é” ‘ .. .J.’ .. . ‘AI. . . 9.3.1.2. . st» ALL); .1“;. la! 9.107» I:..'l)... 5 B. ‘1! 10- .0... ‘3 o)“ i a: 1.. n: a“? ,.Obl.f>‘o (”a .rr v .1 . to! 3.4. . z n .5 . . - {ltlalol‘t V .fl..pnv; ear-I .io~vl 5dr. Yr-IA I’D l3 0‘8 1-. i‘:’? 75.5." 35.57 .1... 45". If»: ‘ It‘ll! l- , 7 FM! f. if! 1 . 3"" ,W.‘" I ‘él I. .14.! ‘.‘ IL in... 3:... ’1‘; )Pa‘.“ 1 x a1).l..$.la.t.xvl {I4 rt. 1.. Xuvliclzezfiguv 1. IS i‘ull" £14! I! . ' 'A :3..me .flhnfikuc. "HU‘ ‘ I If." . lvau ”Lil I» on‘r. It! “lufrrfl .’ it. rfilo(0'n.’l 3.!va n. ‘vnib‘vnll‘lk git fill; 5". to." it! DP!!! .galil .‘II: II in . »4| .Irll. col" .l‘ {I ‘3‘)!!! 3”} Q!!! l 1?. I! '0‘] lllllllllllllllllllllllllllllllllll'llllllllllllllllllll 300908 7762 This is to certify that the I dissertation entitled DERIVATIVES FOR ISOLATION, PURIFICATION AND MASS SPECTROMETRI CHARACTERIZATION OF HOMO AND HETERO OLIGOSACCHARIDES STRUCTURE ELUCIDATION AND SYNTHETIC TRANSFORMATION RELATED TO THE INNER AND OUTER CORE OF THE LIPOPOLYSACCHARIDES presentedby II YUANDA ZHANG has been accepted towards fulfillment of the requirements for Ph.b CHEMISTRY __ _ __ degree in m In Major Vprofesso/ Date ol/ZOIOIX MSU i: an Affirmative Action/Equal Opportunity Institution 0-12771 LIBRARY #— Michlgan State Unlverslty 1 PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. DATE DUE DATE DUE DATE DUE r AUJ zri’fjug I - (1.2 ' ’ [—— [__= _JLJ MSU Is An Affirmative Action/Equal Opportunity lnditution cMMuns-oj PART I: DERIVATIVES FOR THE ISOLATION, PURIFICATION AND MASS SPECTROMETRIC CHARACTERIZATION OF HOMO AND HETERO OLIGOSACCHARIDES PART II: STRUCTURE ELUCIDATION AND SYNTHETIC TRANSFORMATIONS RELATED TO THE INNER AND OUTER CORE OF THE LIPOPOLYSACCHARIDES OF RHIZOBI UM MELILOTI 41 AND LEGUMINOSAR UM BIOVAR VICIAE VF-39 MUTANTS By Yuanda Zhang A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1992 ABSTRACT PART I: DERIVATIVES FOR THE ISOLATION, PURIFICATION AND MASS SPECTROMETRIC CHARACTERIZATION OF HOMO AND HETERO OLIGOSACCHARIDES 13y YuandaZhang A simple, sensitive method for the structural characterization of Oligosaccharides by fast atom bombardment mass spectrometry (FAB-MS) has been designed. Oligosaccharides are labeled with a UV chromophore (which also serves as a charge-stabilizing group) and with a hydrophobic alkyl tail. The chromophore, 2,4-dinitrophenyl or dansyl group, aids UV detection during high performance liquid chromatography (HPLC) and stabilizes ion species formed during analysis by FAB-MS. The hydrophobic tail, provided by an octyl group, enhances the surface activity of the analytes and makes them amenable to separation by reverse-phase chromatography using a C18-bonded phase. This method was applied to the structural analysis of homooligosaccharides, namely a mixture of starch maltodextrins with a degree of polymerization 1-16, potato starch, pure maltohexaose, isomaltohexaose, and N,N’,N”-triacety1chitotriose. The method was also applied to the structure of heterooligosaccharides, namely lacto-N-fucopentaose, lacto-N-difucohexaose (these two Oligosaccharides were from human milk), glycoprotein fetuin, and a previously characterized oligosaccharide from a Rhizobial capsular polysaccharide (1). The method gave a good yield of ions for the derivatized compounds, which in the best cases, were detectable at a level of about 1 picomole. In the case of maltohexaose, four series of sequence anions corresponding to sequential loss of glycosyl residues from the reducing and non-reducing ends by different mechanisms were observed. The mixture of derivatized malto-oligosaccharides was easily separated by HPLC. Based on the relative proportions of the individual oligomers in the mixture calculated from HPLC analysis, even though the higher oligomers were present in amounts of about 0.1%, they were still easily detected in mass spectra of the entire mixture. This represents an improvement in sensitivity of at least IOO-fold over that reported using the aminobenzoic acid alkyl ester method. In the analyses reported here, single scans were taken and no attempts were made to improve signal to noise by signal averaging. I would like to dedicate this thesis to my family, for their support and patient understanding, for their constant encouragement and love. ACKNOWLEDGMENTS I would like to express my most sincere appreciation to my research advisor Dr. Rawle I. Hollingsworth for his support, encouragement and guidance in conducting my graduate research, for sharing his knowledge and experience with me. I would also like to thank Dr. C. C. Sweeley and Dr. J. Preiss for their interest in my research, collaborative work and help. I am also grateful to Dr. David G. McConnell for his useful advice, especially for suggesting that Dr. Rawle I. Hollingsworth be my Ph.D. research advisor, an important step in my scientific career. I am very grateful for the support and encouragement received from Dr. C. K. Chang, Dr. G. Babcock, and Dr. D. Nocera. I wish to thank the members of Dr. Rawle I. Hollingsworth’s group: Mr. Kim K. I, Jung S. H., Debbie L. E., Maria B. B., Chuck Cory Campbell, Tim J. N., Baojen S., Lars P., Robert A. C., Luc B., and Ben H., for many informative discussions. The group meetings, group picnic, group tennis and group pizza lunches will make 116 biochemistry building unforgettable. I also wish to thank the members of the Mass Spectrometry Facility: Dr. J. T. Watson, Dr. D. A. Gage, Bev, Mel, Melinda and Mike, for their hard work with instrumental maintenance and for the free facility lessons, which made the JEOL HXl 10 more enjoyable. Finally I wish to thank the Departments of Chemistry and Biochemistry at Michigan State University for using their facilities. I would like to express my gratitude to the US Department of Energy for their support through a research assistantship from Dr. Rawle I. Hollingsworth’s research grants. TABLE OF CONTENTS Page List of Figures xv List of Abbreviations xxi PART I: DERIVATIVES FOR THE ISOLATION, PURIFICATION AND MASS SPECTRONIETRIC CHARACTERIZATION OF HOMO AND HETERO OLIGOSACCHARIDES CHAPTER 1 BACKGROUNDRI‘BEARCH ......................................................... 1 1.1 Introduction ....................................................................... 1 1.2 The Classical Method of Sequencing Carbohydrates ................ 1 1.3 Electron Ionization Mass Spectrometry .................................. 4 1.4 Chemical Ionization Mass Spectrometry ................................ 5 1.5 Fast Atom Bombardment Mass Spectrometry ........................ 6 1.5.1 Operating Process ...................................................... 7 1.5.2 Sample Preparation .................................................... 8 1.5.3 Matrix ....................................................................... 8 1.6 The Methods in Carbohydrate Analysis .................................. 10 1.7 Classical Methods for Carbohydrate Derivatives ...................... 12 1.8 Derivatization by Chromophore Groups ................................. 13 1.9 Nomenclature ..................................................................... 17 1.9.1 Nomenclature of Carbohydrate Fragments .................... 17 1.9.2 Branched Oligosaccharide Moieties .............................. 19 1.10 Sequences from Oligosaccharides .......................................... Z) 1.11 References .......................................................................... 22 CHAPTER2: ................................................................................ % 2.1 General Methods ................................................................. % 2.1.1 Chemicals .......................................................................... % 2.1.2 The General Procedure for making Derivatives ....................... 5 2.1.2.1 Preparation of N,N-(2,4-dinitrophenyl)octylamine ........ Derivatives .............................................................. % 2.1.2.2 Preparation of N,N-Dansyloctylamine Derivatives ........ % 2.1.3 Instrumentation ................................................................. 27 2.1.3.1 High Performance Liquid Chromatography ................ 27 2.1.3.2 Mass Spectrometry ................................................... 27 2.2 New Approaches to Sequencing Oligosaccharides by negative Mass Spectrometry .............................................................. 29 2.2.1 N,N-(2,4-dinitrophenyl)octylamine Derivatives for the Negative Mass Spectrometric Characterization of Maltohexaose ............ 3 2.2.2 N,N-(2,4-dinitropheny1)octylamine Derivatives for the Isolation, Purification and Negative Mass Spectrometric Characterization of Oligosaccharide Mixture Maltodextrin ............................... $ 2.2.3 N ,N-(2,4-dinitrophenyl)octylamine Derivatives for the Negative Mass Spectrometric Characterization of an Oligosaccharide Isolated from a Bacterial Cell Surface .................................... 40 2.2.4 Sequence Specific Cleavage of N ,N -(2,4-dinitrophenyl)octylamine Derivatives of Hetero and Homooligosaccharides by Negative Ion Fast Atom Bombardment Mass Spectrometry ......................... 42 2.2.4.1 Negative FAB-MS of Isomaltohexaose ......................... 43 2.2.5 2.2.6 2.2.4.2 Negative FAB-MS of N ,N’,N”-Triacetylchitotriose ........ 46 2.2.4.3 Negative FAB-MS of Lacto-N-Fucopentaose ................. 49 2.2.4.4 Negative FAB-MS of Lacto-N-Difucohexaso ................. 51 Sequencing of DNPO Derivatized Starch by Negative FAB-MS. . .54 2.2.5.1 Introduction ............................................................ 54 2.2.5.2 Structure of Amylose ................................................ 54 . 2.2.5.3 Structure of Amylopectin .................................. 55 2.2.5.4 Phosphorus in Starch Fractions ................................. 57 2.2.5.5 Negative FAB-MS Of DNPO Derivatized Oligosaccharides Digested from Potato Starch ............. 59 2.2.5.6 31P NMR Spectra of Phosphorus Contained Starch ....... 61 2.2.5.7 Negative FAB-MS of Synthesized Maltohexaose-6- ........ Phosphate ............................................................... 64 Negative FAB-MS of N,N-(2,4-dinitrophenyl)octylamine 2.2.7 Derivatized Oligosaccharide Isolated from Glycoprotein Fetuin67 N,N-dansyloctylamine Derivatives for the Isolation, Purification and Mass Spectrometric Characterization of Oligosaccharidesw 2.2.7.1 2.2.7.2 2.2.7.3 2.2.7.4 2.2.7.5 The Procedure to Prepare N,N-dansyloctylamine Derivatives of Oligosaccharides .................................. 70 Positive FAB-MS of N,N-dansyloctylamine Derivatized Maltose .................................................................. 70 Positive FAB-MS of N,N-dansyloctylamine Derivatized Maltohexaose ........................................................... 71 Positive FAB-MS of N ,N -dansyloctylamine Derivatized Maltodextrin ............................................................ 72 Positive FAB-MS of N,N-dansyloctylamine Derivatized. Oligosaccharide from the Bacteria Surface .................. 74 viii 2.2.8 Conclusion ......................................................................... 76 2.3 References .......................................................................... 78 CHAPTERS EXPERIMENTAL FOR PART I ...................................................... 81 3.1 Preparation of N,N-(2,4-dinitrophenyl)octylamine Derivatized Maltose .............................................................................. 81 3.2 UV Absorbance of N,N-(2,4-dinitrophenyl)octylamine Derivatized Maltose .............................................................................. 82 3.3 Preparation of N,N-(2,4-dinitrophenyl)octylamine Derivatized Maltohexaose ...................................................................... 82 3.4 Preparation of N,N-(2,4-dinitrophenyl)octylamine Derivatized Maltodextrin ....................................................................... 82 3.5 UV Absorbance of N ,N -(2,4-dinitrophenyl)octylamine Derivatized Maltodextrin ....................................................................... 83 3.6 Preparation of N,N-(2,4-dinitrophenyl)octylamine Derivatized Oligosaccharide 843 EPS Separated from Bacteria Surface ....... 83 3.7 Preparation of N ,N-dansyloctylamine Derivatized Maltose ....... 83 3.8 Preparation of N,N—dansyloctylamine Derivatized Maltodextrin ....................................................................... 84 3.9 Preparation of N,N—dansyloctylamine Derivatized Oligosaccharide Separated From the Bacteria Surface ..................................... 84 3.10 Preparation of N ,N -dansyloctylamine Derivatized Maltohexoase (Method H) ......................................................................... 84 3.11 Preparation of N,N-(2,4-dinitrophenyl)octylamine Derivatized Maltohexoase (Method I) ...................................................... 85 ix 3.12 3.13 3.14 3.15 3.16 3.17 3.18 3.19 3.20 3.21 3.22 3.3 Preparation of N,N—dansyloctylamine Derivatized Maltohexoase (Method I) ........................................................................... 85 Preparation of 6-phosphate Maltohexoase ............................... 85 Preparation of N ,N-(2,4-dinitrophenyl)octylamine Derivatized N,N’,N”-Triacetylchitotriose ................................................. 86 Preparation of N,N-(2,4-dinitrophenyl)octylamine Derivatized Lacto-N-Fucopentoase .......................................................... 86 Preparation of N,N-(2,4-dinitrophenyl)octylamine Derivatized Lacto-N-difucohexoase ......................................................... 86 Preparation of N,N-(2,4-dinitropheyl)octylamine Derivatized Isomaltohexoase ................................................................. 87 Enzyme Digestion of Starch I ................................................ 87 Acid Hydrolysis of Starch ..................................................... 87 Enzyme Digestion of Potato Starch II ..................................... 87 Preparation of N,N-(2,4-dinitr0phenyl)octylamine Derivatized Oligosaccharide isolated form Glycoprotein Fetuin .................. 88 Mass Spectrometry .............................................................. 8 References .......................................................................... as PART II: STRUCTURE ELUCIDATION AND SYNTHETIC TRANSFORMATIONS RELATED To THE INNERAND OUTER CORE 0!" THE LIPOPOLYSACCHARIDES OF RHIZOBIUM MELILOTI 41 AND LEGUMHVOSARUM BIOVAR VICIAE vr-ss MUTANTS Abstract ............................................................................. xxiii CHAPTER 4 STRUCTURE ELUCIDATION AND SYNTHETIC TRANSFORMATION RELATED TO THE INNER AND OUTER CORE IN THE LIPOPOLYSACCHARIDES OF RHIZOBIUM MELILO’TI 4 ....................................................................... 90 4.1 Introduction ....................................................................... so 4.2 Material and General Methods ............................................. 94 4.2.1 Further Purification of the Core Oligosaccharide ........... 94 4.2.2 Composition Analysis ................................................. 94 4.2.3 Methylation Analysis .................................................. 94 4.3 Results and Discussion ........................................................ % 4.3.1 Structure Elucidation of the Inner and Outer Core Rhizobium Meliloti 41 ................................................. % 4.4 Synthesis of 3-Deoxy-Heptitol Derivative ................................. 108 4.4.1 The First Pathway to Synthesize 3-deoxy-heptitol Derivative .................................................................. 108 4.4.2 Second Pathway to Synthesize 3-Deoxy-Heptitol Derivatives ................................................................. 112 4.4.3 The Pathway to Synthesize 3-Deoxy-D-Erythro—2,4-Heptodiulosonic Acid ................................................................................. 113 xi 4.5 References .......................................................................... 115 CHAPTER5 EXPERIMENTAL FORCHAPTER4 ............................................... 117 5.1 General Methods in Structure Elucidation ............................. 117 5.1.1 Microorganisms .................................................................. 117 5.1.2 Methylsulphinyl Methyl Sodium Preparation ......................... 117 5.1.3 Composition Analysis .......................................................... 117 5.1.4 Colorimetric Methods for Carbohydrate Detection ................... 118 5.1.5 Composition Analysis .......................................................... 119 5.1.6 Proton and Carbon-13 NMR Analysis ..................................... 121 5.1.7 GC Analysis ....................................................................... 121 5.1.8 GC-MS Analysis .................................................................. 121 5.1.9 Other Facilities .................................................................... 121 5.2 Synthetic Transformation ..................................................... 122 5.2.1 2-Deoxy-D-Arabino-hexose Propylene Dithioacetal ................... 122 5.2.2 2-Deoxcy-3,4,5,6-Tetrabenzyl-D-Arabino-Hexose Propylene Dithioacetal ................................................................ 123 5.2.3 3-Deoxy-l-Iodo-4,5,6,7-Tetrabenzyl-D-Arabino-Heptose Propylene Dithioacetal ......................................................... 123 5.2.4 3-Deoxy-4,5,6,7-Tetrabenzyl-D-Arabino-Heptulose Propylene Dithioacetal ........................................................................ 124 5.2.5 3-Deoxy-4,5,6,7-Tetrabenzyl-D-Arabino-Heptulose .................... 124 5.2.6 3-Deoxy-2-Deutero—4,5,6,7-Tetrabenzy1-D-G1uco (Manna)- Heptltol .......... 124 5.2.7 3—Deoxy-2-Deutero-D-Gluco (Manno)-Heptitol .......................... 125 xii 5.2.8 3-Deoxy-2-Deutero-1,2,4,5,6,7-hexoacetyl-D-Gluco (Manno)- Heptitol .............................................................................. 1% 5.2.9 2-Deoxy-3,4 : 5,6-Di-0-isopropylidene-D-Arabino-Hexose— Propylene Dithioacetal ......................................................... 1% 5.2.10 3-Deoxy-1-Iodo-4,5 : 6,7-Di-0-isopropylidene—D-Arabino-Heptose— Propylene Dithioacetal ......................................................... 1% 5.2.11 4,6-0-Propylidene-D-Glucose ................................................. 127 5.2.12 2,4-0-Propylidene-Erythrose .................................................. 127 5.2.13 2,4-0-Propylidene-D-Erythrose Propylene Dithioacetal ............. 128 5.2.14 2,4-O-Propylidene-3-0-Acetyl-D-Erythrose Propylene Dithioacetal ........................................................................ 129 5.3 References .......................................................................... 1% CHAPTER6 CHARACTERIZATION OF STRUCTURAL DEFECTS IN THE LIPOPOLYSACCHARIDE OF SYMBIOTICALLY IMPAIRED RHIZOBIUM LEGUMHVOSARUM BIOVAR VICIAE VF-39 MUTANTS 6.1 Abstract ........................................................ . ..................... 131 6.2 Introduction ....................................................................... 131 6.3 Experimental ...................................................................... 133 6.3.1 Culture of Bacteria and Isolation of LPS and LPS Fragments ................................................................. 133 6.3.2 Compositional Analysis .............................................. 133 6.3.3 Structural Analyses .................................................... 134 6.4 Results and Discussion ........................................................ 134 6.5 References .......................................................................... 146 xiii CHAPTER7 7.1 7.2 7.3 7.4 ISOLATION AND TENTATIVE CHARACTERIZATION OF A TRISACCHARIDE CONTAINING A NEW 2-DEOXY-3-C- (HYDROXYMETHYLWENTOFURANOSYL COMPONENT FROM THE LIPOPOLYSACCHARIDE OF A RHIZOBIUM LEGUMHVOSARUM BIOVAR VICIAE LPS MUTANT Introduction ....................................................................... 148 Experimental......................................... ............................. 148 7.2.1 Culture of Bacteria and Isolation of LPS and LPS Fragments .......................................................................... 148 7.2.2 Compositional Analysis .............................................. 149 7.2.3 Methylation Analyses ................................................. 149 7.2.4 Structural Analysis .................................................... 150 Results and Discussion ........................................................ 150 References .......................................................................... 168 xiv Figure Page 1.1 The classical method of sequencing Oligosaccharides ............ 2 1.2 View of the matrix surface during negative FAB-MS process . 9 1.3 Nomenclature of carbohydrate fragments ........................... 18 1.4 The numbering of the ring bonds ........................................ 18 1.5 The nomenclature for branched oligosaccharide .................. 19 1.6 Oligosaccharide fragmentation pathways ........................... 21 2.1 The procedure for making DNPO derivatives ....................... 30 2.2 Negative Mass spectrum of N,N-(2,4-dinitrophenyl)octylamine (DNPO) derivatized maltohexaose ....................................... 31 2.3a The Yn sequence from DNPO derivatized maltohexaose ........ 32 2.3b The sequence 2»“A,, of DNPO derivatized maltohexaose ......... 33 2.3c The sequence 15X,, of DNPO derivatized maltohexaose .......... 34 2.3d The sequence 1:4An of DNPO derivatized maltohexaose ......... 35 2.4a-b Negative Mass spectra of DNPO derivatized maltodextrin ..... 37 2.5 Sequence 0,, of DNPO derivatized maltodextrin .................... {D 2.6 Structure of previously characterized, and derivatized oligosaccharide from Rhizobium trifolii 843 capsular polysaccharide ................................................................. 40 2.7 Negative Mass spectrum of previously characterized, and derivatized oligosaccharide from Rhizobium trifolii 843 capsular polysaccharide .................................................... 41 2.8 Mechanism for rearrangement of acetal bond to produce LIST OF FIGURES pyruvate, formaldehyde and an unsaturated glycosyl residue ............................................................................ 41 2.9 The negative FAB-MS of DNPO derivatized isomaltohexaose .42 2.10 Yn type fragmentation of derivatized isomaltohexaose .......... 44 2.11 Cleavage pattern for 0:4An sequence in maltohexaose ........... 45 2.12 The structure of DNPO derivative of N,N’,N”- triacetylchitotriose ............................................................ 46 2.13 The negative FAB-MS of N,N’,N”-triacetylchitotriose ............ 47 2.14 The cleavage pattern of 0:4Xn Sequence ................................ 48 2.15 The formation of cyanoborohydride adduct .......................... 48 2.16 The cleavage pattern of the DNPO derivative of lacto-N- fucopentaose .................................................................... 49 2.17 The negative FAB-MS of lacto-N-fucopentaose ...................... 50 2.18 The fragmentation pattern of derivatized lacto-N- difucohexaose .................................................................. 52 2.19 The negative FAB-MS of derivatized lacto-N-difucohexaose . ..52 2.20 The molecular structure of amylopectin .............................. 56 2.21 The cluster model of amylopectin ....................................... 57 2.22a The negative FAB-MS of N,N-(2,4-dinitrophenyl)octylamine derivative of Oligosaccharides digested form potato starch, range from 500-1550 .......................................................... fl) 2.22b The negative FAB-MS of N ,N-(2,4-dinitrophenyl)octylamine derivative of Oligosaccharides digested from potato starch, range from 1550-3000 ......................................................... so 2.23 Phosphorus-31 NMR spectrum of phosphorus containing starch ............................................................................. 62 2.24 Non-decoupled phosphorus-31 NMR spectrum of degraded starch ............................................................................. 62 xvi 2.25 Expanded non-decoupled phosphorus-31 NMR spectrum of degraded starch and setting line broadening to 1 Hz ............. 63 2.26 The negative FAB-MS of synthesized phosphorylated maltohexaose............... .................................................... 64 2.27 The negative FAB-MS of N,N-(2,4-dinitrophenyl)octylamine derivatized oligosaccharide isolated from fetuin ................... 68 2.28 Preparation of N,N’-dansyloctylamine derivatives ................ Q 2.29 (+) FAB-MS of N,N-dansyloctylamine derivatized maltose ..... 70 2.30 (+) FAB-MS of N,N-dansyloctylamine derivatized Maltohexaose ................................................................... 71 2.31a (+) FAB-MS of N,N-dansyloctylamine derivatized Maltodextrin range from m/z 520-1800 .................................................... 73 2.31b (+) FAB-MS of N ,N -dansyloctylamine derivatized Maltodextrin range from m/z 1800-3000 .................................................. 73 2.32 N,N—dansyloctylamine derivatized oligosaccharide from a bacteria cell surface .......................................................... 74 2.33 FAB-MS of N ,N-dansyloctylamine derivatized oligosaccharide from bacteria surface ........................................................ 75 3.1 UV absorbance of N,N-(2,4-dinitrophenyl)octylamine derivatized Maltose ........................................................................... 81 3.2 UV absorbance of N,N-(2,4-dinitropheny1)octylamine derivatized maltodextnn ................ 82 4.1 Schematic molecular representation of the E. coli envelope ...91 4.2 Schematic structure of a Salmonella lipopolysaccharide ....... 92 4.3 Structure of trisaccharide from Rhizobium trifolii ANU 843 .33 4.4 Structure of tetrasaccharide from Rhizobium trifolii ANU 843 .................................................................................. 93 xvii 4.5 4.6 4.7 4.8 4.9 4.10 4.11 4.12 4.13 4.14 4.15 4.16 4.17 4.18 4.19 4.20 Proton NMR spectra and structure of trisaccharide from Rhizobium trifolii ANU s43 ................................................ 97 Proton NMR spectrum and structure of trisaccharide from Rhizobium trifolii ANU 843 ................................................ as Proton NMR spectrum of an unusual component in Rhizobium meliloti 41 ........................................................................ Q Deuterium-labeled, carboxyl-reduced component ............... 101 GC trace of 3-deoxy-erythro-heptodiulosomc acid derivatives 102 EI mass spectrum of 3-deoxy-erythro-2,4-heptodiulosonic acid derivatives .............................................................. 102 The structure of 3-deoxy-threo-2,4-heptodiulosonic acid ....... 103 EI mass spectrum of 3-deoxy-erythro-2,4-heptodiulosonic acid deuterated derivatives .............................................. 103 EI mass spectrum of 3-deoxy-heptitol derivatives ................ 104 CI mass spectrum of peracetylated 3-deoxy-erythro-2,4- heptodiulosonic acid ....................................................... 105 CI mass spectrum of 3-deoxy-erythro-2,4-heptodiulosonic aced deuterated derivatives .............................................. 106 Deuterated and undeuterated CI mass fragmentation of 3- deoxy-heptitol derivatives ................................................. 107 The first synthetic pathway for 3-deoxy-heptitol derivatives.. 109 EI mass. spectrum from synthetic 3-deoxy-heptitol derivatives ..................................................................... 111 CI mass spectrum from synthetic 3-deoxy-heptitol derivatives ..................................................................... 111 The second pathway to synthesize 3-deoxy-heptitol derivatives ..................................................................... 112 xviii 4.21 6.1 6.2 6.3 6.4a 6.4b 6.5 6.68 6.6b 6.7 6.8 6.9a 6% 6.10 7.1 The pathway to synthesize 3—Deoxy-D-Erythro—2,4- Heptodiulosonic acid........- ............................................... 114 Gel filtration on Sepharose 4B of the crude LPS isolated from Rhizobium leguminosarum biova viciae VF-39 .................. 135 Gas chromatogram of O-antigen carbohydrate components of Rhizobium Leguminosarum biova viciae VF-39 as the alditol aceates derivatives .......................................................... 1% Proton NMR spectrum of O-antigen .................................. 137 The structure of tetrasaccharide .......... 137 Proton NMR spectrum of the tetrasaccharide .................... 138 Gel filtration on Sepharose 4B Of the crude LPS of Rhizobium leguminosarun biovar viciae VF-39-32 .............................. 139 The structure of the trisacchride ...................................... 140 Proton NMR spectrum of unidentified component from strain VF-39-32 ........................................................................ 141 Gel filtration elution profile on Sepharose 4B of the crude LPS isolated Rhizobium leguminosarum biovar viciae VF-39-86 . 142 Proton NMR spectrum of the material in peak 4 of Figure 6.7 ..................................................................... 143 Proton NMR spectrum of the disaccharide from peak 4 of Figure 6.7 ...................................................................... 144 The structure of the disaccharide mannose and KDO ......... 145 The mass spectrum of the sodium adduct to the sodium carbonated disaccharide a-Man-(1-5)-KDO ......................... 145 Gel filtration on Sepharose 4B of the crude LPS isolated from Rhizobium leguminosarum biova viciae VF-39 .................. 151 xix 7.2 Gel filtration on Sepharose 4B of the crude LPS of Rhizobium leguminosarun biovar viciae VF-39-32 .............................. 152 7.3 Proton NMR spectrum of peak 2 in Figure 7.2 .................... 154 7.4 Proton NMR spectrum of peak 2 from Figure 7 .3 after irradiating atthe signal at 2.02 ppm ............................... 155 7.5 Proton NMR spectrum of the same compound in Figure 7.3 after irradiating the signal at 2.60 ppm ............................. 156 7.6 Partial expanded proton NMR spectra .............................. 158 7.7 Gas chromatogram of the alditol acetate of the unusual glycosyl components ....................................................... 13) 7.8 The CI mass spectrum of the first peak from the Figure 7.7. 161 7.9 The structure of the alditol acetate in the Figure 7.8 ............ 161 7.10 The CI mass spectrum Of the second peak from the first pair in Figure 7.7 ............................................................ 162 7.11 The structure of the alditol acetate for the Figure 7.10 ......... 162 7.12 The total ion chromatogram for the methylation analysis... 164 7.13 Mass spectrum of peak 1 and 2 from Figure 7.12 ................ 165 7.14 The structure and fragmentations for Figure 7. 13 .............. 165 7.15 Mass spectrum of peak 3 from Figure 7.12 ......................... 166 7.16 The structure and fragmentations for Figure 7.15 .............. 166 7.17 Envelope configuration of the furanose ring ....................... 167 7.18 Twist configuration of the furanose ring ........................... 167 XX ABEE C I Dansyl DASO Dimsyl DME ' DMSO DNPO EDC EI FAB-MS GC-MS HMPA HPLC KDO LC-MS LPS's MDO NMR n.O.e. SDS-PAGE TFA TIC LIST OF ABBREVIATIONS p—aminobenzoic ethyl ester chemical impact 5-N,N-dimethylnaphthalene sulfonyl N,N-dansyloctylamine dimethylsulfinyl dimethoxyethane dimethylsulfoxide N,N-(2,4-dinitrophenyl)octylamine 1-ethyl-3(3-dimethylamino-propyl)carbodiimide electron impact fast atom bombardment mass spectrometry gas chromatography-mass spectrometry hexamethylphosphoramide high-performance liquid chromatography 3-deoxy—D-manno-oct-2-ulosonic acid high pressure liquid chromatography interfaced with mass spectrometry lipopolysacchrides membrane-derived Oligosaccharides nuclear magnetic resonance nuclear Overhauser effect sodium dodecyl sulfate-polyacrylamide gel electrophoresis ‘ trifluoroacetic acid total ion chromatogram xxi TLC UV thin layer chromatography ultraviolet xxii PART I: DERIVATIVES FOR THE ISOLATION, PURIFICATION AND MASS SPECTROMETRIC CHARACTERIZATION OF HOMO AND HETERO OLIGOSACCHARIDES CHAPTER 1 BACKGROUND RTEEARCH 1.1 Introduction The carbohydrate-rich glycocomplexes on biosurfaces are among the most important current tOpics of research in biological function and structure. To understand the normal biological and physiological process generally involves the complete characterization of polysaccharides, glycolipids, and glycoproteins, and requires detailed information with regard to the carbohydrate moieties such as molecular weight, sequence, and linkage. The structural elucidation of polysaccharides and larger Oligosaccharides generally involves specific hydrolysis to small Oligosaccharides followed by the separation and identification of the fragments. Normally, the isolated carbohydrate is available only in very small amounts and has multiple functional groups. This makes the structural analysis of carbohydrates one of the most difficult areas in analytical chemistry. 1.2 ThClasdcalMethodsofSequencingCarbohydratae Valent et al. reported a “general and sensitive chemical method for sequencing the glycosyl residues of complex carbohydrates” in 1980 (2). The total procedure is described in Figure 1. Using this method, successful structural characterization of a complex carbohydrate requires a very complicated procedure. The carbohydrate to be characterized must be relatively pure. The purification of large Oligosaccharides is achieved by gel Purification of large Oligosaccharides by gel filtration chromatography Reduction of the glycosyluronic carboxyl groups by sodium borodeuteride 1 Permethylation and fractionation by reversed-phase HPLC 1 Partial hydrolysis to yield di-, tri-, and tetra-saccharides l Ethylation of the small Oligosaccharides to label the carbon atom to which the terminal glycosyl residue had been attached Hydrolysis 0f the small Fractionation of the partially Oligosaccharides, reduction meth 1 . . . . y ated and partially ethylated 0f the reducmg termim or small Oligosaccharides by HPLC the resulting monomer by . - . sodium borodeuteride and and Identificatmn by GC/MS after acetylation, identification by GC/MS example: 4G3 _,. Comparison of the fractions of small CDHO Ac Oligosaccharides and unhydrolyzed CHOMe Oligosaccharides, connection of the EtOCH fraction of the small Oligosaccharides CHOEt “89th“ CHOAc CHZOMe \ Composition analysis to determine the glycosyl linkages and sequences Figure 1: The classical method of sequencing Oligosaccharides. Purification of large Oligosaccharides by gel filtration chromatography Reduction of the glycosyluronic carboxyl groups by sodium borodeuteride l Permethylation and fractionation by reversed-phase HPLC 1 Partial hydrolysis to yield di-, tri-, and tetra-saccharides l Ethylation of the small Oligosaccharides to label the carbon atom to which the terminal glycosyl residue had been attached HydI‘OIYSiS 0f the small Fractionation of the partially Oligosaccharides, reduction meth 1 ° . , , y ated and partially ethylated 0f the reducmg termim 0f small Oligosaccharides by HPLC the resulting monomer by . . . sodium borodeuteride and and Identificatmn by GC/MS after acetylation, identification by GC/MS example: 1G3 -’ Comparison of the fractions of small CDHO Ac Oligosaccharides and unhydrolyzed Me Oligosaccharides, connection of the EtO CH fraction of the small Oligosaccharides CHOEt together CHOAc CHZOMe \ Composition analysis to determine the glycosyl linkages and sequences Figure l: The classical method of sequencing Oligosaccharides. filtration chromatography, which generally uses Biogel P-2 or an ion- exchange resin. The reduction is performed with sodium borodeuteride to identify carboxyl group of glucosyluroic residues present in the complex carbohydrate. The carboxyl group is reduced to the corresponding dideuterio-labeled primary alcohol, which then can be structurally characterized by mass spectrometry. After permethylation, the latter Oligosaccharides are fractionated by HPLC. The larger Oligosaccharides are partially hydrolyzed. The resulting disaccharides, trisaccharides and tetrasaccharides are ethylated in order to label the carbon atom to which the terminal glycosyl residue had been attached. The partially methylated and partially ethylated small oligosacchrides are then fractionated by chromatography on a reversed-phase HPLC column and identified by GC/MS. According to the overlap of the small Oligosaccharides with the unhydrolyzed complex carbohydrate, the oligosaccharide sequences may be pieced together. The fractions from the reversed phase HPLC column are hydrolyzed to monomers, sequentially reduced by sodium borodeuteride, and acetylated to the corresponding partially methylated, partially ethylated alditol acetate. After acetylation, the deuterated alditol acetates are uniquely labeled as the hydroxyl groups at C-1 and C-5 (or C-4 for furanosyl residues) and the positions of linkage are acetylated. The resulting monomers are again identified by GC/MS. The glycosyl-linkage composition of each oligomer is then used to determine the glycosyl sequence. The typical limitations of this classical method are that it is complicated, time consuming and often requires in excess of a year to complete the whole procedure. Another limitation is the quantity of material required. Valent et al. said that this method requires only 4 milligrams of material but the quantities of many isolated biological samples are far below milligram amounts and therefore, cannot be analyzed by this method. 1.3 Electron Ionization Mme Spectrometry The mechanism of electron ionization (EI) mass spectrometry is electron loss from the parent ion thus putting a charge on the molecule so its path can be directed. Sumcient energy is also put into the molecule to break some of its chemical bonds. When the analyte collides with the fast electron, the analyte loses an electron and becomes a particle with an odd number of electrons and a positive charge: M+e'=M+-+2e‘ Most EI mass spectra are recorded at an electron energy of 70 eV. Usually only a small fraction (less than 0.1%) of the analyte molecules are ionized in El (3), and the rest of the sample is pumped away into the vacuum system. Because the pressure is so low in the ion source, the ionized molecules generally do not collide with the walls and other molecules before their fragmentation. The EI mass spectrum usually includes signals from the molecular ion and fragments. The limitations of EI mass spectrometry is that generally EI requires that the sample be in the gas phase. Thus non-volatile, high-molecular- weight carbohydrate materials have remained beyond the scope of the E1 measurement. Because high-molecular-weight carbohydrates are not thermally stable and have to be degraded to the simplest species and chemically derivatized to provide thermally stable condition. The peralkylation methods add mass to the analyte, increasing the complexity of the analysis. The additional energy required for volatilization of the molecule causes unavoidable thermolytic processes, producing high- abundance low-mass fragments, and decreasing the intensity of the molecular ion. Sometime too many fragmentations make the E1 spectra difficult to interpret. However, if the energy of electron beam is lowered to decrease the extra fragments, the entire sensitivity of the spectrum is simultaneously decreased. To obtain complete structural information for a high-molecular-weight carbohydrate by El generally requireds 100-200 ug sample (4). These large sample requirements are impossible to meet in the case of scarce, isolated biological materials. 1.4 Chemical ImpactMass Spectrometry EI mass spectrometry works by colliding molecules with 70 eV electrons but also results in a lot of fragmentation in the process. Because of this fact, there often are not enough surviving molecular ions to indicate the molecular weight of the compound. This is one of the most important pieces of information. The chemical impact (CI) method is a less energetic, softer ionization mode. In the CI process, ionization of the analyte is a result of a gas-phase chemical reaction, not a collision by energetic electrons. Chemical impact uses methane or ammonia as reagent gas. At low pressure in a mass-spectrometer ion source, the reagent gas is bombarded with energetic electrons to produce the expected primary ions. CH4+e'=CH4+°+Ze' CH4“ = CH3+ + H- peralkylation methods add mass to the analyte, increasing the complexity of the analysis. The additional energy required for volatilization of the molecule causes unavoidable thermolytic processes, producing high- abundance low-mass fragments, and decreasing the intensity of the molecular ion. Sometime too many fragmentations make the E1 spectra difficult to interpret. However, if the energy of electron beam is lowered to decrease the extra fragments, the entire sensitivity of the spectrum is simultaneously decreased. To obtain complete structural information for a high-molecular-weight carbohydrate by El generally requireds 100-200 pg sample (4). These large sample requirements are impossible to meet in the case of scarce, isolated biological materials. 1.4 Chemical ImpactMaaa Spectrometry EI mass spectrometry works by colliding molecules with 70 eV electrons but also results in a lot of fragmentation in the process. Because of this fact, there often are not enough surviving molecular ions to indicate the molecular weight of the compound. This is one of the most important pieces of information. The chemical impact (CI) method is a less energetic, softer ionization mode. In the CI process, ionization of the analyte is a result of a gas-phase chemical reaction, not a collision by energetic electrons. Chemical impact uses methane or ammonia as reagent gas. At low pressure in a mass-spectrometer ion source, the reagent gas is bombarded with energetic electrons to produce the expected primary ions. CH4+e'=CH4+-+2e‘ CH4“ = CH3+ + H~ These primary ions will collide with neutral reagent gas, and produce a series of secondary ions: CH4+- + CH4 = CH5+ + CH3- CH3+ + CH4 = C2H5+ + H2 If a small amount of analyte gas is present, the collisions of the secondary ions and the analyte molecules produce gas-phase acid-base chemical reactions. CH5+ is a very strong gas-phase acid, and can readily give up a proton. Most organic molecules are proton acceptors: CH5++M=CH4+(M+H)+ In the CI process, the collision energy of the secondary ions and the analyte is lower than the 70 eV in the E1 process, so in CI spectra, a large proportion of the molecular ions remain and fewer fragment ions are formed than in EI spectra. Although CI is a soft ionization technique, it still requires that the analyte be in the gas phase. There are a lot of reports of CI studies for peralkylated carbohydrates, but there were still restrictions for the high- molecular-weight non-volatile carbohydrate to be analyzed. 1.5 FastAtomBombardmentMassSpecthetry (FAB-MS) The need for effective determination of the structure of large and complex biomolecules has led to the development of several other soft ionization methods, such as secondary ion, fast atom bombardment, electrospray, plasma desorption and laser desorption mass spectrometry. The fast atom bombardment mass spectrometry method was introduced in the last decade (5). 1.5.1 Operating Process In the FAB-MS process, non-volatile and thermally labile compounds can be analyzed. The samples are first dissolved in a solvent and then mixed with a polar, viscous, low-vapor-pressure liquid matrix, which is generally glycerol, or thioglycerol often doped with trifluoroacetic acid (for positive FAB) or triethanolamine (for negative FAB). In the FAB-MS, a neutral accelerated beam of inert xenon, argon, or other suitable gas atoms, with 6 to 10 keV of translational energy, is fired from an atom gun. The stream of neutral gas molecules is generated by ionizing, accelerating and then charge-neutralizing noble gas atoms. The noble gas atoms undergo a collision with a high-energy ion beam to form fast ions which are accelerated. The fast ions capture electrons from an electron-rich saddle- field to become fast atoms and move towards a small metal target which has previously been loaded with the liquid matrix containing the sample to be analyzed. When the atom beam collides with the sample and matrix, kinetic energy is transferred to the surface molecules, causing many molecules to be sputtered out of the liquid into the high vacuum region of the ion source. A large number of the molecules are ionized during the sputtering process. The ionization is achieved by interaction with matrix: addition of a proton or cation (positive ion mode) or loss of a proton (negative ion mode) giving pseudo-molecular ions as major signals in the spectrum. During ionization, some internal energy is imparted to the molecule which leads to fragmentation. 1.5.2 The Sample Preparation Generally, the sample is dissolved in a volatile solvent and 1-1.5 pl of solution is mixed into 1-2 ul of matrix which was previously placed on the target of the probe tip. The probe carrying the loaded target is put into the ion source through a vacuum lock system. The solvent evaporates from the probe in the ion source chamber prior to exposure to the source. 1.5.3 Matrix The viscous liquid matrix plays an important role in the diffusion of sample from the surface of the matrix. It also promotes a stable, reproducible ion current lasting for longer periods of time for the diffusion process (6). A closer view of this process is shown in Figure 1.2. The figure describes the surface of the matrix during bombardment by the neutral fast atom beam and how the desorbed species are produced. When the bombarding beam approaches a 90° angle of incidence, the majority of the energy deposition is limited to the surface layer of the solution, causing intramolecular and intermolecular vibrations deeper in the matrix. The viscous liquid matrix replaces the destroyed area and provides additional analyte to the matrix surface. With this process, using a glycerol matrix, the positive or negative ion signals can be constantly supplied for periods of 20 to 30 min to the detector. Because of the requirement for high viscosity, most of the matrices are naturally hydrophilic. This causes several notable problems. One is poor sensitivity, causing high detection limits. Other problems are intense background ions, intense cluster ions from the matrix, and ion suppression effects. These problems inhibit the formation of certain ions from the sample. Mass Detector I | Secondary Ion Bear;> © @’ ® C9 © % Fast Particle Gun Neutral Fast gm Beam ED 50 32 5° 1;:0 53’ Probe Surface of Glycerol and Sample Solution Figure 1.2 View of the matrix surface during negative FAB-MS process. (M-H)' is a negative charged ion, (G) is a matrix molecular such as glycerol, and (M)0 is a neutral molecule. The interaction between analyte and matrix depends on hydrophobicity, and hydrOphilicity. To improve the quality and intensity of a spectrum, different matrices or matrix additives can be used. For example, in this research, a mixture of glycerol and thioglycerol (1:1) was employed. If glycerol alone is used, it gives strong cluster ions at multiples of 92 + H+ (in positive mode spectra), starting at 93 amu and continuing into the region of m/z 1000. Addition of thioglycerol suppresses cluster ions up to m/z 500. Addition of 0.1 M trifluoroacetic acid (TFA) solution directly to the matrix helps to displace the cations and is employed in positive mode 10 spectra. The addition of triethanolamine to the matrix helps to displace anions. Sensitivity can also be improved by modifying the analyte. Hydrophobic groups have less of an attraction for polar matrix molecules. When analyte is modified with a hydrophobic group, which reduce hydrogen-bonding interactions with matrix, the analyte tends to occupy the surface layer of the polar sample solution. The FAB-MS signal recorded is largely determined by the interactions between hydrOphilicity and hydrophobicity of the matrices and analytes. Some requirements for the liquid matrix are (7, 8): 1. The solubility of sample in the matrix must be very good. 2. The vapor pressure should be as low as possible. 3. The matrix must be chemically inert, and not reactive with the sample. 4. To ensure the diffusion process, the viscosity of the matrix must be reasonably low. 1.6 Methods of Carbolmdrate Analysis Methods for determination of the structures of carbohydrates are quite laborious, insensitive, and time-consuming compared to established sequencing methods for the two other major biopolymers: proteins and nucleic acids. Gas chromatography has been used to separate small oligosaccharides and derivatized oligosaccharide fragments. The problem is that the large oligosaccharides never elute, or at best, do so at a very slow rate because of their high polarity and high molecular weight. In addition, the analyzed material can not be recovered. 11 Gel filtration can be used to isolate both large and small oligosaccharides, and the material can be recovered quantitatively. However, several days are needed to complete the analysis, and detailed structural information still can not be obtained. Although information can be obtained by 1H NMR spectroscopy in some instances using pulse schemes which reveal connectivities involving the anomeric proton or carbon and nuclei across the interglycosidic linkage, ‘it requires a major commitment of personnel and facility resources as well as large amounts of pure materials. These methods also require a fairly high degree of sophistication in experimental set-up and interpretation and are far from routine. The high-performance liquid chromatography (HPLC) technique has the capability of separating, non-destructively, oligosaccharides with high yields. The lack of a fluorophore makes carbohydrate very poorly detectable by UV, which restricts using a conventional UV detector, is a limiting feature for analysis of carbohydrates by HPLC. Mass spectrometry has contributed significantly to carbohydrate structure elucidation and holds much promise for sequencing carbohydrates because the quantity of material required is typically orders of magnitude lower than that required for NMR spectroscopy. In principle, the technique can be applied directly to mixtures if the instrumentation exists for precursor and product ion analysis as is the case with triple quadrupole and four-sector instruments. One of the most popular of current ionization/desorption methods for analyzing non-volatile analytes by mass spectrometry is fast atom bombardment (FAB-MS) (9). FAB-MS data can be obtained from native carbohydrates as long as they are free of salts and other contaminants (10). However, it is very difficult to obtain 12 analytes of such purity. Trace amounts of salts will bear the greater proportion of the total ion current, and surface active components will occupy the matrix surface, displacing the sample and preventing desorption. Currently, several nanomoles of underivatized carbohydrates are required for analysis by FAB-MS (11). The negative ion FAB-MS of y -cyclodextrin was reported in 1981 (12). The spectrum contained a peak for a molecular ion (M-H = 1295), and one for another ion due to loss of one sugar unit. The preliminary data indicate that FAB-MS will be useful in obtaining the molecular weight of complex oligosaccharides. 1.7 ClassicalMethOdsforCarbohydrateDerivatives Strategies for solving the problems in carbohydrate analysis usually involve the classical derivatization methods, such as permethylation and peracetylation. Systematic and reliable procedures have been developed in the early years for these transformations which also allow sample vaporization without thermal decomposition. The permethylated polysaccharides are first degraded by acid-catalyzed hydrolysis (13, 14). The linkage information for the resulting monosaccharides is based on the mass-spectral identification of the reduced acetylated compounds (15, 16). These procedures have been widely used, but mostly for obtaining EI and CI spectra usually by GC-MS (17). The disadvantage of this method is that permethylation and peracetylation are sometimes incomplete for large oligosaccharides. Another drawback is that the molecular weight increases significantly after derivatization, resulting in mass window limitations. Often methods such as the radio labeling of carbohydrates at their reducing termini by reduction of the oligosaccharides with NaB3H4 (18) 13 have also been used. The problems have been also addressed by the use of hydrophobic chromophores and charge localization groups attached to the carbohydrate molecule. This method leads to increased UV sensitivity and better fragmentation in mass spectrometric analyses, as recently reviewed (19). The general strategy for making such derivatives involves the covalent attachment of the chromophore group to the reducing end of oligosacchride by reductive amination. 1.8 Dmivatizationby Chromophore Groups The sensitivity of FAB analysis is strongly influenced by the solubility of the analyte in the matrix. In general for carbohydrates, as the hydrophilicity of the aglycone decreases, solubility and sensitivity also decrease. The poor surface activity of carbohydrates in polar matrices is another one of the features that limits the utility of ion sputtering/ionization techniques such as FAB-MS in the study of carbohydrates. Using reductive amination, the preparation of several derivatives of reducing oligosaccharides has been reported. In 1978, Matsushim et al. (20) reductively aminated oligosaccharides with 2-aminopyridine by means of reduction with sodium cyanoborohydride giving a fluorescent 2- aminopyridine derivative. They used this derivative to determine degree of polymerization by chromatographic properties, sequence of the sugar units, and the linkage points of the sugar units. For the linkage information, they reported the permethylated derivative. The molecule they characterized was a disaccharide. In 1983, Prahash et al. (21) coupled 7-amino-4- methylcoumarin to the aldehyde group of oligosaccharides. One picomole of the derivative can be detected after thin-layer chromatography by 14 illumination with a UV light. The mass spectrometry technique was not utilized. One of the most popular and successful groups for sequencing carbohydrates is the aminobenzoic acid alkyl ester group (22-24). This method has led to enhancement of the ion intensity observed for carbohydrate analytes during FAB-MS, and has allowed the sequence for the carbohydrate chains of several glycoproteins to be determined. In 1984, Sweeley et al. (22) described a method for structural characterization of oligosaccharides by preparing the UV-absorbing oligoglycoaminoalditols from reductive amination with p-aminobenzoic acid ethyl ester to form the so-called ABEE derivatives. They used this method for HPLC and FAB-MS analysis. The mass range they studied had an upper limit of 900 amu which corresponded to derivatized maltotetraose. In 1988, Burlingame et al. (23) applied the ABEE method to the study of the structure of oligosaccharides released fi'om glycoproteins. They reported that the ABEE derivatives provided a deprotonation center enhancing the mass spectra. Positive and negative ion spectra contained moleCular species of similar abundance, but the fragment ion peaks giving the sequence information were significantly more prominent in the negative ion mass spectra. They did not observed cleavages suggesting or confirming branched oligosaccharides in the pattern of fragmentation. Later, Burlingame et al. (24-27), using the heptasaccharide maltoheptaose as a model oligosaccharide, recognized that lengthening the alkyl chain from methyl to n-tetradecyl resulted in a concomitant increase in the molecular ion abundance. They reported that very high mass spectral sensitivities were achieved with n-tetradecyl and n-decyl p-aminobenzoates. However, the yields of derivative obtained were significantly lower than those obtained for 15 n-octyl, n-hexyl, n-butyl, ethyl, and methyl p-aminobenzoates. n-Octyl and n-hexyl p-aminobenzoates were found to be optimal considering both yields of derivatives and mass spectral sensitivity. Using maltoheptaose as a test compound, the ABEE derivative analyte could be detected at levels on the order of 10 pmole with a signal to noise ratio of 8:1. Higher sensitivity could be realized with longer alkyl chain derivatives, but the use of such derivatives was compromised by solubility problems when very long chains were employed. The use of an octyl chain (ABOE derivative) gave significantly higher sensitivity (about 20:1) in comparison to underivatized carbohydrate without severe solubility problems. Limitations of the aminobenzoic acid alkyl ester method are that the derivatizing reagent must be prepared beforehand, the actual derivatization yield is not quantitative for longer chain alkyl esters, and significant sample workup must be performed prior to analysis. Another modification to the ABEE method was reported by Reinhold et al. (28). They attached the aminobenzoic acid pentafluorobenzyl ester group to the mducing end of glycan residues. This imparted fluorescent and UV properties for chromatographic detection, and also allowed the molecule to behave as an electron trap under negative ion chemical ionization. They employed this method using a liquid chromatography interface with mass spectrometry (LC-MS). The advantage of using this derivative is that the sensitivity is very high. Detection levels of 800 fmol were realized with signal-to-noise of 100:1. Even with a 16 fmol injection, a signal-to-noise ration of 3:1 was still observed. The disadvantage of this technique is that only a strong molecular signal can be observed; no sequence or other structural information could be obtained. 16 Recently a new method was developed by Dr. Rawle I. Hollingsworth for the sequencing of peptides by FAB-MS. In this method, the sensitivity of detection and amount of sequence-specific fragmentation of these molecules was increased by attaching a positively charged 2-(triphenylphosphonio) ethyl group to the N - terminus of the molecule (29). The'strategy we adopt here is one that modifies the carbohydrate with a N,N-(2,4- dinitro)phenyloctylamine (DNPO) group for negative mode and N,N- Dansyloctylamine (DASO) derivatives for positive mode detection. These groups serve as the site for charge localization. The dinitrophenyl group was used since it can readily form a negatively charged species. The negative charge is readily delocalized by the two nitro groups on the phenyl ring. In a previous study, it was determined that dinitrophenyl derivatives of secondary amines have molar absorption maxima which were very insensitive to the nature of the alkyl group, the only difference being in the rate of derivatization (30). This makes this derivative ideal for the direct quantitation of molar proportions of components in a mixture by direct integration of peak areas. Also, the positive charge on the DASO derivative (formed by protonation 'of the demethylamino group) is readily delocalized and stabilized. Both the dinitrophenyl and dansyl groups are also excellent chromophores for UV detectiOn and can be attached in a facile manner. The octyl group increases the surface activity of the carbohydrate analytes which have poor surface activity in the hydrophilic matrix. The resulting DNPO and DASO derivatives are therefore designed to overcome problems of low UV visibility, low ionization efficiency, and low surface activity typically observed with carbohydrate compounds during analysis by HPLC and FAB-MS. A large amount of fragmentation with charge retention on the reducing end is also observed. 17 1.9 Nomenclature The fragmentation patterns of the carbohydrate investigated by FAB- MS are generally more complicated than those of other bimolecules, such as peptides and nucleic acids. A systematic nomenclature for assigning and describing the different types of fragmentations of the carbohydrate moieties has been introduced by Domon and Costello (31). 1.9.1 NommclatumofCarbondrateFmgments There are two major modes of cleavage normally observed in FAB- MS of carbohydrates. The simplest one is the fragmentation which results from the cleavage of the glycosidic bond giving information about the oligosaccharide sequence. The other mode of fragmentation is more complex and involves the sugar ring cleavage. Sequences from the latter mode are more difficult to assign, but they contain very important structural information. In Figure 1.3, when the charge is retained on the non-reducing end, Ai, Bi, and Ci labels are used to designate the fragments, where the i indicates the number of the glycosidic bond cleaved counting from the non- reducing end. In the case of fragmentation containing the aglycone or reducing unit, Xj, Y5, and Z5 labels are used to designate the fragments. Here, the subscript j indicates the number of the glycosidic bond cleaved counting from the reducing end or the aglycone unit. In Figure 1.3, only A and X sequences indicate the ring cleavage. The glycosidic bond linked to the aglycone is numbered 0. The fragments formed from the reducing end Y1 21 Yo I, h r’ CHpH Ix : 3-0 l I : l : OHI : I ' l : : —o-:-n I ' I l 5 OH : I I l l I I I I" I 25 / 0,5 As A1 B1 C1 “A2 32 02 33 Figure 1.3. Nomenclature of carbohydrate fragments. with retention of the glycosidic oxygen is the Y sequence. The fragments formed from the non-reducing end with retention of the glycosidic oxygen is the C sequence. The fragmentation without retention of the glycosidic oxygen on the species formed, from the reducing end is the Z sequence. The fragmentation without retention of the glycosidic oxygen, from the non- reducing end, is the B sequence. CHZOH 5 CHZOH O_R 4 o O—R 4 O OH ° OH 3 1 OH 2 2 0“ OH Figure 1.4 The numbering of the ring bonds. The superscripts represent the cleavage involved in the carbohydrate rings. Two superscripts k and 1 (Le. kJAi, kJXj) are used to represent the 19 sugar ring bonds, the ring bonds are numbered according to Figure 1.4. 1.9.2. Branched Oligosaccharide Moieties In the case of the branched oligosaccharide (Figure 1.5), the main body is the longest moiety, the “core”, and the branches are referred to as the “antennae.” The labeling of the core portion is the same as described in section 1.9.1. The Greek letters a, B, 7 are used to represent each branch which are designated in the order of decreasing molecular weight. The fragmentation of the core portion is described without Greek letters. Their Vs / /Ym OH,OH CHZOH O 0 Vote OH OH / Y2 Y1 Y0 \ OH OH 0‘ YG'S Ya‘4 / o OH OH CH20H CHZOH OH,OH / o .0 -OR o o 0 q OH OH OH OH OH OH / / _o -O B 35 13,5 OH OH O \ 4 / a“.2 / Ya‘3 Figure 1.5 The nomenclature for branched oligosaccharide. 2) use begins from the point of connection of the antennae to the core. In the case where there are two branches cleaved on a same ring, the Greek letters representing each cleaved branch can be used together, such as all. The Greek letter a’ and a” are used, the a’, on” are designated in the order of decreasing molecular weight, in the case of sub-branch observed. 1. 10 SequencefromOligosaccharide The fragment ions generated from the ion source reflect the structure of the oligosaccharide. The general sequence information was summarized by Anne Dell (32) in 4 type of cleavage: those are depicted in Figure 1.6. In series A, the charge is retained on the non-reducing end and occurs in both positive and negative modes. In series B, the charge is retained on the reducing end, and the fragmentation yields an unsaturated hexose. This kind of fragmentation occurs in the positive mode only. In series C, the charge is retained on the reducing end, and the ions are 28 amu heavier than those formed in pathway A. This sequence can be observed in both positive and negative modes. Series D gives fragmentation with charge retained on the non-reducing end. Fragments are 42 amu heavier than those from pathway A. This sequence is most commonly observed in negative mode spectra and is rarely observed in the positive mode (32). Understanding the mechanism of fragment formation is very important in identification of the sequence from the derivatized oligosaccharides. In cases using different derivatives, the sequence from internal glucosidic bond cleavage with charge retained on the non-reducing and always gives the same ions. In sequence with charge retained on the reducing end, ions varies depending on the specific type Of derivatives. In 21 very complicated FAB-MS of oligosaccharides, this character can help to identify the sequence as coming from the reducing end or non-reducing end. The reactions may proceed by radical mechanisms (33), or by concerted mechanisms (34). , OH H + o O Senes B 28 CHZOH +H OH OH H OH + O O ——’ OH \ OR H \,H OH OH H + H0 H OH +H 07 +7" + HCHaoH CHZOH -H - 0’1on or 0 O CH -l-l - 0H OR Senes C HO—< + 2:" O OH °”2°” 9 OH OR O H + OH 3 0"on OH,OH - (2“on + 0 O Senes D 0 0' CH20H 0HO OH OR__, OH H‘H‘ + OH 0 CH H H H OH + H OH Figure 1.6 Oligosaccharide fragmentation pathways. 1.11 Refermce 1. 10. 11. 12. Hollingsworth, R. I., and Dazzo, F. B.,Carbohydr. Res., (1988) 112 97- 112. Valent, B. S., Darvill, A. G., McNeil, M., Robertsen, B. K., and Albershem, P., Carbohydr. Res. (1980) 12 165-192. Suelter, C. H., and Watson, J. T., “Biomedical applications of mass spectrometry” (1989) 5, An Interscience Publication, John Wiley & Sons, New York. Reinhold, N. V., and Carr, S. A., Mass Spectrom. Rev. (1983) 2 153- 221. Barber, M., Borodoli, R. S., Sedgwick, R. D., and Tyler, A. N., J. Chem. Soc. Comm. (1981) .7. 325. Barber, M., Bordoli, R. S., Sedgwick, R. D., and Tyler, A. N., Nature (1981) 223 270-275. Gower, J. L. , Biomed. Mass Spectrom. (1985) 12 191. DePauw, E., Mass Spectrom. Rev. (1986) 5, 191. Beth, L. Gillece-Castro and Burlingame, A. L. “Methods in Enzymology” (1990) 123,, 689-712, (Abeison, J. N., and Simon, M. I., Eds,) Academic Press, San Diego. Eggs, H., and Peter-Katalinic, J. Mass Spectrom. Rev. (1987) 5, 331- 393. Kamerling, J. P., Heerma, W., Vliegenthart, J .F.G., Green, B. N., Lewis, I. A. S., Strecker, and Spik, G. Biomed. Mass Spectrom., (1983) 10, 420-425. Fenselau, C., Cotter, R., Hansen, G., Chen, T., Heller, D., J. Chromatog. (1981) 215 21-30. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23 Lindberg, B., Lonngren, J ., Svensson, 8., Adv. Carbhydr. Chem. Biochem. (1975) 3.1, 185-240. Hakamori, S.-I., J. Biochem. (Tokyo) (1964) 55, 205-208. Lindberg, B., Lonngren, J ., “Methods in Enzymology”, (1978) 50 1 Ginsburg, V., Ed., Academic Press, New York. Biermann, C. J ., “Analysis of Carbohydrates by GLC and MS,” (1989) 27-41, Biermann, C. J ., McGinnis, G. D., Eds; CRC Press, Boca Raton, FL. Carpita, N. C., Shea, E. M., “Analysis of Carbohydrates by GLC and MS,” (1989) 157-216, Biermann, C. J ., McGinnis, G. D., Eds; CRC Press, Boca Raton, FL. Turco, S. J. Anal. Biochem. (1981) 118. 278-283. Poulter, L., and Burlingame, A. L. “Methods in Enzymology” (1990) 193, 661-689, Abeison, J. N., and Simon, M. I., Eds, Academic Press, San Diego. Hase, S., Ikenaka, T., and Matsushim, Y., Biochem. Biophys. Res. Comm. (1978) 85 257-263. Prahash, C., and Vijay, I. K. Anal. Biochem. (1983) .128 41-46. Wang, W. T, LeDonne, N. C. JR., Ackerman, B., and Sweeley, C.C. Anal. Biochem., (1984) .141, 366-381. Webb, J. W., Jiang, K., Gillece-Castro, B. L., Tarentino, A. L., Plummer, T. H., Byrd, J. C., Fisher, S. J ., and Burlingame, A. L. Anal. Biochem., (1988) 169 337-349. Poulter, L., Kerre, R., and Burlingame, A. L. Anal. Biochem. (1991) 125 1-13. 26. 27. 29. 30. 31. 32. 24 Hernandez, L. M., Ballou, L., Alvarado, E., Gillece-Castro, B. L., Burlingame, A. L., and Ballou, C. E. J. Biol. Chem. (1989) 264 11849- 11856. Poulter, L., Earnest, J. P., Stroud, R. M., and Burlingame, A. L. Proc. Natl. Acad. Sci. USA (1989) 86 6645-6649. Poulter, L. J ., Earnest, J. P., Strond, R. M., and Burlingame, A. L. Biomed. Environ. Mass Spectrom., (1988) 15, 25-30. Caesar, J. P. J r., Sheeley D. M., and Reinhold V. N. Anal. Biochem. (1990) 121 247-252. Wagner, D. S., Salari, A., Gage, D. A., Leykam, J ., Fetter, J ., Hollingsworth, R. I., and Watson, J. T., Biol. Mass Spectrom., (1991) 29 419-425. McIntire, F. C., Clements, L. M. and Sproull, M. Anal. Chem. (1953) 25 1757-1758. Domon, B., and Costello, C. J. Glycocojugate (1988) 5 397-409. Dell, A., Adv. Carbohydr. Chem. Biochem., (1987) 4.5, 19. Dell, A., and Oates, J. E. T., “Analysis of Carbohydrate by GLC and MS” Biermann, C. J ., McGinnis, G. D., Eds, CRC Press, Boca Raton, FL, (1988) 217-235. Hendrickson, J. B. Angew. Chem. Int. Ed. Engl. (197013,, 47-76. CHAPTER2 2.1. GeneralMethods 2.1.1. Chemicals Glycerol, maltohexaose, Fetuin, N,N’,N”-triacetylchitotriose, and triethanolamine were obtained from Sigma Chemical Co. (St. Louis, MO). Thioglycerol, sodium cyanoborohydride, octylamine, cesium iodide, and all HPLC grade solvents were obtained from Aldrich Chemical Co. (Milwaukee, WI). Ultramark 443 and 6121 were obtained from PCR Inc. (Gainsville, FL). The 2,4-dinitrofluorobenzene was obtained from Eastman Organic Chemicals (Rochester, NY). Glacial acetic acid was obtained from Mallinckrodt Specialty Chemicals Co. (Paris, KY). Methanol was obtained from Baker (Phillipsburg, NJ). Triethylamine was obtained from Fisher Scientific, Inc. (Fair Lawn, NJ). Maltodextrin (degree of polymerization = 1-16) was obtained from Staley Manufacturing, and the Rhizobium capsular oligosaccharide was isolated in another study (1). Potato starch was obtained from Dr. Mirta Sivak (MSU, East Lansing, MI). 2.1.2. TheGeneralProcedureforMaking Derivatives 2.1.2.1. Preparation of N,N- (2,4-dinitropherrvl)octylamine Derivatives (DNPO) MM Maltohexaose ( ~ 0.7 mg) was added to a 1 dram vial, ~3 equivalents octylamine and 4 equivalents sodium cyanoborohydride, then 0.2 mL 5% acetic acid in a solution of water/methanol (3:1) were added. The Teflon-lined vial was capped and heated at 70° C for 1 hour. Subsequently, the solution was blown to dryness. Approximately 3 equivalents triethylamine, 2 equivalents 2,4-dinitrofluorobenzene, and 0.2 mL % methanol/water (4: 1) were added consecutively, and the mixture heated at 45° C overnight. Samples were analyzed directly without purification. W1; Maltohexaose (~ 0.7 mg) was placed to a Teflon-lined screw-cap vial, 3 equivalents octylamine, and 4 equivalents of sodium cyanoborohydride. Acetic acid (5%) 0.2 mL in a solution of water/methanol (3:1) was then added and the vial capped and heated at 70° C for 1 hour, and blown to dryness. Afterward, 0.1 mL of 0.2 M sodium bicarbonate and 2 equivalents of 2,4-dinitrofluorobenzene in 0.5 mL methanol were added (2). The mixture was heated for 20 minutes at 60° C, and subsequently 0.2 mL of 0.2 M sodium hydroxide in dioxane containing trace amounts of water to dissolve the base was added. The mixture was then heated for 1 hour at 60° C. After cooling to room temperature, 0.5 mL water and 0.5 mL ethyl acetate were added. The contents of the aqueous layer were adsorbed on a 018 reversed-phase cartridge column. The column was eluted with 1 mL water, 1 mL water/methanol (1:1), and 1 mL methanol. All three fractions of elutant were analyzed by FAB-MS. Most of the carbohydrate was contained in the water fraction. 2.1.2.2. Preparation ofN,N-Dansyloctylamine Derivatives Maltohexaose ( ~ 0.5 mg) was placed in a 1 dram vial, ~3 equivalents octylamine and 4 equivalents sodium cyanoborohydride, then 0.2 mL 5% acetic acid in a solution of water/methanol (3:1) wrer added; the Teflon- lined vial was capped and heated at 70° C for 1 hour. Subsequently, the solution was reduced to dryness with a stream of nitrogen. Approximately 3 equivalents triethylamine, 2 equivalents dansyl chloride (5-N,N-dimethyl naphthalene sulfuric chloride), and 0.2 mL methanol/water (4:1) were 27 added consecutively, and the mixture heated at 45° C overnight. Samples were analyzed directly without purification. 2.1.3. Instrumentation 2.1.3.1. High Performance Liquid Chromatography Reverse-phase HPLC analysis was accomplished using a Waters (Milford, MA) 600 multisolvent delivery system equipped with a Kratos (Ramsey, NJ) Spectroflow 783 programmable absorbance detector set at 392 nm (for DNPO). The separation of derivatized oligosaccharides was achieved in approximately 100 minutes with a Beckman (San Ramon, CA) Ultrasphere ODS column (4.6 mm x 25 cm) with a 5 um particle size. Elution was performed at a flow rate of 1.0 mL/min at 50° C using an isocratic gradient, with a mobile phase consisting of water and acetonitrile (7:3). 2.1.3.2. Mass Spectrometry FAB-MS was performed on a JEOL HX110 (Peabody, MA) double- focusing mass spectrometer (EB configuration) equipped with a high-field magnet operated in the negative and positive ion mode. Ions were produced by bombardment with a beam of Xe atoms (6 kV). The accelerating voltage was 10 kV, and the resolution was set at 1000 or 3000 according to the mass range of interest. The samples were dissolved in a 1:1 (v/v) H20-MeOH solution; generally 1-1.5 uL sample was mixed with 1 HL of glycerol/thioglycerol (1:1) or triethanolamine on the FAB-MS stainless steel probe tip. The solvent was removed by pumping before analysis. Calibration was performed using Ultramark (443 and 6121) or (CsI)nI" cluster ions, depending on the mass range of interest. A JEOL DA-5000 % data system recorded the spectra. The sample was scanned at a scan slope of 2 min from m/z 0-3000. Data presented were acquired in a single scan and found to be reproducible. % 2.2. NewAppmadiestoSequencingOligosaccharidesbyNegativeMass Spectrometry 2.2. 1. N, N-(2A-dinihophenylhctylamine Derivatives for Negative Mass Spectrometric Characterization of Maltohexaose There are two methods for preparing the DNPO derivatives of oligosaccharides. The process described in Method II for introducing the dinitrophenyl group is a classical one (2). In this method, however, the sodium salts complicate the mass spectra with many salt adduct peaks. In addition, the salt can totally depress the peaks in the mass spectra. This was also the case when too large an excess of sodium cyanoborohydride was employed. Instead of sodium hydroxide and sodium bicarbonate, we used triethylamine (TEA) as the base (Figure 2.1). When an adequate amount of this base was used, very clean sequence peaks were obtained with the elimination of all salt adducts. Stoichiometric control of the chemical reaction made the purification of the sample before analysis unnecessary, thus avoiding loss of material. The experiment was performed on 0.12 mg (~1.9 x 10-7 mole) of maltohexaose; the product was detected at a level of 3 pmole. In order to test for completeness of derivatization, 1H NMR spectra of the derivatized products were obtained. Completeness of derivatization was indicated by the complete disappearance of the anomeric proton signals of the reducing residue of the oligosaccharides. In the case of maltose, these signals appeared at 85.3 (J = 3 Hz) and 84.8 (J = 7 Hz) for a and B anomers respectively. Similar results were obtained with maltohexaose. In the case of maltodextrins complete disappearance of the most downfield protons was also observed. The sensitivity and dynamic range of the NMR spectrum could easily accommodate quantitation of 3) anomeric resonances due to the low abundance, high molecular weight oligomers. 011,011 0 CH20H CH20H o CHZOH OH \HO— F—H N H2(CH2)7CH3 0 H NaCNBH, OH H°——H n H——0H 5%Ac Hm OH OH H— CH30H/1120 3:1 n -0H HO—_H HO-H—H HO C cmchnznca, for maltose n=1, for maltohexaose n=5 F EtaN CH30H/Hao 02N 5:1 N02 I CH“?! cazoa OH HO — -—H OH O —H OH n H — —OH HO — —H CH2N(CH2)10H3 NO, Figure 2.1: The procedure for making DNPO derivatives. The negative ion mode FAB mass spectrum of the N,N-(2,4- dinitro)phenyloctylamine (DNPO) derivative of maltohexaose (molecular weight 1269, Fig. 2.2) showed a major ion at m/z 1268 corresponding to the (M-H)’ ion which arises from the loss of a proton from the parent ion. This 31 M-H Fl ( )‘1aee e l . a , 4 1+ ”As i 1411 t V 1 ° 1 l 2"A4 c . 545 U Y 2.4 n 1 ‘ y A9 458 2 1.4 d 1.5x, 1. ‘A‘ 1:20 707 v, at). 8 49° 15x 702 Y4 y 1251 2 s 2 us 1.) 1 l a 311 i ll l i:i| is"; ll ll! “11‘1th Nil 11, W ill '1. ill“: M“ ii H _ 4, -_ ._ - . ; .1 1.-.__-._.-.;-11_-. . .. ..-_ - . " ' soo soo 700 soo 900 1000 1100 1200 Wicca l Figure 2.2: Negative mass spectrum of N,N-(2,4-dinitrophenyl)octylamine (DNPO)-derivatized maltohexaose. mass spectrum displayed four different types of sequence ions. The mechanisms of formation for these ions are shown in Figure 2.3 a-d. Some of the sequences produced a fragmentation pattern involving two-bond ring cleavages within the cyclic sugar units (3, 4). The predominance of two- bond ring fragmentation over glycosidic bond cleavage following ionization from an infrared laser desorption was observed by Cotter et al. as well (5, 6). The nomenclature for the sequence fragments is based largely on that introduced by Domon and Costello (7). The Yn sequence corresponds to the loss of unhydrohexose residues by an elimination process involving loss of the hydrogen atom at C-2 and cleavage between the glycosidic oxygen and the anomeric carbon of the leaving residue at each glycosidic bond. Each loss corresponds to 162 mass units (Figure 2.3 a). The second sequence, 214A“, originates from cleavage at the non-reducing end by a concerted six 32 centered scission, which can be explained as pericyclic fragmentation of the penultimate ring (8) to give an initial loss of 222 followed by subsequent losses of 162 mass units, the charge being retained on the fi'agment Y Y3 Y2 Y 1 CHZOH '15 CHZOH ,/ CHZOH I CHZOH ,1 CHZOH I : O E O f o : O E CHZOH OH : OH : OH : OH E OH : ”°_—H OH I OH : OH I OH ' OH : I I ' H——OH 1106 944 782 620 459 HO——H CH2N(CH2)7CH3 NO2 CHZOH CHZOH - H ' CHZOH CHZOHI SEOUENOE Yn OH/ Figure 2.3a: The Yn sequence from DNPO-derivatized maltohexaose. containing the non-reducing end (Figure 2.3 b). The sequence 15X“ was in relatively low abundance. Formation of this series of fragments is facilitated by the proton at the 4 position undergoing a 1,3 hydride transfer accompanied by scission of the C1-C2bond and the 05-0 bond of the terminal residue (Figure 2.3c). The peaks representing ions of this series were also separated by 162 mass units, with the charge retained on the fragment containing the reducing end, the ions appeared at a mass-to- charge ratio 28 amu greater than Yn sequence ions. The sequence ions, 14A“, were minor and could be rationalized by formation of an anion at the 33 C-2 hydroxyl group accompanied by a 3, 5 hydrogen transfer (Figure 2.3 11). Again, members of this sequence were characterized by a difference of 162 mass units. In the maltohexaose spectrum, the Yn sequence was the most abundant in the high mass end, but the 25A,, and 114An sequences predominated at the low mass end. CHon CHzOH CHZOH (3sz OH OH 2 OH ..__. QQOl—KLLE L3,; OH IOH Ail—H }869 “As 54: 707 "2.1:: '- A. “A. 0" (CH2),CH, b 02N N02 CH,OH 011,011 _H _ CHon O o SEQUENCE “A“ CH2? ”Ac OH OH " OH + O o O _) o K ‘0 O" o” o OH 5 OH Ho 11 CH OH 'H ' O 20 H’kCHon O OH J” + o H 0 .. _ 1mi- Figure 2.3b: The sequence 2:4An of DNPO-derivatized maltohexaose. CH20H GHzOH (3"on éHo : :\ 0:233— 0:“.‘L .0 CH20H 0" OH OH‘, OH‘ ""“H OH: OH: ———-OH \ l 972 810 ‘1648 GTE: Ho_ —H "Xi ”"3 “"2 5"! 0H 1011970113 ab N02 CH,OH CH,OH SEQUENCE u‘xn CHZOH CHZOH <1CH,OH HS 0 ”:0: Figure 2.30: The sequence 1:5Xn of DNPO-derivatized maltohexaose. CHon 0+1on “3“on : CHzOH ‘1 o 0 RH) :. 1—0 5.2"“. C.,... OH OH ‘OFL‘ ‘5“ a “a, AHO-EHEE-H.-- O O ‘\ 0 ~“ 0 ‘§ OH \‘ ‘ s‘ o—n—H OH OH OH s x . i OH : 0": H——OH 97 ”A6 41.1 573 73.5 HO—-—-H o-r ..... - CH (CH2)7CH3 02" N02 H CH20H CHZOH (31.1on _ . 0 O H o - H SEQUENCE 1"A“ O H H*O(CH2)20H OH OH + OH \OH + O O o CH,OH H O- OH + O CH CHO O CHZOH O '“ H20’u\CHO 0" I O 0 OH Figure 2.3d: The sequence 14A,, of DNPO-derivatized maltohexaose. {E 2.2.2. N, N-(2,4-dinitrophenyl)octylamine Derivatives for the Isolation, Purification and Mass Spectrometric Characterization of OligosaccharideMixtureMaltodexhin A mixture of starch-derived maltodextrin of a degree of polymerization equal to 1-16 was derivatized using essentially the same method as that used for maltohexaose. The experiment was started at 0.14 mg (~7.9 x 10'8 mole). The ease of preparation of this derivative and its visibility by UV detection in HPLC makes it an excellent candidate for routine HPLC isolation of trace amounts of carbohydrate materials. The mixture was analyzed by reverse-phase chromatography using an acetonitrile/water (3:7) solvent system. This allowed separation of the mixture of derivatized oligomers into 16 well-resolved peaks with a degree of polymerization in descending order from 16 to 1 with increasing retention time. The assignments were verified by co-injecting the mixture with derivatized maltohexaose which co-eluted with the eleventh peak from the point of injection. The peak of degree of polymerization 16 was determined to represent less than 0.1% of the total mixture based on peak integration. The mixture was analyzed directly by negative ion FAB-MS (Figure 2.4 a-b). The sensitivity was as low as 0.38 nmole. The most prominent sequence ions observed between 400 and 900 amu correspond to the 214A“ series of ions. For any given component of the mixture, Yn series of ions were isobaric with (M-H)' molecular ions of lower homologues which were also quite prominent. The composite molecular ion Yn series predominated between 900 and 3000 amu, and a (M-H)' ion was observed for the highest degree of polymerization component containing 16 glucose units (m/z 2888). This component represents less than 0.1% of the total mixture, 1 5 l it all i‘ Y ‘I’.’f}.ll";‘! 191‘8 Y., 1 ll. '4 _ 207s ‘ ‘w ‘ ”Ara 2W3 l9 Y'z .0..- $.41 - . .- 14A‘42240 2165 “‘12 1841 (a)_ S $1.64 a 1 I '1 tam“ “A: l 707 v 4 ”50*. y ‘ C3. 2AA6 u 40- «5503 v, ass .2820, cs. 11, c, 3 ”A4 557 827 940 c” “A, C., 1105 c7- Y. 1313 C. a .1 1264V1288 c.- ;‘7fi n 1,4 1426 ca): 1"!" A“ 1305 \ 1436 e ‘1 :M' 145 600 800 1000 1200 1400M,z )- C. Y R L wad Y. 2402:3 l9 L 1592 a . t 6 c i - 10- V , cm 1750 a PZ‘A10 1637 A 4.1517 1529 Y, C". ‘ H ‘ b . 24A 1754 1912 . fill}; ,i u '1 2 «mm n . d a n c e 1600 1800 2000 2200 2400 . 2600 2800 M/Z Figure 2.4 a-b: Negative mass spectra of DNPO-derivatized maltodextrin. 38 which means that it could be detected at levels as low as 0.1 picomole. The limits of sensitivity are far superior to those of the amino benzoic acid ethyl or octyl ester method (9). In Figure 2.4, a cluster of ions was observed at positions corresponding to the isobaric Yn series from the (M-H) ion and the (M-H) ion of lower homologues. These clusters were attributed to the fact that a significant proportion of M' ions are formed by electron capture of parent species. These ions can then break down to give another species two mass units lower by loss of hydrogen, probably in the order of H‘ followed by Ht In fact, with some analytes, using this method we have observed only M species and no (M-H)‘ species. This is not surprising, given the high electron capture cross section of the dinitrophenyl group. A sequence of ions 37 amu greater than the Yn sequence was observed and is due to a cyanoborohydride adduct from the parent species of the Yn sequence. This was not observed in samples purified by ion- exchange chromatography. In the low mass region of the spectrum, a Cu series of ions was quite evident. This series involves cleavage of the glycosidic bond, with retention of the glycosidic oxygen atom by the species formed from the non-reducing end (Figure 2.5). The Cm, series is probably due to an adduct of the analyte with matrix compound triethanolamine followed by the loss of two H20 molecules from the 0,. series. These kinds of adducts are observed very often in large molecules. The spectrum was reproducible, and a similar series was also found during analysis of other starch oligomers. These will be discussed later. H——OH HO——-H 011531 ; cuon : 0112011 : 5 ° : ° : CHzNszwHa OH : OH : OH : -: 04 Of—‘ 02N OH I OH / on I 1313 1475 1637 "02 Ca Cs 019 l , , """" ‘1151 0112011 0112011 : CH2°H : 0'1on : CH20H : 01-1on 07 o u 0 1 O ' O ' o I I I I 5 OH 5 OH 5 OH 5 OH o-; 0; o: 0; 0 OH, OH OH OH 0503 665 327 939 03 C4 05 Co 01-12011 011on SEQUENCE Cn 0*.ng 'H ' OH 0 OH OH Figure 2.5: Sequence Cn of DNPO-derivatized maltodextrin. 40 2.2.3 N,N- (2,4-dinitro)phenyloctylamine Derivative for the Negative Mass Spectrometric Characterization of an Oligosaccharide from a Bacterial Cell Surface As a further test of the efficiency of this derivatization method for enhancing signal intensity, 1 nmol of a previously characterized (1) oligosaccharide from Rhizobium trifolii 843 capsular polysaccharide (Figure 2.6) was derivatized and analyzed. 0112011 01100 CH2 0112<())110H0_|_H HacXO CR H3C 5:}:f": CH (CH2)7CH3 HOOCO R = H3CCHOHCH2CO- no, Figure 2.6: Structure of previously characterized and derivatized oligosaccharide from Rhizobium trifolii 843 capsular polysaccharide. The derivatization was performed on 290 ug (~ 1.7 x 10‘7 mole) of oligosaccharide, and 3 mole of derivatized material was placed on the probe. The derivatization was expected to result in loss of the acetyl and 3- hydroxybutyryl groups because of the basic conditions employed. A sequence of ions at m/z 1742, 1764, 1786, 1808, 1830, and 1852 corresponding to (M- H)’. (M+Na-2H)', (M+2Na-3H)', (M+3Na-4H)', (M+4Na-5H)' and (M+5Na- 41 6H)‘, respectively, was observed (Figure 2.7). Another series of ions 152 amu lower corresponded to the loss of a carbohydrate group by a pericyclic mechanism, most likely the one illustrated in Figure 2.8, and yielding pyruvate, formaldehyde, and an unsaturated glycosyl residue with a shift of 152 amu in the sequence of ions . é°° e 1 I 1 1656 380‘ woe : ' 1034 1670 V J ' ‘7” 17641706 1830 860‘ .' ‘ 17‘2 ‘ I 1852 A 1 ' N ‘ ' b . U40. ' “5° "'00 1730 who 10150 firm n . wz d . a 1 11 c2“. 6 1 C - - --._. . 500 1000 " 15'00' ' ' ' ' ' éooo M/Z Figure 2.7: Negative mass spectrum of previously characterized, and derivatized oligosaccharide from Rhizobium trifolii 843 capsular polysaccharide. HOOC o} ( o .. o H C ;\ OR —-»cn3ccoon + 0150 + {m 3 on on Figure 2.8: Mechanism for rearrangement of acetal bond to produce I pyruvate, formaldehyde and an unsaturated glycosyl residue. 42 2.2.4. Sequence Specific Cleavage of N,N-(2,4-dinitrophenyl)octyl- amine Derivatives of Hetero and Homooligosaccharides by Negative Ion Fast Atom Bombardment Mass Spectrometry N,N-(2,4-dinitrophenyl)octylamine (DNPO) derivatives of N,N’,N”- triacetylchitotriose, lacto-N-fucopentaose, lacto-N-difucohexaose, and isomaltohexaose were analyzed by negative ion FAB-MS. The lacto-N- fucopentaose and lacto-N-difucohexaose were isolated from human milk. Those compounds were used to test the efficiency of the DNPO derivative in providing sequence information on heterooligosaccharides by (-) FAB-MS analysis. Derivatization increased both the ion intensity and efficiency of useful fragmentation. Sensitivities observed from the negative FAB-MS of these naturally occurring oligosaccharides were often better than 1 pmole. Sequence specific cleavage, which is useful both for distinguishing isomers and the analysis of glycosidic linkages, was observed for all of the oligomers. 5 R I 0 . | . a 4- 1 1 l a.“ 3 v 1393 e 3‘ 3 (M41 3 . .2... u 2‘ 04 n - ' As (1 ‘ 77 a Z n 1 0.4A6 c 1 1 11 1.1.;1 11 1 f Y‘ vs a 1.; 1 1111111 31 1111511111111111111 ii” i: 31 . 9“ 1 06 1 H w "‘1 114211-1111, 1.1;» 3‘ 1! ‘J 400 600 800 1000 1200 ”,2 Figure 2.9: Negative FAB-MS of DNPO-derivatized isomaltohexaose 43 2.2.4.1 Negative FAB-MS of Isomaltohexaose The derivatization was performed on 41 ug isomaltohexaose (~ 4.1 x 10‘8 mole), and 0.1 nmol of material was put on the probe for negative FAB- MS analysis. The negative FAB-MS of isomaltohexaose is shown in Figure 2.9. A sharp molecular ion appears at m/z 1268. The lower intensity peak at m/z 1305 is from the cyanoborohydride adduct to the molecule. Comparing the spectrum of isomaltohexaose to Figure 2.2 of maltohexoase, in Yn-type fragmentation, both L4 and 1-6 linked sugars can lead to the same mass losses (-162), regardless of the anomeric configuration (a or B) and, therefore, Yn-type fragmentation does not appear to be a distinguishing sequence between 1-4 and 1-6 linked isomers. The mechanism for fragmentation of isomaltohexaose is shown in Figure 2.10, which can be compared to that of maltohexsaose in Figure 2.3 a. The two isomers undergo Yn-type cleavage to yield the same sequence. The 04A,, sequences in isomaltohexaose (compare Figure 2.3 b) appear at same mass shift with the 23A,, sequence in maltohexaose. Because of the 1-4 ring linkages in the maltohexaose and the 1-6 linked sugar in isomaltohexaose, the Yn-type sequences appearing at the same mass shift can be cleaved from different bond order in different ring linkages (Figure 2.11). Generally, fragmentation and ionization are involved in the two bonds ring cleavage within the cyclic sugar (34-37). The sequences MA“ and 15X“ appearing in the FAB-MS of maltohexaose did not appear in the FAB-MS of isomaltohexaose. The differences in two bonds fragmentation observed for these two isomeric structure, maltohexaose and isomaltohexaose, indicate that this method is potential by quite useful detecting subtle structural differences between oligosaccharides. Y5 Y4 Y3 Y2 Y1 ’1 ’l ’1 1 I I I *1 . I . 0+1on 5 CH 0 5 CH 0 : l E = 2 2 ' CH ' - 0 E 0 E 0 E 2% E CHZO : cuzo I I I OH : OH : 0H 5 0H 5 0H 5 H0-——-H I ._ .— I I I ' HO H 0“ 0H : o” OH : OH : OH : O : . - ' ' 0” H——OH 1106 944 782 520 453 HO———H CHng CH2N(CH2)7CH3 OH H owQ —<-O c220 CHZOH SEQUENCEYn O ”03% ' —’ OH 0 OH Figure 2.10 Yn-type fragmentation of derivatized isomaltohexaose. 45 HO 1 '—'H CH2N(CH2)7CH3 02N N02 CHZOO CH ' H ' o, 4 2°“ SEQUENCE An ocnzc Hero} OH + 11 o ’u‘cuzon "fa—)— Figure 2.11: Cleavage pattern for generation of 0s4An sequence in isomaltohexaose. 46 2.2.4.2. Negative FAB-MS of DNPO Derivative of N,N’,N”- triacetylchitotriose FAB-MS sequence anions were also observed for the DNPO derivative of N ,N’,N”-triacetylchitotfiose, a glucosamine-containing trisaccharide. The structure of the derivative is shown in Figure 2.12. Figure 2.12: The structure of the DNPO-derivative of N,N",N”- triacetylchitotriose. The derivatization was performed on 69 pg (~ 1 x 10'7 mole), and 1 nmol of material was put on the probe. The negative FAB mass spectrum for the DNPO derivative of N ,N’,N”’-tziacetylchitotriose is shown in Figure 2.13. The molecular weight of derivatized N,N’,N”-triacetylchitotriose is 906, and the highest molecular weight species in the spectrum has the same mass. This molecular ion was resulted from electron capture by the dinitrophenyl group. This same phenomenon has been observed for other extremely electron- deficient centers. This appears more favorable than proton loss for this molecule. In Figure 2.13, the Yn sequence is observed, yielding the linkage 47 information for this homotriose. Each Yn sequence peak is separated by 203 amu, an anhydro-N-acetyl glucosamine residue. The Yn sequence ions appeared at m/z 499 and 724, and 906, while the peak at m/z 724 possibly is a sodium adduct to Y2 of m/z 702. R ¢ 9 1 I ‘ uxrl‘bo'“ a w 509 t . i + V 4 e '2“: A ‘ .- Y, b ‘ 499 “1:12:20 M" g 524 m6 d a n c a Mom 500 600 700 800 QOOMIZ Figure 2.13: Negative FAB-MS of the DNPO-derivative of N ,N’,N”’- triacetylchitotriose. Beside the Yn sequence ions, there are other peaks in the mass spectrum. The 2:4A3 ion is at m/z 465, the ion at m/z 524 arose from m/z 465 plus a sodium and a cyanoborohydride. All of ions of X sequence appeared with loss of a water. The 1:5X1 ion with loss of a water is at m/z 509. The 0:4X1 ion with loss of a water is at m /z 624, and 0:4X1 plus a cyanoborohydride appears at m / z 661. The cleavage pattern of 0.4x, sequence is shown in Figure 2.14. The 0:4Xn cleavage proceeds by a 1, 3 hydrogen rearrangement. The cleavage pattern for anion adduct formation by cyanoborohydride is shown in Figure 2.15. 48 0“ o CHZOH CHZOH -€-o } V HO—r-H HO——H HO N Ac 0 H 0-CH2-9-?H-CH2—0—_H H H—r—OH O NHAC H OH ——-> AcHN——H AcHN——H \/\/\/N—CH2 N—CH2 02" 02~ no2 no2 Figure 2.14: The cleavage pattern of 0.4Xn sequence. CHZOH o-CH, ”0““ 'HCNB(O___H O-C - - - _. __ _< H2 9 9HCH2 o H HO-CHz-C-CH CH2—0——H 'H2 H_—OH AcHN“_H + NaCNBH3 AcHN——H /\/\/\/N—CH2 -NaH N CH AM '— 2 02” o2~ N02 +37 amu N02 Figure 2.15: The cyanoborohydride adduct formation. The appearance of peaks corresponding to protonated molecular species is very common in positive FAB-MS. Some investigators have also reported the occurrence of negative salt adducts. Ganguly et al. (10) first reported the anion-adduct formation, the adduct was formed between chloride ions and a glycoside. Prome et. al. (11) also observed abundant chloride- adduct ions mass spectra of glycosides. They observed an ion corresponding 49 to (M -H20 + Cl)‘ in the negative mass spectrum of a trisaccharide. The ion at m/z 661 in Figure 2.13 is of similar origin (0:4X1 - H20 + CNBH)‘. 2.2.4.3. The Negative FAB-MS of Lacto-N-Fucopentaose The structure of its DNPO derivative of Lacto-N-fucopentaose, a heterooligosaccharide from human milk, is shown in Figure 2.16, and its negative FAB-MS is shown in Figure 2.16. Although (as is usually the use for analytes such as these) this spectrum was not obtained from a four-sector tandem mass spectrometer (12), nor by MS/MS link scans, but very good sequence information was obtained from the normal single scan masspectrum. The derivatization was performed Fuc a1—2GalBl-3GlcNAc Bl-3GalB1-4Glc 1.5x1 486 ( x {458 Y1 I 620 Y2 ‘t‘ ‘t‘ CH20H :‘ HO OFNQ ‘ z 323 Y3 0H % “"°__H OH Ovol“ $3)“ 0”" H7_°H g, *190 NHAc ‘ "CF—H HO \ ‘ OH x ‘ " 1'5X3 \A/VV N—CHZ .0 HO ; OzN‘Q HO -, I ' N02 0'13 L"'>985 Y, Figure 2.16: The cleavage pattern of the DNPO derivative of lacto-N- fucopentaose. 50 M 2 j M... | + 1132 a . l l v e A b I; M-H-o d 31H . 111 MONO-H a I! ' Y+NI-H 1154 " t‘ t‘ Hr I ”231“.” 4' ° . W 1‘". ii! 1” O Ilullmhll’ {.H H In 1 . Iii} IL! 1W 1"! I" MI ”I" (1%; $1,} [In ‘1 1‘ ‘1, .1 ‘ _ _ q q ‘ q i ‘ n 1 . . n p p 500 700 800 900 1000 1100 W2200 Figure 2.17: Negative FAM-MS of derivatized lacto-N-fucopentaose. on 33 1.1g (~ 3.9 x 10‘8 mole) of material, and ~ 0.3 nmol of derivative was put on the probe. As in Figure 2.17, the predominant ion in the high mass end of the (-) FAB-MASS spectrum of lacto-N-fucopentaose occurred at m/z 1132. This corresponds to the molecular weight of the compound and indicates that one electron is captured by the molecule to give an (M + a) species. The Yn sequence gave detailed glycosyl sequence information: ion Y4 at m/z 985 corresponds to a loss of a fucosyl residue (147 amu) from the non-reducing terminus, Y4 (loss of a water molecule) appeared at m/z 967, and the peak at m/z 1007 is sodium adduct to Y4, which was rapidly fragmentation from the sodium adduct at the parent ion of m/z 1154. Y3 at m/z 823 is due to the sequential loss of a fucosyl residue (147 amu) and a galactosyl residue (162 amu). Y2 at m/z 620 is due to the further loss of an N-acetylglucosaminyl group (203 amu), and the at m/z 642 is sodium adduct of Y2. Y1 at m/ z 458 is due to the further loss of a galactosyl residue. The peak at m /z 851 is due 51 to 1:5X3 specie. The peak at m/z 572 is due to a sodium adduct to the specie at m/z 550. The peak at m/z 508 is due to a sodium adduct to 1:5X3 ion at m/z 486. The results from FAB-MS analysis of DNPO-derivatized oligosaccharides are quite different from those reported in recent investigations. Cotter et. al. (13), based on their study of permethylated and peracetylated derivatives, concluded that glycosyl components which did not have a reducing end did not undergo two-bond ring cleavages and that chemical derivatization generally suppressed fragmentation. Their conclusion precisely pointed out the disadvantages of permethylation and peracetylation methods. Also, they indicated that only 8 1-4 and 1-6 linked saccharides showed additional peaks due to the loss of water. None of the a- 1-4 and 1-6-linked saccharides showed peaks due to the loss of water. This is not universally true. In our experiments, large molecules showed a greater tendency to lose water. For example, in Figure 2. 17, the additional peaks due to loss of water appeared in the mass range higher than m /z 900. 2.2.4.4 The Negative FAB-MS of Lacto-N-Difucohexaose Lacto-N-difucohexaose is a branched heterooligosaccharide from human milk. The derivatization of lacto-N-difucohexaose was carried out on 41 ug (~ 4.1 x 10'8 mole) of starting material, and 0.1 nmol of final product was put on the probe. The structure of lacto-N-difucohexaose is shown in Figure 2.18 and its (-) FAB-MS in Figure 2.19. In the mass spectrum, the highest ion at m/z 1278 corresponds to the molecular weight of derivatized lacto-N-difucohexaose. This indicates that the 2,4-dinitro phenyl group is again participating in an electron capture reaction. In Figure 2.19, the main sequence ions correspond to the Yn series and gives complete sequence 52 Fuc a1-2GalB 1-3GlcNAc Bl-3GalBl-4Glc | Fucal Yum I458 Y1 HO ’0 ,822 {620 Y2 “2°” HO ' 1 HO 0“ 0 ‘~ Hofi‘“ '. .‘ ’.-‘\‘ 0H H_—OH O I‘ l s‘ ,o ‘ \;0 flax? 2%. .1... :‘. I g ‘ OH : 2’4A3 : \AM/N—LHZ .0 i 2 HO ' *969 : OzN HO ' ' Y 3 : CH3 :0 ‘1’: NO2 : 03A3 "*1131 Y0,4 Figure 2.18: The fragmentation pattern of derivatized lacto-N-difucohexaose R YwNa-H11§3 M.- I9 3‘ 1278 a 1 1 i 1 V ° 2 45 A 1189 3 “Wm v Fuc+Na M’ n ‘Y AA: 845 107 d .4 148.514 0.2Aa Y2 Y a 2 - a3 5 20 2.4 - o . 2 69"2 9 O J . 1 , 1 soo 600 700 800 900 1000 1100 1200 M[2300 Figure 2.19: Negative FAB-MS of derivatized lacto-N-difucohexaose. 53 information. The Ya4 ion at m/z 1131 is due to the loss of an anhydrofucosyl group from the non-reducing terminus. The ion at m /z 1153 is sodium adduct to Ya4. The peak at m/z 1007 arises from the loss of the second fucosyl group. The peak at m /z 969 is thought to arise from the Ya4 ion after loss of an anhydrogalactosyl residue. The ion at m/z 845 is due to the cleavage including both on and [3 branches so-called YaB3 ion. This is due to sequential loss of two fucosyl residues and one anhydrogalactosyl residue from the parent molecule in a metallated (sodium) form. The ion at m /z 620, Y2 ion, arises from the Yam fragment by loss of an anhydroglucosaminyl residue (203 amu). The Y1 ion is due to the reducing terminus glucose linked to the DNPO group (m /z 458). Other peaks such as m/z 480 and 495, are the sodium and cyanoborohydride adducts to Y1, respectively. The peak at m/z 1242 is due to the loss of two water moleccules from the parent ion, and the peak at m/z 1189 is the cyanoborohydride adduct to the species at m/z 1153. Some peaks in the lower mass range belong to An sequences. Several sequences are characterized by sodium adduct ions which probably arise from the molecular species after successive electron and Na+ capture. This species then gives to negative by proton loss (M‘ + Na+ - H+). It has been noted that the carbohydrate moiety in glycoproteins is essential for their function (14). The naturally occurring oligosaccharides in these molecules are often branched. In this study, the sequence determination of branched carbohydrate derivatives analyzed by (-) FAB-MS demonstrates that this method is an excellent technique for providing useful structural information. 54 2.2.5. SequendngofDNPODefivafizedStamhbyNegafive FARMS 2.2.5.1 Introduction Starch, the storage carbohydrate of the majority of higher plants, occurs as water-insoluble granules that vary in size and shape depending on the species and maturity of the plant. Within these granules, there is normally a mixture of two polysaccharides: amylose and amylopectin. The amylose, amounting to 20-30% of total starch, is largely composed of long linear chains of (1-4)-linked a—D-glucopyranose residues. The major starch component is amy10pectin, also a macromolecule containing thousands of shorter branched chains of (1-4) and (1-6) linked a—D-glucose residues. The structural analysis of the isolated amylose and amylopectin components has been carried out by standard methods based on methylation, periodate oxidation, and partial acid hydrolysis studies (15). The chain length can be studied by the enzymic hydrolysis of the (1-6)-oc-D- glucosidic interchain linkages, followed by fractionation of the resulting mixture of linear chains by gel filtration (16). The degree of polymerization is defined as the number of glucose residues per reducing end group, and is normally calculated from physical measurements of the molecular weight. 2.2.5.2 Structure of Amylase The methods of enzymic degradation are widely used in the structural analysis of amylase. The oz-amylase hydrolyses nonterminal (1- 4)-a-D-glucosidic linkages in both amylose and amylopectin, giving a mixture of products, while the B-amylase catalyzes the hydrolysis of alternate linkages, giving maltose as the only low-molecular—weight product. None of the above enzymes have any action on (1-6)—a-D-g1ucosidic interchain linkages; only isoamylase can selectively hydrolyze the 1-6 55 linkages with no effect on the 1-4 linkages. The amylose has the linear (1- 4)-a-D-glucan; most of amylose fraction has secondary chains attached through occasional (1-6)-a branch points and in some species has a few phosphate groups at the C-6 of glucose residues (17). 2.2.5.3. Structure ofAmylopectin The general structure of amylopectin has been known for many years, but the details are still uncertain. Three different diagrammatic representations of the molecule are shown in Figure 2.20 (18). The A chains are linked to the molecule only by the potential reducing group, the B chains are similarly linked and carry one or more A chains, and the C chain ends in the only reducing group R in the molecule. In structure (e), only half of the B chains carry A chains, while the other half of the B chains have their nonreducing end group inside the molecule. A cluster model was further refined by Manners and Matheson (19) in Figure 2.21. In this figure, the majority of the B chains carry more than one A chain, to account for the change in the A:B chain ratio of the partly debranched material. The model in Figure 2.21 is compatible with the physical properties and with various enzymatic degradation studies which show that the branch points are arranged in “tiers” or clusters and are not distributed randomly throughout the macromolecule. The most recent report by Callaghan and Lelievre (20) concludes that the the amylopectin molecules are a large flattened disk shape, consisting of (1-4)-a-D-glucan chains joined by many (1-6)-a-branch points. Takeda et al. (21) reported that the phosphate in D-glucose 6-phosphate residues was analyzed by D- glucose 6-phosphate dehydrogenase. The detailed molecular structure of amylopectin seems to have emerged, but the arrangement of the 56 amylopectin molecules within the starch granule still remains to be determined. (a) Figure 2.20: The molecular structure of amylopectin proposed (18) by (a) Haworth. (b) Staudinger, (c) Meyer, (d) Meyer, (e) revised Meyer structure. R = reducing end. Ali-=1"! B A B vl’ ABA Illa—:11 BCA—T‘lR A___4 AA—IlA——__—_—l_l Figure 2.21: The cluster model of amylopectin. 2.2.5.4. Phosphorus in Starch Fractions Commercially available starches are derived from many botanical sources and contain minor amounts of non-carbohydrate constituents. The content of non-carbohydrate elements associated with starch, in particular phosphorus and nitrogen, show considerable variation. Different phosphorus content in starches is associated with different properties (22). Perhaps this is why considerable research has been conducted to determine the phosphorus content in starches and in their various derivatives. In their classical studies, Kerr and Severson (23) fractionated non- defatted corn and potato starches. They found 0.013% and 0.088% phosphorus in the more butanol-soluble fraction of corn and potato starches, respectively, and 0.009% and 0.014% phosphorus in the less butanol-soluble fractions, respectively, as compared to 0.020% and 0.081% in the respective parent starches. Posternak (24) detected 3% phosphorus in a hexasaccharide isolated from potato starch after hydrolysis by pancreatic oz- amylase. This hexasaccharide yielded D-glucose 6-phosphate upon partial acid hydrolysis, indicating the position of the phosphate. Parrish and Whelan (25) isolated phosphomaltotetraose from potato starch after 58 digestion by salivary oz—amylase. Periodate oxidation showed that the phosphate group is located on the D-glucose unit that is in the third position from the reducing end. Posternak (26) reported that the ratio of phosphate group to D-glucose units was 1:6, which resulted from the calculation that the phosphorus protects a fragment of about 6-7 glucose units against degradation by B-amylase. They also reported that the phosphohexaose is resistant to the action of B-amylase. Hizukuri et al. (27) reported that the phosphorus in starch was estimated to be glucose-6-phosphate, and 60-70% of total phosphorus in potato starch was found to be due to the glucose-6P residue. The rest of the phosphorus was incorporated as glucose-Z-phosphate and/or glucose-3- phosphate. The methods they used were periodate oxidation, borohydride reduction, acid hydrolysis and treatment with alkaline phosphatase. Hizukuri and Tabata (28) isolated three sugar phosphates, 4-0-(oz-D-6- phosphoglucosyl).D-glucose, glucose 3-phosphate, and glucose 6-phosphate, from the a-amylase limit phosphodextrin of potato starch by a partial acid hydrolysis and an ion-exchange resin. From the quantitative yields of glucose 3-phosphate and glucose 6-phosphate, the distribution of esterified phosphate in potato starch was estimated to be 38% on the C-3, and 61% on the C—6 of glucose residues. Their kinetic studies indicated that approximately 1% of the phosphate in starch existed in a form more labile to acid than glucose 3-phosphate, possibly as glucose 2-phosphate. Residues of glucose 6-phosphate are present in some starch fractions, serving as a barrier to B—amylase (29). 59 2.2.5.5. Negative FAB—MS of DNPO Derivatized Oligosaccharides Digested fmmPotatoStarch Analysis of the structure of starch is essential for understanding of phosphorus-containing starch. Many processes and treatments have previously been employed on starches, for instance enzymatic digestion, acid hydrolysis, elemental analysis etc. (30). Applications of techniques such as NMR and FAB-MS to starch analysis have not appeared widely. We have analyzed the N,N-(2,4-dinitrophenyl)octylamine derivative of oligosaccharides resulting from 0.1 M TFA hydrolysis of potato starch. By negative FAB-MS analyses, it is possible to see the heterogeneity introduced by the phosphate and the differences between the unsubstituted and the phosphorus-linked carbohydrate chain among the starch-digested oligosaccharides. The (-) FAB-MS spectra are shown in Figure 2.22 a-b. In Figure 2.22, two main fragmentation patterns, Cu and 2:4An, are observed. Sequence 0,, arises from the glucosidic bond cleavage between two sugar units leaving the oxygen atom on the reducing end and the charge on the non-reducing end. A sequence which is derived from the Cu fragmentation sequence by loss of a water molecule (18 amu lower than the Cu sequence) was accompanied with mass number label and without sequence labeling (due to the limited space in the spectrum). The Cm, sequence is probably due to an adduct of the analyte with the matrix triethanolamine followed by the loss of two water molecules from the 0,, sequence. This type of sequence was also observed in the previous study of maltodextrin with the same kind of derivatives. The 2:4An sequence is another prominent series, which is the most commonly observed series with charge retained on the nonreducing end. The proposed cleavage 8 545 adA‘ ,% 50 C4 665 2.1“ 707 8 OGDNQDCU> ¢<_""fl)—O 20.; 3. A. Figure 2.22 a: (-) FAB-MS of N ,N-(2,4-dinitrophenyl)octylamine derivative of oligosaccharides digested from potato starch, range 500-1550. 003 013:0) o<"~m—e:n ¢ 1 1 600 2200 2400 2600 2800 M [2000 Figure 2.22 b: (-) FAB-MS of N ,N-(2,4-dinitropheny1)octylamine derivative of oligosaccharides digested from potato starch, range 1550-3000. 61 pattern (shown in Figure 2.3 b) involves a glycosidic bond rearrangement by a concerted mechanism. One specific sequence in Figure 2.22 is the phosphorylated glucan series. The peak with mass shift at m/z 538 is the DNPO derivatized terminus glucose with a phosphorus group. The sequence labeled PM is a sodium phosphate saccharide series with relatively low abundance. The sequence can be followed up to eleven sugar units (m /z 2180). 2.2.5.6.Phosphorus-31NMRSpectraofPhosphorusContainingStamh To confirm the existence of the phosphate groups in the oligosaccharides, the acid-hydrolyzed starch was examined by phosphorus- I I I T I l T I I I I I I I I 1 I I I I I I I I I 1 I fi 10 0 -10 -20 -30 -40 PM“ Figure 2.23: 31P NMR spectrum of degraded phosphorus-containing starch (line broadening = 1 Hz) 62 31 NMR spectroscopy. The degraded starch was dissolved in D20, and hexamethylphosphoramide (HMPA) was used as reference. The phosphorus-31 NMR spectrum is shown in Figure 2.23. The line broadening was equal to 1 Hz. In the NMR spectrum, one peak at - 6.5 ppm is possibly due to phosphate group at C-6 position, the lower peaks at - 6.3 with splitting are possibly due to the phosphate groups at C-3 and C-2 positions. The decoupled 31P NMR spectra in pH 4 and pH 3 solutions showed no large variance from the one in neutral conditions. However in the non- decoupled spectrum (Figure 2.24), there are multiple splittings for the peak at - 6.3 ppm, which is possibly due to the overlap of doublet splitting to the ' U l I _ I I r 1— ° ‘5 -1o -15 -1 - PPm Figure 2.24: N on-decoupled phosphorus-31 NMR spectrum of degraded starch (pH = 4). 63 phosphate groups at C-2, and C-3 positions. After expanded the spectrum and setting the line brodening to 1 Hz (Figure 2.25). Material from digested starch still is a macromolecule, the rotation of phosphate groups is slow, the spectrum shows that the peak at -6.5 ppm is an unresolved multiplet probably because of differences in environments of the phosphate species to distributions rotamer populations. The result from N MR spectrum confirmed that starch contains phosphate groups, but the location of phosphate groups till needs further studies. PPm Figure 2.25: Expanded Non-decoupled phosphorus-31 NMR spectrum of degraded starch and setting line broadening to 1 Hz. 2.2.5.7. The Negative FAB-MS of Synthesized MaltohexaoseGPhosphate For further investigation of phosphorylated starches, we attempted to prepare the phosphate-containing oligosaccharides at very low levels of phosphorylation. Study of these synthesized phosphorylated goannsco‘) 0<-'-'m—o:n 64 oligosaccharides and comparison of them to the natural product will greatly enhance our understanding of the structure and properties of the phosphorus-containing starch. The N,N-(2,4-dintrophenyl)octylamine derivatized maltohexaose was treated with phosphorus oxychloride in dimethoxyethane (31). The negative FAB-MS obtained from 0.1 pmole sample is shown in Figure 2.26. The C-6 primary hydroxyl group is the most active position in the pyranose. The synthesized phosphorylated maltohexaose could be the “ ‘- 6 ,L r‘ ‘ The negative FAB-MS indicates the presence of both monophosphorylated and nonderivatized maltohexaose. The low level derivatization could possibly be controlled by the shortness of the reaction time or the hydrolysis of phosphorus oxychloride. The ion at m/z 1269 corresponded to the DNPO derivatized maltohexaose. The ion at m/z 1330 is Mu§P4~hO 445 M" 1330 2AA. 1269 s M" P "454-1110 va 13:9 555 1024 1 V w Ys Y1 who I 2 4 ° '0 - A 1 520 s 856 1106 573 Yam“, ' +a—H Yap_m+Na—H 344 878 2111111111 11111111. ‘.-;'1 111111111 11111191111111111111. I 1 600 800 1 000 1200 M I2 400 Figure 2.26: The negative FAB-MS of synthesized phosphorylated maltohexaose. $ due to the loss of a water molecule form monophosphorylated maltohexaose. The peak at m/z 1349 arises from the monophosphorylated maltohexaose. The peak at m/z 1361 is due to the unphosphorylated material plus glycerol, the matrix compound. In the case of DNPO derivatized maltohexaose, the dinitrophenyl group stabilizes the negative charge, so the negative FAB-MS shows that the ion of the unphosphorylated molecule had the same molecular weight, 1269. Generally, in negative FAB-MS, the mass unit of the adduct from glycerol, which arises by loss of a proton from the glycerol molecule, should be 91 amu higher than M'-. In Figure 2.27, the glycerol adduct peak at m/z 1361 is 92 amu higher than m /z 1269 due to the electron capture of DNPO group. The phosphorylated peaks, Yup ions with loss of a water molecule, appeared at m/z 520, 682, 844, and 1330, corresponding to the one, two, three and six sugar units. The phosphorylated Y4 and unphosphorylated Y5 ions appeamd at m / z 1024 and 1106, respectively. The reason for the incomplete sequence is that the derivatives underwent several steps of chemical reactions (reductive amination, followed by attachment of a chromophore group, and phosphorylation), and that any salt and hydrophobic reagent such as organic compounds will depress the amount of sample sputtered from the probe and reduce the density of the ions. The Yn sequence was relatively weak; only ions Y1 at m/z 458, Y2 at m/z 620, and Y5 at m/z 1106 can be observed, which is not surprising since we were unable to observe the Yn sequence in the underivatized digested starch oligosaccharides. The major sequence from the non-reducing end was the “An sequence of 2AA; at m / z 545, and 2:4A5 at m/z 707, similar to the 2:4An sequence observed in the negative FAB-MS of derivatized oligosaccharides digested from potato starch (Figure 2.21). However, in Figure 2.21 there is better sensitivity and % completion of sequence. The spectrum from synthetic phosphorylated maltohexaose was just a preliminary result. We also attempted to derivatize and phosphorylate the oligosacchrides digested from potato starch by the enzyme a-amylase without much success, even after prolonging the reaction to three days. One possible reason is that the starch was not very soluble in water, and so the reaction mixture was heterogeneous. Another explanation is that the enzyme we used was not active enough. In conclusion, the starch appears to be heterogeneous in nature and contains phosphorus associated compounds. The origin and structure of this phosphate is still unknown. In negative FAB-MS, sequences can be observed both with and without phosphophate linkages, indicating that the starches of some strains contain phosphophate groups, while others do not. This also explained the previous reports in which 3% of phosphorus was found in hexasaccharide, isolated from potato starch (24). The phosphorus- 31 NMR spectra are evidence indicating the satrch contained phosphorus. From the multiplicity of non-decoupling 31P NMR, the phosphorus possibly located at C-6, C-3, and C-2 positions. Because of the complex and delicate structure of starch granules and their derivatives, starch carbohydrate cannot be dissociated from the minor non-carbohydrate components that affect the properties of starches and products containing the starches. Realization of the important efi‘ect of non-starch components such as phosphorus on the properties of starch products will lead to a better understanding of starchy products and will aid in the development of better commercial methods for producing specific products. 67 2.2.6. The Negative FAB-MS of N,N-(2A-DinihoPhenylhctylamine Derivatizedoligosacchm'ideisolatedfmm Glycoprotein Fetuin Glchproteins are an important cell surface component. Determination of the structure of glycoproteins will have a considerable impact on our understanding of the biological function of the cell surface. In 1985, Naik et. al. (32) reported the determination of carbohydrate chains present in glycoproteins. Their method was based on the acetolysis of the intact glycocojugate, extraction of the peracetylated carbohydrate fragments, and analysis by FAB-MS. They reported the molecular ion of fully peracetylated species, and the Yn sequence (which was called A-type oxonium ions). The highest molecular ion they reported, at m/z 1958, was a peracetylated species, possibly from hexopyranose. For further study, the DNPO method was applied to structure elucidation of the carbohydrate moiety of the glycoprotein fetuin. Enzyme degradation was applied to the intact fetuin by Robert A. Cedergren of Dr. R. I. Hollingsworth’s lab at MSU. After enzyme degradation, the sample was derivatized by reductive amination and derivatization, followed by gel filtration to purify the derivatized glycan. The fractions collected from the gel filtration column were tested by the phenol sulfuric calorimetric method. The derivatized oligosaccharides were analyzed by negative FAB- MS. The resultant (-) FAB-MS is shown in Figure 2.27. In Figure 2.27, there are many peaks between m/z 500-1000, some of them possibly due to the unpurified peptide residue. Some higher mass peaks were observed from m / 2 1200-2800, which possibly arise from the derivatized oligosaccharides isolated from fetuin. Because of the limited amount of sample, further analyses were not performed. This preliminary result indicates that the DNPO method can be applied to detect higher molecular 68 weight oligosaccharides from a glycoprotein than the literature reported (32) peracetylation method in glycocojugate analysis. R109, 26 1.8 T 1199-9 2377.9 a t 801 i 1 z 1 as 9.7 ‘:60: b u ‘ n I d 4% 9 4 a 4 n J 1256.1 c I e . B 500 1000 1500 . 2000 2500 ‘ HI§000 Figure 2.27: The negative FAB-MS of N ,N-(2,4-dinitrophenyl)octylamine derivatized oligosaccharide isolated from fetuin. a) 2.2.7. N,N-Dansyloctylamine Derivatives for the Isolation, Purification, and Mom Spectrometric Characterization of Oligosaccharides 2.2.7.1. Preparation of N,N-Dansyloctylamine Derivatives of Oligosaccharies The procedure for preparing the N ,N-dansyloctylamine derivatives of oligosaccharides is similar to that of the N,N-(Z,4-dinitrophenyl)octylamine derivatives and shown in Figure 2.28. 0112011 0 011,011 0112011 0 CHon OH HO— ~—H 11112101121013 0 _ H semen3 OH H°——H OH = o — H 0“ H——OH 5%AO0Hin OH OH n 01130101120 3:1 n H— —OH HO— —H HO — r—H 0H0 0112111110112170113 $0201 ., E1311 C 0 WW ,. 5:1 "(GHQZ‘ 011,011 0 CH2“. OH HO — '-H 1.}. _ 11 OH OH n H— —OH HO— —H CH2N(CH2)7CH3 I 8“ N(CHzah Figure 2.28: Preparation of N,N-dansyloctylamine derivatives. 70 The chromophore is a dansyl [5-(N,N-dimethyl)aminonaphthalene sulfonyl] group which aids UV detection during HPLC and also stabilizes positive ion species formed during analysis by fast atom bombardment mass spectrometry. The hydrophobic tail provided by an octyl group allowss the derivatized oligosaccharides amenable to be separated by reverse-phase chromatography using a C- 18 stationary phase. This method was applied to analyze a mixture of starch maltodextrins of degree of polymerization 3-14, maltohexaose, and maltose. 2.2.7.2. Positive FAB-MS of N,N-Dansyloctylamine Derivatized Maltose The (+) FAB-MS of N,N~dansyloctylamine derivatized maltose is shown in Figure 2.29. The ion at m/z 711 is the sodium salt adduct of N,N- dansyloctylamine derivatized maltose. For this spectrum, sodium hydroxide was used as the base to attach the dansyl group. Because the spectrum contained many sodium adduct peaks, an improved method was used with the triethylamine as base. The peak at m/z 362 is due to the 100- ‘ ‘ R 1 17? 36k 711 e 1 10:! I 1 t . l . V 60‘ e . A 1 b 40‘ u 1 3 . 251 a 20. ' 15.0 n J ° I 13 01.1.1144 1 00 200 300 400 500 600 700 8% Figure 2.29: (+) FAB-MS of N ,N-dansyloctylamine derivatized maltose. 71 derivatizing reagent. The peak at m / z 102 is due to the protonated thioglycerol matrix. 2.2.7.3. Positive FAB-MS of N,N-Damyloctylamine Derivatized Maltohexaose The (+) FAB-MS of N,N-dansyloctylamine derivatized maltohexaose is shown in Figure 2.30. For making the derivatives, 47 mg (~ 4.5 x 10'7 mole) of maltose was used as starting material and 0.1 pmol derivatized material was put on the probe. The peak at m/z 1337 is due to the protonated N ,N -dansyloctylamine derivatized maltohexaose molecule. The next peak is the strong sodium adduct peak at m/z 1359. In the lower mass range, most peaks arise from the matrix and organic reagents. The peak at m/z 102 is protonated thioglycerol, m/z 363 and 454 are from protonated N,N-dansyloctylamine and its glycerol adduct, respectively. In contrast to Figure 2.2 of the negative FAB-MS from DNPO derivatized maltohexaose, the positive FAB-MS showed no sequence peaks at all. In preparing dansyl I 1 111m 2‘?‘ 3?: E4 “‘59 .5 o O L A A A A A A A A A A A A I ooamnaca) o<-n-om h 0 527 728 J-. J. - ‘AL 1 Y f 1 T I I j 200 400 600 800 '1050' ' 7300' ' '14'oo'm GA A A Figure 2.31: Positive FAB-MS of derivatized maltohexaose. 72 derivatives, sodium hydroxide was used as the base. Even after the sample was purified by a C- 18 reversed phase cartridge, the trace amounts of alkali metal ions still decreased the fragmentation abundance. 2.2.7.4. Positive FAB-MS of N,N—dansyloctylamine Derivatized Maltodextrin The (+) FAB-MS of N,N—dansyloctylamine derivatized maltodextrin is show in Figures 2.321a and b. The maltodextrin is a starch-derived mixture of oligomers. In the positive mode, a mixture of dextrin with degree of polymerization 1-15 can be detected. Components present at the level of 0.1 pmol could be detected. This method gave a good yield of cations for the derivatized maltodextrin. In Figure 2.31, the sequence ions observed are protonated Yn ions with an adduct of matrix compound glycerol. Molecular ions from components of the mixture of oligomers which have the same ions as same the Yn ions. In the lower range of 520-1800, the abundance of the Y1 ion is as high as 100% and other Yn ions are still dominant. The other sequence ions are 1:5Xn ions from the protonated molecular species. They are low in abundance and disappear above approximately m/z 1600. There are sequence ions which arise from the main Yn ions by loss of a water. These appear at m/z 600, 762, 924, 1086, 1248, 110, 1572, 1734 without specific labeling. In another sequence, the ions are a little higher than the main ions. This sequence is due to the sodium salt adduct to the main Yn ions and the fragments appear at m /z 640, 802, 964, 1126, 1288, 1450, 1612, 1774, 1936, 2098, 2260, 2422, 2584, 2746 and 2908. A few ions appeared at m/z 647, 809, 971, corresponding to the 1.5x“ ions plus the matrix compound glycerol. The mixture of derivatized oligomers could easily be separated by HPLC. Based on the relative proportions of the 00390.3(:0') O<—""N_¢DI..2 3.. 13.1.81 .3... e13 Y1+Gly+H 1266 Ys+G|y+H 4 Y.+Gly+H 1428 Yz+Gly+H 780 i/ 924 971 1., 964 A n 1 I n 600 800 1000 1 200 1 400 1600 M/Z‘ 800 Figure 2.31 a: (+) FAB-MS of N,N-dansyloctylamine derivatized OOSWQDCU> 0<—"'N_OD A n I maltodextrin. 2562 Y13+Gly+H 2594 Y.,».Glym Y15+Gy+H Y.+Gly+H 2722746 2806 1114 29011 Y1°+GIy+H 2076 1935 2053 2096 Y12+GIY+H 2400 2422 . . . . ‘1 11 ‘ 3 1 ‘ l M ‘ ‘ 2000 2200 2400 2600 2800 M [2000 Figure 2.31 b: (+) FAB-MS of N,N-dansyloctylamine derivatized maltodextrin. 74 individual oligomers calculated from this analysis, the highest oligomers were present in amounts of about 0.1 pmol; however, they could easily be detected in mass spectra of the entire mixture. This represents an improvement in sensitivity of between 100- and 1000-fold compared to the aminobenzoic acid alkyl ester method (16). In our analyses, single scans were taken and no attempt was made to improve signal-to-noise by signal averaging. In some cases, there was no sample work-up after derivatization. 2.2.7.5. FAB-MS of N,N-dansyloctylamine Derivatized Oligosaccharide from the Bacteria Sm'face As a further test of the efficiency of this derivatizing procedure for enhancing signal intensity for positive FAB-MS, a previously characterized oligosaccharide from Rhizobium trifolii 843 capsular polysaccharide ( 1), 011,011 11000 0 011 00 ’ CH2 X 0 0 2%” 00H Pic—h." 11,0 0 on HOOC 0 I OH 011 0 —H X 0 0 11——011 011 CH20 11,0 0 0” 0 110— —H 0001-1 011 0H 0 o 31121110112170113 00011 0A0 011 '8'0 R-H300HOHCH200- / 0 0 011 011 N‘ 011 0115’ 0113 Figure 2.32: N,N-dansyloctylamine derivatized oligosaccharide from a bacterial cell surface. 75 was derivatized and analyzed. The structure of the N,N-dansyloctylamine derivatized oligosaccharide is shown in Figure 2.32. The reaction was performed on 64 ug (~ 3.8 x 10'7 mole) of isolated pure oligosaccharide, and about 3 pmol of derivatized material was put on the probe. The positive FAB-MS of N ,N -dansyloctylamine derivatized oligosaccharide from the bacterial surface is shown in Figure 2.33. Since base treatment was involved in the derivatization procedure, loss of the acetyl and 3-hydroxybutyryl groups in the derivatized molecule was expected. There are four carboxyl groups in the derivatized :2 1689 ." 1711 L 1667 . 1733 3 1757 1797 1925 : 17751811841 1903 1943969 E 1863 199. M/Z Figure 2.33: (+) FAB-MS of N ,N -dansyloctylamine derivatized oligosaccharide from the bacterial cell surface. oligosaccharide; therefore, clusters of sodium salt adducts were observed in the spectrum. The peak at m/z 1903 arises from the derivatized molecule (molecular weight 1810) plus a matrix compound glycerol (93 amu). The clusters at m/z 1925, 1947 1969 and 1991 arise from sodium adducts of the 76 four carboxyl groups on the compound. The peak at m/z 1775 arises from the protonayed derivatized molecule with loss of two water molecules, and the clusters at m/z 1797, 1819, 1841 and 1863 are sodium adducts to the four carboxyl groups. The ion at 1757 is loss of a water molecule from m/z 1775. The ion at m/z 1623 arises from the protonated derivatized molecule loss of a terminal unsaturated hexaose (159 amu) and an aldehyde group. The clusters at m/z 1645, 1667, and 1689 are the sodium adduct to the three carboxyl groups, the ions at m/z 1711 and 1733 are sodium adducts on the molecule. 2.2.8. Conclusion The method described here provides an excellent yield of derivatized material, without underivatization or any side reaction being observed in the spectrum, in a very efficient fashion with a minimum of sample preparation and purification before analysis. The molecular weights of the derivatized compounds are much lower than those obtained from the peralkylation method. This method also has a minimal effect on the sensitivity. Useful structural information was obtained from the ion series observed, without using MS/MS conditions. The sensitivity obtained using this derivative is 100 fold superior to that reported from other methods, and the derivatives are “well-behaved” during HPLC analysis; the peak shapes on HPLC are good with no tailing, and separation can be effected using very simple solvent systems. The sensitivity in FAB-MS analysis is also very good, the limit of detection is as low as 0.01 pmol. The spectra obtained from large oligosaccharide have relatively high ratio of signal to noise. One reason for this is due to the direct analysis of the sample without purification, another reason is that the spectra was done in a single scan 77 without peak average, also large molecules have lower effect from the derivative, the surface activity becomes lower, and the appearance of the sputtered ionization occurs with more difficulty. To distinguish the spectra of isomers, comparision of the spectra of maltohexaose and isomaltohexsaose, Yn sequence was performed. Four sequences were observed from maltohexaose and two sequences were present in the spectrum of isomaltohexaose. The 1'4An from maltohexaose is impossible to obtain in the isomaltohexaose, because after loss of the 1:“A,, species, the isomaltohexaose is still in the aggregated form, like loss of functional groups, which does not release enough energy from the molecule. So the type of fragmentation sequences obtained from the molecules provide information about sequence and some hint of linkages. In the case where there are monosaccharides with different ring size in the polysaccharide, such as furanose and pyranose rings, we can predict the rings with different size are going to give different kind of fragments, even though we have no experince with this case. Also both pyranose and furanose forms will give the same Yn sequence. In the case of large oligosaccharides, such as maltodextrin and starch hydrolyzed oligosaccharides, the most popular sequences observed are (3,, sequences. The Cu sequence is formed from the bond cleavage between the oxygen and C4. The reason we obtain Cn sequence only in large oligosaccharides is possibly due to the fact that this fragmentation sequence can release more energy from the molecules. In most cases, the desorption techniques such as FAB-MS, plasma desorption, and field desorption, generally provide only molecular mass information for many large nonvolatile compounds, revealing only minor structural information from specific fragmentation. The sequence ions 78 obtained in this method are very useful for studying molecular structure, ion structure, and ion fragmentation patterns, which are used to predict and interpret fragmentation of unknown oligosaccharides, glycosides, and also can be extended to more complex hetero-oligosaccharides, especially those from glycoproteins isolated from low-abundance biological samples. This method is that it can give very good sequence information, and some clue for the linkages, but the weakness is to obtain the real linkage information, methylation analysis is still required. Also the cOnfiguration of anomeric proton can not be obtained from this method, but it is easily obtained from NMR spectrometry. 2.3 Reference 1. Hollingsworth, R. 1., and Dazzo, F. B., Carbohydr. Res, (1988) 172 97-112. 2. McIntire, F. C., Clements, L. M. and Sproull, M., Anal. Chem. (1953) 25 1757-1758. 3. Coates, M. L., Wilkins, C. L., Biomed. Mass Spectrom. (1985) 12 424- 428. 4. Coates, M. L., Wilkins, C. L., Anal. Chem. (1987) 5.9 197-200. Takayama, K., Qureshi, N., Hyver, K., Honovich, J ., Cotter, R. J ., Mascagni, P., Schneider, H. J ., J. Biol. Chem. (1986) 26.1 10624-10631. 6. Cotter, R. J ., Honovich, J ., Qureshi, N ., Takayama, K., Biomed. Environ. Mass Spectrom. (1987) 1:1 591-598. 7. Damon, B., and Costello, C., J. Glycocojugate, (1988) 5 397409. 8. Hendrickson, J. B., Angew Chem. Int. Ed. Engl. (197013 47-76. 9. Poulter, L. J ., Earnest, J. P., Strond, R. M., and Burlingame, A. L. Biomed. Environ. Mass Spectrom, (1988)1§ 25-30. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 21. 79 Ganguly, A. K., Capucino, N. F., Fujiwara, H., and Rose, A. K., J. Chem. Soc. Chem. Comm, (1979) 148-149. Prome, D., Prome, J .-C., Puzo, G., and Aurelle, H., Carbohydr. Res. (1985) 1:10 121-129. Biemann, K., “Biological Mass Spectrom.” Burlingame, A. L., and McCloskey, J. A., (Eds), Elsevier Science Publishers B.V., Amsterdam. Spengler, B., Dolce, J. W., and Cotter, J. R., Anal. Chem. (1990) 52, 1731-1737. Sasaki, H., Norinichi, 0., Dell, A., and Fukuda, M., Biochemistry (1988)218618-8626. Williams, J. M., “Starch and its Derivatives” (1968) 91-138, (Radley, J. A., ed.), 4th ed. Chapman & Hall, London. Lee, E. Y. C., Mercier, C., and Whelan, W. J ., Arch. Biochem. Biophys. (1968) 125 1028-1030. Morrison, W. R., and Karkalas, J ., “Methods In Plant Biochemistry“ (1990) 2 323-352. Manners, D. J ., Essays Biochem. (1974) 10 37-71. Manners, D. J ., and Matheson, N. K., Carbohydr. Res. (1981) 20, 99- 110. Callaghan, P. T., and Lelievre, J ., Carbohydr. Res. (1987) 52 33- 40. Takeda, Y., and Hizukuri, S., Carbohydr. Res. (1987) 158 79-88. Manners D. J. “Biochemistry of storage carbohydrates in green plants” (1985) 154, Academic Press, London. Kerr, R. W., and Severson, G. M., J. Am. Chem. 300., (1943) 65, 193. Posternak, T., Helv. Chim. Acta., (1935) 18, 1351. 26. 27. 29. 31. 32. 8) Parrish, F. W., and Whelan, W. J ., Staerke, (1961)13, 231. Postnak, T., J. Biol. Chem., (1951)18_8 317. Hizukuri, S., Tabata, S., Kagoshima, and Nikuni Z., Die Starke (1970)1Q 338-343. Tabata, S., and Hizukuri, S. Die Starke (1971) 8 267-272. Hizukuri, S., Takeda, Y., and Yasuda, M. Carbohydr. Res. (1981) 251 {105213. Westler, R. L., “Methods in Carbohydrate Chemistry” (1964) 4 Westler, R. L., ed., Academic Press, New York. Khorana, H. G., “Some Recent Developments in the Chemistry of Phosphate Esters of Biological Interest” John Wiley & Sons, Inc., New York - London, (1961) 13-43. Naik, S., Oates, J. E., Dell, A., Talor, G. W., Dey, P. M., and Pridham, J. B., Biochem. Biophys. Res. Comm. (1985) 1.32 1-7. CHAPTER3 EXPERIIVIENTAL FOR PART I 3.1 ReparationofNN-(Z4-dinih'ophenylhctylamine Derivatized Maltose Nine mg (2.5 x 10'5 mole) of maltose, 0.1 mL water, 3 equivalents octylamine, and 4 equivalents sodium cyanoborohydride were placed in a Teflon-lined, screw cap, one dram vial. Two-tenths mL 5% acetic acid in a solution of water/methanol (3:1) was then added; the vial was capped and heated at 70° C for 1 hour, then blown to dryness. Afterwards, 0.1 mL of 0.2M sodium bicarbonate and 2 equivalents of 2,4-dinitrofluorobenzene in 2 mL methanol were added (1). The mixture was heated for 25 minutes at 60°C, and subsequently 0.4 mL of 0.2 M sodium hydroxide in dioxane containing trace amounts of water were added to dissolve the base. The mixture was then heated at 60° C for 1 hour. After cooling to room temperature, 5 mL water and 5 mL chloroform were added. The contents of the aqueous layer were adsorbed on a 018 reversed phase cartridge column. The column was eluted with 1 mL water, 1 mL water/methanol (1:1), and 1 mL methanol. All three fractions of elutant were analyzed by NMR and FAB-MS. Most of the carbohydrate was contained in the water fraction. 81 3.2 82 UV Spectrum of N,N-(2,4-dinitrophenyl)octylamine Derivatized Maltose d- d .- «11- d 0.00 s 1 . . . 200 260 320 380 440 500 nm Figure 3.1: UV spectrum of N,N-(2,4-dinitrophenyl)octylamine 3.1. 3.4 3.1. derivatized maltose. Preparation of N,N-(2,4-dinitrophenyl))octylamine Derivatized Maltohexaose Maltohexaose (650 mg, ~ 6.3 x 10'7 mole) was derivatized as in section Preparation of N,N-(2,4-dinih'ophenyl)octylamine Derivatized Maltodextrin Maltodextrin of 17 mg (~ 9.6 x 10'6 mole) was derivatized as in section 83 3.5 UVSpectrumofN,N-(2,4-di_“ .' f} ‘,' ' Derivatized Maltodextrin 1.60" mOZ>CDmOCDUJ> .0 A o l 0.00 5 4 + 5 i t l 5 i i 200 260 320 380 440 500 nm Figure 3.2: UV spectrum of N,N—(2,4-dinitrophenyl)octylamine derivatized maltodextrin 3.6 Preparation of N,N-(2,4-di__" r‘ ff ‘,' ' Derivatized Oligosaccharide Separated metheBactcriaSurface The 850 mg (~ 5.1 x 10'7 mole) of oligosaccharide from the bacteria surface was derivatized as in the section 3.1. 3.7 Preparation of N,N-Densyloctylamine Derivatized Maltose A sample of maltose, 16 mg (~ 4.8 x 10'5 mole), ~3 equivalents octylamine and 4 equivalents sodium cyanoborohydride were added to a one dram vial, after which 0.2 mL 5% acetic acid in a solution of water/methanol (3:1) was added. The Teflon-lined vial was capped and 84 heated at 70° C for 1 hour. Subsequently, the solution was reduced to dryness. Saturated sodium bicarbonate 0.2 mL, 3 equivalents dansyl chloride (5-N,N-dimethyl naphthalene sulfonic chloride), and 0.2 mL methanol/water (4:1) were added consecutively (2), and the mixture heated at 45° C overnight. Samples were analyzed directly without purification. 3.8 Preparation ofN,N-Densyloctylamine DerivatizedMaltodextrin The 93 mg (~ 6.3 x 10'5 mole) of maltose was derivatized as in the section 3.7. 3.9 Preparation of N,N-Densyloctylamine Derivatimd Oligosacclnrlde Separatedfi'omBactm'iaSurfnce The 640 mg (~ 3.8 x 10'7 mole) of maltose was derivatized as in the section 3.7. 3.10 Preparation otNN-Densyloctylamine DerivatizedMaltohexaoee Method II) The 770 pg (~ 7.4 x 10'7 mole) maltose, ~ 5 equivalents octylamine and 4 equivalents sodium cyanoborohydride were added to a 1 dram vial; then 0.1 mL 5% acetic acid in a solution of water/methanol (3:1) was added, the Teflon-lined vial was capped and heated at 70° C for 1 hour. Subsequently, the solution was blown to dryness. Triethanolamine (0.1 11L), 2 equivalents dansyl chloride (5-N,N-dimethy1 naphthalene sulfonic chloride), and 0.2 mL methanol/water (4:1) were added consecutively, and the mixture heated at 45° C overnight. Samples were analyzed directly without purification. 85 3.11 PmmraflondNN-(zwmwhctfimim Derivatized Maltohexaose (MethodI) The 730 ug (~ 7 x 10'7 mole) of maltohexaose, 5 equivalents octylamine, and 6 equivalents sodium cyanoborohydride were placed in a one dram vial. Two-tenths mL 5% acetic acid in a solution of water/methanol (3:1) was then added. The vial was capped and heated at 70° C for one hour, then blown to dryness. Afterwards, 1 mL of triethylamine and 2 equivalents of 2,4-dinitrofluorobenzene in 2 mL methanol, 0.2 mL water were added. The mixture was then heated at 45° C overnight. After cooling to room temperature, 5 mL water and 5 mL ethyl acetate were added. The contents of the aqueous layer were analyzed directly by FAB-MS. 3.12 PreparationofNN-Densyloctylamine DerivatizedMaltohexaoee (MethodI) The 470 11g (~44.5 x 10'7 mole) of maltose, ~5 equivalents octylamine and 4 equivalents sodium cyanoborohydride were placed in a one dram vial; and 0. 1 mL 5% acetic acid in a solution of water/methanol (3:1) was added, the Teflon-lined vial was capped and heated at 70° C for one hour. Subsequently, the solution was blown to dryness. Triethylamine (0.5 1.1L), 0.5 1.1L 2,4-dinitrofluorobenzene, and 0.2 mL methanol/water (4:1) were added consecutively, and the mixture heated at 45° C overnight. Samples were analyzed directly without purification. 3.13 PreparationofG-PhosphateMaltohexoase One-tenth mL of 0.014 M N,N-(2,4-dinitrophenyl)octylamine derivatized maltohexaose was placed in a one dram vial and blown to 86 dryness to remove excess solvent; the vial was kept in an ice bath, and 0.1 mL dimethoxyethane (DME), and 1 11L POC13 were added. The reaction was kept in an ice bath for one hour, then 0.02 mL water was added. The sample was analyzed directly by FAB-MS, without purification. 3.14 Preparation of N,N-(2,4-dinitrophmnrl)octylamine Derivatized N,N’,N”-'h1’acetylchitotriwe The 0.0027 M aqueous solution of N,N’N”-triacetylchitotriose (40 11L, ~ 1.1 x 10'7 mole) was placed in a one dram vial, followed by addition of 25 11L of 24 M octylamine in 5% acetic acid in water/methanol (3:1) solution and addition of 25 11L of 4 M sodium cyanoborohydride in 5% acetic acid water/methanol (3:1) solution. The mixture was heated at 40° C overnight. After cooling to room temperature, 25 11L of 56 11M triethylamine in methanol and 25 11L of 64 11M 2,4-dintrofluorobenzene were added. The mixture was heated at 45° C for two days and analyzed by FAB-MS without further purification. 3.15 Preparation ofN,N-(2,4-dinitznphenyl)octylamineDerivatizedLacto- N-Fucopentome The 7.8 x 10'4 1.1M aqueous solution of lacto-N-fucopentoase (40 1.1L, ~ 1.1 x 10‘7 mole) was derivatized as in section 3.14. 3. 16 Preparation of N,N-(2,4-dinit10phenyl)octylamine Derivatized Lacto- N-Difucohexaoee The 8.2 x 10‘4 M aqueous solution of lacto-N-difucohexaose (40 11L, ~ 1.1 x 10'7 mole) was derivatized as in section 3.14. 87 3.17 Preparation ofN,N-(2,4-dinitznphenyl)octylamine Derivatized Isomaltohexaose The 8.2 x 10‘4 11M aqueous solution of isomaltohexaose (40 11L, ~ 1.1 x 10'7 mole) was derivatized as in section 3.14. 3.18 EnzymeDigestion ofPotato Stamhl Potato starch 54 mg, and 0.7 mL (0.15 unit/mL) of a-amylase were placed in a one dram vial. The a-amylase was suspended in a 2.9 M NaCl solution containing 3 mM CaClz. Acetic acid (0.7 mL of 0.02 11M) buffered with NaOAc (pH 5) was added. The sample was heated at 100° C for 60 minutes and quenched with 5% acetic acid solution. The digested starch was derivatized by N,N-(2,4-dinitrophenyloctyl)amine and analyzed by FAB- MS. The spectrum was not as good as that of maltodextrin, possibly because the starch is not sufficiently water-soluble, and the reaction was not complete. 3.19 AcidHydrolysisof Starch The 72 mg of starch was placed in a one dram vial, followed by 100 1.1L of 0.1 M trifluoroacetic acid. The sample was heated at 75° C for 45 minutes, then N ,N-(2,4-dinitrophenyl)octylamine was derivatized as previously described. The sample was purified by mixed bed resin, and sequence FAB-MS spectrum was obtained. 3.20 EnzymeDigestion of Potato Starchll Potato starch 100 mg and 2 mL DMSO were placed in a one dram vial. The sample was placed at room temperature for 2 days, then 20 11L of 111-amylase (5 units/11L in 50 mM citrate, pH 4.6) was added. The mixture 88 was heated at 45° C and completion of the reaction was indicated by iodine solution (12 0.002%, KI 0.2 % in 0.2 M acetate buffer, pH 4.8). After approximately 20 hours, the color of the iodine solution became pale, the sample was blown to dryness, and derivatized by N,N-(2,4- dinitrophenyl)octylamine using the same procedure described earlier. 3.21 Preparation of N,N-(ZA-dinitrophenylhctylamine Derivatized oligosaccharide isolated fi'om glycoprotein Fetuin Two-tenth mL enzyme digested fetuin was placed in a one dra vial. The sample was lyophilized to dryness and derivatized as in section 3.5. 3.22 MassSpectmmetry FAB-MS was performed on a JEOL HX110 (Peabody, MA) double- focusing mass spectrometer (EB configuration) equipped with a high field magnet operated in the negative ion mode. Ions were produced by bombardment with a beam of Xe atoms (6 kV). The accelerating voltage was 10 kV, and the resolution was set at 1000 or 3000 according to the mass range of interest. The samples were dissolved in a 1:1 (v/v) H20-MeOH solution; generally 1-1.5 11L sample was mixed with 1 11L of glycerol/thioglycerol (1:1) or triethanolamine on the FAB-MS stainless steel probe tip. The solvent was pumped away by the mass spectrometer before analysis. Calibration was performed using Ultramark (443 and 6121) or (CsI)nI‘ cluster ions, depending on the mass range of interest. A JEOL DA-5000 data system recorded the spectra. The sample was scanned at a scan slope of 2 min from m/z 0-3000. Data presented were acquired in a single scan and found to be reproducible. 3.3 Refiemmces 1. McIntire, F. C., Clements, L. M. and Sproull, M (1953) Anal. Chem. 25, 1757-1758 2. Renner, D., and Spiteller, G., Angew. Chem. Int. Ed. Engl. (1985) 24 408-409. PART II: STRUCTURE ELUCIDATION AND SYNTHETIC TRANSFORMATIONS RELATED TO THE INNER AND OUTER CORE OF THE LIPOPOLYSACCHARIDES OF RHIZOBI UM MELILOTI 41 AND LEGULIINOSAR UM BIOVAR VICIAE VF-39 MUTANTS ABSTRACT PART II: STRUCTURE ELUCIDATION AND SYNTHETIC TRANSFORMATION RELATED TO THE INNER AND OUTER CORE IN THE LIPOPOLYSACCHARIDES OF REZOBIUM MELILOTI 41 by Yuanda Zhang The structure of the inner and outer core of the LPS of Rh izobium meliloti 41 has been analyzed by a combination of chemical and spectroscopic methods. An unusual sugar component has been studied. The structure was investigated by J -correlated, 1H-NMR spectroscopy. The composition and structure of the carbohydrate chain was analyzed by methylation analysis using GC-MS. Conformation of the structure (but not configuration) of the discovered compound was attempted by synthesis. The LPS of a wild type strain of Rhizobium leguminosarum biovar viciae (strain VF-39) and two symbiotically defective Tn5 mutants (VF-39-32 and VF-39-86) have been studied. The LPS of the mutants reflected impaired synthesis of the O-antigen. This study clearly defines a role for the bacterial LPS in the proper functioning of the Rhizobium legume symbiosis. A trisaccharide containing an 1:1-linked O-mannopyranosyl residue on the 3-hydroxymethyl group and an 1:1-linked O-galactopyranosyl residue at the 5 position of a new branched pentafuranose sugar has been isolated from strain VF 39-32, a symbiotically defective strain of Rh izob ium leguminosarum biovar viciae impaired in normal lipopolysaccharide synthesis. The branched glycosyl residue has been identified as a 2-deoxy-3- C-(hydroxymethyl)pentofuranose and appears to have the threo- configuration. The structure of the new glycosyl component was determined by 1H-NMR spectroscopy using spin decoupling to determine proton connectivities and by mass spectrometry. The positions of substitution were deduced by proton n.O.e. studies on the native oligomer and the configuration determined by an analysis of coupling constants, n.O.e., and spin decoupling data. The ring size of the branched chain glycosyl component was deduced from coupling constants. Signals for small quantities of this unusual component can be discerned in 1H-NMR spectra of lipopolysaccharide fragments from the parent wild-type organism, suggesting that its synthesis is repressed by that of the normal lipopolysaccharide and significantly repressed in the mutant strain. Traces of a trisaccharide which is likely to be a truncated version of a tetrasaccharide, normally synthesized by the parent strain, containing galactose, galacturonic acid, mannose and 3-deoxy-2- octulosonic acid, were also detected. CHAPTER 4 STRUCTURE ELUC IDATION AND SYNTHETIC TRANSFORMATION RELATED TO THE THE INNERAND OUTER CORE INTHE LIPOPOLYSACCHARIDES OF RHIZOBIUM MELILOTI 41 4.1 Introduction Carbohydrate molecules found naturally or obtained by synthesis are playing a very important role in current chemistry, biochemistry, microbiology, and other areas. The project described in part II has two objectives: the structure elucidation and synthesis of key components of the lipopolysaccharide (LPS) of the nitrogen-fixing bacterium Rhizobium meliloti 41. This organism effects the reduction of dinitrogen gas into ammonia, an essential step in providing plant protein and nutrition. The interest in this research is the chemical structure of the inner and outer core lipopolysaccharide from Rhizobium meliloti 41 LPS mutant. Rhizobium is a species of Gram-negative bacteria which has a nitrogen-fixing symbiotic relationship with legume plants (1). Like most Gram-negative bacteria, Rhizobium has usual surface polysaccharides consisting of lipopolysaccharides (LPSs), extracellular (EPSs) and capsular polysaccharides (CPSs). The major surface polysaccharides of these bacteria are lipopolysaccharides. The symbiosis is characterized by a high degree of specificity with respect to host-range. The bacteria from a given species will selectively infect only one or two host species. In order to assess the role of Rhizobium lipopolysaccharides in the specific recognition between the symbiont bacterium and the host cell, we propose to characterize these molecules structurally. 90 91 Lipopolysaccharides are macromolecules which determine the surface chemistry of Gram-negative bacteria. Lipopolysaccharides are located in the cell envelope (Figure 4.1) (2). They are composed of three main structural / §—-> O-Antlgen Outer Core . 3 -—) Heptose Lipopoly- s sacchande Penn 1 | KDO Lipid A ‘ Outer Membrane Lipoprotein 1 Peptidoglycané— MDO(——°-1:r 10> Penplasm Phospholipids.(_.n “ .- U U Figure 4.1: Schematic molecular representation of the E. coli envelope. Ovals and rectangles depict sugar residues. Circles represent the polar head groups of phospholipids. MDO are membrane-derived oligosaccharide, and KDO is 3-deoxy-D-manno-octulosonic acid. KDO and heptose make up the inner core of LPS. The figure was taken from Christian R. H. Raetz (2) with modifications. regions. The first region of LPS is the O-antigen, which is usually a repeating sequence of carbohydrate molecules which determines the serology of the organism (Figure 4.2) (3). The O-antigen is, typically, highly variable. 92 The O-antigen is attached to a second carbohydrate region, which is much smaller and not a repeating sequence, called the R-core. This core region contains the LPS, marker sugar called KDO, and is usually conserved within a given genus. It might be expected that the core structures of Rhizobium trifolii and Rhizobium meliloti might be the same. The third region of the LPS called lipid A, is made up of a carbohydrate group (usually a disaccharide) to which a variety of 12 to 16 carbon 3-hydroxyalkanic acids are } attached. The lipopolysaccharide molecule is anchored to the outer ‘- membrane of the bacterium via the fatty acids of the lipid A region. .. It has been reported that the lipopolysaccharide of Rhizobium trifolii t genpfatmgl [Outer Core] [Inner Core ] [O-Specific Chain ][ Core ][ LipidA ] [ Polysaccharide ]| Lipid A | 0 Sugar residue, l-Glycero-D-manno-heptose, @2—Keto-3-deoxy-D—manno-octonate, @D—Glucosamine, P Phosphate, NEthanolamine, «wwv 3-Hydroxyacyl group. Wm 3-Acyloxyacyl group Figure 4.2: Schematic structure of a Salmonella lipopolysaccharide. The figure was taken from the review of Westphal et al. (3) with modifications. ANU843 (a strain of a related species) contains a trisaccharide (Figure 4.3) (4) and tetrasaccharide (Figure 4.4) (5) as the components in the core 93 oligosaccharide. According to the above reports, both the O-antigen and the core oligosaccharides contain 3-deoxy-D-manno-2-octulosonic acid (KDO) at their reducing end, and are linked to the rest of the LPS molecule via KDO. In this study, we attempt to characterize the structure of the core oligosaccharide of the LPS of Rhizobium meliloti and describe efforts to determine the structure of one unusual glycosyl component. This work will also allow us to relate host range to LPS structure by studying organisms from two different species. Ho 11,011 OHO,CH :‘lH l Oil—IO COOH OH Figure 4.3: Structure of trisaccharide from Rhizobium trifolii ANU843 (4). HO HOOC 0 H H HO H H H OH H 0 HH H HO H OH 0/ Figure 4.4: Structure of tetrasaccharide from Rhizobium trifolii ANU 843 (5). 94 4.2 Materials and General Methods 4.2.1 Further Purification of the Core Oligosaccharides (4): The core oligosaccharides were initially isolated by mild acid (1% acetic acid) hydrolysis of lipopolysaccharides, followed by gel filtration chromatography of the water-soluble fraction on a Bio-Gel P-2 column using 1% formic acid as eluant. The fractions containing carbohydrate were identified by colorimetric assay, pooled, then lyophilized for NMR. 4.2.2 Composition Analysis (5): The reducing terminus and keto-groups on the oligosaccharide were first reduced to alcohols with sodium borodeuteride. Carboxyl groups were converted to methyl esters by 1% HCl in methanol, then transformed to alcohols by reduction with sodium borohydride or sodium borodeuteride. The reduced oligosaccharides were hydrolyzed to monosaccharides with aqueous trifluoroacetic acid (TFA), reduced to alditols with sodium borohydride, and peracetylated with acetic anhydride/pyridine, followed by GC-MS analysis. 4.2.3 Methylation Analysis (6): The oligosaccharide (1 mg) was pre-reduced with 10% aqueous sodium borohydride, then converted to methyl esters by using 1% methanolic HCl. The methyl esters were reduced with lithium alumina deuteride (LAD). The sample was permethylated using sodium methylsulfenyl anion and methyl iodide. The resulting oligosaccharide was hydrolyzed by 1 M TFA, and converted to alditols by reduction with sodium borodeuteride. Peracetylation was performed by dissolving the oligomers in a minimum volume of TFA and an equal volume of a 2:1 mixture of acetic acid and trifluoroacetic acid anhydride. The mixture was left at room temperature for 6 hours, and T-——-m-A—n _"‘—“—' —-. O 95 concentrated to dryness under a stream of nitrogen. The residue was mixed with chloroform, and partitioned between chloroform and 1 M HCl. The chloroform layer was dried with sodium sulfate and evaporated to dryness for GO and GC-MS analysis. 1H NMR spectra were recorded with a Varian VXR300 and VXR500 spectrometer operated at 300 MHz and 500 MHz respectively, and 13C NMR spectra were recorded with the Varian VXR300 spectrometer operated at 75.4 MHz. All chemical shifts are referenced to Me4Si. GC-MS analyses were performed on a JEOL JMS-AX505H spectrometer interfaced with a Hewlett- Packard 5890A Gas Chromatograph using an ionizing voltage of 70 eV. Electron impact analysis of the per(trideutero)acetylated oligomer was performed on the same instrument at the same voltage by direct insertion probe. 4.3 Results and Discussion 4.3.1 Structure Elucidation of the Inner and Outer Core Rhizobiurn Meliloti 41 From the analysis of inner and outer core components of the lipopolysaccharide of Rhizobium meliloti 41, some primary information was obtained which suggested that one of the components was an unusual sugar. An earlier report (4) indicated that, the core regions of LPS from R. trifolii are composed largely of galacturonic acid (7). More detailed studies of R. trifolii LPSs has shown that both the O-antigen and the core oligosaccharides contain KDO at their reducing end. This result indicates that Rhizobium LPSs are quite different from those from Salmonella and E. col i. The major oligosaccharide from the core region of the lipopolysaccharide from R. trifolii ANU 843 was isolated and its structure determined to be a 96 trisaccharide consisting of two galacturonic acid and a KDO residue (4). The 1H-NMR spectral analysis (Figure 4.5) of the trisaccharide from Rhizobuim trifolii ANU 843 showed that the proton of the methylene group at 3-position of the KDO appeared in the proton spectrum as a doublet of doublets (J = 12.5 and 2.9 Hz) at 81.82 and a triplet (J = 12.5 Hz ) at 82.17. The triplet was assigned to the axial proton which should have a large splitting fi'om the other geminal proton. The equatorial proton should have a comparatively small vicinal coupling, due to being cis to the proton on G4. Whether or not Rhizobium LPSs play a role in the specific recognition between the symbiont bacterium and the host cell is still uncertain (8, 9), but the LPSs may be important in later symbiont events, because the mutants lacking the O-antigen portion of the LPS are defective in nodulation (10). In order to understand the roles of Rhizobium LPSs in symbiosis, the structural characterization of these molecules is very important. A further study reported the second core oligosaccharide from the LPS of R. trifolii ANU 843 is a tetrasaccharide composed of galactose, galacturonic acid, mannose, and a KDO residue. The mannose residue is (IL-linked to the 4-position of the KDO residue, and the galacturonic acid residue is 111-linked to the 6-position of mannose. The galactose residue, acetylated at the 4-position, is attached to the 4-position of mannose by 111-linkage. The 1H-NMR spectra (5) (Figure 4.6) of the tetrasaccharide from R. meliloti 41 sugestion contained a triplet (J = 11.4 Hz) at 82.06 and a distorted doublet of doublets with a larger splitting of 11.4 Hz between 81.74 and 1.86 ppm. These signals were assigned to the axial and equatorial protons, respectively, at the 3-position of a KDO residue. One of the oligosaccharides from the LPS of Rhizobium meliloti 41 had an unusual NMR spectrum. In this component there were no usual peaks between 81.82-82.17 for KDO as in Figures 1 and 2, but instead, the 500 MHz 97 o 011 H30 0 011 0 011 0 H3! 7 "O 0 "Cl ,, .. a€11,011 | c o/‘\ou B I 1 A .1. A 1Trrt11lltvullruvlunvluutullrl1'11-11111111111‘111uln- 5 3 2 1 9.9.111. Figure 4.5: Structure of trisaccharide, and (A) Proton NMR spectrum in D20 from Rhizobium trifolii ANU 843, (B) Spectrum after irradiation of the triplet at 82.17, note the collapse of the doublet of doublets at 81.82 and the perturbation of the signals at 84.05, (C) Spectrum of irradiation at 84.05, note the collapse of the triplet at 82.17 and the doublet of doublets at 81.82, reproduced with permission from Dr. Rawle I. Hollingsworth (4). 98 1H-NMR spectrum showed the protons of a methylene group which appeared only as a doublet of doublets at 8 2.1— 2.25, indicating that they were due to a deoxy group with no neighboring hydrogens and probably, next to a carbonyl group (Figure 4.7). The larger splitting (J = 14.5 Hz) was from geminal proton, and the smaller splitting (J = 1.9 Hz) was due to the long-range coupling from a equatorial proton at the 5-position; therefore, this component should have the galacto-configuration. Jail V 4 T—Tj11l11j111jj jljjfijfi—Illfifijllljljl PM?! 5 4 3‘ 2 Figure 4.6: Proton NMR spectrum and structure of tetrasaccharide from Rhizobium trifolii AN U843, reproduced with permission from Dr. Rawle I. Hollingsworth (5). 1‘fi‘1fi ifiufi 1 99 "DP 4 3 2 t p of:- Figure 4.7: Proton NMR spectrum of an unusual component in Rhizobium meliloti 41 100 The components from Rhizobium meliloti 41 were reduced with 10% aqueous borohydride, which allows the terminal ring to remain in the linear form. The permethylation was effected with 1% methanolic HCl, followed by reduction with borodeuteride. In this stage, the carboxyl group was reduced to CD20H. The hydrolysis was achieved by 1 M trifluoroacetic acid, which cleaved the oligosaccharides to carbohydrate monomer. The peracetylation was performed by pyridine/acetic anhydride (1:1), or pyridine/acetic anhydride-d6 (1:1) (Figure 4.8). Finally the undeuterated and deuterated components were analyzed by GC-MS. From GC/MS analysis of alditol acetates of these oligosaccharides, there were four late-eluting components with similar spectra. The retention times of these components were approximately 40 minutes, higher than that of the hexoses such as glucose (Figure 4.9). The presence of four components with similar spectra is consistent with a diketo sugar, since in the diketo sugar there are two chiral centers, and therefore four possible isomers. In the electron compact (EI) mass spectrum (Figure 4.10), there is a strong cluster at m /z 240. This corresponded to a structure such as the alditol acetate of 3-deoxy-erythro-2,4-heptodiulosonic acid (Figure 4.11). The cluster at m/z 240 corresponds to the starting material with loss of three acetic acid and one formaldehyde molecules. After peracetylation by acetic anhydride-d6 (Figure 4.12), the cluster at m/z 240 shifted to m/z 247. This indicated the presence of three deuterated acetyl groups (Figure 4.13). 101 7C0 H 0“ 7C0 H OH H 0 H H H OH H H COOH NABH HH H COOH OH 0 OH HO H 20Ac l) M0011 szOH H+ A 2) N aBD. HO 2 H2 7C0 H OH H Ac H Py/A H H H 01:1 HCOH HOOAc | HO H OH HCOH H CHZOAc | plus isomer CH2OH m/z 450 plus isomer PylAOzO‘Ds 9 20CCD3 CD3 H2 13 0 IHOCCD, 0 $110001), 0 $110001), 0 01120001), plus isomer Figure 4.8: Deuterium-labeled, carboxyl-reduced component GOD’EDCUI} OCM-fl’O—OZJ 102 RIT M 044547 .1PLJ ‘LZPA‘ ‘Aw. ‘ ‘A4P‘JJJSFA‘ .Jsp‘ “ S. .bundo g TIC II U 3 ._,. LWA - ,x1.8 72.3419 '1 115 0 c e l 4*! ~ m A *r e :33 4.7469 94 ‘ ‘»x22.7 8.6798 145 a - L L 31.8 15.464? 217 1 h H” . .. ”H'HH'HTmAaAsaa a ' who 1500 me 2539 3m 35m 5“,, Figure 4.9: CO trace of 3-deoxy-erythro—heptodiulosonic acid derivatives 1% 2140 J . 1 1- , sa« __ , 1 - _ j 6 : l 1 7 1 Z . ~15: l 5 r . I . l 176 J‘ - - D 5 .II I Wu H. INN I I. I Hi 4.- _ SB 108 150 200 258 300 100 '58- . l 4 8 . 8L 2“ ‘fi ‘ *1 ‘u;‘ r* a4. , . . . . i i v I ‘ v ‘ . I . v 4 ' 350 400 458 500 550 6% W2 Figure 4.10: EI mass spectrum of 3-deoxy-erythro—2,4-heptodiulosonic acid derivatives aoauuacvp OCOAflfifi-‘QJO 103 3; L2 I HOCH HCOH cnon 3-deoxy-threo-2,4- heptodiulosonic acic Figure 4.11: The structure of 3-deoxy-threo-2,4-heptodiulosonic acid. p4.»- . on ”U 58' 4'." '- 4 a . 1 6 1 ' 1 c r 1.4: 1 1 6 I 4. 5 . . H .- ._ I , ls . A . . i _ 58 188 158 288 258 388 188 i 58- .. "T rss'aj ' ' '4éa' ' ' '45'9' 'j '5é91 ' ' rssa' 'saa "/2 Figure 4.12: EI mass fragmentation of 3-deoxy-erythro-2,4oheptodiulosonic acid deuterated derivatives 104 20A0 + T 2 "' + 2 + OAC AC OAC 2 - A H - 0 AC 3 CO | ago + | CHOAC fin in CHOAC (IIOAC EOAC EHZOAE _CH20AC_ OCH3 m/z450 m/zz7o 35/2246 _ ogco: + 0 + _ l — — + 2 0 206033 1 OCCD, 2 0 -3CD COOH -CD 0 o OCCDJ 3 " OCD3 ——3_’ £OCCQ O n l l (’gcn’ f“ “*0 (mod 0 ‘ | 331% coca), I'm 9206(11334 _ll:l‘[2 _ h— 2 _ m/z468 m/2279 m/2247 Figure 4.13: EI mass fragmentation of 3-deoxy-heptitol derivatives a- < mew» am 11 flaODQQZCU’I’ 105 133-“--m -. .. -.-.-... -..- - ...... - . 3:1 . 4 . ‘ t 5% v 1 1 1 ‘ ; ‘ 97" Ii Eng 2 9 a 1 3 9 ' ‘ c 343 gL‘i—H'a- Jill-J [Amt i‘iilsjgh “WWI UL—i—VILLJ-W WA...» “AW-L 58 ' 188 158 288 258 388 358 488 188‘ --- .. ‘ fl . I 50‘ I r- . 449 G’f‘H—L‘ “I“ ‘H‘ L“ -i”'4WWL‘J-fiwtfl°%h‘W--W v v I v r' . 488 493 - see 553 EEG 653 was -7ng . . wz Figure 4.14: CI mass spectrum of peracetylated 3-deoxy-erythro—2,4- heptodiulosonic acid In the chemical ionization (CI) mass spectrum (Figure 4.14), there is a strong molecular cluster at m/z 361 (m/z + H+). This molecular cluster corresponds to the acetylated compound minus acetic acid and formaldehyde. When acetic anhydride-d5 was used, the CI mass spectrum exhibits a strong cluster at m/z 373 (Figure 4.15), showing four deuterated acetyl groups different from the undeuterated one (Figure 4.16). m 4 men» am :0 fl'ODQCLDCU'I) 8 58 188* 58‘ 3.1914. - 88 64 ‘7: Wflu‘ 106 3.73 r L 184 245 97 1 599% 3 '5 are ’ 114.111 Aggie—1m iJuwa—Jt—JL“ LAM-ti 13:3 158 1‘256 333 353 436 L \ L ”4152 3:12 534 57 ' w _ M: 1+ 3‘, w». .4, we .h. . TA. . - - , «11:3 55:3 sea ass “ma . 7511 ”/2 Figure 4.15: CI mass fragmentation of 3-deoxy-erythro-2,4-heptodiulosonic acid deuterated derivatives all L3.JM.1' um .‘I' F DZOAE + P 2 HOAC AC H2 2 8HOAC 'ACOH , , OAC CHOAC CHOAC CHOAC CHOAC _CH20A_C_ _CHzOAg_ m/z 450 m/z 390 9 _* — —+ ZOCCD, 2 o 0 060133 2 2 08GB; 03300011 > 08033 $1108.11. CH0§CD3 £11080), CHOCéID, (ljfizocgficpfl _8H20CC123_ m/z 468 m/z 405 107 -CH20 2 AC 2 OAC £1320 CHOAC EZTC 3 m/z360 2 l}: o 060). | o (:HoCCD3 0 811060); _ D3 m/z 373 d Figure 4.16: Deuterated and undeuterated CI mass fiagmentation of 3-deoxy-heptitol derivatives 108 4.4 Synthesis of 3-Deoxy-Heptitol Derivative 4.4.1 The First Pathway to Synthesize 3-deoxy-heptitol Derivative To further understand the specific genetic character of Rhizobium meliloti 41 (11) and identify the unusual component, a series of syntheses were performed. We started from 2-deoxy-D-glucose (without consideration of the stereochemistry at the four position). The aldehyde group was protected as a 1,3-propylene dithioacetal (12); the structure of the protected product was confirmed by NMR spectroscopy, and peaks were assigned by proton decoupling. A11 hydroxyl groups were then protected by benzyl bromide, and the anomeric proton was replaced by an iodomethyl group. The iodide group was then transformed to a hydroxyl group. The 1,3-dithiane was cleaved by mercury chloride (13, 14). It can also be cleaved with boron trifluoride (15) and N-bromosuccinimide (16). The reduction of the keto group at the 2 position was achieved by sodium borodeuteride. The benzyl ethers were cleaved by catalytic hydrogenation, and the resulting product was peracetylated (Figure 4.17). The final product was purified by TLC, then analyzed by GC and GC-MS. One component had a similar retention time to the component separated from bacterial LPS, but the resulting EI mass fragmentation from GC-MS (Figure 4.18) was not similar to the spectrum obtained from the bacterial component (Figure 4.1). The CI mass spectrum from the synthesized component had a very strong peak at m/z 419, which corresponds to the acetylated compound with loss of formaldehyde (Figure 4.19). These mass spectrometry data from the synthesized compound indicate that the unusual component we identified in the oligosaccharide from the LPS of Rhizobium meliloti 41 possibly is a 3-deoxy—2,4- heptodiulosonic acid. 109 V “0 0 SH SH HO H OH HO] -5° C Figure 4.17: The first synthetic pathway for 3-deoxy-heptitol derivatives. 110 or?l ~‘H ‘\ HgClleacos H? ’ Bz H‘leo Bz (5.) NhBD4 EtOH 2'91, Bl H 032 820 H (1) plus isomer (fl) plus isomer acetic anhydride pyn'dine (8) plus isomer (continued) Figure 4.17: The first synthetic pathway for 3-deoxy-heptitol derivatives. OODWQDCO'J) 0C~“¢*W~'WPO 111 1 33-1 1 14 R h C] . .- a so- , . 1 . 1 :5 so: :41 ~ g 1T6 b . u 3‘ 1319 _ 2 i . 96 * a ‘ 1 9 i ‘ n < 191 ' c 54 1 5 . c 26“ 7 1 9 . a ‘ 3 ”am. 1 237. fl L 41* a, j 1 IA 9: .. -- .1 I 1 n 1 . . .. ' '3 M 3213 we sec 1.90 288 1112 Figure 4.18 EI mass 8 ctrum from synthetic 3-deoxy-heptitol derivatives Pe 1331-..--. ‘ ...._ .. M.-...._..-___.-_.....--.... .-.._-__.-...._-._. .__.--._.._._.T. I . saJ. . ‘- 1 1 2 s 82§1 57 7 E1394 9 1 a “'8 13.51:? 3:1 . 21%3‘2? 42 I 2731’ .L. .1... .111 In 11 11.111 11111111113344--.m1.-..1i2 1111...} 1-.. 18@5121 159 280 258 “ .__..._-- .- 4I19 _ T 4 I. 1 .L . 'r 4 2 3 318 3 4 3 -4 L 1238 I 91"ij: {I . 378379'l319I3jé81811! 4g? 41: 473 i. will”... 11111..-. 1 -ll ,..li--.i,l_.l i. ‘J 1 1L . 1J3 [i 1--.-.- 1_-_,-....-.-.,-l.--.,-_...,._._....,.L aura 351a . 453 smooz / Figure 4.19 CI mass spectrum from synthetic 3-deoxy-heptitol derivatives 112 4.4.2 Second Pathway to Synthesize 3-Deoxy-Heptitol Derivatives: A second pathway for synthesis of 3-deoxy-heptitol derivatives is described in Figure 4.20. The general strategy is similar to the first pathway, except that the protecting groups were changed from benzyl to iSOpropylidene (12). This was intended to make the system more easy to work up. DMF/NaOH (17), AgN03/acetonelwater (18), and tetrabutylammonium hydroxide (19) were tried for the conversion of compound 10, to 11, but the yield was very low. H8O] {C3001 01': .\\ : .s‘“ H’IHZO , OH plus isomer Figure 4.20: The second pathway to synthesize 3-deoxy heptitol derivatives. 113 4.4.3 The Pathway to Synthesize 3-deoxy-D-erythro-2,4- heptodiulosonic Acid The fourth strategy to synthesize the target compound is shown in Figure 4.22. In this pathway, some of the reaction failed and are reported here to avoid future workers making same mistakes. First, oxidization of the methyl-4,6-0-benzylidene-B—D-galactopyranoside (20) was attempted to make D-erythrose, with the periodic acid/sodium bicarbonate, and lead tetraacetate (21, 22) tried as the oxidation reagent. Methyl-4,6-0-benzylidene-B-D- galactopyranoside is not very water-soluble, so it could not be mixed very well with periodic acid/sodium bicarbonate. The conversion of oxidation was very low when using lead tetraacetate. To yield the solubility of the starting material, 4,6-0-benzylidene-D-g1ucose was synthesized. The 4,6-0- benzylidene-D-glucose was oxidized by sodium meta-periodate to the desired 2,4-0-benzylidene-D-erythrose (23), which existed as a dimer. When protection of the aldehyde group in 2,4-0-benzylidene-D-erythrose was attempted with 1,3-propane dithiol, the benzylidene group was cleaved under the acidic condition. To avoid deprotection, a series of different acid conditions was tried. This included HCl etherate solution (24), boron triflouride etherate solution (25), dilute trifluoroacetic acid, and zinc chloride with boron trifluoride. The benzylidene group was always cleaved to form benzaldehyde propylene dithioacetal, no desired product was obtained. 4,6-0-propylidene-D-glucose was synthesized to reduce the lability of the protective acetal group. At low temperature, 1 M HCl etherate solution was used as the catalysts. The 1,3-propyl dithioacetal was formed without disrupting the protective acetal group. 114 H OH HO CHacHZCH : OH _ HO H’v 40°C H s\ s K (15) SH SH OH H_ < H——OH 090 HO] H" o Pyridine Acetic H H anhydrous s s s s \ \ H—K (15) (1) nBuLi H__ CHZCOCOOEt H—"°A° (2) BrCHzcocoosY H— -0Ac H— H H_ H H =0 1)H* NBS 2 :0 00011 a (fuzcocooa HCOH chos 0=0 l aquOH H OH HCOH H OAc 3-deoxy-P-threo—2, . 4-heptod1ulosomc ac1d CH 20H CHZOH Figure 4.21: The pathway to synthesize 3-deoxy-D-erythro-2,4- heptodiulosonic acid 115 An acetyl group was attached to the 3-hydroxyl group in the resulting product at the C3 position. In the alkylation stage, the acetyl group was destroyed by the n-butyl lithium; the propylidene group also underwent elimination. 4.5 References 1. Dazzo, F. B. and Hollingsworth, R. 1., Biol. Cell (1984) 51 267-274. 2. Christian R. H. Raetz, Ann. Rev. Biochem. (1990) 52 129-170. 3. Westphal, 0., Luderitz, 0., Galanos, Ch., Mayer, H. and Rietschel, E. T., “Advance in Immunopharomaco” (1985) 3 13-43. 4. Carlson, R., W., Hollingsworth, R. I. and Dazzo, F. B., Carbohydr. Res. (1988) 115 127-135. 5. Hollingsworth, R. I., Carlson, R W., Garci, F., and Gage, D. A., J. Biol. Chem. (1989) 2% 9294-9299. 6. Prehm, P., Strim, S., Jann, B., and Jann, K., Eur. J. Biochem. (1975) 55 41-55. 7. Carlson, R. W., J. Bacteriol. (1984) 15.8 1012-1017. 8. Law, I. J ., and Strijdom, B. W. J. Biol. Chem. (1977) 2 79-84. 9. Smit, G., Kijne, J. W., and Lugtenberg, B. J., J. Bacteriol. (1937) 162 4294-4301. 10. Noel, K. D., VandenBosch, K. A., and Kulpaca, B., J. Bacteriol. (1986) 168 1392-1401. 11. Williams, M. N. V., Hollingsworth, R. I., Klein, 8., and Signer, E. R., J. Bacteriol. (1990) 112, 2622-2623. 12. Frost J. W. and Knowles, J. R., Biochemistry (1984) 23 4465-4469. 13. Corey, E. J ., and Bock, M. G., Tet. Lett. (1975) 33 3269-3270. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 116 Meyers, A. I., Comins, D. L., Henning, D. M. R., Shimizu, K., J. Am. Chem. Soc. (1979) 101 7104-7105. Corey, E. J ., and Hase T., Tet. Lett. (1979) 4 335-338. Corey, E. J ., and Erickson, B. W., J. Org. Chem. (1971) 3.6 3553-3560. Numazawa, M., and Nagaaoka, M., J. Org. Chem. (1982) 41 4024-4029. Wasserman, H. H., Mariano, P. S., and Keehn, P. M., J. Org. Chem. (1971) 35 1765-1769. McCourt, D. W., Roller, P. P., and Gelboin, H. V., J. Org. Chem. (1981) 45 4157-4161. Bofli, C., and Ferrari, L., Die Starke (1969) 2.1 100. Perlin, A. S., and Brice, C., Can. J. Chem. (1956) 3.4 541-553. Murry, J. E. Mc, adn Blaszczak, L. C., J. Org. Chem. (1974) .32 2217- 2222. Baggett, N ., Buck, K. W., Foster, A. B., Rees, B. H., and Webber, J. M., J. Chem. Soc. (C). (1966) 212-215. Schmidt, U., and Ohler, E., Angew. Chem. Int. Ed. Enl. (1977) 1.6 327. Seebach, D., Jones, N. R., and Cory, E. J., J. Org. Chem. (1968) 3.3 300-305. CHAPTER5 EXPERIMENTAL FORCHAPTER4 5. 1 General Methods in Structure Elucidation 5.1.1 Microorganisms The lipopolysaccharides were obtained from Rhizobium meliloti 41. 5.1.2 Dimethylsulphinyl Sodium Preparation (1) Dry DMSO 30 mL was added from a dry syringe through a serum cap to a three-neck 250 mL flask equipped with a calcium chloride drying tube and containing 1.5 g of sodium hydride (washed twice with 20 mL anhydrous ether). The system was protected under nitrogen. The reaction mixture was stirred at 50° C for 3 hours. After a transparent green color was formed, the reaction was judged to be complete, and was transferred to several one dram vials, flushed with nitrogen, sealed by a Teflon-lined screw-cap, and stored frozen in the dark. 5. 1.3 Hydrolysis of Lipopolysaccharides (2) Acetic acid (1 mL of 1%) was added to lipopolysaccharide (~ 2 mg). and the solution heated at 100° C for 1 hour. Ten mL water/chloroform (1:1) was added to the sample, and the mixture shaken well. The layers were separated and the extraction was repeated three times adding chloroform (5 mL) each time. The chloroform laYers (containing lipid A) were combined together and reduced to dryness. The carbohydrate was left in the water layer which was concentrated and Passed through a Bio-Gel P-2 column (200-400 mesh,16 mm x 90 mm), with 117 118 1% formic acid as eluant. Fractions of 90 drops (2 mL) were collected in each test tube by a fraction collector. 5.1.4 Colorimetric Methods for Carbohydrate Detection 1. Phenol/Sulfuric acid Assay (total carbohydrate) (3): After separation on the Bio-gel P-2 column, 75 uL of sample was taken from each fraction, and 0.5 mL 5% aqueous phenol and 2.5 mL sulfuric acid were added. The absorbance at 490 nm was measured on a Milton Royspectrometer using a tungsten-visible lamp. The absorbance for each fraction was plotted versus fraction number. The samples containing carbohydrate were combined and then lyophilized for instrumental analysis. 2. KDO Assay (4): (1) From each column fraction, 50 uL of sample was transferred to a test tube. (2) 50 11L of 0.5 N H2804 was added to each test tube, and shaken well. (3) All samples were placed in a boiling water bath for 8 minutes, then cooled to room temperature. (4) 50 1.1L H5106 (reagent A) was added to each sample, and shaken well. (5) All samples were left for 10 minutes at room temperature. (6) 200 uL arisenite (reagent B) was added to each sample, and the tubes shaken well. (7) 800 uL thiobarbituric acid (reagent C) was added to each sample. (8) All samples were placed in a boiling water bath for 10 minutes. (9) All samples were cooled to room temperature under running tap water. 119 (10) 2.0 mL butanol (reagent D) was added to each sample. After addition of each reagent, all samples were well mixed. (11) The butanol layer was recovered and its absorbance measured at 552 nm and plotted versus fraction numbers. (12) All samples were combined based on the distribution of carbohydrate, then frozen and lyophilized for NMR. Reagent A: 2.28 g H5106 (O-paraperiodic acid) was dissolved to 100 mL water (0.1N), and stored in a dark bottle. Reagent B: 4 g of NaAst was dissolved in 100 mL of 0.5 N HCl (4.0%). ' Reagent C: 600 mg of 2-thiobarbituric acid dissolved in 100 mL boiling water (0.6%), then cooled to room temperature. This reagent must be prepared prior to use. Reagent D: Five mL concentrated HCl was added to 95 mL n-butanol. 5.1.5 CompositionalAnalysis (2) The reducing terminus was pre-reduced by sodium borohydride: oligosaccharide (0.2 mg) was dissolved in 0.1 mL water, then 0.2 mg of sodium borohydride was added. The mixture was allowed to stand at room temperature for 3 hours, and the excess borohydride was decomposed by adding 10% acetic acid dropwise. The solution was concentrated to dryness under a stream of nitrogen. The carboxyl groups were methylated and then reduced with sodium borodeuteride. In a Teflon-lined, screw-capped vial, the above sample was dissolved in 1 mL of 1% HCl in methanol. The solution was allowed to stand at 60° C for 3 hours and evaporated to dryness. One mL of methanol was added, and the solution concentrated to dryness to remove trace 12) amounts of acid. The residue was dissolved in 0.1 mL water, then 0.2 mg of sodium borodeuteride (98% D; obtained from Aldrich Chemical Company, Inc.) was added. The mixture was allowed to stand at room temperature for 16 hours, and the excess borodeuteride was destroyed by drop-wise addition of 10% acetic acid until decomposition was completed. The solution was then evaporated to dryness. The above sample was hydrolyzed by treatment with 1 mL of 1 M TFA at 100° C for 3 hours. The hydrolysate was concentrated to dryness, then 1 mL water [was added and the solution concentrated to dryness to remove trace amounts of TFA. The aldehyde groups of the monomers were reduced by treatment with 0.2 mg sodium borohydride in 0.1 mL water for 1 hour. The solution was then evaporated to dryness. One mL of methanol was added to the sample, then the evaporation step was repeated. This process of addition of methanol and evaporation was repeated 4 times to remove trace amounts of sodium borohydride and boric acid as the volatile trimethyl ester. ' The hydroxyl group of hydrolyzed and reduced oligOsaccharides was acetylated by acetic anhydride and pyridine (100 11L of a 1:1 mixture). This was accomplished by sonicating for 1 minute and then heating at 100° C for 15 minutes, followed by sonication and heating at 1000 C for an additional 15 minutes. The mixture was evaporated to dryness. Finally, the sample was dissolved in a mixture of 1 mL 1% hydrochloric acid and 1 mL of chloroform to extract the acetylated sugars. This extraction was achieved. twice and the chloroform layer was collected and reduced to dryness for GO and GC-MS analysis. 121 5.1.6 1B and 13C-NMR Analysis For 1H and l3C-nuclear magnetic resonance (NMR) analysis of oligosaccharides: samples (1 mg) were dissolved in 0.7 mL of deuterium oxide (99.96% D, obtained from Cambridge Isotope Laboratories). The spectra were recorded on Varian VXR 300 MHz and Varian VXR 500 MHz (for 1H) instMents. The chemical shifts were measured relative to the water peak (4.65 ppm at room temperature). 5.1.7 GC Analysis The GC analyses were performed on a Varian 5800 gas chromatograph equipped with a flame ionization detector and a Hewlett- packard 3390A digital integrator. All GC analyses were done on DB 225 Fused Silica Capillary Columns (30 m-length, 0.32 mm diameter, film thickness 0.25 mm, from J & W Scientific ). 5.1.8 GC-MS Analysis GC-MS analyses were carried out on: 1) JEOL JMS-AX505H mass spectrometer interfaced with a Hewlett-Packard 5890A Gas Chromatograph or 2) Hewlett-packard 5970 series mass-selective detector interfaced with a 5890A Gas Chromatograph. Data was analyzed with a Hewlett-packard 59970 MS ChemStation. 5.1.9 OtherFacilities UV spectra were obtained with a Varian DMS 200 UV-Visible Spectrophotometer. Gel filtration column chromatography was also followed by UV absorbance using an ISCO V4 Absorbance Detector. 122 An HBI Multistaltic Pump was used in column chromatography. A GILSON Micro Fractionator was used to collect samples from the columns during gel filtration chromatography. A spectronic 21 made by the MILTON ROY Company was used for measuring the UV absorbance in tubes (colorimetric assays). UV absorbance was viewed on TLC (thin layer chromatography) using the UVP. Inc. view box. 5.2 Synthetic'h'ansformation 5.2.1 2-Deoxy-D-Arabino-Hexose Propylene Dithioacetal in Figure 4.20 compound 1 (5): 2—deoxy-D-glucose, 5.03 g (0.031 mole), dissolved in 6 mL cold (0°C) hydrochloric acid (37%) and 3.3 mL (0.033 mole) propylene 1,3-dithiol were added to a 50 mL round-bottom flask. The mixture was stirred in an ice- salt bath at -5° C for 20 minutes. Ice-water was added to the solution, all solvent was removed by a rotary evaporator, and the product was obtained as a white solid. The solid was recrystallized from methanol/acetone (1:10), and filtered. The residue was washed with hexane. 1H-NMR (500 MHz, methanol-d4 ): 81.78 (m, 1 H,w/48 Hz, H-8), 81.96 (dd, 1 H, J = 9.9 and 3.9 Hz, H-2’), 81.98 (dd, 1 H, J = 8.9 and 3.9 Hz, H-2), 82.05 (m, 1 H, w/34.9 Hz, H-8’), 8 2.78 (dt, 2 H, J = 2 and 4 Hz, H—7’ x 2), 8 2.85 (ddd, 2 H, J = 11.5, J = 29.9 and 2 Hz, H-7 x 2), 83.27 (dd, 1 H, J = 1 and 9.9 Hz, H-4), 83.55 (dd, 1 H J = 6.5 and 11.5 Hz, H-6), 83.6 (ddd, 1 H, J = 6.5, 7 .9 and 2.9 Hz, H-5), 83.74 (dd, 1 H, J = 5.5 and 2.9 Hz, H-s’), 54.13 (ddd, 1 H, J = 9.9, 3.9 and 1.9 Hz, H-3), 54.22 (dd, 1 H, J = 9.9 and 4.9 Hz, H-l). 123 5.2.2 2-Deoxy-3,4,5,&Tetrabenzyl-D-Arabino-Hexose Propylene Dithioacetal in Figure 4.17 compound 2(6): The 0.25 g (9.8 x 10’4 mole), 2-Deoxy-D-arabino-hexose propylene dithioacetal compound 1, 2.5 g barium oxide (0.0163 mole), and 15 mL dimethyl sulfoxide were placed in a 50 mL round bottom flask, then 1.8 mL ( 0.015 mole) benzyl bromide was added dropwise. The mixture was stirred at room temperature for 48 hours, then centrifuged, and the residue washed with chloroform. The supernatant and the chloroform extracts were combined. The solution was washed with a 15% aqueous solution of sodium thiosulfate, then washed with water, dried over sodium sulfate, and evaporated to dryness . 5.2.3 8-Deoxy-l-Iodo-4,5,6,7-Tetrabenzyl-D-Arabino-Heptose Propylene Dithioacetal in Figure 4.17 compound 3 (7): A solution of 3 x 10‘4 mole 2-deoxy-3,4,5,6-tetrabenzy1-D-arabino- hexose propylene dithioacetal compound 2 and 10 mL THF (distilled over lithium aluminum hydride) was placed in a 50 mL three-necked round- bottom flask. The system was protected under nitrogen and the mixture stirred in a carbon tetrachloride/ dry ice bath at -30° C. One mL 2.5 M n- butyllithium in hexane was injected at a rate of 3 mL/min. The mixture was stirred for 1.5 hours, and 0.1 mL (~ 0.012 mole) diiodomethane was injected. The mixture was protected from light with aluminum foil, stirred at -5° C (salted ice) for 18 hours, then quenched with two volumes of water, acidified with HCl (1 M) to pH 5-6, and extracted with 3 portions (10 mL each) of ethyl acetate. After washing with 3% aqueous sodium bisulfite and water, the sample was dried with sodium bisulfite. 124 5.2.4 3-Deoxy-4,5,6,7-Tetrabenzyl-D-Arabino-Heptulose Propylene Dithioacetal in Figure 4.17 compound 4 (8): I Approximately 3 x 10" mole 3-deoxy-1-iodo-4,5,6,7-tetrabenzyl-D- arabino-heptose propylene dithioacetal compound 3, and 10 mL THF were placed in a 50 mL round-bottom flask, then 1 mL 1 M tetrabutyl ammonium hydroxide was added. The mixture was allowed to stay at room temperature for 20 hours, then 7 mL 1% HCl in methanol was added. The mixture was washed with 5% sodium bicarbonate, then extracted with ethyl acetate and dried by passage over absorbant cotton. 5.2.5 3-Deoxy-4,5,6,7-Tetmbenzyl-D-Arabino-Heptulose in Figure 4.17 compound 5 (9, 10): Approximately 2.5 x 10'4 mole 3-Deoxy-4,5,6,7-tetrabenzyl-D-arabino- heptulose propylene dithioacetal compound 4 and 10 mL acetonitrile/water (4/1) were placed in a 50 mL round bottom flask, and 0.07 g calcium carbonate (2.5 eq.), 0.17 g mercuric chloride (2.3 sq.) were added. The mixture was stirred at room temperature for 4 hours, filtered through celite, washed with ammonium acetate, and the aqueous layer extracted with ethyl acetate. The organic layers were combined and washed with odium bicarbonate, filtered with cotton and potassium carbonate, and evaporated under nitrogen to dryness. 5.2.6 3-Deoxy-2-Deutero-4,5,6,7-Tetrabenzyl-D-Gluco(1VIanno)-Heptitol in Figure 4.17 compound 5: _ Approximately 2.5 x 10" mole 3-deoxy-4,5,6,7-tetrabenzyl-D-arabino- heptulose compound 5, was treated with 0.2 g sodium borodeuteride in 125 ethanol. The mixture was stirred overnight, quenched with 0.1N HCl, extracted with ethyl acetate, and dried by passage over cotton. 5.2.7 3-Deoxy-2-Deutem-D-GlucoManno)-Heptitol in Figure 4.17 compound 1 (11): Approximately 2.0 x 10“ mole 3-deoxy-2-deutero-4,5,6,7-tetrabenzyl-D- gluco(manno)-heptitol compound 6. 0.05 g 10% palladium chloride on active charcoal, 5 mL cyclohexene, and 10 mL ethanol were placed in a 50 mL round-bottom flask which was stirred and heated at 7 0° C overnight. After cooling to room temperature, the palladium catalyst was filtered off and the filtrate washed with 0.1 M ammonium acetate and water, and then concentrated to dryness on the rotary evaporator. 5.2.3 3-Deoxy-2-Deuteno-l,2,4,5,6,7-Hexacetyl-D-Gluco(Manno)- Heptitol in Figure 4.17 compound 8 (12): . Approximately 1.5 x 10" mole 3-deoxy-2-deutero-D-gluco(manno)- heptitol compound 1, 2 mL pyridine, and 3 mL acetic anhydride were placed in a 25 mL round bottom flask. The mixture was stirred and heated at 80° C for 3 hours. After cooling to room temperature, 5 mL 1 M HCl was added. The mixture was extracted with chloroform, washed with sodium bicarbonate, and dried with sodium sulfate. 5.2.9 2-Deoxy-3,4 : 5,6-Di-0-Isopropylidene-D-Arabino-Hexose Propylene Dithioacetal in Figure 4.20 compound 9 (5): The 2-Deoxy-D-arabino-hexose propylene dithioacetal compound 1 0.5 g, (0.002 mole) was dissolved in acetone (30 mL) in a 100 mL round-bottom flask, and concentrated sulfuric acid (2 drops) was added. The mixture 1% was allowed to stand at 50° C for 12 hours, then concentrated to dryness. The yield was 100% of crude product. After recrystallization from methanol, the yield of final product was 81%. NIP: 103-104° C, 1H-NMR (500 MHz, CDC13): 81.30 (s, 3 H, C-CH3), 81.34 (s, 3 H, C-CH3), 81.36 (s, 3 H, C- CH3), 81.40 (s, 3 H, C-CH3), 81.86 (m, 1 H, w/44.9 Hz, H-8), 81.94 (ddd, 1 H, J = 15, 9.9 and 4.9 Hz, H-2’), 82.08 (m, 1 H, w/34.9 Hz, H-8’), 82.17 (ddd, 1 H, J = 15, 9.9 and 4.9 Hz, H-2), 82.83 (m, 2 H, w/25 Hz, H-7’ x 2), 82.90 (m, 2 H, w/ 40Hz, H-7 x 2), 83.52 (t, 1 H, J = 15 Hz, H-4), 83.89 (dd, 1 H, J = 5 and 9.9 Hz, H-6), 84.0 (dt, 1 H, J = 9.9, and 4.9 Hz, H-5), 84.09 (dd, 1 H, J = 5 and 9.9 Hz, H-6’), 84.19 (ddd, l H, J = 9.9, 3.9 and 4.9 Hz, H-3), 84.24 (dd, 1 H, J = 9.9 and 4.9 Hz, H-l). 13C-NMR (300 MHZ,CD013)I 8 25.3, 26.0, 26.4, 26.6, 26.8, 29.8, 30.2, 39.8, 42.0, 66.8, 76.7, 76.9, 81.3, 109.3, 109.6. IR: (cm-1): 2800-3000 (OH), 1400 (com), 1200, 1100 (CO), 700 (03). 5.2.10 3-Deoxy-l-Iodo-4,5: 6,7-Di-O-Isopropylidene-D-Arabino—Heptose Propylene Dithioacetal in Figure 4.20 compound 10 (7 ): Approximately 3 x 10" mole 2-deoxy-3, 4: 5, 6-di-O-isopropylidene-D- arabino-hexose propylene dithioacetal compound 9_ and 10 mL THF (distilled over lithium alumina hydride) was placed in a 50 mL three-neck round-bottom flask and a nitrogen inlet attached. 0118 mL 2.5 M n- butyllithium in hexane was injected at -30° C, at a rate of 3 mL/min. The mixture was stirred for 2 hours, before 0.1 mL (~ 0.012 mole) diiodomethane was injected. The mixture was protected from light by aluminum foil and stirred at 0° C for 8 hours. The reaction was then quenched by two volumes of water, acidified with HCl (1 M) to pH 5-6 and extracted with 3 portions (10 mL each) of ethyl acetate. The organic layer was washed with 3% aquedus sodium bisulfite and water, and then dried with sodium bisulfite. Product 127 yield was 41%. 1H- NMR (300 MHz, CD013): 81.30 (s, 3 H, C-CH3), 81.34 (s, 3 H, C-CH3), 51.36 (s, 3 H, C-CH3), 51.40 (s, 3 H, C-CH3), 52.13 (quintet, 2 H, J = 9 Hz, H-8, 8’), 82.39 (dd, 1 H, J = 12 and 18 Hz, H-2’), 82.52 (dd, 1 H, J = 15 and 3 Hz, H-2), 83.38 (m, 4 H, w/36 Hz, H-7’ x 2, H-7 1: 2), 83.6 (t. l H, J = 15 Hz, H-4), 83.92 (dd, 1 H, J = 5 and 9.9 Hz, H-6), 84.0-84.12 (m, 3 H, w/36 Hz, H-6’, H3, H5), 55.9 (s, 1 H, H-1,CH21). 13C NMR (300 MHz, CDC13): 525.3, 26.8, 27.2 27.6, 31.2, 31.8, 42.1, 67.2, 76.9, 78.8, 80.6, 109.1, 109.7, 110.8, 120.6, 130.24. IR: (CM‘l): 3000 (C—H), 2400 (C—H, alkane monosubstitute), 1400 (C- CH3), 1500 (C-H, methylene), 1200, 1000 (GO, C-I), 700 (C-S). MS: (JEOL JMS-AX505H): m/z 473 (M+), m/z 346 (lost iodide), m/z 331, m/z 143. 5.2.11 4,6-O-Propylidene-D-Glucose in Figure 4.21 compound 13 (14): Approximately 30 g (0.17 mole) D-glucose and 70 mL propionaldehyde was added to a 250 mL round-bottom flask equipped with a condenser and drying tube. Concentrated sulfuric acid 0.4 mL was added dropwise with vigorous stirring, and the mixture was kept at 40° C overnight. After cooling to room temperature, 20 mL ethanol was added. The pH was adjusted 'to 7 by adding saturated sodium carbonate, then an additional 80 mL ethanol was added. The product was obtained as crystals with a 21% yield. 1H-NMR (500 MHz, .CD30D): 80.72 (t, 3 H, CH3), 81.58 (m, 2 H, CH2), 83.22 (t, l H, C-2), 83.46 (m, 1 H, C-3), 83.57 (m, 1 H, C-4), 83.75 (m, 1 H, C- 5), 84.05 (dd, 1 H, C-6), 84.05 (dd, 1 H, C-6), 84.6 (m, 1 H, propylidene, possibly overlap with B-anomeric proton), 85.16 (d, 1 H, a-anomeric proton). 5.2.12 2,4-O-Propylidene-Erythrose in Figure 4.21 compound 14 (15): - . Approximately 1.67 g (0.0076 mole) 4,6-O-propy1idene-D-glucose compound 13 was dissolved in 50 mL water along with sodium bicarbonate, 128 1.7 g (0.02 mole, 2.7 eq.) in a 500 mL round-bottom flask. Subsequently, potassium meta-periodate (3.8 eq., 0.017 mole) previously dissolved in 200 mL water was also added. The reaction mixture was stirred at 40° C for 3 hours, extracted with hot ethyl acetate, and after concentrating to dryness, 0.9 g product was obtained. The yield was 75%. TLC condition was ethyl acetate/hexane #1. 1H-NMR (500 MHz, CD013): 80.82 (t, 6 H, 2 x CH3), 81.55 (m, 4 H, 2 1: CH2), 83.25 (m, 4 H, C-4), 83.55 (m, 2 H, 2 x C-3), 83.95 (dd, 2 H, 2 1: C-2). 84.59 (d, 1 H, C-l), 84.65 ((1, 1 H C-l’). 13C NMR (300 MHz, CD013): 88.8 (2 1 CH3), 828.5 (2 x CH2), 862.8 (C-2), 862.9 (C-2’), 871.6 (C-3), 871.3 (C- 3’), 883.1 (C-4). 883.4 (C-4’), 897.5 (CH propylidene), 897.8 (C’H, propylidene), 8103.8 (C-l), 8104.2 (C-l’). 5.2.13 2,4-0-Propylidene-D-Erythrose Propylene Dithioacetal in Figure 4.21 compound 15 (16): Approximately 0.53 g (0.0033 mole) 2,4-propy1idene-D-erythrose compound 14 dissolved in 50 mL 1,4-dioxane, and 0.36 mL (3.6 mmol, 1.1 eq.) 1,3-propanedithiol were added to a 100 mL round-bottom flask and cooled to -78° C; 0.7 mL (2 eq.) 1 M HCl etherate solution was then added. The mixture was stirred at room temperature for 3 hours, and the solvent was removed via rotary evaporation. The product was purified by silica gel chromatography (ethyl acetate/hexane 1:1). The product yield was 0.74 g (90% yield). 1H-NMR (500 MHz, CD013): 80.94 (t, 3 H, CH3), 81.65 (m, 2 H, CH2), 82.0 (m, 2 H, -SCH2CHzCH2S-), 82.8 (m, 2 H, axial -SCH_2CHzCH_2S-), 83.0 (m, 2 H, equatorial -SCH20H20H28-), 83.4 (dd, 1 H, C-4), 83.7 (dd, 1 H C-2), 83.9 (m, 1 H, C-3), 84.1 (dd, 1 H, C-4’), 84.25 ((1, 1 H, C-l), 84.45 (t, 1H, propylidene). 13C NMR (300 MHz, CD013): 88.8 (CH3), 827 (CH2), 826 129 (SCHZQHzCst-L 829 (-SQHzCHZQH2S-), 846 (C-3), 864 (C-4’), 870 (C-2). 884.5 (propylidene), 8103.4 (C-l). 5.2.14 2,4-0-Propylidene-3-O-Acetyl-D-Erythrose Pmpylene Dithioacetal in Figure 4.21 compound 15: Approximately 0.37 g (1.48 mmole) 2,4-0-propy1idene D-erythrose propylene dithioacetal was placed in a one dram vial, then 0.5 mL pyridine, 0.5 mL acetic anhydride was added. The solution was left at room temperature for 12 hours before the solvent was blown to dryness under a stream of nitrogen. The TLC eluant was hexane/ethyl acetate 4/ 1. After purification by silica gel chromatography, the yield was 63%. 1H-NMR (500 MHz, CD013): 80.94 (t, 3 H, CH3, propylidene), 81.68 (m, 2 H, CH2), 82.0 (s, 3 H, CH3, acetyl), 82.6 (m, 1 H, axial -SCH2C 1120st-), 82.7 (m, 1 H, equatorial -SCH2CH2CH2$-), 82.8-2.9 (m, 2 H, axial -SCH.ZCH2CH28-), 83.0- 3.1 (m, 2 H equatorial -SCH20HgCH28-), 83.36 (dd, 1 H, C-4), 83.94 (dd, 1 H C-2), 83.97 (d, 1 H, C-l), 84.28 (dd, 1 H, C-4’), 84.48 (t, 1 H, CH, propylidene), 85.0 (m, 1 H, C-3). 13C NMR (80 MHz, CD013): 88.8 (CH3), 821 (acetyl,CH3), 825.6 0, 827.4 (CH2, propylidene), 828.9 (SCH2QH20H2S), 829. 1 (SQHZCHzflH 28), 844.6 (C-3), 864.5 (C-4’), 867.2 (C-2), 884.4 (CH, propylidene), 8103.4(C-1), 8169.6 (carbonyl). 5. 3 References 1. Chaplin M. F. and Kennedy, J. F., “Carbohydrate Analysis” (1988) 182, Practical Approach Series, IRL Press, Oxford, Washington, DC. 2. Hollingsworth, R. I., Abe, M., Sherwood, J. E. and Dazzo, F. B. J. Bacteriol. (1984 ) 1m 510-516. 10. 11. 12. 13. 14. 15. 16. 17. 130 Dubois, M., Gilles, K. A., Hamilton, J. K., Rebers, P. A., and Smith, F., Anal. Chem. (1956) 2.8 350. Waravdekar, V. S. and Saslaw, L. D., J. Biol. Chem. (1959) 234 1945. Frost J. W. and Knowles, J. R., Biochemistry (1984) 23 4465-4469. Srivastava, H. C., Harshe, S. N ., and Singh, P. P., Tet. Let. (1963) 21 1869-1873. Corey E. J. and Seebach, D., Angew. Chem. Internal. Ed. Enl. (1965) 4 1075-1077. Numazawa, M., and N agaoka, M., Steroids (1982) 32 345-355. Corey, E. J ., and Bock, M. G., Tet. Let. (1975) 38 3269-3270. Meyers, A. I., Comins, D. L., Henning, D. M. R., Shimizu, K., J. ‘ Am. Chem. Soc., (1979) 19.1 7104-7105. Anantharamaiah, G. M., and Sivanandaiah, K. M., J. Chem. Soc. Perkin I (1977) 490-491. Zhdanov, R. I., and Zhenodarova, S. M., Synthesis (1975) 222. Ph.D. thesis of Dr. Hollingsworth, R. I., p. 96. Ph.D. thesis of Dr. Hollingsworth, R. I., p. 164. Bonner, T. G., Boume, E. J ., and Lewis, D., J. Am. Chem. Soc. (1965) 81 7453-7458. . Baggett, N., Buck, K. W., Foster, A. B., Rees, B. H., and Webber, J. M., J. Chem. Soc. (C). (1966) 212-215. Schmidt, U., and Ohler, E., Angew. Chem. Int. Ed. Enl. (1977) 15 327 . CHAPTER6 CHARACTERIZATION OF STRUCTURAL DEFECTS IN THE LIPOPOLYSACCHARIDE or SYMBIOTICALLY IMPAIRED RHIZOBIUM LEGUMHVOSARUM BIOVAR VICIAE vr-39 MUTANTS 6. 1 Abstract The lipopolysaccharides (LPS) of a wild-type strain of Rhizobium leguminosarum biovar viciae (strain VF-39) and two symbiotically defective Tn5 mutants (VF-39-32 and VF-39-86) have been studied. The LPS of the mutants reflect impaired synthesis of the O-antigen. In the LPS of one mutant, the core tetrasaccharide was lacking and in that of the other it was truncated to a disaccharide containing mannose and 3-deoxy-D-manno-oct- 2-ulosonic acid (KDO). The latter mutant also synthesized ‘an unusual carbohydrate component containing mannose, galactose, and an unidentified saccharide. The lipid A composition was similar to that found in other strains of R. leguminosarum biovar viciae. The O-antigen of the wild-type bacterium contained 2-0-methylfucose, fucose, 3,6-dideoxy-3- (methylamino)hexose, glucose, 2-amino-2,6-dideoxyhexose and heptose. This study clearly defines a role for the bacterial LPS in the proper functioning of the Rhizobium legume symbiosis. ' 6. 2 Introduction The Rhizobiaceae are a family of Gram-negative bacteria which infect and form a symbiosis with legume plants with consequent conversion or fixation of molecular nitrogen into ammonia. This process is marked by a high level of specificity between the bacteria and the plant, and cell- 131 132 surface carbohydrates are thought to be involved in this process at the levels of initial recognition (1, 2) and nodule development (3-6). It has been demonstrated (3) that mutants of R. japonicum which appeared to be defective in LPS biosynthesis were incapable of forming nitrogen-fixing nodules in plants. This phenomenon was also demonstrated with R. phaseoli mutants (4) and for R. leguminosarum biovar viciae mutants (5). Whereas these three studies lacked chemical analysis of the LPS structure, this aspect was provided in a study of symbiotically-defective R. phaseoli mutants (6). The work now reported extends the results of this study to the R. leguminosarum system and underscores the importance of the synthesis of an intact LPS by the bacterium for viable symbiosis. The rhizobial LPS have unusual features which include transmembrane fatty acids (7,8), unusual sugars including uronic acids in the lipid A (9, 10), and for some LPS, a complete lack of phosphate in the lipid A (10). There is a large degree of structural diversity from strain to strain among the different species (11), and complete structures are known only for two core oligosaccharide components (12-14), namely a trisaccharide containing KDO and two GalA residues and a tetrasaccharide containing KDO, Gal, Man, and GalA residues. R. leguminosarum biovar viciae (strain VF-39), isolated from nodules of Vicia faba, can develop an efficient symbiosis with host plants such as V. hirsuta and Pisum sativum. Mutants of this strain, isolated by Tn5 mutagenesis, were impaired in their ability to elicit normal development of nodules. They were able to induce the formation of nodules and infection threads, but were affected in the release and/or differentiation of the bacteria (5). SDS-PAGE indicates that the symbiotic defect of these mutants is correlated with a failure to produce a normal LPS. This defect 133 was also found for two mutants (including VF-39-32) by preliminary analysis of the LPS carbohydrate (15). The present study was undertaken in order to determine the modified structures of the LPS of the genetically- different mutants, VF-39-32 and VF-39-86, so as to better understand the origin of the changes and to rationalize their effects on symbiosis events. 6.3 Experimental 6.3. 1 CultumofBacteliaandlsolafionofIPSandIPSF‘ragInents Bacteria were grown in a modified Bergensens medium (as described in 16) in shaken-broth cultures. Mutants were grown (5) in the presence of antibiotics. The LPS were isolated by the hot-phenol method (17) and after treatment with DNAse and RNAse, purified on Sepharose 4B using an ammonium formate buffer (9, 10). Column eluates were analyzed by the phenol-sulphuric acid method (18). Fragments were released from the LPS by hydrolysis for 2 h with aqueous 1% acetic acid at 100° C. The hydrolysate was extracted with 5:1 chloroform-methanol in order to remove lipid A, the aqueous layer was lyophilized, and the residue was chromatographed on a column of Biogel P2 using aqueous 0.1% formic acid. Fractions which appeared impure by 1H-N MR spectroscopy and other analyses were subjected to ion-exchange chromatography on DEAE-Sephadex G25 in aqueous 0.01% formic acid adjusted to pH 6.5 with ammonia, eluting with a linear gradient of 0-0.2 M ammonium chloride over two hundred mL. 6. 3. 2 Compositional Analysis Analyses for fatty acids and carbohydrates in lipid A were carried out as described (7-10) after conversion into methyl esters and alditol acetates, respectively. Deuterium-labeled acetate groups were introduced by 134 acylation with acetic anhydride-d5. Gas chromatography of carbohydrates was carried out on a Hewlett Packard 5890 gas chromatograph using a capillary DB225 column and a flame-ionization detector. Mass spectrometry detection was effected with a JEOL 505 mass spectrometer in the El or Cl (ammonia) mode. 6. 3. 3 Structural Analyses Oligosaccharides were methylated after reduction of carboxyl and other carbonyl functions with sodium borodeuteride (as described in 12). 1H-NMR spectrosc0py (500-MHz) was performed on solutions in D20 with a Varian VXR500 spectrometer. FAB-MS was performed with a JEOL HX110-HF instrument in the negative- or positive-ion mode and using a Cs or Xe source. 6. 4 Results and Discussion Chromatography on Sepharose 4B of the LPS from the wild-type strain Rhizobium leguminosarum biovar viciae VF-39 ( Figure 6.1) showed four major fractions of which A-C corresponded to LPS, and the late-eluting peak was composed almost entirely of a glucan. Hydrolysis of each of the fractions A-C with 1% acetic acid followed by extraction of the lipid A with chloroform-methanol (as described under general methods) and chromatography of the components in the aqueous phase on Biogel P2 gave three sub-fractions, and these three profiles were similar (Figure 6.1). 5. leguminosamm VF 39 SEPHAROSE 4B ABSORBANCE (490 nm) FRACTION NUMBER 1 ‘\ \ T . . y . I T - 2 E E E 2 c. g I: BIOGEL 92 :L v 2 8 2 z . 4 g g 2 3 3 g 3 8 3 a: 4 10 3.0 10 30 10 30 rmc. NO. ram. NO. PRAC. no. A 8' C Figure 6.1 Gel filtration on Sepharose 4B of the crude LPS isolated from Rhizobium leguminosarum biovar viciae VF-39. Fractions A-C were each subjected to mild hydrolysis with 1% acetic acid at 100° C for 2 hours and then extracted with 5:1 chloroform-methanol and the aqueous layer then subjected to gel filtration chromatography on Biogel P2. Peaks with the same numbers had similar compositions and NMR spectra, and contained the O-antigen, a tetrasaccharide 1 and a trisaccharide 2. The vertical axes represent absorbance in the phenol-sulphuric acid assay. SIGNAL 2 3 EM 1 if TIME Figure 6.2 Gas chromatogram (see Experimental) of O-antigen carbohydrate components of Rhizobium leguminosarum biovar viciae VF- 39 as the alditol acetates derived from 2-0-methylfucose, fucose, glucose, 2- amino-2,6-dideoxyglucose, 2-amino-2,6-dideoxyhexose, and heptose in the order 1-6, respectively. GC-MS indicated that peak 1 contained 2-0-methyl fucose, fucose, a 3,6- dideoxy-3-(methylamino) hexose (probably the galacto isomer (19)), glucose, amino-2,6-dideoxyhexose and heptose (Figure 6.2). The presence of most of these components was also indicated by 1H-NMR spectroscopy which showed that the O-antigen was heavily acetylated (signals between 1.8 and 2.2 ppm; Figure 6. 3). There were signals at 0.8-1.4 (CHMe), 2.75 (NMe), ~3.28 ppm (OMe). Peak 2 contained the deacetylated version of a tetrasaccharide made up of Man, Gal, GalA and KDO first characterized in R trifolii ANU843 (12) (the linkage of mannose to KDO was later revised (13)). « The composition and linkage positions were confirmed by methylation analysis and the molecular weight by FAB-MS, The structure is shown in Figure 6.4a, and the 1H-NMR spectrum, which is identical to that of the R trifolii tetrasaccharide afier deacetylation, is shown in Figure 137 6.4b Peak 3 contained the trisaccharide 2 identified (6, 13, 14) in R trifolii ANU843. v Y Y T V v v V v V Y Y Y Y Y V V N4 0,4 0'! A U M .5 Figure 6.3 500-MHz 1l-I-NMR spectrum of O-antigen (peak 1 in Figure 6.1). Ho. 0 H COOH HO OH Figure 6.4a: The structure of the tetrasaccharide 1, l 1 V 1 V Figure 6.4b 1H-NMR spectrum of the tetrasaccharide shown in Figure 6.4a. The three H-l signals in the range 4.8-5.4 ppm and are due to a-Gal, Man and Gal, respectively, in decreasing order of chemical shift. Signals for the poorly resolved deoxy function appear at ~2 ppm. fl ‘ “a 0 0| .3 u ”q .5 pm The gel-filtration profile on Sepharose 4B of the total LPS of the mutant VF- 39-32 was radically different from that of the wild type (Figure 6.5). Two major LPS fractions (A and B) which after hydrolysis (with 1% acetic acid as described for the parent strain) gave profiles on Biogel P-2 that were different from those obtained (Figure 6.1) from the wild type. Thus, peak 1 from fraction A contained none of the amino sugars found in the wild type and appeared to contain only a small amount of a capsular antigen comprising mainly glucose. Peaks 2 and 3 from fractions A and B contained similar labeled major components. Peak 1 from fraction B contained small amounts of the O-antigen found in the parent strain. Peak 3 contained the trisaccharide 2. The major peak contained a component, ABSORBANCE (490 nm) R. leguminosarum VF 39- 32 SEPHAROSE FRACTION NUMBER ‘4‘ BIOGEL 92 ABSORBANCE (490 nm) ABSORBANCE (490 nm) 10 30 >10 ' 30 FRAC. mo _ FRAC. m. Figure 6.5 Gel filtration on Sepharose 4B of the crude LPS of Rhizobium leguminosarum biovar viciae VF-39-32. Fractions A and B were each subjected to mild-acid hydrolysis with 1% acetic acid at 100° C for 2 hours and then extracted with 5:1 chloroform-methanol and the aqueous layer then loaded to gel filtration chromatography on Biogel P2. Peak 1 from fraction A, which contained mainly glucose, indicating it might be due to small amounts of capsular material, was devoid of amino sugars. Peak 1 from fraction B contained a small amount of O-antigen. Peak 3 from fraction A and B contained the trisaccharide L and peak 2 contained an unidentified product. 140 OH OH O . O HO OH OH OH 0 93' 5H 0 '-. HO O OH 0min... Figure 6.6a: The structure of the trisaccharide 2, 1H—N'MR spectrum is shown in Figure 6.6b. Two sets of up-field resonances dd at 2.02 (J = 14.5 and 1 Hz) and 2.60 ppm (J = 14.5 and 7.6Hz) indicated the presence of a deoxy sugar. GC (EI and CI)-MS of alditol acetate and trideuteroacetate from this residue, confirmed the presence of Gal and Man and indicated the presence of a 2-deoxyaldose. The signal at 4.53 ppm (dd) was coupled to the methylene group, and the chemical shift, being due to an anomeric proton, is far down-field for H-4 of KDO. The structure of this molecule is still under investigation, but it does not correspond (6) to a truncated form of the tetrasaccharide 1 in Figure 6.4a. The gel-filtration elution profile on Sepharose 4B of the LPS of the mutant, VF-39-86 was also different from that of the wild type (Figure 6.7). The major fractions A-C, when subjected to mild acid hydrolysis followed by gel filtration on Biogel P2, yielded profiles (Figure 6.7) which were different to those obtained from the parent strain. Thus, peak A yielded three peaks of which peak 1 corresponded to traces of capsular material but none of the components present in the O-antigen of the parent strain (Figure 6.2). The quantity of material isolated was too small for 1H-NMR 141 I I I I l I I 'l I I I l I I ‘I 3 l 3.3 3.2 3.1 3.0 2.9 2.8 2,7 2.5 2.5 2.4 2.3 2.2 no: PPM Figure 6.6b 1H-NMR spectrum of unidentified component from strain VF- 39-32 (peak 2 in Figure 6. 5). The peaks labelled with asterisks are due to contaminants. spectroscopy. Peaks 2 and 3 also contained very small amounts of material, and their 1H-NMR spectra contained peaks down-field of 6 ppm indicating that they may contain nucleic acid. Fraction B gave several small peaks each of which contained carbohydrate in quantities too small for study. However, the major peak (4) from fractions B and C contained three aldopyranoside components, as indicated by the presence of the three H-l resonances between 4.8 and 5.2 ppm in the 1H-NMR spectrum (Figure 6.8). The ion-exchange chromatography of this material on DEAE-Sephadex yielded two components; one of them had a 1H-NMR spectrum (Figure 6.9a) that contained only one H-l resonance and which corresponded to a 142 5. Ieguminosarum VF 39—86A SEPHAROSE 4B 20 0 60 80 FRACTION NUMBER ——.‘ 1 BIOGEL P2 1 ML ML E ' ‘ 10 30 10 3o 10 30 FRAC. NO. FRAC. NO. FRAC. NO. ABSORBANCE (490 nm) ABSORBANCE (490 nm) ABSORBANCE (490 nm) .< Figure 6.7 Gel filtration elution profile on Sepharose 4B of the crude LPS isolated from Rhizobium leguminosarum biovar viciae VF-39-86. Fractions A-C were each subjected to mild-acid hydrolysis then to gel filtration on Biogel P2 as described for the parent strain and strain VF-39- 32. Fraction A yielded small amounts of capsular material (peak 1) and possibly nucleic acid material (peaks 2 and 3). Fractions B and C gave one major component (peak 4). 143 disaccharide (Figure 6.9b) of Man and KDO. The same molecule was obtained (6) from the LPS of a R. phaseoli mutant and shown by NMR spectroscopy and methylation analysis to be a-Man-(1-5)-KDO [3]. The molecular weight of this component was confirmed by FAB-MS of m/z 445 for the sodium adduct of sodium carbonated molecule [M+Na]+ (Figure 6.10). The fatty acids found in the lipid A of the parent strain and both mutants were the same and in the same proportions as those found in Rhizobium leguminosarum biovar trifolii ANU843 (7, 8), namely C14, C15, C15 and C13 3-hydroxy acids and 27-hydroxyoctacosanoic acid. Figure 6.8 1H-NMR spectrum of the material in peak 4 of Figure 6.7. 144 The results obtained for the symbiotically deficient mutants VF-39-32 and VF-39-86A of Rhizobium leguminosarum indicate that (a) the ability of the bacterium to synthesize an intact LPS, containing the O-antigen is necessary for proper symbiosis, (b) the synthesis of the core trisaccharide 2 component is independent of the synthesis of the tetrasaccharide 1, and (c) the bacterium can synthesize more than one type of LPS and impairment or suppression of the synthesis of the usual LPS can lead to the synthesis of an alternative LPS which is not normally expressed. This work underlines the importance of bacterial cell surface polysaccharides in infection. Figure 6.9a 1H-NMR spectrum of the disaccharide from peak 4 of Figure 6.7 purified by ion-exchange chromatography. “oiling-5‘36} m<“'rrn>'—‘m:;g 145 (3 HO COOH Figure 6.9b: The structure of the disaccharide of mannose and KDO. 104 5.. 445 I . .I ' . 350 460 430 560 550 600 m/z Figure 6.10: The mass spectrum of the sodium adduct to the sodium carbonated disaccharide a-Man-(1-5)-KDO. 146 6.5 References 2‘5”.“ 5" 10. 11. 12. 13. 14. 15. Bohlool, B. B., and Schmidt, E. L., Science, (1974) 188 269-271. Dazzo, F. B., and Hubbell, D., Appl. Microbial., (1975) 39. 1071-1023. Maier, R., and Brill, W., J. Bacterial, (1978) 133 1295-1299 Noel, K., Vandenbosch, K. and Kulpaca, B., J. Bacterial., (1986)168 1392-1401. Priefer, U., J. Bacterial., (1989) 111 6161-6168. Carlson, R. W., Garcia, F., Noel, D. and R. I. Hollingsworth, Carbohydr. Res., (1989)188 101-110. Hollingsworth, R. I. and Carlson, R. W., J. Biol. Chem., (1989) 284 9300-9303. Bhat, U. R., Mayer, H., Yokuta, A., Hollingsworth, R. I. and Carlson, R. W., J. Bacterial, (1991) 113 2155-2159. Hollingsworth, R. I., and Elghanian, D. L., J. Biol. Chem. (1989) 2.6.4 14039-14042. Hollingsworth, R. I. and Elghanian, D. L., “Endataxin Research” (1990) 1 73-84, Spitzer, N. J. J. and Ziegler, E. J. (Eds.), Elsevier, Amsterdaam. Carlson, R. W., Sanders, R., Napoli, C. and Albersheim, D., Plant Physiol., (1978) 82 912-917. Hollingsworth, R. I., Carlson, R. W., Garci, F. and Gage, D. J. Biol. Chem., (l989)2649294-9299. ' Hollingsworth, R. I., Carlson, R. W., Garci, F. and Gage, D., J. Bial.Chem., (1989) 85 12752. Carlson, R. W., Hollingsworth, R. I. and Dazzo, F. B., Carbohydr. Res., (1988)116 127-135. Carlson, R. W., unpublished data. 16. 17. 18. 19. 147 Hollingsworth, R. I., Abe, M. and Dazzo, F. B., Carbohydr. Res., (1984) 183 Cl-C4. Westphal, O. and Luderitz, O., Angew. Chem. Int. Ed. Enl. (1954) 68 407-417. Dubois, M., Gilles, R. A., Hamilton, J. K., Roberts, P. A. and Smith, F.,Anal. Chem., (1956) 25 350-354. Hollingsworth, R. I., Hrabak, E. M. and Dazzo, F. B., Carbohydr. Res., (1986)1fl 103-113. CHAPTER 7 ISOLATION AND TENTATIVE CHARACTERIZATION OF A TRISACCHARIDE CONTAINING A NEW 2-DEOXY-3-C- (HYDROXYMETHYL)PENTOFURANOSYL COMPONENT FROM THE LIPOPOLYSACCHARIDE OF A RHIZOBI UM LEGUMHVOSARUM BIOVAR VICIAE LPS MUTANT 7.1 Introduction In a recent study(1), we described structural defects of the lipopolysaccharide of two mutants of Rhizobium leguminasarum biovar viciae VF-39 which are impaired in their symbiotic phenotype with respect to infection of their plant host. This impairment is characterized by failure of the bacteria to be released from infection threads into the nodules (2). One of these mutants, VF-39-32, was found to be incapable of synthesizing a tetrasaccharide component normally made by the wild-type organism. This tetrasaccharide contains mannose, galactose, galacturonic acid and 3-deoxy- 2-octulosonic acid. Instead, an oligosaccharide containing an unusual deoxyglycosyl component was isolated. In this report we describe the structure of this unusual glycosyl component and complete the characterization of the effect of the mutation on lipopolysaccharide structure. 7 .2 Experimental 7.2.1 Culture of Bacteria and Isolation of LPS and LPS Fractions Bacteria were grown in a modified Bergensens medium (3) in shaken- broth cultures. Mutants were grown (4) in the presence of antibiotics. The LPS were isolated by the hot-phenol method (5) and after treatment with DNAse and RNAse purified on Sepharose 4B using an ammonium formate 148 149 buffer as described (6, 7). Column eluates were analyzed by the phenol- sulfuric acid method (8). Fragments were released from the LPS by hydrolysis for two hours with aqueous 1% acetic acid at 100° C. The hydrolysate was extracted with 5:1 chloroform-methanol in order to remove lipid A, the aqueous layer was lyophilized, and the residue was chromatographed on a column of Biogel P-2 using aqueous 0.1% formic acid. Fractions which appeared impure by 1H-NMR spectroscopy and other analyses were subjected to ion-exchange chromatography on DEAE-Sephadex G25 in aqueous 0.01% formic acid adjusted to pH 6.5 with ammonia, eluting with a linear gradient of 0-0.2 M ammonium chloride over 200 mL. 7 .2.2 Compositional Analyses Compositional analyses for fatty acids and carbohydrates in lipid A were carried out as described (6, 7, 9, 10) after conversion into methyl esters and alditol acetates, respectively. GC analysis of carbohydrates was carried out on a Hewlett Packard 5890 gas chromatograph using a capillary D3225 column and a flame-ionization detector. Mass spectrometry detection was effected with a JEOL 505 mass spectrometer in the El or C1 (ammonia) mode. 7.2.3 Methylation Analysis After treating with methanol containing a trace of HCl at 55° C followed by evaporation to dryness and reduction by sodium borodeuteride, the product was permethylated with dimsyl anion and methyl iodide, concentrated to dryness and adsorbed on a C-18 reverse-phase cartridge in water. The cartridge was eluted with water, 2:1 methanol/water, and finally with pure methanol. Each eluate was collected separately and subjected to acid hydrolysis with aqueous 2 M TFA at 120° C for 1 hour. The hydrolysates 150 were then reduced with sodium borohydride, acetylated, and subjected to GC/MS analysis. 7.2.4 Structural Analyses Oligosaccharides were methylated after reduction of carboxyl and other carbonyl functions with sodium borodeuteride as described. 1H NMR spectroscopy (500-MHz) was performed on solutions in D20 with a Varian VXR500 spectrometer. FAB-MS was performed with a JEOL HX110-HF instrument in the negative- or positive-ion mode and using a Cs or Xe source. 7 3 Results and Discussion Chromatography on Sepharose 4B of the crude LPS isolated from the wild-type strain Rhizobium legaminasarum biovar viciae VF-39 (top. part of Figure 7 .1) showed four major fractions. A-C corresponded to LPS, and the late-eluting peak was composed almost entirely of a glucan.‘ Hydrolysis of each of the fractions A-C with 1% acetic acid followed by extraction of the lipid A with chloroform-methanol (as described under general methods) and chromatography of the components in the aqueous phase on Biogel P-2 gave three sub-fractions, the three elution profiles were similar (bottom part of Figure 7.1). The gel-filtration profile on Sepharose 4B of the total LPS of the mutant VF- 39-32 was radically different from that of the wild-type (top part of Figure 7.2). Two major LPS fractions (A and B) which, after hydrolysis (with 1% acetic acid as described in general methods) gave elution profiles on BiogelP-Z (bottom part of Figure 7.2) that were different from those. (bottom part of Figure 7.1) obtained from the wild-type. Thus, peak 1 from fraction A contained none of the amino sugars found in the wild type and appeared to 151 5. leguminosarum VF 39 SEPHAROSE 4B ABSORBANCE (490 nM) FRACTION NUHBER . 1\ —( E A A c E g o o o 3 3 $ g, BIOGEL P2 5 U 2 z . «t g 2 g g ' 3 . AA to In I _J ‘ 10 3,0 10 3O 10 3o PRAC. NO. FRAC. NO. PRAC. NC. A B C Figure 7.1: The gel filtration profile on Sepharose 4B of the crude LPS isolated from Rhizobium leguminasarum biovar viciae VF-39 is shown at the top of the Figure. Fractions A-C were each subjected to mild hydrolysis with 1% acetic acid at 100° C for two hours and then extracted with 5:1 chloroform-methanol. The aqueous layer was then subjected to gel filtration chromatography on Biogel P-2. Peaks at bottom of the figure with the same numbers had similar compositions and NMR spectra, and contained the 0- antigen, a tetrasaccharide 1 and a trisaccharide 2. The vertical axes represent absorbance in the phenol-sulfuric acid assay. 152 B. leguminosarum VF 39—32 SEPHAROSE 4 B ABSORBANCE (490 nm) l" 40 60 80 FRACTION NUMBER N o —H N N E E C O a - o v 0‘ v v m " BIOGEL 92 U {I} 2 U 4 fi 2 g 1 3 3 2 3 m I “ 3 10 1 3o 10 ' 30 FRAC- "0- . rmc. no. A s' Figure 7.2: Gel filtration profile on Sepharose 4B of the crude LPS of Rhizobium leguminasarum biovar viciae VF-39-32 is shown at the top of the figure. Fractions A and B were each subjected to mild-acid hydrolysis with 1% acetic acid at 100° C for two hours and then extracted with 5:1 chloroform-methanol. The aqueous layer was subjected to gel filtration chromatography on Biogel P2, and the resulting elution profile is shown at the bottom of the figure. Peak 1 from fraction A, which contained mainly glucose indicating it might be due to small amounts of capsular material, was devoid of amino sugars. Peak 1 from fraction B contained a small amount of O-antigen. Peak 3 from fractions A and B contained the trisaccharide 2., and peak 2 contained an unidentified product. 153 contain only a small amount of a capsular antigen comprising mainly glucose. Peaks 2 and 3 from fractions A and B contained similar labeled major components. Peak 1 from fraction B contained small amounts of the 0- antigen found in the parent strain. Peak 3 from fractions A and B contained the trisaccharide 2, and Peak 2 contained an unidentified product. The 1H NMR spectrum (Figure 7.3) of peak 2 (at the bottom of Figure 7 .2), the major component from the ion-exchange column, suggested the presence of a deoxy function at one of the ring positions by the presence of an up-field doublet of doublets (J = 14.5 + 1 Hz) at 2.02 ppm and another doublet of doublets (J = 14.5 + 7.6 Hz) at 2.60 ppm. The unusually large chemical shift at 2.60 ppm suggested that. the ring might have the furanose form, and the large splitting indicated geminal protons. Irradiation of the more up-field doublet of doublets at 2.02 ppm (Figure 7.4) led to the collapse of the second doublet of doublets at 2.60 ppm as well as loss of the smaller splitting of another doublet of doublets (J = 7.6 + 1 Hz) at 4.53 ppm. This latter signal was assigned to an anomeric proton. This experiment also confirmed that the pair of up-field doublet of doublets were indeed mutually coupled. Irradiation of the doublet of doublets at 2.60 ppm (Figure 7.5) led to the collapse of the more up-field signal at 2.02 ppm and to the collapse of the anomeric proton at 4.53 ppm. This indicated the presence of a 2-deoxy function in the molecule. The lack of perturbation of any other signals besides the anomeric proton when either of the 2-deoxy methylene protons was irradiated indicated that no proton was present on the 3-position. The signal with the large splittings at 2.60 ppm, are very far down-field indicates that is not due to an axial proton with an axial neighbor on an adjacent carbon in a pyranose chair form. The possibility of a pyranose chair form was also ruled out by the very small coupling constant displayed between the most up-field doublet of doublets at 154 PPm Figure 7.3: 500MHz 1H-NMR spectrum of peak 2 (at the bottom part of figure 7 .2), which was the aqueous fraction of the mild, partial acid hydrolysate of peak B (at the top part of Figure 7.2) analyzed by the Biogel P- 2 column. Note the presence of the upfield doublet of doublets (J = 14.5 Hz + 1 Hz) at 2.02 ppm and a second doublet of doublets (J = 14.5 Hz + 7.6 Hz) at 2.60 ppm. The 14.5 Hz mutual coupling indicates the presence of a methylene group and the downfield chemical shift suggests a deoxy function. The unusual chemical shift with the 7.6 Hz coupling indicates that the ring might not be in a pyranose form. 155 T I l I' _.. .. PPm Figure 7.4: 500MHz 1H-N'MR spectrum of peak 2 (at the bottom of Figure 7.2) from the Biogel P-2 column with irradiation of the doublet of doublets at 2.02 ppm. Note the collapse of the other doublet of doublets to a simple doublet with a splitting of 7.6 Hz. The loss of the 14.5 Hz coupling confirms that the two upfield signals of 2.02 and 2.60 ppm are due to geminal protons. Note also that another doublet of doublets (J = 7.6 Hz + 1 Hz) at 4.53 ppm (Figure 7.3) has lost the smaller splitting. The chemical shift of this signal suggests that it is due to an anomeric proton, thus making the methylene group a 2-deoxy function. 156 PPm Figure 7.5: 500MHz 1H-NMR spectrum as in Figure 7.3, after irradiating the signal at 2.60 ppm. Note that the signal at 2.02 ppm has collapsed to a very narrow doublet. Note also that the signal at 4.53 ppm has collapsed to a doublet with a very small splitting. This result confirms that the two upfield signals and the one at 4.53 ppm form a closed-spin system and confirms that the glycosyl residue responsible for them has a 2-deoxy function. The lack of any other coupling to the 2-deoxy-methylene protons indicates that there are no protons on the 3-position. 157 2.02 ppm and the anomeric proton at 4.53 ppm. In a pyranose chair configuration, the most up-field proton would be equatorial and cis- with the anomeric proton, and the expected coupling constant for these cis- protons should be 3-4 Hz instead of the observed 1 Hz. In the Figure 7.3, the anomeric proton must be axial in order to show the 7.6 Hz coupling with one of the neighboring C-2 hydrogen. In addition, this most up-field doublet of doublets shifts too high compared to the other 2-deoxy-methylene signal to be the equatorial component. These facts strongly suggested that a pyranose form did not exist. There are two signals in the anomeric region of the spectrum (Figure 7.3 and expanded in Figure 7.6A). One signal at 4.98 ppm was assigned to an a-anomeric proton of mannose and the other, a doublet at 4.89 ppm, with 1.5 Hz coupling, was assigned to the a-anomeric proton of galactose. Upon irradiation of the doublet of the doublets at 2.60 ppm, an increase in the intensity of the mannosyl residue relative to the more upfield galacosyl residue is noted (Figure 7.5 and expanded in Figure 7.6B). Also on irradiation of the doublet of doublets at 2.02 ppm, enhancement of the mannose signal is seen (Figure 7.4 and expanded in Figure 7.6C). This indicates that the mannosyl residue is in the position adjacent to the 2-deoxy- methylene group. The n. 0. e. to anomeric proton of mannose in the 3 position of the furanose ring can be observed; this is evidence for a five- membered ring, because generally such phenomena are not observed in the pyranose form. After reduction by sodium borohydride, acid hydrolysis, reduction and acetylation, gas chromatography produced the chromatogram of the alditol acetates of the oligosaccharide components shown in Figure 7 .7. There were 158 a“ —-1 - d d -I — —d PPm Figure 7.6: A) Partial 1H-NMR spectrum showing anomeric protons of mannose and galactose residues. B) Spectrum after irradiating of the doublet of doublets at 2.60 ppm. Note the increase in intensity of the mannosyl residue relative to the more upfield galactosyl residue. This indicates that the mannosyl residue is in the position adjacent to the methylene group. C) Spectrum after irradiation of the doublet of doublets at 2.02 ppm. Again note the enhancement of the mannose signal. 159 four peaks appearing as a pair of doublets. Peaks 1 and 2 had identical mass spectra. The chemical ionization mass spectrum, using ammonium as a carrier gas, showed that the first peak of the first pair had a pseudo- molecular ion at m/z 352 (Figure 7.8) which corresponds to a true molecular weight of 334, since analytes form an ammonium adduct under the CI conditions. The structure of the first alditol acetate is shown in Figure 7.9; peracetylation of the first alditol acetate was not complete, because the tertiary hydroxyl group is difficult to acetylate. The pseudo-molecular ion for the second peak of the first pair appeared at m/z 366 (Figure 7.10), which suggested a true molecular weight of 348, and indicated that the two peaks differed by 14 mass units or one methyl group. The structure of the second alditol acetate is shown in Figure 7.11. Methylation at the tertiary hydroxyl group explained the 14 mass units different between the two structures in Figures 7.9 and 7.11; this was due to 1% HCl in methanol applied during the derivatization procedure. Peak 3 in Figure 7.7 is due to mannose, and peak 4 is due to galactose. These later two components were detected in equal proportions; these assignments were confirmed by GC/MS. The GC and GC/MS assignment of the mannose and galactose also confirmed the 1H-NMR spectrum (Figure 7.3), in which two signals appeared: one singlet at 4.98 ppm and the other signal at 4.84 ppm with 1.5 Hz coupling, characteristic of these residues. 160 11.2133 . Retention time Figure 7.7: Gas chromatogram of the alditol acetate of the unusual glycosyl components. Peaks 1 and 2 are due to the component being identified. Peak 3 is mannose, and peak 4 is galactose. 04: 00390.3CU'II 161 too see 303 V were T ' 5%- - ‘7 sea M/Z Figure 7.8: The CI mass spectrum of the first peak from the first pair of doublets in the gas chromatogram of Figure 7.7. CH20Ac CHOAc HO - (ll- CHZOAc CHZOAc m/z 334 Figure 7 .9: The structure of the alditol acetate in Figure 7.8. 00:00.3an acme-tr—mm 162 SB . 56 Jr] 49“ r 36‘ 33. 1e~ , 1 7 273 L ‘ 349 . 133 223 393 433' sec ' ' 'séa' ‘ ' "/2 Figure 7.10: The CI mass spectrum of the second peak from the first pair of doublets in the gas chromatogram of Figure 7.7. CHZOAc CHOAc MeO - (:3-01-120Ac ?“2 CH20Ac m/z 348 Figure 7.11: The structure of the alditol acetate for Figure 7.10. 163 The oligosaccharide was subjected to methylation analysis. After treating with methanol containing a trace of HCl at 55° C followed by evaporation to dryness and reduction by sodium borodeuteride, the product was then permethylated with dimsyl anion and methyl iodide, concentrated to dryness and adsorbed to a C-18 reverse phase cartridge in water. The cartridge was eluted with water, 2:1 methanol/water, and finally with pure methanol. Each eluant was collected separately and subjected to acid hydrolysis with aqueous 2 M TFA at 1200 C for 1 hour. The hydrolysates were then reduced with sodium borohydride, acetylated, and subjected to GC/MS analysis. The total ion chromatogram (TIC) for the methylation analysis appears in Figure 7.12. The fraction contained three major peaks, the small peaks at the very front were due to minor contaminate. Peaks 1 and 2 had identical mass spectra, and were due to terminal mannose and galactose (Figure 7 .13), the structure and fragmentations were showed in Figure 7.14. The peak 3 had a mass spectrum which is unlike that of a normal partially methylated, peracetylated hexose (Figure 7.15). The branching residue in the unusual compound is confirmed by the m/z 261 and 203 fragmentations, and this also supported the NMR evidence that there is no proton at C-3. The proposed structure and fragmentations are shown in Figure 7.16. The 2:1 methanol/water fraction was a major component which contained terminal mannose, terminal galactose and a KDO derivative. 164 ...mu 1 V6.35 4 .1 398‘ 12 :2'6‘8 1 -3113“. ..... .2 L - .11. - r I I ‘T I 23 15 ‘ “7 I S ' 1) IS 3‘0 Retention Time Figure 7.12: The total ion chromatogram for the methylation analysis and fraction contained three major peaks, the small peaks at the very front were due to minor contaminants. 165 5...... 1m 11I7 145 IIISI~ 129 161 ran-v 205 87 71 335... e'rIlIIIIIrJrrIT IIII.II11II1HH1I 6' 6‘ J! 75 OJ 5 S. SS 1.6165110115181181301131t0 145150155!“ 16$INIY‘513I16511I1553NZOS Figure 7.13: Mass spectrum of peak 1 and 2 from Figure 7 .12. I 1 17 CH20MO Figure 7.14: The structure and fragmentations for Figure 7.13. 166 135580 ~ 1 1 9 10113 -‘ 9‘50 -1 161 53625 - 288 73 245 348 . IITT l l I T l T 60flll90milfiilllilifllfllflflllflll30321622125021.2501“Ul2fl2fl5flfllmlflifl Figure 7.15: Mass spectrum of peak 3 from Figure 7.12. 261 (311,013.- CHOAc l AcOCHg—C—OCHa ; .42 _42 CH2 203——> 161 ——> 119 I cnzom - 60 -43 M+=348 —I> 288 ———> 245 Figure 7.16: The structure and fragmentations for Figure 7.15, the branching residue is suggested by the 261 and 203 fragments and is supported by the NMR evidence that there is no proton at C-3. 167 There are two possible configurations of the furanose ring: one is envelope form shown in Figure 7.17, and another is the twisted ring form shown in Figure 7.18. OH OH 0 OH H0 03 OH H4 H1 H0 H0 H28 0 OH H,., Figure 7.18: Twist configuration of the furanose ring. According to the semiempirical calculations (11, 12) of five member ring compounds of butyrolactone and 3-hydroxybutyrolactone, the calculated coupling constants of the envelope form and the twisted ring are very close, and the difference in energy of these two forms is that the envelope form is 0.3 kJ mol'1 higher than the twisted ring. In both Figures 7.17 and 7.18, all of the hydroxyl groups prefer to occupy the pseudo-axial positions in order to 168 reduce the electrostatic effect between the two pseudo equatorial hydroxyl groups. From the current data obtained in this research it is difficult to determine the exact configuration of the furanose ring. Further computational work is necessary to get this part of the work done. 7.4 1. 10. 11. 12. References Zhang, Y., Hollingsworth, R. I., and Priefer, U. B. Carbhydr. Res. (1992) 231 261-271. Maier, R. and Brill, W., J. Bacterial, (1978) 183 1295-1299. Hollingsworth, R. I., Abe, M. and Dazzo, F. B., Carbohydr. Res., (1984) 133 C1-C4. I: Priefer, U., J. Bacterial, (1989) 111 6161-6168. Westphal, O. and Luderitz, 0., Angew. Chem., (1954) 88 407-417. Hollingsworth, R. I., and Elghanian, D. L., J. Biol. Chem. (1989) 254 14039-14042. Hollingsworth, R. I. and Elghanian, D. Lill, “Endataxin Research,” (1990) 1 73-84, Spitzer, N. J. J. and Ziegler, E. J. (Eds.), Elsevier, Amsterdam. Dubois, M., Gilles, R. A., Hamilton, J. K., Roberts, P. A. and Smith, F., Anal. Chem., (1956) 25 350-354. Hollingsworth, R. I. and Carlson, R., J. Biol. Chem., (1989) 284 9300- 9303. Bhat, U. R., Mayer, H., Yokuta, A., Hollingsworth, R. I. and Carlson, R. W., J. Bacterial, (1991) 113 2155-2159. Riggs, N. V., Aust. J. Chem., (1985) 38 1575-1583. Riggs, N. V., Aust. J. Chem., (1985) 38 1585-1589. M 'r w? . ..._ .‘HJ' fifl ‘h‘ "31331333?