UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Investigation of mass spectrometric techniques for the structural determination and the sequencing of… Lam, Zamas 1987

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

Item Metadata

Download

Media
831-UBC_1987_A6_7 L35.pdf [ 5.88MB ]
Metadata
JSON: 831-1.0059809.json
JSON-LD: 831-1.0059809-ld.json
RDF/XML (Pretty): 831-1.0059809-rdf.xml
RDF/JSON: 831-1.0059809-rdf.json
Turtle: 831-1.0059809-turtle.txt
N-Triples: 831-1.0059809-rdf-ntriples.txt
Original Record: 831-1.0059809-source.json
Full Text
831-1.0059809-fulltext.txt
Citation
831-1.0059809.ris

Full Text

INVESTIGATION OF MASS SPECTROMETRIC TECHNIQUES FOR T H E STRUCTURAL DETERMINATION AND T H E SEQUENCING OF SOME BACTERIAL CAPSULAR POLYSACCHARIDES FROM THE FAMILY Enterobacteriaceae; Klebsiella AND Escherichia coli. by Z A M A S L A M B.Sc. (Hons.) Thames Polytechnic, London, England, 1984 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF T H E REQUIREMENTS FOR T H E DEGREE OF MASTER OF SCIENCE IN T H E F A C U L T Y OF G R A D U A T E STUDIES (DEPARTMENT of CHEMISTRY) We accept this thesis as conforming to the required standard T H E UNIVERSITY OF BRITISH COLUMBIA May 1987 © Z A M A S L A M 1987 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department The University of British Columbia 1956 Main Mall Vancouver, Canada V6T 1Y3 Date ZO* J ^ n , /jSl DE-6(3/81) - ii -A B S T R A C T The structural elucidation of bacterial capsular polysaccharides is traditionally performed by using "wet chemical procedures" but instrumental methods such as nuclear magnetic resonance spectroscopy and novel mass spectrometric techniques are coming into prominence. In this thesis three different mass spectrometric techniques were investigated to establish their applicability for the sequencing of bacterial capsular polysaccharides. These techniques included fast atom bombardment (FAB), desorption chemical ionisation (DCI) and laser desorption ionisation Fourier transform ion cyclotron resonance spectroscopy (LDI-FTICR). The soft ionisation produced by these methods allows sequential loss of individual sugar residues without excess thermal decompositon of the ring. Thus sequencing of oligosaccharides could be achieved. The most common of all these three techniques is FAB which is already considered to be a well established form of soft ionisation, although the exact mechanism of ionisation is unknown. The utilisation of DCI has not been thoroughly exploited in carbohydrate research, due to the non-volatility of oligosaccharides and the possible thermal decompostion of the sample in the source. LDI-FTICR due to the general unavailability of the instrument has only been used for model studies of "shelf carbohydrates". In the course of this work it was found that FAB MS and DCIMS complement each other. The sequence of linear oligosaccharides of up to five sugar units can be deduced from either the native or permethylated sample. If the oligosaccharides investigated were generated by phage-borne enzyme, the total sequence of the native polysaccharide can be established. This was illustrated by the use of FABMS on Klebsiella K44 de-O-acetylated oligosaccharide and the reduced oligosaccharide. The sequence of the polysaccharide was shown to be: - iii -HexA deoxyHex — deoxyHex Hex Hex I OAc The location of base-labile acetate group is very difficult to establish chemically. However from the FABMS data, its location was easily recognised by the extra mass carried on the Q-acetylated residue. Simple branched oligosaccharides can also be delineated if derivatised samples were used. The oxonium ions formed are generally neutralised by hydroxyl groups, thus the branch point of the oligosaccharide will be labelled with either two or three hydroxyl groups. Therefore, from the mass difference between the "permethylated" and the labelled sugar residues the sequence and the branch point can be established. This was illustrated by the use of DCIMS on permethylated Klebsiella K46 phage degraded oligosaccharide. The sequence, the branching pattern and the location of the pyruvic acid acetal group for this 4+2 type structure were established to be: Hex Hex HexA Hex Hex Hex > pyr The location of acid-labile pyruvic acid acetal group, like base-labile acetate group, is also difficult to establish chemically. The fragment ions arising from permethylated oligosaccharides were mostly non-reducing end residues. This is due to the stability of oxonium ion formation. However, when an amino sugar was investigated, oxonium ions were not observed. Instead the cleavage took place between the glycosidic oxygen and the reducing end residue. This fragmentation route is in sharp contrast to previously reported spectral data. This may be due to the fact that amino sugars strengthen the glycosidic bond between the oxygen and the carbon-1. LDI-FTICR was investigated for its applicability to a "real" sample. The sequence of the linear Klebsiella K44 de-Q-acetylated, phage degraded oligosaccharide was determined from the spectrum. Furthermore, a few positions of linkage were also deduced from ring cleavage - iv -fragments. Although linkage positions can be obtained from methylation analysis data, some sugar residues such as deoxyhexoses are more labile than others, thus positions of linkage obtained from LDI-FTICR can be used for confirmation. - V -T A B L E O F C O N T E N T S Contents Page A B S T R A C T ii T A B L E O F C O N T E N T S v LIST OF M A S S S P E C T R A viii LIST O F T A B L E S x LIST O F FIGURES xii LIST O F A B B R E V I A T I O N S xiv A C K N O W L E D G M E N T S xvi P R E F A C E xvii A P P E N D I X 108 I I N T R O D U C T I O N 1 I. 1 Immunological importance of bacterial exopolysaccharides 2 II METHODOLOGY IN MASS SPECTROMETRIC A N A L Y S I S OF P O L Y S A C C H A R I D E S 6 II. 1 The mass spectrometer 9 II. 1.1 Ion source 10 II. 1.1.1 Electron ionisation 10 II. 1.1.2 Chemical ionisation 11 II. 1.1.3 Secondary ion mass spectrometry 12 - vi -II. 1.1.4 Fast atom bombardment 13 II. 1.1.5 Laser desorption ionisation 16 II. 1.2 Mass analyser 17 II. 1.2.1 Magnetic analyser 18 II. 1.2.2 Electrostatic analyser 19 II. 1.2.3 Time-of-flight analyser 19 II. 1.2.4 Fourier transform ion cyclotron resonance spectroscopy 20 II. 1.2.5 Quadrupole mass filter 21 11.2 Methylation procedure 21 11.3 Bacteriophage degradation of capsular polysaccharides 22 III MASS SPECTROMETRIC ANALYSIS OF POLYSACCHARIDES 24 III. 1 Laser desorption ionisation Fourier transform ion cyclotron resonance mass spectrometry of the capsular polysaccharide of Klebsiella serotype K44 25 III. 1.1 Abstract .25 III.1.2 Introduction 25 III. 1.3 Experimental 25 III. 1.4 Results and Discussion 26 III. 1.5 Conclusion 36 III.2 Desorption chemical ionisation mass spectrometry of some capsular polysaccharides of Klebsiella and E . coli 42 III.2.1 Abstract 42 - vii -111.2.2 Introduction 42 111.2.3 Experimental 43 111.2.4 Results and Discussion 43 111.2.5 Conclusion 58 111.3 Fast atom bombardment mass spectrometry of some capsular polysaccharides of Klebsiella and E . coli 66 111.3.1 Abstract 66 111.3.2 Introduction 66 111.3.3 Experimental 67 111.3.4 Results and Discussion 68 111.3.5 Conclusion 83 111.4 Discussion 91 111.5 Conclusion 98 VI F U R T H E R W O R K 99 V H B I B L I O G R A P H Y 101 - viii -LIST O F M A S S S P E C T R A Spectra Page III. 1.1 Tabulated positive ion LDI-FTICR spectrum of Klebsiella K44 de-O-acetylated, phage degraded oligosaccharide doped with NaBr 37 III. 1.2 Positive ion FAB spectrum of Klebsiella K44 native phage degraded oligosaccharide 39 III. 1.3 Tabulated positive ion LDI-FTICR spectrum of Klebsiella K44 native polysaccharide doped with K B r 40 III. 1.4 Tabulated positive ion LDI-FTICR spectrum of Klebsiella K44 de-O-acetylated, phage degraded oligosaccharide doped with KBr 41 111.2.1 Positive ion ammonia DCI spectrum of peracetylated raffinose 59 111.2.2 Positive ion ammonia DCI spectrum of permethylated raffinose 60 111.2.3 Positive ion ammonia DCI spectrum of permethylated Klebsiella K79 oligosaccharide derived from Smith degradation 61 III. 2.4 Positive ion ammonia DCI spectrum of permethylated Klebsiella K79 oligosaccharide derived from P-elimination 62 111.2.5 Positive ion ammonia DCI spectrum of permethylated Klebsiella K46 phage degraded oligosaccharide 63 111.2.6 Positive ion ammonia DCI spectrum of reduced, permethylated Klebsiella K46 phage degraded oligosaccharide 64 111.2.7 Positive ion methane DCI spectrum of reduced, permethylated E . coli K44 phage degraded oligosaccharide 65 111.3.1 Positive ion F A B spectrum of raffinose in glycerol 85 111.3.2 Positive ion F A B spectrum of raffinose in thioglycerol 86 111.3.3 Positive ion FAB spectrum of raffinose in mercaptoacetic acid 87 - ix -111.3.4 Positive ion FAB spectrum of raffinose in tetramethyl sulphone 88 111.3.5 Positive ion F A B spectrum of raffinose in tetraethylene glycol 89 III. 3.6 Positive ion FAB spectrum of reduced Klebsiella K46 phage degraded oligosaccharide in thioglycerol 90 i - X -LIST O F T A B L E S Tables Page II. . 1 Wet chemical methods for the determination of sequence in carbohydrates 8 III. 1.1 Satellite peaks of the high intensity structural peaks from LDI-FTICR of de-O-acetylated, phage degraded Klebsiella K44 oligosaccharide 33 III. 1.2 Structures arising from LDI-FTICR of de-O-acetylated, phage degraded Klebsiella K44 oligosaccharide 34 III. 2.1 Fragment structures arising from DCLMS of permethylated Klebsiella K79 oligosaccharide derived from Smith degradation 47 111.2.2 Fragment structures arising from DCLMS of permethylated Klebsiella K79 oligosaccharide derived from P-elimination 49 111.2.3 Fragment structures arising from DCIMS of permethylated Klebsiella K46 phage degraded oligosaccharide 52 111.2.4 Fragment structures arising from DCLMS of reduced, permethylated Klebsiella K46 phage degraded oligosaccharide 55 111.2.5 Fragment structures arising from DCIMS of reduced, permethylated E . coli K44 phage degraded oligosaccharide 57 111.3.1 Structural ions arising from FABMS of underivatised raffinose 69 111.3.2 Structural ions arising from FABMS of permethylated raffinose 71 III. 3.3 Structural ions arising from FABMS of permethylated Klebsiella K79 oligosaccharide derived from Smith degradation 73 III. 3.4 Structural ions arising from FABMS of permethylated Klebsiella K79 oligosaccharide derived from p-elimination 74 III. 3.5 Structural ions arising from FABMS of Klebsiella K67 oligosaccharide derived from chemical degradation 76 -xi-111.3.6 Structural ions arising from FABMS of reduced Klebsiella K46 phage degraded oligosaccharide 78 111.3.7 Structural ions arising from FABMS of Klebsiella K44 de-0_-acetylated, phage degraded oligosaccharide 81 III. 3.8 Structural ions arising from FABMS of reduced Klebsiella K44 de-O-acetylated, phage degraded oligosaccharide 84 111.4.1 Comparison of LDI-FTICR, DCIMS and FABMS. . . 97 - xii -LIST O F F I G U R E S Figure Page 1.1 Schematic representations of Gram-positive and Gram-negative cell wall 4 III. 1.1 cq and 0:2 fragmentation pathways in positive ion LDI-FTICR 27 III. 1.2 Oxonium ion formation 27 III. 1.3 Pi and 02 fragmentation pathways in positive ion LDI-FTICR 28 III. 1.4 Structures derived from a-fragmentation 29 III. 1.5 Structures derived from (3-fragmentation 32 111.2.1 Structure of permethylated raffinose 44 111.2.2 Structure of permethylated Klebsiella K79 oligosaccharide derived from Smith degradation 46 111.2.3 Structure of permethylated Klebsiella K79 oligosaccharide derived from P-elimination 48 111.2.4 Structure of permethylated Klebsiella K46 phage degraded oligosaccharide 51 111.2.5 Structure of reduced, permethylated Klebsiella K46 phage degraded oligosaccharide 53 111.2.6 Structure of reduced, permethylated E. coli K44 phage degraded oligosaccharide 54 III. 2.8 Proposed fragmentation pattern for permethylated E. coli K44 phage degraded oligosaccharide 58 111.3.1 Major fragmentation pathways in positive ion FABMS 67 111.3.2 Structure of Klebsiella K67 oligosaccharide derived from chemical degradation 75 III. 3.3 Structure of Klebsiella K44 de-O-acetylated, phage degraded - xiii -oligosaccharide 80 III.3.4 Structure of reduced Klebsiella K44 de-O-acetylated, phage degraded oligosaccharide 82 - xiv -LIST O F A B B E V I A T I O N S Glc - glucose Gal - galactose Man - mannose Rha - rhamnose Fru - fructose GlcA - glucuronic acid GlcNAc - 2-acetamido-2-deoxyglucose GalNAc - 2-acetamido-2-deoxygalactose Gro - glycerol LPS - lipopolysaccharide Hex - hexose HexA - hexose uronic acid deoxyHex - deoxyhexose HexNAc - acetamido-deoxyhexose Hex r - reduced hexose unit MS - mass spectrometry EI - electron ionisation CI - chemical ionisation DCI - desorption chemical ionisation FD - field desorption FI - field ionisation EHD - electrohydrodynamic ionisation SIMS - secondary ion mass spectrometry FAB - fast atom bombardment - XV -252cf-PD LDI TOF FTICR, FTMS -M Cat amu m/z eV pyr Ac Me mol. wt. NMR G C HPLC DMSO Californium-252 plasma desorption laser desorption ionisation time-of-flight Fourier transform ion cyclotron resonance spectroscopy molecular ion cation atomic mass unit mass to charge ratio electron volt pyruvic acid acetal acetyl methyl molecular weight nuclear magnetic resonance gas-liquid chromatography high performance liquid chromatography dimethyl sulphoxide - xvi -ACKNOWLEDGMENTS I would like to express my sincere gratitude to Professor G.G.S. Dutton and Professor M . Comisarow for their guidance, encouragement, and support throughout the course of this work. I would also like to acknowledge Dr. G. Eigendorf for his help, patience, and technical support, without which this work would not be possible. Thanks are also due to the staff of the M . S. Services, and the Mechanical and Electrical/Electronic Shops. My colleagues in the two different research groups have provided a lot of stimulating discussion, which I gratefully acknowledge. - xvii -.PREFACE* In an effort to familiarise readers who do not work in the particular area of organic chemistry to which this thesis refers, the following explanation of terms used is offered. Fischer projection formulae are used to represent the acyclic modification of sugars. Some examples are shown below. Numbering commences from the carbonyl group at the top of the chain (I). Note that D-glucuronic acid (II) differs from D-glucose (I) only in that C-6 is oxidised to a carboxylic acid group. The C-6 of L-rhamnose (HI) is part of a methyl group and is referred to also by another common name, 6-deoxy-L-mannose. CHO CHO CHO HO-OH O H HO-— O H C H 2 O H D-glucose (I) O H O H HO-HO-O H O H — O H COOH C H 3 D-glucuronic acid (II) L-rhamnose (III) There are four chiral centers in the six-carbon chains (marked with asterisks in structure HI) making it important to appreciate the spatial arrangement of atoms (configuration) that is implied by these Fischer representations. To simplify the nomenclature of all the possible isomers (16 for each of I, n, Ul), all those having the hydroxyl group at the highest - number chiral center (C-5) O H HO-C H 2 O H C H 2 O H D-series L-series - xviii -projecting to the right in the Fischer projection formulae belong to the D-series, and the others to the L-series. Physical and chemical evidence indicates that, in fact, these six-carbon polyhydroxyaldehydes exist in the cyclic form. The ring closure occurs by nucleophilic attack of the oxygen atom at C-5 on the aldehydic carbon atom, generating a new chiral (anomeric) center at C - l . This results in two anomers, represented below in the Tollens formulae. It should be noted that C - l is unique in having two attached oxygen atoms, formally making it a hemiacetal carbon. O H HO HO HO C H 2 O H C H 2 O H ct-D-glucose (TV) (3 -D-glucose (V) Since the Tollens formulae have obvious limitations with their unequal bond lengths, Haworth developed a perspective method of looking at the six-membered ring (VI and VII). This improvement recognises that the ring oxygen atom lies behind the carbon chain and that bond lengths are approximately equal. Often in practise regular hexagons are used in Haworth projections; which he related to such rings as the heterocyclic compound, pyran (VIII) and named them pyranoses. Note that hydroxyl groups not involved in ring formation on the right in Fischer and Tollens formulae point down in the Haworth projections and those on the left point up. Similarly, for aldopyranoses, the group on C-5 points up for D (DC) and down for the L - xix -enantiomer (X). It follows, then, that when sugar residues are attached there are two possible configurations, an a- or P-pyranoside, for each linkage. HO HO H O O O H 2/ O H HO O H O-D-glucopyranose (VI) P-D-glucopyranose (VLT) Pyran (VILT) H O / I ° O H ' C H 3 O H O H oc-D-rhamnopyranose (LX) P -L-rhamnopyranose The true conformation of pyranoid carbohydrates is related to the chair form of cyclohexane. X-ray diffraction analysis has shown that a hexose, such as a-D-glucose (XI), consists of a puckered, six-membered, oxygen-containing carbon ring, with hydroxyl substituents at C - l through C-4, and a hydroxymethyl group at C-5. All substituents on the ring, except for that at C - l , are equatorial. C H 2 O H H O Two isomers (anomers) are possible in relation to the anomeric center (C-l), depending on whether a substituent is axial (a-anomer; XII) or equatorial (P-anomer; XIII), where R = - XX -hydrogen, for monosaccharides, and R = another sugar residue, for di-, oligo, and polysaccharides. Since H - l is in a different chemical environment for the two anomers, nuclear magnetic resonance spectrometry can easily distinguish between them and, thereby, provides invaluable assistance in assigning anomeric configuration. OR H (XII) (XILT) Haworth projections are most useful and will be used in this thesis, even though they give no indication of the three-dimensional molecular shape. There seems to be little justification for the use of formulae which depict states of molecules as well as structures, when the true states are often unknown or variable. * Reproduced with the kind permission of T. E. Folkman from his M. Sc. thesis entitled " Structural Studies on Klebsiella Capsular Polysaccharides", University of British Columbia, April 1979. - x x i -Delicated with love and gratitude to my parents L A M CHE HTM and CHIU KAI YIN -1 -I I N T R O D U C T I O N - 2 -I I N T R O D U C T I O N Carbohydrate-containing macromolecules are found in abundance in all living organisms. These include: a ) exclusively carbohydrate polysaccharide polymers, b ) glycoproteins, proteoglycans and peptidoglycans, c ) glycolipids and lipopolysaccharides (LPS), d ) teichoic acids, and e) nucleic acids 1. Commercial interest in polysaccharides is of long standing in the food, cosmetic, textile, pulp and paper, and the paint industries. Although their thickening and gelling properties have found new application as drilling fluids in the petroleum industry, the main uses of polysaccharides are still in the food industry to control the texture of food as well as its flavour, appearance and colour^. For many years, microbial exopolysaccharides were thought to function simply as energy reserves and structural polymers and also to play a limited role in protection against phagocytosis. It is now well established that complex carbohydrates are also essential in biological recognition where they a) act as receptors for phage and bacteriocins, b) act as highly specific receptors in eukaryotes for viruses, bacteria, hormones and toxins, c) are immunogenic (inducing formation of antibodies), and d ) are specific surface antigens (capable of combining with specific antibodies)^. Other complex carbohydrates act as chemical messengers in plants and are particularly important in regulating growth, development, reproduction and disease control^. 1.1 Immunological importance of bacterial exopolysaccharides In general, the outer most cell component of pathogenic microbes plays a critical role in immune response. Enclosing the plasma membrane of the bacterium is the cell wall, which can be of two types: a) Gram-negative cells have an outer membrane over a peptidoglycan layer, b) -3-Gram-positive cells lack the outer membrane but have an additional component within the peptidoglycan layer. Figure LI is a schematic representation of the Gram-positive and Gram-negative cell walls of bacteria. LPS which exhibit both immunogenicity and full endotoxicity are a main component of the cell wall. It constitutes 10 - 15% (w/w) of the dry cell wall. This species-specific LPS contains the somatic polysaccharide antigen and is called the bacterial O antigen. Many bacteria produce extracellular polysaccharides either in the form of a discrete capsule surrounding the bacterial cell or in the form of loose slime unattached to the cell surface .^ This is the K antigen and constitutes the principle antigen in most of the pathogenic microorganisms. These exopolysaccharides mask the cell wall O antigens and interfere with their serological detection^. Furthermore they also protect the bacteria from phagocytosis and the action of complement. Nevertheless, these exopolysaccharides are also immunogenic and so play an important part in the immune response to bacterial infection. Extensive reviews of the immunology of bacterial capsular polysaccharides have been published^. This immunological defence system is based on recognition of partial structures of the polysaccharide antigens by antibody molecules. These immunodominant structures may be monosaccharide units linked in a specific way, an oligosaccharide in a particular conformation or a non-carbohydrate substituent, eg. pyruvic acid acetal, or N - and O- acetyls^. In acidic polysaccharides the immunodominant sugars are often the charged constituent. In the case of branched polysaccharides the determinant is normally the side chain. Since capsular polysaccharides consist of repeating units, the immunodominant structures are expressed repetitively. Several polysaccharides may share the same antigenic determinant, and can be recognised by the same antibody regardless of their origin. These immunologically related polysaccharides cross-react serologically. Cross-reactions can be used to provide information about the structure of many polysaccharideslO-13. Cell envelope of the Gram-negative cell v a i l Figure 1.1 Schematic representations of Gram-positive and Gram-negative cell wall. -5-Both dead and live "attenuated" bacteria and toxins are commonly used as vaccines for disease prevention. Many exopolysaccharides coupled to carrier proteins can also be used as vaccines**. The first example dates from 1940s, when bacterial polysaccharides were used to protect against pneumococcal infections^. Since then, many microbial polysaccharides have been shown to have noncytotoxic antitumor properties, and hence have been extensively studied in cancer researches, 16. Due to their immunological and serological importance, a fundamental knowledge of bacterial exopolysaccharides structure is essential. This includes their primary structure as well as the three dimensional structure in the solid and solution state. Several laboratories, including ours, have undertaken the structural elucidation of both Klebsiella and E . coli capsular polysaccharides. Generally, this characterisation is performed using mostly "wet chemical" means combined with limited instrumental methods. In this thesis, the possibility of sequencing capsular polysaccharides by using mass spectrometric techniques was explored. These techniques include fast atom bombardment, desorption chemical ionisation, and laser desorption Fourier transform ion cyclotron resonance spectroscopy. M E T H O D O L O G Y IN MASS S P E C T R O M E T R I C A N A L Y S I S O F P O L Y S A C C H A R I D E S - 7 -n M E T H O D O L O G Y IN M A S S S P E C T R O M E T R I C A N A L Y S I S O F P O L Y S A C C H A R I D E S Structural analysis of carbohydrates has always relied heavily on "wet chemical" methods. Instrumental techniques like mass spectrometry (MS) and nuclear magnetic resonance spectroscopy (NMR) are mostly regarded as back up procedures. Nevertheless, in the last two decades the growth in chemical instrumentation has been phenomenal, largely due to advances in the electronics industry. MS and NMR are already playing an important role in carbohydrate research^. A complete analysis of bacterial polysaccharides consists of characterisation and quantification of sugars, the position of linkage, the anomeric nature of the linkage, and the sugar sequence. The conventional strategy using chemical means is to utilise total and partial hydrolysis, methanolysis, carboxyl reduction, deamination, methylation analysis, periodate oxidation, Smith degradation, chromic acid oxidation and uronic acid degradation. Table II. 1 contains a list of common reactions that are utilised in carbohydrate researchlS. The part that MS played was mainly as a gas-liquid chromatography detector for characterising sugar alditol acetates. The quantification, the anomeric nature and the linkage position of the individual monosaccharides are relatively easy to obtain by chemical methods and NMR. The main limination to complete analysis is the sequencing of the polysaccharide. The reasons in the past for the low utilisation of MS in biomolecule research are: a ) ionisation and desorption can only be achieved on thermally stable and volatile molecules, b) the lack of high mass instruments. The first major breakthrough in desorption techniques was in 1969 with the field ionisation source, where the molecular ions of underivatised glucose and sucrose were observed in abundance^. - 8 -Table H I Wet Chemical Methods for the Extermination of Sequence in Carbohydrates Procedure 1. Partial hydrolysis with mineral acids, acetolysis methanolysis, autohydrolysis and hydrofluorinolysis. 2. Smith degradation. Mode of cleavage Different rates of hydrolysis of glycosidic linkages. The C-C bond between vicinal diols are oxidatively degraded. 3. Nitrous acid deamination 5. Degradation of hexuronic acid using lithium in ethylene diamine. 6. Enzymatic methods, including bacteriophage-borne enzymes. Selective degradation of amino sugars through ring contraction on deamination. Selective degradation of hexuronic acids. Selective cleavage at glycosidic linkages. Information A range of oligosaccharides are obtained, from di- to pentasaccharides. Size of oligosaccharide dependent on the linkage sequence of sugar residues. Generation of oligosaccharide without the amino residues. Generation of oligosaccharide without hexuronic residue. Generation of oligosaccharide without hexuronic residue. The reaction is performed on underivatised polysaccharide. Generation of oligosaccharides comprising of one or more repeating units. 4. Base catalysed ^-elimination Selective degradation of from hexuronate residues, hexuronate residues. - 9 -II.l The Mass Spectrometer All mass spectrometers consist of four components: a) the ion source, where the sample under investigation is ionised and transferred into the gas phase, b ) the analyser, where the molecular or fragment ions are mass analysed according to their mass to charge ratio, c ) the detector, where the resolved ions are detected, amplified and their intensity measured, d ) the vacuum system, which provides a stable environment for the above mentioned processes. Sometimes the sample inlet system can be quite elaborate, but it is normally considered as part of the vacuum system. By far the most important method for ion production is collision in a high vacuum between sample molecules and energetic electrons from an electron gun. In the past all mass spectral reference data were collected using electrons with 70eV energy. Nevertheless, in the last two decades a number of novel techniques were developed which complement electron ionisation2^. They can be broadly classified into four main groups: a) chemical ionisation (CI)2*, desorption chemical ionisation ( D O ) 2 2 and thermospray2^. b ) ionisation under strong electrostatic field. This includes field ionisation (FI)24, field desorption (FD)19 and electrohydrodynarnic ionisation (EHD) 2^. c ) ionisation induced by heavy particle bombardment. This includes secondary ion mass spectrometry (SIMS) 2 6 , fast atom bombardment ( F A B ) 2 7 , liquid S I M S 2 8 and Californium-252 plasma desorption ( 2 5 2 CfPD) 2 9 d) ionisation induced by photons, eg., laser desorption ionisation (LDI)30. Mass analysers can be divided into two groups. Static analysers utilise the momentum or energy dispersion properties of a magnetic field or an electrostatic field^l. The instrumental parameters are kept constant in time except for certain small changes required for recording the mass spectrum. Many different field combinations can be used. With dynamic analysers the separation is achieved by a time dependent parameter, for example, the differences in the time of flight of ions down a drift tube^2, or the time dispersion properties of a radiofrequency -10-field33. Ion detectors are mostly based on the conversion of the individual separated ion beams into a proportional electric current which can be amplified and recorded. II. 1.1 Ion Sources Ion production can be achieved by numerous methods and their relative importance is continously changing as new techniques come to light The nature of the sample and the kind of information sought dictate the choice of ionisation technique. The main objective is to produce as many ions as possible, with a beam composition which accurately represents the structural features of the sample. Ion sources have the following characteristics: a ) energy spread, b ) sensitivity, c ) ionic species production, d) memory and background, e) mass riUsaimination, f) ion current stability and noise. II. 1.1.1 Electron Ionisation (EI) This is the most utilised technique for generating ions in the gas phase. When electrons with kinetic energy E^in collide with neutral particles, the collision can be elastic or inelastic. For elastic collision the impact velocity is relatively low, hence no appreciable loss of energy of the electrons is detected. At higher impact velocity the neutral particles are excited and the collision becomes inelastic. However, the neutral particles in such an excited state are not charged and hence not detected. The life-time of such an excited state is approximately 10~8 seconds. At even higher energies, many configurational changes are possible, with ionisation and dissociation being most prominent. At above 25eV doubly charged ions and fragmentation can occur. The variation of ion intensity as a function of electron energy is the ionisation efficiency curve. The maxima of most ion efficiency curves are located approximately at 70eV. The minimium energy required to produce an ion of a particular compound is called the appearance . 11-potential. As ELMS utilises mostly 70eV energy electrons, the excess energy acquired by the molecule causes it to undergo single or multi- stage fragmentation, hence, the molecular ions are often not observed in the case of carbohydrate mass spectrometric analysis. For carbohydrates, EIMS can only be performed on volatile, low mass samples such as alditol acetates. Higher mass samples such as trisaccharides will either thermally decompose or fragment completely with loss of sequence information. II. 1.1.2 Chemical Ionisation (CI) This technique was first introduced by Munson and Field in 19662^. It is directly developed from fundamental studies of ion/molecule reactions. Ionisation is achieved by gas phase ion/molecule reactions. In essence, the reagent gas which is present in large excess is ionised by energetic electrons. This is followed by ion/molecule reactions between the primary ions and the neutral reagent gas which produces the CI reagent ions. Therefore the CI reagent ions have much lower energy than the ionising electrons, consequently CI is a milder form of ionisation, where frequently the molecular ion can be observed in abundance. The most widely used positive ion reagent gas systems are those which yield Bronsted acids (BH + ) . Species of higher proton affinity than B react by proton extraction. Methane gas is probably the most common reagent gas to date. At high pressures (10*1 Torr), methane gas produces a complicated spectrum, with C H 5 + (48%), C 2 H 5 + (41%), and C3H5 + (6%) ions being most abundant. The major ions are formed via the following routes: e ' + CH4 —> CH4+" + 2e" e ' + CH4 —> CH3+" + 2e- + H' e- + CH4 —> C H 2 + ' + 2e" + 2H" CH4+' + CH4 —> CH5+ + C H 3 * CH 3+" + CH4 —> C2H 5 + + H 2 - 12-C H 2 + ' + CH4 —> C2H4+" + H2 C H 2 + ' + CH4 —> C2H 3 + + H2 + H' C2H 3 + + CH4 —> C 3 H 5 + + H 2 Other common reagent gases include hydrogen, isobutane, methanol, water and ammonia. The techniques of Q M S and EIMS complement each other. By choosing the right reagent gas, molecular ions may be observed in the presence of fragment ions. Although CIMS can directly analyse underivatised oligosaccharides, the molecular weight limit of such samples is fairly low, hence derivatisation is normally carried out for CIMS analysis. Sequence determination of permethylated oligosaccharides by CIMS has been achieved. The most common reagent gases used in carbohydrate research are methane and ammonia^^. Another method of increasing the abundance of the molecular weight related ions has been to desorb material from external probes which are positioned direcdy within the CI reagent ion plasma. Since this desorption chemical ionisation (DCI) still involves heating and sample-surface interaction, pyrolysis remains a problem^. To reduce sample-surface interactions, thereby increasing the desorption of intact neutrals, inactive surfaces are used^5. II. 1.1.3 Secondary Ion Mass Spectrometry (SIMS) Mass spectrometric analysis of ions produced by bombarding surfaces with ions in the keV energy range is well established as a surface technique^. The early applications of SIMS were mostly in fundamental studies which utilised the fact that SIMS provides high absolute sensitivity for many surface components. Furthermore SIMS, unlike some other available techniques, could detect hydrogen and could determine the isotopic composition of elements on the surface^. SIMS was first shown to be a very sensitive technique for detection and identification of organic compounds in 1976^6. The main disadvantage of SIMS is the radiation damage caused by the exciting primary ion beam. This damage can be minimised by operating in the "static mode", in which the surface is bombarded at an extremely low primary ion current density. Hence, during -13-the analysis only a small probability exists for a surface area damaged by the impact of an ion to be struck by another primary ion during measurement. The secondary ions formed can be either atomic or molecular. Formation of molecular ions can be accomplished by three different processes: a) canonisation and anionisation (these result from chemical reaction during or following the concurrent production of metal ions and the liberation of organic molecules by sputtering or by thermal vaporisation), b) electron transfer (the resulting anion and cation radicals M"" and M + " , can be observed in SIMS spectra, sometimes associated with fragment ions which correspond to known products of their unimolecular dissociation), c ) the direct sputtering of organic molecular ions from the solid to the gas phased These ions arise by momentum transfer from the primary ions. All SIMS spectra show high abundances of pseudo-molecular ions that allow determination of molecular weight. Secondary ion yields are strongly dependent on the substrate and the properties of the primary beam, with the highest parent-like emissions obtained from compounds deposited on noble metal substrates and utilising a C s + ion beam. With few exceptions, low-energy bombardment of solids has used keV ion beams rather than neutrals, largely as a matter of experimental convenience since ion beams in this energy range are relatively easy to produce^. It is generally understood that the approaching ion is neutralised by long-range electron transfer as it approaches the surface so that the major effects of the impact would be the same for neutrals as for singly charged ions^O. II. 1.1.4 Fast Atom Bombardment (FAB) This technique differs from SIMS in two respects, a) the primary beam is a keV neutral beam produced by charge exchange neutralisation of an ion beam, b) the sample is introduced as a solution or suspension in a relatively involatile l iquid 2 7 . It is now well established that the presence or absence of charge on the impacting primary particle has little effect on the desorption - 14-process, but the use of the neutral beam is somewhat more convenient with magnetic instruments in which the ion source is at high potential. With the use of the liquid matrix the requirement for working in the "static mode" to avoid radiation damage on the surface is removed. The beam current density used is several orders of magnitude larger than those of SIMS and the spectra are stable for up to 30 minutes^0. The reason for this is that the matrix presents a mobile, constantly renewed surface to the bombarding beam, which continously replenishes undamaged sample molecules to be ionised. However, substantial radiation damage to the matrix does occur^l. One disadvantage of the liquid matrix is that a very high background is produced by the ionisation of the matrix and its radiation products. Also clustering of the matrix with sample ions is observed with formation of mixed cluster ions when impure samples or mixtures are analysed. Mass spectral data showed that the same secondary ions are produced whether the sample is in a liquid or solid matrix. The distribution of the secondary ions is representative of the composition of the liquid phase. If odd-electron species are just as stable in solution as even-electron species they can be detected just as easily. Although the ion production mechanism can be considered as identical in FAB and SIMS, a solid matrix usually leads to more fragmentation due to the process of primary energy dissipation. Solutes, molecules and ions are less tightly bound in liquids and the energy dissipation occurs mainly as translation instead of as vibration of the matrix42 This may lead to the ejection of larger solvated clusters where energy dissipation can easily be accomplished by excessive loss of solvent molecules without breaking intramolecular bonds^. Bombarding with lower energy primary particles has been found to increase the extent of fragmentation probably via the ejection of smaller clusters where the energy dissipation by solvent loss is less efficient. The size distribution of the ejected clusters, depending on the primary beam conditions, could be reflected in the degree of fragmentation. The second aspect in which solid and liquid matrix differ is related to the dynamic nature of the liquid sample. The liquid matrix must be considered to be mobile in the spectra-recording - 15-time scale. Hence, F A B can reflect the equilibria of the solution; it has been used to measure acidity constants, stability constants, and to monitor enzyme reactions44>45. A situation of particular importance occurs when the solvent-solute interactions are strong enough to induce heterogeneities in the distribution of species within the liquid matrix. If one solute behaves like a surfactant, a concentration gradient will develop, resulting in the enrichment of the surface with this compound. This behaviour is not confined to certified surfactants such as the long chain quarternary ammonium salts but also occurs with amino acids and dipeptides46,47 A crucial step in the FAB technique is the choice of a suitable matrix composition 48-50. The same solute can produce ions that vary in abundances by orders of magnitude as the matrix composition is altered. Some general requirements concerning the solvent properties of the matrix are summarised as follows: a) The sample must be soluble in the matrix. b ) Only low-vapour pressure solvents can be used in the vacuum of the mass spectrometer. Volatile solvents could in principle be employed with high pumping rates, provided that stable surfaces could be obtained on the time scale of recording a spectrum, c) The viscosity of the solvent must be low enough to ensure the diffusion of the solutes to the surface. d ) Ions from the matrix itself must be as unobstrusive as possible in the secondary ion mass spectrum. e ) The matrix must be chemically inert; if specific ion formation reactions are used to promote secondary ion yield the reaction must be reproducible. Both positive ion and negative ion F A B mass spectra can be utilised for sequence information. Mycobacterial 6-O-methyl-D-glucose polysaccharide, amylase-digested 6-0 methyl-D-glucose polysaccharides and O-methyl-D-glucose lipopolysaccharides have been sequenced by FAB51.52. -16-II. 1.1.9 Laser Desorption Ionisation (LDI) Soon after the advent of the laser several mass spectrometrists realised the potential of a laser mass spectrometer combination^. The nature of the radiation produced by the laser has some specific properties, which makes the laser a highly attractive source for sample desorption. Laser radiation is coherent, monochromatic, directional and intense. The high intensity of a laser beam enables sample vaporation to be on a very short time scale. The direction of the beam is defined by the geometry of the lasing system. Generally, a beam with a diameter of a few millimeters and a very low divergence is generated. This low divergence facilitates the operation of the laser at a relatively long working distance from the sample, without severe radiation loss. Furthermore, the coherence and directional properties together enable focusing of the beam to a very small spot, thereby increasing the power density. The monochromatic nature of the radiation leads to a narrow wavelength band. Although monochromaticity is an important factor in many other laser applications, no effective use of this property can be made in LDI. In principle, with a tunable laser very accurate scanning of wavelength over the absorption bands of the sample can be performed and the effects of selective excitations can be studied. Laser radiation is produced either in pulses or continously. The normal pulse range is from ~lus to ~lms. However, each pulse exhibits a fine structure and is effectively built up of a train of separate pulses 0.1 to lpvs wide. By the so called Q-switching technique part of the total energy can be released in one short pulse of ~10ns duration. Even shorter pulses in the pico second range can be obtained with an additional mode-locking technique. Although the total energy in a pulse is kept low, a lpJ pulse of 10'^ s focused on lO'^cm^ from a Q-switched Nd-Y A G laser leads to an instantaneous high power density of 10^W cm"2. As a result of the energy transferred to the sample material, heating, melting, vaporisation and ionisation processes take placed Under severe conditions large amounts of substrate can be vaporised leaving a crater in the solid. Two classes of power density can be distinguished. With low power density (< 10^W cm-2) evaporation of surface layers can be obtained, in the form of -17-intact neutrals and ionised molecules. This evaporation process is reproducible, and can be looked upon as a thin vapour layer moving into the solid at a steady state. In the high power domain (> lO^W cm"2) a high degree of irreproducibility is observed and mainly atomic and small molecular fragments in an ionised state are desorbed. After vaporisation, collision between neutral and primary ions in the plasma can lead to formation of secondary ions. The dominant process is canonisation by alkali ions^5. Due to the pulse nature of the ion current generated, special mass analysers are required with emphasis on the scan speed and signal registration. Spectrometers equipped with TOF detectors have been most frequently used because of their ability to record a complete spectrum for each laser generated ion pulse and high transmission. The major drawback of this system is the relative low dynamic range of the recording system due to the inherently short sampling time^6. Quadrupole mass filter and scanning magnetic sector type instruments have also been used, but are too slow to obtain a complete spectrum for each pulse^ 7 ^. Therefore high, repetitive rate lasers have been used recording only one mass peak per laser pulse. Magnetic analysers, both single and double focusing instruments, can be efficiently applied employing simultaneous ion recording techniques such as photoplates and channelplate-camera systems59,6G\ Fourier transform ion cyclotron resonance spectroscopy (FTICR), being an inherent pulse experiment, is well suited for LDLMS 6 2 > 6 3. LDI-FTICR feasibility was first demonstrated in 1982 and has been applied to simple carbohydrates 6 6 - 6^ II. 1.2 Mass Analysers The four most important parameters in mass spectromerric analysis of ions are: mass, charge, kinetic eneryg, and intensity. Fundamentally, any mass spectrometer only determines the mass to charge ratio (m/z), hence, if the actual magnitude of the mass is desired, the value of the charge must be determine by other means. Mass analysers have two objectives, a) to resolve an - 1 8 -ion beam of mass, m, from another beam of nearly the same mass, m+Am; b ) to maximise the resolved ion beam intensity. In order to determine the trajectories of the ions, their initial position and velocity are required. For the determination of the m/z ratio, the momentum, the energy and the velocity of the ions are required. II .1.2.1 Magnetic analyser If an ion of mass, m, with a charge, z, is accelerated from the ion source with a voltage, V, to the analyser, the kinetic energy acquired by the ion is E kin = Vz = 1/2 mv 2 1 The force, F, experienced by the ion moving with a velocity, v, across the magnetic field of flux density, B, is F B = zvB 2 According to the Left-hand rule, the ion always describe a circular orbit with radius r. The centripetal force, Frj is equal to F R . Frj = mv 2/r = F R = zvB 3 Hence, mv = rzB 4 Equation 4 shows that the magnetic field acts as a momentum analyser. A momentum distribution may be obtained by placing a photoplate in the exit plane. A more important conclusion can be drawn by combining Equation 1 and 4 to give m/z = B 2 T 2 / Z V 5 Thus, the momentum filter can be converted to a mass filter if the ions are of equal energy and charge. The constancy of ion energy is provided by passing the ion beam through an electrostatic analyser after emerging from the ion source. -19-II. 1.2.2 Electrostatic analyser Using the same principle as with the magnetic analyser, when an ion is subjected to an electrostatic field, E, it describes a circular motion, with the centrifugal force, Fr> equal to the electrostatic force, Fg. F c = mv2/r = Fg = zE zE = mv /^r 6 Combining Equation 1 and 6, r = 2V/E 7 Equation 7 shows that the electrostatic analyser does not analyse mass, it only acts as an energy filter, since any variation of V results in a path of different radius. In other words, ions of any mass will follow the same path for a given accelerating voltage and field strength. The electrostatic analyser is usually used as an energy filter in double-focusing instruments. n. 1.2.3 Time-of-flight analyser (TOF) In a beam of ions of equal energy the heavy ions will travel slower than the light ions, hence, a short pulse will disperse as it moves down a long field-free drift tube into groups of the same mass. The relationship between the time of flight and the m/z ratio of the ions is related to the velocity of the ions passing through an electrostatic analyser. The velocity acquired is, v = (2zV/m)!/2 g And the time of flight, Tf, down a drift tube of length, d, is T f = d/v = d ( l / Z V ^ m / z ) 1 / 2 9 Equation 9 shows that the time of flight of various ions is proportional to the square root of their respective m/z ratios. -20-T O F instruments provide essentially unlimited mass range and determination of nearly every secondary ion. This latter characteristic allows extremely low primary doses for SIMS to be used without sacrificing senstivity. Perhaps, the most important attribute of the TOF method is that it only requires the ions to be stable for the time taken to transverse the accelerating field. As a result, many ions that would go undetected in the scanning instruments may be detected as intact molecular ions by TOF spectrometers. TOF instruments are mosdy used for SIMS, 252cf_ PDMS and LDIMS experiments. II. 1.2.4 Fourier transform ion cyclotron resonance spectroscopy (FTICR or FTMS) In a uniform magnetic field, B, a moving ion of charge, z, and mass, m, will be subjected to Lorentz force, which is perpendicular to the direction of the ion, hence, constraining it to a circular orbit with its radius proportional to its velocity. This orbit will have a characteristic cyclotron frequency, co, and the orbital motion can be described as (0 = zB/m 10 The ICR spectrometer measures the cyclotron frequencies of an ensemble of ions of differing masses. The ICR experiment is divided into two parts, a ) Excitation of cyclotron motion by an alternating electric field induced by a capacitor. In a FTICR spectrometer, all the ions are excited together by using a fast frequency sweep from a radiofrequency oscillator, b) Detection of the excited cyclotron motion. The cyclotroning ions induce an alternating current in an external circuit, the voltage induced across this circuit is the ICR signal^l. This ICR signal is then digitally sampled and Fourier transformed to give a mass spectrum. An FTICR spectrometer has certain advantages over the conventional scanning spectrometer. These are a) high speed^ b) high sensitivity63, c ) ultrahigh mass resolution^, d) high mass ranged e) ease of operation, f) ease of mass calibration, g) suitability for LDI experiments^ h) and the ability to carry out tandem experiments in a single ICR cell^l. -21 -II. 1.2.5 Quadrupole mass filter The quadrupole mass filter operates on the different responses of differing masses to a radio-frequency electric f ield 7 2 . Four stainless steel rods are placed parallel to each other and opposite pairs are electrically connected. Both DC and A C fields are applied to the rods, and mass measurements are done by altering the A C and D C potentials but keeping their ratio constant. Quadrupole instruments use a relatively low source voltage, a fast scan rate and can operate at relatively high pressure as compared to conventional scanning spectrometers, thus are well suited for G C - M S 7 3 . II.2 Methylation procedure Derivatisation is often employed in mass spectrometric analysis to increase sample volatility. Permethylation and peracetylation are two common derivatisation techniques used in carbohydrate research. Although acetylation reactions are relatively easy to carry out, they increase the molecular weight of the sample approximately two fold. Thus, for oligosaccharides with more than four monosaccharide units methylation procedure is normally utilised. The Haworth procedure involves repeated reactions with dimethyl sulphate and sodium hydroxide7^. The partially methylated polysaccharide obtained is then treated with silver oxide and methyl iodide to give the permethylated product7^. This method due to Purdie & Irving was further improved by Kuhn who used N,N-dimethylformamide to increase the solubility of the polysaccharide7^. However, the most convenient procedure was developed by Hakomori 7 7. Complete methylation can be achieved by treating the polysaccharide with the strong base sodium methylsulphinyl methanide (sodium dimsyl) and followed by addition of methyl iodide. If undermethylation is suspected then complete methylation can be obtained by a subsequent Purdie methylation since a second treatment with the Hakomori procedure will cause fJ-elimination of any uronic acid residue present. Hakomori methylation leads to esterification of any uronic acid -22-and pyruvic acid acetals present and also to N-methylation of the acetamide group of amino sugars. The use of a strong base in methylation causes removal of base-labile substituents (eg. acetates), resulting in loss of substitution information. Nevertheless, O-acyl groups can be detected using the Prehm methylation procedure, where the polysaccharide is dissolved in trimethyl phosphate and then methylated with methyl trifluoromethane sulphonate and 2,6-di-(tert-butyl) pyridine as proton scavanger78. The usual cause of undermethylation is incomplete dissolution of the sample. Some polysaccharides like cellulose, are insoluble in DMSO; these can be methylated with a N-methylmorpholine-N-oxide (MMNO) - DMSO mixture7^. The methylated product is removed by partition between water and chloroform. Infra-red spectroscopy is used to check for completeness of methylation (absence of hydroxyl absorption at 3600cm-1). At present a complete structure determination of oligosaccharides by mass spectrometry alone is not possible. However, mass spectrometry coupled with other techniques like methylation analysis will provide not only the sequence of the oligosaccharide but also the sugar composition and linkage positions8^8*. In methylation analysis the free hydroxyl groups of the oligosaccharide are protected by etherification, which on hydrolysis can be distinguished from the linked position. This is accomplished by separation of the sugars as their partially methylated alditol acetate derivatives by G C and identified by GC-MS. II.3 Bacteriophage degradation of capsular polysaccharides. Bacteriophages are serotype-specific viruses that first infect and then propagate in the bacterium thus destroying the host c e l l 8 2 - 8 4 . They are often designated by the Greek letter (p, followed by the number of the serotype that acts as the host strain. The adsorption of the phage to its receptor is highly specific, where the exopolysaccharide acts as a receptor for phage -23-binding8^. The phage associated endoglycanase depolymerises this exopolysaccharide, and subsequent penetration of the host cell by the phage is followed by the release of viral D N A 8 6 . The use of these phage-borne enzymes for depolymerisation of bacterial capsular polysaccharides allows the isolation of selectively cleaved oligosaccharides corresponding to one or more repeating units 8 7. Furthermore, acid- and base- labile non-carbohydrate substituents remain intact on the oligosaccharide, which is usually difficult to achieve using chemical means of degradation. Thus phage degraded oligosaccharides are excellent substrates for mass spectrometric studies. -24-MASS S P E C T R O M E T R I C A N A L Y S I S O F P O L Y S A C C H A R I D E S -25-HI.I Laser Desorption Ionisation Fourier Transform Ion Cyclotron Resonance Spectroscopy (LDI-FTICR) of capsular polysaccharide of Klebsiella serotype K44 111.1.1 Abstract The structural sequence of the de-Q-acetylated capsular polysaccharide from Klebsiella K44 was determined by LDI-FTICR. Glycosidic bond cleavages and ring rupture fragmentations occurred at both ends of the oligosaccharide. From the ring rupture fragments the reducing end and some positions of linkage were determined. The proposed sequence shown below, agreed with the known structure suggesting that LDI-FTICR is a powerful technique for analysing underivatised oligosaccharides. HexA 1 ^deoxyHex 1 ?deoxyHex 1 3 Hex * ?Hex 111.1.2 Introduction The structure of the capsular polysaccharide of Klebsiella K44 has been investigated previously88,89. it i s a linear pentasaccharide repeating unit consisting of a glucuronic acid residue, two glucose residues, two rhamnose residues and an O-acetyl group. FABMS data have also been reported for both the reduced, underivatised and reduced, peracetylated bacteriophage degraded oligosaccharide^0. This particular polysaccharide being a relatively simple linear structure was investigated to assess the feasibility of using LDI-FTICR to study bacterial polysaccharides, and for comparison with other well established ionisation techniques^!. 111.1.3 Experimental The positive ion LDI-FTICR experiment was run by Dr. David Weil of Nicolet Instrument -26-Corp. using the Nicolet FTMS-2000 Fourier transform mass spectrometer with laser desorption ionisation capability. The experimental conditions are similar to FTMS-1000 experiments published previously, with the exception of the dual cell analyser67"6^. FTMS instruments are very sensitive to pressure burst, thus after the laser pulse a delay of ten seconds was incorporated into the experimental pulse sequence. Three separate experiments were carried out: a) Native polysaccharide doped with potassium bromide. b) De-O-acetylated, bacteriophage degraded oligosaccharide doped with potassium bromide. c) De-O-acetylated, bacteriophage degraded oligosaccharide doped with sodium bromide. III. 1.4 Results and Discussion Data analysis incorporated the following assumptions: a ) Peaks with a variance greater than 30ppm from the calculated value were considered to be spurious peaks. Al l the calculated vs observed values for the assigned peaks agreed within 20ppm, with the exception of m/z 553.1681 (27.26 ppm error). b) Background noise level was assumed to be 2.5% relative to the base peak. c) Extensive cationisation of pseudo-molecular and fragment ions occurs. As observed, all the assigned structures are either sodiated or potassiated. d) The ICR cell retains memory of the previous sample. Thus, high intensity sodiated ions will have potassiated counterparts and vice versa. e) Extensive dehydration occurs in pseudo-molecular and fragment ions. f) Two distinctive fragmentation routes were expected. i ) a Cleavage of the glycosidic bonds (Figure m.1.1). The glycosidic oxygen can either be retained on the reducing end fragment, oq fragmentation, or be retained on the non-reducing end fragment, 0:2 fragmentation. With a i fragmentation, an oxonium ion fragment is formed (Figure HI. 1.2). Subsequently, the reducing end fragment picks up a -27-hydrogen, is cationised and detected. However, for 0:2 fragmentation, the non-reducing end fragment is detected instead. The oxonium fragment may not be observable as canonisation is normally the dominant process. O Figure m.1.2 Oxonium ion formation from cq fragmentation route, i i ) P Fragmentation of the ring. Although p* cleavages complicate the spectrum, they have certain advantages. There are also two routes for P fragmentation (Figure HI. 1.3). For pi fragmentation, the positive charge resides on the non-reducing end, while for P2 fragmentation the positive charge resides on the reducing end. This implies that if both Pi and P2 type fragments can be observed the reducing end can be determined. P2 Fragments contain C - l and C-2, while Pi fragments hold C-3, C-4, C-5 and C-6. Figure III. 1.3 Pi and P2 fragmentation routes. In the sodium chloride-doped oligosaccharide experiment (Spectrum III. 1.1) both a and P fragmentations were observed, with molecular weight information indicating a pentasaccharide structure with one hexose uronic acid, two hexoses and two deoxyhexoses. Both the potassiated and sodiated, dehydrated molecular ions were observed, m/z 831.2297 (8.6%) and m/z 815.2431 (24.9%). Loss of carbon dioxide from the sodiated, dehydrated molecular ion was evident, indicating that the oligosaccharide contains an uronic acid, m/z 771.2657 (9.7%). The loss of carbon dioxide was not observed in the corresponding F A B spectrum^O (Spectrum III. 1.2). By far the most prominent ions arose from glycosidic bond cleavage. The base peak of the spectrum is m/z 639.2097 (100%), indicating a sodiated, dehydrated ion containing two hexoses and two deoxyhexoses. The potassiated counterpart is at m/z 655.1926 (26.0%). The non dehydrated ions were also observed, but with lower intensities. These were a sodiated ion at m/z 657.2223, (47.4%) and a potassiated ion at m/z 673.2060, (18.5%). Subsequently, consecutive loss of two 146 amu indicated that the deoxyhexoses are adjacent to each other. Figure HI. 1.4 shows some structures arising from a fragmentation. Hence, a fragmentation provided four pieces of sequence information. These were: i ) An uronic acid residue is either at the reducing or non-reducing end. i i) The two deoxyhexoses are adjacent to each other. iii) The two hexoses are adjacent to each other. -29-HexA~deoxyHex~deoxyHex~Hex~Hex deoxyHex—deoxyHex—Hex-Hex O deoxyHex—deoxyHex—Hex O deoxyHex—Hex—Hex O Figure HI. 1.4 Some structures derived from a fragmentation. -30-iv) A hexose unit is directly linked to one of the deoxyhexoses. Thus from a cleavage, four possible sequences can be assigned to the oligosaccharide. The structures are: HexA—deoxyHex—deoxyHex—Hex—Hex HexA—Hex—Hex—deoxyHex—deoxyHex deoxyHex—deoxyHex—Hex—Hex—HexA Hex—Hex—deoxyHex—deoxyHex—HexA The most prominent p cleavage is at m/z 537.17909 (80.5%). This peak indicates a sodiated structure consisting of two hexoses, a deoxyhexose and a P2 type fragment. From this piece of evidence and the information obtained from a cleavage, the oligosaccharide was determined to be of the following sequence: HexA—deoxyHex—deoxyHex—Hex—Hex P Cleavages occurred at four of the residues. The only residue that did not undergo ring rupture is the middle deoxyhexose. With P2 type fragmentation, one can distinguish a 2-linked residue from other linkage positions as a 2-linked residue will contain one less oxygen in P2 type fragmentation. P2 fragmentation of the uronic acid residue gave a C2H3O2 fragment while P2 fragmentation on the adjacent deoxyhexose generated a C2H3O fragment. Thus the deoxyhexose adjacent to the uronic acid residue was determined as a 2-linked residue. Pi Type fragmentation was observed for the hexose adjacent to the deoxyhexose. The non-reducing end sugars contained an additional unit corresponding to C4H9O3. The reducing end hexose also underwent Pi type fragmentation. In contrast to the other hexose Pi fragmentation the additional unit corresponded to C4H7O2. This may indicate that a C4H7O2 attached unit corresponds to a 4-linked residue and a C4H9O3 attached unit corresponds to a 3-linked residue. A 6-linked residue probably will not undergo P i fragmentation, thus a C4H9O3 containing fragment is likely to be a 3-linked residue. However, dehydration of high mass oligosaccharides in mass spectrometric analysis is commonly observed, thus further model compounds are -31-required to confirm this observation. Therefore two positions of linkage were determined for this oligosaccharide. Figure LTJ.1.5 shows structures derived from P cleavages. The other unassigned "high intensity" peaks are satellite peaks. Table LTJ. 1.1 contains a list of C-l3, H-2 and 0-17 satellites comparing the observed intensities to calculated intensities. There is one possible carbohydrate peak without structural assignment, m/z 391.1241 (15.1%). This peak is likely to be a decomposition product of a trisaccharide unit, however the accurate structure is not known. No oxonium ion related peaks were observed. This may either be due to suppression by cationisation or the oxonium ions may be consumed by ion/molecule reactions during the ten seconds delay between ionisation and detection. Table LTI.1.2 contains a list of complete structures arising from the oligosaccharide by LDI-FTICR. The proposed sequence derived from LDI-FTICR is: HexA 1 DeoxyHex 1 ^deoxyHex 1 3jjex 1 ?Hex The corresponding F A B spectrum gave no indication of ring rupture (Spectrum III. 1.2). Thus for a complete sequencing the oligosaccharide has to be reduced. Both of the potassium bromide-doped experiments gave a and P fragmentation and some sequence information, however they are not detailed enough for structural assignment. This is expected because the native polysaccharide has an extremely high molecular weight (Spectrum III. 1.3). In the case of the oligosaccharide this may be attributable to the strong binding energy for potassium attachment (Spectrum HI. 1.4). By combining MS techniques with methylation analysis more structural information can be obtained. Previous methylation data indicate that the oligosaccharide is comprised of: a 4-linked glucose, a 2-linked rhamnose, a 3-linked rhamnose and a 3-linked glucose8^. As uronic acids are not detected by methylation analysis, only the hexoses and deoxyhexoses are determined. Thus the structure sequence of Klebsiella K44 phage degraded Oligosaccharide obtained by LDFTICR and methylation analysis is: H e x A l - — 2 R h a l 3Rh a l 3GIc1 ^Glc -32-cooH H O - , ; H O - , ; / — ^ i H U i — v r i U / — ^ / OH H O V J u , v l OH* . O H O H \ / O H ' O H O R-l fragmentation C O O H _ H 0 - | possibly 4-linked: C 4 H 7 0 2 7 C 4 O H H O J "^J O H n u - i 0 \ / O H 0 (5-2 fragmentation not 2-linked: CjHiOo I , O H (3-2 fragmentation 2-linked: C 2 H 3 0 O C O O H JJQ £— O H O A— 0 (3"1 fragmentation V C H 3 \ K H 3 ^ 3-linked : C4H0O3 O H \ H O J ° s l O H Figure UL 1.5 Some structures derived from p-fragmentation. Table III. 1.3 Satellite peaks of the intense "structural" peaks -33-Observed Observed Calculated Mass Intensity Intensity 816.2537 8.8 9.7 700.2438 11.2 12.5 682.2036 7.2 8.3 674.1628 5.5 5.7 658.2325 11.6 14.6 656.1817 6.6 8.0 640.2204 27.6 29.6 598.22976 8.4 8.3 554.1746 6.9 8.0 538.1807 19.5 20.7 496.1667 6.6 6.4 494.1551 7.0 6.3 -34-Table in. 1.2 Assigned fragmentation structures of Klebsiella K44 Observed Calculated Relative Mass Proposed Mass Mass Intensity Error Structure amu amu ft ppm 831.2297 831.2167 8.6 15.64 (2Hex,2deoxyHex,HexA)-H20+K 815.2431 815.2428 24.9 0.36 (2Hex,2deoxyHex,HexA)-H20+Na 771.2657 771.2530 9.7 16.50 (2Hex,2deoxyHex,HexA)-H20-C02+Na 741.2286 741.2430 6.0 -18.58 (2deoxyHex,Hex,HexA)+C4H702+Na 715.2131 715.2058 11.1 10.21 (2Hex,2deoxyHex)+C2H302+K 699.2308 699.2318 37.4 - 1.40 (2Hex,2deoxyHex)+C2H302+Na 697.2088 697.1952 5.9 19.39 (2Hex,2deoxyHex)+C2H302-H20+K 681.2240 681.2213 25.0. 3.95 (2Hex,2deoxyHex)+C2H302-H20+Na 673.2067 673.1952 18.5 17.08 (2Hex,2deoxyHex)+K 657.2223 657.2213 47.9 1.59 (2Hex,2deoxyHex)+Na 655.1926 655.1846 26.0 12.22 (2Hex,2deoxyHex)-H20+K 639.2097 639.2107 100.0 - 1.53 (2Hex,2deoxyHex)-H20+Na 613.1672 613.1741 13.4 -11.19 (2deoxyHex,HexA)+C4Ho03+K 597.2062 597.2001 32.8 10.10 (2deoxyHex,HexA)+C4H903+Na -35-Tablem.1.2 Continued Observed Calculated Relative Mass Proposed Mass Mass Intensity Error Structure amu amu % Epm 579.1959 579.1896 9.1 10.89 (2deoxyHex,HexA)+C4Hg.03-H20+Na 553.1681 553,1530 31.2 27.26 (2Hex,deoxyHex)+C2H30+K 537.1790 537.1791 80.5 - 0.17 (2Hex,deoxyHex)+C2H302+Na 519.1643 519.1685 5.6 - 8.03 (2Hex,deoxyHex)+C2H30-H20+Na 511.1629 511.1634 17.6 0.94 (2Hex,deoxyHex)+Na 509.1300 509.1267 6.9 6.40 (2Hex,deoxyHex)-H20+K 495.1692 495.1684 27.4 1.59 (2deoyHex,Hex)+Na 493.1526 493.1526 27.2 - 1.00 (2Hex,deoxyHex)-H20+Na 477.1580 477.1579 13.7 0.28 (2deoxyHex,Hex)-H20+Na 365.0984 365.1055 6.5 -19.28 (2Hex)+Na 349.1127 349.1105 6.9 6.24 (Hex,deoxyHex)+Na 347.0981 347.0949 18.4 9.33 (2Hex)-H20+Na 333.1220 333.1156 9.9 19.21 (2deoxyHex)+Na 315.1083 315.1050 6.3 10.37 (2deoxyHex)-H20+Na -36-In order to identify the uronic acid residue, it has to be reduced to its corresponding hexose, which is then be characterised by methylation analysis. III. 1.5 Conclusion LDI-FTICR is well suited to the investigation of underivatised oligosaccharides. This technique is probably more versatile than other well established mass spectrometric methods for the sequencing of oligosaccharides, primarily due to the fact that it can provide linkage information on underivatised samples. The proposed sequence agreed with previous results obtained from wet chemical methods. HexA 1 2(j e 0 X yHex 1 ^deoxyHex 1 3jjex 1 ?Hex Combining methylation analysis and LDFTICR the hexoses and deoxyhexose were also characterised. The proposed structural sequence is: HexA 1 ^ h a 1 3Rh a l 301c 1 -—*Glc -37-Nominal mess Measured mass Rel . i n t e n s i t y Abs. i n t e n s i t y J 69. r.mu 169.04526 amu 4.1 X J .4507 171 . amu J71 .0612B amu 5 . 7 X 2 . 0 1 1 0 5 73. c'fflU 173.07937 amu 9.6 V 3.3947 IBS. amu 164.95932 amu 4.4 X 1 .5724 amu 1B6.22371 amu 25.4 X 8.9756 i e 7 . amu 167 .513199 amu 3.6 X 1 .2B60 169. amu 169.05701 mmu 5.© X 1.772* 5 9 5 . emu 195.12539 amu 4.4 X 3 .5497 199. amu 19^.0974 3 afliu 11.6 X 4.1066 2*1 . amu 201 . 1 1 2*1 amu 5.4 X .1 .9 J 75 2:3. amu 2 1 3 anu 3.0 X 1,0715 2 \ z . amu 2i5.3B970 amu 9.4 X 3.31B3 I'.'.7. amu » 217.10566 amu 14 . 5 X Jj. J 1 * 1 227.. amu 225.07175 amu 3.4 X 1 .1975 227. emu 2 7 7 .09090 amu 4.5 X 1.5750 amu 229.1! €)5>4 amu 4.1 X 1.46&B J x . amu 233.065B1 amu 3 . 3 V. 1.3B0;i 2 3 3 . amu 233.06301 amu 6.0 X 2.393B 237. amu 237.137236 amu 3.B X 1 .35(52 241 . an:.i 24 1 .10242 amu 3.4 X 1.1B90 243. tfl u 243.11926 amu 6.3 X 2.9299 245. amu 24 5.133 25 e r u 4.7 X 1 .6671 I T 7. amu 257. 5 1 193. amu 4 .2 X I.4935 261 . a m u 261 .13759 amu 9 . 7 X 3.4263 269. amu 26B.57269 amu 3 . 2 X 1.1423 269. i l l l U 269.09B57 amu 5.4 X 1.9242 273. amu 273.C9307 amu 3.2 X 1.1296 2B7. erru 2B7.15124 amu 7.3 X 2.5913 305.. amu 305.16046 anu 4.3 X 1.5155 315. amu 315.J0831 anu 6 . 3 X 2.2390 37*. emu 32ft .0078^. anu 4.1 X 1.4667 321 . amu 321 .07143 amu 3 . 7 X 1.3242 331 . amu 333.12542 amu 4.4 X 3.5437 333. amu 333.32201 amu 9 . 9 X 3.4973 3 4 7 . amu* 347.09B3 1 an 1. IB.4 X 6 . 5 : ? ' . 3 4 9 amu 349.1127K; amu 6.9 X 2.*375 3 f . V . amu 365.O9B40 amu 6.5 X 2.30E3 ft 69. amu 369.1*369 amu 5.5 X 1.9512 369. e*»u 369.3674 3 amu :4.5 X 5.1210 371 . amu 371 .1165B arr.u 6 . 3 X 2.2142 371 . emu 371.34362 amu 3 0 . 3 */ 3.6522 391 . amu * 393 .224J4 arr.u "».*\ 15.1 X 5.3510 Ad.. amu 405.146)69 amu B . 4 X 2.9B61-4 53 . amu 451.14415 amu B.O X 2 . e i 3 5 47?. emu. 475.14552 amu 4.0 V J. .4 365 4 7 7 . amu 477.15B01 amu 13.7 X 4.B353 493. 493 . 1 5 2 5 6 emu -*A<#+ 27.2 X 9.623V 4 9 4 . amu 494.155e>6 amu 7.0 X 2.4601 49!!.. amu 495.16922 amu 27.4 X 9 r6976 4 96. amu 496.16673 amu 6.6 X 2.3365 5<?<9. amu 509.1299B amu 6.9 X 2.43C2 511. amu 511.16267 amu 17.6 X 6.2319 53 9. amu 519.16432 ami. 5.6 X 1.97E1? 524). «'»u 520.11560 amu 3.1 X 1.0B13 32.1 . A m u £21.16682 amu 3.3. V J .096V3 Spectrum HI. 1.1 Tabulated LDI-FTICR spectrum of Klebsiella K44 de-0_-acetylated,phage degraded oligosaccharide doped with sodium bromide -38 -Ti~ 3 . • / V U 523 .12354 •mu 7.2 X 2.559=V 535. •mu 535.17617 6 .4 X 2.2677 r.37. amu» 537 .17B97 •mu 6fr.5 X 26.4626 « « W 53B.1B<>>6#> •mu 19 .5 X 6.B93E 539. •mu 539 .2*732 •mu 3.4 X 1.1666 553. •mu 553.16B0B •mu 31.2 X 11.C25; •mu 554.17456 •mu 6.9 X 2.45^1 577. • P i l i 577.1B706 • mu 3.3 X 1.1696 579. •*u 579.19587 •mu 9.1 X 3.2124 5*:.. •»•.» 595.182** 5.B X 2.C59? 597. « u 597.3*616 ati.j 3?.B X 11 •fB 'ne 59P. M U 59B.22976 •mu e.4 X 2.9670 6«"? . * t i u 6*9.1656* •mu 4 .3 X 1.5266 6i3 . emu 613.16720 nmu 13.4 V 4.7422 614. •ma 6 1 4 , » B 3 2 3 nmu 3.7 X 1 .296(5 6!4. •mu 614.4779P •mu 3.6 X 1.2579 621 . *mu 621.19727 •mu 5.2 X 1.6257 6?9. • ( T I L * 63V.2*971 •mu ief-.B X 35.3*91 *>4e>. •mu 64f>.22fl39 •mu 27.6 X 9.7693 641 . •mu 6*1.22^91 •mu 16.2 X 3.6136 651 . • r:y 651.20490 emu 7.«> X 2.4573 653. 653.22BB7 •mu 3.6 X 1.35*3 655. • I V U 655.19264 •mu 26.0 X 9.1743 fef.fr. •mi'. 656.1fll65 e n j 6.6 V 2.3431 657. emu 6 5 7 . 2 2 2 3 © •mu 47.4 X 16.7411 65B. amu 6 5 B . 2 3 r 4 6 • n-t. 1 1.6 X 4.1152 666. •mu 665.6*59/: •mu 5.3 X 1.8625 669. emu 669.16592 • mj 9.«> v. 3.16*7 673. •my 673 .2*67«> •mu 18.5 X 6 .5iee 674 . 674.1627*, •mu 5.5 X 1 .933?> 677. •mu 676.BB3*>4 •mu ?.4 X 1.1963 6B: . 6E1 .?2395 amu 25. fc X 8.B257 6S2. • r.u 6B2.2C36? « 9 U 7.2 X 2.5536 663. • OH' 6E3.23173 • Ttl 4.7 X 1,64«?7 69-s. • T.i.. 603.6635', 3.6 X 1 .2716 605. *1ll> 6B=.231 IB e.r.j < .5 X 1 .579'.-• T » U 697.2e>B76 •mu 5.9 X 2.«>B2? 6?9 . • («!.•.* 699.2308 'i emu "7.4 V 13.27-7 7C-P. a t T L 705-.24379 •mu 11.2 * 3.9*29 7fc: . •mu 7dl .2326'! 4 .7 V. 1.6777 7J1 . 711 .7e&C»3 a" j. 6.7 X 2.3R32 712. 712.66*52 •mi- 3 .8 X I .3293 "715. •mu 715.21386 •mu 11.1 X 3 .9 7J6. • ftlLt 716.24395 t mu 4.3 X 1.51«>3 74: . • I T i U 741.22662 • mu 6.0 X 2.1CSS •741 . (.mu 744.4&9B9 • m.i 3.6 X 1 .259«> 753. • rru 753 .32793 •mi'. 6.6 X 2.3365 7 7 * . • i . i 776.1919* • mu 3.4 X 1 .1951 771. e r j * 77;.26569 amu 9 .7 X 3.423 1 £ 1 3 . •mi.. 813.2S245 er>:u 6.4 X 2.2556 B : S . « » u ^ B:C.2430B 24.9 X 6.6P03 eifr. •mu B'i fe.2537v •mu B.B X 3.1914 B . 7 . amu Bi7.23439 •mu 4.B X 1.7*52 tte. •mu 816.27903 •mi. 3 .B X i .3397 e29. •mu B29.2B2B3 •mu 4 .2 X 1.4795 B31. • f l > U 831.22972 •mu 8.6 X 3.0363 632. •mu 632.19*67 •mu 3.6 X 1.2646 057. •mu B57.24B7B 6.S X 2.411* B5B. W u B5B.271B5 •mu 3.9 X 1.3694 673. •mu B73.27595 •mu 4.9 X 1.7435 967. •mu 966 .7*565 •mu 3.5 X 1.23«>fr 976. •mu 977.54372 •mu 3.3 X 1.155b 998. emu 996.11365 •mu 8.1 X 2 . 67 :7 Spectrum TH. 1.1 Continued -39-Spectrum LIT. 1.2 F A B spectrum of reduced Klebsiella K44 phage degraded native oligosaccharide in glycerol matrix^O. Nominal mast Measured mass Rel . i n t e n s i t y Abe. i n t e n s i t y emu 107.53605 amu 5.5 X 953.7354 m 123. amu J23.37O90 amu 7.7 X 1.3209 124 . »nu 123.5BB92 amu 7.2 X 1.2507 1B7. amu 1B7.03991 amu 10.1 X 1.7399 1B9. a flu 109.05606 amu 12.2 X 2.107R 199. «rb 199.09491 amu 10.0 X 1.7213 2C1 . t»u 2*1.65704 amu 6.1 X - 1 .05B3 201 . eir .u 201 .1 !297 •mu 4.4 X 766.41B5 m 203, amu 203.07209 amu 15.5 X 2.6701 213. amu 212.0S55B asY_i 5.0 X 661 .93B5 m 215. amu* 215.06972 •mu 100.0 X 17.2622 216. amu 216.5*7337 a<*u 9.1 2 1.5692 217. amu*' 217.08566 amu 24 .6 X 4.2544 219. amu 219.0B77B amu 4 .7 % B10.913J m 277. 227.07051 tmu 6.4 X 1.0975 229. amu 229 .06449 •mu 23.4 X 4 . 0 3 4 4 231 . •mu 231.06571 amu 17.B Y 3.0732 233. amuV" 233.0B149 amu 54.0 X 9.3197 2C<4. amu 234.0B597 amu 4.7 X 612.3169 m 241. amu 241.0B601 amu 6.0 X 1.0327 243. 243.C7456 amu 7.6 X 1.2117 243. amu 243.13334 amu 4.5 X 769.5313 m 217. amu 245.0B062 emu 5.7 X 987 .8541 m 247. amu 247.09667 amu 9.5 X 1.6439 nr « amu 250.9697B arm 6.6 X 1.1461 253! amu 252.96BJ9 amu 7.0 X 1.2028 21:7. amu 2E7.0B41B amu 5.6 X 962.15S3 m 239. 259.09459 amu 46.6 X B.0524 260. emu 26<2.09773 «'nu 6.0 X 1.0360 261 . amu 261.10947 amu 25.6 X 4.4144 271. amu 271 .12119 amu 5.4 X 93B.9649 n, 273. amu 273.11C6B amu 7.9 X 1.3623 275. amu 275.6B757 amu B.6 X 1 . 4 e c : 277. amu 276.72632 amu 7.9 X 1.3577 277. 277 . 1*655 r.mu 17.9 X 3.0846 279. *<*u k*V v 27B.724B7 amu 4.7 X 618.6646 m 393. amu 303.11957 amu IB.4 X 3.1782 305. amu 305.13B4B amu 6.2 X 1.0662 321 . amu 321.13393 amu 7.3 X 1.2527 347. amu 347.141S0 amu 5.0 X 857.6492 m 369. emu 369.34395 amu 7.4 X 1 .279C" 371 . amu 371 .30*46 • mu 5.6 X 962.1583 m 395. 394.60679 •mu 4.9 X 642 .6514 m 45)7. amu* 407.09273 amu 4.9 X B*i .3086 m 537. amu 527.I6B03 amu 10.4 t 1.7937 553. amu* t.53.14176 amu 3';-. 1 X 5.1962 5S4. anu 554.14424 amu 7.5 X 1,29'.7 555.14141 amu 6.7 X 1.1589 amu 595.15199 emu 9.5 X 1.6393 £ , 1 1 . amu 611.13072 amu 4.7 X 608.8379 n an.u * 6r..5.1B5B7 amu 4 .B «. A B23.9136 Hi 667. amu 667.15919 •mu 4.5 X 782.3487 m 697. amu 697.1766E amu 5.6 X 964 .4776 m 715. emu* 715.1547B amu 4 . 1 X 713.9283 tr-7?j7. amu* 757.223B0 amu 9.1 X J . 575*9 Spectrum UJ. 1.3 Tabulated LDI-FTICR spectrum of Klebsiella K44 native polysaccharide doped with potassium bromide -41-Nominal nasv % n»»5ur«tJ »•»« Hel . int»n»;ty Ab*. i n t e n s i t y 1G>9. amu 169.05151 •mu 7.6 X 2B9.550K m 391 . amu 291 .0528* •mu 5 .1 X 194.2139 m 203. ana 2*?.06922 •mu 10.7 X 406.433) m 213. •mu 213.05303 amu 7.1 X 270.6909 m 215. amu 2IS.06726 •mu 100.0 X 3.600< 216. amu 216.07244 amu 9.6 X 370.7886 m 217. amu 217.0D<96 amu 41 . 0 X 1.5573 218. a* iU 218.08901 amu 5 . 7 X 215.27;? m 2J9. •mu 219.00291 amu 4.9 X 186.21?:- n: 223. amu 223.14454 amu 8.5 X 322.7539 m 224. amu 224.17698 amu 7.2 X 272.3999 m 225. amu 225.01775 amu 2.6 X 106.0791 m amu 227.06895 amu B.e X 335.9375 rr> amu 229.0E281 •mu 19.4 X 735.6565; m 2J1. amu 23! .06324 amu 13 .7 X 519.4092 m 233. •mu 233.07620 amu 29.2 X :.10B2 241. amu 241 .08353 •mu 6.6 X 258.3618 m 243. amu 243.07866 amu 6.B X 259.3994 m 245. amu 245.08142 amu 6.2 X 234 .9854 m 247. amu 247.09359 •mu 8.6 X 327.6978 m 251 . •mu 250.96676 amu 23.9 X 52F .442* ni •»r •» amu 252.96769 •mu 11 .0 X 41B.274fc m 259. amu 259.09329 amu 44.6 X 1.6948 26ft. amu 260.09709 •mu 4 . 9 X 1B4.9365 m 2fci. amu 260.75305 amu 4.1 X 156.1279 rT. 261 . amu 261 .03764 •mu B .7 X 331.7261 m 261. amu 261.10973 amu 21 .1 X 603.2637 n. 273. amu 273.10602 •mu 4 . 0 X 151.9775 m 275. amu k«»r. 274 .72674 •mu 17.5 X 662.5743 n, 27?. amu 275.06027 •mu 2.8 X 105.7739 r. 277. •muvnAr v 276.7259P •mu 41 .4 X 1.572V 279. amul * 27e.72306 •mu 33.2 X 1.26:0 277. amu 296.99073 • fliU 3.6 X 137.2681 n. 3*2. amu 3(^3.11658 •mu 5.9 X 224.4263 m 231 . amu 321,0773t •mu e.e X 332.6*16 it. 3<">. amu 3O.0790D •mu 5.5 X 278 .923c y\ 3*9. amu 3^9.0871J •mu 17.S X 657 .9590: m 363. •mu 3h3.07141 •mu 12.2 X 4fc3.B^fc2 rr 379. amu 378.63114 amu 5.9 X 224 .7925* r>' 373. •mu> i 392.60B43 *f l ,U 13.5 X 513.6109 «r. 395. amu W 394.6064* amu 37.7 X 1.434* 397. amu 396.60511 amu 38 . 9 X 1.4794 399. •mu. / 39B.5972e amu 23.5 X B92.7«s£s3 tr 4*7. amu 40*7.09536 amu 21 .6 X B2G.8619 m 421 . •mu 421.10396 amu 5.5 X 210.3SE2 «r 423. amu 423.08464 amu 6 . 0 X 226.3794 «n 493. •mu 493.13273 amu C.5 X 323.1812 tr. 5e=. • * v 509.J3J3B amu 5.3 X 201 .5991 n. 510. ami/ 510.4B411 amu 2 . 7 X 101.4*04 m 512. • 7IU 512.49407 amu 13.6 X 5l6.9fc76 m 51*. amu 514 .4B844 •mu : - A . 7 X 1.015* S i t . •mu 516.47B9ti amu 26.4 X 1.0045 : : B . amu, / 518.47681 emu 10.3 X 389.9536 * . •mu 553.15344 amu 46.8 X 1.779P 5 5 4 . amu 5 M . 16014 •mu 11 .4 X 434 .er :e 555. amu 555.14664 •mu 7 . 9 X 3?1 .6358 m 655. a»u £55.18546 emu 13.2 X 502.6745 n. 752. •mu 752.24063 •mu 4.6 X 174.5606 it) 756. •mu 756.2566C amu 3 .1 X 117.736B m Spectrum m. 1.4 Tabulated LDI-FTICR spectrum of Klebsiella K44 de-O-acetylated, phage degraded polysaccharide doped with potassium bromide -42-III.2 Desorption Chemical Ionisation Mass Spectrometry (DCIMS) of some capsular polysaccharides from Klebsiella and E . coli III.2.1 Abstract DCIMS was used to investigate the fragmentation patterns of some model bacterial oligosaccharides. Using positive ion ammonia DCIMS on the permethylated product, the sequences and the molecular weights of two linear Klebsiella oligosaccharides derived from capsular polysaccharides by chemical degradation were determined. Using the same experimental conditions the molecular weight, the location of a pyruvic acid acetal, and the sequence of a 4+2 type Klebsiella phage degraded oligosaccharide were also determined. This indicates that a combination of DCIMS and phage degradation can be used for the total sequencing of native capsular polysaccharides. Using positive ion methane DCIMS, a partial sequence could also be determined for a linear E . coli phage degraded oligosaccharide containing amino sugars. However the fragmentation observed indicated the site of cleavage is between the glycosidic oxygen and the reducing end residues which is in sharp contrast to that indicated by previously reported spectral data. in.2.2 Introduction CIMS of permethylated oligosaccharides has been used previously to provide sequence and linkage information34,92 The m o s t common reagent gases used are methane, isobutane and ammonia. Ammonia CI mass spectra usually exhibit unambiguous M+NH4 + ions, while isobutane and methane CI mass spectra give considerable information concerning fragment ions34. The major site of bond cleavage occurs at the glycosidic bonds, giving oxonium type -43-ions (Figure HI. 1.2). The non-reducing end fragments may be observed as oxonium ions or be neutralised by hydroxyl groups and subsequently re-ionised by either reagent gas ion attachment, protonation or by abstraction of an electron. The reducing end fragments are generally observed by abstraction of an electron. Thus DCI mass spectra show sequence ions - ions which have lost intact sugar units and can be used for delineating oligosaccharides. Collision activation CI ammonia spectra have also been used to differentiate stereoisomeric permethylated disaccharides with some success^. The following set of experiments was designed to explore the possibility of using DCIMS to sequence "real" oligosaccharides. 111.2.3 Experimental DCIMS analyses of permethylated oligosaccharides were performed on a Delsi Nermag R10-10C quadrupole mass spectrometer interfaced to a D E C PDP 11/23 and a CDC 96 mega bite data system. The reagent gases used were either ammonia or methane with a source pressure of 8X10_5T. The source temperature was between 180-185°C. The samples were deposited on a tungsten filament probe and vaporised by appling a current from 0 to 650mA at a rate of 2mA/second. The methylated samples were methylated by the Hakomori procedure7 7. The product, recovered after extraction with methylene chloride, showed complete methylation (no hydroxyl absorption in the i.r. spectra). The products were further purified by passage through a Sephadex LH20 gel permeation column. 111.2.4 Results and Discussion Data analysis incorporated the following assumptions: a) In the positive ion methane or ammonia experiments the molecular ion is observed as either a -44-M + C H 5 + (M+15 amu), or a M+NH4 + (M+18 amu) peak respectively. If the sample is permethylated or peracetylated then attachment of MeOH (+32 amu) or AcOH (+60 amu) may be observed. b) In negative ion experiments the molecular ion will lose either a H(-l amu) or Me (-15 amu) if the sample is methylated. c ) In permethylated samples, a loss of 174 amu is indicative of a loss of one deoxyhexose unit, a loss of 204 amu is indicative of a loss of one hexose unit, and a loss of 245 amu indicates the loss of one N-acetamido-hexose unit. d) The major cleavage route is the one that forms oxonium ions34 (Figure HI. 1.2). However the oxonium ions formed can be neutralised by hydroxyl groups, and subsequently re-ionised, e ) In positive ion spectra, the base peak is an oxonium type ion, which indicates the non-reducing end residue, with m/z 189 indicative of a permethylated deoxyhexose unit, m/z 219 indicative of a permethylated hexose unit, m/z 260 indicative of a permethylated N-acetamido-hexose unit and m/z 331 indicative of a peracetylated hexose unit. f) Sequence ions are observed from both ends of the oligosaccharides. The major sites of bond cleavage for branched oligosaccharides are at the branch points. 1) Use of derivatised raffinose to establish experimental conditions. OMe OMe OMe Gall 6GIc1 2Fru Figure HI.2.1 Structure of permethylated raffinose. Peracetylated raffinose in positive ion ammonia experiment gave both the pseudo-molecular ion and sequence ions (Spectrum HI.2.1). Although the M+NH4+ peak (m/z 984) and the M --45-Hex related (m/z 696) peak intensities were relatively low (10% and 6% respectively), the spectrum did provide sequence information. The base peak at m/z 331 was deduced as a peracetylated hexose oxonium ion. The addition of an acetoxy group would be fairly facile for a peracetylated sample, thus the M-Hex related ion was deduced as M-Hex+OAc+NH4+ ion. In the negative ion ammonia experiment peracetylated raffinose gave a M - H ion (m/z 965) and M-Hex related (m/z 695) peak with very low intensity (<1%). This M-Hex related peak was deduced as M-Hex+OAc+18-H. In the positve ion methane experiment peracetylated raffinose gave no pseudo-molecular ion, but both the mono- (m/z 331) and di-hexose (m/z 619) oxonium ions were observed. Permethylated raffinose in the positive ion ammonia experiment gave a very intense M + N H 4 + peak (m/z 676, 90%), but no sequence ions were observed except for the mono-hexose oxonium ion at m/z 219 (Spectrum UI.2.2). Both negative ion ammonia and methane experiments showed similar spectra with M-15 peaks at around 10% with no sequence ions, except for the mono-hexose related (m/z 203) oxonium ion. Although the peracetylated product using positive ion and ammonia as the reagent gas gave more sequence information than the permethylated product, peracetylation increased the molecular weight of the sample from 504 amu to 966 amu. This increase in mass limits the analysis of higher peracetylated oligosaccharides. Hence positive ion DCI, using ammonia as the reagent gas, on permethylated samples is the preferred procedure. 2) Permethylated Klebsiella K79 oligosaccharide derived from Smith degradation The positive ion ammonia spectrum of the methylated sample gave a wealth of information on this linear structure (Spectrum m.2.3). The molecular ion was observed with attachment of N H 4 + (m/z 908,7%). This pseudo-molecular ion was observed to lose a glycerol unit (m/z 820, m/z 806), followed by two consecutive losses of a hexose unit (giving m/z 584, m/z 380) and a deoxyhexose unit (giving m/z 206). The oxonium ion fragments of one deoxyhexose unit (m/z -46-189, 100%), two deoxyhexose units (m/z 363) and two deoxyhexoses with one hexose (m/z 567) were observed in abundance. The above observations indicate that the non-reducing end structure is comprised of two deoxyhexoses and one hexose. A m/z 219 (20%) peak was also observed, this cannot be explained by the above structure, which implies that the sample is not pure. This spectrum is very complicated which bears out the fact of an impure sample. However from the major peaks the sequence was deduced as: Rha 1 3Rh a l -—3Gal l 3 G I c 1 - — G r o Figure m.2.2 Structure of permethylated Klebsiella K79 oligosaccharide derived from Smith degradation The glycosidic bond cleavage was only observed from the reducing end. Table m.2.1 contains a list of fragment structures assigned. 3 ) Permethylated Klebsiella K79 oligosaccharide derived from ^elimination In the positive ion ammonia experiment, both pseudo-molecular ions (m/z 790, m/z 776) and sequence ions were observed from this linear structure (Spectrum ni.2.4). Sequential loss of a hexose (-204 amu) and two deoxyhexoses (-174 amu each) from the reducing end gave a deoxyhexose oxonium ion. Consecutive loss of three deoxyhexoses was also observed (gave -47-Table m.2.1 Fragment structures arising from permethylated Klebsiella K79 oligosaccharide derived from Smith degradation. m/z Structure 908 (deoxyHex, deoxyHex, Hex, Hex, gly) + N H 4 + 820 [(deoxyHex, deoxyHex, Hex, Hex) ~ O H + MeOH] + 806 (deoxyHex, deoxyHex, Hex, Hex) ~ O H + N H 4 + 584 [(deoxyHex, deoxyHex, Hex) ~ O H ] + 567 (deoxyHex, deoxyHex, Hex)+ 380 [(deoxyHex, deoxyHex) ~ OH] + 363 (deoxyHex, deoxyHex)+ 206 [deoxyHex ~ OH] + 189 deoxyHex"1" -48-sequence ions of m/z 570, m/z 396, m/z 222), indicating that cleavages occurred from both ends of the oligosaccharide. Table III.2.2 contains a list of structures assigned to these ions. The sequence was deduced to be: 570 396 222 deoxyHex- deoxyHex • deoxyHex-206 189 380 363 Hex~OH 554 537 790 776 758 741 MeO O MeO MeO O MeO / O MeO H , O H MeO OMe ^ OMe OMe OMe R h a l - — 3 R h a l - — 3 R h a l 3Gal Figure III.2.3 Structure of permethylated Klebsiella K79 oligosaccharide derived from (3-eUmination 4) Permethylated Klebsiella K46 phage degraded oligosaccharide This 4+2 structure was chosen to investigate the effect of branching in the study of DCIMS on oligosaccharides, and to determine on which sugar residue the pyruvic acid acetal resides. The primary site of cleavage was expected to be at the branch point, thus the branch point residues would be "labelled" with one extra hydroxyl group in contrast to the methoxy groups on the other sugar residues. Although the M+NH4 + (m/z 1358) was a low intensity peak in the positive ion ammonia experiment (0.5%), it was the only prominent peak with a m/z over 900, and thus was -49-Table IU..2.2 Fragment structures arising from permethylated Klebsiella K79 oligosaccharide derived from P-elimination. m/z Structure 790 [(deoxyHex, deoxyHex, deoxyHex, Hex) ~ O H + MeOH] + 776 (deoxyHex, deoxyHex, deoxyHex, Hex) ~ O H + N H 4 + 758 [(deoxyHex, deoxyHex, deoxyHex, Hex) ~ O H ] + 741 (deoxyHex, deoxyHex, deoxyHex, Hex)"1" 570 [(deoxyHex, deoxyHex, Hex) ~ 20H] + 554 [(deoxyHex, deoxyHex, deoxyHex) ~ O H ] + 537 (deoxyHex, deoxyHex, deoxyHex)4" 396 [(deoxyHex, Hex) ~ 20H] + 380 [(deoxyHex, deoxyHex) ~ O H ] + 363 (deoxyHex, deoxyHex)"1" 222 [Hex ~ 20H] + 206 [deoxyHex ~ OH] + 189 deoxyHex"1" -50-considered a structural peak (Spectrum m.2.5). The m/z 862 peak was indicative of three hexose units and a hexose uronic acid unit with one hydroxyl group. A loss of a hexose unit from the main chain was also observed as a +NH4 + fragment, this fragment has two hydroxyl groups (m/z 662). The m/z 514 peak was indicative of two hexose units and a pyruvic acid acetal unit with one hydroxyl group. A hexose oxonium ion (m/z 219) and a two hexoses oxonium ion (m/z 423) were observed, thus the non-reducing end was determined to be comprised of two hexose units. The m/z 482 peak was indicative of a hexose unit, a hexose uronic unit and a pyruvic acid acetyl unit with three hydroxyl groups attached. Thus this fragment must be the branch point of the oligosaccharide. These observations lead to the following conclusions: a) The non-reducing end sequence is, Hex—Hex— b ) The m/z 514 fragment can be from these structures, —Hex>pyr—Hex, or Hex>pyr—Hex— c) The branch point stucture is one of, —HexU—, or —HexU, or HexU— —Hex>pyr —Hex>pyr- Hex>pyr— Table III.2.3 contains a list of fragment structures arising from this oligosaccharide.Combining the above information the sequence was determined to be, 219 423 662 644 627 1358 Hex Hex HexA Hex 500 482 465 514 r j 862 Hex Hex > pyr Hex Hex > pyr -51-Hex MeO MeO J—I ivieu-j COOMe r- u M e OMe OMe \ QMe OMe MeOOC MeO' OMe Gall 3 G a l l 3 G l c A l 3Man |4 11 Glcl 3Ma n ^y j -Figure ITI.2.4 Structure of permethylated Klebsiella K46 phage degraded oligosaccharide -52-Table DI.2.3 Fragment structures arising from permethylated Klebsiella K46 phage degraded oligosaccharide. COOMe OMe O C H 3 S c / I M e 0 0 C ' N o M ^ y M e O MeO-. \ r C OMe ) OMe m/z Structure 1358 (Hex, Hex, Hex, Hex, Hex>pyr, HexA) + N H 4 + 862 [(Hex, Hex, Hex, HexA) ~ O H ] + 700 [(Hex, Hex>pyr, HexA) ~ 20H] + 662 OHex, Hex, HexA) ~ 20H + N H 4 + 644 [(Hex, Hex, HexA) ~ 20H] + 627 ((Hex, Hex, HexA) ~ OH) + 514 (Hex, Hex>pyr) ~ O H + N H 4 + 500 (Hex>pyr, HexA) ~ 30H + N H 4 + 482 [(Hex>pyr, HexA) ~ 30H] + 465 ((Hex>pyr, HexA) ~ 20H) + 423 (Hex, Hex)+ 219 Hex + -53-5 ) Reduced, permethylated Klebsiella K46 phage degraded oligosaccharide C H 2 O M e Gal 1 -—3Gal l -—3G1CA 1 3 M a n r |4 '1 Glcl 3Man >pyr Figure in.2.5 Structure of reduced, permethylated Klebsiella K46 phage degraded oligosaccharide This reduced oligosaccharide in the positive ion ammonia spectrum showed structural feature peaks similar to the non-reduced sample (Spectrum in.2.5). No pseudo-molecular ion was observed, although there was a prominent peak at m/z 1319, which could be a decomposition product. The reduced main chain gave a prominent peak at m/z 878, corresponding to a reduced hexose unit, two hexose units and a hexose uronic acid unit with one hydroxyl group. Glycosidic bond cleavage occurred at both ends of this main chain. A loss of the non-reducing end hexose +NH4 + peak was observed (m/z 678), as well as the oxonium ion -54-containing two hexose units and a hexose uronic acid unit (m/z 627). The non-reducing end sequence was demonstrated by the presence of mono-hexose (m/z219) and di-hexose (m/z 423) oxonium ion peaks. The side chain fragment (m/z 514) was not prominent. Thus the only definitive information obtained was from the main chain. Table in.2.4 contains a list of fragment structures assigned. 219 H e x -458 440 423 662 644 627 Hex- HexA-678 514 Hex 878 Hex Hex > pyr 7 ) Reduced, permethylated E. coli K44 phage degraded oligosaccharide Me Ac Me Ac GlcAl 3Rh a l ^ l c N A c l - — ^ G a l N A c Figure IH.2.6 Structure of permethylated E. coli K44 phage degraded oligosaccharide. No molecular ion was observed for this positive ion methane experiment. However, the sequence ions were very prominent. The peak at m/z 679 represented a structure comprising one hexose unit and two acetamido-hexose units. Loss of a deoxyhexose unit from sequence ion m/z -55-Table m.2.4 Fragment structures arising from reduced permethylated Klebsiella K46 phage degraded oligosaccharide C H 2 O M e OMe m/z Structure 878 [((Hex, Hex, Hex, HexA) ~ OH) T ] + 678 ((Hex, Hex, HexA) ~ 2 0 H ) r + N I V 662 (Hex, Hex, HexA) ~ 20H + N I V 644 [(Hex, Hex, HexA) ~ 20H] + 627 ((Hex, Hex, HexA) ~ O H ) + 514 (Hex, Hex>pyr) ~ O H + N H 4 + 458 (Hex, Hex) ~ O H + N H 4 + 440 [(Hex, Hex) ~ O H ] + 423 (Hex, Hex)+ 219 Hex + -56-679 was indicated by m/z 505, with subsequent consecutive loss of two acetamido groups (-71 amu each). This was followed by a loss of an acetamido-hexose unit giving m/z 260. Thus, the two acetamido-hexoses were determined to be adjacent to each other and linked to a deoxyhexose unit. The mass at m/z 679, m/z 505, and m/z 260 indicated oxonium type ions Spectrum m.2.7). Hence the non-reducing end is likely to be comprised of two acetamido-hexoses linked to a deoxyhexose unit. Thus the proposed sequence was: HexNAc- HexNAc- •deoxyHex-260 505 679 189 434 363 In contrast to these results, methylation analysis data indicated that one of the acetamido-hexoses was the reducing end residue^4. However, a m/z 233 peak was not observed, this indicates that if the cleavage did originate at the non-reducing end a glucuronic acid oxonium ion was not formed. Thus, the site of cleavage was between the rhamnose residue and the glycosidic oxygen (Figure III.2.7). This fragmentation pattern is in sharp contrast to previously reported spectra^. This discrepancy may be explained by the fact that the acetamido group strengthens the glycosidic bond, thus the direction of cleavage originated from the non-reducing glucuronic acid. Table III.2.5 contains a list of fragment structures assigned. The real sequence of the oligosaccharide is: HexA ~ | — deoxyHex ~ | — HexNAc — | — HexNAc 679 505 260 434 189 363 Furthermore, this DCI experiment showed that the reducing end residue did not undergo reduction. This may be due to a contaminant associated with the reducing end, however this contaminant was removed before methylation. As the molecular ion was not observed, a total sequence could not be determined on the oligosaccharide. Although only a partial sequence can -57-Table IH2.5 Fragment structures arising from permethylated E. coli K44 phage degraded oligosaccharide. COOMe MeO-9 MeO A °. C H 3 n o p 0 M e—o ^ O M e ^ H,OMe y N N Me N A c M e ' ' A c m/z 679 LOMe M e ' N A c M e ' ^ A c MeO Q / o M e — 9 N N f N ^ \ . Me Ac Me Ac OMe 505, 434 (-NMeAc+H), 363 (-2NMeAc+2H) MeO J O OMe V > 0 M e N f N Me Ac 260, 189 (-NMeAc+H) -58-be proposed from the mass spectra, combined with methylation analysis, the sequence of the polysaccharide can be deduced. HexNAc HexNAc deoxyHex HexA COOMe J q MeO ( ( M e MeONJ OMe ° COOMe O MeO /j O • ' O M e ) ( C H 3 MeO N A> \ / OR OMe OMe OMe Figure IH.2.7 Proposed fragmentation pattern for permethylated E.coli K44 phage degraded oligosaccharide III.2.5 Conclusions DCIMS is well suited for the location of pyruvic acid acetal in an oligosaccharide and the sequencing of linear and simple branched oligosaccharides. Although sequence fragments were not always prominent, the use of methylation analysis clarifies the spectral data, and can be used to propose total structural sequence. -59-Spectrum ITJ.2.1 Positive ion ammonia DCI spectrum of peracetylated raffinose -60-o 8F M L ri o f-8 10 in ( 0 n 1 0 ( 0 • , 1 i-s N i i (A -\ a SHE 00 Spectrum HI.2.2 Positive ion ammonia DCI spectrum of permethylated raffinose -61-Spectrum m.2.3 Positive ion ammonia DCI spectrum of permethylated Klebsiella K79 oligosaccharide derived from Smith degradation. Spectrum m.2.4 Positive ion ammonia DCI spectrum of permethylated Klebsiella K79 oligosacharide derived from P-elimination. -63-Spectrum m.2.5 Positive ion ammonia DCI spectrum of permethylated Klebsiella K46 phage degraded oligosacharide Spectrum HI.2.6 Positive ion ammonia DCI spectrum of reduced permethylated Klebsiella K46 phage degraded oligosacharide -65-Spectrum ITJ.2.7 Positive ion methane DCI spectrum of reduced permethylated E. coli K44 phage degraded oligosacharide -66-III.3 Fast Atom Bombardment Mass Spectrometry (FABMS) of some capsular polysaccharides from Klebsiella and E. coli 111.3.1 Abstract F A B M S was used to investigate the fragmentation pattern of some model bacterial oligosaccharides. Using thioglycerol or glycerol as the liquid matrix, the molecular weight and the sequences of five linear Klebsiella structures derived from phage degradation and chemical degradations were determined. For the phage degraded oligosaccharide Klebsiella K44 the complete sequence of the polysaccharide was determined. For branched oligosaccharides, the sequence cannot be determined using underivatised samples as permethylation was required to label the branch point. A partial sequence was also determined for a linear E. coli phage degraded oligosaccharide containing amino sugars. However the fragmentation observed indicated the site of cleavage is between the glycosidic oxygen and the reducing end moiety which is in sharp contrast to previously reported spectral data. 111.3.2 Introduction Although FABMS has only been in existence since 1981, it has already taken a prominent role in carbohydrate research40. it has been used for probing the structures of glycoproteins, glycosphingolipids and other glycoconjugates^5-99- in combination with high field magnets, molecular ions have been observed up to 5000 amu for permethylated glycosphingolipids 100. F A B M S was originally used for analysing underivatised oligosaccharides. It has been established that derivatised samples are more volatile and consequently give more structural information. Intact sugar residues are cleaved from both ends of underivatised oligosaccharides, via either oxonium ion formation or hydrogen transfer at the glycosidic bonds (Figure in.3.1). -67-The most common derivatisation reactions are acetylation and methylation.The major type of fragmentation pathway for a derivatised oligosaccharide is oxonium ion formation. OH Figure m.3.1 Major fragmentation pathways in positive FABMS IH.3.3 Experimental Positive ion FABMS experiments were performed on an in-house-modified Kratos-AEI MS 9 mass spectrometer with an accelerating potential of 6kV, equipped with a saddle field discharge fast atom bombardment gun which uses xenon gas at an accelerating potential of 7.5kV and an ion current of 1mA. The liquid matrices used included glycerol, thioglycerol and mercaptoacetic acid. Underivatised samples were directly applied and mixed in the liquid matrix, while permethylated samples were first dissolved in chloroform or chloroform/methanol (1:1) mixture, and then applied to the matrix. For raffinose experiments, the spectra were recorded using a data system, while all the other experiments were analog recorded on Kodak's photographic Linagraph paper and manually counted. -68-The derivatised samples were methylated by the Hakomori procedure^. The product, recovered after extraction from dichloromethane, showed complete methylation (no hydroxyl absorption in the i.r. spectra). These samples were further purified by passage through a Sephadex LH20 gel permeation column. III.3.4 Results and Discussion Data analysis incorporated the following assumptions: a) Extensive protonation and cationisation of the pseudo-molecular and sequence ions occurs. b) For underivatised samples, cleavage occurs at both ends of the oligosaccharide. A loss of 162 amu is indicative of a loss of a hexose unit, a loss of 164 amu is indicative of a loss of a reduced hexose unit, a loss of 146 amu is indicative of a loss of a deoxyhexose unit, a loss of 176 amu is indicative of a loss of a hexose uronic acid unit, and a loss of 42 amu is indicative of a loss of an acetate group. c ) For permethylated samples, the major type of fragment ions are oxonium ions. 1) Raffinose was used to established the best matrix for both underivatised and permethylated samples (Figure HI.2.1). a) Underivatised raffinose Different liquid matrices were used, these included glycerol, thioglycerol, mercaptoacetic acid, tetraethylene glycol, 1,2,4-butanetriol, triethanolamine, diethanolamine, diallyl phthalate, tetramethyl sulphone, O-n-octyloxynitrobenzene, Triton X-100 a surfactant, PEG 200, PEG 400, and dimethylsulphoxide. The only matrices that gave relatively intense sample peaks were glycerol, thioglycerol, and mercaptoacetic acid (Spectrum III.3.1, Spectrum III.3.2 and Spectrum III.3.3). Table 111.3.1 contains a list of sequence ions arising from raffinose in glycerol, thioglycerol and mercaptoacetic acid matrices respectively. All the other matrices either -69-Table in.3.1 Structural ions arising from underivatised raffinose. Glvcerol CHoOHCHOHCH.OH 92amu Structure m/z Rel. Intensity % M + Glycerol M + FT 597 505 38 54 M + H + - 162 M + H + - 180 343 325 100 94 Thioelvcerol CFkOHCHSHCHoOH 108amu Structure m/z Rel. Intensity % M + K + M + N a + M + H + 543 527 505 4 15 9 M + N a + -162 M + H + -162 M + FT -180 365 343 325 15 64 100 Mercaptoacetic acid HSCH2CQQH 92amu Structure m/z Rel. Intensity % M + K + M + N a + M + H + M + H + - H z O 543 527 505 487 3 11 23 11 M + H + - 162 M + FT- 180 343 325 52 100 -70-gave no sample peaks (e.g., tetramethyl sulphone, Spectrum III.3.4) or the sample peaks had very low intensities (e.g., tetraethylene glycol, Spectrum in.3.5). Thus glycerol or thioglycerol or mercaptoacetic acid was used as the matrix for underivatised oligosaccharides. 543 527 505 487 365 343 325 Hex- Hex- Hex 365 343 325 543 527 505 487 b) Permethylated raffinose Out of all the original liquid matrices, the seven that showed some structural information for underivatised raffinose were investigated for permethylated raffinose. Only thioglycerol and mercaptoacetic acid gave all the sequence ions and molecular ions (Table III.3.2). Although triethanolarnine gave a very prominent pseudo-molecular ion, no sequence ions were observed. Thus, all the permethylated oligosaccharide experiments employed thioglycerol or mercaptoacetic acid as the matrix. Hex- Hex- Hex 219 423 697 681 677 2 ) Klebsiella K79 permethylated oligosaccharide derived from Smith degradation with thioglycerol as the matrix (Figure m.2.2). Most of the sequence ions were observed, the hydrated molecular ion was also observed -71-Table EI.3.2 Structural ions arising from permethylated raffinose OMe OMe OMe Thioelvcerol CH0OHCHSHCH0OH Structure m/z Rel. Intensity % M + K + M + N a + M + H + + H 2 0 697 681 677 46 100 64 (Hex ~ Hex)+ 423 157 Hex + 219 lOOn Mercaptoacetic acid HSCH^COOH Structure m/z Rel. Intensity % M + K + M + N a + M + H + + H 2 0 697 681 677 17 100 42 Hex + 219 lOOn -72-(m/z 908). The non-reducing end sequence was indicated by the mono-, di- and tri-deoxyhexose oxonium ions (m/z 189, m/z 367,and mz/ 567). Table IH.3.3 contains a list of sequence ions arising from this structure. The sequence was deduced as: deoxyHex- deoxyHex • Hex- Hex- Gro 189 363 567 908 3 ) Klebsiella K79 permethylated oligosaccharide derived from ^ elimination with thioglycerol as the matrix figure III.2.3). All the sequence ions and the pseudo-molecular ion were observed. This oligosaccharide cleaved from the non-reducing end giving mono-, di-, tri-deoxyhexose and tri-deoxyhexose plus a hexose oxonium ions (m/z 189, m/z 363, m/z 573, and m/z 741). The pseudo-molecular ion was deduced as the matrix attached ion (m/z 866). Table DI.3.4 contains a list of sequence ions arising from this structure. Thus, the sequence of the oligosaccharide was deduced as: deoxyHex- deoxyHex- • deoxyHex 189 363 —J Hex-j-OH^J 741 537 866 4 ) Klebsiella K67 chemically degraded oligosaccharide Using glycerol as matrix, the pseudo-molecular ions were observed in abundance, these included matrix-, potassium-, sodium-, and proton attachment. A loss of 176 amu followed by 162 amu was observed, leaving m/z 343 which was indicative of a protonated two hexose unit. Table UI.3.5 contains a list of sequence ions arising from this underivatised oligosaccharide. Thus the sequence was limited to two possibilities: -73-Table in.3.3 Structural ions arising from permethylated Klebsiella K79 oligosaccharide derived from Smith degradation. Structure [M + H 2 0 ] + ( deoxyHex, deoxyHex, Hex ) + (deoxyHex, deoxyHex ) + (deoxyHex ) + m/z Rel. Intensity % 908 10 567 36 367 100 189 lOOn -74-Table m.3.4 Structural ions arising from permethylated Klebsiella K79 oligosaccharide derived from {5-elimination. Structure m/z Rel. Intensity % [M + Thioglycerol]+ 866 47 ( deoxHex, deoxyHex, deoxyHex, Hex ) + 741 19 ( deoxyHex, deoxyHex, deoxyHex ) + 531 34 (deoxyHex, deoxyyHex ) + 363 100 (deoxyHex ) + 189 lOOn -75-527 HexA- Hex- Hex- Hex 343 519 519 or 343 Hex- Hex- •Hex 527 773 719 703 681 663 —J—HexAj 773 719 703 681 663 COOH HO HO H,OH O H O H G l c A 1 - — 2 M a n l - — ^ M a n 1 - — 3 G I C Figure ITJ.3.2 Structure of Klebsiella K67 chemically degraded oligosaccharide. 5 ) Klebsiella K46 reduced, phage degraded oligosaccharide. Figure III.2.5. a) Underivatised oligosaccharide in thioglycerol as the matrix (Spectrum m.3.6). Al l the sequence ions were observed. The pseudo-molecular ions could not be determined accurately due to poor resolution of the ion clusters, however, peaks were in the pseudo-molecular ion range. From the sequence ions the structure could be delineated. However, for a -76-Table m.3.5 Structural ions arising from Klebsiella K67 oligosaccharide derived from chemical degradation. O H O H Structue m/z Rel. Intensity % M + glycerol 773 52 M + K + 719 35 M + N a + 703 62 M + H + 681 32 M + H + + H 2 0 663 20 M + Na + - HexA 527 18 M + H + - HexA 505 38 M + H + - H e x 519 16 M + H + - HexA - Hex 343 100 -77-similiar linear structure the sequence ions would have the same mass, thus underivatised samples could not successfully be used for sequencing branched oligosaccharides. Table III.3.6 contains a list of sequence ions arising from this structure. For this model compound the fragmentation is as follow: 541 365 519 1099 1077 Hex- Hex- HexA-521 381 413 [~ Hex J 683 Hex Hex>pyr H O - , H O - i H o y — 9 Hoy—° % O H r - O H COOH O H O H -h - O H O H '— O H Figure III.3.3 Structure of reduced Klebsiella K46 phage degraded oligosaccharide. -78-Table IJX3.6 Structural ions arising from reduced Klebsiella K46 phage degraded oligosaccharide. O H Structure m/z Rel. Intensity % M + N a + 1099 1 M + H + 1077 13 (Hex, Hex, Hexr, HexA) + H + 683 8 (Hex, Hex, HexA) + N a + 541 17 (Hex, Hex, HexA) + H + 519 13 (Hex, Hex1, HexA) + H + 521 14 (Hex>pyr, Hex) + H + 413 100 (HexA, Hex1) + Na + 381 30 (Hex, Hex) + N a + 365 45 -79-b) Permethylated oligosaccharide in thioglycerol matrix Some weak sequence ions were observed corresponding to a di-hexose fragment (m/z 423), a hexose linked to a uronic acid with two hydroxyl groups fragment (m/z 409), a di-hexose linked to a uronic acid with two hydroxyl groups fragment (m/z 627). The pseudo-molecular ion could not be assigned accurately due to poor resolution of the ion clusters. 6 ) E. coli K44 reduced, permethylated phage degraded oligosaccharide with thioglycerol as the matrix. Figure III.2.7. No pseudo-molecular ions were observed. However, peaks corresponding to a mono- and a di-acetamido-hexose were observed in low intensity, which indicated that the non-reducing end sequence of the oligosaccharide was two acetamido-hexoses linked together. These two fragments were also observed in the DCI spectrum, nevertheless, methylation analysis results indicated that the two acetamido-hexoses were from the reducing end. As observed in DCI, this fragmentation pathway is also in sharp contrast to previously reported FAB spectra^. 7 ) Klebsiella K44 de-O-acetylated, phage degraded oligosaccharide. Both glycerol and thioglycerol allowed complete sequencing of this oligosaccharide. However using glycerol as the matrix, an abundance of pseudo-molecular ions were observed (m/z 849, m/z 833, and m/z 811). The protonated molecular ion (m/z 811) was the most intense of all the sample peaks. Bond cleavages occurred at both ends of the oligosaccharide. Consecutive losses of two 162 amu units followed by a loss of 146 amu (giving peaks at m/z 631, m/z 469, and m/z 341) indicated a sequence of two adjacent hexose units and a deoxyhexose unit linked together. From the other end of the oligosaccharide, a loss of 176 amu followed by two 146 amu losses (giving peaks at m/z 635, 489, and m/z 343) indicated a -80-sequence of a hexose uronic acid linked to two adjacent deoxyhexose units. Thus the sequence of the oligosaccharide was limited to two possibilities: 489 635 471 343 HexA- deoxyHex— 341 deoxyHex-469 Hex- Hex 649 849 631 833 811 or 649 631 469 341 Hex- Hex-343 deoxyHex- • deoxyHex- HexA 489 635 849 471 833 811 H O - , H O - , 0 \ ?} ~ ° \ r -}/ 7 H O N / N \ / I O H i O H i O H H , O H G l c A 1 - — 2 R h a ! - — 3 R h a l . — 3 G I c 1 - — 4 G l c Figure DI.3.4 Structure of Klebsiella K44 de-O-acetylated, phage degraded oligosaccharide. Table III.3.8 contains a list of sequence ions arising from this oligosaccharide. As the polysaccharide is made up of repeating oligosaccharide units, the partial sequence of the polysaccharide was deduced as: HexA deoxyHex deoxyHex Hex Hex -81-Table ITJ.3.8 Structural ions arising from Klebsiella K44 de-O-acetylated, phage degraded oligosaccharide. COOH HO"! H O - , Structure m/z Rel. Intensity % M + K + 849 40 M + N a + 833 52 M + H + 811 100 M + H + - Hex 649 9 M + H + - Hex - H 2 0 631 26 M + H + - HexA 635 25 M + H + - HexA - deoxyHex 489 45 M + H + - HexA - deoxyHex - H 2 0 471 43 M + H + - 2Hex - H 2 0 469 69 M + H + - HexA - 2deoxyHex 343 65 M + H + - 2Hex - deoxyHex 341 47 -82-8 ) Klebsiella K44 reduced phage degraded oligosaccharide • - O H - O H G l c U 1 - — 2 R h a ! - — S R h a 1 - — ^ l c 1 - — 4 G l c r 6 - — O A c Figure in.3.5 Structure of Klebsiella K44 reduced phage degraded oligosaccharide. Using glycerol as matrix, the most intense sample peak was the sodiated molecular ion (m/z 877), the potassiated molecular ion was also abundant (m/z 893, 68%). Similar to the de-0_-acetylated oligosaccharide, bond cleavages occurred at both ends, with a loss of 42 amu followed by 164 amu and 162 amu (giving m/z 835, m/z 671, and m/z 509), indicating that the reducing end residue was an acetylated hexose unit linked to another hexose unit. From the non-reducing end, loss of 176 amu was followed by two 146 amu losses (gave m/z 701, m/z 555, and m/z 409), indicating the sequence to be a hexose uronic acid linked to two deoxyhexose units. Thus, the sequence of the oligosaccharide was determined as: 701 555 409 HexA- deoxyHex—j— deoxyHex- Hex-509 67 Hex^j- OAcJ 835 895 877 Table III.3.9 contains a list of sequence ions arising from this oligosaccharide. Due to its base-lability the location of an acetate group is very difficult to establish by means of wet chemical methods. However using FABMS on the underivatised phage degraded oligosaccharide its location can be asssigned to a specific sugar residue. Therefore combining the sequence -83-information from both the de-Q-acetylated and the reduced oligosaccharide, the total sequence of the polysaccharide could be deduced as: HexA deoxyHex deoxyHex Hex Hex I OAc Although the sequencing of the polysaccharide can be achieved from the native phage degraded product, in order to sequence the phage degraded oligosaccharide the reduced sample has to be investigated to locate the reducing end. III.3.5 Conclusions Positive ion FABMS with either glycerol or thioglycerol as the liquid matrix can be used successfully for the sequencing of some oligosaccharides. For branched oligosaccharides the sample needs to be derivatised in order to distinguish branch points. FABMS can also be used for the location of base-labile acetate groups, which are normally very difficult to determine by wet chemical methods. -84-Table m.3.9 Structural ions arising from reduced Klebsiella K44 phage degraded oligosaccharide. , - O H Structure m/z Rel. Intensity % M + K+ M + N a + 895 877 68 100 M + N a + - 42 835 9 M + Na + - HexA 701 14 M + N a + - Hex r- 42 671 8 M + N a + - HexA - deoxyHex 555 28 M + N a + - Hex r- 42 - Hex 509 15 M + N a + - HexA - 2 deoxyHex 409 45 -85-»«. u - * I I I I I I I I I I - » - •» 11II111111 - » a tfl n • I I I I I I I I I I I 111111 n 11 - * s * * » * • » «• 8 S 8 • * « N • j a <B «r N >• _ J a < Spectrum m.3.1 Positive ion FAB spectrum of raffinose in glycerol -86-Spectrum HJ.3.2 Positive ion FAB spectrum of raffinose in thioglycerol Spectrum in.3.3 Positive ion FAB spectrum of raffinose in mercaptoacetic acetic Spectrum 1H.3.4 Positive ion FAB spectrum of raffinose in tetramethyl sulphone -89-Spectrum m.3.5 Positive ion F A B spectrum of raffinose in tetraethylene glycol -90-Spectrum ni.3.6 Positive ion FAB spectrum of reduced, underivatised Klebsiella K46 phage degraded oligosaccharide in thioglycerol matrix -91-III.4 Discussion. The use of mass spectrometric methods for the sequencing of bacterial oligosaccharides is becoming routine in carbohydrate research. Although FD and DCI were both discovered before F A B , the latter is now the most common soft ionisation technique employed in most laboratories. The reasons for this are four fold. a) In the case of FD, the quality of the spectra is largely dependent on sample preparation. Although FD spectra usually give very prominent molecular ions, sequence ions are not always observed. Thus, sequencing of the oligosaccharide via other methods is still required. b) A FAB source can be easily made from a common EI source, hence major rebuilding is not necessary to convert an instrument. This F A B source can be fitted in most double focusing instruments, thus a conventional spectrometer can be easily upgraded into a dedicated FAB ionisation spectrometer. c) The development of stable high field magnets has allowed high sensitivity along with a large mass range. Thus, a high field spectrometer equipped with a F A B source can in theory and practice analyse oligosaccharides containing up to 25 sugar units (~ 5000 amu). d ) Thermal degradation of high molecular weight samples (> 1500 amu) in DCIMS has still be to overcome. The mechanism of ionisation in F A B is not fully understood. At least ten different models have been suggested; ranging from ion sputtering, to surface supply of sample molecules by ion migration in the matrix, and to performance ions modell01-H0 t Nevertheless, it is generally accepted that impurities in the sample surpress the abundance of sample ions observed, thus sample purification is very important. In DCIMS experiments the derivatised sample is deposited on an inert filament, and is then slowly vaporised in the reagent gas ion plasma. During sample heating most volatile impurities -92-are removed prior to the derivatised oligosaccharide which is normally less volatile. Thus for fairly impure samples DCI is probably the better method of ionisation. This thermal-assist procedure has also been applied to FAB* 11. The gradual volatilisation of the sample can also be used to differentiate different oligosaccharides with similar molecular weight by applying selective ion monitoring technique. The fragmentation pathways observed for FAB and DCI of permethylated oligosaccharides are the same. However the energy spread of the secondary ion beam in a FAB source is much higher than the normal EI or DCI source. This intrinsic drawback is due to the "open" configuration of the ion source, the nature of the primary beam and the volatility of the liquid matrix. The distance between each secondary ion generated and the accelerating plate is quite different due to the sputtering effect of the primary beam, thus the kinetic energy spread of the secondary ion is quite wide. This drawback has three effects on the mass spectrum a ) non-reproducibility of mass spectra data, b) as fragmentation is induced by excess energy acquired during ionisation, low energy ions will be observed primarily as pseudo-molecular ions. This leads to loss of structural information, and c ) poor resolution of the ion clusters. The extent of fragmentation in DCI is largely dependent on the reagent gas, the ion plasma density and the rate of sample heating, therefore by keeping these conditions constant DCI is much more reproducible than FAB. No effort has been directed towards identifying ring cleavage fragments in the DCI spectra discussed in this thesis. The low intensities of most of the sample ions above 500 amu indicated that although DCI is a much softer ionisation technique than EI, a considerable amount of thermal degradation does occur. As ring cleavage fragments are less stable than glycosidic bond cleavage fragments, ring cleavage fragment ions will thermally decompose more readily. The upper mass range of a successful DCI experiment where pseudo-molecular ions are observed is probably at around 1400 amu, which is six to seven permethylated sugar units. Although this limit is much less than FAB, most oligosaccharides investigated in our laboratory are below this mass range, so DCI can be utilised. -93-The ability of FAB to analyse native oligosaccharides allows it to locate base-labile acetate groups on specific sugar residues. This does not apply for DCI, as the native oligosaccharides are not volatile enough for direct analysis. Thus, the sample has to be derivatised before DCI mass spectrometric determination. However most common methylation methods will remove base-labile acetate groups therefore total sequencing is not achieved. Nevertheless, this can be overcome by using the Prehm methylation technique which does not remove labile acetate groups** 1. A new derivatisation method has been developed for FAB for analysing O-acetylated oligosaccharides^, in this method the oligosaccharide is acetylated with deuterated acetylating reagent. The original O-acetate group is not replaced during the acetylation process, therefore by the mass difference between the deuterated acetates and the original non-deuterated O-acetate group, its location can be determined. The main advantage of this method is the improved volatility of the derivatised oligosaccharides over the native samples. As seen from the Klebsiella K46 phage degraded oligosaccharide experiments, both F A B and DCI can also be used for the location of the pyruvic acid acetal. Therefore, soft ionisation mass spectrometric techniques applied to permethylated phage degraded oligosaccharides can be used for the total sequencing of polysaccharides containing labile substituents. The advantages of derivatisation are not confined to improving volatility. Derivatisation also helps to label the branch points of the oligosaccharide. As seen from the DCI experiments on permethylated and reduced, permethylated Klebsiella K46 phage degraded oligosaccharide the branch point was labelled with an extra hydroxyl group rather than a methoxy group. Underivatized branched samples would give the same sequence ions as a similiar linear structure. If the branched oligosaccharides are made up of only one type of sugar residue, then the determination of the branch point would not be possible as different fragment structures would have the same mass. More complex branched oligosaccharides may be more difficult to sequence for the same reason. Reducing the oligosaccharide prior to methylation allows the reducing end to be determined. Finding the reducing end is not as important in the case of sequencing a -94-polysaccharide from a phage degraded oligosaccharide. However in order to determine unambiguously the sequence of the oligosaccharide itself, the reducing end has to be established. This is important in the case of locating the site of cleavage for identifing the enzymic properties of the phage borne enzyme. Although only one LDI-FTICR experiment was carried out, the amount of structural information obtained from the underivatised oligosaccharide investigated allowed both sequencing and assignment of possible position of linkages for certain glycosidic bonds. Both F A B and DCI mass spectrometry give sequence information, nevertheless they do not give linkage position. Thus, LDI-FTICR has the advantages of being able to sequence underivatised oligosaccharides as well as providing linkage information. The ability of LDI-FTICR to analyse non-volatile underivatised oligosaccharides is clearly an advantage. Methylation of oligosaccharides usually generates a lot of impurities, which interfere with both FAB and DCI experiments. The most commonly used Hakomori methylation procedure also removes base-labile O-acetate group, thus complete sequencing is only achieved by using the Prehm methylation procedure. Although, LDI-FTICR experiments can in theory locate O-acetate, ion/molecule reactions may consume this labile substituent during the delay after laser excitation. From the observation of ring fragments and the loss of C02 from the uronic acid residue, LDI is a stronger ionisation method than F A B , therefore other labile substituents can also be lost due to excess excitation energy induced by the laser. Methylation for the labelling of branch points may not be required for the sequencing of branched oligosaccharides in LDI-FTICR due to the prominent ring fragments observed in the spectra. Any fragment structures carrying more than one ring rupture units are obviously the branch points. In the FAB experiment, Klebsiella K44 deacetylated, phage degraded oligosaccharide underwent sequential consecutive loss of sugar residues from both ends of the oligosaccharide. In contrast to this, the same oligosaccharide in LDI-FTICR underwent an initial loss of the uronic acid residue, then further sequential consecutive loss of sugar residues from both ends. This may -95-indicate that the uronic acids are less stable in LDI spectra, and preferentially cleaved before other sugar residues. If this preferential cleavage of specific sugar residues can be selectively employed, it may be useful for the sequencing of branched oligosaccharides. Table III.4.1 compares the three different types of ionisation discussed in this thesis. The complete structural characterisation of a bacterial oligosaccharide includes the determination of: a) type and number of sugar components b) sequence and the pattern of branching c) the position of any non-carbohydrate component d) the sites of glycosidic linkages e) the conformation of sugar rings f ) the anomeric configuration of sugars Although mass spectrometry can only be used for the sequencing and the determination of non-carbohydrate components; the types and the glycosidic linkages of the oligosaccharide can be assigned by methylation analysis, whereas the anomeric nature of the linkages can be obtained by NMR. Thus a combination of mass spectrometry, NMR and methylation analysis experiments on the phage degraded oligosaccharides can totally characterise bacterial native polysaccharides. The choice of ionisation technique and derivatisation methods is dependent on the nature of any non-carbohydrate components present and the degree of polymerisation of the oligosaccharide. For an acetate carrying oligosaccharide, the first step in mass spectrometric analysis is to determine the structure of the underivatised oligosaccharide by FABMS. This will sequence the oligosaccharide and locate the acetate group on a specific sugar residue. Then the oligosaccharide can be reduced and methylated by the Prehm method. This reduced permethylated oligosaccharide can then be analysed by either DCIMS or FABMS again for the sequence and the location of the acetate group. Further confirmation of the nature of the sugar residues present and the reducing end residue can be carried out by methylation analysis on the oligosaccharide. The position of acetylation can be determined by comparing the methylation -96-Table III.4.1 Comparision of LDI-FTICR. DCIMS and FABMS. Features Mode of ionisation Ionisation Background noise level High resolution spectra Dependent on matrix/reagent gas Matrix / reagent gas used Underivatised oligosaccharides Permethylated oligosaccharides Linear oligosaccharides Branched oligosaccharides Molecular weight information Structural information Sequence information Non reducing end information Isomeric information Anomeric information Ring size information Labile substituent Positions of linkage Intensity of molecular ion Intensity of structural ions Dehydration LDI-FTICR Laser Photon Very low Yes Yes NaCl KBr Yes ? Yes ? Yes Yes Yes Yes Yes Strong Strong Yes DCIMS FABMS Chemical Fast atom Ion/molecular re. Unknown Low Very high Yes Ammonia Methane Isobutane Yes Yes Yes Yes Yes Yes Yes Yes Weak Weak Yes Glycerol Thioglycerol Mercaptoacetic ac. Yes Yes Yes Yes Yes Yes Yes Yes Yes Weak Weak Yes -97-analysis results of the Prehm permethylated oligosaccharide and the Hakomori permethylated oligosaccharide. For oligosaccharides which have more than six sugar residues, F A B would be used for two reasons: a ) Our DCI instrument has a mass range limit of 1500 amu, thus any permethylated oligosaccharides that contain more than six sugar residues would be outside the instrument range. b ) FAB is a softer form of ionisation than DCI. This is reflected by the fact that in DCI permethylated oligosaccharides fragment from both ends. Thus a relatively high molecular weight sample would be more likely to thermally decompose in the heated DCI source than in the FAB source. For oligosaccharides with less than seven units, DCI is preferred. This is due to the ability of DCI to volatilise the sample slowly, a process which removes many of the volatile non-carbohydrate impurities, which interfere with sample desorption in FAB. -98-III.5 Conclusion Soft ionisation techniques can be used for the sequencing of bacterial polysaccharides. The choice of mass spectrometric methods is dependent on the nature of any non-carbohydrate components present and the degree of polymerisation of the oligosaccharide. When the permethylated phage degraded oligosaccharide is used for analysis, the sequence of the polysaccharide, the branch point and the location of the non-carbohydrate component can be pin-pointed to a specific sugar residue. -99-F U R T H E R W O R K -100-IV F U R T H E R W O R K High performance liquid chromatography (HPLC) is generally used for the separation of complex mixtures of oligosaccharides. FABMS has proven to be a powerful technique for the characterisation of oligosaccharides. A moving belt interface for coupling HLPC and FABMS has been developed 112. This interface has been successful in analysing underivatised oligosaccharides of animal origin H 3 t Another HPLC-FABMS interface has been developed in which the effluent is directly introduced into the FAB sourceH4\ Qur present Delsi-Nermag mass spectrometer has the capability for this interface, thus further work can be carried out on the direct HPLC-FABMS analysis of bacterial oligosaccharides. Further work on the analysis of model oligosaccharides should be performed to establish possible fragmentations observed for different linkage positions. This information can be used to confirm methylation analysis results for acid-labile sugar residues. -101-V B I B L I O G R A P H Y -102-V B I B L I O G R A P H Y 1 Aspinall, G. O. in Aspinall, G. O. (Ed.), "The Polysaccharides", Vol.1, Academic Press, New York, 1982, p l l 2 Sandford, P. A.; Baird, J. in Aspinall, G. O. (Ed.), "The Polysaccharides", Vol 1, Academic Press, New York, 1982, p411 3 Dudman, W. F. in Sutherland, I. (Ed.), "Surface Carbohydrates of the Prokaryotic Cell", Academic Press, London, 1977, p357 4 Albersheim, P.; Davill, A. G.; McNeil, M. ; Valent, B. S.; Sharp, J. K.; Nothnagel, E . A.; Davis, K. R.; Yamazaki, N.; Gollin, D. J.; York, W. S.; Dudma, W. F.; Davill, J. E.; and Dell, A. in Ciferri, O. and Dure, L.ITI (Eds.), "Structure and Function of Plant Genomes", Vol. 63, N A T O ASI Series, Plenum Press, New York,1982, p293 5 Bishop, C. T.; Jennings, H. J. in Aspinall, G. O. (Ed.), "The Polysaccharides", Vol 1, Academic Press, New York, 1982, p 291 6 Jann, K.; Jann, B. Prog. Allergy 1983,33,53 7 Jenning, H. J. Adv. Carbohydr. Chem. Biochem. 1983, 44, 155 8 Jann, K.; Westphal, O. in Sela, M . (Ed.), "The Antigens", Vol 3, Academic Press, N. Y. 1975, p i 9 Heidelberger, M. ; Dutton, G. G. S. / . Immunol. 1973, 111, 857 10 Heidelberger, M . ; Nimmich, W. Immunochem. 1976, 13, 67 11 Heidelberger, M . ; Lew, Y. J. Carbohydr. Res. 1976, 52, 255 12 Heidelberger, M . Res. Immunochem. Immunobiol. 1973, 3, 1 13 Heidelberger, M . Trends Biochem. Sc. 1982, 7, 261 14 Whistler, R. L.; Bushway, A. A.; Singh, P. P.; Nakahara, I.; Tokuzen, R. Adv. Carbohydr. Chem. Biochem. 1976, 32, 235 15 Crook, J. R.; Otto, W. K.; Jones, R. S. Proc. Soc. Exp. Biol. Med. 1962, 109, 552 -103-16 Miyamoto, K.; Koshiura, R.; Hasegawe, T.; Kato, N. Jpn. J. Pharmacol. 1984, 36, 51 17 Reinhold, V. N.; Carr, S. A. Mass Spectrom. Rev. 1983, 2, 153 18 Aspinall, G. O. in Aspinall, G. O. (Ed.), "The Polysaccharides", Vol.1, Academic Press, New York, 1982, p36 19 Beckey, H. D. Int. J. Mass Spectrom. Ion Phys. 1969, 2, 500 20 Vestal, M . L . Mass Spectrom. Rev. 1983, 2, 447 21 Munson, B. Anal. Chem. 1971, 43, 28A 22 Cotter, R. J. Anal. Chem. 1980, 32, 1589A 23 Blakely, C. R.; Carmody, J. J.; Vestal, M . L . J. Am. Chem. Soc. 1980, 102, 5931 24 Beckey, H. D. "Field Ionisation Mass Spectrometry" Pergamon Press, Oxford, 1971 25 Evans, C. A. Jr.; Hendricks, C. D. Rev. Sci. Instrum. 1972, 43, 1527 26 Benninghoven, A.; Sichtermann, W. K. Anal. Chem. 1978, 50, 1180 27 Barber, M . ; Bordoli, R. S.; Elliott, G. J.; Sedgwick, R. D.; Tyler, A. N. Biomed. Mass Spectrom. 1981, 8, 492 28 Aberth, W.; Straub, K. M. ; Burlingame, A. L . Anal. Chem. 1982, 54, 2029 29 Macfarlane.R. D.; Torgerson, D. F. Science 1976, 191, 920 30 Honig, R. E. "Laser Interaction and Related Plasma Phenomena" Plenum Press, NY, 1971, p85 31 Roboz, J. " Introduction to Mass Spectrometry Instrumentation and Techniques" Interscience Publishers, N. Y. , 1968, p26 32 Blanth, E . W. "Dynamic Mass Spectrometers ", Elsevier, Amsterdam, 1966 33 Beauchamp, J. L . Ann. Rev. Phys. Chem. 1971, 22, 527 34 Harada, K. -L; Ito, S.; Takade, N.; Suzuki, M. ; Tatematsa, A. Biomed. Mass Spectrom. 1983, 10, 5 35 Reinhold, V. N.; Carr, S. A. Anal. Chem. 1982, 54, 499 36 Liebl, H. Anal. Chem. 1974, 46, 22A 37 Benninghoven, A. Surf. Sci. 1973, 35, 427; Surf. Sci. 1975, 53, 596 - 104-38 Day, R. J.; Unger. S. E. ; Cooks, R. G. / . Am. Chem. Soc. 1979, 50, 1180 39 Devienne, F. M . ; Roustan, J. C Org. Mass Spectrom. 1982,17, 173 40 Barber, M . ; Bordoli, R. S.; Elliott, G. J.; Sedgewick, R. D.; Tyler, A .N. Anal. Chem. 1982, 54, 645A 41 Field, F. FL; J. Phys. Chem. 1982, 26, 5115 42 Johnstone, R. A. W.; Lewis, I. A. S.; Rose, M . E . Tetrahedron 1983, 9, 1597 43 , Puzo, G.; Prome, J.-C. Org. Mass Spectrom. 1984, 19, 448 44 Caprioli, R. M . Anal. Chem. 1983, 55,2387 45 Smith, L . A.; Caprioli, R. M . Biomed. Mass Spectrom. 1984, 11, 392 46 Barber, M . ; Bordoli, R. S.; Elliot, G. J.; Sedwick, R. D.; Tyler, N. / . Chem. Soc, Faraday Trans. 1 1983,79,1249 47 De Pauw, E.; Pelzer, G.; Dao, V. D.; Marien, J. Biochem. Biophys. Res. Comm. 1984, 123, 27 48 Gower, J. L . Biomed. Mass Spectrom. 1985, 12, 191 49 Arita, M . ; Iwamori, M . ; Higuchi, T.; Nagai, Y . / . Biochem. 1983, 93, 319 50 Martin, S. A.; Costello, C. E.; Biemann, K. Anal. Chem. 1982, 54, 2362 51 Forsbury, L. S.; Dell, A.; Walton, D. J.; Ballou, C. E. J. Biol. Chem. 1982, 257, 3555 52 Dell, A.; Ballou, C. E. Biomed. Mass Spectrom. 1983,10, 50 53 Dell, A.; Ballou. C. E. Carbohydr. Res. 1983,120, 95 54 Ready, J. F. " Effect of High-Power Laser Radiation" Academic Press, N. Y.; 1971 55 Rollgen, F. W.; Schulten, H. R. Z. Naturforch. 1975, 30a, 1683 56 Hillenkamp, F.; Unsold, E.; Konfmann, R.; Nitsche, R. Appl. Phys. 1975, 8, 341 57 Heresch, F.; Schmid, E. R.; Huber, J. F. K. Anal. Chem. 1980, 52, 1803 58 Hardin, E. D.; Vestal, M . L . "Proceedings 28th Annual Conference on Mass Spectrometry and Allied Topic" ASMS, N. Y. 1980, p616 59 Davis, G . D. "Advances in Mass Spectrometry" 1980, 8A, 1012 60 Posthumus, M . A. Anal. Chem. 1978, 50, 985 -105-61 Comisarow, M . B J . Chem. Phys. 1978, 69, 4097 62 Marshall, A. G.; Comisarow, M . B. Chem. Phys. Lett. 1974, 25, 282 63 Marshall, A. G.; Comisarow, M . B. Chem. Phys. Lett. 1974, 26,486 64 Comisarow, M . B. Adv. Mass Spectrom. 1976, 7, 1042 65 Comisarow, M . B. Adv. Mass Spectrom. 1981, 8, 1698 66 McCrery, D. A.; Ledford, E. B. Jr.; Gross, M . L. Anal. Chem. 1982, 54, 1435 67 Coates, M . L.; Wilkin, C. L . Biomed. Mass Spectrom. 1985, 12, 8, 424 68 Coates, M . L.; Wilkin, C. L. Biomed. Environ. Mass Spectrom. 1986, 13, 199 69 Coates, M . L.; Wilkin, C. L . Anal. Chem. 1987, 59, 197 70 Ledford, E. B. Jr.; Rempel, D. L.; Gross, M . L . Anal. Chem. 1984, 56, 2744 71 Freiser, B. S. Anal. Chim. Acta 1985, 178, 125 72 Paul, W.; Raether, M . Z. Physik 1955, 140, 262 73 Hammer, C. -G. ; Holmstelt, B.; Ryhage, R. Anal. Biochem. 1968, 25, 532 74 Haworth, W. N. J. Chem. Soc. 1915, 107, 8 75 Purdie, T.; Irvin, J. C. J. Chem. Soc. 1903, 83, 1021 76 Kuhn, R.; Trischaman, H.; Low, I. Angew. Chem. 1955, 67, 32 77 Hakomori, S. I. / . Biochem. (Tokyo) 1964, 55, 205 78 Prehm, P. Carbohydr. Res. 1980, 78, 372 79 Jenning, H. J.; Lugowski, C. / . Immunol. 1981, 127, 1011 80 Bjorndal, H.; Hellerquist, C. G.; Lindberg, B.; Svensson, S. Angew. Chem. Int. Ed. Engl. 1970, 9, 610 81 Lindberg, B. Methods Enzymol. 1972, 28, 178 82 Adams, M . "Bacteriophages", Interscience Publishers Inc., N.Y. 1957 83 Douglas, J. "Bacteriophages", Chapman and Hall, London, 1975 84 Mattews, C. K. "Bacteriophages Biochemistry", Van Nostrand Reinhold Co., N.Y. 1971 85 Lindberg, A. A. in Sutherland. I. (Ed.), "Surface Carbohydates of the Prokaryotic Cell" Academic Press, London, 1977, p289 -106-86 Rieger-Hug, D.; Stirm, S. Virology 1981, 113, 363 87 Geyer, H.; Hirnmelspach, K.; Kwiatkowski, N.; Schlect, S.; Stirm, S. Pure Appl. Chem. 1983, 55, 637 88 Dutton, G. G. S.; Folkman, T. E . Carbohydr. Res. 1980, 78, 305 89 Kaurnaratne, D. N. Ph.D. Thesis, 1985 90 Dell, A. personal communications 91 Dell, A.; Tiller, P. R. Biochem. Biophys. Res. Comm. 1986, 135, 3, 1126 92 de Jong, E. G.; Heerma, W.; Sicherer, C. Biomed. Mass Spectrom. 1979, 6, 242 93 de Jong, E. G.; Heerma, W.; Dijkstra, G. Biomed. Mass Spectrom. 1980, 7, 127 94 Lim. A. V. S. Ph. D. Thesis, 1986 95 Egge, H.; Peter-Katalinic, J. in Burlingame, A. L.; Castagonli, N. Jr.(Eds.), "Mass Spectrometry in the Health and Life Sciences", Elsevier Science Publishers, Amsterdam, 1985, p401 96 Oates, J.; Dell, A.; Fukuda, M. ; Fukuda, M . N. Carbohydr. Res. 1985, 141, 149 97 Kamerling, J. P.; Heerma, W.; Vliegenthart, J. F. G.; Green, B.; Lewis, I. A. S.; Strecker, G.; Spik, G. Biomed. Mass Spectrom. 1983, 10, 7, 420 98 Morris, H. R.; Dell, A.; Panico, M. ; McDowell, R. A. in Burlingame, A. L.; Castagonli, N. Jr.(Eds.), " Mass Spectrometry in the Health and Life Sciences", Elsevier Science Publishers, Amsterdam, 1985, p363 99 Dell, A.; Oates, J.; Morris, H. R.; Egge, H. Int. J. Mass Spectrom. Ion Phys. 1983, 46, 415 100 Hanfland, P.; Dabrowski, U.; Kordowicz, M. ; Egge, H.; Peter-Katalinic, J.; Dabrowski, J. in Chester, M . A.et al (Eds.), "Glycoconjugates", p414 101 Bojesen, G.; Moller, J. Int. J. Mass Spectrom. Ion Phys. 1986, 68, 239 102 Barber, M. ; Bordoli, R. S.; Elliott, G. J.; Sedgwick, R. D.; Tyler, A. N. / . Chem. Soc, Faraday Trans. I, 1983,79,1249 103 Ledmann, L.; Kessler, M . ; Konig, W. A. Biomed. Mass Spectrom. 1984, 11, 5, 217 -107-104 Sunner, J. A.; Kulatunga, R.; Kebarle, P. Anal. Chem. 1986, 58, 2009 105 Freas, R. B.; Ross, M . M. ; Campana, J. E . / . Am. Chem. Soc 1985, 107, 6195 106 Clayton, E.; Wakefield, A. J. / . Chem. Soc, Chem. Commun. 1984, 969 107 Wong, S. S.; Rollgen, F. W.; Manz, I.; Przybylski, M . Biomed. Mass Spectrom. 1985, 12, 43 108 Schronk, L . R.; Cotter, R. J. Biomed. Environ. Mass Spectrom. 1986, 13, 395 109 Cooks, R. G.; Busch, K. L . Int. J. Mass Spectrom. Ion Phys. 1983,53, 111 110 Renner, D.; Spiteller, G. Biomed. Environ. Mass Spectrom. 1986,13, 401 111 Ackerman, B.; Watson, J. T.; Holland, J. F. "Proceedings of33tn Annual Conference on Mass Spectrometry and Allied Topics" ASMS 1985, p968 112 Stoch,, J. G.; Cooks, J. C ; Rhinehart, Jr. K. L. Anal Chem. 1985, 57, 985 113 Her, G.; Santikan, S.; Reinhold, V . N. "Proceedings of34tn Annual Conference on Mass Spectrometry and Allied Topics " ASMS 1986, p353 114 Caproili, R. M . ; Fan, T.; Cottrell, J." Proceedings of34tn Annual Conference on Mass Spectrometry and Allied Topics " ASMS 1986, p381 -108-A P P E N D I X -109-A P P E N D I X The following experiments were also carried out. They were not successful for the following stated reasons. 1) FABMS of permethylated Klebsiella K67 oligosaccharide derived from chemical degradation, with thioglycerol as the matrix. Medium intensity peaks were observed at around 870 amu and 600 amu. Due to the fact that the peaks were not resolved, accurate masses could not be determined. Nevertheless, these two peaks are likely to be the pseudo-molecular ion and the corresponding minus one hexose ion as the molecular weight of the permethylated oligosaccharide is 876 amu. 2 ) FABMS of reduced, permethylated Klebsiella K44 phage degraded oligosaccharide with thioglycerol as the matrix. Two medium intensity peaks were observed at around 1030 amu, these two peaks are likely to be the pseudo-molecular ions as the molecular weight of the permethylated oligosaccharide is 1036 amu. No sequence ions were observed. 3 ) FABMS of Klebsiella K46 phage degraded oligosaccharide with thioglycerol as the matrix. Two medium intensity peaks at around 1020 amu, which are possiblity the pseudo-molecular ions (M + N a + - pyr, m/z 1023 and M + H + - pyr, m/z 1005). No sequence ions were observed. -110-4) FABMS of permethylated Klebsiella K46 phage degraded oligosaccharide with thioglycerol as the matrix. No sample peaks were observed due to insufficient sample. 5) DCIMS of permethylated Klebsiella K67 oligosaccharide derived from chemical degradation. Due to insufficient sample very weak sample peaks were observed, which are too weak for accurate assignment 6) DCIMS of reduced, permethylated Klebsiella K44 phage degraded oligosaccharide. Due to insufficient sample very weak sample peaks were observed, which are too weak for accurate assignment 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0059809/manifest

Comment

Related Items