Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Mechanistic investigations of enzyme-catalyzed peptide and carbohydrate epimerization Murkin, Andrew Stewart 2004

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

Item Metadata

Download

Media
831-ubc_2004-931609.pdf [ 16.02MB ]
Metadata
JSON: 831-1.0061206.json
JSON-LD: 831-1.0061206-ld.json
RDF/XML (Pretty): 831-1.0061206-rdf.xml
RDF/JSON: 831-1.0061206-rdf.json
Turtle: 831-1.0061206-turtle.txt
N-Triples: 831-1.0061206-rdf-ntriples.txt
Original Record: 831-1.0061206-source.json
Full Text
831-1.0061206-fulltext.txt
Citation
831-1.0061206.ris

Full Text

Mechanistic Investigations of Enzyme-Catalyzed Peptide and Carbohydrate Epimerization by A N D R E W STEWART M U R K I N B.Sc., The University of British Columbia, 1998 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE F A C U L T Y OF G R A D U A T E STUDIES Department of Chemistry We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH C O L U M B I A April, 2004 © Andrew Stewart Murkin, 2004 Abstract 11 The inversion of stereochemistry in biomolecules is catalyzed by enzymes called racemases and epimerases. One such epimerase, isolated from the venom of the funnel-web spider Agelenopsis aperta, interconverts two peptide neurotoxins that differ only by the absolute stereochemistry at Ser46. Because enzymatic epimerization of peptides is otherwise unknown, the mechanism by which this occurs was investigated. Various substrate analogues were synthesized, in which the functional groups about the epimerizable serine were modified. One of these, containing chlorine in place of the hydroxyl, was found to undergo enzyme-catalyzed elimination of HC1 to generate a dehydroalanine derivative. This dehydroalanine peptide was independently synthesized and found to be a potent inhibitor of the epimerase, exhibiting a sub-micromolar IC50. These results support a deprotonation-reprotonation mechanism that proceeds through an enolate intermediate. The planarity of this intermediate is apparently mimicked by the sp2 character of the a-carbon of dehydroalanine, resulting in the observed inhibition. The biosynthesis of sialic acids is initiated by the conversion of UDP-7V"-acetylglucosamine (UDP-GlcNAc) to 7V-acetylmannosamine (ManNAc), a process involving epimerization at C-2. Whereas a hydrolyzing UDP-GlcNAc 2-epimerase is known to catalyze this reaction in mammals, a similar enzyme in bacteria had not been previously identified. The gene product from neuC, part of the Escherichia coli K l gene cluster responsible for sialic acid synthesis, was shown by H - and P-NMR spectroscopy and mass spectrometry to convert UDP-GlcNAc to UDP and 2-acetamidoglucal. The homologous gene from Neisseria meningitidis B, siaA, was cloned and expressed as a histidine-tagged fusion protein. This protein, previously identified erroneously in literature in its untagged form as an A^-acetylglucosamine-6-phosphate 2-epimerase, was shown to be a hydrolyzing UDP-GlcNAc 2-epimerase, converting UDP-GlcNAc to a-ManNAc and UDP. Similar to the non-hydrolyzing UDP-GlcNAc 2-epimerase, I l l SiaA was found to be allosterically regulated by its substrate, exhibiting sigmoidal kinetics with £Cat = 4.7 s"1, K' = 1.5 mM, and Hi l l coefficient, n a p p = 1-9. Additionally, in the presence of UDP, 2-acetamidoglucal was found to be a substrate, thereby implicating it as an intermediate. Incubations with (2"- 2H)UDP-GlcNAc showed no significant kinetic isotope effect on kcat and kcJKm, indicating that deprotonation at C-2" is not rate limiting. A solvent isotope discrimination experiment, wherein either UDP-GlcNAc or 2-acetamidoglucal and UDP were incubated with the enzyme in 50% deuterated water, failed to show a preference for proton transfer over deuteron transfer during glycal hydration, indicating that this step is also not rate limiting. Finally, the 1 8 0 label from the incubation of [ l" - 1 8 0]UDP-GlcNAc with SiaA was shown to depart with UDP, confirming that the reaction proceeds with C - 0 bond cleavage. iv Table of Contents Abstract ii Table of Contents iv List of Figures viii List of Tables xiii Abbreviations and Symbols xiv Acknowledgements xxii Dedication xxiii Chapter 1: Enzyme-Catalyzed Racemization and Epimerization of Amino Acids and Peptides 1 1.1 Chirality in Nature 2 1.2 Peptides Containing D-Amino Acids 3 1.2.1 Non-ribosomally Derived Peptides 3 1.2.2 Ribosomally Derived Peptides 6 1.3 Amino Acid Racemases and Epimerases 10 1.3.1 Cofactor-Dependent Amino Acid Racemases 11 1.3.2 Cofactor-Independent Amino Acid Racemases and Epimerases 15 1.4 Posttranslational Epimerization of Peptides 20 1.5 Acidity of the a-Proton 24 1.5.1 Intermediate Stabilization 25 1.5.2 Abstraction of the a-Proton 26 1.5.3 Acidity of the a-Proton in a Peptide 27 1.6 Project Goals 27 Chapter 2: Mechanistic Studies on Spider Venom Peptide Epimerase 29 2.1 Protein Homology to Serine Proteases 31 2.1.1 Protein Structure and Sequence Determination 31 2.1.2 Modified Deprotonation/Reprotonation Mechanism 33 2.2 Enzyme Purification and Activity 35 2.2.1 Enzyme Source and Activity 35 V 2.2.2 Metalloprotease Activity 37 2.3 Depsipeptide Analogue ; 41 2.3.1 Synthesis of Depsipeptide Analogue 5 42 2.3.2 Evaluation of Depsipeptide Analogue 5 45 2.4 Chloroalanine Analogue 47 2.4.1 Synthesis of Chloroalanine Analogue 18 48 2.4.2 Enzymatic Reaction and Inhibition by Chloroalanine Analogue 18 49 2.5 Dehydroalanine Analogue 53 2.5.1 Inhibitors of Cofactor-Independent Amino Acid Racemases and Epimerases 53 2.5.2 Synthesis of Dehydroalanine Analogue 24 54 2.5.3 Inhibition of Epimerase by Dehydroalanine Analogue 24 56 2.6 Phosphinate Analogue 58 2.6.1 Phosphinate Analogues as Intermediate and Transition State Mimics 58 2.6.2 Synthesis of Phosphinate Analogue 37 60 2.6.3 Evaluation of Phosphinate Analogue 37 62 2.7 Steps toward Further Understanding of the Peptide Epimerase's Structure and Mechanism 63 2.7.1 Active Site Labelling 64 2.7.2 Ketone Analogue 67 2.7.3 Affinity Chromatography 70 2.8 Conclusions 71 2.9 Experimental Procedures 73 2.9.1 General Synthetic Methods 73 2.9.2 Synthesis of Depsipeptide Analogue 5 76 2.9.3 Synthesis of Chloroalanine Analogue 18 80 2.9.4 Synthesis of Dehydroalanine Analogue 24 84 2.9.5 Synthesis of Phosphinate Analogue 37 85 2.9.6 Synthesis of Enzyme-Labelling Compounds 44, 45, and 46 88 2.9.7 General Enzyme Methods 89 2.9.8 Enzyme Purification 90 2.9.9 Epimerase and Metalloprotease Assays 90 2.9.10 Determination of Extinction Coefficients for 1,18, and 24 91 2.9.11 Enzymatic Studies on Substrate Analogues 91 2.9.12 Attempted Covalent Modification of the Enzyme 93 vi Chapter 3: Enzyme-Catalyzed Epimerization of Carbohydrates 94 3.1 Inversion of Stereochemistry in Amino Acids Versus Carbohydrates 95 3.2 Epimerization at "Activated" Centres .'. 96 3.3 Epimerization at "Unactivated" Centres 99 3.3.1 NAD+-Dependent Epimerases 99 3.3.2 Cofactor-Independent Epimerases 103 3.4 Project Goals 115 Chapter 4: Mechanistic Studies on a Bacterial Hydrolyzing UDP-A -^Acetylglucosamine 2-Epimerase 116 4.1 Enzymatic Activity of NeuC 118 4.1.1 Affinity Purification of NeuC 118 4.1.2 N M R Assays with NeuC 121 4.2 Cloning and Expression of siaA, a neuC Homologue from Neisseria meningitidis 125 4.2.1 Previous Work Done 125 4.2.2 Ligation Independent Cloning of siaA and Affinity Purification of Histidine-tagged SiaA 127 4.2.3 Identification of the Substrate 132 4.3 Characterization of SiaA Activity 133 4.3.1 Characterization of the Product: Stereochemistry and Isotope Incorporation 133 4.3.2 Catalytic Competence of 2-Acetamidoglucal and Possible Alternative Substrates 135 4.3.3 Kinetic Characterization of SiaA by a Continuous Coupled Assay 137 4.4 Kinetic Isotope Effect Studies 144 4.4.1 Kinetic Isotope Effect on & c a t 145 4.4.2 Kinetic Isotope Effect on kcJKm by Intermolecular Competition 147 4.4.3 Solvent Isotope Discrimination During Glycal Hydration 150 4.5 Positional Isotope Exchange (PIX) Studies 155 4.5.1 Introduction 155 4.5.2 PIX Experiment 160 4.6 Related Studies and Future Directions 163 4.7 Conclusions 166 4.8 Experimental Procedures 169 4.8.1 Materials 169 4.8.2 General Methods 171 vii 4.8.3 Overexpression of Intein-tagged NeuC and Affinity Purification of NeuC 174 4.8.4 N M R Assays with NeuC 174 4.8.5 Cloning of siaA Gene _• 175 4.8.6 Overexpression and Affinity Purification of 6xHis-SiaA 176 4.8.7 N M R Assays with 6xHis-SiaA 177 4.8.8 Kinetic Characterization of 6xHis-SiaA 178 4.8.9 Synthesis of Isotopically Substituted and Labelled UDP-GlcNAc 179 4.8.10 Determination of Kinetic Isotope Effects (KIEs) 181 4.8.11 Positional Isotope Exchange (PIX) Experiment and Test for C - 0 versus P - 0 Bond Cleavage 183 4.8.12 Protein Sequence Alignment 183 Appendix: 'H-NMR Spectra of Final, Deprotected Peptides 184 References 191 List of Figures V l l l Figure 1.1 Structure of peptidoglycan = 4 Figure 1.2 Antibiotics derived from D-amino acids 5 Figure 1.3 Scheme used to show that dermorphin and deltorphin are ribosomally derived peptides containing D-amino acids 7 Figure 1.4 Isomeric peptide toxins isolated from the venom of Agelenopsis aperta 9 Figure 1.5 Generalized reactions catalyzed by racemases and epimerase 10 Figure 1.6 One- and two-base mechanisms of amino acid racemases 11 Figure 1.7 Mechanism of PLP-dependent racemases 12 Figure 1.8 Models of alanine racemization 13 Figure 1.9 Proline metabolism in Clostridium sticklandii 15 Figure 1.10 Models of proline racemization 16 Figure 1.11 Interconversion of enzyme protonation states 17 Figure 1.12 Tracer perturbation experiment 18 Figure 1.13 Elimination of HC1 from ^reo-(3-chloroglutamate catalyzed by glutamate racemase mutants 19 Figure 1.14 Formation of lanthionine, a dimeric amino acid with D-configuration 20 Figure 1.15 Proposed mechanisms for serine epimerization by spider venom peptide epimerase 22 Figure 1.16 Conversion of an amino acid zwitterion to enol zwitterion 24 Figure 1.17 Enol intermediates of proline and glutamate and corresponding analogues 25 Figure 2.1 Primary structure of the peptide epimerase 31 Figure 2.2 Sequence homology between the peptide epimerase and serine proteases 32 Figure 2.3 Acyl-enzyme mechanism of peptide epimerization 34 Figure 2.4 Funnel-web spider, Agelenopsis aperta 35 Figure 2.5 Typical epimerization assay 36 Figure 2.6 Purification of crude spider venom by cation-exchange chromatography 38 Figure 2.7 Purification of ion-exchange purified epimerase by size-exclusion chromatography 39 Figure 2.8 Specificity of the metalloprotease activity 40 Figure 2.9 Strategy to trap the purported acyl-enzyme adduct using a depsipeptide 42 Figure 2.10 Initially designed scheme towards the synthesis of depsipeptide 43 Figure 2.11 Corrected synthesis of depsipeptide 5 44 Figure 2.12 Racemization of activated amino acid residues during convergent peptide coupling 45 Figure 2.13 Enzyme-catalyzed hydrolysis of depsipeptide 5 46 Figure 2.14 Strategy to trap the purported acyl-enzyme adduct using a 3-chloroalanine derivative 47 Figure 2.15 Synthesis of 3-chloroalanine analogue 18 48 Figure 2.16 Non-enzymatic hydrolysis of chloroalanine analogue 18 and enzymatic elimination of HC1 , 50 Figure 2.17 Enzyme-catalyzed elimination of HC1 from chloroalanine 50 Figure 2.18 Plot of percent inhibition of epimerization versus concentration of chloroalanine analogue 18 52 Figure 2.19 Substrate and intermediate analogues of a) proline racemase, b) Dap epimerase, and c) glutamate racemase that inhibit via planarity at C-2 54 Figure 2.20 Synthesis of dehydroalanine analogue 24 55 Figure 2.21 Plot of percent inhibition of epimerization versus concentration of dehydroalanine analogue 24 56 Figure 2.22 Phosphinate formation from 1-aminophosphinic acid and acrylate 60 Figure 2.23 Synthesis of phosphinate 37 61 Figure 2.24 Covalent inhibition by benzoxazinone 44 65 Figure 2.25 Electrospray mass spectrum of the epimerase following incubation with 44 66 Figure 2.26 Strategy to label the active site using iV-haloacetyl peptides 66 Figure 2.28 Possible preparation of ketone analogue 47 69 Figure 2.29 Preparation of affinity chromatography resin 71 Figure 3.1 Reaction catalyzed by D-ribulose-5-phosphate 3-epimerase 96 Figure 3.2 Proposed mechanism of D-ribulose-5-phosphate 3-epimerase 97 Figure 3.3 Proposed mechanism for epimerization at C-2 of glucopyranoses 98 Figure 3.4 Epimerization at an 'unactivated' centre by transient oxidation 99 Figure 3.5 The mechanism of the reaction catalyzed by UDP-galactose 4-epimerase 100 Figure 3.6 Evidence for 4-ketohexose as an intermediate in the UDP-Gal 4-epimerase reaction 101 Figure 3.7 Epimerization at C-3 and C-5 102 Figure 3.8 Mechanism of the reaction catalyzed by L-ribulose-5-phosphate 4-epimerase 104 Figure 3.9 Reactions catalyzed by hydrolyzing and non-hydrolyzing UDP-GlcNAc 2-epimerases 104 Figure 3.10 Structure of the UDP-GlcNAc 2-epimerase dimer 105 Figure 3.11 Proposed mechanisms of the reaction catalyzed by the non-hydrolyzing UDP-GlcNAc 2-epimerase 106 Figure 3.12 Positional isotope exchange catalyzed by the non-hydrolyzing UDP-GlcNAc 2-epimerase 107 Figure 3.13 Mammalian biosynthetic pathway for the sialic acid 7V-acetylneuraminic acid 109 Figure 3.14 Possible mechanisms for the reaction catalyzed by the mammalian hydrolyzing UDP-GlcNAc 2-epimerase 110 Figure 3.15 The sialic acid biosynthetic pathway in Escherichia coli K l 112 Figure 3.16 Autoradiogram showing activity of NeuC 114 Figure 4.1 Intein-mediated affinity purification of NeuC 118 Figure 4.2 Mechanism of protein splicing involving the intein from the Saccharomyces cerevisiae V M A 1 gene 119 Figure 4.3 Proposed mechanism of DTT-induced cleavage of NeuC from the intein tag 120 Figure 4.4 SDS-PAGE showing purification of NeuC .-. 121 Figure 4.5 N M R assay of NeuC activity 122 Figure 4.6 3 1 P N M R assay of NeuC activity 123 Figure 4.7 Discontinuous coupled assay for ManNAc formation 126 Figure 4.8 Possible mechanisms for epimerization of GlcNAc6P 127 xi Figure 4.9 Ligation-independent cloning 128 Figure 4.10 pET-30 Xa/LIC cloning/expression region 129 Figure 4.11 Agarose gel from colony PCR of NovaBlue E. coli cells transformed with pAM04 130 Figure 4.12 SDS-PAGE gels of affinity purification of 6xHis-SiaA 131 Figure 4.13 ' H N M R assay of SiaA activity 132 Figure 4.14 Enzymatic formation of a-(2-2F£)ManNAc in D 2 0 by ] H N M R spectroscopy 134 Figure 4.15 Enzymatic conversion of 2-acetamidoglucal to (2- 2H)ManNAc in the presence of UDP 136 Figure 4.16 Coupled assay used to measure the rate of formation of UDP 138 Figure 4.17 Enzyme kinetics for 6xHis-SiaA 138 Figure 4.18 Typical rate curves for two different enzymes with the same F m a x 139 Figure 4.19 Monod-Wyman-Changeux (MWC) model for the binding of substrate molecules to a tetrameric protein 141 Figure 4.20 Koshland-Nemethy-Filmer (KNF) model for the binding of substrate molecules to a tetrameric protein 142 Figure 4.21 pH and buffer dependence of activity 144 Figure 4.22 Reaction coordinate diagram illustrating the origin of primary kinetic isotope effects 145 Figure 4.23 Enzymatic synthesis of (2"- 2H)UDP-GlcNAc 146 Figure 4.24 Relative change in the ratio of deuterated substrate to protiated substrate as a function of the extent of reaction 148 Figure 4.25 Solvent isotope discrimination experiment 153 Figure 4.26 Positional isotope exchange in carboxylate and phosphate esters 156 Figure 4.27 Mechanism of glutamine synthetase as determined by PIX •; 156 Figure 4.28 Measurement of the extent of PIX by enzymatic and chemical derivatization 157 Figure 4.29 3 1 P N M R spectrum of inorganic phosphate labelled with 44% 1 8 0 158 Figure 4.30 3 1 P N M R spectra showing the P p signals of 1 80-labelled UDP-GlcNAc and UDP-ManNAc 159 Figure 4.31 Experiment to test for C - 0 bond cleavage and PIX 160 xii Figure 4.32 Synthesis of [ l"- 1 8 0]UDP-GlcNAc 161 Figure 4.32 3 1 P - N M R spectra showing the conversion of [1 "- 1 8 0]UDP-GlcNAc to [(3-1 8 0]UDP by SiaA 162 Figure 4.34 Elimination of UDP via an E l mechanism 163 Figure 4.35 Sequence alignment of UDP-GlcNAc 2-epimerases 165 List of Tables Table 4.1 Summary of D(V/K) calculations 150 Table 4.2 Primers used in ligation-independent cloning, colony PCR, and sequencing of siaA 176 xiv Abbreviations and Symbols 6xHis-SiaA SiaA fused with an N-terminal peptide tag containing a six histidine repeat 5 chemical shift (ppm) £220 extinction coefficient at 220 nm £262 extinction coefficient at 262 nm </> fractionation factor (ground-state) $ transition-state fractionation factor A220 absorbance of light at 220 nm ^262 absorbance of light at 262 nm ^280 absorbance of light at 280 nm Ac acetyl AcOH acetic acid Ad 1-adamantyl Ado adenosine ADP adenosine diphosphate alle allo-isoleucine ATP adenosine triphosphate Bn benzyl Boc tert-butoxycarbonyl bp base pair BSA bovine serum albumin (protein work); 7V,0-bis(trimethylsilyl)acetamide (synthesis) CBD chitin-binding domain CI Ala 3-chloroalanine CMP-Neu5Ac cytidine 5'-monophospho-Af-acetylneuraminic acid COSY correlation spectroscopy X V C V d D Da DCC DCI-MS dd DE52 D E A D Dha DIEA D20, h^yd D M A P DMF DMSO D N A dNTP DTT D V U(V/K) E. coli E. coli K l EDC EDTA ESI-MS Et 3 N column volume(s) doublet (NMR) deuterium (2H) Dalton AfTV'-dicyclohexylcarbodiimide desorption chemical ionization mass spectrometry doublet of doublets (NMR) diethylaminoethyl-cellulose ion-exchange resin diethyl azodicarboxylate dehydroalanine diisopropylethylamine solvent kinetic isotope effect on the rate of glycal hydration Af-dimethyl-4-aminopyridine AyV-dimethylformamide dimethylsulfoxide deoxyribonucleic acid deoxyribonucleotide triphosphate 1,4-dithio-DL-threitol deuterium kinetic isotope effect on £ c a t deuterium kinetic isotope effect on kcJKm Escherichia coli K l strain of Escherichia coli l-ethyl-3-(3'-dimethylaminopropyl)carbodiimide hydrochloride ethylenediaminetetraacetate, disodium salt electrospray ionization mass spectrometry triethylamine xvi E t 2 0 diethyl ether EtOAc ethyl acetate EtOH ethanol FH fractional conversion of a protiated compound to products Fmoc fluoren-9-yloxycarbonyl g ' 2 gravitational acceleration constant (9.8 m s" ) Gal galactose Glc glucose GlcNAc Af-acetylglucosamine GlcNAc6P N-acetylglucosamine 6-phosphate Glu-P y-glutamyl phosphate H B T U 0-(benzotriazol-1 -yl)-N, N, N', N '-tetramethyluronium hexafluorophosphate Hex hexanes (mixture of hexane isomers) HOBt 1-hydroxybenzotriazole hydrate HPLC high performance liquid chromatography i c 5 0 inhibitory concentration at which reaction rate is half maximal IGFBPs insulin-like growth factor-binding proteins iPrOH isopropyl alcohol IPTG isopropyl 1 -thio-P-D-galactopyranoside IVB co-Agatoxin IVB IVC co-Agatoxin IVC J coupling constant (NMR); subscripts indicate coupling partners kb kilobases(lOOObp) &cat catalytic rate constant •^cat/-^ m specificity constant; second-order rate constant kDa kilodalton X V I I KIE KL20 Km L B L B H B LC-MS LPS LSI-MS m M A L D I Man ManNAc ManNAc6P meDap MeOH mRNA MurNAc M W C O "app NA, Nc, NG, Nj N A D + N A D H N A D P + NBS enzyme-inhibitor complex dissociation constant kinetic isotope effect equilibrium constant for a H2O-HDO-D2O mixture Michaelis constant Luria-Bertani medium low-barrier hydrogen bond liquid chromatography mass spectrometry lipopolysaccharide liquid secondary ionization mass spectrometry multiplet (NMR) matrix-assisted laser desorption ionization mannose Af-acetylmannosamine iV-acetylmannosamine 6-phosphate Tneso-diaminopimelic acid methanol messenger ribonucleic acid TV-acetylmuramic acid molecular weight cut-off number of substrate binding sites (allostery); atom fraction of deuterium in a H2O-HDO-D2O mixture Hi l l coefficient number of given nucleotide (A, C, G, or T) present in a sequence nicotinamide adenine dinucleotide, oxidized form nicotinamide adenine dinucleotide, reduced form nicotinamide adenine dinucleotide phosphate, oxidized form Af-bromosuccinimide XV111 neuC NeuC N. meningitidis N M R O D 6 0 0 OH PBS PCR PDC PEP pet. ether PG PhLac Ph 3P pi Pi PLP PMSF PPi ppm PSA psi PyBroP R gene in E. coli K l that encodes NeuC, a hydrolyzing UDP-GlcNAc 2-epimerase protein from expression of neuC; hydrolyzing UDP-GlcNAc 2-epimerase from E. coli K l Neisseria meningitidis nuclear magnetic resonance optical dispersion at 600 nm carboxylic acid terminus (in peptide sequences) phosphate buffered saline polymerase chain reaction pyridinium dichromate phosphoenol pyruvate petroleum ether protecting group L-3-phenyllactic acid triphenylphosphine isoelectric point; pH at which a molecule has no net charge inorganic phosphate pyridoxal phosphate phenylmethylsulfonyl fluoride pyrophosphate parts per million polysialic acid pounds per square inch bromotripyrrolydinophosphonium hexafluorophosphate final ratio of isotopically labelled to unlabelled compound initial ratio of isotopically labelled to unlabelled compound X I X RffE R N A rpm rt Ru5P SDS-PAGE siaA SiaA SPPS Taq TBDPS tBu TEA TFA THF tR Tris tRNA UDP UDPDH UDP-GlcNAc UDP-ManNAc UDP-ManNAcUA Urd UTP non-hydrolyzing UDP-GlcNAc 2-epimerase from E. coli ribonucleic acid revolutions per minute room temperature ribulose 5-phosphate sodium dodecylsulfate polyacrylamide gel electrophoresis gene in N. meningitidis that encodes SiaA, a hydrolyzing UDP-GlcNAc 2-epimerase protein from expression of siaA; hydrolyzing UDP-GlcNAc 2-epimerase from N. meningitidis solid-phase peptide synthesis Thermophilics aquaticus ter/-butyldiphenylsilyl tert-buty\ triethanolamine trifluoroacetic acid tetrahydrofuran retention time 2-amino-2-(hydroxymethyl)-1,3-propanediol transfer ribonucleic acid uridine diphosphate UDP-ManNAc dehydrogenase UDP-A^-acetylglucosamine UDP-N-acetylmannosamine UDP-A^-acetylmannosaminuronic acid uridine uridine triphosphate U V ultraviolet v initial reaction velocity (rate) F m a x maximal reaction velocity (rate) Xu5P xylulose 5-phosphate Z benzyloxycarbonyl Common Amino Acid Abbreviations A Ala alanine C Cys cysteine D Asp aspartate E Glu glutamate F Phe phenylalanine G Gly glycine H His histidine I He isoleucine K Lys lysine L Leu leucine M Met methionine N Asn asparagine P Pro proline Q Gin glutamine R Arg arginine S Ser serine T Thr threonine V Val valine w Trp tryptophan Y Tyr tyrosine Nucleotide Base Abbreviations A adenine cytosine guanine thymine uracil X X l l Acknowledgements I owe the majority of my gratitude to my supervisor, Dr. Martin Tanner, whose combination of knowledge, leadership, enthusiasm, and wit is one-of-a-kind. My decision to undertake Ph.D. studies and to pursue a career in this field is due in no small part to the experiences under his guidance. It has been an absolute pleasure to work in the Tanner lab. I have had the unique opportunity to know every member of the group, past and present, and I must state that I am truly fortunate to have had such talented, helpful, and frequently comedic companions. I am specifically grateful to Dr. Jason Galpin for advice and assistance with synthesis and HPLC and to Dr. Robert Campbell for being a model researcher of whom I am still in awe. The bulk of my molecular biology knowledge is attributable to the assistance of Louis Luk and Wayne Chou. I am thankful to the Sherman group for granting me access to their peptide synthesizer and chemicals, to Dr. Robert Volkmann and Charles Kristensen for donations of spider venom, and to Dr. Warren Wakarchuk for providing bacterial genomic DNA. Additionally, the provision of synthetic carbohydrates by Wayne Chou and purified non-hydrolyzing UDP-GlcNAc 2-epimerase by Dr. Jomy Samuel was invaluable. Finally, I am thankful for the endless support of my family and friends. Though most of you may not understand more than a few sentences in this thesis, your contribution is greater than the words within. Dedicated to my family 1 Chapter 1 Enzyme-Catalyzed Racemization and Epimerization of Amino Acids and Peptides 2 1.1 Chirality in Nature In 1848, Louis Pasteur made a fundamental discovery that would profoundly change the scientific community's view of molecules; in fact, it has recently been regarded as "the most beautiful experiment in the history of chemistry."1 While studying the crystal forms of synthetic tartrates, Pasteur noticed that an optically inactive mixture of crystals could be separated by hand into two sets, distinguished by a non-superimposable mirror-image shape. Each set was found to be optically active, but only one shared identical properties with naturally occurring (+)-tartrate isolated from wine. He then hypothesized that this molecular asymmetry, later termed chirality, is one of the essences of life. To paraphrase his words, living organisms only produce molecules that are of one specific orientation, and these molecules are always optically active. Although there are exceptions to this rule, as wil l be described in the following sections, Pasteur's statement generally holds. Best illustrating this is the fact that nature employs almost exclusively L-amino acids and D-sugars. Chemical evolutionists largely agree that at some point in the early development of the universe, a physical trigger created an excess of these enantiomers over their D and L counterparts. The resulting selection allowed nucleic acids, composed of D-ribose and D-deoxyribose, to self-replicate and to direct the synthesis of proteins, composed only of L-amino acids. This homochirality is a hallmark of life, a requisite for an efficient living system; in fact, the incorporation of a single D-amino acid in a protein or L-deoxyribonucleotide in D N A can cause sufficient structural disruption to render the molecule inactive. As organisms further evolved, they developed a need to outcompete for survival, a problem that was overcome by accommodating new resources or by establishing unique defensive strategies. One way this may have been achieved is by modifying and expanding their 3 stereochemical repertoire, allowing them to consume nutrients that their competitors could not and providing protection against various harmful substances, such as antibiotics, degradative enzymes, and immunochemicals. Hence, nature as we know it today displays broad stereochemical diversity, a result, in part, of inverting the stereochemistry in many classes of biomolecules including amino acids, peptides, and carbohydrates. The following sections describe some of these modifications and the mechanisms by which they are produced. 1.2 Peptides Containing D-Amino Acids Perhaps the best studied molecules containing inverted stereochemistry are peptides and peptide derivatives containing D-amino acids. Such peptides have been detected in many classes of living organisms, ranging from the simple, unicellular (e.g., bacteria) to the complex, multicellular (e.g., mammals). These peptides can be divided into two categories based on their origin: ribosomally derived and non-ribosomally derived. 1.2.1 Non-ribosomally Derived Peptides Although most peptides in living organisms are assembled from a pool of twenty amino acids by ribosomes, some microorganisms are also capable of producing peptides using large, multifunctional enzymes known as peptide synthetases. These peptides often display diverse structures and exhibit a broad spectrum of biological activities. Specifically, compounds such as lipopeptides, peptidolactones, and depsipeptides (peptides containing ester linkages) are assembled from a variety of precursors including hydroxy acids, polyamines, amino acids with 4 meso-Dap Figure 1.1 Structure of peptidoglycan. The red, dashed line marks the union between the glycan (top) and peptide (bottom) portions. unusual side chains, iV-methylated amino acids, and - most relevant to this discussion - D-amino acids.2 The most studied non-ribosomal peptide bearing D-amino acids is peptidoglycan (Figure 1.1), a complex macromolecule composed of polymeric strands of alternating N-acetylglucosamine (GlcNAc) and Af-acetylmuramic acid (MurNAc) that are highly cross-linked by peptide chains. Peptidoglycan is an essential component of bacterial cell walls, providing rigidity by means of its intertwined framework and virulence by means of its uncommon amino acids. In particular, D-alanine, D-glutamate, and meso-diaminopimelic acid (meso-Dap) serve to protect the bacterium from lysis by foreign degradative enzymes. The formation of these amino acids (see Section 1.3) is of particular interest in pharmaceutical research, as their presence is unique to bacteria. D-Amino acids not only play a role in defensive mechanisms, they are also responsible for empowering bacteria with a form of biochemical weaponry. The bacterium Bacillus brevis produces the antibiotic gramicidin S, a cyclic peptide bearing two D-phenylalanine residues (Figure 1.2a). An X-ray crystal structure study suggests that the configuration of these residues 5 actinomycin D Figure 1.2 Antibiotics derived from D-amino acids. D-phenylalanine and D-valine are shown in red. contributes to the formation of an ordered assembly that may be adopted in the lipid bilayer, resulting in disruption of the cell membrane and increased susceptibility to lysis.3 Other antibiotics are derived from precursors containing D-valine. Penicillins and cephalosporins are formed from the common tripeptide 8-(L-a-aminoadipoyl)-L-cysteinyl-D-valine (Figure 1.2b) and serve to inhibit transpeptidase enzymes involved in peptidoglycan synthesis. Actinomycin D, which functions to inhibit transcription of DNA, consists of a phenoxazone ring that anchors two cyclic depsipeptides containing D-valine (Figure 1.2c). 6 1.2.2 Ribosomally Derived Peptides Whereas bacterial peptide synthetases are capable of assembling elaborate peptides with seemingly few restrictions, ribosomes are limited to a selection of only twenty amino acids, all of which (except the achiral amino acid glycine) are of the L-configuration. The fate of ribosomally derived peptides, however, seldom concludes at this stage; a number of posttranslational modifications are often made, such as acylation, alkylation, glycosylation, phosphorylation, truncation, addition of prosthetic groups, and formation of disulfide bridges. These transformations often dramatically alter the charge and mass of the initially synthesized peptide, and as a result, they can often be detected by standard proteomic methods, like electrophoresis and mass spectrometry. Occasionally, however, subtle, seemingly invisible changes can occur for which these methods cannot be used. This is the case for peptides that contain D-amino acids. The first example of a eukaryotic peptide found to incorporate a D-amino acid was dermorphin, a heptapeptide isolated from the skin of the Argentinean tree frog Phyllomedusa sauvagei4 Dermorphin exhibits opioid analgesic activity, particularly on the central nervous system, and is one thousand times more potent than morphine.5 Initial attempts to synthesize the peptide, which has the sequence Tyr-Ala-Phe-Gly-Tyr-Pro-Ser-NH2, yielded material that was inactive. Hence, a more careful examination of the natural peptide revealed that the second amino acid was in fact D-alanine. Consequently, the question as to its origin was raised: was it assembled by non-ribosomal machinery, as in the bacterial peptides discussed in Section 1.2.1, or was it the result of an otherwise unknown posttranslational modification? To address this, a technique known as complementary D N A (cDNA) cloning was employed.6 In brief, this involves the construction of a D N A library using P. sauvagei frog skin messenger R N A (mRNA) as a template and the enzyme reverse transcriptase. Because mRNA is only produced from genes that are being actively transcribed, the cDNA represents a much smaller subset of the genomic DNA, excluding all but the coding regions of actively expressed genes. This library was then screened for the presence of codons for the Phe-Gly-Tyr-Pro-Ser portion of dermorphin. The resulting positive clones provided a sequence that encodes a much larger precursor peptide containing four copies of dermorphin bearing L-Ala at the D-Ala position (Figure 1.3). This discovery confirmed that the active peptide results from a posttranslational modification of a ribosomally derived precursor. Further examination of the precursor revealed another heptapeptide with the predicted sequence Tyr-Met-Phe-His-Leu-Met-Asp-NH2.6 The corresponding peptide, termed deltorphin, was isolated from skin extracts, and similar to dermorphin, the second amino acid (Met) was found to have the D-configuration. Since these findings, many dermorphin and deltorphin analogues have been identified, varying in their specific sequence, but generally about seven residues in length and always with a D-amino acid (Ala, Met, Leu, or alle) at the second frog skin mRNA cDNA library screened cDNA clones transcription and translation Met 1 J Met 197 T Tyi I -Lys-Arg-Tyr-Ala-Phe-Gly-Tyr-Pro-Ser-Gly-Glu-Ala-Lys-Lys-Posttranslational processing Tyr-DAIa-Phe-Gly-Tyr-Pro-Ser-NH2 Dermorphin -Lys-Arg-Tyr-Met-Phe-His-Leu-Met-Asp-Gly-Glu-Ala-Lys-Lys-Posttranslational processing IIC-Tyr-DMet-Phe-His-Leu-Met-Asp-NH2 Deltorphin Figure 1.3 Scheme used to show that dermorphin and deltorphin are ribosomally derived peptides containing D-amino acids. The initial peptide possesses four copies of the direct precursor to dermorphin (black) and a single copy of the direct precursor to deltorphin (grey). 8 position.4 This work influenced the discovery of ribosomally derived, D-amino acid-containing peptides in other animals. The African giant snail Achatina fulica was found to produce a neuroexcitatory tetrapeptide called achatin-1 (Gly-D-Phe-Ala-Asp).7 The all-L isomer, achatin-2, was found to lack any excitatory activity, indicating the importance of the stereochemistry at the second position. A second peptide, called fulicin (Phe-D-Asn-Glu-Phe-Val-Ntb), was later isolated from the snail.8 As with achatin-1, the D-configuration was found to be essential for activity. A cDNA study similar to that described above was also undertaken to show that fulicin is genetically encoded.9 To this point it may appear that incorporation of D-amino acids is restricted to the second position of small peptides (indeed, there are over a dozen such examples). However, a hyperglycemic hormone, which regulates blood glucose levels, was isolated from crayfish and lobsters and was shown to be 78 amino acids long and to contain a D-phenylalanine residue at the third position.4 More striking than this is the discovery of two isomeric peptides from the venom of the American funnel-web spider Agelenopsis aperta. These peptides, named oo-Agatoxin IVB and IVC, are 48-residue neurotoxins that function to paralyze prey by blocking whole cell C a 2 + current. They share identical amino acid sequences, have C-terminal carboxylates (as opposed to amides in most of the examples above), and have the same disulfide bridge motifs.10 Reversed-phase HPLC analysis, however, confirmed that they are unique species, and it wasn't until a peptidase digestion was performed that their distinction was realized." The C-terminal fragments were analyzed, revealing that IVB contains a D-serine at position 46, while IVC contains L-serine (Figure 1.4). Again, cDNA cloning confirmed that a larger, genetically encoded precursor becomes cleaved to give IVC, which in turn becomes modified to give I V B . 1 2 co-Aga IVC C O O H C O O H Figure 1.4 Isomeric peptide toxins isolated from the venom of Agelenopsis aperta. D-Ser46 (blue), L-Ser46 (red), other amino acid residues (green), and disulfide bridges (yellow) are indicated. Illustration has been modified from Krei l . 1 3 In this section, we have seen several examples of how nature has expanded its stereochemical library by introducing D-amino acids into peptides, both from genetic and non-genetic origins. What has been omitted until this point, however, is the means by which these inversions occur. Whereas in some cases, such as the majority of ribosomally derived peptides, there is very little knowledge as to the cause of these modifications, there are several examples of enzymes that have been shown to generate amino acids and peptides possessing the D configuration. 10 1.3 Amino Acid Racemases and Epimerases Enzymes that catalyze the inversion of a stereocentre* in a molecule are called racemases (Figure 1.5a) i f the substrate and product are enantiomers and epimerases (Figure 1.5b) i f they are epimers (diastereomers that differ in configuration at only one centre). In the context of peptides and amino acids, the former, containing multiple stereocentres, would be inverted by epimerases, while the latter, having side chains with or without a stereocentre, would be inverted by epimerases or racemases, respectively, a) b) Figure 1.5 Generalized reactions catalyzed by racemases and epimerases. A, B, C, and D are unique atoms or groups in which a) all lack a stereogenic centre or b) at least one group (A*) has a stereogenic centre. A l l amino acid racemases and epimerases currently known utilize deprotonation and reprotonation reactions to bring about stereochemical inversion.14 A question that was originally presented by Rose was whether this deprotonation/reprotonation is conducted by the same enzymatic residue (one-base mechanism; Figure 1.6a) or two distinct residues (two-base mechanism; Figure 1.6b).15 In either case, a basic enzymatic residue abstracts the a-proton to produce a planar carbanion intermediate. The opposite face is then protonated by an acidic residue to give the inverted stereocentre. In addition to categorization by the number of bases, amino acid racemases and epimerases are commonly divided into two groups depending on whether they utilize a cofactor to facilitate catalysis. Accordingly, representative examples of * A stereocentre (or stereogenic centre) is an atom bearing groups of such a nature that the interchange of any two groups yields a stereoisomer. In order to satisfy the correct definitions of "racemase" and "epimerase", this term has been restricted in this thesis to only tetrahedral centres (thereby excluding cis-trans isomers). 11 a) b) ENZ ENZ BiH ENZ •Bo Figure 1.6 One- and two-base mechanisms of amino acid racemases. a) The one-base mechanism requires repositioning of the enzymatic base (B) and/or intermediate to reprotonate, whereas b) the two-base mechanism utilizes distinct bases (B-i, B2). cofactor-dependent and cofactor-independent racemases and epimerases will be described, including the establishment of their mechanisms and categorization as one-base or two-base. 1.3.1 Cofactor-Dependent Amino Acid Racemases Because the a-proton of amino acids is relatively non-acidic, some racemases utilize the electron-withdrawing cofactor pyridoxal phosphate (PLP) to lower the pKa. The cofactor, which is initially bound to the enzyme's active site lysine residue as an imine (Schiff s base), undergoes a transimination to form an iminium ion with the amino acid (Figure 1.7).16 This adduct can 12 Figure 1.7 Mechanism of PLP-dependent racemases. then be readily deprotonated to generate a resonance-stabilized carbanion. Reprotonation from the opposite face and subsequent transimination yields the enantiomeric product. There exist several amino acid racemases throughout nature that utilize PLP for catalysis. These include arginine racemase from Pseudomonas graveolens,11 which preferentially 18 racemizes both arginine and lysine, serine racemase from Streptomyces garyphalus , silkworm, 1 9 and rat brain, 2 0 and racemases with broad specificity from Aeromonaspunctata21 Pseudomonas striata22 and Pseudomonas putida.23 In some cases, attempts have been made to determine the number of bases employed, but the results are largely inconclusive.14 13 Alanine racemase The best understood amino acid racemase that uses PLP is alanine racemase. This enzyme is common to many species of bacteria and is responsible for generating one of the D-amino acids present in peptidoglycan (Section 1.2.1). Alanine racemase from Streptococcus faecalis was first isolated in 1951, wherein its dependence on PLP was realized.2 4 Subsequently, much work has been done in attempts to determine whether the mechanism involves one or two bases. Two models have been proposed, each of which fulfills certain stereochemical criteria: both enantiomers must be accommodated in one active site, both faces of the planar PLP-adduct must participate in proton exchange with the enzyme, and the labile C - H bond must be oriented perpendicular to the cofactor.25 The model suggested by Henderson and Johnston is a "swinging Figure 1.8 Models of alanine racemization. a) Henderson and Johnston's one-base "swinging-door" mechanism. Ring substituents have been omitted for clarity, b) Adams' two-base mechanism, viewed along the C a - N bond. The planar PLP-adduct is represented by a rectangle. 14 door" one-base mechanism (Figure 1.8a), whereby the enzymatic base deprotonates the alanine-PLP complex to give a resonance stabilized carbanion, which rotates on an imaginary hinge to Oft expose the opposite face and is then reprotonated by the same enzymatic residue. In contrast, the two-base model proposed by Adams (Figure 1.8b) requires no reorientation of the intermediate, as the opposite face can be reprotonated by an appropriately positioned residue.16 A distinguishing characteristic of racemase mechanisms that is routinely explored is the nature of the hydrogen in the product with respect to substrate and solvent. Evidence for a one-base mechanism is obtained i f a phenomenon known as internal return is observed. Internal return occurs when the hydrogen that is removed from the substrate is reattached to form the product. Thus, i f substrate specifically substituted with deuterium or tritium at the alpha carbon were incubated with the enzyme in H2O and the product analyzed, then the presence of an isotopic label would confirm a one-base mechanism (equivalently, unlabelled substrate may be incubated with enzyme in deuterated or tritiated water). When L-[a- HJalanine was tested with 27 alanine racemase, no label was observed in the D-alanine product. Although this result may seem to support a two-base mechanism, which necessarily involves incorporation of hydrogen from the solvent by means of the second enzymatic base, it is indistinguishable from a one-base mechanism in which solvent exchange is much faster than catalytic turnover, allowing the proton on the enzymatic base to wash out during the lifetime of the intermediate. More recent work has removed much of the earlier mechanistic uncertainty, ultimately favouring a two-base mechanism. The X-ray crystal structures of alanine racemase from Bacillus stearothermophilus have implicated specific residues as serving catalytic roles. In the absence of a substrate or analogue, Lys39 was observed covalently bound to PLP via an aldimine, whereas when the enzyme was crystallized in the presence of a substrate analogue, Lys39 was found to be positioned appropriately to act as a general acid/base residue. Additionally, the structure showed that Tyr265 from the other monomer of this dimeric enzyme 15 could serve as the second acid/base residue.29 The one/two-base question has essentially been put to rest through the site-directed mutagenesis of a neighbouring arginine residue, which is linked to Tyr265 through a hydrogen-bond network. The mutant enzymes exhibited decreased reactivity in the proton transfer to and from only one enantiomer, thereby demonstrating that the role of Tyr265 is to act as a general acid/base for this substrate isomer. 1.3.2 Cofactor-Independent Amino Acid Racemases and Epimerases Although the pyridoxal phosphate cofactor greatly enhances racemization by stabilizing a carbanionic intermediate, many enzymes are still capable of efficiently catalyzing racemization and epimerization without it. Proline racemase The earliest studies on such cofactor-independent enzymes involved proline racemase, which is produced in Clostridium sticklandii and serves a degradative role in proline metabolism. Upon conversion from the L-enantiomer, D-proline is ring-opened to 8-aminovalerate (5-aminopentanoate) by D-proline reductase (Figure 1.9), a process that is physiologically important in the anaerobic metabolism of amino acids.31 H 2 proline H 2 D-proline L-Pro D-Pro Figure 1.9 Proline metabolism in Clostridium sticklandii. 16 The majority of mechanistic details of proline racemization were determined by Abeles and coworkers. 3 2 ' 3 3 Not unexpectedly, incorporation of solvent isotope into the alpha position was observed when substrate was incubated in D2O, indicating that a 19 deprotonation/reprotonation mechanism was operative. Analysis of product and recovered substrate in the early stages of reaction, however, revealed that 100% of the product contained C O , a) -IAAALAA/VP b) L-Pro C O , H, I H' B H B BD aJuvuvrLruinri D CO H / K H H D-[a-2H]Pro (p 1 # ' JL (CH 2) 3 V (CH 2 ) 3 BH H + BD C O , H L-[a-2H]Pro g - (not observed) C O , H BH B" C0 2 H^^H (CH2)3 BH BD L A I U V V U \ A / J A T I D > < ,SH+ co 2 _ H^^H (CH2)3 BD BD L T J V V V V V V I A / ' ' i/\J\n/uvruvlrvr> M D D-[a-2H]Pro C O , H. I H L-[a-2H]Pro (not observed) B~ B Figure 1.10 Models of proline racemization. a) One-base mechanism and b) two-base mechanism. Proline is presented as a projection along the C a - N bond. 17 deuterium and 100% of the substrate retained protium, regardless of reaction direction. These results exclude the possibility of a one-base mechanism (Figure 1.10a). To illustrate this, consider the racemization of L-proline in D2O. In order to account for complete incorporation of deuterium (i.e., lack of internal return) into D-proline using a one-base mechanism, rapid proton exchange of the enzymatic base with solvent must occur. The absence of deuterium in L-proline would require a skewed partitioning of the intermediate towards D-proline. If these conditions hold, then when studying the reaction in the reverse direction, deuterium should be incorporated into the substrate, D-proline, much faster than into the product, L-proline. Although the findings rule out this possibility, they are compatible with a two-base mechanism in which no solvent isotope exchange occurs during the lifetime of the carbanionic intermediate (Figure 1.10b). Further evidence in support of a two-base mechanism came from examination of the two enzyme protonation forms ( E L and E D ) involved in catalysis (Figure 1.11). These forms are capable of interconverting either by direct exchange of enzyme-bound protons with solvent (top path) or via binding and catalysis (bottom path). Rudnick and Abeles studied the kinetic barrier to interconversion by incubating varying concentrations of DL-[ct-3H]proline with the racemase and monitoring the release of tritium into solvent. As the proline concentration increased, the B " BH _ BH ET : l 1 _ L — J L L-Pro H, D-Pro C 0 2 C 0 2 ,H H ' V ^ ^ . ^ " V _ ^ H H ^ H E L -L -P ro ? " B H BH B- E [ ) . D . P r o Figure 1.11 Interconversion of enzyme protonation states. E L and E D represent the protonation forms capable of inverting L-proline and D-proline, respectively. The carbanionic intermediate has been omitted for clarity. 18 rate of tritium loss decreased, indicating that the substrate-derived proton is protected from solvent and only exchanges after product is released. When DL-proline is present in high concentrations, catalytic turnover in the reverse direction outcompetes direct interconversion of the two enzyme states, resulting in the recapture of tritium by proline and the concomitant reduction in the rate of tritium release into solvent. A similar investigation of the two enzyme forms was conducted by incubating DL-[14C]proline with proline racemase in a "tracer perturbation experiment."34 When a large excess of unlabelled L-proline was added to a sample of DL-[1 4C]proline in the presence of the racemase, a flux of radioactive material from D-proline to L-proline occurred, even though the overall flux of proline was in the opposite direction (L —> D) (Figure 1.12). This phenomenon can be explained by considering that the large excess of L-proline effectively shifts all of the enzyme into the E D form, whereupon it may revert to the E L form either by direct exchange with solvent or by binding and inverting D-[14C]proline. Because increased label was detected in the L-proline pool, it is clear that there is a significant kinetic barrier to direct interconversion of E L and E D - By measuring the magnitude of the flux, a rate constant of 105 s"1 was estimated for this interconversion. Time (min) Figure 1.12 Tracer perturbation experiment. 19 Glutamate racemase Glutamate racemase is a cofactor-independent racemase that is responsible for generating the D-glutamate that is incorporated into peptidoglycan in bacteria (Section 1.2.1). In a similar vein as with proline racemase, isotope labelling experiments were used to show that glutamate racemase utilizes a two-base mechanism. 3 5 ' 3 6 Site-directed mutagenesis of two cysteine residues 36 resulted in loss of racemase activity, implicating these as the catalytic bases. Furthermore, each of these Cys—> Ala mutants was capable of eliminating HC1 from a single enantiomer of the substrate analogue //zra?-P-chloroglutamate (Figure 1.13). Because this process requires only one base, each cysteine could be designated as serving to deprotonate one particular substrate enantiomer. 0,C H f\Y = C 0 2 C NH3 r + T3 C73 ovuuwi C184A OoC H,N' .H + HCI CO, CI 3 0 2 C = NH3 S CH 3 I I C73A l A A / u v m C184 Figure 1.13 Elimination of HCI from ^reo-P-chloroglutamate catalyzed by glutamate racemase mutants. The CI84A mutant operates only on the 25,35' enantiomer (left), while the C73 A mutant operates only on the 2R,3R enantiomer (right). 20 1.4 Posttranslational Epimerization of Peptides Whereas the racemization of amino acids has been largely well-characterized, the mechanism of inversion of stereocentres in peptides, with few exceptions, has remained uncertain. As outlined in Section 1.2.2, the presence of D-amino acids in peptides of genetic origin has strongly indicated that a posttranslational process is involved. The identification of the corresponding enzymes that bring about these modifications, however, has been elusive. Examples of ribosomally derived peptides that are synthesized via an inversion of stereochemistry are the dermorphins and deltorphins (Figure 1.3) and the antibiotic peptides known as lantibiotics. Lantibiotics are cyclic peptides that often incorporate the unusual amino acids lanthionine and P-methyllanthionine, which are dimeric amino acids that result from the modification of precursor peptides (prelantibiotics) containing serine and threonine residues, respectively (Figure 1.14).37 It has been suggested that the prelantibiotics undergo a dehydration to give dehydroalanine and dehydroaminobutyric acid residues, to which a nearby cysteine residue may add to yield a cyclic peptide. This process results in the formation of a residue that bears the D-configuration and thereby is an overall inversion of stereochemistry. Figure 1.14 Formation of lanthionine, a dimeric amino acid with D-configuration. 21 Spider Venom Peptide Epimerase While exploring the peptide components of the venom isolated from the North American funnel-web spider Agelenopsis aperta, Heck et al. discovered the isomeric toxins o A g a - I V C channels when tested on rat cerebellar Purkinje cells. The functional consequence of this blockage is that nerve conduction in the cerebellum, the brain tissue responsible for motor control, is inhibited and the spider's prey becomes paralyzed. The two toxins, however, exhibit differing potency and stability; IVB, which bears a D-serine residue as the result of a posttranslational modification from IVC, is fivefold more effective and is more resistant to degradation by a peptidase also present in the venom. Thus, it is evident that the introduction of an unusual amino acid by means of a posttranslational epimerization represents a unique strategy that enables the venom's toxin to evade the host's defences and elicit an enhanced effect. Further examination of the protein components of the venom revealed an unprecedented discovery: the spider also produces an enzyme that catalyzes the interconversion of the epimeric peptides, thereby marking the first identified polypeptide epimerase.12 The novelty of this finding prompted Heck et al. to explore the nature of the epimerization in terms of substrate specificity and catalytic mechanism. In determining specificity, it was found that the natural substrates IVC and IVB could be optimally truncated to 1 and 2, respectively, which possess only the final five residues including the epimerizable serine. These shortened substrates provided a more convenient means by which to assay activity and also facilitated synthesis of and IVB (Figure 1.4). 11, 12 These peptides were found to block specific voltage-dependent Ca' 2+ . O H o y o yy .OH .OH Ac-Glv-Leu-LSer-Phe-Ala 1 Ac-Gly-Leu-DSer-Phe-Ala 2 22 analogues containing other amino acids in place of serine. In particular, threonine, cysteine, alanine, and O-methylserine were examined. Whereas the threonine analogue - the only one bearing a hydroxyl group - was inactive, the others were epimerized to varying extents. Outlined in Figure 1.15, three mechanistic models were suggested and probed using various methods. The dehydration-rehydration model (a), similar to the formation of lanthionine, involves a base-assisted elimination of H2O to generate a planar dehydroalanine intermediate. Re-addition of H2O with concomitant protonation on the opposite face of the C a -Cp double bond yields the inverted stereocentre. A retroaldol/aldol mechanism (b) proceeds via cleavage of the C a -Cp bond to liberate formaldehyde. Nucleophilic attack by the opposite face of the planar enolate intermediate would re-form the C-C bond with the opposite stereochemistry. Figure 1.15 Proposed mechanisms for serine epimerization by spider venom peptide epimerase. Inversion of the serine residue by a) dehydration/rehydration, b) retroaldol/aldol, and c) deprotonation/reprotonation mechanisms. The remainder of the peptide chain is represented by wavy bonds. 23 A third possibility is the familiar deprotonation/reprotonation mechanism (c). Because the venom epimerase was shown to operate independently of PLP or metals, a two-base mechanism analogous to those utilized by cofactor-independent racemases was suggested.38 The dehydration/rehydration and retroaldol/aldol mechanisms were disfavoured by two experiments. The former was weakened by the absence of 1 8 0 label in the recovered epimers when the epimerization was performed in H2 O. Provided that the active site is solvent accessible when it is occupied by the intermediate, a labelled water molecule should be able to replace the unlabelled one during the lifetime of the intermediate, leading to 1 8 0 incorporation into the two epimers. The latter mechanism was weakened by the inability of hydroxylamine, a known scavenger of free formaldehyde, to intercept the hypothetical formaldehyde intermediate. Furthermore, the fact that some peptide analogues lacking a hydroxyl group are competent substrates (vide supra) strongly disfavours these two mechanisms. Only the deprotonation/reprotonation mechanism could be supported by experimental evidence. In a similar approach as used in studying cofactor-independent racemases, solvent isotope incorporation into product was observed, regardless of reaction direction. Additionally, substantial primary kinetic isotope effects were observed when incubating the enzyme with substrate containing [2,3,3-2H3]serine or with unlabelled substrate in D2O. A kinetic isotope effect is a manifestation of the difference in mass between two isotopes, such as protium and deuterium. Bonds to heavier isotopes require more energy to break; thus, there is an increased kinetic barrier involved in cleaving a C - D (or O-D) bond relative to a C - H (or O-H) bond. Consequently, i f the bond is broken in a rate-limiting step of a reaction, a decrease in reaction rate would be measured in the case of a heavier isotope. The findings by Heck et al. indicate that epimerization involves both deprotonation (C-H bond cleavage) and reprotonation (O-H bond cleavage) and that these steps are partially rate limiting. 24 1.5 Acidity of the a-Proton The crux of all amino acid racemases and epimerases is their ability to efficiently remove the a-proton in their substrates. With catalytic rate constants as high as 103 s"1 (versus 10"11 s"1 for the uncatalyzed reaction),39 these enzymes are capable of lowering the activation energy barrier (AG*) from -32 to 13 kcal/mol, as determined by transition state theory (Equation 1.1). AG* =-RT\n\ th^\ h = Planck's constant k& = Boltzmann's constant 1.1 The large kinetic barrier in the uncatalyzed deprotonation may be primarily attributed to the same factors responsible for the large thermodynamic barrier involved in the conversion of the amino acid zwitterion to its enol zwitterion form or enolate monoanion form (pKen0\ = 20, ( P ^ C H ) O = 29; Figure 1.16). This 27 kcal/mol barrier (as given by the equilibrium free energy equation, AG° = 2.303 RT pKeno\) exceeds the observed AG* for the enzymatic reaction by 14 kcal/mol; thus, in order for racemases and epimerases to liberate the a-proton, they must lower the thermodynamic barrier by selectively stabilizing the enol intermediate. •A H 3 N amino acid 0 zwitterion enolate monoanion OH e n o 1 zwitterion Figure 1.16 Conversion of an amino acid zwitterion to enol zwitterion. Associated pK values are indicated, as determined by Williams et al. for proline.3 9 25 1.5.1 Intermediate Stabilization Several mechanisms by which the enol may be preferentially bound have been suggested. It is likely that the enzyme's active site accommodates a planar environment at C-2 better than a tetrahedral one. This notion has been supported through the use of substrate analogues with sp hybridization at C-2, such as 3 and 4 (Figure 1.17), which have been found to bind orders of magnitude stronger to the enzymes proline racemase and glutamate racemase, respectively, than the corresponding substrates.33'40 Another major contributor to the increased stabilization of the intermediate is hydrogen bonding. It is assumed that the intermediate of the enzymatic reaction is largely the enol zwitterion (i.e., monoprotonated at oxygen) rather than the negatively charged enolate because protonation of the dianion (pA^ oH = 9; Figure 1.16) is favoured at neutral pH . 3 9 Consequently, this protonated enol may form an important hydrogen bond to the enzyme that is not present in the substrate complex. It has been proposed that such a hydrogen bond may be unusually strong, such that the hydrogen is shared nearly equally between the enolic oxygen and the general acid catalyst in what is termed a low-barrier hydrogen bond (LBHB). 4 1 If the pKa of the general acid matches the pKa of the enol, the L B H B interaction is maximally enhanced, resulting in estimated glutamate 4 enol Figure 1.17 Enol intermediates of proline and glutamate and corresponding analogues. 26 bond strengths exceeding 20 kcal/mol, thereby significantly lowering AG°. The existence of LBHBs, however, remains hotly contested; nevertheless, in the hydrophobic, low-dielectric environment of an enzyme active site, hydrogen bonds may have strengths of 5 kcal/mol or more.42 A third factor that may provide stability to the enol zwitterion is a strengthening in the intramolecular electrostatic interactions between the positively and negatively charged atoms. In comparison to the amino acid zwitterion, the enol form places the negative charge closer to the positive ammonium group, an interaction that is favourable, particularly at an active site of low dielectric constant.39 1.5.2 Abstraction of the a-Proton Although amino acid racemases and epimerases rely on the above strong binding interactions with the enol intermediate, they also utilize various chemical strategies to assist removal of the a-proton. Perhaps the most important factor in enzyme-catalyzed deprotonation reactions is the enzyme's ability to perform simultaneous acid and base catalysis. Whereas keto-enol tautomerization free in solution is a stepwise process that proceeds through a higher energy intermediate, an enzyme may catalyze a direct tautomerization by a concerted process involving deprotonation of the a-proton and protonation of the carbonyl.43 Another important method of enhancing a-proton removal is by first acidifying the a-amino and carboxylate groups. Although the neighbouring carbonyl offers resonance stabilization of the carbanionic intermediate, deprotonation at the a-carbon is greatly enhanced by an adjacent electron-withdrawing cationic ammonium group. Furthermore, neutralization of the carboxylate by 27 protonation greatly lowers the pKa of the a-proton. As outlined in Figure 1.16, the pKa of this form of an amino acid (22) is 7 units lower than that of the zwitterion (29). 3 9 ' 4 4 1.5.3 Acidity of the a-Proton in a Peptide The factors described above may be sufficient to account for the observed enzymatic rates for amino acid racemization and epimerization. A slight complication, however, may result when the a-proton is located within a peptide. Two significant differences between a free, fully protonated amino acid (pKa = 22) and the corresponding residue in a peptide may lead to a significant decrease in acidity. First, the inductive influence of the positively charged ammonium group is absent when replaced with a neutral amide N H . Second, replacement of the carboxylic acid with an amide introduces a less electron-withdrawing carbonyl. These changes can each be estimated to increase the pKa by up to 5 units, 4 4 , 4 5 resulting in a p ^ that may be as high as 32. Thus, the deprotonation of a peptide a-proton presents a peptide epimerase with a challenge greater than that faced by an amino acid racemase. 1.6 Project Goals The overall aim of this thesis is to probe the mechanisms by which peptides and carbohydrates are epimerized by enzymes. Although carbohydrate and peptide epimerization share many similarities, such processes are distinct enough to warrant separate analyses. Accordingly, a detailed account of carbohydrate epimerases is presented in Chapter 3, followed by the mechanistic characterization of such an epimerase in Chapter 4. In order to provide the 28 appropriate context, goals related to this specific project have been deferred until the end of Chapter 3. The uniqueness of the spider venom peptide epimerase makes it an attractive target for investigation. Furthermore, the enzyme's ability to liberate the weakly acidic peptide a-proton adds additional intrigue. Therefore, Chapter 2 is dedicated to addressing the mechanism of this enzyme and the nature of the enzyme-intermediate complex. First, pertinent aspects of the protein including its purification are discussed. A n alternative deprotonation/reprotonation mechanism that employs covalent catalysis is suggested and evaluated by means of several substrate analogues. The synthesis of these modified peptides is described, and their efficacy as inhibitors is analyzed. The inhibitor studies allow for a better understanding of the structure of the enzyme-bound intermediate, ultimately favouring the deprotonation/reprotonation mechanism shown in Figure 1.15c. 29 Chapter 2 Mechanistic Studies on a Spider Venom Peptide Epimerase 30 As described at the end of Chapter 1, the primary goal of this project was to understand the mechanism by which the spider venom peptide epimerase inverts its peptide substrates. Presented by Heck et al. and illustrated in Figure 1.15 were three mechanisms that could potentially account for this process. Two of these, the dehydration/rehydration mechanism and the retroaldol/aldol mechanism, were strongly disfavoured through their experiments, but the third, the deprotonation/reprotonation mechanism, was ultimately supported by their findings and by the precedence of amino acid racemases. Accordingly, this mechanism was considered in evaluating the results of the work presented in this chapter. Also discussed at the end of Chapter 1 is that the a-proton in the enzyme's peptide substrates is weakly acidic. The removal of this proton potentially poses a challenge to the epimerase, one that a simple deprotonation/reprotonation mechanism may be insufficient to overcome. The method by which it accomplishes this is not immediately apparent. It is clear that it follows a pathway that lowers the energetic barriers preventing enolization, but the lack of any associated cofactors limits the options available. Described in the next section, however, an intriguing discovery by another research group prompted us to suggest an alternative mechanism involving covalent catalysis. Thus, the majority of this project has focused on the evaluation of these possible mechanisms through the synthesis of a number of substrate and intermediate analogues. These peptides, described in Sections 2.3 - 2.6, introduced modifications that ultimately enabled us to favour one of the mechanisms. 31 2.1 Protein Homology to Serine Proteases 2.1.1 Protein Structure and Sequence Determination In the short period between the initial report by Heck et al. on the existence of the spider venom epimerase12 and their follow-up analysis on its mechanism,38 a Japanese group independently isolated and sequenced the enzyme. Following a similar purification as that employed by their American counterparts, Shikata et al. observed epimerase activity in a size-exclusion column fraction containing a protein of molecular weight 29 418 Da, as determined by matrix-assisted laser desorption ionization (MALDI) mass spectrometry.46 The protein appeared to be homogeneous by electrophoresis. The protein was subjected to amino acid sequencing, revealing some peculiar features. First, N-terminal sequencing yielded two amino acid sequences with approximately equal recovery, indicating that the enzyme consists of two polypeptide chains (Figure 2.1). This conclusion was confirmed by disulfide bond reduction with dithiothreitol (DTT), followed by S-alkylation of the free cysteine side chains and HPLC separation. Second, limited proteolytic digestion and Edman degradation identified four disulfide linkages, three which are internal to the heavier polypeptide chain and one which serves as the bridge joining the heavy and light 0 Mans f U C X-V / ^ V ^Man-GlcNAo—GlcNAc j j ( ) NH2 C-C G-N C""C C '"C COOH 1 31 47 | 127 159 181 19G 219 243 NH2-—C — — COOH T 9' 18* Figure 2.1 Primary structure of the peptide epimerase. Amino and carboxy termini are indicated by NH2 and COOH, and cysteine (C) and asparagine (N) residues are labelled. Disulfide bridges are indicated by dashed lines. Figure has been reproduced from Shikata et al. chains. Third, further degradation revealed that Asnl27 is covalently attached to an oligosaccharide chain composed of mannose, A^-acetylglucosamine, and fucose residues. Such a posttranslational modification is not uncommon in eukaryotic proteins possessing an Asn-Xaa-Thr motif (where Xaa is any amino acid except Pro). By far the most unusual revelation regarding this protein was illuminated by a protein Epimerase Kallikrein Thrombin Flavoxobin Batroxobin Epimerase Kallikrein Thrombin Flavoxobin Batroxobin Epimerase Kallikrein Thrombin Flavoxobin Batroxobin Epimerase Kallikrein Thrombin Flavoxobin Batroxobin Epimerase Kallikrein Thrombin Flavoxobin Batroxobin Epimerase Kallikrein Thrombin Flavoxobin Batroxobin 10 20 30 40 IVGGKTAKFGDYPWMVSIQQKNKKGTFDHI CGGAIINVNWILTAAHC IVGGTNS SWGEWPWQVSLQVKLTAQ IVEGQDAEVGLSPWQVML FRKSP VIGGDECNINEHPFLVALYDAWSG VIGGDECDINEHPFL AF MYYSP 60 RHL CGGSLIGHQWVLTAAHC QELLCGASLISDRWVLTAAHCLL RFL CGGTLINPEWVLTAAHC RYF CGMTLINQEWVLTAAHC 50 70 80 90 FDQPIVKS DYRAYVGLRSILHTKENTVQRLELSKIVLHPGYK P FDGLPLQD VWRIYSGILNLSDITKDTPFS Q I K E I I I H QNYKV YPPWDKN FTVDDLLVRI GKHSRTRYERKVEKISMLDKIYIHPRYNWK DSKNF KMK LGAHSKKVLNEDEQIR NPKEKFICPNKKND NRR F MRIHLGKHAGSVANYDEWR YPKEKFICPNKKKN 100 110 120 130 KKDPDDIALIKVAKPIVIGNYANGICVP KGVT NPEG NA TVIGWGK SEGNHDIALIKLQAPLNYTEFQKPICLPSKGDT STIYTNC WVTGWGF ENLDRDIALLKLKRPIELSDYIHPVCLPDKQTAAKLLHAGFKGRVTGWGN EVLDKDIMLIKLDSPVSYSEHIAPLSLPSSPPS VGSVC RIMGWGS VITDKDIMLIRLDRPVKNSEHIAPLSLPSNPPS VGSVC RIMGWGA 140 150 ISS GGKQ VNTLQEV T I P I SKE KGEI QNILQKV NIPL VTNEECQK RRE TWTT S VAE VQ P S VLQ W N L PL VE RP VCKA ITP VEET FPDVPHCANINL LDDVECKP ITT SEDT YPDVPHCANINL FNNTVCRE 180 190 200 210 220 ICAGA E G KDSCQADSGGPLFQID ANGVATLIGTVANG ADCGY VCAGYK EGG KDACKGDSGGPLVCKH NGMWRLVGITSWG EGCAR FCAGYKPGEGKRGDACEGDSGGPFVMKSPYNNRWYQ MGIVSWG EGCDR LCAGVL QGG IDTCGFDSGTPLIC NGQFQ GIVSYGGHPCGQ LCAGVL QGG IDTCGGDSGGPLIC NGQFQ 230 240 KHYPGVYMKVSSYTNWMSKNMV REQPGVYTKVAEYMDWILEKTQSSDGKAQMQ S PA DGKYGFYTHVFRLKKWIQKVIDRLGS SRKPGIYTKVFDYNAWIQSIIAGNTAATCLP PRKPAFYTKVFDYLPWIQSIIAGNKTATC P 160 170 IPWKKCKEIYGDEFSEFEYSQITPYM RYQDYKITQRM STRI RITDNM GYPELLPEYRT AYNGLP AKT GILSWGSDPCAE Figure 2.2 Sequence homology between the peptide epimerase and serine proteases. Alignments were made by Shikata et al. to maximize homology; spaces indicate residues that are absent. Identical residues are blue, and the residues comprising the putative catalytic triad are red. 33 sequence database search. Although the light chain, only 18 amino acids in length, shows no homology with other proteins, the heavy chain shares up to 35% sequence identity with several serine proteases (Figure 2.2).4 6 Of particular note is that the ubiquitous protease catalytic triad, composed of Serl94, His46, and Asp96, and several flanking residues are conserved in the epimerase. Adding further intrigue is the observation by Heck et al. that the enzyme is inhibited by phenylmethylsulfonyl fluoride (PMSF), a serine protease inhibitor. 2.1.2 Modified Deprotonation/Reprotonation Mechanism We postulated that these findings were indicative of the epimerase's mechanism, and accordingly, we proposed a mechanism that accommodates both a deprotonation/reprotonation reaction and the serine protease machinery. As depicted in Figure 2.3, step 1 proceeds in the same way as the first step utilized by proteases, wherein the serine residue, made nucleophilic by a hydrogen bond network with histidine and aspartate, attacks the amide carbonyl of the substrate's serine residue. Completion of the step occurs following collapse of the tetrahedral intermediate (not shown), which results in displacement of the C-terminal peptide. At this point, the acyl-enzyme species, by virtue of its more electron-withdrawing ester functionality, possesses a more acidic a-proton (approximately 5 pATa units lower;4 5 recall Section 1.5.3). Consequently, it may be possible to invoke a deprotonation mechanism (step 2) to generate a planar enolate. Non-stereospecific reprotonation (step 3) enables the formation of the inverted stereocentre. Finally, the transiently displaced amine of the C-terminal peptide adds back to the carbonyl to generate the epimeric product (step 4). Figure 2.3 Acyl-enzyme mechanism of peptide epimerization. For clarity, only the epimerizable serine is depicted; remaining residues are represented by wavy bonds. Tetrahedral intermediates in steps 1 and 4 have also been omitted for simplicity. Much of this project, therefore, has been dedicated to evaluating the mechanism of the peptide epimerase in light of the possibility that it proceeds via covalent catalysis. The overall strategy involves the synthesis and testing of substrate analogues designed to trap the enzyme in a covalently modified state which could be detected by mass spectrometry. If an increase in mass were observed, proteolytic digestion of the enzyme into smaller, manageable fragments would enable the determination of the nature of the modification, and thereby allow us to confirm or alter our view of the enzymatic epimerization. Furthermore, regardless of whether or not a covalent adduct were detectable, the use of appropriate analogues may provide insight as to the character of the intermediate and transition state involved in isomerization. 35 The sections that follow, therefore, describe the approaches we used to assess the nature of the spider venom peptide epimerase. First, enzyme purification and activity issues are explored. This is followed by the rational synthesis and evaluation of two substrate analogues designed to probe the two major mechanistic models - the simple deprotonation/reprotonation mechanism and the covalent catalysis mechanism. Then, influenced by the results of these studies, the design and testing of two intermediate analogues are discussed. Finally, based on our findings, a clearer picture of the mechanism is presented along with a number of suggestions for future investigations. 2.2 Enzyme Purification and Activity 2.2.1 Enzyme Source and Activity Before delving into mechanistic studies, it is appropriate to address certain aspects of the enzyme that present some limitations and complications to the experimental approaches available to us. The source of the peptide epimerase studied in this thesis is the venom obtained from the spider Agelenopsis aperta (Figure 2.4). The suppliers of this venom obtain it by passing a non-destructive current across the anesthetised spider's head and collecting the secreted venom into capillary tubes.47 Because the average yield of crude venom is only ~0.2 \xL, several hundred milkings must be pooled to obtain sufficient quantities for experimentation. Figure 2.4 Funnel-web spider, Agelenopsis aperta. Photograph is approximately actual size. 36 Accordingly, the crude venom was supplied in 100 uL portions, which, following purification, provided only enough enzyme (typically ~ 0.5 mg) for 30 - 40 assays (vide infra). Further complicating studies on the epimerase is its very low specific activity. As described by Heck et a/. 3 8 and used throughout our studies, the epimerase activity was assayed by incubation with one of its substrate isomers at 37 °C in the presence of 5 mM ethylenediaminetetraacetate (EDTA; see Section 2.2.2). At various time points, aliquots were analyzed by reversed-phase HPLC and the extent of reaction was determined by peak integration. Consistent with the findings of Heck et al, the enzyme was found to interconvert the epimers at a very slow rate (1 x 10"4 |j.mol min 1 mg"1 for substrate 1); as Figure 2.5 illustrates, typically only 25% conversion of 1 had occurred after 21 h of incubation. Because of this, it was necessary to perform lengthy incubations (5 - 24 h) in order to obtain sufficient product to measure reliably by HPLC. Thus, this time requirement, combined with the limitation on the number of possible assays mentioned above, placed a restriction on the range of experimental options available to us. Cloning is a tool commonly utilized by molecular biologists to overcome such problems of low protein quantities. By obtaining the gene encoding the protein of interest and applying it 2.0 1.5 8 1.0 0.5 a) X Substrate 1 X b) 0 2 4 6 8 10 12 14 16 18 20 22 0 2 4 Time (min) Substrate 1 8 10 12 14 16 18 20 22 Tims (min) Figure 2.5 Typical epimerization assay. Reversed-phase HPLC profiles of the incubation of 100 u.M substrate 1 (LSer) with the enzyme in the presence of 5 m M EDTA at 37 °C for a) t = 2 min andb) t = 21 h. 37 to an appropriate expression system, it is theoretically possible to obtain unlimited amounts of recombinant protein. This strategy is not likely to be an option in this case, however. The most significant obstacle in expression of eukaryotic proteins is ensuring that all necessary posttranslational modifications are made. Hence, as mentioned in Section 2.1.1, two separate polypeptides would have to be produced, one of them enzymatically glycosylated, and the correct disulfide bonds must be formed, including the interchain bridge. Such a challenge is beyond the scope of this project; therefore, studies have been formulated to avoid the requirements for large quantities of protein. 2.2.2 Metalloprotease Activity Another complication inherent in the previous studies on this enzyme is that there is a metalloprotease activity associated with the protein that persists after chromatographic purification. It was observed that in the absence of a divalent metal chelator, such as EDTA, rapid proteolysis of the substrate occurred.38 Although this secondary activity on normal peptide substrates could be controlled by the inclusion of 5 mM EDTA in the assay buffer, it was unclear what effect it may have on modified peptides that were to be used to explore the epimerase mechanism. This uncertainty led us to attempt to separate the two activities by performing an orthogonal purification. In discussions with the discoverer of the spider venom epimerase, Dr. Robert Volkmann, it was communicated that several attempts had been made to further purify the enzyme with no success.48 These methods include hydrophobic interaction chromatography, reversed-phase HPLC, and anion-exchange chromatography. It is apparent, however, that Dr. Volkmann and coworkers overlooked another possibility: the theoretical isoelectric point, pi, of the enzyme was 38 calculated from the sequence of the heavy chain and found to be 8.71.4 9 This basic pi suggested to us that cation-exchange chromatography would be a more sensible option, since at neutral pH, the protein would be positively charged. Accordingly, a sample of crude venom was applied to a column containing sulfopropyl strong cation exchange resin and eluted with increasing salt gradients. Fractions containing protein of the expected molecular weight (29.5 kDa) were collected, pooled (Figure 2.6), and assayed to confirm the presence of epimerase activity. These fractions were then applied to a 2^80 b) Time (min) Figure 2.6 Purification of crude spider venom by cation-exchange chromatography. a) Chromatogram depicting absorbance at 280 nm vs. time. Selected fractions are labelled and pooled fractions are shaded, b) SDS-PAGE gel showing the corresponding fractions from the chromatogram and molecular weight markers (MW) in kDa. 39 a ) 1 2 jj |||lja / W \ 3 V ' \ , • . i l l , Time {min) Figure 2.7 Purification of ion-exchange purified epimerase by size-exclusion chromatography. a) Chromatogram depicting A2&0 vs. time. Selected fractions are labelled and the desired fraction is shaded, b) SDS-PAGE gel showing the corresponding fractions from the chromatogram and molecular weight markers (MW) in kDa. Sephadex G-75 column, which is the size-exclusion gel that had been used to purify the enzyme in previously reported applications. The desired protein eluted as a single peak (Figure 2.7), showed a single band on an SDS-PAGE gel, and had a molecular weight (29 473 ± 4 Da, 29 314 ± 6 Da) consistent with the epimerase (calculated weight = 29 471 Da, 29 317 Da*), as determined by electrospray ionization mass spectrometry (ESI-MS). When the orthogonally purified enzyme was tested with substrate 1 in the presence or absence of EDTA, both epimerase and metalloprotease activities were observed. Thus, the additional chromatographic step failed to improve the purification of the epimerase, and so, in all further work, enzyme that had been purified by a single size-exclusion chromatographic step was utilized. * The 29 471 Da peak is expected of 65.5% of all isozyme forms isolated from the venom. 5 0 The 29 317 Da peak is based on a weighted average of the 29 308 Da isozyme (carbohydrate moiety missing a mannose residue; 14.1% of all forms) and the 29 324 Da isozymes (missing a fucose residue; 19.4%). 40 The inability to separate the two activities can be explained by two scenarios: either the epimerase, through misfortune, shares similar size and ionie properties with another protein having protease activity, or a single protein possesses both activities. This latter possibility was tested by incubating both peptide epimers 1 and 2 with the enzyme in the absence of EDTA. Presumably, i f a protein of moderately low molecular weight were bifunctional, it would share a common binding site for both activities. Since it is known that the D-epimer 2 is a substrate for the epimerase, it would be reasonable to expect it also to serve as a substrate for the protease. On the other hand, i f the protease activity was from an impurity, the D-epimer would not be expected to be hydrolyzed since most proteases are specific for peptides containing only L-amino acids. As Figure 2.8 shows, the L-epimer was largely cleaved into smaller, faster-eluting fragments within 40 minutes, whereas the D-epimer remained completely intact (note that the metalloprotease activity is much faster than the epimerase activity, and as such, the extent of epimerization during this relatively short time period is negligible). Thus, the two activities appear to be catalyzed by different proteins. With respect to the possibility that an impurity co-elutes with the epimerase, the issue can become more complex. One must put into question the identity of the major protein that has - 4 o a. 1 3 § C4 a) / Peptide fragments Substrate 0 2 4 6 8 10 12 14 16 18 20 22 0 2 4 6 8 10 12 14 16 18 20 22 Tims (mln) Tlmafmin) Figure 2.8 Specificity of the metalloprotease activity. Reversed-phase H P L C profiles after incubation of 100 u M a) substrate 1 (LSer) and b) substrate 2 (DSer) with the enzyme in the absence of EDTA at 37 °C for 40 min. 41 been isolated and sequenced. It is conceivable that this protein may in fact be the metalloprotease and that the epimerase is actually a small impurity. This notion could account for the extremely low specific activity of the epimerase and is supported by the precedence -albeit scarce - for metal-dependent serine proteases, such as insulin-like growth factor-binding proteins (IGFBPs). 5 1 ' 5 2 The nature of these two activities, however, could be laid to rest i f covalent modification of the enzyme by the use of appropriately designed substrate analogues were found to correlate with inhibition of epimerase activity. Thus, the synthesis and testing of such analogues, which comprised the majority of this project, were undertaken to attempt to answer this question, but more importantly, to address our ultimate goal of understanding the mechanism of peptide epimerization. 2.3 Depsipeptide* Analogue The most compelling evidence for a mechanism that involves an acyl-enzyme intermediate is the isolation and characterization of the covalently bound enzyme adduct. Usually, though, the lifetime of this species is too short to allow for direct detection by conventional biochemical means. Instead, it is often necessary to employ a substrate analogue that has altered reactivity to extend the existence of this intermediate. A mechanism-based inhibitor utilizes the enzyme's catalytic tools to "trick" the enzyme into performing its normal role, and in the process, leads to a build-up of an abortive complex that may then be isolated and analyzed. Depsipeptides are oligomers of amino acids and hydroxy acids. The depside bond, the ester linkage in a depsipeptide, is implied between the amino acid and hydroxy acid (e.g. Ser-PhLac) as recommended by the International Union of Pure and Applied Chemistry ( IUPAC) 5 3 but can also be found in literature symbolized by inserting if^COO) between the amino acid and the corresponding amino acid of the hydroxy acid (e.g. Ser- y/(COO)-Phe). 42 Figure 2.9 Strategy to trap the purported acyl-enzyme adduct using a depsipeptide. Covalent catalysis in the cleavage of 5 could be demonstrated by detection of adduct 6 by ESI-MS and of alcohol 7 by HPLC. One approach that we followed was to introduce an analogue that could acylate the enzyme but whose leaving group would have reduced nucleophilic character such that it could not add back to regenerate the full-length molecule. A functional group that has the potential to satisfy these requirements is an ester, as illustrated by the depsipeptide analogue Ac-Gly-Leu-Ser-PhLac-Ala-OH (5, Figure 2.9). The ester linkage, being inherently more scissile than its amide counterpart, should be readily cleaved by the enzyme's catalytic serine, but the resulting alcohol (7), being a poorer nucleophile than the natural amine, may not be able to re-add, thereby trapping the epimerase in its covalently modified state (6). 2.3.1 Synthesis of Depsipeptide Analogue 5 Although the use of solid-phase synthesis of depsipeptides has been demonstrated in recent years, 5 4 ' 5 5 the requirement for different sets of conditions for the formation of the depside and peptide bonds made this option less attractive. Instead, a convergent solution-phase synthesis of 13, a direct precursor to 5, was initially undertaken as outlined in Figure 2.10. Dipeptide 8 was easily prepared by acetylation of the commercially available peptide H-Gly-Leu-OH. The side-chain-protected serine was Fischer esterified to give 9, which was then 43 o R = H 8, R = Ac o . HO. J l , + ^ A . X>fBu Y ^ O H + H 3 N - ^ Y Cl" J Ac 2 0 93% R = H 1 AcCI, MeOH 9, R = Me 1 A 96% EDC, HOBt, Et 3N, CH 2 CI 2 63% O OBn O EDC, HOBt, Et 3N, CH 2 CI 2 85% H II 1 H 10, R = Me 1 11, R = H ' OR O 1   1 LiOH, MeOH/H 20 82% T o i H O ^ X N X / O f f i u \ JL O 12 ^ O B n O \ s O \ o 13 Figure 2.10 Initially designed scheme towards the synthesis of depsipeptide 5. The final, dashed arrow indicates reaction was unsuccessful (see text for details). coupled to 8. The resulting tripeptide 10 was saponified to yield the carboxylic acid portion (11) of the desired depsipeptide. Phenyllactic acid was coupled without protection of the hydroxyl group to the ^-butyl ester of alanine to give alcohol 12. Unfortunately, attempts to form 13 using a variety of coupling reagents including N,N'-dicyclohexylcarbodiimide (DCC) 56 l-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC), isopropenyl chloroformate,57' 5 8 and bromotripyrrolydinophosphonium hexafluorophosphate (PyBroP) 5 9 were unsuccessful. A n analysis of the formation of depside bonds by Davies et al. confirmed the demands of this reaction.60 In their study, which also employed phenyllactic acid as the model alcohol, they concluded that the secondary hydroxyl group is particularly difficult to acylate. Furthermore, it has been noted that the size of the 44 .OfBu OH 12, EDC, 0.15 equiv DMAP, CH2CI2 81% .Offiu O \ ^ O 14 1) H 2, Pd/C, EtOAc 2) 8, EDC, HOBt, DMF/CH2CI2 16% .OR 15, R = fBu 5, R = H • HCOOH 85% Figure 2.11 Corrected synthesis of depsipeptide 5. activated carboxylic acid fragment can lead to a reduction in the coupling rate.61 Thus, by simplifying the carboxylic acid from a tripeptide to a single amino acid, we were able to overcome the problem. Figure 2.11, therefore, shows the modified route that led to the successful preparation of 5. Z-Ser(OriBu) was esterified with 12 using EDC and D M A P to yield 14. This tripeptide was hydrogenolyzed and coupled with 8 to provide the pentapeptide 15. The /-butyl groups were finally removed by acid treatment, and the desired depsipeptide 5 was purified to 85% by HPLC. Attempts to prepare the D-serine epimer of 5 were unfruitful. It was apparent that after hydrogenolysis of the corresponding tripeptide [D-Ser]14, condensation with 8 led to an inseparable mixture of isomers. A similar result was obtained when attempting to couple 8 with another tripeptide bearing an N-terminal residue with D-configuration ([D-ClAla]22; vide infra, Section 2.4.1). This finding suggests that there may be a difference in reactivity in forming this peptide bond; it is possible that significant steric interactions occur between the bulky leucine residue and the incoming peptide when the attacking residue possesses the D-configuration, thereby increasing the lifetime of the activated form of 8 (intermediate 16; Figure 2.12). Prolonged existence of this activated species may lead to cyclization to oxazolone 17, which by 45 O AcHN Figure 2.12 Racemization of activated amino acid residues during convergent peptide coupling. virtue of its increased a-proton acidity, may lead to base-catalyzed racemization, ultimately resulting in a peptide with an epimerized leucine. It is well-documented that bulky amino acids, particularly 7V-acyl residues (as opposed to carbamate-protected amino acids), are prone to racemization in this manner. 6 1' 6 2 2.3.2 Evaluation of Depsipeptide Analogue 5 Incubation of 5 with the enzyme in the presence of EDTA was monitored by reversed-phase HPLC (Figure 2.13). It was found that within one hour, most of the depsipeptide had disappeared, while two new, faster eluting peaks were apparent. Suspecting that these may be indicative of the carboxylic acid and alcohol that would result from ester hydrolysis, a sample of 5 was saponified in 1 M NaOH and then analyzed by HPLC; thus, by comparison to the peak retention times of the base-hydrolyzed sample, the species generated by the enzyme were identified as Ac-Gly-Leu-Ser-OH and H-PhLac-Ala-OH. To distinguish which peptide corresponded to each peak, they were also prepared individually by deprotection of their 46 a) \ Figure 2.13 Enzyme-catalyzed hydrolysis of depsipeptide 5. Reversed-phase HPLC profiles following incubation of 100 uM 5 with enzyme and 5 mM E D T A at 37 °C for a) 5 min, b) 2 h, and c) 4 h; x = Ac-Gly-Leu-Ser-OH, y = H-PhLac-Ala-OH, * = impurity (apparently inert) from synthesis of 5. immediate precursors, 11 and 12, using hydrogenolysis and acid hydrolysis, respectively. A control sample confirmed that the hydrolysis of the depsipeptide was in fact enzyme-catalyzed. Because it was still possible that a portion of the ester had acylated the enzyme despite the dominant hydrolytic activity, reaction mixtures were analyzed by ESI-MS at early and late stages. The spectra appeared unchanged from the native enzyme, so no evidence for covalent adduct formation could be obtained using the depsipeptide. In retrospect, these results are not overly surprising. The ester function is inherently more susceptible to hydrolysis due to increased electrophilicity over its amide counterpart. It is likely that EDTA, which is capable of inhibiting the metalloprotease activity on the more robust amide, is insufficient to block this activity on more reactive substrates. Given our inability to establish a link between the epimerase and metalloprotease activities, however, it is not possible to conclude that the protein responsible for epimerization is also responsible for hydrolysis of the depsipeptide. 47 2.4 Chloroalanine Analogue An alternative strategy is to use an analogue that possesses a leaving group, such as a halide, p to the carbonyl. Such a compound would not alter the carbonyl's reactivity and should not be subject to protease activity in the presence of EDTA. The strategy in using this analogue is outlined in Figure 2.14. Based on the fact that minor substitutions for the serine's hydroxyl group - for example, thiol (cysteine) and methoxy (O-methylserine) - still act as substrates for epimerization, the chloroalanine derivative 18 should bind within the epimerase's active site. Following acyl-enzyme formation (step 1), a-proton abstraction by a general base may cause |3-elimination of the adjacent chloride (step 2). Since the resultant dehydroalanine adduct is not an intermediate in the normal reaction pathway, it may halt any further catalytic steps; alternatively, this species may serve as a Michael acceptor for a neighbouring enzymatic residue (step 3), leading to covalent labelling of the protein at two positions. Figure 2.14 Strategy to trap the purported acyl-enzyme adduct using a 3-chloroalanine derivative. Following acylation, general-base-catalyzed P-elimination of HCI may occur, yielding a dehydroalanine species. 1,4-Addition by another active site nucleophile (Nu) may also occur to label the enzyme in two positions. 48 2.4.1 Synthesis of Chloroalanine Analogue 18 Again, a solution-phase synthesis was chosen to prepare the chloroalanine analogue, primarily to avoid potential substitution or elimination side reactions that may occur with the primary alkyl chloride under solid-phase synthesis conditions. A convergent route was chosen (Figure 2.15), analogous to that used in the preparation of depsipeptide 5. The C-terminal dipeptide 19 was prepared by condensation of Z-Phe and Ala-O^Bu. The key amino acid, 3-chloro-L-alanine (20), was formed from chlorination of L-serine methyl ester following the procedure of Fischer and Raske,6 3 followed by amino-protection as the benzyl carbamate (21)6 4. Dipeptide 19 was deprotected by hydrogenolysis and coupled to 21 to give tripeptide 22. 1) H 2, Pd/C, MeOH 2) 8, EDC, HOBt, DMF/CH2CI2 28% Figure 2.15 Synthesis of 3-chloroalanine analogue 18. 49 Hydrogenolysis, coupling to dipeptide 8 already in hand, and hydrolysis of the /-butyl ester (23) yielded the target pentapeptide 18. During workup and H P L C purification, it was discovered that 18 is sensitive to mild heating in solution. Hence, this compound was cautiously handled at or below ambient temperature whenever possible. As with the depsipeptide analogue (recall Section 2.3.1), attempts to synthesize the 3-chloro-D-alanine epimer of 18 were unsuccessful. Following hydrogenolysis of [D-ClAla]22, coupling with 8 also resulted in epimerization. 2.4.2 Enzymatic Reaction and Inhibition by Chloroalanine Analogue 18 Prior to testing with enzyme, it was necessary to examine the stability of 18 to incubations at 37 °C by HPLC. The chloroalanine analogue initially eluted as a single peak (Figure 2.16a), but after 23 h of incubation, it had become partially converted to a second species whose retention time is consistent with the L-serine-containing peptide 1 (Figure 2.16b). This observation suggests that chloroalanine is susceptible to non-enzymatic hydrolysis over extended incubations at an elevated temperature; however, since >80% remained intact, the minor side reaction was not expected to interfere significantly with enzymatic studies. Incubation of 18 with the enzyme in the presence of E D T A was monitored by reversed-phase HPLC (Figure 2.16c). It was found that, aside from the same degree of hydrolysis, the chloroalanine derivative was converted to another, closely eluting compound (24 in Figure 2.16c). L C - M S analysis of the reaction mixture confirmed the identity of the hydrolysis product as 1 (m/z 535 Da) and revealed that the mass of the new species (517 Da) is 36 units lower than the mass of 18 (553 Da). This decrease suggests that the enzyme recognizes 18 as an alternative 50 14 12 f o 10 a. 8 a) i—i—i 20 -<—i—I— 22 ->—i—I—r-24 -T—1 26 C) 18 A 24 Time (mfn> 22 Time (irtn) i—i 20 22 24 Time (rrin) 26 Figure 2.16 Non-enzymatic hydrolysis of chloroalanine analogue 18 and enzymatic elimination of HC1. Reversed-phase H P L C profiles following incubation of 100 uM 18 at 37 °C in the absence of enzyme at a) t = 0 and b) t = 23 h and c) in the presence of the epimerase at t = 23 h. Peaks have been assigned to corresponding compound numbers (1 = L-Ser epimer of substrate, 18 = chloroalanine analogue, 24 = dehydroalanine analogue). substrate, but instead of inverting its stereochemistry (since no peak attributable to the D-epimer was detected), it catalyzes the elimination of H Q to form a dehydroalanine residue (Figure 2.17). Attempts to determine the kinetics of this reaction were hampered by the facts that (i) the peaks from 18 and 24 overlap and (ii) the extinction coefficient of 24 at 220 nm (£220) is much larger (see Section 2.5.3); nevertheless, it proceeded noticeably faster than epimerization of 1 or 2. As with the depsipeptide study, it was still possible that catalysis had proceeded through an acylated intermediate. Alternatively, a portion of the enzyme may have become covalently modified by Michael addition to the dehydroalanine derivative (24). Thus, reaction mixtures Ac-Gly-Leu-N H :B-Enz N-Phe-Ala-OH CI" Ac-Gly-Leu-N epimerase H HB-Enz H N-Phe-Ala-OH 24 Figure 2.17 Enzyme-catalyzed elimination of HC1 from chloroalanine. 51 were analyzed by ESI-MS at early and late stages, but again, the spectra appeared unchanged from the native enzyme. Although 18 was found not to modify the enzyme covalently, its resemblance to substrate implicated it as a potential non-covalent inhibitor of epimerization, either directly or upon formation of 24. Ideally, a formal kinetic analysis would be performed, whereby a fixed concentration of inhibitor is incubated with the enzyme in the presence of varying concentrations of substrate. By repeating these assays with different concentrations of the inhibitor, a series of kinetic curves can be plotted to determine the dissociation constant, K\, which serves as a measure of the enzyme's affinity for the inhibitor. However, because this procedure requires dozens of data points and therefore also a large quantity of enzyme, it is not feasible with the peptide epimerase. An alternative analysis that does not require many points but still allows for quantification of inhibition is to determine the concentration of inhibitor at which epimerase activity is 50%, IC50. For a competitive inhibitor, the percent inhibition, i%, as a function of inhibitor concentration, [I], is given by: 6 5 100[I] % ( r c n ^ 2.1 K, + [I] Solving this equation at 50% inhibition shows how IC50 is related to K\. ICj. = 2.2 Substituting Equation 2.2 into Equation 2.1 gives: . 100[I] * % ~IC 5 o+[I] 2 3 A similar derivation for noncompetitive and uncompetitive inhibitors also yields Equation 2.2. Thus, a plot of/% vs. [I] gives a hyperbolic curve from which IC50 can be determined with the aid of a graph-fitting program. 52 Accordingly, varying concentrations (0.1 uM to 100 uM) of 18 were incubated with the epimerase in the presence of EDTA, followed by the addition of a fixed concentration (100 uM) of substrate 1 (Km = 8 m M ) 3 8 to initiate the reaction. Monitoring by HPLC revealed that, after 23 h of incubation, significant inhibition of epimerase activity had occurred; in fact there was almost no detectable product when the concentration of 18 was tenfold less than 1. A plot of i% vs. [18] (Figure 2.18) gave a hyperbolic curve from which IC50 was found to be 0.43 ± 0.04 uM. Hence, it can be concluded that the chloroalanine analogue or its reaction product 24 is a potent inhibitor of the epimerase, with a 50% inhibitory concentration that is 4 orders of magnitude lower than the Km for the corresponding substrate. Given that the vast majority of 18 would be converted into 24 during the incubation periods, the dehydroalanine peptide may be serving as the actual inhibitory species. 0 2 4 6 8 10 Concentration of 18 (uM) Figure 2.18 Plot of percent inhibition of epimerization versus concentration of chloroalanine analogue 18. The point at 100 uM is not shown for clarity. 53 2.5 Dehydroalanine Analogue 2.5.1 Inhibitors of Cofactor-independent Amino Acid Racemases and Epimerases Similar studies on amino acid racemases and epimerases have shown that introduction of a leaving group at the |3-carbon or on the nitrogen of the a-amino group leads to enzyme-catalyzed elimination to create an sp2-hybridized a-carbon. These and other molecules containing a trigonal planar centre at the a position have been found to be strong competitive inhibitors (recall Section 1.5.1). The first such inhibitors were designed to mimic the sp2 character at C-2 of the intermediate and transition state formed in the proline racemase reaction. Pyrrole-2-carboxylate (3) 3 2 and Al-pyrroline-2-carboxylate (25)6 6 (Figure 2.19a) are planar about C-2 and as such, were found to be potent competitive inhibitors of proline racemase. These early compounds served as models for developing similar inhibitors of other amino acid racemases and epimerases. The enzyme diaminopimelate (Dap) epimerase converts LL-Dap (Km = 260 uM) into meso-Dap (26), which, along with D-Ala and D-Glu, is an unusual amino acid that is incorporated into peptidoglycan (Figure 1.1). The enzyme, which is a cofactor-independent epimerase similar to proline racemase, was found to be inhibited by the halogenated substrate analogues 27 and 28, with I C 5 0 values of 4 - 25 u M (Figure 2.19b). 6 7 ' 6 8 In a manner analogous to the chloroalanine analogue 18, these compounds were converted to the planar enamine 29 by elimination of H X . Similarly, /V-hydroxy-Dap (30) was found to inhibit the epimerase (^j = 5.6 uM), likely by dehydration to form the planar imine 31. 6 9 Using the same strategy, 7V-hydroxy-54 NH2 NHOH NH2 NH 30 31 Figure 2.19 Substrate and intermediate analogues of a) proline racemase, b) Dap epimerase, and c) glutamate racemase that inhibit via planarity at C-2. D-glutamate (32) was found to be a good competitive inhibitor (K\ = 56 uM) of glutamate racemase (Km = 300 uM), presumably via imine 4. 4 0 In light of these inhibitor studies, the findings that the spider venom peptide epimerase converts chloroalanine to dehydroalanine and that incubation with 18 inhibits epimerization strongly suggest that the dehydroalanine derivative 24 is the active species that is responsible for inhibition. We therefore explored this possibility by synthesizing 24. 2.5.2 Synthesis of Dehydroalanine Analogue 24 A recent methodology for the site-specific, chemoselective incorporation of dehydroalanine into peptides was developed by van der Donk and coworkers. Previous 55 O O DEAD, PPh3 " Y ^ O H THF,-78°C B o c N » . . . ^ o N a H B ( 0 M e ) 3 Y OH \ 0 H 34 ^ S e P h 35, R = BOC | 1)JFA, CH 2 CI 2 33, R = Fmoc 2) Et3N, H20 MeCN, Fmoc-OSu o H O Phe-Ala-OH N A H2°2 Ji Ac-Glv-Leu"" Y ^Phe-A la -OH 9 3 % ' Ac-Gly-Leu ^ S e P h 36 24 Figure 2.20 Synthesis of dehydroalanine analogue 24. Abbreviations used: Boc = tert-butoxycarbonyl, D E A D = diethyl azodicarboxylate, Fmoc = 7V-(fluoren-9-ylmethoxycarbonyl), Fmoc-OSu = Fmoc-succinimide, Sec(Ph) = Se-phenylselenocysteine, Dha = dehydroalanine. methods had involved procedures such as activation and elimination of serine derivatives, Hoffmann elimination from 2,3-diaminopropionic acid residues, and oxidative elimination of S-alkyl or 5-aryl cysteines; however, these strategies are limited because they lack selectivity or are incompatible with many pre-existing amino acid residues in the peptide. Van der Donk avoided these problems by using the amino acid Af-(fluoren-9-ylmethoxycarbonyl)-Se-phenyl-L-selenocysteine (Fmoc-Sec(Ph), 33). This residue is incorporated into the desired position in a peptide via solid-phase peptide synthesis (SPPS) and is unmasked by oxidative elimination to reveal the dehydroalanine residue. Amino acid 35 was prepared from Boc-Ser by cyclization to the P-lactone 34,71 followed 70 by ring opening with phenylselenide anion (Figure 2.20), as previously described. Protecting group exchange provided 33, which was then incorporated into pentapeptide 36 in 56% yield using standard SPPS. Finally, oxidation by treatment with aqueous hydrogen peroxide gave 24. 56 2.5.3 Inhibition of Epimerase by Dehydroalanine Analogue 24 To be sure that the dehydroalanine residue did not undergo reaction with the epimerase, samples of 24 were incubated with the enzyme for 3 days, and aliquots were analyzed by reversed-phase HPLC and ESI-MS. No changes were observed, and the HPLC retention time confirmed our assignment of the peak that was formed by the enzymatic reaction of 18 (Figure 2.16c). Additionally, the £220 of 24 was found to be (1.14 ± 0.06) x 104 M " 1 cm"1, nearly threefold larger than that of 18 ((3.9 ± 0.4) x 103 NT1 cm"1) or 1 ((4.2 ± 0.3) x 103 M " 1 cm"1), a difference that can be attributed to the conjugated 71-system of dehydroalanine. As previously mentioned, it is this increased £220 that interfered with integration and therefore quantification of the enzymatic formation of 24 from 18. We then proceeded to test 24 as an inhibitor of epimerization. Again, varying concentrations (0.1 uM to 10 uM) of 24 were incubated with the epimerase in the presence of EDTA, followed by the addition of a fixed concentration (100 uM) of substrate 1. As / o 0 6-0 2 4 6 8 10 Concentration of 24 (uM) Figure 2.21 Plot of percent inhibition of epimerization versus concentration of dehydroalanine analogue 24. 57 anticipated, monitoring by HPLC revealed that, after 21 h of incubation, significant inhibition of epimerase activity had occurred. A plot of i% vs. [24] (Figure 2.21) gave a hyperbolic curve from which IC50 was found to be 0.47 ± 0.08 p.M. This value is in agreement with the IC50 determined for 18. Thus, it appears that our assumptions regarding the mode of inhibition by 18 were correct: in keeping with the substrate analogues of amino acid racemases and epimerases, the chloroalanine analogue 18 is enzymatically converted to the dehydroalanine species 24, which by virtue of its sp character at C-2, mimics the planar enolate intermediate in the epimerization reaction and results in substantial inhibition. The results of the studies involving 18 and 24 allow us to draw some mechanistic conclusions. Although it is not possible to exclude covalent catalysis, no evidence in support of such a mechanism has been obtained aside from the reported sequence homology, which itself is questionable (recall Section 2.2.2). Furthermore, the dehydroalanine analogue essentially eliminates the dehydration/rehydration mechanism (Figure 1.15a). Recall that this mechanism proceeds through 24 as an intermediate; thus, i f it were operative and assuming a water molecule was active-site accessible, 24 should have become hydrated to 1 and 2. Therefore, considering our inhibition findings, together with the previous studies by Heck et al., it is apparent that the epimerase employs a simple deprotonation/reprotonation mechanism. 3 \ OH 3 enolate intermediate 24 58 2.6 Phosphinate Analogue 2.6.1 Phosphinate Analogues as Intermediate and Transition State Mimics The dehydroalanine studies provided strong evidence for planar geometry at C-2 in the intermediate of the epimerase reaction. Further support for an enolate intermediate could be obtained from a peptide analogue bearing an oxyanion in a position equivalent to the enolic oxygen. The amide group cannot be simply replaced with an enolate or alkoxy group, however, since these species are unstable at physiological pH. To overcome this, the oxyanion can be introduced as a resonance-stabilized analogue such as a phosphinate (Ri-P02-R2)-Phosphinates have been used extensively in the design of peptide analogues for the purpose of forming very tightly bound inhibitors of several classes of proteases. They have been used largely in blocking the actions of metallopeptidases72' 7 3 and metalloproteases including angiotensin converting enzyme (ACE) , 7 4 " 7 6 endothelin converting enzyme (ECE), 7 6 " 7 8 and matrix metalloproteases.79 Phosphinates have also been valuable in studying aspartate proteases such as 80 81 82 pepsin ' and human immunodeficiency virus (HIV-1) protease and the serine proteases chymotrypsin83 and trypsin.8 4 Additionally, these compounds have been used to inhibit the ATP-dependent ligases y-glutamylcysteine synthetase (y-GCS) 8 5 and glutathionyl spermidine synthetase (GSPS) ; these enzymes differ from the proteases in that they catalyze the assembly of peptides, rather than their degradation. The potency of phosphinates to all of these enzymes is largely attributable to two closely related factors. The enzymes employ similar mechanisms whereby a nucleophile attacks a peptide carbonyl, forming a tetrahedral, anionic intermediate. The permanent negative charge and tetrahedral geometry of the phosphinates closely resemble 59 this intermediate and the transition state leading to its formation, thereby making these compounds among the most effective inhibitors known. The use of phosphorus analogues in examining racemase mechanisms is limited. Perhaps 87 the best example is (l-aminoethyl)-phosphonate (36), which differs from a phosphinate by an additional P-0 bond. It was found that this alanine mimic, following formation of an aldimine with PLP, was bound within the active site of alanine racemase in such a way that its release was extremely slow (tm = 25 days), thereby essentially causing inactivation.88 The efficacy of this inhibitor was later made clear by a crystal structure of the enzyme-phosphonate complex.29 The structure revealed hydrogen bonds from four enzyme residues and two water molecules to the phosphonate oxyanions, causing the enzyme to adopt an unusual conformation that hindered the molecule's release. We believed that phosphinate 37 could be a good inhibitor of the peptide epimerase in one of two possible modes. First, although our originally proposed mechanism invoking covalent catalysis seemed less favourable in view of our recent results, it could not be excluded. Consequently, as has been demonstrated with the protease inhibitors mentioned above, it is conceivable that the tetrahedral geometry of 37 could model a covalent intermediate. enolate intermediate 60 Alternatively, and perhaps more likely, inhibition could be achieved through mimicry of the enolate formed from a simple deprotonation/reprotonation mechanism by means of introducing a stabilized phosphinate anion. Recall that the dehydroalanine analogue 24 possesses an sp centre at C-2 but bears no charge; in this case, however, 37 lacks the sp centre but carries a negative charge. 2.6.2 Synthesis of Phosphinate Analogue 37 Ideally, the phosphinate should be designed to match the natural enolate as closely as possible, both in terms of amino acid sequence and stereochemistry. Major difficulties in synthesizing these types of compounds often arise, however, when attempting to incorporate certain side chain functional groups and when stereocontrol is desired. The most common route is to prepare the 1-aminophosphinic acid that corresponds to the residue on the N-terminal side of the phosphinate linkage and to add this to the 2-substituted acrylate that corresponds to the residue on the C-terminal side (Figure 2.22). Because this step is non-stereoselective, a mixture of diastereomers is necessarily obtained. This shortcoming is not easily avoidable, and a mixture must be carried forward in subsequent steps. It is often possible, though, to separate the isomers of the final, deprotected peptides using HPLC. Although the preparation of the racemic serine phosphinic acid analogue has been reported, the use of the alanine analogue 38 was preferred to avoid the need for an additional O R 2 O R 2 Figure 2.22 Phosphinate formation from 1-aminophosphinic acid and acrylate. PG = protecting group. 61 protecting group (recall Ac-Gly-Leu-Ala-Phe-Ala-OH can serve as an alternative substrate; see p. 22). Additionally, chiral resolution of 38 was deemed unnecessary since either isomer should be accepted within the enzyme's active site. The pentapeptide 37 was synthesized as a diastereomeric mixture as outlined in Figure 2.23. As previously described in the literature, phosphinic acid 38 was made nucleophilic through the action of 7V,0-bis(trimethylsilyl)acetamide (BSA) and after 1,4-addition to acrylate Z - N ' H 38 Z - N ' H OH 4 - " O ,H C0 2 Et Ph 39 CO,Et Ag 2 0 , CHCI 3, A 2) NaOH, MeOH 3) H 3 0 + 81% 42 Ph 1) H 2 , Pd/C 2) 8, EDC, HOBt, DMF/CH 2 CI 2 70% EDC,HOBt, H DIEA 75% O 41 Nf "C0 2 R 2 H 43, = Ad, R2 = Me —i 37, R L R2 = H 1) NaOH, MeOH 2) TFA/CH 2 CI 2 59% Figure 2.23 Synthesis of phosphinate 37. Abbreviations: B S A = N,0-bis(trimethylsilyl)acetamide; A d = 1-adamantyl; DIEA = diisopropylefhylamine. C 0 2 H Ph 62 39 yielded the pseudodipeptide 40.90 Initial attempts were made to deprotect the ethyl ester and couple to Ala-OtBu without protecting the phosphinic acid moiety. Despite the precedence of similar routes,90 the condensation step proved troublesome. Consequently, this problem was alleviated by protection of the acid as the 1-adamantyl ester 4191' 9 2 prior to coupling with Ala-OMe to give 42. Note, however, that this phosphinate ester introduced two disadvantages: not only did it require extra protection and deprotection steps, the phosphorus was made stereogenic and therefore added complexity to analysis of spectral data and purification of intermediates. Nevertheless, hydrogenolysis of 42 and coupling to the versatile dipeptide 8 formed pentapeptide 43, which was subsequently deprotected to give the target 37. Composed of up to four stereoisomers (i.e., either L or D at Alai and at Phe), 37 was purified by reversed-phase HPLC. The two major fractions that eluted, 37c and 37d, each contained two inseparable diastereomers that gave mass and N M R spectra consistent with the product. Although the absolute stereochemistry of the purified isomers could not be determined without great difficulty, it was assumed that two of these, either together as one fraction or separate, possessed the desired L-configuration at the phenylalanine residue (recall that the configuration at the central alanine residue is likely unimportant for enzyme binding). Thus, it was reasoned that at least one of 37c and 37d would contain an isomer that could be bound in the enzyme's active site and possibly exhibit inhibition of activity. 2.6.3 Evaluation of Phosphinate Analogue 37 Analogue 37 was tested as an inhibitor of the epimerase reaction as before. Whether as the crude mixture prior to HPLC purification or as the individual fractions 37c and 37d, however, no inhibition was detected at concentrations up to 100 uM (equimolar with substrate 63 1). Because phosphinates are capable of inhibiting many metalloproteases, it was also tested as an inhibitor of the enzyme's metalloprotease activity by incubation in the absence of EDTA. Again, no decrease in the rate of substrate cleavage was observed. Although drawing solid conclusions from a negative result is always questionable, one may speculate that the failure of 37 to inhibit either activity indicates that it is bound poorly, i f at all, within the active site(s). It is possible that the tetrahedral geometry of the phosphorus is actually a poor mimic of the intermediate and transition state, thereby further weakening the covalent catalysis mechanism. Despite the presence of the negative charge on the phosphinate oxygen, it is apparent that the enzyme's geometric constraints dominate. Likely, C - l must remain planar, a condition that is satisfied by the enolate intermediate of the deprotonation/reprotonation mechanism. 2.7 Steps toward Further Understanding of the Peptide Epimerase's Structure and Mechanism The analogues described in this chapter have offered substantial insight into the structure of the reaction intermediate and transition state, and therefore, have allowed us to adjust our view of the reaction mechanism, ultimately favouring the simple deprotonation/reprotonation mechanism originally proposed by Heck et al. Nevertheless, some details regarding the mechanism and structure of the epimerase could be addressed to clarify or reinforce our understanding of this unique enzymatic process. This section, therefore, outlines examples of such approaches that could be pursued in future work. 64 2.7.1 Active Site Labelling Further mechanistic insight could be garnered by identifying enzymatic residues that play a role in catalysis. Modern techniques such as protein crystallography and cloning and mutagenesis are routinely used to obtain a three-dimensional view of the enzyme-substrate complex and to determine the residues that are important in binding and catalysis. As previously mentioned in Section 2.2.1, though, these options are currently unavailable in studying the spider venom epimerase. An alternative option, however, is the classical method of chemical modification of active site residues. An appropriately chosen compound may target the active site and react covalently with a particular enzymatic residue, which can be subsequently identified by proteolysis, peptide sequencing and mass spectrometry. Successful covalent tagging of the epimerase could also serve as a means of removing the uncertainty linking the reported sequence of the enzyme with its function (recall Section 2.2.2). After confirming that the protein had been altered, it could then be assayed for epimerase activity. If inhibition of epimerization correlated with the degree of covalent modification, then it could be concluded that the published sequence indeed corresponds to the epimerase. Since the deprotonation/reprotonation mechanism is strongly favoured, chemicals that have been used to label amino acid racemase residues may also be effective with the peptide epimerase. Based on the knowledge that proline and glutamate racemases utilize cysteines as general acid/base residues, Heck et al. tested this hypothesis by treating the enzyme with N-ethylmaleimide, a thiol scavenger agent.38 The chemical reduced epimerization but failed to inactivate the enzyme, so it is unlikely that a cysteine plays a critical role in catalysis. Although a chemical used to label functionally similar enzymes was unsuccessful when applied to the peptide epimerase, perhaps a chemical used to label structurally related enzymes would work. The purported homology of the epimerase to serine proteases implicated specific 65 residues as possibly serving a catalytic role. As previously mentioned, Serl94, His46, and Asp96 were suggested as operating as a triad in a covalent mechanism. Interestingly, Heck et al. had reported that the commonly used serine protease inhibitor PMSF was also capable of inhibiting epimerase activity, though no attempts to determine the nature of the inhibition were reported. When we repeated this and determined the mass of the enzyme by ESI-MS, however, no change was found. Apparently the observed inhibition was not due to covalent modification of the sequenced protein. We decided to test a known covalent inhibitor of thrombin, as this enzyme is one of the most homologous to the epimerase. 2-Ethoxy-4H-3,l-benzoxazin-4-one (44) had been found to inhibit thrombin via ring-opening as shown in Figure 2.24. This compound has the added advantage of producing a chromophoric adduct, which could assist in identifying the labelled fragment upon proteolytic cleavage. Following incubation of 44 with the epimerase at 0.1 m M for 15 min, ESI-MS analysis indicated that partial labelling had occurred (Figure 2.25). We then desired to conduct an epimerase assay to determine i f the modification was active-site directed. Because 44 was found to co-elute with 1 in the H P L C chromatogram, it was first necessary to remove excess 44 by ultrafiltration. Unexpectedly, however, when the resulting enzyme solution was analyzed by ESI-MS, the covalent adduct had disappeared. Thus, the compound appears to act reversibly, and with no way of measuring epimerase activity in the presence of excess 44, we were forced to abandon further studies on this compound. Figure 2.24 Covalent inhibition by benzoxazinone 44. The labelled product possesses a chromophore with maximal absorbance at 319 nm.44 66 22000-20000-18000-16000-S 14000 2 - 10000 8000-6000-4000 2000-1 29476.0 Free enzyme Enzyme + adduct 25783.0 32625.0 34606.0 26000 32000 34000 Figure 2.25 Electrospray mass spectrum of the epimerase following incubation with 44. Calculated masses of free enzyme and covalently modified enzyme are 29 471 and 29 662 Da, respectively. The failure to demonstrate modification of the active site of the epimerase led us to synthesize another potential labelling compound that bears some resemblance to the substrate. Influenced by the use of N-haloacetyl amino acids and peptides in labelling several proteases94'95 we prepared the JV-chloroacetyl (45) and ./V-bromoacetyl (46) derivatives of the C-terminal peptide, Phe-Ala. It was hoped that the proximity of the halogen to a catalytic residue would result in displacement of the halide and tagging of the enzyme (Figure 2.26). The attachment of the haloacetyl groups to the dipeptide was accomplished by treatment C 0 2 H C 0 2 H 46, X = Br Figure 2.26 Strategy to label the active site using N-haloacetyl peptides. 67 with chloroacetyl chloride or bromoacetyl bromide according to the literature.95 Unfortunately, incubation at concentrations as high as 7 m M 45 or 1 m M 46 failed to alter the enzyme as determined by ESI-MS. Chemical labelling of the peptide epimerase has proven to be a difficult task. If a successful chemical is to be found, it may not be the result of rational design. Rather, "brute force" methods such as using chemical libraries may be needed. 2.7.2 Ketone Analogue To solidify the deprotonation/reprotonation mechanism and effectively eliminate the covalent catalysis mechanism, it would be desirable to design a substrate analogue that could not be cleaved to yield an acylated enzyme but would still behave as a substrate for the epimerase activity. Perhaps the best model that would strengthen the putative scissile amide bond linking the serine and phenylalanine residues is a ketone, such as compound 47. Under the covalent mechanism, the tetrahedral species that would form from attack by the enzymatic serine could not collapse with C-C bond cleavage; under the deprotonation/reprotonation mechanism, however, the ketone is expected to be readily epimerized, potentially faster than its amide counterpart due to the lower pKa of the a-proton. The stereoselective synthesis of ketomethylene isosteres has proven to be difficult, as there are few simple ways to form a ketone linkage while maintaining the desired stereochemistry in the amino acid analogues that are to be joined. Two similar methods that 68 a) C0 2 Et 1) H 2 , 50 psi, Pd/BaS0 4 2) HOAc, PhCH 3, A Boc-N H 53 (47%) + cis lactone (3%) b 2 ) J U 3) H 2 , Pd/C 52 (78%) + C-4 epimer(17%) Q " A X Pfr 54 Me 1) LiN(SiMe3)2 2) BrCH2OBn 3) Ca(BH4)2 4) PPh 3 , NBS "Ph 5)Mg,Et 20 BrMg' Boc-Val-N H C 0 2 H 1) HCI, dioxane 2) (Boc-Val)20 3) Ac 2 0, DMAP B o c " N 4) H 2 , Pd/C 5) PDC, DMF 1) 50 2) H 3 0 + OBn Figure 2.27 Strategies for synthesizing ketomethylene peptides, a) Kleinman's method via lithium ethyl propiolate (48) involves preferential alkylation of lactone 52 to trans lactone 53. b) Rich's method via Grignard 49. Abbreviations used: NBS = 7V-bromosuccinimide, PDC = pyridinium dichromate. have been used to accomplish this are the organolithium (Figure 2.27a) and Grignard (Figure 2.27b) additions reported by Kleinman et al96 and Rich et al.,91 respectively. In either case, the C - C linkage is generated by nucleophilic addition of a carbanionic amino acid analogue (48 or 49) to the aldehyde form of the other amino acid (50). The resulting alcohol is protected in an acylated form to allow for subsequent chemical manipulation and peptide chain extension (exception: 51 is not acylated during N-terminal extension to avoid rapid O—»N transacetylation97). Deprotection with base (e.g. K2CO3) and oxidation (e.g. pyridinium dichromate, PDC) can then be used to unveil the ketone functionality (not shown in Figure 2.27). 69 The former strategy is synthetically simpler in that there are fewer steps and all reagents are commercially available on a large scale (though the organolithium reagent 48 must be generated in situ from the alkyne); however, it suffers from the need to introduce the side chain of the residue on the methylene side of the ketone, a process that relies upon fortuitous separation from the minor diastereomers of lactones 52 and 53. The latter method is potentially hampered by the requirement for the enantioselective preparation of Grignard reagent 49, a process that is lengthy and difficult to produce on a large scale due to the cost of the chiral auxiliary used to prepare 54. Taken in another light, however, these additional steps make this method attractive since the desired stereochemistry is assured at an early stage and no special separation steps are needed. In synthesizing the substrate analogue 47, the dipeptide synthon 55, therefore, could be no prepared by one of these methods using an appropriately protected serinal (56) (Figure 2.28). The phenylalanine residue could be introduced via Rich's Grignard 49 or by stereoselective alkylation with benzyl halide following lactonization analogous to the preparation of 53." The only additional consideration with the use of serine is that the side chain alcohol must employ a protecting group strategy different from that of the backbone alcohol to allow regioselective oxidation. The tert-butyldiphenylsilyl (TBDPS) group is compatible with the mild acid and base conditions used to remove the Boc and acetyl groups and can be displaced with fluoride in the final step. In testing 47 with the enzyme, one must be cautious in interpreting the results; i f epimerization is observed, a direct deprotonation/reprotonation mechanism is clearly supported, but this does not necessarily mean that the epimerase could not utilize a covalent mechanism in 47 Figure 2.28 Possible preparation of ketone analogue 47. 70 its normal reaction. Because the ketone is already more acidic than the corresponding acyl-enzyme adduct, it is conceivable that the epimerase could employ an alternative mechanism. Nevertheless, the result would add to the already substantially favoured deprotonation/reprotonation mechanism, which does not require any caveats unlike the covalent mechanism. 2.7.3 Affinity Chromatography Because of the failure to separate the epimerase and metalloprotease activities by the orthogonal techniques described in Section 2.2.2, another method is desired. Based on the strong binding exhibited by the dehydroalanine analogue 24, it may be possible to use affinity chromatography. The enzyme could be applied to a column composed of the dehydroalanine peptide attached to a solid support, thereby causing the epimerase to bind specifically and ideally allowing the protease to pass through. A convenient support for attaching peptides is Affi-Gel 15 (Bio-Rad). This resin is composed of an agarose backbone with a cationic linker that terminates in a succinimide ester, which serves as an activator for peptide coupling (Figure 2.29). When a peptide is introduced, its N-terminus forms an amide bond with the resin. This reaction is enhanced with acidic peptides, such as 57, due to attractive electrostatic interactions with the linker. As done previously for its pentapeptide counterpart 24, heptapeptide 57 could be synthesized from the SPPS-derived phenylselenocysteine-containing peptide 58 by oxidation with H2O2. This longer peptide, which contains the next two residues in the sequence of the natural substrate, distances the critical residues from the agarose core in order to increase the opportunity for interaction with the epimerase's active site. Hopefully, the acrylate moiety of 57 71 N H Affi-Gel 15 C H 3 N H linker H-Met-Glu-Gly-Leu-Sec(Ph)-Phe-Ala-OH 58 H 2 0 2 H-Met-Glu-Gly-Leu-Dha-Phe-Ala-OH 57 Affi-Gel 15, DMSO K A g g r ^ ^ p - O ^ ^ Lin k e r / W V / H H C-N-Met-Glu-Gly-Leu-N I I O O I I -C-Phe-Ala-OH Figure 2.29 Preparation of affinity chromatography resin. would remain inert to any side-reactions, allowing the dehydroalanine-peptide to be attached to the resin by displacing the succinimide group. Either crude venom or singly purified enzyme could be applied to the affinity column followed by thorough washing to remove any impurities including the protease. Elution of pure enzyme could then be achieved by introducing a competitively high concentration of 24 into the column buffer. Finally, dialysis would allow removal of the inhibitor to give pure epimerase, which would be assayed for both activities and analyzed by electrophoresis and ESI-MS. 2.8 Conclusions The spider venom peptide epimerase proved to be a challenging but rewarding project. Based on the intriguing report on the enzyme's sequence homology to serine proteases, we were inspired to formulate a mechanism that employs covalent catalysis using the protease machinery 72 to enhance removal of the a-proton. Unfortunately, many of our approaches to assess this mechanism and other protein aspects met with inconclusive results. An attempt was made to purify the epimerase activity from the metalloprotease activity using cation-exchange chromatography. The failure to separate the activities means that the possibility that the epimerase is actually a small component in a sample of the sequenced protease remains real. In light of this finding, our observation that the enzyme catalyzes the hydrolysis of the depsipeptide analogue 5 could not be attributed to either activity. On the other hand, the synthesis and testing of the chloroalanine (18) and dehydroalanine (24) analogues enabled us to draw mechanistic conclusions, as we have recently published.1 0 0 The enzyme was found to catalyze the elimination of HC1 from 18 to form 24. This species caused potent inhibition of epimerization, exhibiting an IC50 of 0.5 uM, several orders of magnitude lower than the Km values of the substrates. By comparison to the inhibitors of several amino acid racemases and epimerases, the sp2 hybridization of C-2 was believed to mimic the planar enolate intermediate involved in the simple deprotonation/reprotonation mechanism. We were therefore compelled to favour this mechanism over our proposed covalent catalysis mechanism. The ability of the spider venom peptide epimerase to liberate the weakly acidic a-proton using such a mechanism seems perplexing. However, after our research on the epimerase was concluded, J. P. Richard and coworkers published the first report to quantify the acidity of such a proton.1 0 1 They reported that the pKa of the internal residue of the tripeptide Gly-Gly-Gly is only 25.9, which is 6 units lower than what we had estimated. Consequently, the penalty for introducing an a-proton into a peptide versus a fully protonated amino acid is only 4 units. Thus, it is reasonable that a deprotonation/reprotonation mechanism could be employed; the very low specific activity of the epimerase may simply be a reflection of this four-unit increase in pKa. 73 2.9 Experimental Procedures 2.9.1 General Synthetic Methods 2.9.1.1 Materials A l l solvents, reagents, and buffers were purchased from Sigma-Aldrich or Fisher Scientific except where noted. Alanine-loaded Wang resin, amino acid starting materials used in solution phase synthesis, and diisopropylethylamine (DIEA) were purchased from Advanced ChemTech. Fmoc-Phe, Fmoc-Leu, and 0-(benzotriazol-l-yl)-iV,A^,A^',jV-tetramethyluronium hexafluorophosphate (HBTU) (Advanced ChemTech) and piperidine (Sigma-Aldrich) were graciously provided by Prof. J. Sherman. L-(-)-3-Phenyllactic acid (PhLac) was purchased from Lancaster. A^-(Fluoren-9-ylmethoxycarbonyl)-.S'e-phenyl-L-selenocysteine (33) was prepared from Boc-Ser according to Okeley et al.70 Glycyl-L-leucine and L-phenylalanyl-L-alanine were purchased from Sigma-Aldrich. The pentapeptide substrates 1 and 2 were purchased from the Nucleic Acids Protein Services Unit at the University of British Columbia (UBC). Triethylamine (Et 3N), dichloromethane (CH2CI2), and ethyl acetate (EtOAc) were distilled under N2 over CaH2. Methanol (MeOH) and ethanol (EtOH) were distilled under N2 over magnesium methoxide or magnesium ethoxide, respectively. Tetrahydrofuran (THF) was distilled under N2 from sodium and benzophenone. Anhydrous A^N-dimethylformamide (DMF) was dried over 4 A molecular sieves and stored under argon. 2.9.1.2 Routine Methods Reaction progress was routinely monitored by thin-layer chromatography (TLC) on aluminum-backed sheets of silica gel 60, 0.2 mm thickness, with fluorescent indicator U V 2 5 4 (Macherey-Nagel, Germany). Visualization of compounds was achieved by spraying the plate with a solution of ammonium molybdate tetrahydrate ((T^FLOeMovC^^FbO, 24.0 g), ammonium 74 cerium(IV) nitrate ((NH 4) 2Ce(N0 3)6, 0.5 g), H 2 S 0 4 (28 mL), and water (500 mL), followed by charring with a heat gun. UV-absorbing compounds were also detected prior to treatment with molybdate spray by 254 nm U V light. Compounds containing free amine groups were detected by dipping the plates in ninhydrin followed by charring with a heat gun. Flash chromatography was performed on 230-400 mesh silica gel from SiliCycle (Quebec). Deprotected peptides were dissolved in HPLC-grade MeOH and purified by preparative reversed-phase HPLC on a Waters DeltaPak CI 8 column (19 x 300 mm, 15-um particle size, 100-A pore) using a Waters 600E system with detection at 220 nm using a Waters 484 tunable absorbance detector. Elution was achieved in mixtures of 0.1% TFA in water and 0.05% TFA in C H 3 C N (both solvents were sparged with helium prior to use) using linear gradients optimized for each peptide with a flow rate of 14 mL/min. Analytical reversed-phase H P L C of purified peptides and enzyme assays was performed on a Waters DeltaPak CI8 column (3.9 x 150 mm) using either the same system as for preparative HPLC or a Waters 625 L C system with a Waters 486 tunable absorbance detector. Elution was achieved as described for preparative HPLC using a flow rate of 1.0 mL/min. Retention times were corrected using compound 1 as an internal standard. Peak integration was determined using Millennium 2010 software (Millipore). Solid-phase peptide synthesis was performed on an Applied Biosystems 431A Peptide Synthesizer using Fmoc-protected amino acids, H B T U , and HOBt. ' H nuclear magnetic resonance (NMR) spectra were obtained on a Bruker AC-200E, AV300, or AV400 spectrometer at a field strength of 200, 300, or 400 MHz, respectively. Assignments were based on COSY spectra and literature precedent from related compounds. 3 1 P - N M R spectra were recorded either on a Bruker AV300 spectrometer at 121.5 MHz or on a Bruker AV400 at 162 MHz, and assignments were based on literature precedent. Mass spectrometry was performed by the Mass Spectrometry Centre at U B C by liquid secondary ionization mass spectrometry (LSI-MS), desorption chemical ionization (DCI), or electrospray 75 ionization (ESI-MS). Elemental analyses were performed by Mr. Peter Borda in the Microanalytical Laboratory at U B C and by Canadian Microanalytical Service, Ltd. High-resolution N M R spectra of final, deprotected peptides for which elemental analysis was not achievable have been included in the Appendix. 2.9.1.3 Solution-phase Peptide Coupling Unless noted otherwise, amino acids and peptides were coupled by the following procedure. A solution of the /V-protected amino acid (1.0 equiv.), 3-(3-dimethylaminopropyl)-l-ethylcarbodiimide (EDC, 1.05 equiv.), and 1-hydroxybenzotriazole hydrate (HOBt, 1.1 equiv.) in 4 - 6 mL mmol"1 dry CH2CI2 was stirred 10 min under argon, a) In the case of coupling to the hydrochloride salt of a carboxy-protected amino acid, the amino acid (1.0 equiv.) and E t 3 N or DIEA (2.0 equiv.) were added, and the solution was stirred at it for 18 - 24 h. b) In the case of coupling to a neutral carboxy-protected peptide, a solution of the peptide (1.0 equiv.) in 1 mL mmol"1 dry CH2O2 and E t 3 N or DIEA (1.0 equiv.) were added, and the mixture was stirred at rt for 18 - 48 h. The mixture was diluted threefold or fourfold with CH2CI2 and washed with equal volumes of water, 5% KHSO4, sat. N a H C 0 3 , and brine. Occasionally during extraction, it was necessary to remove precipitated triethylamine hydrochloride salt by filtration. The organic layer was dried over MgS04, filtered, and rotary evaporated. The crude peptide was purified by recrystallization or flash chromatography. 2.9.1.4 Hydrogenolysis A solution of the JV-(benzyloxycarbonyl)-protected peptide in 3 mL mmol"1 distilled MeOH containing 5% Pd/C (10% w/w peptide) was stirred under 1 atm of hydrogen at rt for 20 h. The mixture was filtered through Celite, and the filtrate was rotary evaporated. The resulting amines were used without further purification. 76 2.9.1.5 Hydrolysis of /-Butyl Esters and Ethers A solution of the protected peptide in ~5 mL mmof1 88% formic acid was stirred at rt for 48 h. After removal of solvent by rotary evaporation under reduced pressure at < 35 °C, the residue was dissolved in HPLC-grade MeOH and evaporated to dryness. The resulting colourless foam was dissolved in 15 -20 mL mmof1 HPLC-grade MeOH and purified in ~2 mL portions by preparative reversed-phase HPLC, followed by concentration and lyophilization. 2.9.2 Synthesis of Depsipeptide Analogue 5 2.9.2.1 L-3-Phenyllactyl-L-alanine /-butyl ester (12) PhLac (2.00 g, 12.0 mmol) and Ala-O/BuHCl (2.69 g, 14.8 mmol) were coupled as described in Section 2.9.1.3. Recrystallization from MeOH gave 2.98 g (85%) of 12 as white needles: *H N M R (200 MHz, CDC13) 5 7.25 (s, 5H, C 6 H 5 PhLac), 7.01 (d, 1H, JHNa = 7.1 Hz, N H Ala), 4.44 (quintet, 1H, JHNa = 7.1 Hz, a -CH Ala), 4.31 (dd, 1H, JaPa = 3.9 Hz, Japb = 8.1 Hz, a -CH PhLac), 3.22 (dd, 1H, JaPa = 3.9 Hz, Jpp= 13.9 Hz, p - C H 2 a PhLac), 2.90 (dd, 1H, Japb = 8.1 Hz, Jpp= 13.9 Hz, p - C H 2 b PhLac), 2.50 (bs, 1H, OH), 1.45 (s, 9H, C(CH 3 ) 3 ), 1.31 (d, 3H, Jap= 7.1 Hz, C H 3 Ala); +DCI-MS (NH 3) m/z 294 (M + H + , 22%), 238 (M + H + - H 2 C=C(CH 3 ) 2 , 100%); Anal. Calcd for C i 6 H 2 3 N 0 4 : C, 65.51; H, 7.90; N , 4.77. Found: C, 65.40; H , 7.91; N , 4.99. 2.9.2.2 A^-(Benzyloxycarbonyl)-0-(/-butyl)-L-seryl-L-3-phenyllactyl-L-alanine /-butyl ester (14) Z-Ser(OriBu) (1.76 g, 5.97 mmol) and 12 (1.75 g, 5.97 mmol) were coupled as described in Section 2.9.1.3 except that HOBt was omitted and D M A P (73 mg, 0.60 mmol) was added. Flash chromatography (2:1 pet. ether/EtOAc) gave 2.76 g (81%) of 14 as a colourless oil: ' H 77 N M R (300 MHz, CDCI3) 5 7.38-7.27 (m, 5H, C 6 H 5 Z), 7.26-7.08 (m, 5H, C 6 H 5 Phe), 6.98 (d, 1H, JHNa = 7.3 Hz, N H Ala), 5.60 (d, 1H, JHNa = 7.8 Hz, N H Ser), 5.50 (dd, 1H, JaPa = 4.7 Hz, Jap = 5.7 Hz, a -CH PhLac), 5.09 (d, 1H, Jgem = 12.2 Hz, C H 2 a Z), 5.01 (d, 1H, Jgem = 12.2 Hz, C H 2 b Z), 4.42-4.28 (m, 2H, a -CH Ala, a -CH Ser), 3.81 (dd, 1H, JaPa = 3.2 Hz, Jpp = 9.2 Hz, p-C H 2 a Ser), 3.55 (dd, 1H, Japb = 3.6 Hz, Jpp = 9.2 Hz, p - C H 2 b Ser), 3.19 (dd, 1H, JaPa = 4.6 Hz, Jpp= 14.3 Hz, p - C H 2 a PhLac), 3.12 (dd, 1H, Japb = 5.8 Hz, Jpp = 14.3 Hz, p - C H 2 b PhLac), 1.41 (s, 9H, C0 2 C(CH 3 ) 3 ) , 1.18 (d, 3H, Jap=l.\ Hz, C H 3 Ala), 1.09 (s, 9H, Ser-OC(CH 3) 3); +LSI-MS (3-nitrobenzyl alcohol matrix) m/z 571 (M + H + , 16%), 515 (M + H + - H 2 C=C(CH 3 ) 2 , 19%), 459 (M + H + - 2 x H 2 C=C(CH 3 ) 2 , 79%), 415 (M + H + - 2 x H 2 C=C(CH 3 ) 2 - C 0 2 , 75%); Anal. Calcd for C 3 i H 4 2 N 2 0 8 : C, 65.24; H , 7.42; N , 4.91. Found: C, 65.14; H, 7.52; N , 5.23. 2.9.2.3 A^Benzyloxycarbonyl)-0-(/-butyl)-D-seryl-L-3-phenyllactyl-L-alanine /-butyl ester ([D-Ser]14) The title compound was prepared in 85% yield (2.47 g) as described above for 14 using Z-DSer(OrBu) (1.50 g, 5.08 mmol): *H N M R (200 MHz, CDC13) 8 7.38-7.04 (m, 10H, C 6 H 5 Z, C 6 H 5 Phe), 6.77 (d, 1H, J m a = 7.3 Hz, N H Ala), 5.58 (d, 1H, JHNa = 7.8 Hz, N H Ser), 5.46 (dd, 1H, JaPa = 4.7 Hz, Japb = 5.7 Hz, a -CH PhLac), 5.13-4.98 (m, 2H, C H 2 Z), 4.49-4.20 (m, 2H, a-C H Ala, a -CH Ser), 3.62 (dd, 1H, Ja/h = 3.2 Hz, Jpp=9.2 Hz, p - C H 2 a Ser), 3.50 (dd, 1H, Japb = 3.6 Hz, Jpp=9.2 Hz, p - C H 2 b Ser), 3.21-2.90 (m, 2H, p -CH 2 PhLac), 1.38 (s, 9H, C0 2 C(CH 3 ) 3 ) , 1.18 (d, 3H, JaP = 7.1 Hz, C H 3 Ala), 1.09 (s, 9H, Ser-OC(CH 3) 3); +LSI-MS (3-nitrobenzyl alcohol matrix) m/z 571 ( M + H + , 17%), 515 ( M + H + - H 2 C=C(CH 3 ) 2 , 19%), 459 (M + H + - 2 x H 2 C=C(CH 3 ) 2 , 57%), 415 (M + H + - 2 x H 2 C=C(CH 3 ) 2 - C 0 2 , 60%). 78 2.9.2.4 N-Acetylglycyl-L-leucine (8) To a stirred solution of glycyl-L-leucine (5.42 g, 28.8 mmol) in 20 mL 0.1 M AcOH was added Et 3 N to pH 6. Acetic anhydride (15.9 mL, 144 mmol) was added in small portions while pH 6 was maintained by addition of Et 3 N. Amberlite® IR-120 resin (H + form) was added to pH < 2, then filtered off. The filtrate was rotary evaporated under reduced pressure to give 6.19 g (93%) of 8 as a colourless glass: ; H N M R (200 MHz, D z O) 8 4.37 (dd, 1H, Ja/h = 5.7 Hz, Japb = 8.9 Hz, a -CH Leu), 3.88 (s, 2H, a - C H 2 Gly), 2.00 (s, 3H, C H 3 Ac), 1.73-1.50 (m, 3H, P-CH 2 , y-C H Leu), 0.89 (d, 3H, Jy5, = 6.1 Hz, 8 r C H 3 Leu), 0.84 (d, 3H, Jy52 = 6.1 Hz, 8 2 -CH 3 Leu); +DCI-MS (NH 3) m/z 248 (M + N H 4 + , 35%), 231 (M + H + , 100%); Anal. Calcd for CioH 1 8 N 2 04: C, 52.16; H, 7.88; N , 12.17. Found: C, 51.99; H , 7.81; N , 12.01. 2.9.2.5 A^-Acetylglycyl-L-leucyl-0-(/-butyl)-L-seryl-L-3-phenyllactyI-L-alanine /-butyl ester (15) Tripeptide 14 (3.41 g, 5.98 mmol) was hydrogenolyzed in distilled EtOAc to give 2.22 g (85%) of the amine as a pale yellow gum. Due to its insolubility in CH2CI2, dipeptide 8 (1.35 g, 5.85 mmol) was dissolved in D M F and then coupled to the amine (2.55 g, 5.85 mmol) as usual. Flash chromatography (washed with 1:1 pet. ether/EtOAc and eluted with 5:95 MeOH/EtOAc) gave 0.59 g (16%) of 15 as a pale yellow foam: *H N M R (400 MHz, CDC13) 8 7.28-7.15 (m, 5H, C 6 H 5 PhLac), 6.88 (d, 1H, JHNa = 1.4 Hz, N H Ser), 6.70 (d, 1H, JHNa = 7.1 Hz, N H Leu), 6.57 (d, 1H, Jma = 8.2 Hz, N H Ala), 6.26 (t, 1H, JHNa = 4.9 Hz, N H Gly), 5.49 (dd, 1H, Jafia = 4.9 Hz, Japb = 6.0 Hz, a -CH PhLac), 4.54 (dt, 1H, Japa = Japb = 3.6 Hz, JHNa = 7.2 Hz, a -CH Ser), 4.48 (q, 1H, JHNa = Japa = Japb = 7.7 Hz, a -CH Leu), 4.40 (quintet, 1H, J m a = Jap = 7.2 Hz, a -CH Ala), 3.93 (d, 2H, JHNa = 5.1 Hz, a - C H 2 Gly), 3.82 (dd, 1H, JaPa = 3.5 Hz, Jpp= 9.3 Hz, P-CH 2 a Ser), 3.55 (dd, 1H, Japb = 3.7 Hz, Jpp=9A Hz, p - C H 2 b Ser), 3.22 (dd, 1H, JaPa = 4.7 Hz, Jpp = 14.4 Hz, p - C H 2 a PhLac), 3.17 (dd, 1H, Japb = 6.1 Hz, Jpp= 14.4 Hz, p - C H 2 b PhLac), 2.02 (s, 3H, 79 C H 3 Ac), 1.71-1.57 (m, 3H, y-CH, (3-CH2 Leu), 1.42 (s, 9H, C0 2 C(CH 3 ) 3 ) , 1.23 (d, 3H, Jap=12 Hz, C H 3 Ala), 1.12 (s, 9H, Ser-OC(CH 3) 3), 0.92 (d, 3H, Jy51 = 5.6 Hz, 8 i -CH 3 Leu), 0.90 (d, 3H, JyS2 = 5.7 Hz, 8 2 -CH 3 Leu); +LSI-MS (glycerol matrix) m/z 649 (M + H + , 28%), 593 (M + H* -H 2 C=C(CH 3 ) 2 , 28%), 537 (M + H + - 2 x H 2 O C ( C H 3 ) 2 , 63%); Anal. Calcd for C 3 3 H 5 2 N 4 0 9 : C, 61.09; H , 8.08; N , 8.64. Found: C, 61.23, H , 8.15; N , 8.50. 2.9.2.6 A'-Acetylglycyl-L-leucyl-L-seryl-L-S-phenyllactyl-L-alanine (5) Compound 15 (0.318 g, 0.490 mmol) was deprotected with formic acid in the usual manner to give 223 mg (85%) of crude 5. Following H P L C purification, 186 mg (71%) was obtained as a colourless, brittle solid: *H N M R (400 MHz, D 2 0 ) 8 7.32-7.15 (m, 5H, C 6 H 5 PhLac), 5.30 (dd, 1H, JaPa = 5.5 Hz, Japb = 6.7 Hz, a -CH PhLac), 4.54-4.50 (m, 1H, a -CH Ser), 4.34 (dd, 1H, JaPa = 4.6 Hz, Japb = 9.4 Hz, a -CH Leu), 4.25 (q, 1H, Jap = 7.3 Hz, a -CH Ala), 3.89 (dd, 1H, JaPa = 5.1 Hz, Jpp = 11.8 Hz, (3-CH2 a Ser), 3.83 (s, 2H, a - C H 2 Gly), 3.77 (dd, 1H, Japb = 4.1 Hz, Jpp = 11.8 Hz, (3-CH2 b Ser), 3.15 (dd, 1H, JaPa = 5.3 Hz, Jpp= 14.2 Hz, (3-CH2 a PhLac), 3.08 (dd, 1H, Jafib = 6.8 Hz, Jpp= 14.2 Hz, P - C H 2 b PhLac), 1.95 (s, 3H, C H 3 Ac), 1.60-1.42 (m, 3H, y-CH, p -CH 2 Leu), 1.22 (d, 3H, Jap= 7.2 Hz, C H 3 Ala), 0.84 (d, 3H, JyS, = 6.1 Hz, 8 i -CH 3 Leu), 0.80 (d, 3H, JyS2 = 6.1 Hz, 8 2 -CH 3 Leu); +LSI-MS (thioglycerol matrix) m/z 537 ( M + H+, 100%), 559 ( M + Na + , 91%), 575 ( M + K + , 63%); Analytical H P L C tR = 20.3 min (85%), 19.7 min (15%). 80 2.9.3 Synthesis of Chloroalanine Analogue 18 2.9.3.1 ^-(BenzyloxycarbonyO-L-phenylalanyl-L-alanine f-butyl ester (19) Z-Phe (6.60 g, 22.0 mmol) and Ala-OrBuHCl (4.00 g, 22.0 mmol) were coupled as described in Section 2.9.1.3. Recrystallization from EtOH/H 2 0 gave 8.4 g (90%) of 19 as white crystals. Spectral properties were in agreement with literature.102 2.9.3.2 3-Chloro-L-alanine hydrochloride (20) and 3-chloro-D-alanine hydrochloride (D-20) The title compounds were prepared by a slight modification of the procedure of Fischer and Raske. A suspension of Ser-OMeHCl (15.1 g, 97.1 mmol) in 50 mL distilled acetyl chloride was cooled to 0 °C and stirred vigorously. PCI5 (22 g, 110 mmol) was added in ~4 g portions every 5 min. As the mixture became viscous from the precipitation of a white solid, 100 mL additional acetyl chloride was added in small, periodic portions. Following 30 min of additional stirring, the suspension was filtered through a medium-pore glass Buchner funnel and washed with 200 mL cyclohexane to give 14.8 g (88%) of 3-chloro-L-alanine methyl ester, hydrochloride salt. After thorough air drying, the ester was dissolved in 70 mL 6 M HCI and refluxed gently overnight. Solvent was removed by rotary evaporation under reduced pressure. The resulting residue was redissolved in a minimum of warm H 2 0 , an equal volume of toluene was added, and the mixture was rotary evaporation under reduced pressure. The crude product was recrystallized from z'-PrOH/Et20 to give 10.7 g (69%) of 20 as colourless crystals. Spectral properties were identical to a commercial sample (Sigma-Aldrich). D-20 was prepared from DSer-OMeHCl in the same way. 81 2.9.3.3 ^ -(BenzyoxycarbonyO-S-chloro-L-alanine (21) and A^-(Benzyoxycarbonyl)-3-chloro-D-alanine (D-21) The title compounds were synthesized from 20 and D-20 according to a published procedure.64 2.9.3.4 iV-(Benzyloxycarbonyl)-3-chloro-L-alanyl-L-phenylalanyl-L-alanine /-butyl ester (22) Dipeptide 19 (2.50 g, 5.86 mmol) was hydrogenolyzed to give 1.71 g (100%) of the amine as a colourless oil. A portion of the amine (1.33 g, 4.55 mmol) was then coupled to 21 (1.17 g, 4.55 mmol). The residue was purified by flash chromatography (40:40:1 pet. ether/EtOAc/MeOH) to give 1.31 g (54%) of 22 as a colourless glass: ! H N M R (400 MHz, CDC13) 5 7.35 (s, 5H, C 6 H 5 Z), 7.28-7.13 (m, 5H, C 6 H 5 Phe), 6.76 (d, 1H, JHNa = 7.7 Hz, N H Phe), 6.25 (d, 1H, J m a = 7.1 Hz, N H Ala), 5.45 (d, 1H, JHNa = 7.3 Hz, N H ClAla), 5.11 (s, 2H, C H 2 Z), 4.69-4.61 (m, 1H, a -CH Phe), 4.55-4.48 (m, 1H, a -CH ClAla), 4.32 (quintet, 1H, JHNa = Jap = 7.1 Hz, a -CH Ala), 3.96 (dd, 1H, JaPa = 3.9 Hz, Jpp - 11.2 Hz, p -CH 2 a Ala), 3.68 (dd, 1H, Japa = 5.1 Hz, Jpp= 11.2 Hz, p -CH 2 b ClAla), 3.09 (dd, 1H, Japa = 6.4 Hz, Jpp= 13.7 Hz, p -CH 2 a Phe), 3.05 (dd, 1H, JaPa = 7.1 Hz, Jpp= 13.8 Hz, p -CH 2 a Phe), 1.43 (s, 9H, C(CH 3 ) 3 ), 1.29 (d, 3H, Jap= 7.1 Hz, C H 3 Ala); +LSI-MS (thioglycerol matrix) m/z 532 (M + H + , 6%), 476/478 (M + H + - H 2 C=C(CH 3 ) 2 , 3 5C1/ 3 7C1: 79/28%); Anal. Calcd for C 2 7H34C1N 30 6: C, 60.95; H , 6.44; N , 7.90. Found: C, 61.29; H , 6.57; N , 8.20. 2.9.3.5 A^-(Benzyloxycarbonyl)-3-chloro-D-alanyl-L-phenylalanyl-L-alanine /-butyl ester ([D-ClAla]22) The title compound was prepared as described above for 22 with the following change. After the final rotary evaporation, the crude product was recrystallized from EtOAc/Hex to give 1.75 g (62% based on 1.39 g D-21) of [D-ClAla]22 as a white solid: ' H N M R (400 MHz, CDC13) 5 7.35 (s, 5H, C 6 H 5 Z), 7.26-7.12 (m, 5H, C 6 H 5 Phe), 6.79 (d, 1H, JHNa = 7.6 Hz, N H Phe), 6.28 (d, 1H, JHNa = 7.1 Hz, N H Ala), 5.57 (d, 1H, J m a = 7.4 Hz, N H ClAla), 5.12 (s, 2H, C H 2 Z), 82 4.68-4.62 (m, 1H, a -CH Phe), 4.55-4.48 (m, 1H, a -CH ClAla), 4.32 (quintet, 1H, JHNa = JaP = 7.1 Hz, a -CH Ala), 3.95 (dd, 1H, Ja/h = A A Hz, Jpp = 11.2 Hz, p -CH 2 a ClAla), 3.66 (dd, 1H, Japb = 4.9 Hz, Jpp = 11.2 Hz, p -CH 2 b ClAla), 3.12 (dd, 1H, Ja/h = 6.1 Hz, Jpp = 13.6 Hz, p -CH 2 a Phe), 3.00 (dd, 1H, Japb = 7.1 Hz, Jpp= 13.8 Hz, p -CH 2 b Phe), 1.43 (s, 9H, C(CH 3 ) 3 ), 1.29 (d, 3H, JaP = 7.1 Hz, C H 3 Ala); +DCI-MS (NH 3) m/z 532 (M + H + ) , 476/478 (M + H + - H 2C=C(CH 3) 2); Elemental analysis was not obtained since subsequent coupling with 8 was accompanied by epimerization. 2.9.3.6 A^-Acetylglycyl-L-leucyl-3-chloro-L-alanyl-L-phenylalanyl-L-alanine /-butyl ester (23) Tripeptide 22 (3.01 g, 5.66 mmol) was hydrogenolyzed to give 2.10 g (93%) of the amine as a pale yellow foam. Dipeptide 8 (1.22 g, 5.28 mmol) in D M F was coupled to a portion of the amine (1.33 g, 4.55 mmol) in CH 2 C1 2 . Instead of the usual workup, the solvent was removed by rotary evaporation under reduced pressure, and the resulting amber residue was triturated in 50 mL CH 2 C1 2 , leaving 23 as a white solid which was collected by filtration. The filtrate was concentrated and partitioned between 50 mL water and 50 mL EtOAc. The organic layer was washed with 50 mL portions of 5% K H S O 4 , sat. NaHC0 3 , and brine, then dried over MgSCv The mixture was filtered and rotary evaporated, and the resulting residue was triturated with E t 2 0 to give additional 23 (total yield: 0.866 g, 27%): ! H N M R (400 MHz, CD 3 OD) 5 7.27-7.16 (m, 5H, C 6 H 5 Phe), 4.63 (dd, 1H, JaPa = 5.2 Hz, Japb = 9.0 Hz, a -CH Phe), 4.54 (dd, 1H, Ja/h = 5.2 Hz, Japb = 7.3 Hz, a -CH ClAla), 4.38 (dd, 1H, JaPa = 5.8 Hz, Japb = 9.4 Hz, a -CH Leu), 4.26 (q, 1H, Jap=1.2> Hz, a -CH Ala), 3.86 (s, 2H, a - C H 2 Gly), 3.80 (dd, 1H, JaPa = 4.9 Hz, Jpp= 11.3 Hz, p -CH 2 a ClAla), 3.74 (dd, 1H, Japb = 7.3 Hz, Jpp= 11.3 Hz, p -CH 2 b ClAla), 3.20 (dd, 1H, JaPa = 5.2 Hz, Jpp = 14.0 Hz, p -CH 2 a Phe), 2.93 (dd, 1H, Japb = 9.1 Hz, Jpp = 14.0 Hz, p -CH 2 b Phe), 1.98 (s, 3H, C H 3 Ac), 1.74-1.61 (m, 1H, y-CH Leu), 1.61-1.54 (m, 2H, P-CH 2 Leu), 1.45 (s, 9H, 83 C(CH 3 ) 3 ) , 1.35 (d, 3H, Jap = 7.3 Hz, C H 3 Ala), 0.96 (d, 3H, Jy5, = 6.7 Hz, S j - C ^ Leu), 0.92 (d, 3H, JyS2 = 6.4 Hz, 8 2 -CH 3 Leu); +LSI-MS (thioglycerol matrix) m/z 632 (M + Na + , 8%), 610/612 (M + H + , 3 5C1/ 3 7C1: 51/21%), 554/556 (M + H* - H 2 C=C(CH 3 ) 2 , 3 5 C1/ 3 7 C1: 56/26%); Anal. Calcd for C2 9H44C1N507: C, 57.09; H , 7.27; N , 11.48. Found: C, 56.95; H , 7.36; N , 11.51. 2.9.3.7 A^-Acetylglycyl-L-leucyl-3-chloro-L-alanyl-L-phenylalanyl-L-alanine (18) A solution of 23 (20 mg, 0.033 mmol) in 3 mL 88% formic acid was stirred at rt for 48 h. After removal of solvent by rotary evaporation under reduced pressure at < 35 °C, the residue was dissolved in HPLC-grade MeOH and evaporated to dryness. The resulting colorless foam was dissolved in 9 mL HPLC-grade MeOH and purified in ~2 mL portions by preparative reversed-phase HPLC, followed by concentration and lyophilization to give 14 mg (75%) of 18 as a white solid: ! H N M R (400 MHz, CD 3 OD) 8 7.24-7.13 (m, 5H, C 6 H 5 Phe), 4.60 (dd, 1H, Jajh = 5.2 Hz, Japb = 8.8 Hz, a -CH Phe), 4.52 (dd, 1H, JaPa = 5.2 Hz, Japb = 7.3 Hz, a -CH ClAla), 4.39-4.29 (m, 2H, a -CH Leu, a -CH Ala), 3.84 (s, 2H, a - C H 2 Gly), 3.78 (dd, 1H, JaPa = 5.2 Hz, Jpp = 11.3 Hz, (3-CH2 a ClAla), 3.72 (dd, 1H, Japb = 7.6 Hz, Jpp= 11.3 Hz, P-CH 2 b ClAla), 3.18 (dd, 1H, = 5.2 Hz, J^= 14.0 Hz, p - C H 2 a Phe), 2.91 (dd, 1H, Japb = 8.8 Hz, Jpp= 14.0 Hz, p-C H 2 b Phe), 1.96 (s, 3H, C H 3 Ac), 1.70-1.59 (m, 1H, y-CH Leu), 1.59-1.50 (m, 2H, p-CH 2 Leu), 1.37 (d, 3H, Jap= 7.0 Hz, C H 3 Ala), 0.93 (d, 3H, JyS, = 6A Hz, 8,-CH 3 Leu), 0.89 (d, 3H, Jy52 = 6.4 Hz, 8 2 -CH 3 Leu); +LSIMS (thioglycerol matrix) m/z 554/556 (M + H + , 3 5C1/ 3 7C1: 68/25%); Analytical HPLC tR = 15.9 min. 84 2.9.4 Synthesis of Dehydroalanine Analogue 24 2.9.4.1 TV-Acetylglycyl-L-leucyl-^e-phenyl-L-selenocysteinyl-L-phenylalanyl-L-alanine (36) The title peptide was synthesized on a 0.25-mmol scale by standard solid phase peptide synthesis protocols using the synthetic amino acid Fmoc-Sec(Ph) (33). After cleavage from the resin with 95% TFA, the peptide was dissolved purified by reversed-phase HPLC to give 94 mg (56%) of 36 as a white solid: ! H N M R (300 MHz, CD 3 OD) 8 7.53-7.46 (m, 2H, o-C 6H 5-Se), 7.29-7.14 (m, 8H, m,p-C6U5 Sec, C 6 H 5 Phe), 4.58 (dd, 1H, JaPa = 5.0 Hz, Japb = 8.9 Hz, a -CH Phe), 4.41-4.27 (m, 3H, a -CH Ala, a -CH Leu, a -CH Sec), 3.86 (s, 2H, a - C H 2 Gly), 3.25-3.14 (m, 2H, (3-CH 2 a Sec, p -CH 2 a Phe), 3.05 (dd, 1H, Japb = 9.2 Hz, Jpp = 12.3 Hz, p -CH 2 b Sec), 2.90 (dd, 1H, Japb = 9.2 Hz, Jpp = 13.9 Hz, p -CH 2 b Phe), 1.97 (s, 3H, C H 3 Ac), 1.71-1.60 (m, 1H, y-C H Leu), 1.60-1.50 (m, 2H, P-CH 2 Leu), 1.39 (d, 3H, JaP = 7.3 Hz, C H 3 Ala), 0.95 (d, 3H, JyS, = 6.2 Hz, 8 i -CH 3 Leu), 0.91 (d, 3H, Jy52 = 6.2 Hz, S 2 - C H 3 Leu); +LSI-MS (thioglycerol matrix) m/z 674/676 (M + H + , 7 8Se/ 8 0Se: 39/68%). 2.9.4.2 A^-Acetylglycyl-L-leucyl-2,3-didehydroalanyl-L-phenylalanyl-L-alanine (24) To a solution of 36 (40 mg, 0.059 mmol) in 10 mL HPLC-grade MeOH was added 30% aqueous H 2 0 2 (27 uL, 0.24 mmol), and the mixture was stirred at rt for 30 min. Reaction completion was confirmed by analytical reversed-phase HPLC, and the mixture was purified by preparative HPLC to give 28 mg (93%) of 24 as a white solid: ' H N M R (400 MHz, D 2 0 ) 8 7.31-7.17 (m, 5H, C 6 H 5 Phe), 5.51 (s, 1H, p -CH 2 a Dha), 5.46 (s, 1H, p -CH 2 b Dha), 4.59 (dd, 1H, JaPa = 6.4 Hz, Japb = 8.5 Hz, a -CH Phe), 4.30-4.23 (m, 2H, a-CH Ala, a -CH Leu), 3.84 (s, 2H, a-C H 2 Gly), 3.14 (dd, 1H, JaPa = 6.4 Hz, Jpp = 14.0 Hz, p -CH 2 a Phe), 2.97 (dd, 1H, Japb = 8.5 Hz, Jpp= 14.0 Hz, p -CH 2 b Phe), 1.93 (s, 3H, C H 3 Ac), 1.61-1.48 (m, 3H, y-CH, p-CH 2 Leu), 1.31 (d, 85 3H, JaP = 7.3 Hz, C H 3 Ala), 0.85 (d, 3H, JyS, = 5.8 Hz, 8 , -CH 3 Leu), 0.80 (d, 3H, JyS2 = 5.8 Hz, 5 2 -CH 3 Leu); +LSI-MS (3-nitrobenzyl alcohol matrix) m/z 518 (M + H + , 61%), 540 ( M + Na + , 26%); Analytical HPLC tR = 16.5 min. 2.9.5 Synthesis of Phosphinate Analogue 37 2.9.5.1 ^-(BenzyloxycarbonylJ-alanyl-YfPOi-CHil-phenylalanine ethyl ester (40) The pseudodipeptide analogue 40 was prepared as a diastereomeric mixture from racemic l-[(benzyloxycarbonyl)amino]ethylphosphinic acid (38) and ethyl ot-methylenebenzene-propanoate (39) according to Chen et al.90 Spectral properties were identical to those published elsewhere.91 2.9.5.2 A^^BenzyloxycarbonyO-alanyl-^KIPO^-l-adamantyO-CHil-phenylalanine (41) The hydroxyphosphinyl function of 40 was protected as the 1-adamantyl ester, and the carboxy ester was subsequently saponified as outlined in literature.91 2.9.5.3 ^-(BenzyloxycarbonyO-alanyl-TIPO^-l-adamanty^-CHzl-phenylalanyl-L-alanine methyl ester (42) Compound 41 (2.17 g, 4.01 mmol) and Ala -OMeHCl (0.588 g, 4.21 mmol) were coupled as described in Section 2.9.1.3. Flash chromatography (washed with 1:1 pet. ether/EtOAc and eluted with 15:15:1 pet. ether/EtOAc/AcOH) gave 1.88 g (75%) of 42 as a colourless brittle foam: *H N M R (400 MHz, CDC13) 5 7.36-7.08 (10H, C 6 H 5 Z, C 6 H 5 Phe), 6.82-6.50 (1H, N H Ala 2 ), 5.63-5.33 (1H, N H Ala,), 5.19-4.95 (2H, C H 2 Z), 4.49-4.33 (1H, a -CH Ala 2), 4.13-3.79 (1H, a-CH Alai), 3.67-3.53 (3H, C H 3 OMe), 3.10-2.60 (3H, P C H 2 , a -CH Phe), 2.40-2.16 (1H, p - C H 2 a Phe), 2.16-1.65 (10H, 3xCH Ad, 3xCH 2 Ad, p - C H 2 b Phe), 1.64-1.41 (6H, 3xCH 2 Ad), 86 1.39-1.03 (6H, C H 3 Ala , , C H 3 Ala 2 ); 3 1 P N M R (162 MHz, CDC13) 5 49.9, 49.6, 49.3, 48.6; +LSI-MS (3-nitrobenzyl alcohol) m/z 625 (M + H + , 54%). 2.9.5.4 A^-Acetylglycyl-L-leucyl-alanyl-^IPO^-l-adamanty^-CHzJ-phenylalanyl-L-alanine methyl ester (43) Phosphinate 42 (0.547 g, 0.876 mmol) was hydrogenolyzed to give 0.375 g (87%) of the amine as a brittle, colourless foam. Dipeptide 8 (0.176 g, 0.762 mmol) in D M F was coupled to the amine (0.374 g, 0.762 mmol) in CH 2 C1 2 . Flash chromatography (9:1 CHCl 3 /MeOH) gave 0.432 g (81%) of 43 as a pale amber foam: *H N M R (400 MHz, CDC13) 5 7.33-7.14 (7H, C 6 H 5 Phe, 2xNH), 7.11-6.41 (2H, 2xNH), 4.58-4.17 (3H, 3xct-CH), 4.16-4.03 (1H, a-CH), 4.03-3.58 (1H, a-CH), 3.75-3.61 (3H, C H 3 OMe), 3.29-2.64 (3H, P C H 2 , a -CH Phe), 2.41-2.24 (1H, p-C H 2 a Phe), 2.22-1.68 (13H, 3xCH Ad, 3xCH 2 Ad, P-CH 2 b Phe, y-CH, P-CH 2 Leu), 1.68-1.49 (6H, 3xCH 2 Ad), 1.46-1.07 (6H, C H 3 A l a h C H 3 Ala 2 ), 0.97-0.81 (6H, 2xCH 3 Leu); 3 1 P N M R (162 MHz, CDC13) 5 49.36, 49.26, 48.88, 48.77; +LSI-MS (glycerol) m/z 703 (M + H + , 58%), +ESI-MS (MeOH) m/z 725 ( M + Na + , 100%); Anal. Calcd for QeHssN^gP: C, 61.52; H , 7.89; N , 7.97. Found: C, 61.26; H, 7.99; N , 7.01. Note: Although N content is low, negative results from initial studies with the deprotected peptide 37 deemed further purification futile. 2.9.5.5 A^-Acetylglycyl-L-leucyl-alanyl-^IPOi-CHzl-phenylalanyl-L-alanine (37) The pentapeptide phosphinate was deprotected following a protocol used on a similarly protected compound.103 Thus, to a stirring solution of 43 (118 mg, 0.168 mmol) in 2 mL MeOH was added 1 mL of 4 M NaOH. The initially yellow solution was stirred at rt for 6 h, by which time it had become colourless. MeOH was removed by rotary evaporation, and the remaining solution was diluted with 5 - 1 0 mL distilled H 2 0 . Amberlite IR-120 (H) resin was added until pH ~2, resulting in the formation of a white precipitate. An equal volume of EtOAc was added, and the resin was removed by decantation. The organic layer was separated, washed with an 87 equal volume of brine, and dried over Na2S04. Filtration and rotary evaporation yielded 104 mg (90%) of the crude carboxylic acid as a colourless foam. The residue was dissolved in 5 mL of a 1:1 mixture of TFA/CH2CI2, and after stirring at rt for 10.5 h, the solvent was removed by rotary evaporation to give a colourless oil. Addition of 5 mL of 10% Na2CC>3 resulted in the formation of a precipitate that was later identified as 1-adamantyl alcohol. The solid was dissolved in 5 mL EtOAc and the layers were separated. The aqueous layer was acidified with Amberlite, filtered, and rotary evaporated. The residue was redissolved in 10 mL distilled H 2 0 and lyophilized to give 113 mg crude 37 as a brittle solid. The material was separated into four fractions, labelled in order of elution as 37a, b, c, and d, by preparative reversed-phase HPLC. Concentration and lyophilization of each fraction yielded the following: 37a, 2.0 mg hygroscopic solid: *H N M R and +ESI-MS inconsistent with product. 37b, 5.1 mg hygroscopic solid: ' H N M R and +ESI-MS inconsistent with product. 37c, 28.3 mg (30%) white solid: ' H N M R (300 MHz, D 2 0) 5 7.29-7.08 (m, 5 H , C 6 H 5 Phe), 4.24 (dd, 1H, JaPa = 4.9 Hz, JaPb = 9.1 Hz, a -CH Leu), 4.19-4.02 (m, 2H, a -CH A l a u a-C H Ala 2), 3.81 (s, 2H, a - C H 2 Gly), 2.91-2.69 (m, 3 H , a -CH Phe, PCH 2 ) , 1.95 (s, 3 H , C H 3 Ac), 1.93-1.87 (m, 1H, p - C H 2 a Phe), 1.76 (dd, 1H, Japb = 5.0 Hz, Jpp = 13.0 Hz, p - C H 2 b Phe), 1.60-1.41 (m, 3 H , y-CH, p-CH 2 Leu), 1.25-1.14 (m, 6 H , C H 3 A l a b C H 3 Ala 2), 0.84 (d, 3 H , JyS1 = 6.0 Hz, 5]-CH 3 Leu), 0.79 (d, 3 H , JyS2 = 6.0 Hz, 5 2 -CH 3 Leu); 3 1 P (121.5 MHz, D 2 0) 8 47.62; +ESI-MS (MeOH + 0.1% formic acid) m/z 555 (M + H + , 100%), 577 (M + Na + , 45%); -ESI-MS (MeOH) m/z 553 (M - H + , 100%). 37d, 26.5 mg (29%) white solid: ' H N M R (300 MHz, D 2 0 ) 8 7.30-7.09 (m, 5 H , C 6 H 5 Phe), 4.31-4.21 (m, 1H, a -CH Leu), 4.13-3.99 (m, 2H, a -CH Alai , a -CH Ala 2), 3.82 (s, 2H, a-C H 2 Gly), 2.99 (dd, 1H, JPCHaa= 4.6 Hz, Jgem = 12.7 Hz, PCH 2 a ) , 2.94-2.78 (m, 1H, a -CH Phe), 2.65 (dd, 1H, Jpcma = 10.5 Hz, Jgem = 12.8 Hz, PCH 2 b ) , 1.93 (dd, 1H, Japa = 4.7 Hz, Jpp = 13.6 88 Hz, p -CH 2 a Phe), 1.92 (s, 3H, C H 3 Ac), 1.76 (dd, 1H, Japb = 6.5 Hz, Jpp = 13.5 Hz, p -CH 2 b Phe), 1.61-1.42 (m, 3H, y-CH, p-CH 2 Leu), 1.19 (dd, 3H, Jap = 12 Hz, JHP = 13.9 Hz, C H 3 AlaO, 0.95 (d, 3H, JaP= 7.3 Hz, C H 3 Ala 2), 0.84 (d, 3H, JyS1 = 5.7 Hz, 8,-CH 3 Leu), 0.78 (d, 3H, Jy52 = 5.6 Hz, 8 2 -CH 3 Leu); 3 1 P (121.5 MHz, D 2 0 ) 8 44.99; +ESI-MS (MeOH + 0.1% formic acid) m/z 555 ( M + H + , 100%), 577 (M + Na + , 45%); -ESI-MS (MeOH) m/z 553 (M - H + , 100%). 2.9.6 Synthesis of Enzyme-Labelling Compounds 44, 45, and 46 2.9.6.1 2-Ethoxy-4H-3,l-benzoxazin-4-one (44) The title compound was synthesized from anthranilic acid and ethyl chloroformate as previously described.93 2.9.6.2 A'-Chloroacetyl-L-phenylalanyl-L-alanine (45) The neutral dipeptide Phe-Ala was chloroacetylated using a procedure similar to that reported for the preparation of iV-chloroacetyl-L-phenylalaine.95 A solution of chloroacetyl chloride (17 uL, 0.21 mmol) in 800 uL EtOAc was added to a solution of Phe-Ala (25.1 mg, 0.106 mmol) in 425 uL of 1 M NaOH in a glass vial. The vial was capped and vortexed for 2 min before being acidified with 141 uL of 6 M HCI, resulting in the formation of a white precipitate at the interface between the two layers. The solid was collected by filtration and washed with E t 2 0 to give 17.1 mg (52%) of 45: ' H N M R (400 MHz, DMSO) 8 8.44 (d, 1H, JHNa = 6.8 Hz, NH), 8.32 (d, 1H, J m a = 8.0 Hz, NH), 7.26-7.12 (m, 5H, C 6 H 5 Phe), 4.54 (m, 1H, a-C H Phe), 4.19 (quintet, 1H, JHNa = Jap=l.\ Hz, a -CH Ala), 3.99 (d, 1H, Jgem = 13.3 Hz, C H 2 a ClAc), 3.95 (d, 1H, Jgem = 13.4 Hz, C H 2 b ClAc), 3.01 (dd, 1H, JaPa = 3.0 Hz, Jpp = 13.3 Hz, p-89 C H 2 a Phe), 2.73 (dd, 1H, Japb = 10.1 Hz, Jpp = 13.1 Hz, p - C H 2 b Phe), 1.26 (d, 3H, Jap = 7.0 Hz, C H 3 Ala); +CI-MS (NH 3) m/z 313/315 (M + H + , 3 5C1/ 3 7C1: 100/33%). 2.9.6.3 Af-Bromoacetyl-L-phenylalanyl-L-alanine (46) Compound 46 was prepared as described for 45 using bromoacetyl bromide. Following acidification, however, no precipitate formed; thus, the organic layer was separated, washed twice with an equal volume of H 2 0 , and rotary evaporated: ! H N M R (400 MHz, DMSO) 8 8.46 (d, 1H, JHNa = 8.4 Hz, NH), 8.44 (d, 1H, JHNa = 7.2 Hz, NH), 7.26-7.14 (m, 5H, C 6 H 5 Phe), 4.55 (dt 1H, Jmia = 4.1 Hz, JaPa = Japb = 9.0 Hz, a -CH Phe), 4.21 (quintet, 1H, J m a = JaP = 7.3 Hz, a-CH Ala), 3.81 (s, 1H, C H 2 ClAc), 3.03 (dd, 1H, JaPa = 4.2 Hz, Jpp = 14.0 Hz, P-CH 2 a Phe), 2.73 (dd, 1H, Japb = 9.8 Hz, Jpp= 14.0 Hz, p - C H 2 b Phe), 1.29 (d, 3H, JaP=7.3 Hz, C H 3 Ala). 2.9.7 General Enzyme Methods Lyophilized and frozen spider venom were either purchased from SpiderPharm, Inc. or obtained by donation from Dr. Robert Volkmann (Pfizer Central Research) and Charles Kristensen (SpiderPharm). A l l buffers were purchased from Sigma-Aldrich or Fisher Scientific. Protein concentrations were determined by Bradford assay1 0 4 on a Cary 3E UV-Vis spectrophotometer using bovine serum albumin (BSA) in phosphate buffered saline (PBS; 4.3 m M Na 2 HP0 4 -7H 2 0, 1.4 mM K H 2 P 0 4 , 137 mM NaCl, and 2.7 mM KC1, adjusted to pH 7.4) as the standard. Protein purity was assessed using sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE), stained with Coomassie blue according to Laemmli. 1 0 5 Molecular weight standards for SDS-PAGE were B S A (66 kDa) and carbonic anhydrase (29 kDa), both purchased from Sigma. ESI-MS and L C - M S were performed by Mr. Shouming He on a Perkin-Elmer Sciex API300 electrospray mass spectrometer. 90 2.9.8 Enzyme Purification 2.9.8.1 Routine Enzyme Purification For routine assays the epimerase was purified by size-exclusion chromatography on Sephadex G-75sf gel (Amersham Biosciences, formerly Pharmacia) according to literature with the following change: in lieu of lyophilization, the isolated enzyme was concentrated by centrifugation through a membrane (10 000 MWCO) to a volume of 1.0-1.5 mL, divided into 50 JUL aliquots, flash-frozen in liq. N 2 and stored at -80°C. 2.9.8.2 Orthogonal Enzyme Purification For attempted removal of metalloprotease activity, cation exchange chromatography followed by size exclusion chromatography was performed by the following procedure. A 1 mL HiTrap SP HP column (Amersham Biosciences, formerly Pharmacia) connected to a BioSepra FPLC was equilibrated with 20 m M potassium phosphate buffer, pH 6.8. Two 100 uL samples of lyophilized venom were each dissolved in 500 uL distilled H 2 0 , pooled, and loaded onto the column at 0.2 mL min"1. Washing and elution were achieved at 1.0 mL min"1 by linear gradients of phosphate buffer containing 0.4 M NaCl. The relevant fractions were pooled and concentrated by centrifugation to 400 uL, which was then treated to size exclusion chromatography as described in the "Routine enzyme purification" above. Protein molecular weight was determined roughly by SDS-PAGE and accurately by ESI-MS and appeared unchanged (29 472 ± 3 Da, 29 312 ± 6 Da) from the theoretical weight (29 471 Da, 29 317 Da). 2.9.9 Epimerase and Metalloprotease Assays To assay epimerase activity, typically, a 62.5 uL (0.022 mg) solution of enzyme in PBS was equilibrated with 7.5 uL of 50 mM EDTA at 37 °C for 30 min, followed by the addition of 5 91 uL of 1.6 m M (100 uM final) substrate 1. Aliquots were analyzed after 5-23 h by reversed-phase HPLC. Extent of reaction was determined by integration of the peaks corresponding to 1 (relative retention time, tR = 13.8 min) and 2 (tR = 11.1 min). Enzyme activity (typically 1 x 10"4 umol min"1 mg"1) was calculated from the percent epimerization at early reaction stages (i.e., less than 10% conversion) so that the reverse reaction was negligible. Reaction reversibility was verified by incubating the enzyme with substrate 2 in place of 1. To assess metalloprotease activity, the same procedure as above was followed with the exclusion of EDTA. Because of the higher rate of protease activity relative to epimerization, aliquots were analyzed within 1 h. 2.9.10 Determination of Extinction Coefficients for 1,18, and 24 Varying concentrations (0 - 93 uM) of 1, 18, or 24 in deionized H2O containing 0.1% TFA were prepared and placed in a quartz cuvette with a 1 cm path length. Their absorbances at 220 nm (A220) were measured at 25 °C, and the corresponding extinction coefficients (£220) were determined from the slope of a plot of 2^20 vs. concentration. 2.9.11 Enzymatic Studies on Substrate Analogues 2.9.11.1 Enzymatic Reaction of 5 The standard epimerase assay was performed using 5 uL of a 1.6 m M (100 uM final) solution of 5 in H2O in the place of substrate. As a control, the 62.5 uL enzyme sample was substituted with buffer. New peaks detected in the HPLC profile of the enzymatic reaction were identified by retention time comparison to a sample of Ac-Gly-Leu-Ser-OH and H-PhLac-Ala-OH generated from the saponification of 5 in 1 M NaOH. Additionally, Ac-Gly-Leu-Ser-OH 92 was prepared in 91% yield by hydrogenolysis of 11 (91.6 mg, 0.225 mmol) and was found to correspond to the first peak ( t R = 2.6 min). H-PhLac-Ala-OH was prepared by hydrolysis of 12 (23.2 mg, 0.0791 mmol) in 3 mL of 1:2 TFA/CH2CI2 over 48 h and was found to correspond to the second peak ( t R = 7.9 min). 2.9.11.2 Enzymatic Reaction of 18 The standard epimerase assay was performed using 5 uL of a 1.6 m M (100 uM final) solution of 18 in DMSO in the place of substrate. As a control, the 62.5 [xL enzyme sample was substituted with buffer. New peaks detected in the HPLC profiles after 23 h were identified by retention time comparison to 1 and 24 as standards and by LC-MS. 2.9.11.3 Inhibition of Epimerase by 18 and 24 Enzyme samples were pre-incubated with EDTA for 30 min as described above in the typical assay. Five microlitres of 1.6 - 1600 u M (0.1 - 100 uM final) of 18 in DMSO or 24 in H2O was added and incubated for 1 h. The epimerase assays were subsequently initiated by the addition of 5 uL of 1.6 mM substrate 1, and aliquots were analyzed after 5-23 h. Samples containing 5 uL DMSO or H 2 0 in place of 18 and 24, respectively, served as controls. IC50 values were determined using GraFit 4.0 (Erithacus Software) by fitting data to a hyperbolic curve. 2.9.11.4 Attempted Inhibition by 37 Crude 37 as well as 37c and 37d were individually dissolved in H2O to two concentrations (15 u M and 1.5 mM). Enzyme samples (25 uL) were pre-incubated with 3 uL of 50 m M EDTA, followed by 2 |uL 37, 37c, or 37d, giving final concentrations of 1.0 or 100 uM. After incubation for 1 h, 2 uL of 1.6 m M substrate 1 was added, and the mixtures were incubated 93 for 18 h before being analyzed by HPLC. A control was prepared by substituting 2 uL H2O for 37. 2.9.12 Attempted Covalent Modification of the Enzyme 2.9.12.1 2-Ethoxy-4H-3,l-benzoxazin-4-one Compound 44 was dissolved in DMSO and incubated at final concentrations of 0.1 and 1.0 mM with the enzyme and E D T A at 37 °C for 15, 30, and 60 min. Aliquots were flash-frozen in liq. N 2 and submitted for ESI-MS analysis. Excess 44 was then removed by three rounds of ultrafiltration through a centrifugal membrane (10 000 MWCO), each time diluting the sample to 4 mL with PBS and reducing it to its original volume (125 uL). Again, aliquots were submitted for ESI-MS. 2.9.12.2 W-Haloacetyl Dipeptides Chloroacetyl peptide 45 was dissolved in DMSO and incubated at final concentrations of 0.01, 1, and 7 mM with the enzyme and EDTA as usual. Bromoaceryl peptide 46 was treated similarly to final concentrations of 0.01 and 1 mM. After 1 h, the samples were flash-frozen in liq. N 2 and submitted for ESI-MS. 94 Chapter 3 Enzyme-Catalyzed Epimerization of Carbohydrates 95 3.1 Inversion of Stereochemistry in Amino Acids Versus Carbohydrates The previous chapters have illustrated that by inverting the stereochemistry in amino acids and peptides, nature has developed a strategy for molecular diversification. The resulting derivatives bearing these D-amino acids possess unique properties that differ from their homochiral counterparts, a characteristic that can empower an organism with survival advantages. Because amino acids generally have only one stereocentre, however, the inversion of stereochemistry in peptides is largely limited to the a-carbon. Carbohydrates, on the other hand, can possess anywhere from one to greater than ten 1 0 6 stereocentres in each monosaccharide unit. This bestows upon nature an enormous degree of stereochemical diversity from which to assemble molecules with increased complexity and specialized properties. Whereas the common amino acids possess the L-configuration and become inverted to the D-form by racemases or epimerases, carbohydrates, which commonly possess the D -configuration, are not necessarily inverted to L-sugars; in fact, most carbohydrate epimerases do not form L-sugars. This characteristic is attributable to the D/L nomenclature convention, which is dependent on the configuration of the highest numbered stereocentre, called the "configurational atom". Thus, carbohydrate epimerases interconvert either a D-sugar and an L -sugar or a pair of D-sugars, depending on whether the isomerization occurs at the configurational atom or at another position. Further differing from amino acids is the environment of the centre of inversion. In amino acids and peptides, the racemizable/epimerizable position is almost invariably the ct-carbon, a position that can be described as "activated" in that the C - H bond is somewhat acidified by the adjacent electron-withdrawing and resonance-stabilizing carbonyl. In contrast, the epimerizable centre in a sugar may or may not lie in proximity to a carbonyl. As will be 96 described in the following sections, carbohydrate epimerases can generally be divided into two categories based on whether the inversion occurs at an "activated" or "unactivated" centre. 3.2 Epimerization at "Activated" Centres As was illustrated by the amino acid racemases in Section 1.3, the decreased pKa of a proton a to a carbonyl group allows for a deprotonation/reprotonation mechanism to invert a stereocentre. Likewise, some carbohydrate epimerases make use of such a mechanism to isomerize centres a to a carbonyl. D-Ribulose-5-phosphate 4-epimerase Arguably the most studied enzyme in this category is D-ribulose-5-phosphate 3-epimerase, which interconverts D-ribulose 5-phosphate (DRu5P) and D-xylulose 5-phosphate (DXu5P), as shown in Figure 3.1. These sugars are involved in the nonoxidative portion of the pentose phosphate pathway, which serves as an alternative to glycolysis in glucose metabolism. Early studies on the mechanism of the epimerase showed that tritium was incorporated at C-3 when the reaction was performed in H2O. Additionally, a primary kinetic isotope effect was detected through the use of [3- 2H]-DXu5P. 1 0 8 Reaction of this labelled substrate under irreversible conditions also revealed that the product incorporated a proton solely from the solvent, while recovered substrate retained deuterium. This lack of internal return suggested that HO. H OH 3-epimerase DRu5P DXu5P Figure 3.1 Reaction catalyzed by D-ribulose-5-phosphate 3-epimerase. 97 His His H i s x His His^ His DRu5P enediolate D X U 5 P Figure 3.2 Proposed mechanism of D-ribulose-5-phosphate 3-epimerase. Figure has been adapted from Jelakovic et al.m a two-base mechanism was operative. More recent work has allowed for a more detailed mechanism to be established. Based on X-ray crystal structural data of the enzyme from potato chloroplasts110 and rice 1 0 9 and using substrate modeling, the reaction mechanism in Figure 3.2 was proposed. In the forward direction, Asp38 abstracts the proton from C-3 of DRu5P, generating a cis-enediolate intermediate. The enzyme-bound Z n 2 + serves to polarize the carbonyl and to stabilize the oxyanion in the intermediate. Protonation by Asp 178 then yields the epimeric product. N-Acetylglucosamine 2-epimerase and similar enzymes Other examples of epimerization of free, "activated" sugars are the reactions catalyzed by iV-acetylglucosamine (GlcNAc) 2-epimerase, Af-acetylglucosamine-6-phosphate (GlcNAc6P) 2-epimerase, and cellobiose 2'-epimerase. These enzymes invert the configuration at C-2 in their respective glucose-based substrates to the corresponding mannose-based products. Although details are lacking for these enzymes, it has been suggested that they share a common mechanism involving deprotonation of the ring-opened forms of the sugars, reprotonation from the opposite face of the enol intermediate, and ring-closure (Figure 3.3).1 1 1 Thus, the inversion steps resemble the D-ribulose-5-phosphate 3-epimerase reaction. Evidence for such a mechanism 112 includes the observations that solvent isotope is incorporated at C-2' of cellobiose and that 98 : B - E n z RiO i) GlcNAc: X = NHAc; R-,, R 2 = H ii) GlcNAc6P: X = NHAc; R 1 = H; R 2 = P 0 3 2 iii) Cellobiose: X = OH R-, = ; R 2 = H RiO B—Enz H B - E n z i) ManNAc ii) ManNAc6P iii) p-D-Glcp-(1-^4)-D-Man H B - E n z RiO H B - E n z B - E n z b) NHAc NHAc 59 60 Figure 3.3 Proposed mechanism for epimerization at C-2 of glucopyranoses. a) The proposed mechanism of i) GlcNAc 2-epimerase, which interconverts Af-acetylglucosamine (GlcNAc) and Af-acetylmannosamine (ManNAc), ii) GlcNAc6P 2-epimerase, which interconverts A^-acetylglucosamine 6-phosphate (GlcNAc6P) and A^-acetylmannosamine 6-phosphate (ManNAc6P), and iii) cellobiose 2'-epimerase, which interconverts (3-D-glucopyranosyl-(l—>4)-D-glucose (cellobiose) and P-D-glucopyranosyl-(l—->4)-D-mannose. b) Inhibitors of GlcNAc 2-epimerase. compounds 59 and 60, which are analogues of the open-chain form of GlcNAc, are inhibitors of GlcNAc 2-epimerase.113 GlcNAc 2-epimerase has recently been shown to serve a catabolic role in the metabolism 113 of sialic acid, a carbohydrate that plays an important role in cellular recognition (vide infra: page 108). Until very recently, little information on the activity of GlcNAc6P 2-epimerase was known. First detected in the bacterium Enterobacter cloacae four decades ago," 4 no further reports had been made on the enzyme until 2000, when a recombinant protein from Neisseria 99 meningitidis was claimed to have this activity;1 1 5 however, no mechanistic aspects were addressed. Because it shares sequence homology with the enzyme that initiated the second project of this thesis (NeuC; see page 113), the N. meningitidis protein was selected for further studies, which are detailed in Chapter 4. 3.3 Epimerization at "Unactivated" Centres 3.3.1 NAD+-Dependent Epimerases In cases where the carbohydrate lacks a carbonyl or hemiacetal group, epimerases must utilize additional or alternative methods to invert the stereocentre. The most common strategy in dealing with such "unactivated" centres is the transient oxidation of the substrate using the cofactor N A D + . Thus, rather than catalyzing the removal of a proton at the epimerizable centre, these enzymes catalyze the removal of a hydride, which reduces N A D + to N A D H , and concomitantly results in the oxidization of the alcohol to a carbonyl (Figure 3.4; step 1). At this H OH H OH Figure 3.4 Epimerization at an "unactivated" centre by transient oxidation. Note that step 2b presumably involves an enol(ate) intermediate that is not shown. 100 point, one of two epimerization mechanisms can be used. In the upper pathway (step 2a), epimerization occurs at the site of oxidation via reduction on the opposite face of the carbonyl. In the lower pathway (step 2b), deprotonation, followed by reprotonation on the opposite face of the enol(ate) intermediate (not shown) and final reduction (step 3), results in an overall epimerization a to the site of oxidation. UDP-Galactose 4-epimerase The best understood example of a direct oxidation/reduction mechanism is the reaction catalyzed by UDP-galactose 4-epimerase. The enzyme, which is a participant in the Leloir pathway that interconverts galactose (Gal) and glucose (Glc), inverts the stereochemistry at C-4 of the uridine diphosphate (UDP) derivatives of these sugars (Figure 3.5). 1 1 6 A key step in accomplishing this epimerization is a rotation of the ketone intermediate 61 about the C - 0 and P-O bonds at the anomeric oxygen. With the UDP moiety held fixed in the enzyme's binding site, this rotation exposes opposite faces of the carbonyl to the reduced cofactor. A great number of studies on the enzyme's structure and mechanism have been reported and reviewed. 1 1 1 ' I 1 6" 1 1 9 Early experiments indicated that the epimers did not incorporate solvent-derived 2 H , 3 H , and 1 8 0 upon extended incubations with the enzyme. " This information UDP-GIc Figure 3.5 The mechanism of the reaction catalyzed by UDP-galactose 4-epimerase. 101 a) N a B 3 H 4 E * N A D + - E«NAD 3H Figure 3.6 Evidence for 4-ketohexose as an intermediate in the UDP-Gal 4-epimerase reaction, a) Chemical reduction of enzyme-bound N A D + and b) enzyme-catalyzed reduction of 4-ketohexose 62. eliminated mechanisms involving deprotonation/reprotonation or dehydration/rehydration. Support for a direct oxidation mechanism came from the observation of a primary kinetic isotope effect on the epimerization of [4"- 3H]UDP-Gaf or [4"- 3 H]TDP-Glc . 1 2 3 , 1 2 4 Perhaps the strongest evidence for the mechanism in Figure 3.5 was obtained from analysis of the 4-ketohexose intermediate (61) and its analogues. Prolonged incubations with an equilibrated mixture of UDP-Gal and UDP-Glc were accompanied by the release of 61 into solution. 1 2 5" 1 2 7 Additionally, treatment of the reduced form of the epimerase (E«NAD3H), which was generated by reaction with N a B 3 H 4 (Figure 3.6a), with the intermediate analogue UDP-6-deoxy-a-D-xy/o-hexos-4-ulose (62), followed by acid hydrolysis, produced a mixture of 6-deoxy-glucose (63) and 6-1 98 deoxy-galactose (64) labelled with tritium at C-4 (Figure 3.6b). These studies also served to 127 128 show that the hydride is transferred to and from one face (termed P or si) of the cofactor. ' Homologous epimerases are presumed to utilize N A D + through the same direct oxidation/reduction mechanism as UDP-galactose 4-epimerase. These enzymes include UDP-JV-acetylglucosamine 4-epimerase, UDP-glucuronate 4-epimerase, and UDP-xylose 4-epimerase.16' 129-131 * The numbering convention for sugar nucleotide diphosphates specifies locants of the nucleotide base unprimed numerals, those of the ribosyl ring single primed numerals, and those of the glycosyl ring double primed numerals. 102 GDP-D-mannose 3,5-epimerase Several other epimerases utilize N A D + to catalyze a transient oxidation at C-4 to bring about epimerization at an unactivated centre; however, instead of inverting the centre at which oxidation occurs, these enzymes take advantage of the increased acidity at C-3 and C-5 (recall pathway b in Figure 3.4). One such enzyme, GDP-D-mannose 3,5-epimerase, catalyzes the interconversion of GDP-D-mannose and GDP-L-galactose via epimerizations at C-3 and C-5 (note, however, that since the isomers are not epimers, the enzyme is technically not an epimerase). Incubation of either substrate with the enzyme in tritiated water led to incorporation of 3 H at C-3 and C-5 in both isomers. This observation is consistent with a mechanism involving an initial oxidation, followed by two successive deprotonation/reprotonation steps via enediol intermediates 65 and 66 and final reduction of the ketone (Figure 3.7a). The biosyntheses of dTDP-L-rhamnose from dTDP-D-glucose and of GDP-L-fucose from a) dTDP-GIc dTDP-rhamnose GDP-Man GDP-fucose Figure 3.7 Epimerization at C-3 and C-5. a) Proposed mechanism for the enzymatic interconversion of GDP-D-mannose and GDP-L-galactose. b) dTDP-GIc is converted to dTDP-rhamnose by the action of RlmB (dehydratase), RlmC (3,5-epimerase), and RlmD (reductase), and GDP-Man is converted to GDP-fucose by G M D (dehydratase) and GFS (epimerase/reductase). 103 GDP-D-mannose each involve the action of multiple enzymes that collectively function in a manner that resembles the mechanism of GDP-D-mannose 3,5-epimerase (Figure 3.7b). 1 4 ' 1 1 8 A principal difference, however, is that formation of the initial ketone intermediate in these reactions does not involve oxidation by N A D + ; rather, it is under the direction of a dehydratase activity, which additionally serves to deoxygenate C-6. 3.3.2 Cofactor-Independent Epimerases The enzymes described above illustrate how a potentially difficult stereoinversion can be accomplished with the aid of an oxidative cofactor. However, there is another group of carbohydrate epimerases that invert "unactivated" centres but do not employ a cofactor. As a result, the mechanisms by which these enzymes epimerize their substrates are not immediately clear, and therefore, they have become attractive targets for further study. L-Ribulose-5-phosphate 4-epimerase The enzyme L-ribulose-5-phosphate 4-epimerase plays a role in arabinose metabolism and catalyzes the interconversion of L-ribulose 5-phosphate (LRu5P) and D-xylulose 5-phosphate (DXu5P; this is also the product of D-ribulose-5-phosphate 3-epimerase described in Section 3.2). In labelling experiments using 3 H 2 0 and H 2 1 8 0 , neither isotope was found to be incorporated,107 strongly disfavouring a mechanism involving deprotonation or dehydration. Furthermore, incubations with [4- H]-DXu5P yielded no significant kinetic isotope effect. These findings and the fact that the enzyme utilizes a divalent metal ion, rather than an N A D + cofactor, 1 3 4 ' 1 3 5 suggested that epimerization involves C - C bond cleavage (Figure 3.8). A retro-aldol reaction proceeds with fission of the bond between C-3 and C-4. Rotation of the aldehyde intermediate (67) to expose the opposite face, followed by an aldol addition gives the epimeric 104 1 C H 2 O H :<X X H 2 O H . - O A CH 2 OH CH 2 OH 2 + M ' ^ 2 + M ' | 2 + M ' 3 K / C . 2 + M | HO—C- ) -H . H O ^ H - H O f H Enz . H O — C — H — fiR ^ \ h / 0 ^ - 4 C - H O ^ / H , U I M ' U " H % ^ H B H - C - O H " \ 5 C H 2 O P 0 3 2 - ^ L o P O 2 - C H 2 O P 0 3 2 -p,./ rw npn 2- UH2UPU3 B- LRu5P CH2OPO3 D X u 5 p Enz Figure 3.8 Mechanism of the reaction catalyzed by L-ribulose-5-phosphate 4-epimerase. M + = divalent metal cation. product. Recent isotope effect, structural, and mutagenesis studies have confirmed this mechanism and identified the catalytic acid/base residues.136"138 Interestingly, L-ribulose-5-phosphate 4-epimerase was also shown to catalyze a slow aldol reaction between glycolaldehyde phosphate (67) and dihydroxyacetone, which is a conjugate acid of the putative enolate intermediate (68).139 Non-hydrolyzing UDP-N-acetylglucosamine 2-epimerase The other cofactor-independent epimerase that inverts the stereochemistry at an "unactivated" centre is UDP-iV-acetylglucosamine 2-epimerase. Two versions of this enzyme have been found in nature, distinguishable by the overall reaction they catalyze (Figure 3.9). One of these, found in mammals and some bacteria, converts UDP-GlcNAc to ManNAc, a W U L " 2-Epimerase UDP-GlcNAc UDP-ManNAc Figure 3.9 Reactions catalyzed by hydrolyzing and non-hydrolyzing UDP-GlcNAc 2-epimerases. process that involves both stereoinversion and hydrolysis. The other, found exclusively in bacteria, catalyzes a true epimerization by interconverting UDP-GlcNAc and UDP-ManNAc. To avoid confusion, these enzymes will be referred to hereinafter as hydrolyzing and non-hydrolyzing UDP-GlcNAc 2-epimerase, respectively. The non-hydrolyzing epimerase functions to provide bacteria with a source of UDP-ManNAc for use in the biosynthesis of cell wall polysaccharides. Most studies on the enzyme have been conducted on the protein isolated from Bacillus cereus or Escherichia coli, and from both sources, it was found to require neither a metal nor a cofactor for activity. 1 4 0 ' 1 4 1 It was discovered, however, that UDP-GlcNAc is essential for catalytic turnover; in fact, the reverse reaction (i.e., UDP-ManNAc —» UDP-GlcNAc) could only proceed in the presence of small quantities of UDP-GlcNAc. Kinetic assays established that as well as a substrate, UDP-GlcNAc serves as a regulatory modulator (a description of allosteric regulation is given in Section 4.3.3, page 139). 1 4 0 ' 1 4 1 Implications on the nature of this regulation came from the recently reported structure of the E. coli epimerase, obtained by X-ray crystallography (Figure 3.10).1 4 2 Each of the homodimeric enzyme's subunits is composed of two domains, at the centre of which is a cleft where the substrate binds. The structure revealed that the two subunits exist in slightly different conformations, "closed" and "open", differing Figure 3.10 Structure of the UDP-GlcNAc 2-epimerase dimer. 106 by a 10° interdomain rotation. On the basis of the extensive hydrogen bonding interactions detected between the substrate-derived UDP moiety and residues in the "closed" subunit, it was suggested that the "closed" chain is the catalytically active subunit, whereas the "open" chain plays a regulatory role. It is possible that substrate binding in the "open" subunit induces a conformational change across the dimer interface that converts the neighbouring subunit into an active form. Until the past decade, most of the details surrounding the mechanism of the non-hydrolyzing UDP-GlcNAc 2-epimerase remained unclear. Incubation of UDP-GlcNAc with the epimerase in tritiated buffer led to incorporation of 3 H at C-2" of both epimers.143 This result excluded a direct oxidation/reduction mechanism, such as that used by UDP-Gal 4-epimerase, and clearly established that epimerization involves deprotonation and reprotonation at this position. To account for this finding, Salo proposed a mechanism that proceeds by transient s-: B—Enz NAD + NADH H O - ~ ^ £ AcHN I ^AcHN a) HO OUDP UDP-GlcNAc H 0 ~ - - \ N H A c NAD + NADH H 0 H OUDP UDP-ManNAc NHAc 1 \ OUDP H O — J t l A C ^ BH \ 0 U D P -O ^ NHAc Enz OUDP b) Enz—B:-x HO--^. \ HO HO ~Q AcHN/1 O V X I / O U M P UDP-GlcNAc Enz H ° ^ \ NHAc if cr H1 O-- O . I .OUMP O ^ I / O U M P Y n O UDP-ManNAc O Figure 3.11 Proposed mechanisms of the reaction catalyzed by the non-hydrolyzing UDP-GlcNAc 2-epimerase. a) Transient oxidation at C-3" by a tightly bound N A D + cofactor, followed by deprotonation/reprotonation and reduction, b) anti elimination of UDP to generate a glycal intermediate, followed by syn addition of UDP. 107 oxidation at C-3" (Figure 3.11a), similar to that employed by GDP-D-mannose 3,5-epimerase (recall Figure 3.7). Despite the fact that exogenous N A D + did not affect activity, it was possible that the cofactor was tightly bound within the enzyme's active site. However, more recent experiments using U V spectroscopy, proteolytic digestion, and amino acid sequence analysis on the recombinant enzyme from E. coli (this protein is referred to hereinafter as RffE, the product of the gene rffE) conclusively showed that N A D + is not present.144 An alternative mechanism that is consistent with solvent isotope incorporation involves reversible elimination of UDP (Figure 3.1 lb). In the UDP-GlcNAc -> UDP-ManNAc direction, anti elimination of UDP with H-2" removal would generate 2-acetamidoglucal as an intermediate. UDP could then add back with protonation from the opposite face of the glycal to give the epimeric product. Substantial evidence in support of such a mechanism was obtained by Morgan et al. using isotopically labelled substrates.144 Kinetic assays using (2"- 2H)UDP-GlcNAc revealed a primary kinetic isotope effect on £ c a t of 1.8, indicating that H-2" is removed in a rate-limiting step. More importantly, incubations with [ l" - 1 8 0]UDP-GlcNAc resulted in isotopic scrambling between the bridging (i.e., linking the anomeric carbon and P-phosphorus) H 0 ^ \ NHAc HO HO AcHN O - O ACHN I y M p ' , 0 U M p o UDP-GlcNAc M U H O A ^ 1 \ - # ° " 0 U M P UDP-ManNAc H 0 ^ \ ^ AcHN o \ H 0 - A NHAc H C r - \ A - — o ^ H O - ' v M - ^ HOA--*11-^ , HO--X--*J»*A AcHN I O " I 0~ O. | .OUMP O . | .OUMP Figure 3.12 Positional isotope exchange catalyzed by the non-hydrolyzing UDP-GlcNAc 2-epimerase. • = 1 8 0 . Incubation of UDP-GlcNAc containing 1 8 0 at the bridging position (top left) with the epimerase produced a mixture of both epimers with the label scrambled between bridging (top) and nonbridging (bottom) positions. 108 and nonbridging positions on the P-phosphate of the epimers (Figure 3.12). Such a finding indicates that C - 0 bond cleavage occurs and implicates UDP as an intermediate, whereby bond rotation would allow the oxygen formerly at the bridging position to exchange with one of the two oxygens formerly at a nonbridging position. Further support for the intermediacy of the glycal and UDP was provided by their direct observation by N M R spectroscopy, mass spectrometry, and HPLC. After extended incubations with high concentrations of the epimerase, the UDP-sugar mixture was eventually converted almost entirely to 2-acetamidoglucal and UDP, indicating they are thermodynamically more stable than either UDP-sugar. Mammalian hydrolyzing UDP-N-acetylglucosamine 2-epimerase Prior to the identification of the bacterial non-hydrolyzing epimerase, a similar enzyme was isolated from rat liver. 1 4 5 Although the two epimerases share many similarities, there are several structural and functional features that distinguish the two. Whereas the non-hydrolyzing epimerase is a homodimer with a subunit molecular weight of 42 kDa, the mammalian enzyme is a much larger homohexamer with a subunit molecular weight of 75 kDa. 1 4 6 In addition, the mammalian protein contains two catalytic domains that catalyze consecutive reactions: the N -terminus contains the epimerase activity that converts UDP-GlcNAc to ManNAc and UDP, whereas the C-terminus contains a kinase activity that phosphorylates ManNAc to ManNAc 6-phosphate (ManNAc6P). 1 4 7 These reactions are the first two steps in the mammalian biosynthesis of a class of polyhydroxylated a-keto acids known as sialic acids (Figure 3.13). Subsequent condensation of ManNAc6P with phosphoenol pyruvate (PEP; step 3), followed by phosphate hydrolysis (step 4), generates the nine-carbon carbohydrate 7V-acetylneuraminic acid (Neu5Ac), which is the most prevalent member of the sialic acids. Following activation of Neu5Ac as the CMP derivative (step 5), the sialic acid becomes transferred to the termini of cell-surface oligosaccharides (step H O - ^ H O -\ H 2 0 UDP " \ NHAc H<fis: AcHN 1 1 ManNAc OH OUDP UDP-GlcNAc 109 ATP ADP " 3 ^ " - , NHAc ManNAc6P OH PEP-" 3 4 2-O3POS C 0 2 H f ^\ Pi H 2 0 AcHN C 0 2 H HO Neu5Ac9P HO CMP-Neu5Ac sialylated oligosaccharides Figure 3.13 Mammalian biosynthetic pathway for the sialic acid iV-acetylneuraminic acid. The numbered reactions are catalyzed by the following enzymes: 1) Hydrolyzing UDP-./V-acetylglucosamine 2-epimerase, 2) Af-acetylmannosamine kinase, 3) 7V-acetylneuraminic acid 9-phosphate synthase, 4) A'-acetylneuraminic acid 9-phosphate phosphatase, 5) C M P - j V -acetylneuraminic acid synthetase, 6) sialyltransferases. The epimerase reaction is boxed to emphasize that it serves as the key regulator in the pathway. P, = inorganic phosphate; PP, = pyrophosphate; S A = sialic acids (e.g., Neu5Ac); O = monosaccharide. 6). It is widely held that the display of these sialic acids serves critical roles in cellular recognition and adhesion processes;148 in fact, their function in these processes is believed to account for the observed correlation between cell surface sialylation and the incidence of metastasis of cancerous cells. 1 4 9 Within the biosynthetic pathway, the epimerase reaction has been found to be the rate-determining step, and therefore, it is considered the key regulator in the production of sialic acid. 1 5 0 This conclusion is illustrated by the discovery that the final compound, CMP-Neu5Ac, acts as a feedback inhibitor of the epimerase.151 Despite the fact that the reaction catalyzed by the mammalian hydrolyzing epimerase is technically not an epimerization (since the substrate and products are not isomeric), it was presumed to be mechanistically similar to the reaction catalyzed by the non-hydrolyzing 110 epimerase.143 For many years, this contention had been largely speculative, primarily supported by the observation of incorporation of solvent-derived tritium at C-2 in the ManNAc product. However, the recent success in cloning the enzyme revealed 22% sequence identity of the N -terminal region to RffE , 1 4 7 further suggesting a common mechanism. This discovery prompted Chou et al. to examine the conversion of UDP-GlcNAc to ManNAc and UDP in light of three possible mechanisms (Figure 3.14).153 While one involves hydrolysis prior to epimerization (Path A) and therefore does not invoke the homology with RffE, the other two (Paths B and C) could proceed via elimination and addition steps similar to the enzyme's non-hydrolyzing counterpart. These potential mechanisms presumably share the same first step (i.e., elimination of UDP to generate a glycal intermediate) but differ in that Path B completes the epimerization to UDP-ManNAc prior to hydrolysis, whereas Path C foregoes re-addition of UDP and instead hydrates the glycal to give ManNAc directly. To distinguish between the three potential mechanisms, the putative intermediates in each case were separately studied for catalytic competence. Incubations of the enzyme with GlcNAc AcHN 2-acetamidoglucal Figure 3.14 Possible mechanisms for the reaction catalyzed by the mammalian hydrolyzing UDP-GlcNAc 2-epimerase. I l l failed to yield ManNAc, 1 4 5 thereby further weakening Path A . To test Path B , UDP-ManNAc was treated with the epimerase, which catalyzed hydrolysis to ManNAc. 1 5 4 Although this result would seem to support Path B, examination of the ManNAc product from incubations in 3 H20 revealed tritium incorporation at C-2, whereas tritium was absent from the UDP-ManNAc recovered from an incomplete reaction. These observations argue against both steps in Path B, suggesting that UDP-ManNAc simply serves as an alternative substrate. In contrast to the previous two mechanisms, Path C was supported by the introduction of its putative reaction intermediate. When incubated with the epimerase, 2-acetamidoglucal was converted to a compound that co-migrated with ManNAc by paper electrophoresis.155 Chou et al. eventually demonstrated that the identity of this compound was in fact ManNAc by following the reaction by 'H-NMPv spectroscopy.153 Furthermore, when the epimerase reaction was conducted in H 2 1 8 0 , no label could be detected in the UDP product. Additionally, [ l " - 1 8 0]UDP-GlcNAc was shown to be cleaved such that the l s O label departed with UDP. These labelling experiments indicate that the reaction proceeds with cleavage of the bond between the anomeric carbon and the bridging oxygen, as had been observed with RffE (vide supra). Thus, it was concluded that Path C is operative. Bacterial hydrolyzing UDP-N-acetylglucosamine 2-epimerase Although most studies on sialic acids have focused on mammalian sources, there are some strains of bacteria that also synthesize and display these special carbohydrates on their surfaces. 1 5 6 ' 1 5 7 In contrast to animal cells, which incorporate sialic acids at the termini of cell surface oligosaccharides, bacteria tend to introduce sialic acids (usually Neu5Ac) as internal residues in lipopolysaccharides (LPS) or as a homopolymer called polysialic acid (PSA) in capsular polysaccharides.156 Many of the bacteria that express LPS or PSA are pathogens that cause human diseases such as meningitis (Escherichia coli strain K l and Neisseria meningitidis 112 group B ) 1 5 8 and food-borne gastroenteritis (Campylobacter jejuni)}59,160 Sialic acids on the cell surface of these bacteria are thought to provide a protective barrier that enables evasion of the host's immune response. Advances in the past decade have allowed for the nearly complete elucidation of the pathways involved in the biosynthesis of sialic acids in several bacterial species. The genes responsible for the production and polymerization of Neu5Ac in E. coli K l are localized within a single gene cluster, termed neu (Figure 3.15a).161"163 A similar organization of genes (termed polysialic acid (PSA) Figure 3.15 The sialic acid biosynthetic pathway in Escherichia coli K l . a) The neu gene cluster encoding the enzymes involved in the synthesis and polymerization of Neu5Ac. b) The known enzymes in the pathway are indicated: NeuB = 7V-acetylneuraminic acid synthase; NeuA = CMP-N-acetylneuraminic acid synthetase; NeuS = CMP-Neu5Ac: poly-a-2,8-sialosyl sialyltransferase. NeuC (red) is a putative hydrolyzing UDP-GlcNAc 2-epimerase. Alternatively, a GlcNAc6P 2-epimerase and ManNAc6P phosphatase (blue) have been suggested to produce ManNAc (see text below). 113 sia) is found in N. meningitidis group B . 1 6 4 The function, i f known, of each gene product in the biosynthetic pathway is indicated in Figure 3.15b. Overall, the pathway closely resembles that of mammals (compare to Figure 3.13) with notable exceptions. First, Neu5Ac is produced from the direct condensation of ManNAc with PEP, rather than through ManNAc6P and Neu5Ac9P as intermediates. Second, the activated form of Neu5Ac, CMP-Neu5Ac, is polymerized to PSA under the direction of a polymerizing transferase, rather than attached to the terminus of another oligosaccharide chain. Finally, and most relevant to this project, the enzyme(s) responsible for the generation of ManNAc had not been conclusively identified in either bacterial species as of the beginning of the research presented in Chapter 4. In fact, it has been argued that ManNAc is produced by the action of two enzymes, GlcNAc6P 2-epimerase and ManNAc6P phosphatase,115 though this claim could not be substantiated with sufficient evidence.165 In view of the mammalian pathway, however, it would be reasonable that some bacteria may utilize UDP-GlcNAc as a source of ManNAc through the action of a hydrolyzing UDP-GlcNAc 2-epimerase. Indeed, the product (NeuC) of the neuC gene in E. coli K l was selected by Dr. Willie Vann and coworkers, with whom we eventually collaborated, as a candidate for fulfilling this function.166 Vann and coworkers noted that NeuC is homologous to the N-terminal region of the bifunctional hydrolyzing UDP-GlcNAc 2-epimerase/ManNAc kinase from human, mouse, and rat liver. This region, which contains the epimerase activity, exhibits 53% similarity and 27% identity to NeuC. Thus, it was reasonable to suspect that the two enzymes could catalyze a common reaction, and so, attempts were made to demonstrate this activity. The importance of NeuC to sialic acid biosynthesis was shown by a genetic complementation experiment.166 A n E. coli cell line containing a deletion in the neuC gene was constructed, resulting in an acapsular phenotype, indicating that PSA was not being expressed. When a plasmid containing neuC was supplied in trans (i.e., as a D N A molecule separate from the genomic DNA), however, capsule synthesis was restored. Likewise, when a derivative of 114 this cell line containing a deletion in the nanA gene, which encodes an enzyme that degrades Neu5Ac, was provided with exogenous Neu5Ac, capsule synthesis was also restored. These findings therefore established a link between NeuC and the synthesis and display of sialic acid. Prior to our collaborative efforts, the only chemical assay for NeuC activity was performed by incubations with radiolabeled substrate and analysis of the products by paper chromatography. Purified NeuC was incubated with UDP-[ 1 4C]GlcNAc, resulting in the generation of two spots in the paper chromatogram (Figure 3.16). Under these elution conditions, the major spot (B) appeared to co-migrate with a GlcNAc standard, while the minor spot (A) co-migrated with a ManNAc standard. The apparent generation of GlcNAc was attributed to substrate hydrolysis, whereas the formation of small amounts of ManNAc was attributed to the enzymatic activity of NeuC. This evidence, however, was _. „ deemed insufficient to satisfy their conclusions, as a Figure 3.16 Autoradiogram showing activity of NeuC Lane m a n u s c r i t c o n t a i n i n g t h e s e f i n d i n g s w a s r e j e c t e d f o r 1 reveals two reaction products S ^ ^ T ? * 0 0 . 0 / , publication. Thus, although it would appear that the activity [ 1 4 C]GlcNAc with NeuC. A and 6 K F J B were assigned as ManNAc and o f N e u C [ % a § a h d r o l { U D P - G l c N A c 2-epimerase, a GlcNAc, resp., by comparison to standards (not shown). Lane 2 is u . « a m o r e c o n v i n c i n g a s s a y would be necessary, a control using boiled enzyme. 115 3.4 Project Goals Encouraged by the prior success in studying the bacterial non-hydrolyzing and mammalian hydrolyzing UDP-GlcNAc 2-epimerases in our laboratory, we attempted to assist Vann and coworkers by applying more diagnostic techniques to monitor the reaction catalyzed by NeuC. The first section of Chapter 4 describes our use of N M R spectroscopy and mass spectrometry to characterize the products of the reaction. Our findings enabled us to conclude that NeuC catalyzes an anti elimination of UDP to generate 2-acetamidoglucal as the major product. This reaction is identical to the first step of the proposed mechanism of the mammalian epimerase, and therefore, we were able to establish a functional link between the two enzymes. Based on the results with the NeuC enzyme from E. coli, we proceeded to investigate the homologous protein SiaA from N. meningitidis. The remaining sections of Chapter 4 outline the cloning and expression of the siaA gene, followed by N M R , mass spectral, and kinetics studies that allowed for the identification and characterization of the enzyme's activity. Additionally, in an attempt to shed further light on the mechanism of SiaA, several experiments using isotopically labelled substrates are outlined. Finally, mechanistic similarities and differences among the four UDP-GlcNAc 2-epimerases - the non-hydrolyzing enzyme (RffE), the mammalian hydrolyzing enzyme, and the two newly identified bacterial enzymes (NeuC and SiaA) - are discussed. 116 Chapter 4 Mechanistic Studies on a Bacterial Hydrolyzing UDP-/V-Acetylglucosamine 2-Epimerase 117 As mentioned at the end of Chapter 3, the role that the hydrolyzing UDP-GlcNAc 2-epimerase from mammals plays in sialic acid biosynthesis and the details of its catalytic mechanism have been well established. The existence of sialic acids in some bacteria has prompted the search for a similar enzyme in these organisms. A bacterial version of the enzyme would be of interest, as it would likely share a similar structure and mechanism with the mammalian enzyme but would not suffer from the known instability that limits the types of experiments that can be performed with that enzyme. 1 5 2" 1 5 4 For instance, numerous attempts to crystallize the mammalian epimerase have failed, 1 6 8 but as is frequently the case, a bacterial enzyme would likely be readily crystallized. The initial work by Vann and coworkers in isolating the product NeuC from the Escherichia coli K l gene neuC indicated that this enzyme is likely a hydrolyzing UDP-GlcNAc 2-epimerase. However, the shortage of experimental evidence supporting this conclusion provided us with an opportunity for collaboration. Presented in Section 4.1 is our approach to identify the activity of NeuC, primarily through the use of N M R spectroscopy, and our discovery that NeuC acts to form UDP and 2-acetamidoglucal instead of catalyzing the expected reaction. Because the major product observed with NeuC was the proposed glycal intermediate rather than ManNAc, we sought to discover a homologue with uncrippled hydrolyzing UDP-GlcNAc 2-epimerase activity; based on sequence comparison, the siaA gene from Neisseria meningitidis B was targeted. The remaining sections of this chapter, therefore, illustrate the cloning and expression of this gene in E. coli and unambiguously demonstrate the activity of the protein SiaA. Additionally, characterization of the enzyme through N M R spectroscopy and kinetics is described, and attempts to obtain a more detailed picture of the mechanism through kinetic isotope effect and positional isotope exchange experiments are presented. 118 4.1 Enzymatic Activity of NeuC 4.1.1 Affinity Purification of NeuC To complete the details of the reaction catalyzed by NeuC, Dr. Willie Vann provided us with the plasmid pSR647, 1 6 6 which contains the neuC gene inserted into a plasmid that had been obtained as part of the IMPACT (Intein Mediated Purification with an Affinity Chitin-binding Tag) system from New England Biolabs. 1 6 9 pSR647 encodes a protein consisting of NeuC fused at its ('-terminus to what is termed an intein tag. This fusion » ^ _ J — — ^ ^ ^ ^ M protein facilitates purification by introducing an affinity promoter Expression domain that allows the protein to bind specifically to a resin. V native protein in a single chromatographic step. Hence, this technique allows for rapid, simple purification of a from the column while the intein-CBD fusion remains bound. column by incubation in the presence of a thiol reagent, such as is induced to undergo intein-mediated self-cleavage on the impurities are washed through. At this point, the fusion protein protein (NeuC), a central intein protein, and a C-terminal applied to a chitin column, the CBD binds to the resin, while all chitin-binding domain (CBD). When crude cell extracts are The principle, illustrated in Figure 4.1, begins with the expression of a three-domain protein: the N-terminal target 1,4-dithio-DL-threitol (DTT). The target protein is then eluted Figure 4.1 Intein-mediated affinity purification of NeuC. 119 The chemical cleavage event is based on a natural posttranslational processing event, known as protein splicing, that discards a central protein segment (i.e., the intein) of a larger precursor protein, allowing the N-terminal and C-terminal domains (i.e., the exteins) to be joined. Shown in Figure 4.2 is the proposed mechanism for protein splicing involving the intein from the V M A gene of Saccharomyces cerevisiae}1 Following an N-S acyl shift between the N-extein . A Cys1 O Intein Cys455 N-S Acyl Shift Asn454 Mr H NH 2 C-extein O N-extein o Cys1 HS. O H ?N Intein Cys455 Transthioesterifi cation Asn454 V O Ml H NH 2 C-extein HSv HoN' Cys1 N-extein i A Intein Cys455 Asn454 Peptide Cleavage and Succinimide Formation N-extein O A . Y N H NH 2 C-extein Cys455 HS. H,N Cys1 Intein H ?N HS. C-extein S-N Acyl Shift Asn454 Excised Intein o N-extein . A Cys455 C-extein Spliced Exteins Figure 4.2 Mechanism of protein splicing involving the intein from the Saccharomyces cerevisiae V M A 1 gene. 120 intein's N-terminal Cysl residue and the N-extein's amide carbonyl, transthioesterification with the intein's C-terminal Cys455 residue occurs. Irreversible peptide cleavage then occurs via cyclization of the adjacent Asn454 residue, and a final S-N acyl shift yields the modified protein composed of the spliced exteins. In the case of the plasmid supplied with the IMPACT system, however, the V M A gene has been modified to exclude Cys455, preventing intramolecular transthioesterifcation. Thus, introduction of DTT allows the S-acylated intein to be intercepted, resulting in N-terminal protein cleavage to liberate the target extein as a thioester, which spontaneously hydrolyzes to yield a free C-terminus (Figure 4.3). Additionally, the IMPACT system V M A gene has been modified with an Asn454Ala mutation to prevent cleavage between the intein and CBD, which would interfere with purification. Accordingly, we initiated steps toward the production of NeuC by transforming E. coli NeuC Cys1 Intein Ala454 C H 3 N-S acyl shift CBD Intein CBD >r N I H Ala454 C H 3 DTT-induced N-terminal cleavage H S ^ Cys1 H 2 N ' Intein Ala454 C H 3 CBD Figure 4.3 Proposed mechanism of DTT-induced cleavage of NeuC from the intein tag. 121 1 2 3 Res MW Figure 4.4 SDS-PAGE showing purification of NeuC. Res = resin sample after cleavage, M W = molecular weight markers (29 and 66 kDa). Fractions 1 and 2 were pooled and used in subsequent experiments. DH5a with pSR647. Since this plasmid also carries a gene for ampicillin resistance, cultures of the transformed cells were grown in the presence of ampicillin. After inducing protein expression by the addition of isopropyl 1-thio-P-D-galactopyranoside (IPTG), the cells were lysed, and the crude protein was applied to a chitin column. Following overnight incubation with DTT, the 45 kDa NeuC was collected and judged to be >95% pure by SDS-PAGE (Figure 4.4). The presence of only one band, corresponding to the 55 kDa intein-CBD tag, from a sample of the chitin resin (Figure 4.4, Res lane) indicates that the cleavage reaction was efficient. 4.1.2 NMR Assays with NeuC The enzymatic reaction of the affinity-tag purified NeuC was assayed by H - and P-N M R spectroscopy. Following buffer exchange into a sodium phosphate buffer in D2O, a 1 ^ 1 sample of the enzyme was incubated with 5 mM UDP-GlcNAc, and H - and P-NMR spectra were immediately acquired (Figure 4.5a and Figure 4.6a, respectively). After extensive incubation at 37 °C, a new singlet at 6.59 ppm in the 'H-NMR spectrum was observed (Figure 4.5b). This signal is consistent with the formation of 2-acetamidoglucal (Figure 4.5c), whereby the anomeric proton of UDP-GlcNAc at 5.40 ppm has been converted to a vinylic proton. Integration of the H - l signals indicates that 8% conversion had occurred over the course of 15 h, 122 a) t = 3 min HO' 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 b)t = 43.5 h J 1 J I ' 1—r—•—1—•—'—i—'—'—• • I—1—1—•—•—I—•—'—1—l—•—•—'—'—i '—•—'—i 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 c) Pure 2-acetamidoglucal HO-— k 8'.0 7'5 To a5~ 6.0 5!5 s!o" 45 ppm Figure 4.5 l H-NMR assay of NeuC activity. Spectra of the incubation with 5 mM UDP-GlcNAc for a) 3 min and b) 43.5 h. c) Spectrum of an authentic sample of 2-acetamidoglucal. The position of the anomeric proton (bold face) is indicated by the arrow. and 19% conversion had been achieved by 43.5 h. Repeated measurements confirmed that the rate of conversion was dependent on enzyme concentration and control experiments lacking enzyme showed no detectable formation of 2-acetamidoglucal under otherwise identical conditions. No signals attributable to ManNAc (5.02 ppm for H - l of the a-anomer and 4.92 ppm for H - l of the P-anomer153) were observed in the spectrum, presumably due to the inherent insensitivity of the technique (<1% product would not be detected by this method). Additionally, the spectrum revealed a new doublet at 7.87 ppm, adjacent to the doublet from H-6 of the uracil ring in the substrate, indicative of the formation of UDP. The 3 1 P - N M R spectrum 123 r-^ -~-~-r->-•—-*-»• ^ - . - . - . - . - ^ - i — — - ! i-...,...r - | . . , . , v , , ,—. , . • - 5 - 6 - 7 - 8 -9 -10 -11 -12 -13 c) Pure UDP -5 -7 -8 5 -7o -11 -12 -13 ppm Figure 4.6 3 I P-NMR assay of NeuC activity. Spectra of the incubation with 5 mM UDP-GlcNAc for a) 3 min and b) 43.5 h. c) Spectrum of an authentic sample of UDP. (Figure 4.6b) further supports this, as a pair of doublets appeared at -6.26 and -9.75 ppm, downfield from those of UDP-GlcNAc. A spectrum of UDP was obtained for comparison (Figure 4.6c). To further confirm that 2-acetamidoglucal had been formed, the enzyme solution was treated with an anion-exchange resin to remove all phosphate-containing species. The l H - N M R and mass spectra of the resulting material were in agreement with that of authentic 2-acetamidoglucal.171 The failure to detect ManNAc formation prompted us to explore the possibility that a regulatory molecule or cofactor is necessary for full activity. It has been shown that CMP-iV-124 acetylneuraminic acid (CMP-Neu5Ac) is a feedback inhibitor of the mammalian UDP-GlcNAc 2-epimerase.146 Therefore, a sample of NeuC in deuterated buffer was incubated with 5 m M UDP-GlcNAc in the presence of 1 mM CMP-Neu5Ac, and the reaction was monitored by N M R ; however, no difference in the products or extent of the reaction was detected when compared to a control lacking CMP-Neu5Ac. In a similar study, the enzyme was assayed in the presence of either 5 m M N A D + or 5 m M N A D P + and no detectable changes in activity were observed. Hence, an oxidative cofactor is not required, a finding that is not surprising since both UDP elimination and glycal hydration are redox neutral processes. These N M R studies clearly indicated that the major products from NeuC catalysis in vitro are 2-acetamidoglucal and UDP. Furthermore, our findings assisted in establishing the identity of the major radiolabeled product spot in Vann and coworkers' paper chromatogram (recall Figure 3.16, page 114) as 2-acetamidoglucal. They demonstrated this by co-migration with an authentic sample of 2-acetmidoglucal we provided and by using conditions that clearly distinguished this material from GlcNAc, which they hypothesized had formed from substrate hydrolysis. Supported by the sequence homology with the well-characterized non-hydrolyzing (RffE) and hydrolyzing (mammalian) UDP-GlcNAc 2-epimerases, it is clear that the function of NeuC is also as a hydrolyzing UDP-GlcNAc 2-epimerase. The complementation experiments by Vann indicate that this enzyme is required for sialic acid biosynthesis in vivo, and as such, it is expected that it converts UDP-GlcNAc to ManNAc and UDP. The inability to detect more than very low levels of ManNAc by the paper chromatography assay indicates that there is something missing in our in vitro model of the enzymatic reaction, a notion that is reinforced by the extremely slow rates of catalysis that were observed. As mentioned above, it is possible that a regulatory molecule or cofactor is necessary, but this role is not served by CMP-Neu5Ac, a known regulator of the mammalian enzyme. Alternatively, the complete conversion to ManNAc 125 could be fulfilled by a second enzyme such as a glycosidase, which are known to hydrate 172 glycals. A third possibility is that the cloning of the neuC gene by Vann and coworkers could have been performed erroneously. Because the only published sequence of this gene was from them,1 7 3 there is no way of knowing i f the cloned neuC used in subsequent studies is an accurate representation of the E. coli K l D N A without independently re-cloning and sequencing from these cells. As will be described in more detail in Section 4.6, this suggestion is strengthened by protein sequence alignments and mutagenesis of homologous enzymes. As these studies were drawing to a close and the publication of our results was in preparation,166 the difficulties encountered with NeuC inspired us to explore a homologue from another species of bacteria. Described in the remaining sections, therefore, are cloning and mechanistic studies on SiaA, the Neisseria meningitidis counterpart of NeuC. As will be made apparent, the opportunity to study this enzyme met with good fortune, allowing us to form a clearer picture of sialic acid biosynthesis in bacteria. 4.2 Cloning and Expression of siaA, a neuC Homologue from Neisseria meningitidis 4.2.1 Previous Work Done Among the many homologues of NeuC identified by Vann 1 6 6 and others by sequence alignments is the protein SiaA, found in N. meningitidis serogroup B. SiaA exhibits 52% similarity and 32% identity to NeuC. Because this organism produces a polysialic acid capsule that is identical to that of E. coli K l , 1 7 4 it is reasonable to expect SiaA to serve a role in sialic acid synthesis, likely in the formation of ManNAc. 126 0 3 P 0 v HO v - ^ ^ ° \ H O ^ ^ ° \ AcHN 0 H AcHN 0 H GlcNAc6P y* G lcNAc HO HO pyruvate ManNAc6P O H neuraminic acid ManNAc " O H aldolase C 0 2 H HO Neu5Ac Figure 4.7 Discontinuous coupled assay for ManNAc formation. After epimerization with SiaA, the phosphorylated sugars were hydrolyzed with phosphatase. Coupling of ManNAc with pyruvate using neuraminic acid aldolase provided Neu5Ac, which could be quantified by a colourimetric thiobarbituric acid assay.1 5 At the commencement of this project, however, the only report on the function of this enzyme concluded that it was an Af-acetylglucosamine-6-phosphate (GlcNAc6P) 2-epimerase.115 The researchers drew their conclusion from a discontinuous coupled assay using crude cell extracts and GlcNAc6P, as shown in Figure 4.7. Thus, the authors proposed that SiaA converts GlcNAc6P to ManNAc6P and that a second enzyme specifically hydrolyzes ManNAc6P to ManNAc, which becomes incorporated into sialic acid by subsequent condensation. Their approach, however, suffers from the fact that crude extracts were employed and that there was no direct spectroscopic evidence for the formation of ManNAc, only a colourimetric indicator of 3-deoxy-2-keto-aldonic acids. 1 7 5 Although we viewed their studies with some scepticism, we envisioned two possible mechanisms that could account for the interconversion of the two phosphorylated sugars. Recall from Section 3.1 that the open-chain form of GlcNAc6P possesses an acidified H-2 due to the adjacent aldehyde group, and therefore, a deprotonation/reprotonation mechanism, followed by ring closure, could yield the epimeric product (Figure 4.8, upper pathway). Alternatively, the sequence homology with NeuC indicates that a similar elimination could occur to form 2-127 AcHN \ deprotonation H AcHN 2 " 0 3 P O -H O ^ H O - . AcHN OH GlcNAc6P "0 3 PO HO HO-dehydration A c H N H rehydration 2 0 3 PO- NHAc H O ^ HO— OH ManNAc6P Figure 4.8 Possible mechanisms for epimerization of GlcNAc6P. The upper pathway depicts a deprotonation/reprotonation mechanism; the lower pathway invokes a dehydration/rehydration mechanism. acetamidoglucal (Figure 4.8, lower pathway). Rehydration with protonation on the opposite face would then give ManNAc6P. We desired to perform N M R assays similar to what had been used with NeuC in order to verify the activity of SiaA and to discriminate between the possible mechanisms. Therefore, the cloning of the siaA gene from N. meningitidis was undertaken. 4.2.2 Ligation Independent Cloning of siaA and Affinity Purification of Histidine-tagged SiaA As exemplified by the intein-tagged NeuC in Section 4.1.1, cloning a gene into a vector that expresses as an affinity fusion is a fast, convenient way of obtaining large quantities of purified proteins. Many such systems available commercially require the use of restriction enzymes to cut the gene of interest from the parent D N A and a ligase to join this insert to the vector, which must also be treated with the corresponding restriction enzymes. Novagen's ligation-independent cloning (LIC) system, however, does not require either of these steps, and 128 as a result, is much simpler and more efficient; in fact, this method has been used successfully in our laboratory to clone at least five genes, all of which yielded active enzymes with high levels of expression. The LIC method, illustrated in Figure 4.9, begins with the polymerase chain reaction (PCR) of the target gene using primers containing extensions that are complementary to the vector. To generate single-stranded overhangs for insertion into the vector, the PCR product is TARGET GENE 1 PCR 5'-GGTATTGAGGGTCGCATG-3'-CCATAACTCCCAGCGTAC-TARGET GENE •TAAGGCTCTAACTCTCCTCT-3' -ATTCCGAGATTGAGAGGAGA-5' 1 T4 DNA polymerase + dGTP only GGTATTGAGGGTCGCATG-GCGTAC-TARGET GENE • TAAGG • ATTCCGAGATTGAGAGGAGA TCC; AGGCCATAACTCCCA LIC Insert + LIC Vector CTCTAACTCTCCTCT ftfti Annealing ITC C G GTATTGAG GGTCGCATG" IAGGCCATAACTCCCAGCGTAC- TARGET GENE • TAAG G CTCTAACTCTC CTCTI • ATTCCGAGATTGAGAGGAGA1 Transformation, Repair TCCGGTATTGAGGGTCGCATG" AGGCCATAACTCCCAGCGTAC- TARGET GENE • TAAGGCTCTAACTCTCCTCT • ATTCCGAGATTGAGAGGAGA Recombinant Plasmid Figure 4.9 Ligation-independent cloning. The first G on each 3'-end is shown in red. Nicked D N A is indicated by red arrows. 129 treated with T4 D N A polymerase in the presence of dGTP (but not dATP, dCTP, or dTTP). This enzyme possesses two functions that are dependent on the reaction conditions: the polymerase activity assembles D N A in the presence of nucleotide triphosphates (dNTPs), whereas the 3' to 5' exonuclease activity digests D N A in their absence. Thus, the polymerase removes nucleotides from one strand of each end of the PCR product until it encounters a residue corresponding to the dGTP present in the reaction mixture. Similarly, the vector is treated with T4 D N A polymerase using dCTP to provide the complementary overhangs. The insert and vector then anneal, and the resulting nicked plasmid is transformed into competent E. coli cells. Repair enzymes within the host cells ligate the nicked strands to yield a covalently linked recombinant plasmid. The vector chosen for LIC, the pET30 Xa/LIC vector, contains a sequence encoding an N-terminal histidine tag (His-tag) and fluorescence tag (S-tag) upstream from the insertion site (Figure 4.10). When a plasmid formed from this vector is transformed into host cells, expression of a fusion protein results, whereby the two tags are attached to the recombinant protein of a ) thrombin Xa/LICsite T7 promoter > i  / I  His-tag W S-tag MCS His-tag b) ATG CAC CAT CAT CAT CAT CAT TCT TCT GGT CTG GTG CCA CGC GGT TCT GGT ATG Met|His His His His His His|Ser Ser Gly|Leu Val Pro Arg|Gly Ser|Gly Met His-tag thrombin cut site AAA GAA ACC GCT GCT GCT AAA TTC GAA CGC CAG CAC ATG GAC AGC CCA GAT CTG |l_ys Glu Thr Ala Ala Ala Lys Phe Glu Arg Gin His Met Asp Ser [Pro Asp Leu S-tag GGT ACC GGT GGT GGC TCC GGT ATT GAG GGT CGC , / c 7 Gly Thr Gly Gly Gly Ser G l y , l i e Glu Gly A r g | Xa cut site Figure 4.10 pET-30 Xa/LIC cloning/expression region, a) Vector schematic. Note that the region downstream from the LIC site, containing the multiple cloning site (MCS) and second His-tag, is not translated i f a stop codon is included in the gene insert (as is the case with siaA). b) Nucleotide and amino acid sequences of the region that is translated as an N-terminal tag. 130 interest. While the S-tag can optionally be used for protein detection and quantification, the His-tag is more commonly utilized for affinity purification. The imidazole groups of the six consecutive histidine residues chelate to column-bound nickel ions, allowing the protein to bind to the resin while impurities are washed through. After eluting with imidazole buffer, the protein can be used as is, or the His-tag and/or S-tag can be cleaved by the proteases thrombin or factor Xa. Accordingly, siaA was cloned using LIC from chromosomal D N A isolated from Neisseria meningitidis subgroup B (strain MC58). The presence of the gene in the resulting plasmid, pAM04, was verified by colony PCR (Figure 4.11) and by D N A sequencing.* The recombinant plasmid was then transformed into an expression cell line (BL21 (DE3) E. coli), 1 2 3 4 5 6 7 8 9 10 11 12 13 3 . 0 -2.0-1.5-1.0-Lane 2 3 4 5 6 7 8 9 10 11 12 13 Colony A A B B c c D D E E F F T7 + - + + - + - + + T7Rev - + - • - + - + • - • Figure 4.11 Agarose gel from colony PCR of NovaBlue E. coli cells transformed with pAM04. 1% agarose gel containing 0.5 ug mL"1 ethidium bromide, illuminated with U V light. Lane 1 = 1 kb markers (New England Biolabs); 3.0 kb, 2.0 kb, 1.5 kb, and 1.0 kb (too faint to see on photograph) markers are labelled. A l l other lanes were loaded with PCR mixtures from individual colonies (A - F) using A M I 102 (the gene-specific, reverse primer used in LIC; see Experimental Methods, Section 4.8.5) and either T7 (a universal, vector-specific forward primer) or T7Rev (a universal, vector-specific reverse primer), as indicated. For correct orientation of the gene in the plasmid, only PCR with A M I 102 and T7 would yield a D N A band of 1377 bp. Note that three independently published sequences are in agreement,17 but one is erroneous. 131 MW CL FT CW 1 2 3 4 5 MW 6 7 8 9 10 11 Figure 4.12 SDS-PAGE gels of affinity purification of 6xHis-SiaA. Selected molecular weight (MW) markers are labelled (kDa). C L = cell lysate, FT = flow-through, CW = column wash. Fractions were eluted with 2 column volumes containing imidazole concentrations as follows: 1, 2, 3 = 25 mM; 4, 5 = 50 mM; 6 = 75 mM; 7 = 100 mM; 8, 9 = 250 mM; 10, 11 = 500 mM. which overexpressed histidine-tagged SiaA (6xHis-SiaA), as seen in the SDS-PAGE of the cell lysate (CL; Figure 4.12). The protein was purified on a nickel-Sepharose® column using an increasing step gradient of imidazole buffer and was judged to be >95% pure by SDS-PAGE (Figure 4.12, fractions 10 and 11). To remove imidazole and to concentrate the protein, attempts were made to buffer exchange the pooled column fractions by centrifugation through a membrane (10 kDa MWCO). Unfortunately, this procedure resulted in a significant loss (>80%) in total protein, apparently caused by adsorption to the membrane. To circumvent this problem, another batch of 6xHis-SiaA was prepared, dialyzed against an imidazole-free buffer, flash-frozen, and lyophilized. The resulting dry protein was either reconstituted for immediate use by addition of water or stored at -20 °C, where it remained stable for at least three months, as determined by kinetic assays (vide infra). Using these procedures, >10 mg of pure protein was typically obtained from 1 litre of cell culture. 132 4.2.3 Identification of the Substrate With purified 6xHis-SiaA in hand, the activity of the enzyme was determined by a ' H -N M R assay using 7V-acetylglucosamine 6-phosphate in a deuterated buffer. Because there was some doubt as to the reported GlcNAc6P 2-epimerase activity, GlcNAc and UDP-GlcNAc were also separately incubated with the enzyme. It was observed that within 30 min at 37 °C, only UDP-GlcNAc underwent reaction; in fact, neither GlcNAc nor GlcNAc6P showed any degree of reaction after 2 days. As depicted in Figure 4.13, the H - l " signal of UDP-GlcNAc at 5.4 ppm had completely disappeared, while two new signals at 5.0 and 4.9 ppm appeared, consistent with the H - l signals of the a- and (3-anomers of ManNAc, respectively.153 Unlike NeuC but similar to the mammalian enzyme, no 2-acetamidoglucal was observed during this process. A 3 1 P - N M R spectrum (not shown) also indicated the complete conversion from UDP-GlcNAc to free UDP. This information was sufficient to conclude that the activity of SiaA is in fact as a a)t = 0 11111111111111111 M 11111111111111111 II 111 5.4 5.3 5.2 5.1 5.0 4.9 4.8 4.7 ppm Figure 4.13 'H-NMR assay of SiaA activity. Spectra showing signals of the anomeric protons (bold face) from the incubation of 5 mM UDP-GlcNAc with SiaA in deuterated buffer at a) t = 0 and b) t = 30 min. Solvent-derived deuterium is in red. Assignment of the product anomers was based on literature precedence.153 133 hydrolyzing UDP-GlcNAc 2-epimerase, in contrast to previous reports.115' 1 7 6 Thus, having established that SiaA catalyzes this reaction to completion, unlike its E. coli counterpart, NeuC, we had successfully identified the first bacterial equivalent of the mammalian epimerase that was functionally competent in vitro. With this in mind, a series of experiments similar to those used to characterize the mammalian enzyme1 5 3 was undertaken. The remainder of this chapter details these experiments, as well as two novel kinetic isotope effect studies in Sections 4.4.2 and 4.4.3. 4.3 Characterization of SiaA Activity 4.3.1 Characterization of the Product: Stereochemistry and Isotope Incorporation To establish the stereochemistry of the reaction, it was necessary to determine which anomer (a or P) of ManNAc is the initially formed product. Complicating the analysis, however, is the fact that the two anomers readily, non-enzymatically interconvert via mutarotation, leading to an approximately equal proportion of each isomer (Figure 4.13b; Keq = 1.1 in favour of the a-anomer at 25 °C). Additionally, it was observed that this process is buffer-catalyzed by dissolving a sample of solid ManNAc (predominantly the (3-isomer) in either D2O or 50 mM deuterated sodium phosphate, pD 8.0. After 3 min, only 4% a-anomer existed in the unbuffered solution, as determined by ' H - N M R spectroscopy, whereas an equilibrium mixture (i.e., 52% a) was obtained in the buffered sample. Therefore, in order to identify the first formed anomer before substantial mutarotation could occur, the buffer concentration was reduced to 10 mM. As can be seen in Figure 4.14, it was revealed that the first-formed product in the enzymatic reaction is a-ManNAc. As the reaction continued, mutarotation produced (3-134 a ) t = 0 b ) t = 1.5 c ) t = 4 i i i i i i i i i i i i i i i i i i _ 1 _ i i i i | i i i i [ i i 5.5 5.4 5.3 5.2 5.1 5.0 4.9 ppm Figure 4.14 Enzymatic formation of a-(2-2H)ManNAc in D 2 0 by 'H-NMR spectroscopy. a) Before addition of enzyme, b) Within the first 90 s, only the a-anomer was apparent, but c) after 4 min, the P-anomer had begun to form as a result of non-enzymatic mutarotation. Signals from the anomeric protons (bold face) are indicated. Solvent-derived deuterium is in red. ManNAc, eventually leading to an equilibrium mixture of the two anomers (not shown). Thus, the stereochemistry of the reaction involves a net retention of configuration at C - l . Furthermore, the N M R experiment also indicated that the reaction proceeds with >97% incorporation of solvent-derived deuterium at C-2. This finding was determined by the absence of H-2 signals from ManNAc (not shown) and by the appearance of the H - l signals (Figure 4.14). The anomeric proton signal of UDP-GlcNAc appears as a doublet of doublets due to coupling to H-2" and the P-phosphorus of UDP. When UDP departs and deuterium becomes incorporated at C-2, the anomeric proton signals appear as singlets due to negligible J H I , D 2 values and to the absence of phosphorus. These observations imply that the inversion of stereochemistry involves deprotonation at C-2, followed by reprotonation in the opposite configuration with a solvent-derived proton. The lack of observable internal return of the 135 original protium of UDP-GlcNAc into the product is consistent with a two-base mechanism (recall discussion of amino acid racemases in Section 1.3) in which distinct enzymatic residues serve as the catalytic base and acid. The H - l " splitting pattern in the UDP-GlcNAc remaining after 75% conversion appeared unchanged from the initial material, indicating that no solvent-derived proton is incorporated into the substrate. This suggests either that the deprotonation of substrate involves an irreversible step or that the enzymatic residues in the two-base mechanism do not exchange protons with solvent during the lifetime of the intermediates. 4.3.2 Catalytic Competence of 2-Acetamidoglucal and Possible Alternative Substrates The results and interpretations presented in the previous section are entirely consistent with those of Chou et al. for the mammalian enzyme.1 5 3 Their work additionally implicated 2-acetamidoglucal as an intermediate by incubating this compound with the enzyme and following the H - l signals by ' H - N M R spectroscopy. They observed slow, partial turnover to ManNAc (6% in 29 h) and slightly increased conversion (12%) in the presence of UDP. Thus, a synthetic sample of 2-acetamidoglucal was tested with 6xHis-SiaA in a similar fashion. In the presence of 5 m M UDP, 37% conversion occurred in 4.5 h at 37 °C (Figure 4.15), showing that hydration of the glycal, although slower than the full reaction with the natural substrate, is a catalytically competent process. The rate and extent of this reaction are much greater for SiaA than for the mammalian enzyme; in fact, extended incubation (>2 days) led to complete conversion to ManNAc. Also differing from the mammalian enzyme is that UDP is absolutely obligatory for catalytic turnover; no reaction was observable in its absence, even after 4 days at 37 °C. This 136 a)t = 0 I ppm Figure 4.15 Enzymatic conversion of 2-acetamidoglucal to (2-2H)ManNAc in the presence of UDP. ' H - N M R spectra of the incubation of 2-acetamidoglucal and UDP with 6xHis-SiaA at a) t = 0 and b) t = 4.5 h. observation suggests that the hydration step may proceed while UDP is still bound in the epimerase's active site. In view of the overall reaction, therefore, one may conclude that the chemical steps are identical to those of the mammalian epimerase. Thus, in terms of stereochemistry, SiaA catalyzes an arcri-elimination of UDP, followed by a syrc-addition of water. The ability of the enzyme to catalyze elimination and addition reactions on opposite faces of the glycal intermediate additionally bears some similarity to the non-hydrolyzing UDP-GlcNAc 2-epimerase, RffE, which adds UDP instead of a molecule of water. The ability of the enzyme to convert the glycal to product indicates that the enzyme is capable of binding and turning over compounds other than the natural substrate, UDP-GlcNAc. 137 Granted that the glycal may be a special case since it is by all accounts also the natural intermediate, it is conceivable that the enzyme could accept other compounds. For instance, glycosyl fluoride 69 introduces a leaving group at the anomeric carbon, which may serve as an adequate replacement for UDP, and thus, the compound may function as a substrate mimic. The oxazoline of ManNAc (70), on the other hand, may be an intermediate in the hydration step, and therefore, it may be enzymatically hydrolyzed directly to ManNAc. These compounds had already been prepared by Wayne Chou in our lab according to literature procedures. 1 8 0 ' 1 8 1 Thus, they were separately incubated with SiaA at 37 °C either in the presence or absence of UDP. In control samples lacking enzyme, 69 and 70 were found to undergo slow hydrolysis (t\a ~ 24 h) to GlcNAc and ManNAc, respectively. With SiaA, no elevated levels of hydrolysis were observed even in the presence of large quantities of the epimerase or when UDP was added. 4.3.3 Kinetic Characterization of SiaA by a Continuous Coupled Assay To quantify the activity of SiaA kinetically, a coupled assay dependent on the production of UDP was employed (Figure 4.16). This assay, previously used with the mammalian epimerase,153 indirectly measures UDP formation by first reacting it with phosphoenol pyruvate (PEP) under the direction of pyruvate kinase. The resulting pyruvate is then reduced by one equivalent of N A D H using lactate dehydrogenase. The consumption of N A D H ( £ 3 4 0 = 6220 M " 1 HO HO HO ^ -AcHN I U D p SiaA ManNAc + UDP UDP UTP " 0 2 C . . O P 0 3 2 - V y 0?C Pyruvate kinase NADH NAD + .0 V / "°2C> 138 PEP Lactate C H 3 dehydrogenase C H 3 pyruvate lactate Figure 4.16 Coupled assay used to measure the rate of formation of UDP. cm'1) is then monitored spectrophotometrically by recording the decrease in absorbance at 340 nm. The substrate concentration was varied between 0.20 and 10.0 mM, and the initial reaction rates were measured in sodium phosphate buffer at pH 7.5 (Figure 4.17). The plot of initial reaction velocity (v) against [UDP-GlcNAc] is sigmoidal, which is indicative of an enzyme that is allosterically regulated by its substrate. In order to interpret these data, therefore, 20 c "E 1 1 1 1 — i 1 <i> /f c a t = 4.7 ± 0.5 s"1 K'= 1.49 ± 0 . 1 2 mM kcJK' = (3.2 ± 0.4) x 10 3 M" 1 s"1 napp= 1.94 ± 0 . 1 4 _L 4.0 6.0 8.0 [UDP-GlcNAc] (mM) 10.0 Figure 4.17 Enzyme kinetics for 6xHis-SiaA. Initial reaction rate (v) vs. [UDP-GlcNAc]. Kinetic parameters as determined by fitting the data to the Hi l l equation are indicated. a brief description of the theory underlying allostery and cooperativity is warranted, of the kinetic results follows on page 142. 139 Discussion Allostery and Positive Cooperativity Many enzymes are multimeric and possess multiple substrate-binding sites. In the case of non-regulatory multimeric enzymes, each subunit operates independently, such that substrate binding and catalysis in one site has no effect on other subunits. Thus, these enzymes behave kinetically as though they were monomeric, exhibiting hyperbolic v versus [S] plots (Figure 4.18, curve a), hallmarks of classical Michaelis-Menten kinetics governed by Equation 4.1. v [S] — — — 4.1 If the binding of a substrate molecule in one site affects the binding of subsequent molecules, however, the system exhibits cooperativity, and the substrate is called an allosteric modulator.* Specifically, in positive cooperativity, the allosteric activator increases the affinity of the enzyme for additional substrate molecules. The effect of positive cooperativity is a Figure 4.18 Typical rate curves for two different enzymes with the same VmAx. a) Hyperbolic curve from a non-allosteric enzyme and b) sigmoidal curve from an allosteric enzyme. * The analysis here is restricted to substrates simply for convenience; however, any ligand that exhibits an effect, either positive or negative, on the binding of another ligand (which may or may not be another molecule of the same ligand) can be considered an allosteric modulator. 140 deviation from Michaelis-Menten kinetics, as expressed in sigmoidal v versus [S] plots (Figure 4.18, curve b). The degree of sigmoidicity in these curves can provide insight as to the number of binding sites and degree of cooperativity between them. To understand how the mechanism of binding relates to the curve shape, it is convenient to use models, the most common of which are the concerted or Monod-Wyman-Changeux (MWC) model 1 8 2 and the sequential or Koshland-Nemethy-Filmer (KNF) model. 1 8 3 The M W C model, originally designed to explain the binding of oxygen by hemoglobin but can be applied to other allosteric proteins including enzymes as described here, is based on the following assumptions: 1. The enzyme is oligomeric. 2. The enzyme exists in one of two conformational states: T (tense), the inactive form, which is predominant when no substrate is bound, or R (relaxed), the active form. The two states are in equilibrium and differ in the energies and numbers of bonds between subunits. 3. The T state has a lower affinity for the substrate than the R state. 4. A l l binding sites in each state are equivalent and have identical binding constants (i.e., the symmetry assumption). Thus, the transition between one conformation and another is an all-or-none event. The schematic form of this model is depicted in Figure 4.19 using a tetrameric protein as an example. In the absence of substrate, the equilibrium strongly favours the inactive T state, such that the initial ratio of the states, L = [T] 0/[R] 0 » 1. With the addition of low amounts of substrate, the substrate will bind preferentially to the R state molecules, and by Le Chatelier's principle, this will pull more enzyme molecules from T to R. At these low substrate concentrations, however, there is little opportunity for the R states to encounter another substrate molecule nor for many R state molecules to be formed; hence, the majority of enzyme molecules exists in the inactive state, resulting in slow initial reaction rates (see early stages of sigmoidal plot, Figure 4.18b). As the substrate concentration is increased, however, enzyme molecules have a much greater chance of binding additional substrate molecules, so the higher R states 141 T states T 1 S S s CL\ (§) (D T 3 T 4 Rr R 2 R states c3L ^ o o R 3 c 4 L _ (SXS) R 4 Figure 4.19 Monod-Wyman-Changeux (MWC) model for the binding of substrate molecules to a tetrameric protein. • = vacant subunit, O = occupied subunit, T = tense conformation, R = relaxed conformation, S = substrate, Kj and KR = dissociation constants of the T and R states, resp., L = [T0]/[Ro] (initial equilibrium ratio of states), c = K^IKj. Note 1) that all four protein subunits share the same conformation at any time (symmetry assumption) and 2) that the equilibrium ratio of each T-R pair is multiplied by the ratio of the dissociation constants, c, causing the relative population of the R states to increase with each substrate molecule bound. become populated, and the T states decrease accordingly. Thus, a slow increase in activity is observed with increasing [S], resulting in a rise in the slope of the rate curve (see middle stages of sigmoidal plot). Eventually, the enzyme becomes saturated with substrate and a plateau in the reaction rate is approached, as in the case of Michaelis-Menten kinetics (see late stages of both plots). The K N F model offers an alternative view that avoids the symmetry assumption of the M W C model. Instead, it assumes that the progress from T to the substrate-bound R state is a sequential process. Specifically, the assumptions are: 1. In the absence of substrate, all subunits exist in one conformation (A). 2. Upon binding, the substrate induces a conformational change (from A to B) in the subunit to which it is bound. This change may be transferred to the adjacent empty subunits, thereby influencing the binding constants for subsequent substrate molecules. The schematic form of the K N F model is illustrated in Figure 4.20. Because this model does not assume an all-or-none event, the dissociation constant for the first substrate molecule 142 A4 A 3 B A 2 B 2 A B 3 B 4 Figure 4.20 Koshland-Nemethy-Filmer (KNF) model for the binding of substrate molecules to a tetrameric protein. • = vacant subunit with conformation A, O = occupied subunit with conformation B. Note that all four protein subunits need not possess the same conformation at any time, as is the case in the M W C model. may be greater than that for a second substrate molecule, and so forth. Although the lack of symmetry makes the mathematical representation of the binding events more complex, the KNF model is more general and is often considered to provide a better description of some proteins than the M W C model. 1 8 4 Regardless of the model chosen, it is possible to relate catalysis to substrate binding in an allosteric enzyme by means of the Hi l l equation (Equation 4.2): y = [sy r m a v K'+[sy 4.2 where n is the number of substrate binding sites and K is a constant comprising interaction factors and intrinsic dissociation constants (i.e., an apparent overall binding constant). The derivation of this equation can be found elsewhere.65 When experimental data are fitted to this equation, the calculated value of n (often denoted as napp and referred to as the Hi l l coefficient) is nearly always less than the actual number of binding sites. The next higher integer above this « a p p value represents the minimum number of actual sites. For instance, i f a Hi l l coefficient of 1.8 is obtained, it is likely that the enzyme possesses two sites with relatively strong cooperativity (though it is also possible that the enzyme possesses more than two sites with poor cooperativity). With the realization that SiaA is an allosteric enzyme, the kinetic data were fitted to the Hi l l equation, and the resulting kinetic parameters are indicated in Figure 4.17 (page 138). The 143 value of the Hil l coefficient was found to be 1.9, which is consistent with the presence of two UDP-GlcNAc binding sites exhibiting strong cooperativity. This value compares to the Hi l l coefficient of the homologous UDP-GlcNAc 2-epimerase RffE. Because RffE is dimeric, with one substrate binding site per monomer, it is likely that SiaA is also dimeric. The allosteric regulation of SiaA by its substrate marks a distinction from the mammalian enzyme, which exhibits negative cooperativity («app = 0.45).1 4 6 In the case of negative allostery, the enzyme displays essentially normal activity at low substrate concentrations but exhibits marked substrate inhibition at high concentrations. The mammalian epimerase is additionally feedback inhibited by another allosteric modulator, CMP-Neu5Ac, a product further along the sialic acid biosynthetic pathway. It has been suggested that these two sources of negative allostery are important in directing the fate of UDP-GlcNAc by ensuring that the amounts required for oligosaccharide biosynthesis are not diverted to sialic acid biosynthesis.146 In contrast, because SiaA displays positive allostery, under higher concentrations of UDP-GlcNAc, the synthesis of PSA is promoted. It appears that the sialic acid needs of N. meningitidis and other bacteria that synthesize PSA differ from those of mammals. To determine the pH for optimal activity, the coupled assay was performed over a range of pH values at a high concentration (7.5 mM) of UDP-GlcNAc. Additionally, because phosphate is a poor buffer at pH values above 8, Tris was substituted in this region. The pH-rate profile shown in Figure 4.21 reveals two factors regarding the epimerase activity: first, optimal activity, given by the curve maximum, was at pH -8.6. Second, the discontinuity in the profile in the region of overlap between the two curves (pH 7.6 to 8.0) indicates that there is a substantial buffer effect on reaction rate. It is possible that this is the result of weak inhibition by phosphate. 144 Figure 4.21 pH and buffer dependence of activity. Initial reaction rate (v) divided by total enzyme concentration ([E]T) was determined in NaH2P04 buffer (o) and in Tris buffer (•) as indicated. pH values of Tris solutions have been corrected to account for the known temperature dependence of its pKa.lS5'186 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0 PH 4.4 Kinetic Isotope Effect Studies It is possible to gain mechanistic information by comparing the kinetic parameters (usually Vmax and Vmax/Km) for an unlabelled substrate with those of a substrate substituted with a heavier isotope at a position that is involved in the rate-limiting step of a reaction. Such kinetic isotope effects (KIEs), as they are termed, are the result of differences in the kinetic barrier to catalysis brought about by differences in bond strength involving the isotope of interest. In particular, a primary deuterium KIE is observed when the C - H bond that is broken in the rate-limiting step is replaced by a C - D bond. The greater mass of deuterium over protium gives the C - D bond a lower vibrational zero-point energy (Figure 4.22). The result is an increased activation energy needed to remove the deuterium and reach the transition state, which is identical in energy for both isotopologues. * As defined by IUPAC, isotopologues are molecular entities that differ only in isotopic composition (number of isotopic substitutions).'87 This term is in contrast to isotopomers, which are isomers having the same number of each isotopic atom but differing in their positions. "Isotopomer" is frequently used incorrectly in the literature to describe an isotopologue. 1 8 8 - 1 9 0 145 E [ C L B]* Figure 4.22 Reaction coordinate diagram illustrating the origin of primary kinetic isotope effects. A E H , A E D = activation energy required to break a C - H or C - D bond, respectively. Zero-point and transition state energy levels are indicated. B = enzymatic base; L = H or D. Reaction coordinate The N M R studies in the previous section support a mechanism involving removal of the proton at C-2" of UDP-GlcNAc in the formation of the intermediate 2-acetamidoglucal. If this C - H bond is broken in a rate-limiting step, it may be possible to observe a primary KIE by comparing the kinetics of UDP-GlcNAc with protium or deuterium at C-2". The next two subsections describe experiments designed to probe KIEs on Fi^x and Vmax/Km, respectively, for this reaction, while the third subsection examines hydration of the glycal. 4.4.1 Kinetic Isotope Effect on kcat To explore the possibility of a KIE on the catalytic rate constant, £ c a t , the two isotopologues (2"- 1H)UDP-GlcNAc and (2"- 2H)UDP-GlcNAc could be treated with the epimerase at saturating concentrations. The deuterium KIE on kcat, symbolized DV, is simply calculated as the ratio of the rate constants as given by: 4.3 (2"-2H)UDP-ManNAcUA Separate by ion-exchange chromatography on DE52 Figure 4.23 Enzymatic synthesis of (2"-2H)UDP-GlcNAc. U D P D H = UDP-ManNAc dehydrogenase. Deuterated UDP-GlcNAc was synthesized enzymatically from the protiated compound as described by Morgan et al. (Figure 4.23). 1 4 4 Following treatment with the bacterial epimerase RffE in deuterated buffer, an equilibrium mixture (1:9) of (2"- 2H)UDP-ManNAc and (2"-2 H)UDP-GlcNAc was obtained. Because these epimers have similar chromatographic properties, it was necessary to remove UDP-ManNAc enzymatically. Thus, after heat inactivation and removal of the epimerase by filtration, this epimer was specifically oxidized to UDP-7V-acetylmannosaminuronic acid (UDP-ManNAcUA) using UDP-ManNAc dehydrogenase (UDPDH), and the resulting mixture was separated by ion-exchange chromatography. The purified sugar was desalted by size-exclusion chromatography and exchanged to the sodium salt by passage through Amberlite resin. The deuterated substrate was obtained in 43% yield and shown to have incorporated 2 H at the desired position to >97% by ' H - N M R spectroscopy and mass spectrometry. 147 The two isotopologues were then individually subjected to the coupled assay at 7.5 m M (determined spectrophotometrically using £262 = 9890 M " 1 cm"1). From the average of three measurements, DV was found to be 1.15 ± 0.04. Because primary deuterium KIEs typically range from 2 to 10, the results indicate an absence of a KIE. This is in agreement with the findings of Chou et al. for the mammalian epimerase, wherein a DV of 1.0 ± 0.1 was determined.153 Thus, cleavage of the C - H bond is not a rate-limiting step in the reaction mechanism. 4.4.2 Kinetic Isotope Effect on kcJKm by Intermolecular Competition An alternative analysis is to measure a KIE on the second-order rate constant, kcJKm, symbolized D(V/K). Whereas a direct comparison method was employed above to determine DV, an intermolecular competition experiment can be used to determine D(V/K). In this case, a mixture of two isotopologues of known isotopic composition is treated with the enzyme. After a certain fractional conversion to products has occurred, the isotopic composition of the recovered starting material is determined. If a KIE exists, the faster reacting species containing the lighter isotope will be depleted to a greater extent than the slower species containing the heavier isotope. By accurate measurement of the extent of reaction and of the initial and final isotopic ratios of the substrate, it is possible to calculate D(VIK) using the equation derived by Melander and Saunders:188 4.4 148 where Fu is the fractional conversion of the protiated species to products, and R and Ro are, respectively, the final and initial ratios of deuterated substrate molecules, a®, to protiated substrate molecules, au, as given in Equation 4.5: It is instructive to view a graphical representation of Equation 4.6, which is the exponential form of Equation 4.4, as depicted in Figure 4.24. It can be seen that for primary isotope effects, which are always greater than 1, the isotopic enrichment with the heavier isotope in the remaining starting material increases exponentially as the reaction progresses. As F H is brought closer to unity, R/Ro begins to rise sharply. Thus, it is especially beneficial to allow the reaction to proceed further for smaller KIEs in order to maximize the vertical spread of the curves and thereby determine the isotope effect accurately. 1 8 8 ' 1 9 1 4.5 D - 1 (VIK) 4.6 10 RIRo 5 0 0 0.5 Figure 4.24 Relative change in the ratio of deuterated substrate to protiated substrate as a function of the extent of reaction. D(V/K) is indicated for each curve. 149 It is important to note that Equation 4.4 is valid only under irreversible reaction conditions. This has been found to be the case for the hydrolyzing reaction catalyzed by SiaA, since no UDP-GlcNAc could be detected by N M R or mass spectral methods following extended incubations with the products, ManNAc and UDP. A n equimolar (based on A2ei) solution of (2"- 1H)-UDPGlcNAc and (2"- 2H)-UDPGlcNAc was prepared in D2O, and ! H - N M R and mass spectra were acquired. The epimerase in 40 m M sodium phosphate buffer, pD 8.0, was then introduced to this mixture, and reaction progress was monitored by ' H - N M R spectroscopy. After >80% completion, the enzyme was inactivated with dilute HC1, and following neutralization with NaOH, another set of ' H - N M R and mass spectra was obtained. The mass spectra were used to determine the initial (RQ) and final (R) ratios of the deuterated (m/z 607 M-H*) and protiated (m/z 606 M-H*") substrates as follows. In order to relate the observed intensities of the mass spectral peaks (ko6, koi, and z^ os) to the relative number of molecules of the two isotopologues, an and ao, it was necessary to account for the natural abundance of heavier isotopes in the two molecules using Equations 4.7-4.9: '606 = « H X ( { ~ P ) 4 J i607 = aHxP + aDx(l-P) 4.8 *608 = aD X P 4.9 where P is the fraction of molecules containing a 1-amu heavier isotope. Solving the system of three equations in three variables allows an and ao to be calculated (Equations 4.10 and 4.11). aH 2*606 + *607 V^607 4/ 6 0 6j ' 6 0 8 4.10 *607 + 2*608 "\r607 4*606*6 n _ - - • ^ _ 4.11 " D 2 Thus, by determining an and a D , RQ and R were calculated from the mean of five measurements using Equation 4.5. 150 The N M R spectra allowed for the determination of the fraction of remaining (2"-^ U D P - G l c N A c , 1 -FH, as follows. The initial (S0) and final (S) ratios of the H - l " signals (5 5.38 ppm) to the uracil H-6 signal* (8 7.82 ppm) were each calculated based on the mean of three integrations. Because the H - l " signals contained contributions from both isotopologues, it was necessary to adjust S and So using R and Ro. Thus, the fraction of protiated substrate was calculated using Equation 4.12. 1 8 8 It is noteworthy that this adjustment is frequently omitted in , _ F l I . i . x l ± i 4.12 H S0 l + R the literature136' 1 9 1 ' 1 9 2 because one of the isotopologues is usually present in trace amounts, enabling to be determined directly. Finally, the KIE on kcJKm was calculated using Equation 4.4. A summary of the calculations from two independent experiments is given in Table 4.1. An average D(V/K) of 1.26 ± 0 . 1 5 was determined. In agreement with the °V of 1.15 ± 0.04 described in the previous subsection, this indicates that proton removal is not involved in a cleanly rate-limiting step. Table 4.1 Summary of °(V/K) calculations So S Ro R RMo 1 - F H °(VIK) 0.929 ± 0.006 0.168 ±0.004 0.992 ±0.017 1.36 ±0.25 1.38 ±0.25 0.149 ±0.016 1.20 ±0.15 0.934 ± 0.006 0.399 ± 0.005 0.84 ± 0.06 1.074 ±0.027 1.27 ±0 .10 0.379 ±0.015 1.33 ± 0.16 A V E R A G E 1.26 ± 0 . 1 5 4.4.3 Solvent Isotope Discrimination During Glycal Hydration Having established that formation of the glycal intermediate is not a rate-limiting step in the overall reaction, we examined the possibility that the subsequent step, namely hydration of the glycal, is rate limiting. In contrast to the elimination step, however, a C - H bond is formed, * This signal was used as an internal standard for integration since the response from this proton in the substrates and product UDP is assumed to be unchanged throughout the reaction. 151 rather than broken, so it is not possible to use the same forms of primary KIE experiments. Under such circumstances, one approach that is commonly used is to measure a secondary KIE. A secondary KIE is exhibited when a change in hybridization of an isotopically labelled atom occurs during the rate-limiting step. To test for such an effect, one could conceivably introduce a deuterium label at the anomeric position of UDP-GlcNAc or 2-acetamidoglucal, since this undergoes rehybridization from sp2 to sp3 upon hydration. Several factors may complicate such an approach, however. First, the spectrophotometric coupled assay employed above for labelled UDP-GlcNAc cannot be used since this assay is dependent on UDP formation rather than ManNAc. Second, the unprecedented synthesis of ( l"- 2 H)UDP-GlcNAc or (l- 2H)-2-acetamidoglucal would need to be designed and carried out. Third, in the case of labelled UDP-GlcNAc, the anomeric position undergoes two opposing changes in hybridization (sp —>sp , then sp2—>sp3) throughout the conversion to ManNAc, a process that may mask the KIE on the hydration step. Finally, in the case of labelled glycal, incubation of the enzyme with the putative intermediate does not truly reflect the normal reaction, and accordingly, non-chemical steps (i.e., binding or conformational changes) may become rate limiting (the ' H - N M R assay described in Section 4.3.2 indicates this is the case, as turnover of 2-acetamidoglucal to ManNAc was significantly slower). In light of the drawbacks with attempting to measure a secondary KIE, a preferable alternative is to measure a solvent isotope effect, which could derive from the exchange of an enzyme-bound, solvent-derived isotope to the unlabelled glycal intermediate. One method to measure such a KIE is to determine the relative rates in a buffer prepared in D2O versus the same buffer in H 2 0 . A potential problem with this approach, however, is that because all exchangeable protons in the enzyme would become deuterated, steps other than the chemical step of interest may also exhibit kinetic isotope effects. To avoid this possible complication, we 152 decided to perform a solvent isotope discrimination experiment that probes the transfer of the solvent-derived proton that becomes incorporated at C-2 of the ManNAc product. The isotope discrimination approach stems from fractionation factor theory, which describes the relative tendency for a particular exchangeable site to bind protium versus deuterium. As outlined by Albery, 1 9 3 the fractionation factor, (/>, represents the equilibrium composition of the X - L site (L signifies either H or D) with respect to the isotopic composition of the solvent: X - H + 1/2 D 2 0 X - D + 1/2 H 2 0 4.13 [ X - H ] [ D ! o ] * [ X - H ] n V2 4 1 4 where n is the atom fraction of deuterium in the solvent. Albery points out that the rightmost factor in Equation 4.14 is a small correction that accounts for the marginal difference between the theoretical value of 4 for the equilibrium constant, KL0, for the H 2 0 - H D O - D 2 0 equilibrium, H 2 0 + D 2 0 2 H D O KLi0 4 1 5 and the experimentally determined value of 3.78. 1 9 4 Gold 1 9 5 and Kresge 1 9 6 have shown that by applying fractionation factors to transition state theory, it is possible to derive an equation with which one can calculate the solvent KIE. For a general-acid-catalyzed enzymatic reaction, the conversion of the enzyme-substrate complex (in our case, the substrate is the glycal intermediate) to the transition state can be described by the hypothetical equilibrium, K* ± [ B - L S] =^=^= [ B — L — S ] • [B L-P] 4.16 where B is the enzymatic site (i.e., the general acid), L is the proton/deuteron being transferred, S is the substrate site, P is the product site, and K* is the equilibrium constant for formation of the transition state. The rate constant, k, is related to the equilibrium constant, K\ by the Eyring equation: 153 k = K W K * 4 . 1 7 h h [ B - L S] where K = transmission coefficient, ke = Boltzmann constant, h = Planck's constant, and T = temperature. Recognizing that the feT/h factor is independent of the isotope being transferred and assuming that KIS not subject to an isotope effect, the solvent KIE, D 2°A: h y d , can be written in terms of fractionation factors: kBT [ B - - H - - S ] * [ B - D S] k H _ _ h [ B - H S] _ [ B - H S] _ ^ K D.O . Z1L = *" " ~ J + = ^—: -L- = ^- 4.18 hyd kD kBT [ B - - D - - S ] t [ B - - D - - S ] t <t>% h [ B - D S] [ B - - H - - S ] 1 where <j> and $ are the ground-state and transition-state fractionation factors. Thus, it is possible to calculate D 2 ° £ h y d i f ^ and $ can be determined. To determine the relevant fractionation factors involved in the hydration of the glycal, we designed a solvent isotope discrimination experiment (Figure 4.25). In this experiment, it is assumed that the ground-state fractionation factor, <f>, is approximately 1, since this is the case for HO- AcHN (2-H)ManNAc Enz Figure 4.25 Solvent isotope discrimination experiment. In 50% D2O, the glycal has a statistically equal opportunity to abstract a proton or a deuteron from the general acid. If a KIE were to exist for this step, however, there would be a greater tendency to form protiated ManNAc. 154 most amino acid residues.197 Hence, by reconstituting the enzyme in a 50:50 mixture of D 2 0 and H2O, approximately half of all enzyme molecules should possess deuterium on the general acid responsible for providing the proton that becomes H-2 of ManNAc. This supposition is valid for all residues except lysine, which is polyprotic, and cysteine, for which <f>& 0.5. 1 9 7 Consequently, if protonation of C-2 is subject to an isotope effect (i.e., involved in a rate-limiting step), there would be a greater tendency for the glycal to accept a proton over a deuteron, an occurrence that would be reflected in the transition-state fractionation factor, Because the hydration reaction is irreversible and involves the delivery of a solvent-derived proton/deuteron onto a non-labile site in the ManNAc product, $ can be evaluated directly from the deuterium content of the recovered ManNAc using Equation 4.14. 1 9 3 ' 1 9 5 It should be noted that this approach does not differentiate between a stepwise or concerted mechanism, nor does it exclude the possibility that more than one enzymatic residue contributes to any observed isotope discrimination. In the latter case, the calculated $ would actually represent the product of the individual fractionation factors involved in the transition state. Accordingly, two such experiments were performed, one using UDP-GlcNAc as the substrate, the other using 2-acetamidoglucal as the substrate in the presence of UDP. After incubation at 37 °C for 40 min and 7.5 h, respectively, UDP was removed by the addition of ion-exchange resin, and the resulting ManNAc samples were submitted for mass spectral analysis. Using equations equivalent to 4.10 and 4.11, the ratio of (2- 2H)ManNAc to (2-*H)ManNAc was calculated to be 0.806 for UDP-GlcNAc and 0.794 for 2-acetamidoglucal. As mentioned above and using Equation 4.14 with n = 0.5, these values represent the transition-state fractionation factor for protonation of C-2. Therefore, using (f> = 1, the solvent isotope effect, D 2 °& h y d , was calculated by Equation 4.14 to be 1.24 and 1.26, respectively. Thus, there appears to be no significant solvent kinetic isotope effect associated with hydration. 155 4.5 Positional Isotope Exchange (PIX) Studies 4.5.1 Introduction Isotope effect studies are designed to probe the nature of the bonds, frequently those involving hydrogen, involved in a reaction mechanism by a measurable change in rate. In the previous section, these studies focused on the epimerizable centre in terms of bond-breakage in the conversion of UDP-GlcNAc to the intermediate glycal and bond-formation in the conversion of the glycal to ManNAc. Another approach in studying the individual steps in a reaction mechanism is to follow the movement of an isotope from a particular position in the starting material either to the products or to another position within the recovered starting material. The former can often enable the ready discrimination between mechanistic possibilities in which the anticipated location of the isotopic label differs. The latter case, characterized by the scrambling of the label within the substrate, is termed positional isotope exchange (PIX), and its ability to provide evidence for enzyme-bound intermediates has made it a particularly useful technique that merits further discussion. PIX is a phenomenon that can occur i f a substrate contains a carboxylate or phosphate ester group. As shown in Figure 4.26, i f O is labelled at the bridging position and the O-R2 bond is broken to give an intermediate carboxylate or phosphate, then the two C - 0 bonds or three P - 0 bonds may become torsionally equivalent via rotation about the R i - C or R iO-P bond. Provided that this bond rotation and the reverse reaction are both kinetically significant relative to a subsequent step, then it is possible for the starting material to re-form with the label at a non-bridging position. 156 O O • • y _ y — y y X V N « — R 2 R+ ^ O " R ^ R I ^ ^ O — R 2 o o • • i i i i — y — y Figure 4.26 Positional isotope exchange in carboxylate and phosphate esters. • = 1 8 0 , O = i e O . R 2 + may exist either as a discrete cation as depicted (i.e., as an intermediate of an S N I or E l reaction) or as a neutral species (i.e., resulting from an SN2 or E2 reaction). The first PIX experiment, which analyzed the reaction catalyzed by glutamine synthetase (Figure 4.27a), was reported in 1976 by Midelfort and Rose. 1 9 8 The enzyme was incubated with two of its substrates, glutamate and [y- 1 8 0 4 ]ATP (71; Figure 4.27b), labelled at the (3-y bridge, in the absence of the third substrate, ammonia. Analysis of recovered ATP revealed that isotopic scrambling to the P nonbridging positions (72; Figure 4.27b) had occurred. Thus, it was a ) O O ATP ADP + Pi 0 Q _ V S 0 * glutamine synthetase +NH3 Gin - + NH4+ . ^ , — O Y NH2 0 2 C Y -Glu 71 0 • • II II II A d o - O — P — O — P — O — P — + 0 2C-Glu 1 I I CT O" 72 0 0 • II J l _ _ II A d o - O — P — 0 3 - P — • • — P — 0 2 C - G l u 1 V I + I I O"  O" I rotation A d o - 0 — P — O — P — O I I 0" O" E G l u A T P . ' E G l u - P A D P •>"• E G l u - P A D P N H 4 + . ' E G l n A D P P j Figure 4.27 Mechanism of glutamine synthetase as determined by PIX. a) Overall reaction catalyzed by glutamine synthetase, b) PIX experiment showing scrambling of 1 8 0 label from bridging to nonbridging positions; Ado = adenosine, c) Reaction mechanism supported by PIX; E = enzyme. 157 concluded that the reaction involves transient formation of enzyme-bound ADP and glutamyl phosphate (Glu-P) prior to reaction with ammonia (Figure 4.27c). The positions of the 1 8 0 labels in the recovered ATP were determined by mass spectrometry (MS). However, the MS methods available at the time were insufficient to determine the location of a particular isotope without the use of enzymatic and chemical steps to derivatize the labelled materials. Consequently, the terminal diphosphate of the two ATP forms, 71 and 72, was displaced by the enzyme acetyl-CoA synthetase (Figure 4.28). The resulting pyrophosphate forms, 73 and 74, were free to dissociate from the enzyme, at which point the two phosphoryl groups could exchange by rotation before re-associating with the enzyme to re-form ATP as two new forms, 75 and 76, in additional to the original forms, 71 and 72. At this point, the degree of PIX could be distinguished by the appearance of Y-PO3 groups containing only one A M P - O — P — • — P — • _ u . _., „ I I C H 3 C O 2 C H 3 C 0 2 A M P cr 71 AM P - O — P — O — P — • " acetyl-CoA synthetase 0 II • II • • II II II 0 — p — • -i II - p — i II II 0 — P — 0 — p — • i i i cr i i i cr 73 74 cr I 72 • O II II AM P-«—P—•—P—O" I I CT 75 AMP-«—P—O—P—O" I I • _ cr 76 rotation acetyl-CoA synthetase C H 3 C O 2 " C H 3 C 0 2 A M P 9 O II II _ " • — P — • — P — O I I 9- o-»—P—o—P—o I I • - o-•»> L -OH 76 ADP | OH _OH glycerokinase 77 N=o • II . -o—P—o I o-78 1) KOH || • C H 3 O — P — O C H 3 2) C H 2 N 2 I O C H 3 79 Figure 4.28 Measurement of the extent of PIX by enzymatic and chemical derivatization. a) Scrambling of 1 8 0 labels from y- to (3-phosphoryl group of ATP. b) Derivatization of y-PC>3 for MS analysis. 158 18 atom of i 0 0 , as in 76. Isolation of this group was achieved by treating the equilibrated ATP mixture with dihydroxyacetone (77) and glycerokinase to generate dihydroxyacetone phosphate (78) possessing the y-PCb group. Finally, the phosphate group was liberated by base hydrolysis and esterified to trimethylphosphate (79), which was readily analyzed by gas chromatography-MS. Since Midelfort and Rose's work, analysis of PIX has been greatly aided by N M R 31 spectroscopy. Cohn and Hu utilized P-NMR spectroscopy to monitor PIX by showing that the number of 1 8 0 atoms on a phosphorus atom could be determined by the observed spectral pattern.199 They found that each substitution of 1 8 0 for 1 6 0 in a phosphate or phosphate ester causes a small upfield shift (~0.02 ppm). The 3 1 P - N M R spectrum of inorganic phosphate randomly labelled with 44% 1 8 0 illustrates this effect (Figure 4.29). The five peaks correspond to the statistical population of molecules containing between 0 and 4 atoms of 1 8 0 . Aside from indicating the number of isotopically labelled atoms, the chemical shift can 1 P. also indicate the P - 0 bond order. Cohn and Hu synthesized ATP and ADP with O labelled at various positions in the triphosphate and diphosphate chain and determined the changes in chemical shift for the 3 1 P resonances relative to unlabelled material.2 0 0 It was found that the 1 6 o 2 1 8 o 2 16 0 18 0 I Figure 4.29 3 1 P-NMR spectrum of inorganic phosphate labelled with 44% 1 8 0 . Spectrum was obtained | <~—| i 1 1 1 r 1 from Cohn and H u . 1 9 9 -0.04 0.00 0 04 0.08 0.12 ppm 159 bridging positions, which have a bond order of one, increase the chemical shift by 0.017 ppm on average, whereas the nonbridging positions, which have a bond order of 1.5 (non-terminal phosphate) and 1.33 (terminal phosphate), increase the chemical shift by 0.028 ppm and 0.022 ppm, respectively. Thus, i f a substrate labelled with 1 8 0 at a bridging position undergoes PIX to nonbridging positions, the 3 1 P - N M R spectrum will show a new, up field signal. The application of this N M R approach to an enzymatic PIX reaction in the context of this project is best exemplified by the non-hydrolyzing UDP-GlcNAc 2-epimerase, RffE. As mentioned briefly in Chapter 3, when [ l" - 1 8 0]UDP-GlcNAc was incubated with RffE and the equilibrated mixture of epimers analyzed by 3 1 P - N M R spectroscopy, the 1 8 0 label was found to scramble between the bridging and nonbridging positions of the two sugars (Figure 3.12, page a) o II u (unlabelled) = sugar—O—lj>—O—UMP O" O b (bridging) = sugar—•—lj>—O—UMP o~ II n (nonbridging) = sugar—O—(j>—O—UMP O" b) JU -12.6 —r— -12.8 J J l -13.0 -13.2 ppm -13.4 -13.6 -13.8 -14.0 Figure 4.30 "P -NMR spectra showing the Pp signals of ,80-labelled UDP-GlcNAc and UDP-ManNAc. a) [1 " - O J U D P - G l c N A c before treatment with RffE and b) equilibrium mixture of both epimers after treatment with RffE. Peaks exist as doublets due to coupling to P a . UDP-ManNAc signals are indicated by primes. Figure has been modified from Morgan et al.144 160 107).1 4 4 This scrambling was determined by observing a decrease in the Pp signal of [1"-1 8 0]UDP-GlcNAc (signal b in Figure 4.30) with the concomitant generation of an upfield 3 1 P signal corresponding to [p\,b- 1 80]UDP-GlcNAc (signal n in Figure 4.30). The relative intensity of the new, upfield signal to the original signal at equilibrium was found to be 2:1, as expected for the statistical distribution of nonbridging to bridging sites. Similar patterns were also observed for the product, UDP-ManNAc. These findings were taken as evidence for cleavage of the anomeric C - 0 bond and for the transient formation of enzyme-bound UDP. 4.5.2 PIX Experiment In view of the hydrolyzing epimerase reaction of SiaA, an analysis similar to that used for Rfffi could be employed to monitor the fate of the 1 8 0 label in [ l " - 1 8 0]UDP-GlcNAc (Figure 4.31). Together with the evidence described earlier in this chapter for 2-acetamidoglucal as an intermediate, the observation of 1 8 0 in the product UDP would provide irrefutable evidence for 31 C - 0 bond cleavage. Additionally, i f the reaction is monitored by P-NMR spectroscopy, then it may be possible to detect PIX in the starting material. Provided that the elimination of UDP is [Pnb- 1 80]UDP-GlcNAc Figure 4.31 Experiment to test for C-O bond cleavage and PIX. pnb = nonbridging position of the (3-phosphate. 161 A c O - ^ A c O ^ t ^ — C AcHN AcO--. A c O - ^ V - ^ C A c O ~ \ ^ - T * . AcHN 1) Et3NP(OBn)2, triazole, CH 2CI : AcO-OH CH 3 CN 80-85 °C H 2 « • H 2) H 2 0 2 , THF HO-•PO(OBn) 2 1) H 2 , Pd/C 2) NaOMe, MeOH HO t H O ^ UMP-morpholidate, 1/-/-tetrazole, HO trioctylamine, pyridine •PO(ONa) 2 OH OH Figure 4.32 Synthesis of [l"-180]UDP-GlcNAc. reversible and proceeds at a rate comparable to or faster than hydration of the glycal, then the enzyme-bound UDP may be capable of scrambling the 1 8 0 label by bond rotation before adding back to give UDP-GlcNAc labelled at the nonbridging position. In order to explore this strategy, [ l" - 1 8 0]UDP-GlcNAc was synthesized as previously reported (Figure 4.32). 1 4 4 The material was found to contain only 30% l s O by ESI-MS. The low incorporation is attributable to the fact that the exchange reaction was performed at a temperature lower than that used in the literature (80-85 °C vs. 90-95 °C). This conclusion is consistent with the findings of Sala, who observed that although decreased temperatures led to cleaner product 181 mixtures, they suffered from lowered label incorporation. The degree of incorporation was also visible by 3 1 P - N M R spectroscopy, which further indicated the position of the 1 8 0 label. As Figure 4.33a shows, Pp of the labelled and unlabelled material appeared as two doublets that were separated by 0.012 ppm, consistent with 1 8 0 at the bridging position. The labelled substrate 1 31 was incubated with 6xHis-SiaA, and after 30 min, H - and P-NMR spectra indicated that the reaction had proceeded to 55% completion (Figure 4.33b). The UDP product gave a pair of doublets for Pp with a separation of 0.021 ppm, showing that the 1 80-label had departed with UDP. Therefore, as anticipated, this confirms that the enzymatic reaction involves C - 0 bond cleavage. Inspection of the Pp signals of the remaining UDP-GlcNAc revealed no difference. If a) o o AcHN I J] _ J] _., , •—P—O—P—OUrd |P |a UDP-GlcNAc ° - ° --4 b) -i 1 1 r~ -4.4 -4.5 -4.6 -4.7 J o o II II ' • — P — O — P — O U r d ( p ja O - O -UDP Pa 162 —I 1 1— -11.6 -11.7 -11.8 - 1 0 -11 ppm I " —I 1 1 -11.7 -11.8 -11.9 L - 1 0 -11 ppm Figure 4.33 3 1 P -NMR spectra showing the conversion of [l"- l sO]UDP-GlcNAc to [0-l s O ] U D P by SiaA. a) Before addition of enzyme, b) After 55% reaction completion (30 min). P a and Pp signals of the pictured compounds are indicated. Signals due to unlabelled (u) and 1 80-labelled (1) species can be seen in the selected expansions. Urd = uridine. scrambling of the label had occurred within the limits of detection (it has been reported that 10% scrambling could be observed by this method153), then a new peak should have appeared upfield from the signal corresponding to the bridging position (i.e., approximately 0.029 ppm upfield from that of the unlabelled material). 1 4 4 ' 1 5 3 These findings are in complete agreement with those of Chou et al, who had also used this approach in studying the mammalian hydrolyzing epimerase.153 Consequently, the mechanistic implications of our results are common. As mentioned above, the observation of 163 HO-HO' - UDP HO - H + HO Figure 4.34 Elimination of UDP via an E l mechanism. 1 8 0 in the UDP product indicates that the enzyme cleaves UDP-GlcNAc to yield UDP via breakage of the anomeric C - 0 bond. Further, the lack of observed PIX suggests that once formed, UDP does not partition back to the pool of UDP-GlcNAc; in other words, the generation of UDP is irreversible (or a prior step such as binding or a conformational change is irreversible). This irreversibility implies that C - 0 bond breakage may be rate limiting, a scenario that requires a stepwise, El-type mechanism likely involving an oxocarbenium intermediate that is rapidly deprotonated to give the glycal (Figure 4.34). Another possibility, however, is that within the confines of the enzyme's active site, interactions between enzymatic residues and the oxygens of the p-phosphate may impede rotation. Under this restriction, even though the reaction may be reversible, the same oxygen atom will be positioned to re-form the sugar-nucleotide linkage. However, because PIX is observable with the homologous non-hydrolyzing epimerase, RffE, this possibility is questionable. In fact, under irreversible conditions with RffE, isotopic scrambling was found to be unencumbered with each turnover to product.181 4.6 Related Studies and Future Directions The findings presented in this work clearly illustrate that there is a functional link connecting the two bacterial hydrolyzing UDP-GlcNAc 2-epimerase enzymes, NeuC and SiaA, with the mammalian homologue and the bacterial non-hydrolyzing epimerase, RffE. A l l four enzymes employ anti elimination and syn addition steps in their reaction mechanisms, resulting 164 in the inversion of stereochemistry at C-2. As such, the removal and addition of the proton at this position are apparently under the direction of a two-base mechanism, as concluded from solvent isotope incorporation studies. Thus, identification of the active site residues in SiaA that are responsible for general base/acid catalysis could be pursued. In the case of RffE, several candidate residues were identified by solving the enzyme's crystal structure.138 In a similar vein, one could use X-ray crystallography to determine the arrangement of the active site residues of SiaA. Unfortunately, attempts to crystallize the protein by the research group of Dr. Natalie Strynadka in the Department of Biochemistry, U B C , have met with failure.2 0 1 It was observed that SiaA, even without its polyhistidine tag, tended to aggregate as a polydispersed protein, thereby preventing crystallization, which requires single protein forms. The sequence homology with RffE (Figure 4.35), however, may render unnecessary the need for a separate crystal structure for SiaA. In the case of RffE, the enzyme structure implicated three residues as general acids or bases involved in the transfer of the proton at C-2. 1 4 2 Thus, the corresponding mutant proteins D95N, E117Q, and E131Q were prepared and assayed for activity. 2 0 2 A l l three mutant enzymes exhibited dramatically reduced activity (kcat was reduced ~10 4 compared to wild-type), but because the mutations also disrupted the enzyme's allostery, it was not possible to assign a catalytic role to each residue. In SiaA, these three residues are conserved as D100, El22, and D131, respectively. When these were mutated and assayed by Wayne Chou in our lab, the enzymes were catalytically crippled, so much so that they could not be kinetically characterized using the coupled assay described in Section 4.3.3. Therefore, these residues are critical to the proper function of the enzyme, though, again, it is unclear whether allostery or catalysis is compromised by mutation. 165 i R f f E MKJVI S i a A M K R I I NeuC K K I I R a t M E K N G N N R K L R V t 1 O 2 0 3 0 T V F Q T , C I T G T YVT<3S VjATQN P E A I J A E Y G I 3&dy|S H A L A K D 3 A Y I e h h | T M L R E T ] F G I K T E 9 . F F E A K V . D L E L H L X . E I Q L D L A F F E L D V 4 0 T ^ q H r e m l j d TGJMRMMKTY ;M0CDN a y Gam I H 1 1 : pL ID D|y qHTTi y R Mjl| 5 0 JVJtJKLFHlVlP iRJTlYKE|v)TEE EQD EQjD 7 0 a o xo o R f f E S i a A NeuC R a t R f f E S i a A NeuC R a t R f f E S i a A NeuC R a t FYCjRI A I L S S R DYTSLN. . . . I M 0 P J G ^ g l t e ( l T C R I L E G p Q K P I L | a E F K P . . . H V J V h V ^ ^ ^ ' T E T p i T ^ L . N Y a K T Y L F S N C j I Q j G E P M G A V L G N T I T F I S R L S D E I E P . . . *MVMI Hf rKRILBALAG)AA' N F M I I K V V D I N I N J T T S H T H I L H S M S V C L N S F G C F F S N N T Y « A V M V I ^ ^ Y H I F S V ^ I A a s H H N I OffMjT HTff T.HTT|yJR|nT^nR& All^VR R V G T . A[ t |VKT.. .pjn |vT . iPJR .LKP^im I VH^KRiFiDAlLAlllftiTSiAAlliMMll 1 2 0 1 3 0 I S O t I S O 1 7 0 P M G E V E A B U R T G D L Y j a P W P E Efl LJS|GTVDDS jANYDEF G T I D D S GM E J A M YS e i s h i F'.HSKLE KL a h Li ULt Y 3 V FS (LA L T C pfr EJT SRQNfiTLJRJEJNfUJAD S RJllF IT VA jNEQAVTRLVoMGE K R K H I H I I SIT E E Y K K R VliCJLjGE K P G S V F N I CJTRJS AEQ HJLlisMqEjD H D Rf fL L A g ' |sLGA|Ej t 1 B 0 1 9 0 2 0 0 2 1 0 2 2 0 2 3 0 A L L W V R D Q V M S S D K L R S E L A A N Y P F I i V M A S S T L P S L E E V K E Y Y G . . . . L P Y E f N A L S L H L P N K Q E L E L K Y G . . . . S L L I K L L S A K N K D Y M S I I R M W L G D D . V K C I S K J S l E ^ H H R R E S F G R G F f f i ^ I C H l J f i A D I A T T H I I S M F H H V X I T E A H L I M P Q Y A A C J Y F K H I J e l s g . . q YJFV V V F H P l E T f L S T Q S V N D Q I D E j L L s B l S F F K N T H MABQBgVBJTDI KHS i KMFJELJTLDEL I i SFHKRTlljVJLJ 2 4 0 2 5 0 2 6 0 2 7 0 2 8 0 2 9 0 3 0 0 Q a V Y H v H L N P | S R E P V N R I L G R V K H B i L I D P 0 E § B P | | v w B M N f i ^ L t t I * r M a < ^ Q E g A f e L @ l S s i y p h n d t g t e s i l q e l l k y q s d . k f i a f p s F | I G S M A D T G S | D I I Q R K V K Y F C K E Y K F R Y L I S FlPtNlirJAGSKEIMviRVMRKKGI E H H PNiPlR A V KHTVI PlF dQlFJ I dlivAjHAlG OM I GMBSlClGlVRBVJqAFlGiT RJFEJYJ^LVferJLKlHA|KFkVGNi SAjdlRgAjPiLYkSVj] RSEDSSLAM I K Y S C G L I G N B S | S | G | L I 0 V J R S L K | V | a ! 3 1 0 3 2 0 3 3 0 3 4 0 3 5 0 3 6 0 R f f E V p v M r ^ E R l S E A l V f f ^ S i a A S I D vfeJTRC SNRHMGKS 1 1 H T D Y E T K N I F D A I Q Q A C S L G K F E A D D T F N G G D T R T S T E R . F A E V I N NeuC T I N I jdDRQKGRVRidASV I D V P V E K N A I V R G I N I S Q D E K F I S V V Q S S S N P Y F K E N A L I N A V R I I K R a t V l l W w T R c l l G l R l E T l d E M M L H V R D A D T Q D K I L o l A l L H L Q F G KQYPClSlKI Y G D G N A V P R I L K F L K j S I E 3 7 0 R f f E A L K N N R I S L S i a A NPETWNVSAOKRFIDLNL NeuC D F I K S K N K D Y K D F Y D I P E C T T S Y D . R a t L Q E P L Q K K F C F P P V K E N I S Q D I D H I Figure 4.35 Sequence alignment of UDP-GlcNAc 2-epimerases. RffE = non-hydrolyzing fromE. coli; SiaA = hydrolyzing from N. meningitidis; NeuC = hydrolyzing fromE. coli K l ; Rat = hydrolyzing from Rattus norvegicus (only the first 403 residues). Numbering is based on RffE sequence. Yellow boxes contain conserved residues; residues in white text on red share identity; residues in red text share similarity. Arrows indicate potential general acid/base residues. Intriguingly, when the E122Q and D131N SiaA mutants were incubated with UDP-GlcNAc and monitored by ' H - N M R spectroscopy, very slow, partial formation of 2-166 acetamidoglucal was detected. This finding is reminiscent of NeuC, which could only be observed to catalyze the elimination step using this assay. Inspection of the amino acid sequence of NeuC revealed one significant difference from its homologues: one of the targeted glutamate residues (El 17 in RffE) is a histidine residue in NeuC (Figure 4.35). This difference could account for the restricted activity of the enzyme, and it would be interesting to see i f mutation to glutamate could restore the ability to hydrate the glycal intermediate.* Another extension of this work is to explore possible homologues in other species. In particular, the bacteria Streptococcus agalactiae group B and Campylobacter jejuni have been found to possess genes that encode proteins showing substantial homology to SiaA and NeuC. 1 6 6 ' 2 0 3 At the same time we chose to clone siaA, we also targeted the C. jejuni gene, neuCl. Using the LIC method described for siaA, neuCl was successfully amplified from genomic DNA, inserted into an expression vector, and transformed into BL21(DE3) E. coli cells. Unfortunately, initial attempts to express the protein were unfruitful, and so, efforts on this protein were abandoned in favour of SiaA. Thus, the C. jejuni enzyme remains a potential target for future study. 4.7 Conclusions The study of the two bacterial hydrolyzing UDP-GlcNAc 2-epimerase enzymes presented in this chapter evolved from an unexpected opportunity initiated by communications with our collaborators. The enzyme from Escherichia coli, NeuC, had been studied by Vann and coworkers using largely biological methods. Their inability to convincingly characterize the products of the reaction motivated us to pursue a chemical approach aided primarily by N M R Prior to doing so, however, the neuC gene should be re-cloned from the genomic D N A of E. coli K l and sequenced. Because only one sequence of the gene has been reported,173 one must be sure that the histidine residue is not the result of erroneous cloning by the authors. 167 spectroscopy and mass spectrometry. By comparison to an authentic sample, the major products were shown to be 2-acetamidoglucal and UDP, which by analogy to the homologous mammalian epimerase and non-hydrolyzing epimerase, RffE, is the result of an anti elimination reaction. Vann was able to use radioactive isotope chromatography to show that the enzyme additionally forms a small amount of the epimeric sugar ManNAc; hence, it was concluded that NeuC is a hydrolyzing UDP-GlcNAc 2-epimerase and that the inability to detect ManNAc as the major product in vitro suggests there may be a regulatory molecule or protein necessary for full activity. Encouraged by these results and intrigued by the literature claim of GlcNAc 6-phosphate 2-epimerase activity, the homologous gene from Neisseria meningitidis, siaA, was cloned and expressed. The enzyme was shown to catalyze the same reaction as NeuC but with complete conversion to ManNAc and UDP. Although the putative glycal intermediate could not be observed directly, incubation of 2-acetamidoglucal with SiaA in the presence of UDP resulted in gradual conversion to the ManNAc product. Using a continuous coupled assay, the enzyme was found to be allosterically regulated by its substrate, exhibiting positive cooperativity; this is in contrast to the mammalian homologue, which displays negative cooperativity. The corresponding Hi l l coefficient of 1.9 is suggestive of a strongly cooperative dimeric enzyme, as in the case of RffE. In agreement with the mammalian enzyme, SiaA was found to transfer solvent-derived deuterium into the product with each turnover, consistent with a two-base epimerization mechanism. Isotopically substituted substrates were used to probe the nature of the rate-limiting step. The absence of a significant kinetic isotope effect using (2"- 2H)UDP-GlcNAc indicates that C - H bond cleavage is not rate limiting. Additionally, no solvent isotope discrimination was detected when either UDP-GlcNAc or 2-acetmidoglucal and UDP were reacted with 50% deuterated water. Assuming the catalytic acid residue is monoprotic with no intrinsic isotopic preference, 168 this implies that proton transfer during hydration of the glycal is not rate limiting. Incubation with [ l" - 1 8 0]UDP-GlcNAc resulted in the generation of 1 80-labelled UDP, confirming that the nucleotide diphosphate departs with C - 0 bond cleavage. The absence of PIX suggests that this step may be rate limiting, a scenario that requires a stepwise, El-type mechanism likely involving an oxocarbenium intermediate that is rapidly deprotonated to give the glycal. Alternatively, the lack of PIX may be indicative of restricted rotation of the p-phosphate in the enzyme-bound UDP. In this case, it is possible that a subsequent step such as product release is rate limiting. A third possibility that cannot be excluded from these results is that substrate binding or a conformational change prior to the first chemical step is rate limiting. After the completion of this project and during final revisions to our recent report on NeuC, 1 6 6 an article addressing the metabolism of ManNAc in E. coli K l was published by Ringenberg et al.}65 They examined the flux of ManNAc in the sialic acid biosynthetic pathway and demonstrated that ManNAc6P is not an obligate precursor. On this basis, they stated that the previous claim by Petersen et al. that the N. meningitidis enzyme SiaA or its E. coli homologue NeuC is a GlcNAc6P 2-epimerase115 is invalid, a conclusion we also made through direct means. They further suggested that the activity Petersen et al. had attributed to SiaA was likely due to background activity from the epimerase NanE produced by the host E. coli cells in which siaA was overexpressed. This notion is supported by the observation that a purified sample of SiaA exhibited minimal GlcNAc6P 2-epimerase activity compared to a crude extract.115 Ringenberg et al. also independently purified histidine-tagged NeuC, though no source was cited, and monitored the epimerase reaction using radiolabelled UDP-GlcNAc. Using a chromatographic approach similar to that of Vann et al. in our manuscript, they found that a compound was formed that co-migrated with an authentic sample of GlcNAc and not the anticipated product ManNAc. They concluded that this compound was either GlcNAc or a GlcNAc-related intermediate (i.e., 2-acetmidoglucal). Although they incorrectly stated it was the result of 169 substrate hydrolysis, their speculation was confirmed by our more convincing N M R assays with NeuC. 4.8 Experimental Procedures 4.8.1 Materials A l l substrates and buffers were purchased from Sigma-Aldrich or Fisher Scientific except where noted. 1 80-enriched water (95%) was purchased from Icon Isotopes, New Jersey. 2-Acetamido-l,2-dideoxy-D-tfra£z'rco-hex-l-enopyranose (2-acetamidoglucal),171' 1 7 2 2-acetamido-2-deoxy-a-D-glucopyranosyl fluoride (69), 1 8 0 and 2-methyl-4,5-dihydro-(l,2-dideoxy-P-D-I & 1 mannopyranoso)[2,l-£/]-l,3-oxazole (70) were synthesized by Wayne Chou in our lab. Cytidine 5'-monophospho-A/-acetylneuraminic acid (CMP-Neu5Ac) was provided by Dr. Willie Vann, originally purchased from Sigma-Aldrich. L B medium components (tryptone and yeast extract) and DE52 ion-exchange resin were purchased from Difco Laboratories. Microbiological agar was purchased from Marine BioProducts. Ampicillin and kanamycin sulfate were purchased from Fisher Biotech and Gibco, respectively. Aprotonin and pepstatin A were purchased from Boehringer Mannheim. Taq polymerase, 10X PCR buffer, 10 mM dNTP mix, and the pET Directional TOPO Expression Kit (One Shot® TOP 10 and BL21 Star™ (DE3) One Shot® chemically competent cells, pET TOPO vector, 12.5 mM dNTP mix, salt solution, and sterile water) were purchased from Invitrogen. The pET-30 Xa/LIC Vector Kit (NovaBlue GigaSingles™ chemically competent cells, Xa/LIC vector, T4 D N A polymerase, 10X T4 D N A polymerase buffer, 100 mM DTT, 25 mM EDTA, 25 mM dGTP, nuclease-free water, and SOC medium) was purchased from Novagen. 170 Centrifugal filters (4 mL, 10 000 MWCO) were purchased from Millipore. Spectra/Por® Biotech RC dialysis tubing (12 - 14 000 MWCO) was purchased from Spectrum Laboratories, Inc. Acryl-cuvettes for use in enzyme kinetic assays were purchased from Sarstedt. Chitin resin was purchased from New England Biolabs. Chelating Sepharose Fast Flow resin was purchased from Pharmacia Biotech. Chelex® 100 resin (200-400 mesh, N a + form), AG® 1-X8 resin (100-200 mesh, formate form), and Bio-Gel® P-2 resin were purchased from Bio-Rad Laboratories. Amberlite® IR-120 resin ( H + form) and Dowex® 1X4-200 resin (CI" form) were purchased from Sigma-Aldrich. D N A primers for cloning and sequencing were obtained from the Nucleic Acids Protein Services (NAPS) Unit at UBC. The plasmid pSR647, corresponding to neuC inserted into the intein-fusion plasmid pCYB4 from the IMPACT™ 1 Kit (New England Biolabs), was provided by Dr. Willie F. Vann at the Laboratory of Bacterial Toxins, Center for Biologies Evaluation and Research, Food and Drug Administration, Bethesda, Maryland. Chromosomal D N A from N. meningitidis (serogroup B, strain MC58) was provided by Dr. Warren Wakarchuk at the Institute for Biological Sciences, National Research Council of Canada, Ottawa, Ontario. The plasmid pUSOl, encoding UDP-ManNAc dehydrogenase (UDPDH), was provided by Dr. Paul Morgan in our lab. DH5a and JM109(DE3) E. coli cells were purchased from Invitrogen and Promega, respectively, and made chemically competent as described in Section 4.8.2.3. Purified UDP-GlcNAc 2-epimerase (RffE; non-hydrolyzing), used in the preparation of (2"- 2H)UDP-GlcNAc (Section 4.8.9.1), was provided by Dr. Jomy Samuel in our lab. 171 4.8.2 General Methods 4.8.2.1 General Enzyme Methods Unless otherwise stated, proteins were handled on ice or at 4 °C. Protein concentrations were determined by Bradford assay1 0 4 on a Cary 3E UV-Vis spectrophotometer using bovine serum albumin (BSA) in 20 mM sodium phosphate buffer, pH 7.5 as the standard. Protein purity was assessed using SDS-PAGE, stained with Coomassie blue according to Laemmli. 1 0 5 Molecular weight standards for SDS-PAGE were B S A (66 kDa) and carbonic anhydrase (29 kDa), both purchased from Sigma. Deuterated buffers were prepared by dissolving buffer components in H2O, adjusting to a desired pH with dilute HCI or NaOH, followed by concentration to dryness (e.g., lyophilization or vacuum centrifugation) and reconstitution with an equal volume of D2O. The pD values of the resulting buffer solutions were determined by adding 0.4 units to the observed pH meter readings, as has been empirically established by Glasoe and Long. 2 0 4 One unit of enzyme activity is defined as the amount of enzyme needed to convert 1 umol of substrate into product in 1 min at 37 °C. 4.8.2.2 General DNA Methods Primer concentrations were determined by measuring A260 according to Equation 4.19, . . . . A 100 ^ M ) = A™ Xl.5NA+0JNc+l.2NG+0MN7 4 1 9 where N\, NQ, NQ, NJ are the number of the respective bases in the primer. Plasmid and chromosomal D N A concentrations were determined by measuring ^260 according to the Equation 4.20, j(ug/mL) = A260 x 50 4.20 172 D N A purity was assessed by agarose gel electrophoresis using 100 bp and 1 kb D N A standards (New England Biolabs). Polymerase chain reaction (PCR) products were purified using the SpinPrep™ PCR Clean-up Kit from Novagen. Plasmids were isolated from 3 - 5 mL cell cultures using Wizard® Minipreps D N A Purification Resin and Wizard® Minicolumns (Promega), according to the manufacturer's protocol. D N A sequencing of recombinant genes and selected PCR products was performed by NAPS, U B C . 4.8.2.3 Preparation of Chemically Competent E. coli A n adaptation of the method of Inoue et al.205 was employed. A 5 mL culture (in LB) was inoculated from a single colony of cells and grown overnight at 37 °C with shaking at 225 rpm. The overnight culture was used to inoculate a 250 mL culture (LB), which was incubated at 37 °C with shaking at 200 rpm until an OD600 of 0.4 was reached. The culture was then transferred to an ambient temperature shaker (22 °C, 135 rpm) until the OD600 reached 0.6. The cells were cooled on ice for 10 min before being centrifuged at 2500 xg at 4 °C for 5 min. The cells were resuspended in 80 mL transformation buffer (10 mM PIPES, 15 mM CaCl 2 , 250 mM KC1, adjusted to pH 6.7 with 1 M K O H , followed by 55 mM MnCl 2 ) , cooled on ice for 10 min, and centrifuged as before. The cells were resuspended in 20 mL transformation buffer containing 7% v/v DMSO and cooled on ice for 10 min. Cells were aliquoted in 200 uL portions into 1.6 mL microcentrifuge tubes, frozen in liq. N 2 , and stored at -80 °C. 4.8.2.4 Polymerase Chain Reaction (PCR) To a 200 uL PCR tube were added 5.0 uL of 10X PCR buffer, 1.0 u-L of 10 mM dNTP mix, 1.5 uL of 50 mM M g C l 2 , 25 ng of template DNA, 25 pmol of each primer, 0.25 uL of 5 U/uL Taq polymerase, and distilled H 2 0 to a total volume of 50 uL. D N A was amplified using 173 an iCycler™ Thermal Cycler (Bio-Rad) according to the following cycles: one cycle of 3 min at 94 °C; thirty cycles of 45 s at 94 °C, 30 s at 55 °C, and 90 s at 72 °C; and one cycle of 10 min at 72 °C. 4.8.2.5 Colony PCR To verify that cells were successfully transformed with plasmid containing the desired gene, colony PCR was performed, essentially as reported. ' A master mix, enough for 12 reactions, was prepared by combining 25.0 u.L of 10X PCR buffer, 5.0 uL of 10 m M dNTP mix, 7.5 uL of 50 mM MgCh, 50 nmol each of a gene-specific primer and a vector-specific primer (see Table 4.2 in Section 4.8.5 for primer sequences), and distilled H 2 0 to a total volume of 250 uL. Twenty microlitres was aliquoted into 200 uL PCR rubes, and using a sterile micropipette tip, a sampling of cells from each colony was added. After resuspending tne cells by pipetting up and down several times, the tubes were boiled for 10 min in a thermal cycler. To each tube was added 0.25 \ih of 5 UlyL Taq polymerase, and the PCR cycles described in 4.8.2.4 were performed. 4.8.2.6 NMR Spectroscopy and Mass Spectrometry ' H - N M R spectra were obtained on a Bruker AV300 or AV400 spectrometer at a field strength of 300 or 400 MHz, respectively. Proton-decoupled 3 1 P - N M R spectra were recorded on these spectrometers at 121.5 MHz or 162 MHz, respectively. Mass spectrometry was performed by the Mass Spectrometry Centre at UBC by ESI-MS using a Waters Micromass LCT mass spectrometer. 174 4.8.3 Overexpression of Intein-tagged NeuC and Affinity Purification of NeuC The plasmid pSR647, consisting of neuC ligated into the intein fusion plasmid pCYB4, was transformed into E. coli DH5a competent cells, which were incubated at 37 °C on L B agar plates containing 100 pg mL"1 ampicillin. A single colony was used to inoculate a 10 mL culture, which was grown overnight at 37 °C with shaking at 225 rpm. The overnight culture was diluted 1:100 in L B medium containing 100 pg mL"1 ampicillin and grown at 37 °C with shaking at 225 rpm until an OD600 of 0.6 was reached. Cells were induced for overexpression by addition of IPTG to 0.5 mM, and the culture was grown for 16-20 h at 22 °C with shaking at 135 rpm. Cells were harvested at 5000 rpm in a Sorvall® SLC-1500 rotor, resuspended in lysis buffer (20 mM HEPES, pH 7.8, 500 m M NaCl, 1 m M EDTA, 1 u.g mL"1 pepstatin A , and 1 ug mL"1 aprotinin), and lysed at 20 000 psi in an ice-cooled French pressure cell. The cell lysate was clarified by centrifugation (5000 rpm, 30 min) and filtration through 0.45 um syringe filters, and loaded onto a 12 mL column containing chitin resin. The resin was washed with 10 column volumes (CV) of column buffer (lysis buffer minus pepstatin and aprotinin), followed by 3 C V of cleavage buffer (column buffer containing 50 mM DTT). The enzyme was cleaved from the fused chitin binding domain by allowing the resin to stand overnight and was eluted with 12 mL of column buffer. 4.8.4 NMR Assays with NeuC Affinity-purified NeuC was exchanged into deuterated sodium phosphate buffer (10 mM, pD 8.0) by four rounds of concentration and dilution using 4 mL centrifugal filters (10 000 MWCO). The enzyme was diluted to 665 uL and transferred to an N M R tube containing Chelex® 100 resin (20 mg, previously rinsed with D2O). The reaction was initiated by addition of 35 uL of 100 m M UDP-GlcNAc in D 2 0 , and the resulting solution (5 m M UDP-GlcNAc) was 175 incubated at 37 °C for two days. ] H - and 3 1 P - N M R spectra were obtained at timed intervals on a 300 MHz spectrometer. The final solution was stirred with 1 mL AG® 1-X8 resin for 1 h, filtered through glass wool, and lyophilized to dryness. The sample was dissolved in 700 uL D2O, and a ' H - N M R spectrum was obtained, indistinguishable from a standard of 2-acetamidoglucal. The sample was again lyophilized to dryness and analyzed by mass spectrometry: +ESI-MS m/z 226 (M + Na + , 100). From a separate preparation of NeuC, two 665 uL samples (0.3 mg mL"1) were prepared in deuterated buffer in N M R tubes, as described. To one tube, 0.5 mg CMP-Neu5Ac (1 mM) was added, and then to both tubes, 35 uL of 100 mM UDP-GlcNAc was added. The reactions were incubated at 37 °C with monitoring by 1 H -NMR spectroscopy. From another preparation of NeuC, four 665 uL samples were prepared in deuterated buffer in N M R tubes. To two of the tubes was added 17.5 uL of 200 m M solutions in D 2 0 of N A D + or N A D P + (each 5 mM final), and to the other two was added 17.5 pX of D 2 0 alone. To one of these D 2 0 tubes was added 17.5 uL of 200 mM GlcNAc6P (5 m M final), and to the remaining three was added 17.5 pL of 200 mM UDP-GlcNAc (5 mM final). The tubes were incubated at 37 °C with monitoring by 1 H-NMR spectroscopy. 4.8.5 Cloning of siaA Gene Primers (Table 4.2) were designed by combining an overhang sequence complementary to the pET-30 Xa/LIC vector with a gene-specific sequence corresponding to the 5'-end of the sense strand or the 3'-end of the antisense strand. The siaA gene was amplified by PCR as described in 4.8.2.4 using N. meningitidis chromosomal DNA. Ligation-independent cloning (LIC) was then used as directed by the kit supplier (Novagen) to insert the amplified gene into 176 the pET vector. The presence of the gene in the resulting plasmid, pAM04, was confirmed by colony PCR (4.8.2.5) and D N A sequencing. Table 4.2 Primers used in ligation-independent cloning, colony PCR, and sequencing of siaA Primer Sequence* NA A t NG AM1101 5'-GGT ATT G A G GGT C G C A T G A A A A G A ATT CTT TGC ATT AC-3 ' 11 5 10 12 A M I 102 5'-AGA G G A G A G TTA G A G C C T T A A A G A TTC A A A T C G A T A A-3 ' 16 4 9 8 T7 5'-TAA T A C G A C T C A C T A T A G GG-3' 7 4 4 5 T7Rev 5'-GCT A G T TAT TGC T C A G C G G-3' 3 4 6 6 * Bases in red represent overhang region. Bases in black represent gene-specific region. Indicated in bold face are the start codon (ATG) and the reverse complement (TTA) of the ocher stop codon (TAA). 4.8.6 Overexpression and Affinity Purification of 6xHis-SiaA The plasmid pAM04, encoding SiaA fused at the N-terminus with a peptide linker containing a six-histidine repeat, was transformed into BL21 Star™ (DE3) One Shot® chemically competent E. coli cells, which were incubated at 37 °C on L B agar plates containing 30 pg mL"1 kanamycin. Single colonies were used to inoculate 4 x 1 0 mL cultures, which were grown overnight at 37 °C with shaking at 225 rpm. The overnight cultures were poured into 4 x 500 mL L B medium containing 30 pg mL"1 kanamycin and grown at 37 °C with shaking at 225 rpm until an OD600 of 0.6 - 0.7 was reached. Cells were induced for overexpression by addition of 120 mg L" 1 (0.5 mM) IPTG, and the cultures were allowed to continue growth until an OD600 of 1.6 - 1.8 was reached (approx. 4 h). Cells were harvested at 4000 rpm in a Sorvall® Heraeus 6445 rotor, resuspended in lysis buffer (20 mM sodium phosphate, pH 7.5, 10 mM imidazole, 0.4 M NaCl, 1 pg mL"1 pepstatin A, and 1 pg inL"1 aprotinin), and lysed at 20 000 psi in an ice-cooled French pressure cell. The cell lysate was clarified by centrifugation in a Sorvall SLC-1500 rotor at 10 000 xg for 40 min. 177 A 9 mL column containing Chelating Sepharose® Fast Flow resin was charged with 2 C V of 100 m M N i S 0 4 , followed by washing with 2 C V of distilled H 2 0 and 3 C V of start buffer (lysis buffer minus aprotinin and pepstatin). The clarified cell lysate was loaded at 2 mL min"1, and start buffer (approx. 8 CV) was passed through the column at 3 mL min"1 until no more flow-through protein eluted, as determined by monitoring A2&0 using a Spectra/Chrom™ Flow Thru U V Monitor (Spectrum®). To remove non-specifically bound proteins, a step gradient was then applied, consisting of 2 C V each of 5%, 10%, 15%, 20%, and 50% elution buffer (20 mM sodium phosphate, pH 7.5, 500 mM imidazole, and 0.4 M NaCl). 6xHis-SiaA was finally eluted with 3 - 4 C V of 100% elution buffer. Initial attempts to remove imidazole and concentrate the enzyme by buffer exchange using centrifugal filters resulted in substantial loss of protein. Instead, fractions containing the enzyme were pooled and dialyzed overnight against a 1:100 volume of dialysis buffer (20 mM or 50 m M sodium phosphate, pH 7.5). The enzyme solution was then divided into 665 pL or 1.33 mL aliquots, flash-frozen in liq. N 2 , and lyophilized. The enzyme could be stored in this state at -20 °C for at least three months without significant loss in activity. Samples were reconstituted as needed by adding H 2 0 or D 2 0 and gently inverting several times. 4.8.7 NMR Assays with 6xHis-SiaA 4.8.7.1 Initial Assays Three 1.15 mL samples of 6xHis-SiaA (0.21 mg each) in 20 m M deuterated sodium phosphate buffer, isolated as described in Section 4.8.6, were incubated at 37 °C with GlcNAc, GlcNAc6P, or UDP-GlcNAc (5 mM each). ' H - N M R spectra were recorded at intervals between 30 min and 2 d. 178 4.8.7.2 Stereochemistry and Solvent Deuterium Isotope Incorporation A 665 pL sample of enzyme (35 pg) in 10 mM deuterated sodium phosphate buffer, pD 8.0, was placed in an N M R tube, followed by 35 pL of 100 mM UDP-GlcNAc. ' H - N M R spectra were acquired after 1.5 min, 4 min, and 2 d. 4.8.7.3 Catalytic Competence of 2-Acetamidoglucal and Alternative Substrates To two N M R tubes was added 665 pL of enzyme (170 pg) in 50 mM deuterated buffer, followed by 17.5 pL of D 2 0 or 200 mM UDP in D 2 0 (5 m M final). Reactions were initiated by the addition of 17.5 pL of 200 m M 2-acetamidoglucal in D 2 0 (5 m M final), and 1 H-NMR spectra were acquired after 4.5 h and 2 d. A third tube containing 5 mM 2-acetamidoglucal in 50 m M deuterated phosphate buffer served as a control. Glycosyl fluoride 69 and ManNAc oxazoline 70 were treated as described for 2-acetamidoglucal using 20 mM buffer. 4.8.8 Kinetic Characterization of 6xHis-SiaA Enzyme kinetics were measured using a continuous coupled assay for UDP formation, as described by Chou et al.,153 using 0.20 - 10.0 mM UDP-GlcNAc. UDP-GlcNAc stock concentration was determined by measuring ^4262 using £262 = 9890 M " 1 cm"1. Reactions were initiated by addition of 10 pL 6xHis-SiaA (2.6 pg), and the decrease in ^340 (due to N A D H ; £340 = 6220 M " 1 cm"1) was monitored at 37 °C. Kinetic parameters were determined from initial velocities fit to the Hi l l equation using the computer program GraFit. 2 0 8 The pH dependence of activity was assessed by conducting the coupled assay in buffers prepared at various pH values as follows. Stock solutions of 360 m M sodium phosphate buffer were adjusted to pH 6.0, 6.5, 7.0, 7.5, and 8.0 at 22 °C using dilute aqueous HCI or NaOH. To test basic pH values, stock solutions of 360 m M Tris were adjusted to pH 8.0, 8.5, 9.0, 9.5, and 179 10.0 at 22 °C. Accounting for the known temperature dependence of the pKa of Tris-tT (-0.027 units per °C increase), ' the corresponding pH values of the solutions in the cuvettes at 37 °C were approximated as 7.6, 8.1, 8.6, 9.1, and 9.6. Assays were performed in the presence of 7.5 mM UDP-GlcNAc as described previously by substituting the appropriate buffer. 4.8.9 Synthesis of Isotopically Substituted and Labelled UDP-GlcNAc 4.8.9.1 Disodium uridine 5'-(2"-acetamido-2"-deoxy-(2"-2H)-a-D-glucopyranosyl diphosphate) ((2 "-2H)UDP-GlcNAc) (2"- 2H)UDP-GlcNAc was prepared according to the procedure of Morgan et al.m with the following changes. Overexpression and purification of UDPDH. After transformation of chemically competent JM109(DE3) E. coli with pUSOl, cells were grown at 37 °C on an L B agar plate containing 50 pg mL"1 ampicillin. Single colonies were used to inoculate 4 x 10 mL L B containing 50 pg mL"1 ampicillin, which were grown overnight at 37 °C with shaking at 225 rpm. The overnight cultures were used to inoculate 4 x 500 mL cultures, which were incubated at 37 °C with shaking at 225 rpm until an OD600 of 0.6 was reached. The cells were induced for overexpression with 0.5 m M IPTG and were allowed to continue growth to an OD600 of 1.8. The cells were harvested by centrifugation at 5000 rpm in a Sorvall® SLC-1500 rotor, resuspended in 20 mL lysis buffer (50 m M triethanolamine, 10% glycerol, 2 m M DTT, 1 pg mL"1 pepstatin, and 1 ), and lysed at 20 000 psi in an ice-cooled French pressure cell. The cell lysate was clarified by centrifugation at 7000 rpm for 10 min, and the supernatant was brought to 40% saturation with (NFL)2S04 (0.226 g mL"1 at 0 °C). The precipitated protein was pelleted by centrifugation at 5000 rpm for 30 min, and the supernatant was brought to 60% saturation with (NIL^SC^. (additional 0.120 g mL"1 at 0 °C). The precipitated protein, containing UDPDH, was separated 180 by centrifugation at 5000 rpm for 30 min and redissolved in 10 mL storage buffer (50 m M Tris-HC1, pH 8.8, 10% glycerol, and 4 mM DTT). The resulting U D P D H solution was used as described144 without further purification. Purification of (2"-2H)UDP-GlcNAc. The product was purified using a 19 mL column of AG® 1-X8 resin by eluting with a linear gradient of 0 - 1 M L i C l (pH 3.5, 400 mL total volume). Portions of the fractions were transferred to N M R tubes, and 5% D 2 0 was added. Fractions containing (2"- 2H)UDP-GlcNAc, identified by 3 1 P - N M R spectroscopy (H 2 0 as lock solvent), were concentrated to near dryness in a SpeedVac® Plus SC110A vacuum centrifuge (Savant) at 43 °C. The fractions were pooled and desalted by applying to a column of Bio-Gel® P-2 resin (2.6 x 47 cm), eluted with distilled H 2 0 . A column of Amberlite® IR-120 resin (10 mL, H + form) was converted to the Na + form by passing through 1 M NaOH until basic to litmus, then H 2 0 until neutral. The product was then applied and eluted as the sodium salt with distilled H 2 0 . Extent of 2 H incorporation was determined to be >97% by 1 H -NMR spectroscopy and -ESI-MS. 4.8.9.2 Disodium uridine 5'-(2"-acetamido-2"-deoxy-[l"-180]-a-D-glucopyranosyl diphosphate) ([1 "-180]UDP-GlcNAc) [ l " - I 8 0]UDP-GlcNAc was prepared according to the procedure of Morgan et al.144 The product was purified as described above for (2"- 2H)UDP-GlcNAc except that a 1.6 x 27 cm column of Dowex® 1X4-200 resin (CI form) was substituted for the AG® 1-X8 column. The 1 18 H-NMR spectrum was identical to the unlabelled compound. The extent of O incorporation was determined to be 30% by mass spectrometry: -ESI-MS (H 2 0) m/z 606 (M - H + , 1 6 0 , 100), 608 (M - tf", 1 8 0 , 40.6). The location of the 1 8 0 label was confirmed by 3 1 P - N M R spectroscopy: (121.5 MHz, D 2 0) 5 -10.008 (d, J = 20.5 Hz, P a ) , -11.714 (d, J = 20.5 Hz, Pp- 1 60), -11.726 (d, J = 20.5 Hz, P p - 1 8 0) . 181 4.8.10 Determination of Kinetic Isotope Effects (KIEs) 4.8.10.1 V : KIE onkcat Rates of reaction were determined in triplicate for (2"-'H)- and (2"- 2H)UDP-GlcNAc (both at 7.5 mM) using the coupled assay in Section 4.8.8. The primary KIE on kcat (DV) was calculated by Equation 4.3 (page 145). 4.8.10.2 D(V/K): KIE on keJKm by Intermolecular Competition Step 1: Calculation ofS0. A 1:1 molar solution (700 pL total, 8.13 mM each) of (2"-'H)-and (2"- 2H)UDP-GlcNAc in D 2 0 was prepared from stock solutions of 17.1 m M and 15.5 mM, respectively, as determined by ^ 2 6 2 - A ' H - N M R spectrum (400 MHz) was acquired, and the ratio (So) of the H - l " signals (5 5.38 ppm) to the uracil H-6 signal (8 7.82 ppm) was calculated based on the mean of three integrations. Step 2: Calculation of Ro. A 35 pL sample was removed, frozen in liq. N 2 , lyophilized, and submitted to mass spectral analysis (-ESI-MS, H 2 0) . The relative numbers, an and ao, of the two isotopologues were calculated by Equations 4.10 and 4.11. The ratio (Ro) of (2"-2 H)UDP-GlcNAc to (2"-'H)UDP-GlcNAc was calculated by Equation 4.5 based on the mean of five measurements. Step 3: Calculation of S. The remaining 665 pL of the sample was added to a lyophilized sample of 6xHis-SiaA (340 pg in 40 m M sodium phosphate buffer, pH 7.5), and reaction progress was monitored by ' H - N M R spectroscopy. After 3.5 h (approximately 85% completion), 5 pL of 6 M HCI was added to inactivate the enzyme, followed by neutralization with 30 pL of 1 M NaOH. Again, a 1 H-NMR spectrum was acquired, and the ratio (S) of the H -1" signals of the remaining substrates to the overlapping signals from the uracil H-6 of the remaining substrates and UDP was calculated based on the mean of three integrations. 182 Step 4: Calculation ofR. The sample was frozen in liq. N 2 , lyophilized, and submitted to mass spectral analysis. The ratio (R) of (2"- 2H)UDP-GlcNAc to (2"-'H)UDP-GlcNAc was calculated by Equation 4.5 based on the mean of five measurements. Step 5: Calculation of 1 - FH and D(V/K). The fraction of remaining (2"-'H)UDP-GlcNAc (1 - F H ) was calculated by Equation 4.12. Finally, the primary KIE on VIK (D(V/K)) was calculated by Equation 4.4. 4.8.10.3 Solvent Isotope Discrimination During Glycal Hydration A 2 mL sample of 50% D 2 0 was prepared by mixing 0.901 g of H 2 0 (0.500 mol) and 1.000 g of 99.9% D 2 0 (0.499 mol). This labelled water was used to reconstitute a sample of lyophilized enzyme (340 ug, 1.33 mL; 20 m M sodium phosphate buffer) and to prepare stock solutions of 2-acetamidoglucal (200 mM), ManNAc (200 mM), and UDP (200 mM). The enzyme solution was divided into two 665 uL portions; to one were added 17.5 uL UDP and 17.5 uL ManNAc, and to the other were added 17.5 uL UDP and 17.5 uL glycal. The mixtures were incubated at 37 °C for 7.5 h, at which time, ~1 mL of AG® 1-X8 resin was added to remove anionic species (i.e., UDP and phosphate buffer). The mixtures were filtered through glass wool, frozen, lyophilized, and submitted for mass spectral analysis (+ESI-MS). The relative numbers of molecules, an and a®, of (2-'H)ManNAc and (2- 2H)ManNAc, respectively, were calculated from Equations 4.10 and 4.11 using z'244 (corresponding to m/z 244, M + Na +), z245, and i2^- The ratio of protiated to deuterated product was calculated as AH/^D- The sample containing ManNAc as the starting sugar served as a control and gave no 246 peak, indicating the reaction is irreversible. As described above, 50% D 2 0 was used to reconstitute a sample of lyophilized enzyme (170 u.g, 665 uL; 20 m M sodium phosphate buffer) and to prepare a stock solution of UDP -GlcNAc (100 mM). To the enzyme solution was added 35 uL UDP-GlcNAc. The mixture was 183 incubated at 37 °C for 40 min, at which time, AG® 1-X8 resin was added. The mixture was filtered, concentrated to dryness in a SpeedVac centrifuge, and submitted for mass spectral analysis. The same calculations as above were used to determine the ratio of ManNAc isotopologues. 4.8.11 Positional Isotope Exchange (PIX) Experiment and Test for C-O versus P-O Bond Cleavage A 700 pL solution of 10 m M [l"- 1 8 0]UDP-GlcNAc in 20 mM sodium phosphate buffer, pH 7.5, was placed in an N M R tube, and Chelex® 100 resin (~20 mg, previously rinsed with D2O) was added. A 3 1 P - N M R spectrum was obtained using the following acquisition parameters: spectral frequency = 121.5 MHz, sweep width = 2437 Hz, acquisition time = 13.4 s, pulse delay = 2.0 s, pulse width = 10 ps. The solution was removed from the tube, mixed with lyophilized enzyme (340 pg; 20 mM sodium phosphate buffer), and returned to the tube. Reaction progress at 25 °C was monitored by ' H - and 3 I P - N M R spectroscopy. After -55% completion (30 min), the enzyme was inactivated by addition of 5 pL of 6 M HCI, followed by 30 pL of 1 M NaOH, and a 3 1 P - N M R spectrum was acquired: 5 -4.559 (d, J =22 Hz, P p - 1 6 0 of UDP), -4.580 (d, J = 22 Hz, Pp- l sO of UDP), -9.099 (d, J= 22 Hz, P a of UDP), -10.049 (d, J = 21 Hz, P a of UDP-GlcNAc), -11.792 (d, / = 21 Hz, Pp- 1 60 of UDP-GlcNAc), -11.805 (d, J= 21 Hz, P p - 1 8 0 of UDP-GlcNAc). 4.8.12 Protein Sequence Alignment The multiple sequence alignment of RffE (Swiss-Prot primary accession number P27828), SiaA (Q57141), NeuC (Q47400), and mammalian (035826) UDP-GlcNAc 2-epimerases was made using C L U S T A L W 2 0 9 using the B L O S U M weight matrix and gap opening, ending, extension, and distance penalties of 10, 10, 7.5, and 8, respectively.210 184 Appendix ^ - N M R Spectra of Final, Deprotected Peptides Current Data Parameters NAME L-ester EXPNO 4 PROCNO 1 F2 - Acqu Date_ Tine INSTRUM PROBHD PULPROG TO SOLVENT NS DS SWH FIDRES AQ RG DW DE TE 01 i s i t i o n Parameters 20001117 23.01 av400 mm BBI 1H-zg30 16384 CDC 13 16 2 5592.841 Hz 0.341360 Hz 1.4647796 sec 181 89.400 usee 6.00 usee 300.0 K 1.00000000 sec f o CD ======= CHANNEL f l NUC1 1H PI 10.00 usee PL1 6.20 dB SF01 400.1320007 MHz F2 - Processing parameters SI 32768 SF 400.1300124 MHz WDW EM SSB 0 LB 0.30 Hz GB 0 PC 1.00 ID NMR plot parameters CX 20.00 cm F1P 8.000 ppm Fl 3201.04 Hz F2P 0.000 ppm F2 0.00 Hz PPMCM 0.40000 ppm/cm HZCM 160.05200 Hz/cm I—I—r—I—I—I—I—1—I—I—I—I—I—I 1 00 Current Data Parameters NAME Cl-pept-OtBu EXPNO 3 PROCNO 1 OJ ; n •*T OJ tn r> to o cn o OJ (J3 F2 - Acqui Date_ Time INSTRUM PROBHD PULPROG TD SOLVENT NS OS SWH FIDRES AQ RG DW OE TE 01 s i t i o n Parameters 20011105 13.21 av400 mm BBI 1H-zg30 32768 MeOH 33 2 4990.020 Hz 0.152283 Hz 3.2834036 sec 2048 100.200 usee 6.00 usee 300.0 K 1.00000000 sec „ „ „ „ CHANNEL f l ======== NUC1 1H PI 10.00 usee PL1 4.70 dB SF01 400.1320000 MHz F2 - Processing parameters SI 32768 SF 400.1300171 MHz WDW EM SSB 0 LB 0.20 Hz GB 0 PC 1.00 ID NMR plot parameters CX F1P Fl F2P F2 PPMCM HZCM 20.00 cm 8.000 ppm 3201.04 Hz 0.000 ppm 0.00 Hz 0.40000 ppm/c 1B0.05200 Hz/cm C u r r e n t Data Parameters NAME Sec-pept EXPNO 1 PROCNO 1 F2 - Acqu Oate_ Time INSTRUN PROBHD PULPROG TD SOLVENT NS OS SWH FIDRES AQ RG DW DE TE Dl i s i t i o n Parameters 20020111 19.18 av300 5 mm QNP 1H/ zg30 32768 MeOH 32 2 6313.131 Hz 0.192661 Hz 2.5952756 sec 574.7 79.200 usee 6.00 usee 300.0 K 1.00000000 sec CHANNEL f l ======== PI 13.00 usee PL1 0.00 dB SF01 300.1316000 MHz F2 - P r o c e s s i n g parameters SI 16384 SF 300.1300073 MHz HDW EM SSB 0 LB 1.00 Hz GB 0 PC 1.00 ID NMR p l o t parameters CX 30.00 cm F1P 8.000 ppm F l 2401.04 Hz F2P 0.000 ppm F2 0.00 Hz PPMCM 0.26667 ppm/cm HZCM .80.03467 Hz/cm ID Current Data Parameters NAME Oha-pept EXPNO 3 PROCNO 1 F2 - A c q u i s i t i o n Parameters D a t e . 20020124 Time 14.08 INSTRUM av400 PflOBHD 5 mm BBI 1H-PULPBOG zg30 TD 32768 SOLVENT 020 NS 32 DS 2 SWH 4990.020 Hz FIDRES 0.152283 Hz AO 3.2834036 sec RG 90.5 DW 100.200 usee DE 6.00 usee TE 300.0 K Dl 1.00000000 sec ======== CHANNEL f l ======== NUC1 1H PI 10.00 usee PL1 4.80 dB SF01 400.1317530 MHz F2 - P r o c e s s i n g parameters SI 32768 SF 400.1300115 MHz WDW EM SSB 0 LB 0.20 Hz GB 0 PC 1.00 ID NMR p l o t parameters CX 30.00 cm F1P 8.000 ppm F l 3201.04 Hz F2P 0.000 ppm F2 0.00 Hz PPMCM 0.26667 ppm/cm HZCM 106.70133 Hz/cm 00 OO C u r r e n t Data Parameters NAME P-pept2.4 EXPNO 1 PROCNO 1 F2 - A c q u i s i t i o n Parameters D a t e _ 20020812 Time 1.23 INSTRUM av300 PROBHD 5 mm QNP 1H/ PULPROG zg30 TD 32768 SOLVENT D20 NS 32 DS 2 3WH 4496.403 Hz FIDRES 0.137219 Hz AO 3.6438515 sec RG 114 DW 111.200 usee OE 6.00 usee IE 300.0 K Dl 1.00000000 sec LHANINLL r 1 NUC! 1H PI 13.00 usee PL1 0.00 dB SFOl 300.1316000 MHz F2 - P r o c e s s i n g parameters SI 16384 SF 300.1300119 MHz WDW EM SSB 0 LB 0.20 Hz GB 0 PC 1.00 ID NMR p l o t parameters CX 30.00 cm F1P 8.000 ppm F l 2401.04 Hz F2P 0.000 ppm F2 0.00 Hz PPMCM 0.26667 ppm/cm HZCM 80.03467 Hz/cm . ppm T 7 T 4 Current Data Parameters NAME P-pept2.5 EXPNO 1 PROCNO 1 F2 - Acqu Date_ Time INSTRUM PROBHD PULPROG TD SOLVENT NS DS SWH FIDRES AQ RG DW OE TE 01 i s i t i o n Parameters 20020812 0.59 av300 5 mm ONP 1H/ zg30 32768 020 32 2 1496.403 Hz 0.137219 Hz 3.6438515 sec 114 111.200 usee 6.00 usee 300.0 K 1.00000000 sec - « « « - - CHANNEL f l — — = » - -NUC! 1H PI 13.00 usee PL1 0.00 dB SF01 300.1316000 MHz F2 - Process i n g parameters SI 16384 SF 300.1300119 MHz WOW EH SSB 0 L8 0.40 Hz GB 0 PC 1.00 10 NMR p l o t parameters CX 30.00 cm F1P 8.000 ppm F l 2401.04 Hz F2P 0.000 ppm F2 0.00 Hz PPMCM 0.26667 ppm/cm HZCM 80.03467 Hz/cm k KJ K.A.J \ J LI o References 191 1. Freemantle, M . Chem. Eng. News 2003, 81, 27-30. 2. Marahiel, M . A. ; Stachelhaus, T.; Mootz, H . D. Chem. Rev. 1997, 97, 2651-74. 3. Tishchenko, G. N . ; Andrianov, V . I.; Vainshtein, B. K. ; Dodson, E. Bioorg. Khim. 1992, 18, 357-73. 4. Volkmann, R. A . ; Heck, S. D. in D-amino acids in sequences of secreted peptides of multicellular organisms; Jolles, P., Ed.; Birkhauser Verlag: Basel; Boston, Mass, 1998, pp 87-105. 5. Erspamer, V . ; Melchiorri, P.; Falconieri Erspamer, G.; Montecucchi, P. C.; de Castiglione, R. Peptides 1985, 6 Suppl. 3, 7-12. 6. Richter, K. ; Egger, R.; Kreil, G. Science 1987, 238, 200-2. 7. Kamatani, Y . ; Minakata, H. ; Kenny, P. T.; Iwashita, T.; Watanabe, K. ; Funase, K ; Sun, X . P.; Yongsiri, A . ; Kim, K . H. ; Novales-Li, P. Biochem. Biophys. Res. Commun. 1989, 160, 1015-20. 8. Ohta, N . ; Kubota, I.; Takao, T.; Shimonishi, Y . ; Yasuda-Kamatani, Y . ; Minakata, H. ; Nomoto, K. ; Muneoka, Y . ; Kobayashi, M . Biochem. Biophys. Res. Commun. 1991, 775, 486-93. 9. Yasuda-Kamatani, Y . ; Nakamura, M . ; Minakata, H. ; Nomoto, K.; Sakiyama, F. J. Neurochem. 1995, 64, 2248-55. 10. Yu, H . ; Rosen, M . K ; Saccomano, N . A . ; Phillips, D.; Volkmann, R. A . ; Schreiber, S. L . Biochemistry 1993, 32, 13123-9. 11. Heck, S. D.; Kelbaugh, P. R.; Kelly, M . E.; Thadeio, P. F.; Saccomano, N . A. ; Stroh, J. G.; Volkmann, R. A . J. Am. Chem. Soc. 1994,116, 10426-36. 12. Heck, S. D.; Siok, C. J.; Krapcho, K. J.; Kelbaugh, P. R.; Thadeio, P. F.; Welch, M . J.; Williams, R. D.; Ganong, A . H. ; Kelly, M . E.; Lanzetti, A . J.; Gray, W. R.; Phillips, D.; Parks, T. N . ; Jackson, H. ; Ahlijanian, M . K. ; Saccomano, N . A. ; Volkmann, R. A . Science 1994, 266, 1065-8. 13. Kreil, G. Science 1994, 266, 996-7. 14. Tanner, M . E.; Kenyon, G. L. in Comprehensive Biological Catalysis: A Mechanistic Reference; Sinnott, M . , Ed.; Academic Press: San Diego, 1998; Vol . 2, pp 7-41. 15. Rose, I. A . Annu. Rev. Biochem. 1966, 35, 23-56. 16. Adams, E. Adv. Enzymol. Relat. Areas Mol. Biol. 1976, 44, 69-138. 17. Yorifuji, T.; Ogata, K . J. Biol. Chem. 1971, 246, 5085-92. 192 18. Svensson, M . L . ; Gatenbeck, S. Arch. Microbiol. 1981,129, 213-15. 19. Uo, T.; Yoshimura, T.; Shimizu, S.; Esaki, N . Biochem. Biophys. Res. Commun. 1998, 246,31-4. 20. Wolosker, FL; Sheth, K . N . ; Takahashi, M . ; Mother, J. P.; Brady, R. O., Jr.; Ferris, C. D.; Snyder, S. H . Proc. Natl. Acad. Sci. USA 1999, 96, 721-5. 21. Inagaki, K. ; Tanizawa, K. ; Tanaka, H. ; Soda, K . Prog. Clin. Biol. Res. 1984, 144A, 355-63. 22. Soda, K. ; Osumi, T. Biochem. Biophys. Res. Commun. 1969, 35, 363-8. 23. Lim, Y . FL; Yokoigawa, K. ; Esaki, N . ; Soda, K . J. Bacteriol. 1993,175, 4213-7. 24. Wood, W. A. ; Gunsalus, I. C. J. Biol. Chem. 1951,190, 403-16. 25. Dunathan, H . C. Proc. Natl. Acad. Sci. USA 1966, 55, 712-6. 26. Henderson, L . L. ; Johnston, R. B. Biochem. Biophys. Res. Commun. 1976, 68, 793-8. 27. Shen, S. J.; Floss, H . G.; Kumagai, H. ; Yamada, H. ; Esaki, N . ; Soda, K. ; Wasserman, S. A. ; Walsh, C. J. Chem. Soc. Chem. Commun. 1983, 82-3. 28. Shaw, J. P.; Petsko, G. A . ; Ringe, D. Biochemistry 1997, 36, 1329-42. 29. Stamper, C. G. F.; Morollo, A . A. ; Ringe, D. Biochemistry 1998, 37, 10438-45. 30. Sun, S.; Toney, M . D. Biochemistry 1999, 38, 4058-65. 31. Seto, B.; Stadtman, T. C. J. Biol. Chem. 1976, 251, 2435-9. 32. Cardinale, G. J.; Abeles, R. H . Biochemistry 1968, 7, 3970-8. 33. Rudnick, G.; Abeles, R. H. Biochemistry 1975,14, 4515-22. 34. Fisher, L. M . ; Albery, W. J.; Knowles, J. R. Biochemistry 1986, 25, 2538-42. 35. Gallo, K . A. ; Tanner, M . E.; Knowles, J. R. Biochemistry 1993, 32, 3991-7. 36. Tanner, M . E.; Gallo, K . A. ; Knowles, J. R. Biochemistry 1993, 32, 3998-4006. 37. Jung, G. Angew. Chem., Int. Ed. Engl. 1991, 30, 1051-68. 38. Heck, S. D.; Faraci, W. S.; Kelbaugh, P. R.; Saccomano, N . A. ; Thadeio, P. F.; Volkmann, R. A . Proc. Natl. Acad. Sci. USA 1996, 93, 4036-9. 39. Williams, G.; Maziarz, E. P., 3rd; Amyes, T. L. ; Wood, T. D.; Richard, J. P. Biochemistry 2003, 42, 8354-61. 40. Glavas, S.; Tanner, M . E. Bioorg. Med. Chem. Lett. 1997, 7, 2265-70. 193 41. Gerlt, J. A. ; Gassman, P. G. J. Am. Chem. Soc. 1993,115, 11552-68. 42. Gerlt, J. A. ; Kreevoy, M . M . ; Cleland, W.; Frey, P. A . Chem. Biol. 1997, 4, 259-67. 43. Gerlt, J. A. ; Kozarich, J. W.; Kenyon, G. L. ; Gassman, P. G. J. Am. Chem. Soc. 1991, 113, 9667-9. 44. Rios, A. ; Amyes, T. L.; Richard, J. P. J. Am. Chem. Soc. 2000,122, 9373-85. 45. Bordwell, F. G.; Zhang, S.; Zhang, X . - M . ; Liu, W.-Z. J. Am. Chem. Soc. 1995,117, 7092-6. 46. Shikata, Y . ; Watanabe, T.; Teramoto, T.; Inoue, A. ; Kawakami, Y . ; Nishizawa, Y . ; Katayama, K. ; Kuwada, M . J. Biol. Chem. 1995, 270, 16719-23. 47. Jasys, V . J.; Kelbaugh, P. R.; Nason, D. M . ; Phillips, D.; Rosnack, K . J.; Saccomano, N . A. ; Stroh, J. G.; Volkmann, R. A . ; Forman, J. T. J. Org. Chem. 1992, 57, 1814-20. 48. Volkmann, R. A. , Efforts to Purify the Isomerase, Personal communication to Tanner, M . E., 1999. 49. Compute pI/Mw Tool, Expert Protein Analysis System (ExPASy): http://ca.expasy.org/tools/pi tool.html 50. Shikata, Y . ; Ohe, H. ; Mano, N . ; Kuwada, M . ; Asakawa, N . Biosci., Biotechnol, Biochem. 1998,62,1211-5. 51. Fowlkes, J.; Freemark, M . Endocrinology 1992,131, 2071-6. 52. Cheung, P. T.; Wu, J.; Banach, W.; Chernausek, S. D. Endocrinology 1994,135, 1328-35. 53. Liebecq, C , Ed. Biochemical Nomenclature and Related Documents; 2nd ed.; Portland Press, 1992. 54. Bianchi, E.; Steinkuhler, C ; Taliani, M . ; Urbani, A. ; Francesco, R. D.; Pessi, A . Anal. Biochem. 1996, 237, 239-44. 55. Kuisle, O.; Quinoa, E.; Riguera, R. J. Org. Chem. 1999, 64, 8063-75. 56. Gilon, C ; Klausner, Y . ; Hassner, A. Tetrahedron Lett. 1979, 40, 3811-4. 57. Sefler, A . M . ; Kozlowski, M . C ; Guo, T.; Bartlett, P. A . J. Org. Chem. 1997, 62, 93-102. 58. Zeggaf, C; Poncet, J.; Jouin, P.; Dufour, M . - N . ; Castro, B. Tetrahedron 1989, 45, 5039-50. 59. Coste, J.; Frerot, E.; Jouin, P. J. Org. Chem. 1994, 59, 2437-46. 60. Davies, J. S.; Howe, J.; Le Breton, M . J. Chem. Soc. Perkin Trans. 2 1995, 2335-9. 61. Humphrey, J. M . ; Chamberlin, A . R. Chem. Rev. 1997, 97, 2243-66. 194 62. Su, S.; Kakeya, H. ; Osada, H. ; Porco, J. A . Tetrahedron 2003, 59, 8931-46. 63. Fischer, E.; Raske, K. Ber. 1907, 40, 3717-24. 64. Wilchek, M . ; Zioudrou, C ; Patchornik, A. J. Org. Chem. 1966, 31, 2865-7. 65. Segel, I. H . Enzyme Kinetics: Behavior and Analysis of Rapid Equilibrium and Steady-State Enzyme Systems; Wiley-Interscience: New York, 1975. 66. Keenan, M . V. ; Alworth, W. L. Biochem. Biophys. Res. Commun. 1974, 57, 500-4. 67. Gelb, M . H. ; Lin, Y . ; Pickard, M . A. ; Song, Y . ; Vederas, J. C. J. Am. Chem. Soc. 1990, 112, 4932-42. 68. Baumann, R. J.; Bohme, E. H. ; Wiseman, J. S.; Vaal, M . ; Nichols, J. S. Antimicrob. Agents Chemother. 1988, 32, 1119-23. 69. Lam, L. K. ; Arnold, L . D.; Kalantar, T. H. ; Kelland, J. G.; Lane-Bell, P. M . ; Palcic, M . M . ; Pickard, M . A. ; Vederas, J. C. J. Biol. Chem. 1988, 263, 11814-9. 70. Okeley, N . M . ; Zhu, Y . ; van Der Donk, W. A . Org. Lett. 2000, 2, 3603-6. 71. Pansare, S. V. ; Arnold, L. D.; Vederas, J. C. Org. Synth. 1992, 70, 10-7. 72. Jiracek, J.; Yiotakis, A . ; Vincent, B.; Checler, F.; Dive, V . J. Biol. Chem. 1996, 271, 19606-11. 73. Jackson, P. F.; Cole, D. C ; Slusher, B . S.; Stetz, S. L. ; Ross, L . E.; Donzanti, B. A. ; Trainor, D. A . J. Med. Chem. 1996, 39, 619-22. 74. Karanewsky, D. S.; Badia, M . C ; Cushman, D. W.; DeForrest, J. M . ; Dejneka, T.; Loots, M . J.; Perri, M . G.; Petrillo, E. W., Jr.; Powell, J. R. J. Med. Chem. 1988, 31, 204-12. 75. Krapcho, J.; Turk, C ; Cushman, D. W.; Powell, J. R.; DeForrest, J. M . ; Spitzmiller, E. R.; Karanewsky, D. S.; Duggan, M . ; Rovnyak, G.; Schwartz, J.; Natarajan, S.; Godfrey, J. D.; Ryono, D. E.; Neubeck, R.; Atwal, K. S.; Petrillo, E. W., Jr. J. Med. Chem. 1988, 31, 1148-60. 76. McKittirck, B. A . ; Stamford, A . W.; Weng, X . ; Ma, K . ; Chackalamannil, S.; Czamiecki, M . ; Cleven, R. M . ; Fawzi, A . B. Bioorganic & Medicinal Chemistry Letters 1996, 6, 1629-34. 77. Lloyd, J.; Schmidt, J. B. ; Hunt, J. T.; Banish, J. C ; Little, D. K. ; Tymiak, A. A . Bioorg. Med. Chem. Lett. 1996, 6, 1323-6. 78. Chackalamannil, S.; Chung, S.; Stamford, A. W.; McKittrick, B . A . ; Wang, Y . ; Tsai, H . ; Cleven, R.; Fawzi, A . ; Czarniecki, M . Bioorg. Med. Chem. Lett. 1996, 6, 1257-60. 79. Caldwell, C. G.; Sahoo, S. P.; Polo, S. A. ; Eversole, R. R.; Lanza, T. J.; Mills, S. G.; Niedzwiecki, L. M . ; Izquierdo-Martin, M . ; Chang, B. C ; Harrison, R. K. ; Stein, R. L. ; Hagmann, W. K . Bioorg. Med. Chem. Lett. 1996, 6, 323-8. 195 80. Bartlett, P. A. ; Kezer, W. B. J. Am. Chem. Soc. 1984,106, 4282-3. 81. Bartlett, P. A. ; Hanson, J. E.; Giannousis, P. P. J. Org. Chem. 1990, 55, 6268-74. 82. Grobelny, D.; Wondrak, E. M . ; Galardy, R. E.; Oroszlan, S. Biochem. Biophys. Res. Commun. 1990,169, 1111-6. 83. Walker, B.; Wharry, S.; Hamilton, R. J.; Martin, S. L. ; Healy, A. ; Walker, B. J. Biochem. Biophys. Res. Commun. 2000, 276, 1235-9. 84. Cui, J.; Marankan, F.; Fu, W.; Crich, D.; Mesecar, A . ; Johnson, M . , E. Bioorg. Med. Chem. 2002, 10, 41-6. 85. Katoh, M . ; Hiratake, J.; Kato, H. ; Oda, J. Bioorg. Med. Chem. Lett. 1996, 6, 1437-42. 86. Chen, S.; Lin, C.-H.; Walsh, C. T.; Coward, J. K . Bioorg. Med. Chem. Lett. 1997, 7, 505-10. 87. Badet, B. ; Walsh, C. Biochemistry 1985, 24, 1333-41. 88. Badet, B. ; Inagaki, K. ; Soda, K. ; Walsh, C. T. Biochemistry 1986, 25, 3275-82. 89. Baylis, E. K. ; Campbell, C. D.; Dingwall, J. G. J. Chem. Soc. Perkin Trans. 1 1984, 2845-53. 90. Chen, H. ; Noble, F.; Mothe, A. ; Meudal, H . ; Coric, P.; Danascimento, S.; Roques, B. P.; George, P.; Fournie-Zaluski, M . C. J. Med. Chem. 2000, 43, 1398-408. 91. Vassiliou, S.; Mucha, A. ; Cuniasse, P.; Georgiadis, D.; Lucet-Levannier, K. ; Beau, F.; Kannan, R.; Murphy, G.; Knauper, V. ; Rio, M . C ; Basset, P.; Yiotakis, A . ; Dive, V . J. Med. Chem. 1999, 42, 2610-20. 92. Georgiadis, D.; Matziari, M . ; Yiotakis, A . Tetrahedron 2001, 57, 3471-8. 93. Krantz, A. ; Spencer, R. W.; Tarn, T. F.; Liak, T. J.; Copp, L . J.; Thomas, E. M . ; Rafferty, S. P. J. Med. Chem. 1990, 33, 464-79. 94. Shaw, E. in The Enzymes; 3rd ed.; Boyer, P. D., Ed.; Academic Press: New York, 1970; Vol . 1, pp 91-146. 95. Weiss, B.; Hui, K. S.; Toth, E.; Hui, M . ; Manigault, I.; Kenessey, A. ; Banay-Schwartz, M . ; Lajtha, A . Res. Commun. Mol. Pathol. Pharmacol. 1990, 15, 105-14. 96. Fray, A . H. ; Kaye, R. L. ; Kleinman, E. F. J. Org. Chem. 1986, 51, 4828-33. 97. Holladay, M . W.; Salituro, F. G.; Rich, D. H . J. Med. Chem. 1987, 30, 374-83. 98. Gennari, C ; Longari, C ; Ressel, S.; Salom, B.; Mielgo, A . Eur. J. Org. Chem. 1998, 945-59. 99. Ghosh, A . K. ; Fidanze, S. J. Org. Chem. 1998, 63, 6146-52. 196 100. Murkin, A . S.; Tanner, M . E. J. Org. Chem. 2002, 67, 8389-94. 101. Rios, A. ; Richard, J. P.; Amyes, T. L . J. Am. Chem. Soc. 2002,124, 8251-9. 102. Kaminski, Z. J. Synthesis 1987, 917-20. 103. Yiotakis, A . ; Vassiliou, S.; Jiracek, J.; Dive, V . J. Org. Chem. 1996, 61, 6601-5. 104. Bradford, M . M . Anal. Biochem. 1976, 72, 248-54. 105. Laemmli, U . K . Nature 1970, 227, 680-5. 106. Jarosz, S.; Salanski, P.; Mach, M . Tetrahedron 1998, 54, 2583-94. 107. McDonough, M . W.; Wood, W. A. J. Biol. Chem. 1961, 236, 1220-4. 108. Davis, L. ; Lee, N . ; Glaser, L . J. Biol. Chem. 1972, 247, 5862-6. 109. Jelakovic, S.; Kopriva, S.; Suss, K . H . ; Schulz, G. E. J. Mol. Biol. 2003, 326, 127-35. 110. Kopp, J.; Kopriva, S.; Suss, K. H. ; Schulz, G. E. J. Mol. Biol. 1999, 287, 761-71. 111. Samuel, J.; Tanner, M . E. Nat. Prod. Rep. 2002,19, 261-77. 112. Amein, M . ; Leatherwood, J. M . Biochem. Biophys. Res. Commun. 1969, 36, 223-7. 113. Luchansky, S. J.; Yarema, K. J.; Takahashi, S.; Bertozzi, C. R. J. Biol. Chem. 2003, 278, 8035-42. 114. Ghosh, S.; Roseman, S. J. Biol. Chem. 1965, 240, 1525-30. 115. Petersen, M . ; Fessner, W.; Frosch, M . ; Luneberg, E. FEMSMicrobiol. Lett. 2000,184, 161-4. 116. Frey, P. A. FASEB J. 1996,10, 461-70. 117. Tanner, M . E. in Comprehensive Biological Catalysis: A Mechanistic Reference; Sinnott, M . , Ed.; Academic Press: San Diego, 1998; Vol . 3, pp 76-82. 118. Tanner, M . E. Curr. Org. Chem. 2001, 5, 169-92. 119. Frey, P. A . in Pyridine Nucleotide Coenzymes: Chemical, Biochemical, and Medical Aspects; Dolphin, D., Poulson, R., Avramovic, O., Eds.; Wiley: New York, 1987; Vol . 2, pp 461-77. 120. Anderson, L. ; Landel, A . M . ; Diedrich, D. F. Biochim. Biophys. Acta 1956, 22, 573-5. 121. Kalckar, H . M . ; Maxwell, E. S. Biochim. Biophys. Acta 1956, 22, 588-9. 122. Koshland, D. E., Jr.; Kowalsky, A . Biochim. Biophys. Acta 1956, 22, 575-7. 123. Nelsestuen, G. L. ; Kirkwood, S. Biochim. Biophys. Acta 1970, 220, 633-5. 197 124. Adair, W. L. , Jr.; Gabriel, O.; Ullrey, D.; Kalckar, H . M . J. Biol. Chem. 1973, 248, 4635-9. 125. Maitra, U . S.; Ankel, H . Proc. Natl. Acad. Sci. USA 1971, 68, 2660-3. 126. Wee, T. G.; Davis, J.; Frey, P. A . J. Biol. Chem. 1972, 247, 1339-42. 127. Wee, T. G.; Frey, P. A. J. Biol. Chem. 1973, 248, 33-40. 128. Nelsestuen, G. L. ; Kirkwood, S. J. Biol. Chem. 1971, 246, 7533-43. 129. Creuzenet, C.; Belanger, M . ; Wakarchuk, W. W.; Lam, J. S. J. Biol. Chem. 2000, 275, 19060-7. 130. Mufloz, R.; Lopez, R.; de Frutos, M . ; Garcia, E. Mol. Microbiol. 1999, 31, 703. 131. Fan, D.-F.; Feingold, D. S. Methods Enzymol. 1972, 28, 422-5. 132. Barber, G. A. J. Biol. Chem. 1979, 254, 7600-3. 133. Salo, W. L.; Fossitt, D. D.; Bevill, R. D., I l l ; Kirkwood, S.; Wood, W. A . J. Biol. Chem. 1972, 247, 3098-100. 134. Deupree, J. D.; Wood, W. A . J. Biol. Chem. 1972, 247, 3093-7. 135. Deupree, J. D.; Wood, W. A . J. Biol. Chem. 1970, 245, 3988-95. 136. Lee, L . V . ; Vu, M . V. ; Cleland, W. W. Biochemistry 2000, 39, 4808-20. 137. Samuel, J.; Luo, Y . ; Morgan, P. M . ; Strynadka, N . C.; Tanner, M . E. Biochemistry 2001, 40, 14772-80. 138. Luo, Y. ; Samuel, J.; Mosimann, S. C.; Lee, J. E.; Tanner, M . E.; Strynadka, N . C. Biochemistry 2001, 40, 14763-71. 139. Johnson, A . E.; Tanner, M . E. Biochemistry 1998, 37, 5746-54. 140. Kawamura, T.; Kimura, M . ; Yamamori, S.; Ito, E. J. Biol. Chem. 1978, 253, 3595-601. 141. Kawamura, T.; Ishimoto, N . ; Ito, E. / . Biol. Chem. 1979, 254, 8457-65. 142. Campbell, R. E.; Mosimann, S. C ; Tanner, M . E.; Strynadka, N . C. J. Biochemistry 2000, 39, 14993-5001. 143. Salo, W. L. Biochim. Biophys. Acta 1976, 452, 625-8. 144. Morgan, P. M . ; Sala, R. F.; Tanner, M . E. J. Am. Chem. Soc. 1997,119, 10269-77. 145. Comb, D. G.; Roseman, S. Biochim. Biophys. Acta 1958, 29, 653. 146. Hinderlich, S.; Stasche, R.; Zeitler, R.; Reutter, W. J. Biol. Chem. 1997, 272, 24313-8. 198 147. Effertz, K. ; Hinderlich, S.; Reutter, W. J. Biol. Chem. 1999, 274, 28771-8. 148. Rutishauser, U . Curr. Opin. Cell Biol. 1996, 8, 679-84. 149. Yogeeswaran, G.; Salk, P. L . Science 1981, 212, 1514-6. 150. Keppler, O. T.; Hinderlich, S.; Langner, J.; Schwartz-Albiez, R.; Reutter, W.; Pawlita, M . Science 1999, 284, 1372-6. 151. Kornfeld, S.; Kornfeld, R.; Neufeld, E. F.; O'Brien, P. J. Proc. Natl. Acad. Sci. USA 1964, 52, 371-9. 152. Glaser, L . Biochim. Biophys. Acta 1960, 41, 534-6. 153. Chou, W. K. ; Hinderlich, S.; Reutter, W.; Tanner, M . E. J. Am. Chem. Soc. 2003,125, 2455-61. 154. Salo, W. L. ; Fletcher, H . G., Jr. Biochemistry 1970, 9, 882-5. 155. Sommar, K . M . ; Ellis, D. B. Biochim. Biophys. Acta 1972, 268, 590-5. 156. Angata, T.; Varki, A . Chem. Rev. 2002,102, 439-69. 157. Schauer, R. Sialic Acids: Chemistry, Metabolism and Function; Springer-Verlag: New York, 1982. 158. Glode, M . P.; Sutton, A . ; Robbins, J. B. ; McCracken, G. H. ; Gotschlich, E. C ; Kaijser, B.; Hanson, L. A . J. Infect. Dis. 1977,136Suppl, S93-7. 159. Gilbert, M . ; Karwaski, M.-F.; Bernatchez, S.; Young, N . M . ; Taboada, E.; Michniewicz, J.; Cunningham, A . - M . ; Wakarchuk, W. W. J. Biol. Chem. 2002, 277, 327-37. 160. Alios, B. M . Clin. Infect. Dis. 2001, 32, 1201-6. 161. Bliss, J. M . ; Silver, R. P. Mol. Microbiol. 1996, 21, 221-31. 162. Vimr, E.; Steenbergen, S.; Cieslewicz, M . J. Ind. Microbiol. 1995,15, 352-60. 163. Whitfield, C ; Roberts, I. S. in Mol. Microbiol.; Blackwell Publishing Limited, 1999; Vol . 31, p 1307. 164. Frosch, M . ; Edwards, U . ; Bousset, K. ; Krausse, B.; Weisgerber, C. Mol. Microbiol. 1991, 5, 1251-63. 165. Ringenberg, M . A. ; Steenbergen, S. M . ; Vimr, E. R. Mol. Microbiol. 2003, 50, 961-75. 166. Vann, W. F.; Daines, D. A. ; Murkin, A . S.; Tanner, M . E.; Chaffin, D. O.; Rubens, C. E.; Vionnet, J.; Silver, R. P. J. Bacteriol. 2004,186, 704-12. 167. Daines, D. A. ; Vann, W. F.; Chaffin, D. O.; Rubens, C. E.; Guerry, P.; Silver, R. P. "The NeuC Protein of Escherichia coli K l is a Uridine Diphosphate N-Acetylglucosamine 2-Epimerase," unpublished manuscript, 2002. 199 168. Reutter, W., Personal communication to Tanner, M . E., 2002. 169. IMPACT I: One-Step Protein Purification System Instruction Manual, v. 1.8, New England Biolabs, Inc. 170. Chong, S.; Shao, Y . ; Paulus, H. ; Benner, J.; Perler, F. B.; Xu, M . Q. J. Biol. Chem. 1996, 277,22159-68. 171. Pravdic, N . ; Franjic-Mihalic, I.; Danilov, B. Carbohydr. Res. 1975, 45, 302-6. 172. Lai, E. C ; Withers, S. G. Biochemistry 1994, 33, 14743-9. 173. Zapata, G.; Crowley, J. M . ; Vann, W. F. J. Bacteriol. 1992, 174, 315-9. 174. Kasper, D. L. ; Winkelhake, J. L. ; Zollinger, W. D.; Brandt, B. L. ; Artenstein, M . S. J. Immunol. 1973,110, 262-8. 175. Warren, L . J. Biol. Chem. 1959, 234, 1971 -5. 176. Swartley, J. S.; Stephens, D. S. J. Bacteriol. 1994,176, 1530-4. 177. Ganguli, S.; Zapata, G.; Wallis, T.; Reid, C ; Boulnois, G.; Vann, W. F.; Roberts, I. S. J. Bacteriol. 1994,176, 4583-9. 178. Tettelin, H. ; Saunders, N . J.; Heidelberg, J.; Jeffries, A . C ; Nelson, K. E.; Eisen, J. A . ; Ketchum, K . A. ; Hood, D. W.; Peden, J. F.; Dodson, R. J.; Nelson, W. C ; Gwinn, M . L. ; DeBoy, R.; Peterson, J. D.; Hickey, E. K. ; Haft, D. H. ; Salzberg, S. L. ; White, O.; Fleischmann, R. D.; Dougherty, B . A . ; Mason, T.; Ciecko, A . ; Parksey, D. S.; Blair, E.; Cittone, H. ; Clark, E. B.; Cotton, M . D.; Utterback, T. R.; Khouri, H. ; Qin, H. ; Vamathevan, J.; Gi l l , J.; Scarlato, V . ; Masignani, V. ; Pizza, M . ; Grandi, G.; Sun, L. ; Smith, H . O.; Fraser, C. M . ; Moxon, E. R.; Rappuoli, R.; Venter, J. C. Science 2000, 287, 1809-15. 179. Edwards, U . ; Muller, A. ; Hammerschmidt, S.; Gerardy-Schahn, R.; Frosch, M . Mol. Microbiol. 1994,14, 141-9. 180. Thiem, J.; Wiesner, M . Synthesis 1988, 124-7. 181. Sala, R., Mechanistic studies on UDP-N-acetylglucosamine 2-epimerase, Ph.D. Thesis, University of British Columbia, 1999. 182. Monod, J.; Wyman, J.; Changeux, J. P. J. Mol. Biol. 1965,12, 88-118. 183. Koshland, D. E., Jr.; Nemethy, G.; Filmer, D. Biochemistry 1966, 5, 365-85. 184. Fersht, A . Enzyme Structure and Mechanism; 2nd ed.; W. H . Freeman and Company: New York, 1985. 185. Good, N . E.; Winget, G. D.; Winter, W.; Connolly, T. N . ; Izawa, S.; Singh, R. M . M . Biochemistry 1966, 5, 467-77. 200 186. Bates, R. G.; Hetzer, H . B. J. Phys. Chem. 1961, 65, 667-71. 187. McNaught, A. D.; Wilkinson, A . Compendium of Chemical Terminology; 2nd ed.; Blackwell Science: Oxford, 1997. 188. Melander, L. C. S.; Saunders, W. H. Reaction Rates oflsotopic Molecules; Wiley-Interscience: New York, 1980. 189. Zyryanov, A . B. V. , A . V . ; Salminen, A. ; Goldman, A. ; Lahti, R.; Baykov, A . A . Biochemistry 2004, 43, 1065-74. 190. Paneth, P. Acc. Chem. Res. 2004, 36, 120-6. 191. Singleton, D. A. ; Thomas, A. A . J. Am. Chem. Soc. 1995,117, 9357-8. 192. Parkin, D. W. in Enzyme Mechanism from Isotope Effects; Cook, P. F., Ed.; CRC Press: Boca Raton, 1991, pp 269-90. 193. Albery, J. in Proton-Transfer Reactions; Cladin, E., Gold, V. , Eds.; Chapman and Hall: London, 1975, pp 263-315. 194. Albery, W. J.; Davies, M . H . Trans. Farad. Soc. 1969, 65, 1059. 195. Gold, V . Adv. Phys. Org. Chem. 1969, 7, 259-331. 196. Kresge, A . J. Pure Appl. Chem. 1964, 8, 243. 197. Schowen, R. L. in Isotope Effects on Enzyme-Catalyzed Reactions; Cleland, W. W., O'Leary, M . H. , Northrop, D. B., Eds.; University Park Press: Baltimore, 1977, pp 64-99. 198. Midelfort, C. F.; Rose, I. A . / . Biol. Chem. 1976, 251, 5881-7. 199. Cohn, M . ; Hu, A. Proc. Natl. Acad. Sci. USA 1978, 75, 200-3. 200. Cohn, M . ; Hu, A. J. Am. Chem. Soc. 1980,102, 913-6. 201. Gunawan, J.; Strynadka, N . , Essential Oligosaccharide Metabolizing Enzymes in Pathogenic Bacteria, Personal communication to Tanner, M . E., 2003. 202. Samuel, J.; Tanner, M . E. Biochim. Biophys. Acta 2004, in press. 203. Guerry, P.; Ewing, C. P.; Hickey, T. E.; Prendergast, M . M . ; Moran, A. P. Infect. Immun. 2000, 68, 6656-62. 204. Glasoe, P. K.; Long, F. A . J. Am. Chem. Soc. 1960, 60, 188-90. 205. Inoue, H. ; Nojima, H. ; Okayama, H. Gene 1990, 96, 23-8. 206. Sandhu, G. S.; Precup, J. W.; Kline, B. C. Biotechniques 1989, 7, 689-90. 207. Gussow, D.; Clackson, T. Nucleic Acids Res. 1989,17, 4000. 201 208. GraFit, v. 4.0.18, Erithacus Software, Ltd. 209. Thompson, J. D.; Higgins, D. G.; Gibson, T. J. Nucleic Acids Res. 1994, 22, 4673-80. 210. Henikoff, S.; Henikoff, J. G. Proc. Natl. Acad. Sci. USA 1992, 89, 10915-9. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0061206/manifest

Comment

Related Items