunately, the radical quenches at temperatures that might allow more clear manifestation of motionally averaged hyperfine structure, limiting the available signal-to-noise rati o. Nevertheless, several lines of evidence establish this species as a tyrosyl radical. Figure 6-1. EPR spectrum of OxdC frozen 30 seconds after oxalate addition. Perpendicular mode, 5 K, 0.1 mW, 4 scans. Sample was 150 ÂµM OxdC subunits, generated by addition of 100 mM potassium oxalate pH 5.2. The lineshapes at low or higher temperatures are inconsistent with other radicals that have been observed in proteins, such as glycyl (Wagner et al., 1992), cysteinyl (Lassmann et al., 2003), or tr yptophanyl (Lendzian et al., 199 6). Some similarity exists between the spectral envelope of the presen t species and that of a sulfinyl radical observed during oxygen-dependent inactivation of pyruvate-formate lyase (Reddy et al., 1998). However, when the cysteine residue wa s mutated (Chapter 2, there is only one Cys in the OxdC sequence), the C383S mutant formed the observed radical at a level and
76 rate similar to unaltered enzyme (data not s hown) and therefore ruled out formation of cysteinyl radical. The goals of the research presented in this chapter was to give more through characterization of observed or ganic radical, determine its kinetic competence and test the hypothesis that observed organic radical is tyrosyl radical. Results and Discussion Variation of Radical Intensit y as a Function of Time Quantitation of the observed organic radica l against a 1,1-diphenyl-2-picrylhydrazil (DPPH) standard allow an intensity estimate equivalent to ~4 to 8 % of potential manganese sites in the sample (manganese o ccupancy varied in the range ~50% to 100% for different enzyme preparations used in EPR experiments). Samples containing formate and bicarbonate without oxalate did not show a radical spec ies, consistent with the presence of an effectively irreversible step subsequent to the pr oduction of the radical species in the forward reaction. A thorough examination of radical intensity and substrate consumption versus time (Figure 6-2) shows development of both to be at least biphasic. Radical intensity increases monotonically during steady-state turnover, and peaks approximately as turnover becomes substrate-limited. As form ate production approaches its asymptotic limit, the radical decays back toward zero in tensity. In addition, the lineshape of the signal was observed to become gradually narrower as the signal decayed, possibly reflecting a distribution of hyperfine c ouplings arising from multiple radical conformations. In order to extract an approxi mate rate of radical formation, several potential kinetic mechanisms were numerically simulated. The solid lines overlaid on the radical intensity data and formate concentrations of Figure 6-2 represent the simulated
77 intensities and concentrations, respectively, of the simplest model that simultaneously fit both data sets, shown in Figure 6-3. The onl y explicit constraints on the model were the initial oxalate concentration of 100 mM a nd the enzyme concentration in the EPR samples. Figure 6-2. The steady-state OxdC radical intensity as a function of reaction time. Intensity was calculated from double-in tegration of the simulated radical signal and comparison with a DPPH concentration standard. Reaction conditions were 0.5 M piperazine/piperÂ·H+ pH 5.2, [oxalate]0 = 100 mM, [OxdC] = 15.9 ÂµM. All spectra were ta ken at 4 K with 0.2 mW microwave power, 10 G modulation amplitude, and were the sum of 4 scans. Inset : formate as a function of reaction time. Data were simulated (solid lines) according to the kinetic mechan ism presented in Figure 6-3. K1 (Figure 6-3) is not a simple dissoci ation constant but instead reflects all forward and reverse rate constants up to the hypothetical branch point in which the steady-state radical is formed. Numeri cal simulation gave a value of 14 mM-1Â·min-1 for K1. However, the dependence of the overall mo del on this value was found to be low as reflected by large standard deviations in the constituent effective rate constants k1 and k-1. Thus, the particular value for K1 has little chemical meaning or kinetic consequence. On the other hand, first-order rate constants for radical formation and decay within the model
78 presented in Figure 6-3 were found to be k3 = 0.02 Â± 0.001 min-1 and k-3 = 0.16 Â± 0.013 min-1, respectively. The former is of special in terest in connection with overall enzymatic turnover. Figure 6-3. Kinetic model for the steady-stat e radical observed dur ing OxdC catalysis. E is free enzyme, F represents all kineti cally equivalent enzyme species on the catalytic pathway, and R is the form of the enzyme containing the observed steady-state radical. Kinetic constants from numerica l simulations using this general mechanism are k2 = 632 Â± 1.7 min-1; k3 = 0.020 Â± 0.001 min-1; and, k-3 = 0.16 Â± 0.01 min-1. The enzymatic kcat of 420 min-1 implies that radical formation is not kinetically competen t and thus not on the main catalytic pathway. If the presence of the radical is required duri ng the catalytic cycle, either directly or indirectly as a marker for an obligate intermed iate, then it is expected that the radical intensity should be proportional to the rate-ofchange of oxalate with time, i.e. the slope of the formate production curve s hould increase as more radical is formed. This is not the case, as seen in the Figure 6-2 inset. The magnitude of the slope in the Figure 6-2 inset is maximal at zero time, whereas the radical reaches maximum intensity only after ten minutes. Although the formate concentrations at early time points might suggest a positive curvature and thus an activation process, the inflection of the [formate] vs. time curve does not occur at the maximum of ra dical intensity, as would be expected. Therefore it can be conclude d that the radical species doe s not represent a by-product of an obligatory activation pr ocess. In addition, kcat under the experimental conditions is
79 calculated to be ~420 min-1 based on the slope of the Figure 6-2 inset at t=0. The rate of primary radical formation (0.02 min-1) is thus approximately f our orders of magnitude lower, ruling out its participation as a kine tically competent intermediate in the reaction mechanism. The existence of a paramagnetic in termediate is of course not ruled out by the present study; however, observed radical is not the steady-state radical, nor does it accumulate to observable levels during stea dy-state turnover under these experimental conditions. Effect of Temperature and Microw ave Power on Radical Signal Figure 6-4 shows the spectra of the ra dical observed during turnover at two temperatures 5 K (Figure 6-4A) and 140 K (Figure 6-4B). At higher temperatures additional structure can be seen, consistent with increased molecular motion averaging out hyperfine anisotropy and the associat ed broadening that manifests at lower temperatures (Weil et al., 1994). At temp eratures higher than 200 K the radical was observed to quench rapidly, consistent with increased protein flexib ility allowing electron transfer to or from this radical. Notably, in experiments in which the reaction is frozen within a few minutes after oxalate addition a nd then annealed, the radical re-forms upon thawing, brief incubation, and re-freezing, a nd likewise anneals at high temperature (data not shown). Repeated cycles in this fashi on eventually fail to produce a radical signal because all oxalate has been consumed, but upon addition of a sma ll amount of oxalate the signal reappears. This behavior clear ly illustrates a connection between radical formation and enzymatic turnover. Frozen annealing does not permit exchange of products and substrates between the enzyme active site(s) and the bulk medium, but allows the radical to quench through enhanced molecular mo tion and interaction with a
80 redox-active species. Upon thawing, fresh oxala te can bind to the enzyme, and initiate further turnover with concomitant production of the radical species. Figure 6-4. Effects of temp erature and isotopic editing on the observed radical spectrum in OxdC. (A) 5 K without isotopic ed iting, 1 G M.A., 0.1 mW, 4 scans; (B) 160 K without isotopic editing, 2 G M. A., 2 mW, 10 scans; (C) 120 K with enzyme containing tyrosine residues uni formly on the ring with deuterium, 4 G M.A., 0.8 mW, 16 scans. Higher temp eratures allow sufficient motion for partial averaging of hyperfine anisot ropy, narrowing the hyperfine lines and allowing subtle features to be obser ved that would be hidden at lower temperatures. The instability of the observed radical species even in frozen samples was intriguing, and the temperature dependence inve stigated further. The tyrosyl radical was
81 generated by limited enzymatic turnover, and a sample frozen for spectral analysis. Cryostat temperatures were slowly raised in 10 or 20 K steps, and ra dical signal intensity monitored in time (Figure 6-5). Contrary to finding a temperature at which the radical decayed completely to zero, or temperat ure dependence of a single observed quenching rate, the radical signal intensity was found to quench quickly and irreversibly, but only partially as temperature was raised, reaching an asymptotic intensity at each temperature. Upon lowering the cryostat temperature to 100 K, where radical signal intensity was not observed to decrease with time relative to 10 K spectra, signal intensity did not recover Figure 6-5. Effects of temperature on the obs erved radical spectrum in OxdC. Intensity is defined as the peak-to-peak difference of the derivative signal. All spectra were taken with 0.6315 mW microwav e power, 10 G modulation amplitude, and 1 scan. ruling out temperature broadening as an explan ation. The existence of asymptotic value of radical concentration that decrease as temperature is raised suggests that only subpopulations of the radical quench at each te mperature. Thus, at low temperatures, the radical and its motionally coupled environm ent can adopt an arrangement that is conducive to electron transfer in only a small fraction of protein molecules. Each time temperature is increased, a greater fraction of the proteins can access the conformation 50 100 150 200 250 300 350 120140160180200IntensityT / K
82 required for electron transfer and radical quenching. Therefore at a given temperature the intensity of the radical signal decreases until it reaches an asymptotic value, which corresponds to all of the pr otein molecules in which th e radical does not adopt a conformation suitable for electron transfer. The response of radical signa l intensity to applied micr owave power at 10 K is shown in Figure 6-6. Fitting to equation 6-1 gives a characteristic half-saturation power of 1.1 Â± 0.1 mW and an inhomogeneity parame ter value of 1.34 Â± 0.02. These values are consistent with an organic radical. The quality of the fit in Figu re 6-6 suggests that only one radical species is present at the reaction time examined here ( 2 minutes). The fitted half-saturation power of 1.1 mW is lower than would be expected for a transition metal or for a radical tightly coupled magnetically to one. Figure 6-6. Power saturation of the g = 2.0 si gnal formed by OxdC in the presence of oxalate and dioxygen. The stable radical formed in ribonucleo tide reductase exposed to the cancer therapeutic gemcitabine (2Â´,2Â´-d ifluoro-2Â´-deoxycytidine) exhi bits a PÂ½ of 1.0 Â± 0.5 mW (van der Donk et al., 1998), whereas the tyrosyl radical in the R2 protein that is in close proximity to the dioxo-diferric activating metal cluster has a characteristic half-saturation
83 power of 47 Â± 12 mW (Gerfen et al., 1998). The latter is greater than the former due to the dipolar-enhanced relaxation provided by its proximity to the diiron center. By analogy, low measured PÂ½ of 1.1 mW suggest s that the radical formed in OxdC is relatively isolated magnetically from any Mn center, which may have initiated the immediate reactive precursor. Identification of a Tyrosyl Radical To investigate if the observed radical is a tyrosyl radical OxdC was prepared with all tyrosine residues substituted with L-4-hydroxyphenyl-2,3,5,6d4-alanine ( d4-tyrosine). Isotopically labeled protei n was prepared by growing E. coli cells in a minimum media supplied with the N -phosphonomethyl glycine, an inhibi tor of biosynthesis of aromatic amino acids (Kim et al., 1990), and phenylalanine, tryptophan, and d4-tyrosine (Figure 67). Incorporation of deutereted tyrosine was confirmed by tryptic digestion and mass spectroscopy. Spectral narrowi ng upon deuteration of the tyrosyl ring gave the most convincing and direct evidence that observed radical is tyro sine. Substitution of tyrosyl ring protons for deuterons collap ses the observed splittings cleanly, so much so that an inflection becomes visible at high observat ion temperatures even without resolution enhancement (Figure 6-4C). A sim ilar spectrum was observed for the YZ tyrosyl radical of the Mn-depleted Synechocystis 6803 photosystem II upon deuteration of the ring positions orthoto the hydroxyl group (pos itions 3 and 5) (Tommos et al., 1995). In Search for Tyrosine Residues Forming the Radical In the mechanism catalyzed by OxdC propos ed in Chapter 4 there is no tyrosine residues involved in the catalysis. Data pres ented here (Figure 6-2) have demonstrated that formed radical is not catalytically competent and mechanism of its formation or its function are unclear. Identification of specifi c residue forming the tyrosyl radical and
84 effects its mutation might have on the reaction rate, specificity or enzyme stability can help explain role and the mechanism of radi cal formation. In the OxdC sequence there are 14 tyrosines (Figure 6-8) . According to sequence ali gnment of several OXDCase (Dunwell et al., 2000) only few of tyrosine s are conserved: Y104, Y133, Y200, Y305 (Figure 6-8). Tyrosines closest to the Mn-bi nding site (about 5 Ã…) ar e Y200 in N-terminal domain and Y340 in C-terminal domain. Fo r both single mutants Y200F and Y340F and also double mutant Y200F/Y340F unimpaired tyrosyl radical EPR signal was observed during the turnover. Figure 6-7. Optical density indicating growth of E. coli cells in time. The first stationary phase is reached when all unlab eled tyrosine was consumed. E. coli cells continued growing after addi tion of labeled tyrosine ( d4-Tyr). Chemical structure of d4-Tyr and glyphosate was shown on left. This result, together with the modestly perturbed half-saturation power, implies some distance between the radical and the ac tive-site manganese ions. A seemingly welldefined tyrosyl ring torsion angle does not support radical formation on the protein surface, which could arise from non-speci fic oxidation by reactive side products. 0 0.5 1 1.5 0 5 10 15 20 25 time / h Glyphosate (1g/L) Phe, Trp (50 mg/L) Tyr (7 mg/L) d 4 -Tyr (50 mg/L) Induction NH2O H D CO2H D D N P HO2C H OH O OH D
85 For the other two conserved Tyr residues: Y104 and Y133, mutants were prepared and tested for radical formation. For each mu tant unperturbed tyrosyl radical EPR signal appeared during the turnover. Interestingly, for the Y104F half power saturation was 3.5 fold higher than half power saturation for the wild type tyrosyl radical (Table 6-1), which might suggest that alternative tyrosyl radica l was formed or mutation caused change in protein environment in proximity to the radical. Figure 6-8. OxdC monomer show ing distribution of tyrosine and tryptophan residues. Conserved Tyr residues are shown in red and Trp residues in yellow. Although prepared mutants did not help iden tify the residue that forms the radical, they gave some interesting results: for some of the mutants specific activity significantly dropped relative to the wild type (Table 6-1). The highest decrease was observed for Y200F, which might be explained by its proxim ity to the active site where it participate in hydrogen bonding network and helps orient bonded oxalate as proposed by Just and co-workers (2004). For oxalate decarboxylase it is in teresting that a ll Trp residues (7 total) are conserved and positioned along both domains. Trp residue closest to the N-
86 terminal Mn-binding site was mutated a nd the mutant W132F has shown significant decrease in the oxalate decarboxylation and dye oxidation specific activity (Table 6-1). This result can be explained by looking in the OxdC crystal structur e, which shows that W132 residue hydrogen-bonds Glu101, a ligand for Mn ion in N-terminal domain. Therefore W132F mutation can indirectly affect redox properties of Mn centar. Table 6-1. Results of kinetic and EPR char acterization of some of the Tyr and Trp mutants (Â“tbdÂ”-to be determined). Mutant SA/ %wt KM/ Dye oxidation EPR Signal P1/2 I.U./mg mM as %wt mW Wt 79 2 100 5 1 100 + 1.1 0.1 Y200F 26 3 33 5 2 120 + tbd Y340F 56 11 70 5 1 90 + tbd Y200F/Y340F 12 4 15 6 2 90 + tbd Y104F 49 9 62 4.5 0.5 tbd + 3.9 0.4 Y133F 63 5 80 4.3 0.5 tbd + 1.2 0.1 W132F 14 2 18 5 2 17 tbd tbd Spin-trapping experiments The formation of free radicals by OxdC during turnover was monitored by EPR spin-trapping studies. The addi tion of substrate to OxdC assay under aerobic conditions in the presence of 5,5-dimethyl-l-pyr roline N-oxide (DMPO) as well as 5diethoxyphosphoryl-5-methyl-1-pyrro line N-oxide (DEPMPO) lead s to the appearance of a sharp, free radical solution EPR spectrum (Apendix C, Figures C-1). In the control experiments where OxdC or oxalate was omitted, the signal did not appear, which imply that the trapped radical is produced only under turnover conditions . To clarify whether the radical formed was enzyme bound, OxdC was separated from the small molecules using a G-25 gel filtration column. For the fractions that contained enzyme, which was detected using UV-Vis spectrometry, there was no EPR signal and therefore no trapped radical present (Figure C-2). It was in frac tions with no enzyme that the spin-trapped
87 radical was found (Figure C-2). The species tr apped thereby cannot be the tyrosyl or any other protein-based radical. Sp in-trapping experiments have given another evidence that formed tyrosyl radical is not formed on the protein surface. Since OxdC in the solution forms hexamers, the radical could be also formed between the interfaces of two monomers, thereby hindering the spin trapÂ’s access. Another possibili ty is that tyrosyl radical is not formed in the presence of spin-traps, if trapped radical is tyrosyl-radicalforming species. This hypothesis could have been checked if EPR spectrum was taken below 100 K. Spin-trapping experiments described here s howed that such a technique can be used as a tool for detecting radical species form ed during OxdC turnover on main or auxiliary catalytic pathway. Spectral shape of the EPR si gnal of the trapped ra dical suggest that it can be hydroxyl, or other, oxygen based radical, but this should be further confirmed by analyzing spin-trapping product by mass sp ectroscopy. The OxdC activity assay performed in the presence of DMPO shows th at with increasing amounts of spin trap reagents present during the reaction, the ra te of formate production is not significantly altered (Figure C-3). If the species trapped was an intermediate, theoretically the rate of product formation should decrease, as unrea ctive spin-trapped intermediate cannot proceed to form the desired product. A confoundi ng factor that could explicate the lack of decline in formate production is that concentr ation of DMPO used in the assay was too low. Potential Functional Implications of the Observed Radical A tyrosyl radical is conspicuously missing in the mechanism proposed in Chapter 4 (Figure 4-4). The necessity for such a species in the catalysis of oxalate decarboxylation is not clear at present, and none of the pr esented data speaks to a requirement for a
88 protein radical in the OxdC catalytic mechan ism. However, such a requirement is not ruled out, and there are now several well-ch aracterized examples in the literature of enzymes, which employ an obligate protein ra dical that is only observable as a more stable species. The type I ribonucleotide reductase from E. coli forms a catalytic cysteinyl-439 radical in its R1 subunit (Stubbe et al., 2003). This radical is formed from hydrogen atom abstraction by tyrosine-122 in the R2 subunit, itself formed by action of the essential diferric cluster. Py ruvate-formate lyase (PFL) from E. coli initiates each turnover from a stable glycyl radical at re sidue 734, and like the type I ribonucleotide reductase employs the more stab le radical as a Â“holsterÂ” for a more reactive one, in this case cysteinyl-419. The E. coli type III RDPR employs an analogous scheme, with glycyl-681 (Sun et al., 1996) serving the role of Gly-734 in the PFL. The sulfur-based radicals do not accumulate during turnover, but instead are seen as the more stable delocalized glycyl or tyrosyl radicals that ar e corollary to the primary reduction reaction. Identification of the paramagnetic species as a tyrosyl radical advances mechanistic reasoning substantially. The radical forms only under turnover conditions, which might imply that a residue near the active site is involved. However, as discussed above, an obligatory radical-mediated activation or pa rticipation of the observed radical as a reaction intermediate contradi ct the data. Furthermore, fo rmation of radical in Y200F, Y340F, and Y200F/Y340F active site mutants t ogether with the modestly perturbed halfsaturation power imply some distance between the radical and the active-site manganese ions. A seemingly well-defined tyrosyl ring torsion angle does not support radical formation on the protein surface, which c ould arise from non-specific oxidation by reactive side products. Indeed, th e data taken together raise the possibility of long-range
89 electron transfer through the pol ypeptide framework. In the Chapter 5 it was established that two Mn binding sites in the Na nd C-terminal domain communicate somehow. Although, data suggested that the nature of this communication is st ructural and that Cterminal Mn-binding site has predominantly st ructural role, those da ta did not role out possible electronic interaction between the two Mn-binding sites. If such an electron pathway exist than the radical-forming tyro syl residue might be on the pathway or in proximity of main electron transfer pathway. The precise mechanistic rationalization of the identified OxdC tyrosyl radical, its catalytic connection with the divalent metal centers and possibility for l ong-range electron tran sfer await further study. Especially interesting will be to determine if any kineti cally competent radical is formed during the OxdC turnover in pre-steady state conditions. Experimental Section Materials. Potassium oxalate, oxalic acid, and Amicon ultrafiltration membranes were obtained from Fisher Chemical. d4-L-Tyrosine, the ProteoProfile Trypsin In-Gel Digest Kit, N,N-bis(2-hydroxyethyl)-2-aminoe thanesulfonic acid (BES) and piperazine were from Sigma-Aldrich. All chemicals were used without furthe r purification. Sitedirected mutagenesis of OxdC tyrosine a nd tryptophan residues to phenylalanine was carried out using the QuikChange as described in previous chapters. Preparation of native and [d4-Tyr]-labeled oxalate decarboxylase. Isotopically neutral OxdC protein was over-expressed, purif ied and assayed as described in Chapter 2. Incorporation of ring-deuterat ed tyrosine was achieved by gr owing cells in M9 minimal medium supplemented with 50 mg/L phe nylalanine, 35 mg/L tryptophan, 5 mg/L tyrosine, and 1 g/L N-(phosphonomethyl)glycine (glyphosate) in order to inhibit de novo biosynthesis of aromatic amino acids (Kim et al., 1990). The culture was initiated with a
90 limiting concentration of unlabeled tyrosine ra ther than the isotopically labeled form in order to minimize the possibility of deuter ium incorporation into other amino acids through metabolic scrambling during bioma ss accumulation. Medium was inoculated with a freshly transformed si ngle colony of bacterial cells and shaken at 250 RPM at 37 Â°C. When the cell culture star ted to leave the exponential growth phase due to limiting tyrosine (A600 ~ 0.8), 50 mg/L of L-4-hydroxyphenyl-2,3,5,6-d4-alanine ( d4-tyrosine) was added. Since the yield of expression a nd specific activity of OxdC depends on the ability of cells to express heat shock proteins before induction, the culture was allowed to reach exponential growth phase again (A600 ~ 1.2) before heat-shocking at 42 Â°C for 18 min and subsequent addition of isopropyl -D-galactothiopyranoside (IPTG) and MnCl2 to final concentrations of 1 and 5 mM, respectively. Addition of MnCl2 caused formation of a white precipitate most likely due to fo rmation of the manganous glyphosate complex. Cultures were harvested by cen trifugation after three hours of induction at 37 ÂºC, and enzyme manipulated and purified as for the unlabeled protein. Tryptic digestion and mass spectroscopy. d4-Tyrosine incorporation was verified by tryptic digestion of labeled OxdC and mass sp ectral analysis of the fragments. Purified OxdC (18 mg) was loaded onto a 12% SDS pol yacrylamide gel, and prepared according to protocols supplied with the ProteoProfile kit. The mass profile of the peptide mixture thus prepared was analyzed on a PerSeptive Biosystems Voyager DE-PRO matrixassisted laser desorption i onization time-of-flight mass spectrometer at the Protein Chemistry Core Facility, Biotechnology Program, University of Florida. Preparation of samples for EPR spectroscopy. Unless otherwise noted, EPR sample pH was buffered with 50 mM sodium acetate at pH 5.2. Enzyme preparations
91 were dialyzed, chromatographed over G-25 size ex clusion resin, or di afiltered to remove free Mn2+. Reactions were initiated by addition of oxalate. After transfer of the necessary reaction components to the EPR sample tubes, samples were frozen within 1-2 seconds by immersion into an isopentane bath prep ared by cooling with a secondary liquid nitrogen bath until rapid nitrogen boiling subsided. Temporal evolution of the OxdC ra dical signal under turnover conditions. A 5 mL reaction mixture was prepared contai ning 2 mg/ml OxdC, 50 mM potassium acetate buffer pH 5.2. The reaction was begun by additi on of potassium oxalate to 100 mM final concentration. Aliquots (500 ÂµL) were quenc hed at different ti mes by cryogenic flashfreezing for EPR samples, or by adding NaOH to 0.1 M final concentration for oxalate consumption measurements. Radical quenching in the frozen state. To measure the lability of the observed radical in the temperature range used for Â“high-temperatureÂ” spectral acquisition, a sample of enzyme undergoing turnover was fl ash-frozen two minutes after introduction of substrate oxalate. Temperature was raised to 100 K, a reference spectrum taken, and radical intensity monitored for 30 minutes at each temperature 120 K, 130 K, 140 K, 150 K, 160 K, 180 K, and 200 K. Single scans were acquired at 0. 6 mW with a 10 G modulation amplitude and a 10.5 s sweep tim e.In order to rule out temperature broadening or reversible ch emical processes as the cause of the decreased signal intensity, after acquiring the la st scan at 200 K, cryostat te mperature was lowered to 100 K and a spectrum acquired under the same conditions as at the beginning of the experiment. The sample was then thawed br iefly and refrozen to test for continued turnover. The presence of an inte nsified radical signal relative to that seen in the final 200
92 K scan confirmed the presence of oxalate in the solution, showing that the enzyme was initially frozen under turnover conditions prio r to complete consumption of substrate. Spectral acquisition and analysis. CW perpendicularly polarized X-band spectra were acquired on either a Bruker Elexsys model E580 spectrometer operating in TE101 mode, or a Bruker model ECS106 with a dual-mode cavity opera ting in TE102 mode. Temperature control on both instruments was maintained with an Oxford Instruments model ESR900 helium flow cryostat coupled to an Oxford ITC 503 temperature controller. Time course data for the devel opment and decay of the OxdC radical were numerically simulated to test potential kine tic models using the program package Gepasi (Mandes, 1993, 1997). The rate constants thus derived were compared to the turnover number in EPR samples, which was calculate d based on the initial slope of a plot of [formate] vs. time. For measurement of PÂ½, the radical signal intensity was taken as the magnitude of the peak-to-trough difference. Data were tr ansformed and plotted in order to fit parameters of the standard equation (Rupp et al., 1978; Padmakumar and Banerjee, 1995; van der Donk et al., 1998): 2 / 11 log 5 . 0 log log P P b K P S (6-1) where S is the peak-to-peak derivative signal amplitude, P is the microwave power, K is a factor accounting for vertical displacements of the plot and depends on the sample and instrument (van der Donk et al., 1998), b is an inhomogeneity pa rameter varying from 1 (completely inhomogeneous broadening) to 2 (completely homoge neous broadening), and PÂ½ is the characteristic half -saturation power of the signa l. Fitting was done with the commercial plotting package Kaleidagra ph (Synergy Software, Reading, PA).
93 Spin trapping. Samples for free radical spin-trapping experiments were prepared with 3.45 mg/mL OxdC enzyme, 0.15 M acetic acid buffer pH 4.2, 450 mM 5,5dimethyl-l-pyrroline N-oxide (DMPO) or 107 mM 5-diethoxyphosphoryl-5-methyl-1pyrroline N-oxide (DEPMPO), and 75 mM oxalate pH 4.2. Controls were made, with the same composition, except OxdC storage bu ffer were added without the enzyme. EPR spectra were taken at room temperature. To assess whether the radical trapped by DMPO was enzyme bound, gel filtration using a G-25 column was performed. A r eaction mixture containing 3.45 mg/mL OxdC, 0.15 M potassium acetate buffer pH 4.2, 450 mM DMPO, and 75 mM oxalate pH 4.2 was incubated at room temperature for 2 minutes before examined through EPR to ensure that a radical was trapped. The mixture was then poured onto the column. Fractions were collected and their absorbances measured using a UV-Vis spectrometer. EPR spectra of all the fractions were obtained to determ ine which fraction contained the spin-trap (Figure C-2).
94 APPENDIX A PROTEIN PURIFICATION CHROMATOGRAMS The following figures show typical chroma tograms obtained during the purification of OxdC wild-type and mutants. A) DEAE-Sepharose Fast Flow, washed with 50 mM imidazole hydrochloride buffer pH 7, developed with 500 ml NaCl gradient (0 1 M) 0 500 1000 1500 2000 mAu 0 20 40 60 80 100 120 min Method Run 10/12/2004, 4:51:10 PM, Method : , Result : C:\...\prime\Manua Figure A-1. DEAE-Sepharose Fast Flow ch romatogram. Arrow points out to the peak that showed OxdC activity and was colle cted and purified further. Blue line represents absorbance, red line conductiv ity and green percentage of high-salt buffer (0-100%). B) phenyl-Sepharose Hi-Performance, washed with 50 mM imidazole hydrochloride buffer pH 7, developed with (NH4)2SO4 gradient (1.7 0 M). Representative chromatogram for wild type and all mutants (top) except R270K, E333D, E333Q, R92K/R270K and E162Q/E333Q (bottom) are given in Figure A-2. Blue arrow indicates the least hydrophobic peak and one with the highe st OxdC specific activity,
95 which was collected and purified further. Black arrows indicate peaks with some OxdC activity. -50 0 50 100 mAu 0 20 40 60 80 100 120 min 1 2 3 4 5 6 7 8 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65 67 69 71 73 75 77 79 81 83 0 50 100 150 mS/cm 0 20 40 60 80 100 min Method Run 9/22/2004, 3:30:51 PM, Method : , Result : C:\...\prime\Manual 1 2 3 4 5 6 7 8 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65 67 69 Figure A-2. phenyl-Sepharose Hi-Performance ch romatogram for the wild type and most of the mutants (top) a nd R270K, E333Q, E333D and double mutants (bottom). Arrows point out to the peaks that sh owed some OxdC activity. Blue arrow point to a peak with the highest OxdC specific activity that was collected and purified further. Blue line represents absorbance, read conductivity and green percentage of the buffer with no AS (0-100%). C) Q-Sepharose Hi-Performance, washed with 50 mM imidazole hydrochloride buffer pH 7, developed with 400 ml NaCl gradient (0 1 M).
96 Manual Run 0:1_UV Manual Run 0:1_Cond Manual Run 0:1_Conc Manual Run 0:1_Fractions 0 500 1000 mAu 0 50 100 150 200 250 300 350 ml 1 3 4 5 6 7 8 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 5960 Figure A-3. Q-Sepharose Hi-Performance chro matogram. Arrow points out to the peak that showed OxdC activity and was colle cted and purified further. Blue line represents absorbance, read conductivity and green percentage of high-salt buffer 0-100%.
97 APPENDIX B DERIVATION OF EQUATIONS FOR HEAVY ATOM ISOTOPE EFFECT Although oxalate (A) is symmetrical, the tw o ends of the molecule when bound to OxdC are not equivalent and therefore yi eld different products, formate (B) and CO2 (C). If the heavy isotope is at th e end of the substrate that becomes formate (shown as *A below), this gives an IE of kb*. Equally, if the heavy isotope is at the end of the substrate that becomes CO2 (shown as A* below), this gives an IE of kc*: A B C BC k + *A A* kb* kc* If we define A = [oxalate], B = [formate], and C = [CO2], we can write A + B = A0 B = C A*+B*+C* = A0* R0 = A0*/A0 f = B/A0 = C/A0 1-f = (A0-B)/A0 The conversion of oxalate to products is desc ribed by the following differential equation: -dA/dt = kA = dB/dt = dC/dt (B-1) and the effect of isotopic substitution leads to dB*/dt = ( 1/2 ) kb*A* and dC*/dt = ( 1/2)kc*A* (B-2) Hence, dC*/dB* = kc*/kb*, which integrates to C */ B * = kc*/ kb* = x (B-3)
98 From eq B-3, it is clear that x remain s constant during oxalate breakdown and is therefore best determined by letting the reac tion go to completion. In this case, x = Rc/Rb, where Rc and Rb are the mass ratios in C and B, respectively. Case I: Heavy atom isotope effects by analysis of residual oxalate. The effect of heavy atom substitution in the carboxyl group that becomes formate is described by the following differential equation: dB */ dB = ( 1/2 ) kb*A* /( kA ) which, by use of the relationships give n above, can be rearranged to yield dB */ A* = ( kb*/ 2k )[ dB /( A0-B )] (B-4) Using the fact that A* = A0* B* C* = A0* B*(1 + x), we can write ) 1 ( * * *0x B A dB = ( kb*/2k )[ dB /( A0-B )] which integrates to give ) 1 ln( * ) 2 / 1 ( * ) 1 ( * * ln 1 10 0f k k A x B A xb (B-5) If we let Rs and R0 be the mass ratios in the residual a nd initial oxalate, respectively, then Rs = B A x B A 0 0) 1 ( * * (B-6) f A x B A A A B A x B A R Rs1 1 * ) 1 ( * * * ) ( ) 1 ( * *0 0 0 0 0 0 0 (B-7) Combining eqs B-5 and B-7, we obtain ) 1 ln( 2 * / 1 ln 1 10f k k R R f xb s (B-8) and so
99 apparent isotope effect = ) / )( 1 ( ln ) 1 ln( * ) 1 ( 20R R f f k x ks b Use of eq. B-3 to replace kb* in eq. B-8 then gives apparent isotope effect = * ) 1 ( 2ck x xk (B-9) Thus k/kb* = 2 ) 1 ( x (apparent isotope effect) (4-13) k/kc* = x x2 ) 1 ( (apparent isotope effect) (4-12) The value of x, needed to obtain the heavy atom isotope effects, is determined as described above. Case II: Heavy atom isotope effect s by analysis of formate and CO2. If we define A = [oxalate], B = [formate], and C = [CO2], we can write A0*-B0*(1+x) = A0*-B*C* R0 = A0*/A0 Rb = B*/B Rc = C*/C Therefore f(Rb/R0) = (B/A0) (B*/B) (A0/A*) = B*/A0* and f(Rc/R0) = C*/A0* Note that R0 in these expressions is for the whole molecule. Then 0 0 0) ( 1 * * * * R R R f A C B Ac b (B-10) and hence we obtain apparent isotope effect = 01 ln ) 1 ln( R R R f fc b (B-11)
100 APPENDIX C EPR SPIN-TRAPPING OF A RADICAL DURING OXALATE DECARBOXYLASE TURNOVER EPR spectra obtained when spin trap reagents DMPO and DEPMPO were added to OxdC reaction mixture during the turnover we re shown below. Implication of observed spectra and experimental conditions are discussed in Chapter 6. Figure C-2. EPR spectra showing that trapped radical is not tyrosyl radical. EPR spectra of fractions collected after running 3.45 mg/mL enzyme, 0.15 M potassium acetate buffer pH 4.2, 450 mM DMPO, and 75 mM oxalate pH 4.2 reaction mixture on G-25 column. Samples 7 & 8 contain enzyme while samples 11 to 15 do not. The spectrum of th e reaction mixture prior to filtration is also shown.
101 A B Figure C-1. EPR spectra of the spin-traped radical formed during OxdC turnower. (A) OxdC in 0.15 M acetic acid buffer pH 4.2, 450 mM DMPO, 75 mM oxalate pH 4.2 plus controls. (B) OxdC in 0.14 M acetic acid buffer, 428 mM DEPMPO, 71 mM oxalate pH 4.2 plus controls. Instrumental parameters: center field, 3484 G; sweep width, 60 G; microwave power, 20 mW; modulation amplitude, 0.5 G; time constant, 40.96 ms; and gain, 60 dB.
102 0 1 2 3 4 5 6 00.040.4220Activity / mM/minConcentration (DMPO) / mM Figure C-3. Rate of formate production versus concentra tion of DMPO present during OxdC catalysed oxalate deacrboxyla tion. Experiments were done in triplicates. Figure shows slight increase of OxdC activity in the presence of 0.04-0.4 mM DMPO.
103 LIST OF REFERENCES Allison, M.J., Dawson, K.A., Mayberry, W.R., and Foss, J.G., (1985) Arch. Microbiology 141(1): 1-7 Anand, R., Dorrestein, P.C., Kinsland, C ., Begley, T.P., and Ealick, S.E. (2002) Biochemistry 41: 7659-7669 Anantharaman, V., Aravind, L., and Koonin, E.V. (2003) Curr.Opin. Chem. Biol. 7:12-20 Andrieux, C.P., Gonzalez, F., and Saveant, J.M. (2001) J. Electroanalytical Chemistry 498: 171-180 Azam, M., Kesarwani, M., Natara jan, K., and Datta, A. (2001) Biochem. Biophys. Res. Commun 289: 807-812 Azam, M., Kesarwani, Chakraborty, S., Natarajan, K., and Datta, A. (2002) Biochem. J. 367: 67-75 Bayles, J.W., Bron, J., and Paul, S.O. (1976) J. Chem. Soc. Faraday Trans. 1: 1546-1552 Berti, P. J. (1999) Methods Enzymol. 308: 355-397 Bucher, G., Halupka, M., Kolano, C., Schade, O., and Sander, W. (2001) Eur. J. Org. Chem. 4: 545-552 Butler, J., and Henglein, A. (1980) Radiat. Phys. Chem. 15: 603-609 Chakraborty, S., Chakraborty, N., Jain, D., Salunke, M.D., and Datta, A. (2002) Protein Sci. 11: 2138-2147 Chang, C.H., SvedruÂži , D., Ozarowski, A., Walker, L., Yeagle, G., Britt, R.D. Angerhofer, A., and Richards, N.G.J. (2004) J. Biol. Chem. 276(47): 43627-43634 Chang, C.H., and Richards, N.G.J. (2005) J. Chem. Theory. Comput. in press Cleland, W.W. (1977) Adv. Enzymol. 45: 273-287 Cleland, W.W. (1979) Methods Enzymol. 63, 103-138 Cleland, W.W. (1982) CRC Critical Reviews in Biochemistry. 13, 385-400 Cook, P.F., and Cl eland W.W. (1981) Biochemistry 20: 1790-1797
104 Costa, T., Steil, L., Martins, L.O., Volker,U., and Henri ques, A.O. (2004) J. Bacteriology 186(5): 1462-1474 De Pauw, L., and Toussinet, C., (1996) Rev. Med. Brux. 17: 67-74 Dunwell, J.M.and Gane, P.J. (1998) J. Mol. Evol. 46: 147-154 Dunwell, J.M., Khuri, S., and Gane, P.J. (2000) Microbiol. Mol. Biol. Rev. 64(1): 153179 Dunwell, J.M., Culham, A., Carter, C.E ., Sosa-Aguirre, C.R., and Goodenough, P.W. (2001) Trends. Biochem. Sci. 26: 740-746 Dunwell, J.M., Purvis, A., and Khuri S. (2004) Phytochemistry 65: 7-17 Dutton, M.V., Kathiara, M., Gallagher, I.M., and Evans, C.S (1994) FEMS Microbiol. Lett. 116: 321-326 Dutton, M.V., and Evans, C.S. (1996) Can. J. Microbiol. 42: 881-895 Edwards, R.A., Baker, H.M., Whittaker, M.M., Whittaker, J.V., Jameson, J.B., and Baker, E.N. (1998) J. Biol. Inorg. Chem. 3: 161-171 Emiliani, E., and Bekes, P. (1964) Arch. Biochem. Biophys. 105: 488-493 Emiliani, E., and Riera, B. (1968) Biochim. Biophys. Acta 167: 414-421 Escutia, M.R., Bowater, L., Edwards, A., Bo ttrill, A.R., Burrell, M.R., Polanco, R., Vicuna, R., and Bornemann, S. (2005) App. Environ. Microbiol. 71(7): 3608-3616 Franceschi, V.R., and Horner, H.T.Hr. (1980) Bot. Rev. 46(4): 361-427 Galus, Z. (1985) In Standard Potentials in Aqueous Solution; Bard, A.J., Parsons, R., Jordan, J., Eds.; Marcel Dekker: New York GrÃ¤slund, A., Lubitz, W., and Lassmann, G. (1996) J. Am. Chem. Soc. 118: 8111-8120 Gerfen, G. J., van der Donk, W.A., Yu, G., Mc Carthy, J.R., Jarvi, E.T., Matthews, D.P., Farrar, C., Griffin, R.G., and Stubbe, J. (1998) J. Am. Chem. Soc. 120: 3823-3835 Gopal, B., Madan, L.L., Betz, S. F., and Kossiakoff, A.A. (2005) Biochemistry 44: 193201 Headley, G.W., and OÂ’Leary, M.H. (1990) J. Am. Chem. Soc. 112: 1894-1896 Hermes, J.D., Morrical, S.W., OÂ’Leary, M.H ., and Cleland, W.W. (1984) Biochemistry 23: 5479-5488 Hodgkinson, A. (1977) Oxalic Acid in Biology and Medicine; Academic Press: London
105 Isse, A.A., Gennaro, A., and Maran, F. (1999) Acta Chem. Scand. 53: 1013-1022 Ito, K., and Bernstein, H. J. (1956) Can. J. Chem. 34: 170-178 Jonsson, M.; Wayner, D.D., and Lusztyk, J. (1996) J. Phys. Chem. 100: 17539-17543 Just, V.J., Stevenson, C.E.M., Bowater, L., Tanner, A., Lawson, D.M., and Bornemann, S. (2004) J. Biol. Chem. 279(19): 19867-19874 Kanoufi, F., and Bard, A.J. (1999) J. Phis. Chem. B 103: 10469-10480 Kathiara, M., Wood, D.A., and Evans, C.S. (2000) Mycol. Res. 104: 345-350 Kehres, D.G., and Maguire, M.E. (2003) FEMS Microbiology Rev. 27: 263-290 Kesarwani, M., Azam, M., Natarajan, K., Mehta, A., and Datta, A. (2000) J. Biol. Chem. 275, 7230-7238 Kim, H.W., Perez, J.A., Ferguson, S.J., and Campbell, I. D. (1990) FEBS Letters 272: 3436 Klinman, J.P. (1978) Adv. Enzymol. Relat. Areas. Mol. Biol. 46: 415-94 Kotsira, V.P., and Clonis, Y.D. (1997) Arch. Biochem. Biophys. 340: 239-249 Kuwan, R.Y., Kasahara, M., Fujibayashi, H., Takamatsu, N., Ogasawara, K., and Watabe, K. (2002) Microbiology 148: 3971-3982 Lassmann, G., Kolberg, M., Bleifuss, G., Gr Ã¤slund, A., SJÃ¶berg, B.-M., and Lubitz, W. (2003) Phys. Chem. Chem. Phys. 5: 2442-2453 Lendzian, F., Sahlin, M., MacMillan, F., Bittl, R., Fiege, R., PÃ¶tsch, S., SjÃ¶berg, B.-M., Graslund, A., Lubitz, W., and Lassmann, G. (1996) J. Am. Chem. Soc. 118: 81118120 Libert, B., and Franceschi, V.R. (1987) J. Agric. Food Chem. 35: 926-938 Lillehoj, E.B., and Smith, F.G. (1965) Arch. Biochem. Biophys. 109: 216-220 Loewus, F.A., Saito, K., Suto, R.K., and Maring, E. (1995) Biochem. Biophys. Rcs. Commun. 212:196-203 Lowry, O.H., Rosbrough, N.J., Farr, A.L., and Randall, R.J. (1951) J. Biol. Chem. 193: 265 Magro, P., Marciano, P., and Di Lenna, P. (1988) FEMS Microbiol. Lett. 49: 49-52 MÃ¤kelÃ¤, M., Galkin, S., Hatakka, A. and Lundell, T. (2002) Enzyme and Microbial Technology 30: 542-549
106 Mendes, P. (1993) Comput. Applic. Biosci. 9: 563-571 Mendes, P. (1997) Trends Biochem. Sci. 22: 361-363 Mehta, A., and Datta, A. (1991) J. Biol. Chem. 266: 23548-23553 Micales, J. A. (1997) Biodeterioration & Biodegradation 39: 125-132 Michaelis, M., and Menten, M.L. (1913) Biochem. Z 49:333-337 Mizuno, K., Whittaker, M.M., Bachi nger, H.P., Whittaker, J.W. (2004) J. Biol. Chem. 279: 27339-27334 Northrop, D.B. (1975) Biochemistry 14: 2644-2654 Northrop, D.B. (1977) in Isotope Effects on Enzyme-Catalyzed Reactions; Cleland W.W., OÂ’Learly, M.H., Northrom D.B., Eds.; University Park Press: Baltimore, MD; pp 122-148 OÂ’Leary, M.H. (1980) Methods in Enzymol. 64: 83-104 OÂ’Leary, M. H. (1989) Annu. Rev. Biochem. 58: 377-401 Padmakumar, R., and Banerjee, R. (1995) J. Biol. Chem. 270: 9295-9300 Pelle, K., Wittmann, M., Lovrics, K., and Noszticzius, Z. (2004) J. Phis. Chem. 108: 7554-7562 Quayle, J.R. (1963) Biochem. J. 87: 368-373 Raquena, L., and Bornemann, S. (1999) J. Inorg. Bio. Chem. 74: 275-280 Reddy, S. G., Wong, K.K., Parast, C.V., Peisach , J., Magliozzo, R.S., and Kozarich, J.W. (1998) Biochemistry 37: 558-563 Reinhardt, L.A., SvedruÂži , D., Chang, C.H., Cleland, W. W., and Richards, N.G.J. (2003) J. Am. Chem. Soc. 125(5): 1244-1252 Rupp, M., Rao, K.K., Hall, D.O., and Cammack, R. (1978) Biochim. Biophys. Acta 537: 255-269 Sahin, N. (2003) Reseach in Microbiology 154: 399-407 Scelonge, C.J., and Bidney, D.L. 1 Octobe r 1998, PCT patent a pplication WO 98/42827 Schutte, H., Flossdorf, J., Sahm, H., Kula, M.R. (1976) Eur. J. Biochem. 62: 151-160
107 Schwarzenbacher, F. von Delft, L., Jaroszew ski, P., Abdubek, E., Ambing, T., Biorac, L.S., Brinen, J.M., Canaves, J., Cambell, H.J., Chiu, X., Dai, A.M., Deacon, M., Didonato, M.A., Elsliger, S., Eshagi, R., Floyd, A., Godzik, C., Grittini, S.K., Grzechnik, E., Hampton, C., Karlak, H.E., Klock, E., Koesma, J.S., Kovarik, A., Kreusch, P., Kuhn, S.A., Lesley, I., Levi n, D., McMullan, T.M., McPhillips, M.D., Miller, A,. Morse, K., M oy, J., Ouyang, R., Page, K., Quijano, A., Robb, G., Spraggon, R.C., Stevens, H., van der Bedem J., Velasquez, J., Vincent, X., Wang, B., West, G., Wolf, Q., Xu, K.O., Hodgson, J., Wooley, I.A., and Wilson (2004) Proteins: Struct. Funct. Genet. 56: 392-395. Shimada, M., Akamtsu, Y., Tokimatsu, T., Mii, K., and Hattory, T. (1997) J. Biotechnology. 53: 103-113 Shimazono, H. (1955) J. Biochem. (Tokyo) 423:21-340 Shimazono, H., and Hayaishi, O. (1957) J. Biol. Chem. 227: 151-159 Sims, L. B.; Lewis, D. E. (1984) In Isotopes in Organic Chemistry, Vol. 6: Isotopic Effects. Recent Developments in Theory and Experiment; Buncel. E.; Lee, C. C., Eds.; Elsevier: Amsterdam, pp.161-259. Speakman, J.C. (1940) J. Chem. Soc. 855-859 Stubbe, J., Nocera, D.G., Yee, C. S., and Chang, M.C.Y. (2003) Chem. Rev. 103: 21672201 Sun, X., Ollagnier, S., Schmidt, P.P., Atta , M., Mulliez, E., Lepape, L., Eliasson, R., GrÃ¤slund, A., Fontecave, M., Reichard, P., and SjÃ¶berg, B.-M. (1996) J. Biol. Chem. 271: 6827-6831 Tanaka, N., and Araki, M. (1985) J. Am. Chem. Soc. 107: 7780-7781 Tanner, A., and Bornemann, S. (2000) J. Bacteriol. 182: 5271-527 Tanner, A., Bowater, L., Fairhurst, S. A. and Bornemann, S. (2001) J. Biol. Chem. 276: 43627-43634 Tommos, C., Tang, X.-S., Warncke, K., Hoga nson, C.W., Styring, S., McCracken, J., Diner, B.A., and Babcock, G.T. (1995) J. Am. Chem. Soc. 117: 10325-10335 van der Donk, W.A., Yu, G., PÃ©rez, L., Sa nchez, R.J., Stubbe, J., Samano, V., and Robins, M.J. (1998) Biochemistry 37: 6419-6426 Vetting, M.W., Wackett, L.P., Que, L.Jr., Lipscomb, J.D., and Ohlendorf, D.H. (2004) J. Bacteriol 186(7): 1945-1958 Wagner, A. F.V., Frey, M., Neugebauer, F. A., SchÃ¤fer, W., and Knappe, J. (1992) Proc. Natl. Acad. Sci. USA 89: 996-1000
108 Waldrop, G. L., Braxton, B.F., Urbauer, J. L., and Cleland, W.W. (1994) Biochemistry 33: 5262-5267 Weil, J. A., Bolton, J. R., and Wertz, J. E. (1994) in Electron Paramagnetic Resonance: Elementary Theory and Practical Applications, pp. 321-323, Wiley-Interscience, New York Whittaker, M.M., and Whittaker, J.W. (1999) J. Biol. Chem. 274: 34751-35757 Whittaker, M.M., and Whittaker, J.W. (2000) J. Biol. Inorg. Chem. 5: 402-408 Whittaker, M.M., and Whittaker, J.W. (2002) J. Biol. Inorg. Chem. 7: 136-145 Williams, H.E., and Wandzilak, T.R. (1989) J. Urol. 141: 742-747 Woo, E.J., Dunwell, J.M., Goodenough, P.W ., Marvier, A.C., and Pickersgill R.W. (2000) Nat. Struct. Biol. 7:1036-1040.
109 BIOGRAPHICAL SKETCH DraÂženka SvedruÂži was born in Osijek, Croatia, in 1974. She holds a B.Sc. degree in chemistry from the College of Natural Sciences and Mathemathics at the University of Zagreb. DraÂženka did her undergraduate resear ch in biochemistry at the Institute of Medical Research and Occupational Health unde r the direction of Dr. Elsa Reiner. From 1997 till 2000 she was enrolled as a graduate student in the College of Chemical Engineering and Technology at the University of Zagreb. The work on organic synthesis of hetrocyclic compounds with potential anti viral and antitumor activities was done under the supervision of Pr of. Mladen Mintas. DraÂženka st arted graduate studies in the Chemistry Department of the University of Florida in August 2000, where she joined Dr. Nigel G. J. RichardsÂ´ research group