UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Structural polysaccharide chemistry with analytical applications to lemon gum Gibney, Kelly Blair 1971

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

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata


831-UBC_1971_A1 G53.pdf [ 11.59MB ]
JSON: 831-1.0060083.json
JSON-LD: 831-1.0060083-ld.json
RDF/XML (Pretty): 831-1.0060083-rdf.xml
RDF/JSON: 831-1.0060083-rdf.json
Turtle: 831-1.0060083-turtle.txt
N-Triples: 831-1.0060083-rdf-ntriples.txt
Original Record: 831-1.0060083-source.json
Full Text

Full Text

STRUCTURAL POLYSACCHARIDE CHEMISTRY WITH ANALYTICAL APPLICATIONS TO LEMON GUM by KELLY BLAIR GIBNEY B.Sc.; University of British Columbia, 1963 M.Sc, University of British Columbia, 1967 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in the Department of Chemistry We accept this thesis as conforming to xhe required standard THE UNIVERSITY OF BRITISH COLUMBIA April, 1971 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of Br i t ish 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 representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of CMB*\I*TM The University of Br i t ish Columbia Vancouver 8, Canada Date AfglL 3 1 , Mf\ ABSTRACT Chairman: Professor G.G.S. Dutton A gas chromatographic method has been developed whereby the following compounds may be determined quantitatively and simultaneously: ethylene glycol, glycerol, erythritol, threitol, arabinose, xylose, monnose, galactose and glucose. These products correspond to the constituent parts of periodate oxidized, borohydride reduced polysaccharides. Quantitative analysis of the fractions obtained on total and partial hydrolysis of such polyalcohols pro-vides a new and powerful analytical tool in polysaccharide chemistry. Optimum conditions for the Smith degradation of a polysaccharide have been illustrated by a detailed analysis of lemon gum. A new method for the reduction of uronic acids in poly- and oligosaccharides has been developed. Reduction can be carried out after only esterification of the uronic acids. Reaction at this stage avoids possible degradation or fractionation during esterification and subsequent saponification of poly-saccharide hydroxyl groups. Good levels of reduction [75$] have been achieved with reduction of highly branched lemon gum. Silylated polyhydroxy compounds and galacturonic acid were examined by proton magnetic resonance spectroscopy. The trimethylsilyl [TMS] protons of such ethers and esters show good separation from one another. Analysis of the trimethylsilyl proton peaks provides information concerning the number and degree of subsitution of hydroxyl and acid functions present in the parent molecule. Signal strength due to the nine protons on each trimethyl s i l y l group provides a chemical means of signal enhancement. Analysis was carried out on. 2 mg. of a minor TMS glucofuranose derivative. [ I l l ] Lemon gum has been analysed by the method of Lindberg et a l [8l]. Analysis of lemon gum is the fi r s t example of the application of partially methylated alditol acetate mass spectroscopy to a plant gum. The results of this analysis are in good agreement with previous analysis of this gum. [iv] TABLE OF CONTENTS Pap.e No. 'INTRODUCTION . . . .' 1 PERIODATE OXIDATION OF CARBOHYDRATES AND EVALUATION OF REACTION PRODUCTS BY GAS LIQUID CHROMATOGRAPHY 2 PERIODATE OXIDATION IN STRUCTURAL STUDIES OF POLYSACCHARIDES 2 SIMULTANEOUS ESTIMATION OF POLYHYDRIC ALCOHOLS AND SUGARS . 5 CONTROL DURING THE SMITH DEGRADATION l6 THE OPTIMUM HYDROLYSIS OF OXIDIZED LEMON GUM 19 SMITH FRAGMENTS 28 REDUCTION OF URONIC ACIDS JO METAL HYDRIDE REDUCING AGENTS . 56 EVALUATION OF NEW ROUTE TO REDUCED POLYSACCHARIDES 59 SIGNAL ENHANCEMENT USING TRIMETHYLSILYL [TMS] DERIVATIVES AND UTILIZATION OF TMS AS A PROBE FOR DETERMINATION OF NUMBER OF HYDROXYLS IN A COMPOUND 1+5 SIGNAL ENHANCEMENT k-9 TRIMETHYLSILYLATION OF GALACTURONIC ACID 51 METHYLATION OF LEMON GUM 62 MASS SPECTROMETRY OF METHYLATED POLYOL ACETATES 67 METHYLATION ANALYSIS OF LEMON GUM 71 METHYLATED POLYOLS FROM LEMON GUM 7k METHYLATED DEGRADED LEMON GUM 80 METHYLATION ANALYSIS OF LEMON GUM AFTER PERIODATE OXIDATION. AND BOROHYDRIDE REDUCTION . 82 Cv] TABLE OF CONTENTS [Cont'd] Pafte No. METMLATION OF LEMON GUM [Cont'd] METHYLATION OF LEMON GUM SI 85 EXPERIMENTAL . . . 88 GENERAL METHODS • 88 I] HAKOMORI METHOD 95 II] PURDIE METHOD 95 III] KUHN METHOD . 96 METHYLATION OF OLIGO-AND MONOSACCHARIDES 96 PURDIE AND KUHN METHODS . 96 HAKOMORI METHOD . . . . . . . . . 97 PERIODATE OXIDATION 97 ESTIMATION OF PERIODATE 98 SIMULTANEOUS ESTIMATION OF POLYHYDRIC ALCOHOLS AND SUGARS . 98 PURIFICATION OF LEMON GUM . . . . 101 OPTIMUM HYDROLYSIS OF LEMON GUM - PERIODATE DEGRADED . . . 102 SMITH FRAGMENTS 105 • THE REDUCTION OF ACIDIC POLYSACCHARIDES . . . . ' 1 0 5 ESTERIFICATION 105 REDUCTION OF ETHYLENE GLYCOL ESTER OF LEMON GUM . . . 106 SIGNAL ENHANCEMENT 107 TRIMETHYL SILYLATION OF GALACTURONIC ACID . . . . . . . . . 107 METHYLATION ANALYSIS 108 [vi] TABLE OP CONTENTS. [Cont'd] To,f\e No. APPENDIX I . . . . 110 ESTIMATION OF DEGREE OF POLYMERIZATION 110 EXPERIMENTAL 112 * APPENDIX II . . . . . . . . . . •. i Ilk ENZYMATIC HYDROLYSIS OF LEMON GUM Ilk EXPERIMENTAL 117 ENZYMATIC HYDROLYSIS OF LEMON GUM 117 FRACTIONATION OF PECTINASE, CELLULASE AND HP-150 . . . 118 PECTINOL kl-V CONCENTRATE . . . . . . Il8 CELLULASE 36 . . . . . . . 119 HP-150 120 APPENDIX III . .• 121 CARBON-CARBON BOND CLEAVAGE, DIALYSIS AND DESTRUCTIVE . PURIFICATION, AN UNDERGRADUATE ORGANIC EXPERIMENT 121 CHARACTERIZATION 125 APPENDIX IV 126 LEMON GUM: DEGRADATION WITH 0.1U SULFURIC ACID 126 EXPERIMENTAL 126 AUTOHYDROLYSIS OF LEMON GUM 126 EXPERIMENTAL . . 128 OLIGOSACCHARIDES 128 APPENDIX V 130 A TECHNIQUE FOR HANDLING MICRO SAMPLES COLLECTED FROM THE GAS CHROMATOGRAPH , ., 130 [vii] TABLE OF CONTENTS [Cont'd] Pape No. APPENDIX VI 131+ COMPUTER SEARCHING THE CURRENT CARBOHYDRATE LITERATURE . . 1J'+ PROFILE NUMBER 2, EVALUATION ' 139 PROFILE NUMBER 3 1^ 1 APPENDIX VII . . . 150 LEMON GUM ,. . 150 1/lON H2S04 LEMON GUM 160 SMITH POLYALCOHOL LEMON GUM . . . . . . 163 LEMON GUM SI 167 BIBLIOGRAPHY 172 [ v i i i ] LIST OF TABLES Table No. Pa.qe Wo. TABLE I Possible Reaction Products from Periodate Oxidized Polysaccharides 6 TABLE II Analysis of Reaction Products from an Arabinoxylan 8 TABLE III Ana,lysis of Reaction Products of a Glucomannan . . 10 TABLE TV Analysis of Reaction Products of a Galactoglucomannan 10 TABLE V Analysis of Threitol and Erythritol Mixtures . . . 10 TABLE VI Analysis of Simulated Crude Hemicellulose Hydrolysates 10 TABLE VII Molar Response Factors of Polyols and Sugars . . . 12 TABLE VIII Composition of Peaks in Gas Pha.se Chroma, to gram of a Mixture of Arabinose, Xylose, Galactose, Glucose and Mannose 15 TABLE IX Number of Sugar Anomers Detected by Gas-Liquid Chromatography and the Percentage Composition of Equilibrium Solutions 15 TABLE X Analysis of Lemon Gum Smith Polyalcohol and its Hydrolysis and Oxidation Products 26 TABLE XII Analysis of the Polysaccharides Obtained from Alginic Acid by Different Reduction Procedures . . 55 TABLE XIII Carbohydrate Analysis of Lemon Gum and its Reduced Product K2 TABLE XTV Yields and Degrees of Methylation of Various Polysaccharides 65 [ix] LIST OF TABLES [Cont'd] Table No. Paffe No, TABLE XV Recently Developed Methylation Procedures . . . . 63 TABLE XVI Methylation Analysis of Lemon Gum 79 TABLE XVII Methylation Analysis of Degraded Lemon Gum . . . . 8 l TABLE XVIII Methylation Analysis of Lemon Gum Smith Polyalcohol 83 TABLE XIX Methylation Analysis of Lemon Gum SI 86 TABLE XX Anomeric Ratios for BTMSA Silylation of Galacturonic Acid in Different Solvents 108 TABLE XXI Enzyme Preparations Examined for Activity on Lemon Gum Ilk TABLE XXII Variable Steps in Preparation of Erythritol . . . 122 TABLE XXIII Oligosaccharides Isolated on Hydrolysis of Lemon Gum 127 TABLE XXIV Neutral and Acidic Components Obtained on Hydrolysis of Lemon Gum Oligosaccharides . 129 TABLE XXV Computer Print Out of Profile One 136 TABLE XXVI Computer Print Out of Profile Two 138 TABLE XXVII Computer Print Out of Profile Three IkO TABLE XXVIII Computer Print Out of Profile Four Ikl TABLE XXIX Evaluation of Search Profiles ikk [x] LIST OF FIGURES Figure No. Pap;e No. FIGURE 1 Model Periodate Oxidation Products Before and After Reduction k FIGURE 2 Separation of Products as Trimethylsilyl Derivatives from an Arabinogalactan . . . . . . . . 8 FIGURE 3 Separation of Products as Trimethylsilyl Derivatives from an Arabinoxylan 8 FIGURE k Separation of Products as Trimethylsilyl Derivatives, from a Glucomannan 9 FIGURE 5 Separation of Products as Trimethylsilyl Derivatives from Galactoglucomannan 9 FIGURE 6 Separation of Products as Trimethylsilyl Derivatives from Crude Hemicellulose 9 FIGURE 7 Gas Chroma to grams of Trimethylsilyl Derivatives from Lemon Gum Smith Polyalcohol Hydrolysed With Various Sulfuric Acid Strengths for 2k Hours at 20°C 21 FIGURE 8 Gas Chromatograros of Trimethylsilyl Derivatives from Lemon Gum Smith Polyalcohol Hydrolysed for Different Periods of Time With 0.5N Sulfuric Acid at 20°C 23 FIGURE 9 Relative G.L.C. Peak Areas for Components 2,3,4 from Lemon Gum Smith Polyalcohol 2k FIGURE 10 Relative G.L.C. Peak Areas for Components 1 [Glycerol] and 5 [Arabinose] from Lemon Gum Smith Polyalcohol 2k Cxi] . LIST OF FIGURES [Cont'd] Figure No. Page No. FIGURE 11 Gas Chromatogram of Lemon Gum Smith Polyalcohol Hydrolysed for 32 Hours With 0.5N Sulfuric Acid at 20°C After Dialysis 25 FIGURE 12 Gas Chromatogram of the Supernatant From the Isolation of Lemon Gum SI Before and After Total Hydrolysis 30 FIGURE 13 The Reduction of Esters to Ethers with an Excess of Diborane 32 FIGURE 14 The /S -Elimination Mechanisms for Uronic Acids . kl FIGURE 15 100 MHz P.M.R. Spectrum for the Eryrthritol Protons in Trimethylsilylated Erythritol k6 FIGURE 16 Compounds Subjected to P.M.R. [100 MHz] Analysis of Trimethylsilyl Protons 47 FIGURE 17 Proton Magnetic Resonance Spectra for the T r i -methylsilyl Protons of Silylated Galacturonic Acid 54 FIGURE 18 P.M.R. Spectrum of Trimethylsilylated Galacturonic Acid [G.L.C. Peak 1] 55 FIGURE 19 P.M.R. Spectrum of Trimethylsilylated Galacturonic Acid [G.L.C. Peak 2] 56 FIGURE 20 P.M.R. Spectrum of Trimethylsilylated Galacturonic Acid [G.L.C. Peak 4] 57 FIGURE 21 Mass Spectrum of the Trimethylsilylation Product of Galacturonic Acid [G.L.C. Peak 1] 58 [xii] LIST OF FIGURES [Cont'd] Figure No. - Page No. FIGURE 22 Mass Spectrum of the Trimethylsilylation Product of Galacturonic Acid [G.L.C. Peak 2] 59 FIGURE 23 Mass Spectrum of the Trimethylsilylation Product of Galacturonic Acid [G.L.C. Peak 3] 60 FIGURE 24 Mass Spectrum of the Trimethylsilylation Product of Galacturonic Acid [G.L.C. Peak k] 6l FIGURE 25 Gas Chroma to gram of Lemon Gum After Methylation, Lithium Aluminum Hydride Reduction, Hydrolysis, Borohydride Reduction and Acetylation 72 FIGURE 26 Thin Layer Chromatogram of the Ten Fractions of Lemon Gum Methylated Polyol Acetates 72 FIGURE 27 Gas Chromatogram of Degraded Lemon Gum Methylated Polyol Acetates 80 FIGURE 28. Gas Chromatogram of Lemon Gum Smith Polyalcohol Methylated Polyol Acetates . . . 83 FIGURE 29 Gas Chromatogram of Lemon Gum SI Methylated Polyol Acetates '. 86 FIGURE 30 Gas Chromatogram of Glucitol Hydrolysed with 1.0N H2S04 for 2k Hours 111 FIGURE 31 Gas Chromatogram of Glycitol Showing Undersilylation After 10 Min. of Reaction . 112 FIGURE 32 Collection of Gas Chromatography Samples in Glass Capillaries and Oven Design for Their Micromanipulation 132 [ x i i i ] LIST OF FIGURES [Cont'd] Fip<u.re No. Pap;e No. FIGURE 33 Oven and Gas Supply for Micromanipulation . . . . 133 FIGURE 34 Lemon Gum Peak 3 .- 150 FIGURE 35 Lemon Gum Peak 4 151 FIGURE 36 Lemon Gum Peak 5 152 FIGURE 37 Lemon Gum Peak 6B . . . . . . . . . . 153 FIGURE 38 Lemon Gum Peak 6T . . . . . . . 154 FIGURE 39 Lemon Gum Peak 7B 155 FIGURE 40 Lemon Gum Peak 7T . 156 FIGURE 41 Lemon Gum Peak 8 157 FIGURE 42 Lemon Gum Peak 9 158 FIGURE 43 Lemon Gum Peak 10 l/lON H2S04 Lemon Gum Peak 4 Smith Polyalcohol Lemon Gum Peak 6 Lemon Gum SI Peak 6 159 FIGURE 44 l/lON HsSO* Lemon Gum Peak 1 160 FIGURE 45 l/lON HeS04 Lemon Gum Peak 2 . . . l 6 l FIGURE 46 l/lON H£S0 4 Lemon Gum Peak 3 162 FIGURE 47 Smith Polyalcohol Lemon Gum Peak 2 . 163 FIGURE 48 Smith Polyalcohol Lemon Gum Peak 3 164 FIGURE 49 Smith Polyalcohol Lemon Gum Peak 4 165 FIGURE 50 Smith Polyalcohol Lemon Gum Peak 5 166 FIGURE 51 Lemon Gum SI Peak 1 167 FIGURE 52 Lemon Gum SI Peak 2 . . . . . . . . . 168 FIGURE 53 Lemon Gum SI Peak 5 . . . . . ' 169 FIGURE 54 Lemon Gum SI Peak 4 170 • FIGURE 55 Lemon Gum SI Peak 5 . . . 171 ACKNOWLEDGEMENTS The author wishes to express his grateful ap-preciation for the help, encouragement and advice given him by Professor G.G.S. Dutton, who directed the work described in this thesis. He wishes to thank Dr. P.E. Reid, Dr. J.J.M. Rowe, Dr. K. Hunt, Dr. L.D. Hall and Dr. J.N.C. Whyte for their valued counsel, the University of British Columbia for the award of the MacMillan Bloedel Graduate Scholarship and Miss N. Housenga for her diligence during the preparation of this manuscript. [xv] To my wife. 1 INTRODUCTION Lemon gum is a highly branched polysaccharide exudate. During the course of structural investigations of this polymer, a detailed structure was published by Jones and Stoddart [ l j . Although work continued to con-firm the structure proposed, the direction of investigation was altered. Methods were designed to improve the quality of polysaccharide structural determinations while decreasing the quantities of material required. The isolation of pure polysaccharides has been greatly stimulated by several important improvements in methods and techniques. Investigators can now isolate small quantities of extremely pure polysaccharides using preparative moving boundary electrophoresis, starch block electrophoresis, ultracentrifugation using density gradients, gel permiation chromatography, and various types of ion exchange chromatography. The general drawback of these methods is an inability to isolate large quantities of polymer. This problem can be further complicated by a lack of available sample. Biological materials may contain polysaccharides in low concentrations or as very com-plex mixtures. Time can be a c r i t i c a l factor since a glyco protein may de-nature. However, while the techniques of polymer isolation have improved so have techniques for the investigation of small quantities of carbohydrates. In the ensuing sections new methods detailing improvements in techniques already applied to carbohydrates and analytical procedures not previously described will be presented. Their, application to model compounds and lemon gum [as an example of a complex polysaccharide] w i l l be developed. PERIODATE OXIDATION OF CARBOHYDRATES AND EVALUATION OF REACTION PRODUCTS BY GAS LIQUID CHROMATOGRAPHY PERIODATE OXIDATION IN STRUCTURAL STUDIES OF POLYSACCHARIDES Periodate oxidation results In a specific attack on a carbohydrate polymer. Periodic acid, or its salts w i l l oxidize vicinal hydroxy 1 functions, the cleaved glycols forming a dialdehyde. The existence of three contiguous hydroxyls v i l l result in the formation of a mole of formic acid from the central carbon atom. Periodate oxidation is com-plicated by steric effects, electronic effects and over oxidation. Since this field has been v e i l outlined in several reviews [2,3 ] i t v i l l be sufficient to say that the dialdehyde that results from an oxidation forms an extremely complex system. Direct hydrolysis of the dialdehyde is virtu-ally impossible to interpret. Severe degradation of the small component units formed in the oxidation occurs. In previous work in this laboratory [ 4 ] , experiments aimed at preparation of partially methylated tetroses by oxidation between C-2 and C-3 of two hexoses were unsuccessful because of side reactions, lack of hydrolysis and degradation. I f this simple system does not yield clear results, the concept of analysis of complex polysac-charides at the polyaldehyde stage becomes impossible. The Barry degradation [ 5] was developed to analyse the polyaldehyde oxidation product. The polyaldehyde was treated with phenylhydrazine in dilute acetic acid solution. An insoluble product formed, containing ap-proximately one molecule of phenylhydrazine condensed with each dialdehyde group. This complex broke down when heated in an aqueous or ethanolic solution 3 of phenylhydrazine and acetic acid. The products were phenylosazones and unoxidized parts of the polysaccharide a l l of which may be Identified. There were, however, two areas of difficulty with this procedure. The products of the oxidized portion of the molecule were osazones and hence structural asymmetry of a portion of the molecule was lost. Secondly i t was relatively difficult to purify and quantitatively estimate the osazones produced. A method was developed by Smith [ 6] to simplify the interpretation of results from periodate oxidation of polysaccharides. The procedure involved reduction with metal borohydrides and subsequent controlled hydrolysis of the polyalcohol which was formed. Dialdehydes formed on periodate oxidation do not exist as free aldehydes but like sugars themselves, undergo cyclization and hydration [ 7 ] * Some examples of cyclized model compounds [I, II, III] which have been analysed are shown in Figure 1. Alcohols such as IV are readily hydrolysed with dilute mineral acid at room temperature. Cyclic acetals such as I, II, and III are relatively stable. Smith realized that instability of alcohols such as IV toward dilute acid at room temperature, compared with glycosides, provided a new and more powerful tool for controlled degradation of polysaccharides. When a sugar moeity of a polysaccharide is cleaved by periodate and reduced, the resulting alcoholic derivative, being a true acetal, is sensitive to acid. Sugar units which are not cleaved and are attached glycosidicly to cleaved or unattached units remain intact being relatively stable to acid. Using the marked differences in stability of glycosides over true acetals a vide variety of oligosaccharides have been obtained. The structures of the stable units provide an indication of the polysaccharide subunits. FIGURE 1 MODEL PERIODATE OXIDATION PRODUCTS BEFORE AND AFTER REDUCTION IE W Named the Smith degradation this procedure has been reviewed and utilized extensively [ 8, 9, 10, 11]. Periodate oxidation, reduction and total hy-drolysis of a polysaccharide, often incorrectly called a Smith degradation, also provides a considerable source of information regarding fine structure in the polymer. Two limitations exist which detract from the use of total or partial hydrolysis of the Smith polyalcohol. One of these problems is the inability of standard analytical techniques to determine quantitatively the totally reduced products of the oxidation, that is to say ethylene glycol, glycerol, erythritol and threitol. Ethylene glycol is especially difficult to qual-itatively identify on paper Cihrona to grams because of diffusion in chromato-graphic solvents, streaking and lack of reaction. The second problem is the 5 wide variety of rates of hydrolysis of different glycosidic linkages and different tx-ue acetals. These factors cause difficulties in carrying out partial hydrolysis. The solutions to these problems w i l l be outlined in the following two subsections. They w i l l demonstrate a whole new approach and usefulness of Smith oxidations involving partial and total hydrolysis. SIMULTANEOUS ESTIMATION OF POLYHYDRIC ALCOHOLS AND SUGARS As mentioned before, one method of investigating the structure of a polysaccharide involves periodate oxidation, borohydride reduction and com-plete hydrolysis [12]. For example, in the case of a linear 1 —5-4 linked pentosan, the products w i l l be ethylene glycol from the non reducing end and glycerol from the interior units. When any of the interior units carry side chains these units are immune to periodate oxidation and thus subsequent hydrolysis gives monosaccharides as well as polyhydric alcohols. Much useful information can be obtained from a knowledge of the ratio between non reducing end groups, interior units and branch points. Two different analytical methods have normally been used in this connection after separation of the components by paper or thin-layer chromatography. Thus, polyhydric alcohols have been estimated by chronotropic acid [13] and monosaccharides by phenol sulfuric [14] or other colorimetric method. The necessity of using two distinct analytical procedures in addition to the separations involved is tedious and introduces possibilities of error. The present work reports a method using gas-liquid chromatography [G.L.C] whereby mixtures of polyhydric alcohols and monosaccharides may be analysed simultaneously and accurately. 6 TABLE I POSSIBLE REACTION PRODUCTS FROM PERIODATE OXIDIZED POLYSACCHARIDES Polysaccharide Possible Products A rab inoxylan Ethylene glycol, glycerol, xylose, arabinose Glucomannan Glycerol, erythritol, glucose, mannose Galactoglucomannan Glycerol, erythritol, threitol, glucose, mannose, galactose Arab ino galactan Ethylene glycol, glycerol, threitol, arabinose, galactose This investigation arose from the need to have a simple method for the analysis of polysaccharides occurring in wood and plant gums. Table I illustrates the possible products from four such typical polysaccharides. The method involves transformation of the mixture of sugars and poly-hydric alcohols resulting from the periodate oxidation of a polysaccharide into derivatives sufficiently volatile for separation by G.L.C. In order to test the viability of the method, several test mixtures were prepared cor-responding to the possible products shown in Table I. For each system studied the relative proportions of the components were varied over wide limits thus covering the different situations encountered in various naturally occurring polysaccharides. The components were separated as their t r i -methylsilyl [TMS] derivatives [15] on a column employing SF-96 as the liquid phase. 7 The separation of the products arising from an arabinogalactan is shown in Figure 2 and the data obtainable from an arabinoxylan are given in Table II with the separation shown in Figure 3. Similarly the simulated gluco-mannan is represented by Table III and Figure k. In a like manner the simu-lated galactoglucomannan system is shown in Table IV and Figure 5. These results need l i t t l e explanation except in the last case. According to the current view [16 ] of the structure of a wood galactoglucomannan one would not expect galactose to survive the periodate oxidation. However, galactose was included as a check on hexose analysis. The model experiments reported here were an opportunity to test the validity of the proposed method and therefore the most general cases were studied. It is clearly seen from the figures that the separations are ex-cellent and the agreement between calculated and found values is good. The worst separation in the present study is that between threitol and erythritol in the galactoglucomannan system. But even here i t is possible to calculate the relative amounts [17] in fair agreement with theory [Table V]. This problem w i l l only commonly arise where a polysaccharide contains galactose units linked 1 - 4 together with glucose and/or mannose similarly linked. Examination of Figure 6 w i l l show that the pentoses have a lower re-tention time [R^ ,] than the hexoses. This means that the same column and procedure may be used to separate a l l of the polyhydric alcohols and sugars mentioned in Table I. More importantly i t means that this same system may be used to separate and estimate a l l five of the sugars commonly occurring in wood polysaccharides. This separation has now been described by others whose results appeared while our work was in progress [ 18-21 ]. Particularly FIGURE 2: SEPARATION OF PRODUCTS AS TRIMETHYLSILYL DERIVATIVES FROM AN ARABINOGALACTAN ETHYLENE GLYCOL BUTANE-1,4-OIOL u THREITOL I -//• r-YA-r—ih ARABINOGALACTAN — i — 34 I 4 4 10 16 28 52 FIGURE 3: SEPARATION OF PRODUCTS AS TRIMETHYLSILYL DERIVATIVES FROM AN ARABINOXYLAN ARABINOXYLAN TABLE II ANALYSIS OF REACTION PRODUCTS FROM AN ARABINOXYLAN • % composition by weight Ethylene glycol Glycerol , Arabinose Xylose Found 15-6 18.8 29-3 36.0 Calculated . 15-8 19.0 .. 29.9 , 35-5 Found . 3°-5 38.1 29.2 " ' 3-15 Calculated 1 - 30.8 : 37-7 ' 28.9 3-44 Found 42.9 5-2° 3-31 48.5 Calculated 42.8 5-14 4.07 47.8 FIGURE k: SEPARATION OF PRODUCTS AS TRIMETHYLSILYL DERIVATIVES FROM A GLUCOMANNAN GLUCOMANNAN GLYCERCI. ERYTHRITOL J v. -If—i it 1 / / -16 26 43 60 •56 FIGURE 5: SEPARATION OF PRODUCTS AS TRIMETHYLSILYL DERIVATIVES FROM GALACTOGLUCOMANNAN GALACTOGLUCOMANNAN 43 46 50 FIGURE 6: SEPARATION OF PRODUCTS AS TRIMETHYLSILYL DERIVATIVES FROM CRUDE HEMICELLULOSE ETHYLENE GLYCOL GLYCEROL BUTANE-1,4-DIOL I ARABINOSE I v J l J u l j \ — - J u THRE ll ERYTH 1 — i t — r - " - i — o — i 1 1 ' 1 r— 0 4 10 16 26 34 38 43 46 50 56 10 TABLE I I I " ANALYSIS OF REACTION PRODUCTS OF A GLUCOMANNAN % composition by weight Glycerol Erythritol Mannose Glucose Found 19.2 9.20 35-5 36-4 Calculated 19.2 8.70 36.0 3<>.4 Found 3i-4 1.48 65.1 2.22 Calculated 33-2 1.46 62.8 3.11 Found 6.78 29.6 6.20 . 57-5 Calculated 6-34 28.6 5.85 5 9 * TABLE IV ANALYSIS OF REACTION PRODUCTS OF A GALACTOGLUCOMANNAN % composition by weight Glycerol Galactose Mannose Glucose Found 17.6 •• 25-7 29.38 28.2 Calculated ' 5 3 27.8 28.4 28.5 Found 4.12 , 53-2 39.1 3-71 Calculated 4.02 54-8 37-4 3-74 Found 4.08 2.85 38.6 54-7 Calculated 3-94 302 36.8 55-7 Found — 47-5 2.61 50.2 Calculated — - * 47-2 4.81 48.4 TABLE V -ANALYSIS OF THREITOL AND ERYTHRITOL MIXTURES % composition by weight Thrcitol Erythritol Found 54-5 455 Calculated 5i-5 48.5 Found 83.2 16.8 • Calculated 84.1 15-9 Found 20.8 79.2' Calculated 17-5 82.5 TABLE VI ANALYSIS OF SIMULATED CRUDE HEMICELLULOSE HYDROLYSATES % composition by weight Arabinose Xylose Galactose Mannose Glucose Found 17.9 . 23.1 , 20.4 •; 16.95 ' 22.0 Calculated 18.4 . ... 21.6 21.1 17-25 21.7 Found 17-75 22.6 ' 21.6 34-° 4.07 Calculated 18.2 21.7 21.3 34-4 4-37 Found 3-15 39.8 . 3.03 ' 16.62 39.0 Calculated 3.28 38.7 3-8i >-' 15-3 39-0 Found 33-8 4.07 38.5 3-64 ' 20.0 Calculated 33-6 3-85 38.9 3-94 19.92 11 notable Is the excellent Swedish paper to which further reference wil l be made [22J. The results presented in Table VI are further confirmation of the accuracy of the method used in this laboratory. The results quoted in Tables II - VI were obtained using an instrument with a thermal conductivity detector and systems containing free sugars. Each of these factors requires brief comment. Firstly, different substances cause such a detector to respond to a varying degree, thus for identical molar amounts of different compounds the areas under the peaks are not the same. It is therefore necessary to determine the Molar Response Factor [M.R.F.] for each compound to be determined. This is done relative to an internal standard, which in the present work was butane -1,4-diol. This compound is readily available, easily purified and has a convenient retention time, as Figure 6 shows. The view has recently been expressed that an inert com-pound is perferable [23] and terphenyl has been suggested. Although this compound has certain advantages, solubility in carbohydrate hydrolysis mix-tures can limit its use as a quantitative internal standard. When working with compounds of similar molecular complexity i t may be acceptable to assume that a l l the response factors are identical. However, when the compounds examined range from Ca to C 6 this assumption is incorrect [24], as shown in Table VTI. These values are presented here to show the vari ation which may be expected. It is our experience that the exact M.R.F. applic able to any analysis depends on the precise experimental conditions used. The value is influenced by such factors as the, rate of temperature programming. A contrary view has been expressed [25] but at this time i t would seem ad-visable for each worker to determine suitable response factors using exactly 12 the same experimental conditions as in the analytical determinations. It would be most unwise to assume without verification that the figures quoted in Table VII are valid for other systems. TABLE VII MOLAR RESPONSE FACTORS OF POLYOLS AND SUGARS Compound Molar Response Factor Ethylene glycol O.832 + 0.033 Butane -1,4-diol 1.00 + 0.000 Glycerol 1.26 + 0.014 Threitol 1.78 + 0.088 Erythritol 1.72 + 0.061 Arabinose 1.74 + 0.061 Xylose 1.77 + 0.064 Galactose 2.08 + 0.126 Glucose 2.02 + 0.091 Mannose 2.02 + 0.059 Secondly, in solution any one sugar exists as an equilibrium mixture of the anomars of the furanose and pyranose forms. When certain peaks overlap i t w i l l be necessary to use a peak corresponding to one anomeric form as a measure of the total amount of that particular sugar. This fact has been clearly discussed in the Swedish paper cited previously [22]. Since the r e l -ative equilibrium concentrations of the different forms are highly dependent on the solvent used [26 ] i t follows that the figures quoted in that paper [22 ] are only valid for that particular system. For accurate results i t is thus necessary for each worker to adopt a standardized procedure for the treatment 13 of polysaccharide hydrolysates. The composition of the sugar peaks obtained in the present work is given in Table VIII and shown in Figure 6. The equilibrium composition of the solutions is shown in Table XX. TABLE VIII COMPOSITION OF PEAKS IN GAS PHASE CHROMATOGRAM OF A MIXTURE OF ARABINOSE, XYLOSE, GALACTOSE, GLUCOSE AND MANNOSE Major Component Minor Component[s] Arabinose 1, 2, 3 Xylose 3 + 4 Mannose 1 Galactose 2 Galactose 3 Glucose 2 Xylose 1 + 2 Galactose 1 Mannose 2 Glucose 1 TABLE IX NUMBER OF SUGAR ANOMERS DETECTED BY GAS-LIQUID CHROMATOGRAPHY AND THE PERCENTAGE COMPOSITION OF EQUILIBRIUM SOLUTIONS Sugar Anonym Percentage Compos i t i o n a Arabinose 3 Xylose 4 1 + 2 3-58 + 0.28 3 + 4 96.4 + 0.28 Galactose 3 l 8.19 + O.80 2 29.3 + 1.27 3 62.6 + 1.02 Mannose 2 1 7^5 + 1.21 2 25.5 + 1.21 Glucose 2 b 1 38.3 + 0.33 2 61.7 + O.36 a Anomers numbered in order of elution. b It is possible that glucose shows a third anomer present in very small concentration. Ik The analysis of a periodate oxidized glucomannan by gas-liquid chrom-atography of the derived erythritol and glycerol acetates was reported in I960 by Bishop and Cooper [27]. More recently Zinbo and Timell [28] have used the s i l y l derivatives for the analysis of a xylan. In neither case has the fate of glycol aldehyde, obtainable from carbons one and two, been studied, although the former authors noted the existence of an unidentified peak with a greater retention time than glycerol triacetate and suggested this might be due to glycol aldehyde diacetate. In an attempt to discover what happens to the glycol aldehyde a solution of this compound in pyridine was silylated and examined using the same chrom-atographic conditions as before. There resulted a single peak indistinguish-able from glycerol and attributed to the dimer [29]. When an aqueous solution of glycol aldehyde was concentrated to dryness and silylated, two peaks were obtained. One of these had the same retention time as glycerol and the other a shorter [presumably the monomer]. If an aqueous solution of glycol aldehyde was f i r s t reduced with sodium borohydride only one peak corresponding to ethylene glycol was obtained. This suggested one way of dealing with the problem. When, however, synthetic mixtures of glycol aldehyde and the neutral products listed in Table I were analysed, several new and unexpected peaks were obtained. These peaks were assumed to be due to acetals formed between glycol aldehyde and the polyhydric alcohols [29]. These observations did not accord with those made on periodate oxidized amylose, where the amount of glycol aldehyde should equal the sum of glycerol and erythritol. Furthermore, the gas chromatographic analysis of the products from mesquite gum gave results in good agreement with those obtained by paper chromatographic separations [50]. 15 The difference in treatment of the model systems and the polysaccharides was an acid hydrolysis step. When the model systems containing glycol aldehyde -were heated at 100° for 2k hours with IN sulfuric acid and then an-alysed* no unidentifiable peaks were obtained. There was no change in the ratio of the sugars present. The hydrolysates were dark brown and contained a brown precipitate. There was a small increase in the glycerol concentration and a small decrease in ethylene glycol. The former is no doubt due to traces of glycol aldehyde dimer since in a separate experiment in which glycol aldehyde alone was treated with aoid a small peak occurred in the glycerol region. The apparent loss of ethylene glycol cannot at present be explained since i t was verified that this compound was not lost on ion exchange resins nor on concentration of its aqueous solutions. It is concluded that the glycol aldehyde is almost entirely destroyed in the total hydrolysis of the polyalcohol and is thus not a serious factor in such analyses. Where a highly accurate value for glycerol is necessary this may be obtained by reduction of the hydrolysate with borohydride before analysis to convert any remaining glycol aldehyde dimer to ethylene glycol. The f i r s t applications of gas-liquid chromatography to polysaccharide chemistry were reported for methylated sugars [31] because of their volatility. This technique is now widely applied and has also been used in conjunction with periodate degradation to facilitate the resolution of a complex mixture of isomeric sugars [32-34]. With the introduction of trimethylsilylation by Sweeley et a l . [15] this technique has been extended to the separation of sugars [18-22 ] and glycosylalditols obtained by Smith degradation [3^ and references therein]. Although the work reported here arose out of our interests 16 in wood polysaccharides and plant gums It is clear that the procedure is a general one and may be used for many other types of polysaccharides. It has already been applied with success to a study of sapote gum [35] and mesquite gum [30]. Work has been carried out to extend the method to systems con-taining deoxy and amino sugars [36] thus permitting the study of a wider ( variety of polysaccharides. , . CONTROL DURING THE SMITH DEGRADATION In addition to using gas-liquid chromatography for the examination of total hydrolysis of a Smith polyalcohol, this analytical method has provided a new and most powerful tool in the structural elucidation of polysaccharides by Smith degradation. As mentioned before, Smith degradations involve u t i l i -zation of differences in the ease of hydrolysis of various acetal structures generated during the periodate oxidation and subsequent reduction. Unfortun-ately the Smith degradation is not as simple as would f i r s t appear. The problem arises from the different rates at which different glycosides and acetals hydrolyse. Two extreme examples may be used to illustrate this point. Pyranosidic linkages are considerably more stable than corresponding furanosidic linkages. It has been possible to utilize this property to isolate L-arabinose from mesquite gum [37]. L-arabinose in polysaccharides often occurs as furanoside and hence may be readily hydrolysed by O.OIN sulfuric acid at 95°C. Hydrolysis of mesquite gum for 36 hours under these conditions produced a re-sidual polysaccharide containing no L-arabinose. At the other extreme, 4-0-methyl-D-glucuronic acid when attached glycosidically to other carbohydrate residues is extremely difficult to hydrolyse. Under conditions of hydrolysis sufficiently drastic to hydrolyse a l l other pyranosidic linkages, two thirds 17 of the 2-0-[4-0-methyl-D-glucopyranosyluronic acid]- o< -g-xylose of a soft-wood xylsn wi l l remain unhydrolysed [38]. This resistance to hydrolysis can probably be extended to the acetal resulting from cleavage between carbons two and three of the uronic acid. The ability to choose correct hydrolysis conditions, acid strength and duration, without an adequate analytical technique for a Smith polyalcohol^ is severely hampered. One must contend with glycosidic linkages which are easily hydrolysed as well as stable acetals. No set conditions appear to be employed in preparing Smith degraded residual polysaccharides. An oat /3 -D-glucan, oxidized and reduced, was hydrolysed with O.5N hydrochloric acid for 8 hours at room temperature [6 ]. Since there are no uronic acids in the glucan the acid treatment might seem rather severe. Brome grass hemicellulose on the other hand is known to contain uronic acid. Its Smith polyalcohol was hydrolysed with 0.1N hydrochloric acid at room temperature for 6 hours [6 ]. Although the isolation of small glycosylalditols was the object of the latter experiment, these conditions might not be expected to completely hydrolyse a l l acetals generated by the oxidation and reduction. A further example where possibly too much acid was used was. in the hydrolysis reported for lemon gum [1 ]. - Experimental data shox-js that optimum acid concentration is 0.5N rather than 1.0N. As well," optimum duration of acetal hydrolysis with minimum loss of arabinose is 16 hours rather than the reported kQ hours. A normal acetal resulting from a periodate oxidized and borohydride re-duced monosaccharide glycoside can be hydrolysed completely by 0.1N sulfuric Smith polyalcohol wi l l imply a periodate oxidized and sodium borohydride reduced polysaccharide. 18 acid at room temperature in eight hours. These conditions indicate the minimum acid strength required for analysis of Smith polyalcohols. Glyco-sidic linkages v i l l be preserved. This property v i l l allow isolation of intact glycosylaiditoIs released during hydrolysis vhile retaining as much as possible of the unoxidized polysaccharide. The degree of polymerization [D.P.] v i l l remain high preserving important structural features for subsequent ex-amination. Linkages other than the above may not be hydrolysed under the specified conditions. Methyl 4-0-methyl- c< -D-glucuronate methyl ester vas oxidized vith periodate and reduced with borohydride. The resulting polyol, although easier to hydrolyse than the related neutral parent glycoside, required hydro-lysis with 0.5K sulfuric acid for 16 hours at room temperature. This is more than sufficient acid strength to hydrolyse a considerable proportion of L-arabinofuranosyl linkages. Some pyranosidic linkages w i l l be broken as well. A problem exists therefore in the structural analysis of some polymers. Mild acid hydrolysis w i l l not cleave a l l oxidized and reduced uronic acids. Subunits which would be potential sites of periodate oxidation or methylation w i l l s t i l l be blocked. Stronger acid on the other hand w i l l cleave glycosidic bonds which have formed a constituent part of the overall structure of the polysaccharides. There are two potential solutions available for this problem. The f i r s t involves use of optimum hydrolysis conditions. These conditions involve control of hydrolysis parameters to achieve maximum acetal cleavage with minimal glycosidic hydrolysis. The second solution requires the re-duction of uronic acids in the polymer. This reduction eliminates those acetal structures which were difficult to hydrolyse. Thus as much as possible of the. intact original structure v i l l be retained by mild hydro ly t i c conditions. 19 The analysis of polyols and sugars outlined in the f i r s t portion of this thesis provides the f i r s t practical method for carrying out detailed analysis of any Smith polymer for optimum hydrolysis conditions. A l l parameters may be varied; time, temperature, acid strength. Since only small amounts of Smith polysaccharide are required for hydrolysis, in normal practice only 5 milligrams or less, the limit is set by the ability to handle precipitation of the residual polysaccharide. Precipitating the residual polysaccharide as hydrolysis proceeds, produces a fraction containing only high molecular weight products. This fraction may then be dialysed and freeze dried i f desired. The resultant residual polysaccharide can be subjected to total hydrolysis, thus avoiding complications which result from glycol aldehyde acetal formation. This technique has been successfully applied to lemon gum. Much milder con-ditions than those previously used are necessary for maximum yield of residual polysaccharide with a minimum amount of unnecessary degradation. THE OPTIMUM HYDROLYSIS OF OXIDIZED LEMON GUM Determination of optimum hydrolysis conditions for periodate oxidized and borohydride reduced lemon gum involves examination of parameters affecting the degree of hydrolysis. Lemon gum polyalcohol was prepared as outlined in the experimental section. Small samples of the polyalcohol were dissolved in water. The normality was adjusted to a series from 0.05N to 1.0N by the . addition of concentrated acid. Each sample was allowed to hydrolyse for twenty-four hours at room temperature. The samples were neutralized and 20 residual polysaccharide was isolated. Analyses were carried out to deter-mine remaining unhydrolysed acetal linkages. The sample hydrolysed with 0-5N sulfuric acid showed almost total hydrolysis of linkages which could be residual acetal links. 1 Figure 7 shows the reduction in glycerol as acid strength was increased from 0.05N to O.5K. Removal of a l l glycerol can be seen to require a 0-5N acid strength. A lower acid strength would have required a prolonged hydrolysis. The acid strength having been chosen i t was necessary to determine optimum hydrolysis time. Optimum hydrolysis time wil l minimize loss of L-arabinose with respect to elimination of glycerol. Lemon gum polyalcohol was dissolved in 0.5N sulfuric acid and samples were taken at intervals to follow removal of com-ponents x-)ith respect to time. To a certain extent this is a difficult thing to accurately assess. Fragments may be lost during hydrolysis which contain unoxidized carbohydrates. These fragments result from two forms of hydrolysis. Hydrolysis occurring at acetal linkages may liberate Smith fragments, for ex-ample, arabinosyl glycerol or galactosyl glycerol. In addition there w i l l be hydrolysis which xtfill break normal glycosidic bonds such as arabinofuranose linked units. The assay of how well acetals have been cleaved wi l l be in-dicated by residual glycerol since only glycosidic glycerol should be stable. New gas-liquid chromatography columns were prepared in an identical manner •to those used i n i t i a l l y . Peaks eluted prior to glycerol were obscured by a pyridine tailing peak. The peak resulted from some form of adsorption or reaction between copper and pyridine. It is at this time s t i l l not - clear what difference exists between the two tubing samples. Results with the second tubing were only slightly affected since the major peak produced by the polyol on hydrolysis and attributable to an acetal was the glycerol peak. It was possible to eliminate the tailing peak by switching to stain-less steel columns. Since only ethylene glycol was affected extensive use was s t i l l made of copper tubing columns. It has been subsequently dis-covered that copper tubing may be made passive by silvering the inner side with a mirror silvering solution. FIGURE 7: GAS CHROMATOGRAMS OF TRIMETHYLSILYL DERIVATIVES FROM LEMON GUM SMITH POLYALCOHOL HYDROLYSED WITH VARIOUS SULFURIC.ACID '. STRENGTHS FOR 2k HOURS AT 20°C . . . 22 Additional end groups w i l l be formed i f large fragments are released by the action of acetal and glycoside hydrolysis. Under these conditions an increase in stable glycerol w i l l occur. There exists however a large difference in rate of hydrolysis between acetal and glycosidicly bound glycerol. By plotting the degree of hydrolysis against time, i t is possible to follow the decrease in components as hydrolysis proceeds. Examination of the plot w i l l indicate the optimum hydrolysis time for the polymer. Figure 8 shows gas chromato-grams indicating progressive loss of glycerol. Figures 9 and 10 show loss of components plotted with respect to a response for galactose. There was, however, loss of galactose. This fact was clearly seen on hydrolysis of the soluble fraction from the precipitation solvent. The graphs indicate the time for optimum yield of residual polysaccharide with a minimum of contaminating acetal structures. During the course of these hydrolysis studies i t was confirmed that re-sidual glycerol present in the precipitated polysaccharide was chemically bound and not mechanically entrained or hydrogen bonded. This confirmation was de-termined in the following manner. The polysaccharide obtained after 32 hours of hydrolysis was dissolved in water and dialysed against running water for 2k hours. The polymer isolated by freeze drying was shown to contain the same amount of residual glycerol. Clearly the glycerol was covalently bonded or i t would have passed freely from the dialysis tubing. Analysis of the graphs which plot decrease in components, indicated that removal of a l l but a residual amount of glycerol required 8 - 16 hours. Hy-drolysis for that period formed a polysaccharide containing minimal glycerol. These conditions are considerably milder than those reported by Jones and 23 GLYCEROL(l) 0 HR. (Log often.) ETHYLENE GLYCOL GALACTOSE (6) - i — v | 1 i 1 1 r 2 HR. FIGURE 8: GAS CHROMATOGRAMS OF TRIMETILYLSILYL DERIVATIVES FROM LEMON GUM SMITH POLYALCOHOL HYDROLYSED FOR DIFFERENT PERIODS OF TIME WITH 0.5N SULFURIC ACTJD AT 20°C 20r 0 12 4 8 16 TIME (HR) FIGURE 9: RELATIVE G.L.C. PEAK AREAS FOR COMPONENTS 2,3,4 FROM LEMON GUM SMITH POLYALCOHOL FIGURE 10: RELATIVE G.L.C. PEAK AREAS FOR COMPONENTS 1 [GLYCEROL] AND 5 [ARABINOSE] FROM LEMON GUM SMITH POLYALCOHOL 25 Stoddart [ l ] . Their hydrolysis of borohydride reduced - periodate oxidized lemon gum involved treatment at room temperature with one normal sulfuric acid for forty-eight hours* FIGURE 11 GAS CHROMATOGRAM OF LEMON GUM SMITH POLYALCOHOL HYDROLYSED FOR 32 HOURS WITH 0.5N SULFURIC ACID AT 20°C AFTER DIALYSIS 601 50-40 30 20-10 -0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 MIN Evaluation of data presented indicates that even at the milder con-ditions used in our work degradation of the polysaccharide has occurred. Only trace quantities of galactose were present in the supernatant confirming that the galactan portion of the polysaccharide had been subjected to only mild degradation. I f this were not the case one would have expected signif-icantly larger losses of free galactose from the polymer. The galactose which was lost has been in the form of low molecular weight Smith fragments such as galactosyl glycerol* Arabinose was found to a considerable extent in the supernatant. Since arabinose is furanosidicly linked, one would conclude that i t had been produced from the polymer by two mechanisms* Some free 26 arabinose resulted from direct hydrolysis of araban portions of lemon gum. The second portion of arabinose resulted from hydrolysis of Smith fragments since subunits such as arabinofuranosyl glycerol are susceptible to mild acid hydrolysis. TABLE X ANALYSIS OF LEMON GUM SMITH POLYALCOHOL AND ITS HYDROLYSIS AND OXIDATION PRODUCTS G l y c e r o l Mole $ A r a b i n o s e G a l a c t o s e Lemon gum S m i t h p o l y a l c o h o l 27.1 14.0 8.8 Lemon gum S I [16 h r . - 0.5N H 2 S O a - 20°C] 2.8 14.6 82.7 Lemon gum S m i t h p o l y a l c o h o l s u p e r n a t a n t a f t e r J2 h r s . [TOTAL HYDROLYSIS] , 21.1 21.8 57-1 Lemon gum S m i t h p o l y a l c o h o l s u p e r n a t a n t a f t e r 32.hrs. [NO FURTHER HYDROLYSIS] 75-3 20.8 4.0 Lemon gum S I p o l y a l c o h o l 32.8 11.1 49.2 Lemon gum S I I [16 h r s . - 0.5N H 2S0 4. - 20°C] 3.2 13-7 83.I A sample of lemon gum was oxidized, reduced and hydrolysed under optimum conditions to produce lemon gum SI. The composition of the polymer was as in-dicated above in Table X. . Lemon gum SI was oxidized, reduced and hydrolysed in a similar manner to that used in its own preparation. This product was designated lemon gum SII. Analysis of the polyalcohol SII indicated a lower percentage of arabinose than had been observed by Jones and Stoddart [ l ] . The 27 retention of some arabinose is to be expected from reported substitution patterns. The large difference observed may be related to increased acid strength during their hydrolysis. More side chains v i l l have been removed from the galactan framework, exposing additional sites for oxidation by periodate. With more acetals being formed during the oxidation and reduction the polymer's galactan framework v i l l undergo greater depolymerization. Additional losses vere encountered during both the i n i t i a l and second oxidation-reduction operations. Each of the two operations resulted in ap-proximately forty percent weight losses* These decreases cannot be attributed to oxidation weight loss due to formic acid or other minor constituents. They must reflect a degree of autohydrolysis and base catalysed depolymerization occurring during the periodate oxidation and borohydride reduction. Polyalde-hydes are known to be susceptible to basic hydrolysis and depolymerization [39]. The polyalcohol can be degraded by two acid sources. Acetic acid was added to destroy the excess borohydride while oxidized uronic acids vere present in the polymer during reduction and dialysis. 28 SMITH FRAGMENTS The optimum yield of residual polysaccharide after Smith oxidation w i l l not necessarily produce an optimum yield of Smith fragments* Degradation of substituent units during hydrolysis is reflected in the appearance of free sugars. For example, any Smith fragment containing arabinofuranosyl linkages w i l l undergo hydrolysis under mildly acidic conditions. Examination of the alcohol-water soluble fraction from mild acid hydrolysis to lemon gum SI gave results presented in Table X. The alcohol-water soluble fraction obtained on hydrolysis to lemon gum SI contained large quantities of glycerol. This fragment arises from cleavage of arabinofuranose, arabino-pyranose [blocked at position four] and galactopyranose [blocked at position two and/or six]. It is known from our own and previous work that galactose in lemon gum is not substituted at C-2 hydroxyl. The amount of glycerol present in intact lemon gum SI alcohol-water solubles may be divided into two distinct classes. The largest portion of glycerol results from multiple oxidations. For example, when side chains terminated in k-Q[4-0-methyl- d-D-glucopyranosyluronic acid]-L-arabinose are oxidized, both sugars w i l l be cleaved and free glycerol [from arabinose] w i l l be released on mild hydrolysis. The second source of free glycerol on mild hydrolysis arises from glycosidic hydrolysis of labile furanosidic linkages. Two complications affect the yield of glycerol. Stable uronic acid de-rived subunits may effect liberation of glycerol while s t i l l forming part of the soluble fraction. Secondly, glycolaldehyde formed by periodate cleavage can readily complex with free hydroxyIs to form an acetal blocking group. 29 The complexity of this problem was mentioned earlier in discussion of the need for total hydrolysis during analysis of periodate oxidized polymers* Substitution such as this results in chromatographic complexity. These acetals, as new components, have unique retention times and molar responses. A possible but untried solution to this problem may be to hydrolyse in the presence of a large amount of vicinal diol [i.e. butane -2,3-diol]. It is possible that added diol w i l l act as a scavenger for available glycol aldehyde. Glycerol also exists in the alcohol-water soluble fraction from lemon gum SI in the form of glycosidicly bound glycerol. It is this material which may be regarded as structurally significant. Two conclusions may be drawn from studies of this material. A portion of the soluble fraction was dissolved in acetone-methanol [1:1]. This soluble material contained low molecular weight monosaccharide fragments in a free and glycol aldehyde combined form as well as small amounts of arabinose, galactose and a single glycerol con-taining Smith fragment. The Smith fragment was a galactopyranosyl glycerol. No other Smith fragments could be positively identified among the other trace components. The remainder of the original SI soluble fraction [insoluble in acetone-methanol] was shown by GLC to contain no low molecular weight Smith fragments. It contained small amounts of low molecular weight arabinose and galactose oligosaccharides but the bulk of the sample was higher molecular weight oligosaccharides terminated by a glycerol moeity* This was clearly seen when samples of acetone-methanol insolubles were hydrolysed and checked for reducing power. 20 The presence of high molecular weight galactan oligosaccharides termin-ated by glycerol provides further proof that the internal galactan frame-work of lemon gum contains some galactose units substituted at positions six and one only. FIGURE 12: GAS CHROMATOGRAM OF THE SUPERNATANT FROM THE ISOLATION OF LEMON GUM SI BEFORE AND AFTER TOTAL HYDROLYSIS [at 52 min. the temperature was raised at 3°/min. to a final temperature of 250°C] COMPONENT (2) THREITOL (4) SOLUBLE TOTAL HYDROLYSIS AltABINOSE GALACTOSE A J 24 12 36 44 48 52 56 GLYC (2) THRE ARAB SOLUBLE FROM MILD HYDROLYSIS GAL-GLYC GAL — i i I I — 20 24 28 32 i 36 40 52 • i 60 • / / • 64 90 12 16 56 REDUCTION OF URONIC ACIDS As indicated earlier uronic acids were a potential source of error in evaluation of polysaccharide structure by mild acid hydrolysis of the Smith polyalcohol. There exists a solution to the problem of resistance to hydrolysis of uronic acids or acid containing subunits. Reduction of the uronic acids prior to oxidation w i l l result in an acetal which is easily hydrolysed. 31 Two basic approaches exist for the reduction of an acidic polysaccharide. They involve reduction of either the free uronic acid or an ester of the acid. The most common reduction of a polysaccharide is performed on the fully methyl-ated product. This is readily accomplished with lithium aluminum hydride [LAH] in tetrahydrofuran [THF] or similar dry ether-type solvents. It does not how-ever provide a route to free unsubstituted polysaccharides. One use that has been made of an ether blocking group is contained in a paper by Rees and Samuel [40]. Alginic acid was esterified with diazomethane, trimethylsilylated with hexamethyldisilazane and trimethylchlorosilane [15] and then treated with LAH in cold THF. The product separated slowly as re-duction proceeded. Unfortunately LAH is a poor reagent to use with sugar derivatives containing or forming unblocked hydroxyl groups. LAH forms strong aluminate complexes with these free bydroxyls. It is often necessary to esterify the product to free the substrate from residual reducing agent. Some workers have used diborane as their reducing agent. It was noted by Smith and Stephen [hi] that metal hydrides failed to reduce acylated acidic polysaccharides. Ester groups were attacked f i r s t and as a result acidic polysaccharide was precipitated from the reaction mixture. Brown [42] however has shown that diborane reduces carboxyl groups in preference to ester groups. Smith and Stephen [41] f i r s t showed that diborane could be an effective re-ducing agent for acylated acidic polysaccharides. They also noted that the propionate derivative of alginate was more soluble in diglyme and hence re-duced more easily. Other workers have used this procedure to reduce acidic polysaccharides. Hirst et a l [43] reduced alginic acid by Smith's method of in situ generation of diborane. This reagent reduced 90f> of the uronic acids and produced $.2$ 32 residual n-propoxy group. The degree of reduction was slightly higher than that which Smith and Stephen found. Ross and Thompson [44] treated an acetyl-ated 4-0-methyl-glucuronoxylan with externally generated diborane, reducing 90$ of the carboxyl groups. Manning and Green [45] compared the reduction of alginic acid by dif-ferent methods. One of these methods was the use of diborane generated externally on di-O-propionyl alginic acid. A large excess of diborane [6.5 moles per mole of carboxyl group] was used. The same polymer was also re-duced by in situ generation of diborane from the addition of boron trifluoride to sodium borohydride. Thus any difference between these two reduced polymers could be attributed to the effect of sodium borohydride and boron trifluoride in the polymer solution. This use of externally produced diborane showed for the f i r s t time that an ester could be reduced to an ether by diborane alone. They explained this reduction by a modification of the mechanism suggested by Pettit and Kasturi [46 ] to account for the reduction of esters to ethers with diborane and boron trifluoride. The proposed mechanism ex-plained why more n-propionyl esters were reduced to n-propyl ethers with diborane generated in situ than externally. As the sodium borohydride con-centration was reduced, boron trifluoride was felt to catalyse the following reaction: FIGURE 13: THE REDUCTION OF ESTERS TO ETHERS WITH AN EXCESS OF DIBORANE [adapted from Pettit and Kasturi [46]] 9 D H 3 R-C-OR 5==i B H 3 B H 3 b R-C-OR R-C-OR • • B H 2 0 H 2 O n , . H R - C - O R 1 •BH 2 RCHgOR * R-CH-OR S^*- R-CH-OR B H , \ !' B H 2 ck,: __ B H 3 r 4 h . c TABLE XII ANALYSIS 0? THE POLYSACCHARIDES OBTAINED FROM ALGINIC ACID BY DIFFEREHT REDUCTION PROCEDURES Functional Group Reduction Procedure Polymer Anhydro-uronic acid [<f>f b n-Propionyl ester [$] Total propoxy-c group [$>] Alkali-labile n-propoxy-group [#] Esterified uronic^ acids Degree of polymsrisatii Methyl di-O-propionyl alginate with LiBH^ Bl 6.3 0.00 o.ooe 0.00 Negative 103 Dl with LiBH 4 B2 10.4 0.00 3.9 0.25 Negative . 83 Di-O-propionyl alginic acid with B2HS Dl 11.0 7.6 3-5 2.3 Negative 111 Di-O-propionyl alginic acid with NaBHj + BF3 D2 I6.3 3.2 7.5 2.7 Negative 66 a Determined by CO2 evolution. Determined by the hydroxylamine-ferric perchlorate colour test. e cl. Determined by a Zeisel method. Determined by osn»metry of the triacetate in 1,1,2-trichloroethane. Contained Q.kQfja methoxy-group from diazomethane methylation of di-O-propionyl alginic acid. 34 Less uronic acid is reduced when diborane is generated in situ. Since boron trifluoride catalyses the reduction of n-propionyl esters to ethers, i t may also catalyse reductive cleavage of n-propionyl esters* An increased rate of reductive cleavage would have caused the polymer to precipitate sooner thereby lowering the reactivity of the carboxyl group. Sodium ions present in the in situ diborane generation may form sodium carboxylate which is not re-duced by diborane. Manning and Green [45] suggested that the differences in D.P. were a result of the strong Lewis acid, boron trifluoride. The degree of depolymer-ization was dependent on how fast and in what quantity boron trifluoride was added. The use of diborane for the reduction of acidic polysaccharides prior to periodate oxidation has several serious drawbacks. Even when diborane was generated externally the D.P. of di-O-propionyl alginic acid was reduced from 158 to HI, a drop of 30 percent. Although data are not available, a corres-ponding drop in D.P. might be expected to occur on reduction of other acidic polysaccharides, neutral linkages being more susceptable to acidic hydrolysis. The introduction of ether blocking groups is the most serious drawback to utilization of this method of reduction. In the best diborane case there was 3*5$ pro poxy ether substitution. This represents in its lowest case an iQjo conversion of original n-propionyl ester to n-propyl ether. Alginic acid is a linear 1— • 4 linked polyuronide. Its di-O-propionyl derivative w i l l have two out of ten esters reduced to ethers. For every five uronic acid groups oxidizable by periodate in the original polymer only slightly greater than three of the five w i l l be oxidizable after reduction with diborane. In addition esterification must be employed to produce a soluble polymer and saponification is necessary for removal of the ester groups. Some fraction-ation may occur on esterification, altering slightly the composition of the polymer. Saponification may produce a small amount of alkaline degradation. It is important to stress again that diborane suffers severely as a potential reagent for preparing neutral polysaccharide for periodate oxidation. The most serious drawback is the introduction of ether linked n-propyl groups. The second method employed commonly for conversion of acidic into neutral polysaccharides involves reduction of uronic acid esters \*ith a metal hydride. Two main esters are used, methyl and hydroxyethyl. Methyl esters are generally formed in one of two ways. A polysaccharide can be esterified by dispersion in methanol containing concentrated sulfuric acid or hydrogen chloride. Bishop and Zitko [47] in an analysis of sun-flower pectic acid esterified the galacturonan in methanol containing suf-ficient sulfuric acid to make the concentration 2N. The mixture was kept at 0-2°C for nineteen days before isolation of the esterified product. This pro-cedure resulted in a significant reduction in D.P. of the polysaccharide. A procedure such as this would be extremely detrimental to a polymer containing labile furanosidic linkages. The second method for forming the methyl ester of an acidic polysaccharide involves the use of diazomethane in an ethereal suspension of the polymer. Direct esterification with diazomethane is not a practical route in preparation of a neutral polysaccharide. Methylation w i l l occur at a number of alcoholic functions in the polymer, interfering with sub-sequent periodate oxidation of the reduced polysaccharide. Esterification with diazomethane is best carried out on the acetylated or propionylated poly-saccharide. The esterified polymers have very limited numbers of free hydroxyl groups capable of methylation. 56 Hydroxyethylation is an alternate method of esterifying uronic acids. This reaction involves esterification of the uronic acid vith ethylene oxide. -which reacts directly vith the polysaccharide acid in aqueous solution. There is reported to be no evidence for hydroxyalkylation of alcohol groups by ethylene oxide. An investigation on a model compound would however be a useful undertaking since isolation of polyethylene glycol esters has been reported in the literature [48], METAL HYDRIDE REDUCING AGENTS Sodium borohydride has been reported to give good reductions of ethylene oxide esters [49 ]• Attempts to utilize this reagent have generally been unsuccessful in this laboratory. Sodium borohydride saponifies the esters and the acids and salts produced are not readily reducible. Potassium borohydride has also been used to reduce esterifled uronic acids [50]. Bishop and Zitko [47] in their paper on the reduction of galacturonan form sunflower pectic acid analysed in detail the effect of solvent on potassium borohydride efficiency. They examined water, Q0% aqueous methanol, and &0$ aqueous dimethyl sulfoxide. Their results indicated that reductions in 80# aqueous methanol were most efficient. De-esterification was suppressed, a reasonably high proportion of galacturonic acid was reduced and recoveries were good. In water, a high proportion of galacturonic acid was reduced but there was considerable de-esterification and recoveries were low. The low recovery was felt to result from degradation of esterifled gal-acturonan in alkaline media [51]. In BOJ» aqueous dimethyl sulfoxide [DMSO], de-esterification was greatly suppressed. The greatest reduction was obtained 37 when a partially reduced product was reacted. These advantages however were offset by low recovery. Indeed i t i?as reported that when this solvent was used for the f i r s t reduction of a fully esterified galacturonan, complete de-gradation occurred. No product precipitable by methanol could be obtained. Reductions were found to be only slightly more efficient with methyl than hy-droxyethyl esters. Bishop and Zitko [47] examined the efficiency of esterification. As galacturonan was reduced heterogeneous esterification with diazomethane be-came less efficient. At a galacturonic acid content of 11$, no more than 45$ of the carboxyl groups could be esterified. The substrate was even freeze dried to provide a large surface area. Esterification with ethylene or pro-pylene oxide gave good results but the products when precipitated by organic solvents, tended to form gels which were difficult to handle. It was possible however to esterify residual carboxyl groups in highly reduced galacturonans with diazomethane in dimethyl sulfoxide. In this procedure the esterified product did not need to be isolated before reduction since the same solvent could be used. However, the main drawback to using the method proposed by Bishop and Zitko was the accumulated decrease in degree of polymerization [from 270 to 21]. Lithium borohydride has also been used to reduce acidic polysaccharides. This reagent was f i r s t used by Rees and Samuel [40 3 for the reduction of a l -ginic acid. A heterogeneous reaction in boiling THF was carried out on both 2-acetoxy-ethyl di-O-acetylalginate and methyL-di-0-acetylalginate. Special precautions were taken to ensure that the polysaccharides were completely ac-cessible. The authors report that reduction yielded a polysaccharide contain-ing 4.7$ anhydro uronic acid. 38 Manning and Green [45] in their examination of the reduction of alginic acid, reduced methyl di-O-propionyl alginate by the method of Rees and Samuel. Methyl di-O-propionyl alginate [89$ of its uronic acids esterified] was however soluble in tetrahydrofuran. A l l n-propionyl and methyl esters were reductively cleaved in the reduced polysaccharide. A l l reducible terminal sugar units were converted to glycitols. In addition the authors reported a kyf> reduction of uronic acid carboxyl group to primary alcohols* Rees and Samuel suggested that reduction of functional groups on poly-saccharides under heterogeneous conditions is incomplete because molecules of the solid phase are inaccesible. Manning and Green [45] however showed that 94$ of the hemiacetal end group were reduced under heterogeneous conditions. It was felt that lowered reactivity was not due to lack of accessibility. Iheir alternate explanation is based on activation energies. Addition of solvation energy for heterogeneous reaction to the activation energy s t i l l makes the total energy for reaction of aldehydes lower than that for carboxylic acids. Hence aldehydes react more rapidly than carboxylic acids under hetero-geneous conditions. It should be noted that there may be vast differences in accessibility. The reducing end group of a linear molecule such as alginate should, i f well dispersed, be accessible to reduction. Uronic acids on the other hand can be strongly associated with hydroxyl functions of other polymer chains. Areas of polysaccharide may thus be rendered inaccessible. Aspinall and McKenna [52] esterified Araucaria bidwilll gum with ethylene oxide. Acetylation and subsequent reduction by the method of Rees and Samuel gave a product polysaccharide with less than 1$ uronic acid. 39 EVALUATION OF NEW ROUTE TO REDUCED POLYSACCHARIDES The method of Rees and Samuel [kO ] for reducing acidic polysaccharides is obviously effective. It does however involve several reactions; poly-saccharide esterification, uronic acid esterification. reduction and sapon-ification. Each reaction may have to be repeated and possible fractionation or degradation of the polysaccharide can occur. Removal of any of these steps might increase the chance of obtaining a neutral polysaccharide which retains potential periodate oxidation sites and is closely related to the original polymer. To prevent unnecessary degradation, esterification should not be carried out in methanol with an acid catalyst because of D.P. reduction. Diazomethane is unsuitable on the unsubstituted polymer due to ether formation. Ethylene oxide has shown none of these detrimental effects. It has certain advantages; water soluble, volatile, i f hydrolysed to ethylene glycol, i t may be readily removed by dialysis, and i t also inhibits any bacterial degradation. Its two disadvantages are long reaction time for esterification as mentioned previously and possible ether formation. The latter point requires further evaluation, however no etherification has thusfar been reported. Certain reducing agents are unacceptable in the reduction of acidic polysaccharides. Diborane to be effective requires a homogeneous reaction. But, soluble ester derivatives are reduced in part to ethers which block sub-sequent periodate oxidation. S i l y l blocking groups which also make a soluble derivative, bydrolyse as the reaction proceeds producing a heterogeneous re-action. Lithium aluminum hydride complexes free hydroxyls, the result being a heterogeneous reaction. The complex is difficult to break down, introducing ko problems in isolating the reduced polysaccharide. Thus the choice of re-ducing agents is limited to the borohydrides. Their ability to reduce esters and acids is increased with increasing covalent character. Lithium boro-hydride best fits this requirement [53]. In the work conducted in this laboratory, choice of solvent was a departure from those previously employed. It was shown by Bishop and Zitko [47] that 8C$ aqueous DMSO greatly suppressed de-esterification and a high proportion of reduction occurred. Since only the hydrolysis of reduction complex involves water in a borohydride reduction i t was decided to use pure DMSO. Addition of water after complex formation w i l l complete the reaction. A large number of polysaccharides are soluble in this reagent and tests showed only a slow reaction with lithium borohydride. The only disadvantage in the use of DMSO was reported decomposition of polysaccharide [47]- An explanation as to why depolymerization occurs was not indicated, however there are two possibil-ities. Alkaline cleavage by /3 -alkoxycarbonyl elimination could lead to depolymerization. DMSO is effective in preserving methyl ester necessary for this elimination. However, i t should not greatly increase the ability to co-ordinate transition states or leaving groups in comparison to water or methanol. When methyl 4-0-methyl c< -D-glucuronic acid methyl ester was reduced with lithium borohydride in DMSO there was gas chromatographic evidence for the presence of small amounts of by-product. The by-product was tentatively assigned the structure of the 4,5-dehydro-D-glucuronate [54]. Confirmation of this structure w i l l require further analysis. The proposed mechanisms for the elimination in uronic acids are outlined in Figure 14. in FIGURE Ik THE /3 -ELIMINATION MECHANISMS FOR URONIC ACIDS [It has been recently suggested that the eliminations proceed via an ElcB mechanism [ 5 5 3 3 E2: ElcB: EtO * H X-EtO H b — OPh X-OPh ;= X- OPh The second possible explanation is that DMSO altered the precipitation characteristics of pectin. Work in our laboratory on isolation of methylated polysaccharides from DMSO solutions indicated changes in ease of precipitation. Methylated polysaccharides in DMSO occasionally failed to precipitate on addition to water. Methylation mixtures poured directly into dialysis tubing with no addition of water to the DMSO resulted in cases of almost complete loss of partially methylated polysaccharides. Dilution of methylation mixture with water, followed by dialysis to remove DMSO resulted in easily isolated methy-lated polysaccharide in excellent yield. Lemon gum was utilized as a model substance to evaluate reduction of esterified polysaccharides dissolved in DMSO. Lithium borohydride was the reducing agent employed. Polysaccharide reduction was monitored by G.L.C. 42 Inclusion of erythritol during polysaccharide hydrolysis gave quantitative results on retrieval of constituent monomers. The results indicated additional difficulties in obtaining a neutral polysaccharide by reduction. Esterification with ethylene oxide requires reaction for a period of ten days [4-7 3 • However, loss of arabinose occurred during the esterification of lemon gum. TABLE XIII CARBOHYDRATE ANALYSIS OF LEMON GUM AND ITS REDUCED PRODUCT Arabinose Rhamnose * 4-0-Methyl Glucose Galactose Glucose Neutral Sugar Recovery Lemon Gum 32.7 3.4 — 63.9 — 66.6 Lemon Gum OCH2CH2OH 29.9 3-3 — 66.8 — — Lemon Gum OCH2CH2OH 1 x LiBH 4 25.0 1.5 15-9 51.3 5.3 84.0 Lemon Gum 2 x OCH^ CHaOH 2 x LiBH 4 26.25 1.7 18.8 48.6 4.8 89.5 Calculations based on aldobiouronic acid survival during hydrolysis in-dicated quantitative levels of reduction. These calculations were based on survival of aldobiouronic acids during hydrolysis as well as degradation of individual sugars. The results accounted for a l l sugars present prior to hy-drolysis and correlated well with recovery of neutral sugar in hydrolysate. The loss of some arabinose during esterification was not unexpected. The polysaccharide was deionized hence a l l uronic acids were free to cause hy-drolysis. Although the large amount of ethylene oxide present provided a substrate for reaction of ionized acids, there was s t i l l water present and cleavage of labile arabinose linkages could occur. As the reaction proceeded the pH decreased to 7 and hydrolytic action was reduced. Hydrolysis could be suppressed by substitution of DMSO for water during esterification. It was necessary however to isolate the polymer before reduction as residual ethylene oxide reacted with lithium borohydride. Lithium borohydride was added as a DMSO solution to lemon gum 2-hydroxy -ethyl ester in DMSO. Hydrogen was released during the i n i t i a l vigorous re-action through a mercury gas trap to exclude moisture. By the 15th hour of reduction a thick gel had formed. This gel could be partially broken down by stirring or gentle shaking. A slow evolution of hydrogen continued for several days. The solution was then treated with small portions of dilute acetic acid to destroy excess lithium borohydride. The solution was diluted with water and dialysed. Saponification occurred during isolation making re-es terif ication necessary for continued reduction. The i n i t i a l ethylene oxide esterification did not fully block a l l uronic acids. The lack of complete reduction with an excess of lithium borohydride suggested either a problem of accessibility or hydrolysis of esters during re-duction. Similar results in subsequent reductions could be attributed to effects of either of these factors. Some losses occurred as a result of work-up and dialysis. However yields were generally good. The mild nature of the reaction was evident from the yields of L-arabinose in relation to D-ga lactose on subsequent reduction. Yields of arabinose in-creased in relation to galactose as the polymer was further reduced. In addition, losses which one might expect in arabinose should have been accompanied by losses in 4-0-methyl-D-glucuronic acid, since this acid is the major acid lost on mild hydrolysis [1 ], Acid hydrolysis of lemon gum resulted in re-duction of 4-0-methyl compared to unsubstituted glucuronic acid. On reducing lemon gum ester vith lithium borohydride in DMSO, the percentage of 4-0-methyl-D-glucose increased as reduction proceeded. Had any oligosaccharides been formed during reduction, they would have been lost during dialysis. This would have resulted in lower 4-0-methyl-D-glucose levels as was the case on mild acid hydrolysis. The reduction method as evaluated on lemon gum was successful in reducing 75$ of available uronic acids. It seems probable that this method would be more successful on smaller polymers. Central portions of a large molecule like lemon gum may be inaccessible to a reagent such as lithium borohydride. Borohydrides can complex readily with free hydroxyIs, blocking access to the interior of this highly branched polymer. Reduction with lithium borohydride in DMSO appears to be very useful for convenient detection of the presence of different uronic acids such as glu-curonic and 4-0-methyl glucuronic. If the structures of the major aldobiouronic acids are known, then accurate carbohydrate analyses are also conveniently ob-tained. The method involves a minimum number of reaction steps and hence would appear to be a very convenient procedure for i n i t i a l examination of polymer uronic acids. If higher degrees of reduction occur in lower molecular weight or linear polysaccharides, increased use of this reduction sequence could be envisioned. SIGNAL ENHANCEMENT USING TRIMETHYLSILYL [TMS] DERIVATIVES AND UTILIZATION OF TMS AS A PROBE FOR DETERMINATION OF NUMBER OF HYDROXYLS IN A COMPOUND During gas-liquid chromatographic examination of products from total hydrolysis of periodate oxidized carbohydrates, samples of per silylated polyol were examined by proton magnetic resonance [p.m.r.]. Spectra for car-bohydrate protons were very complex because of similar electronegativity and environment [see Figure 15]. It was noted however that s i l y l methyl proton resonances were very sharp and in most cases, unless s i l y l ethers were in identical positions, resolved from one another. This discovery indicated a possible technique for characterization of polyhydroxy compounds. Introduction of TMS ethers increases the number of hydrogens on each hydroxyl by a factor of nine. Consequently detection of the number of hydroxyIs in a compound is possible even though the remaining proton signals are lost in background noise. In addition, compounds which are of unknown structure may be given some partial assignment of structure. It may be possible to determine number of hydroxyIs, their positions and degree of substitution at the parent carbon atom. Any compound containing a s i l y l ether would be amenable to this analytical procedure. In examinations of carbohydrate molecules, knowing the number of hydroxyIs can be a useful addition to information about that compound. The analysis of compounds on G.L.C. can result in a confused picture because of wide variation in retention times for partially methylated sugars and polyols. Although i t is possible to bypass many difficulties with mass spectrometric analysis, i t may be dif f i c u l t to distinguish between components by their mass spectra alone. k6 For example, two methyl sugars may have similar G.L.C. retention times. The mixed mass spectra might be very difficult to interpret. The s i l y l p.m.r. spectra however may show two anomers and indicate something about their sub-stitution. FIGURE 15: 100 MHZ P.M.R. SPECTRUM FOR THE ERYRTHRITOL PROTONS IN TRIMETHYLSILYLATED ERYTHRITOL 4.0 ERYTHRITOL .' PROTONS (TMS ERYTHRITOL) - J.: 100MHz 1 X 3,0 p. p.m. I l l ' I —L. _ l Li_L_J I I I I Figure 16 is a l i s t of compounds which have been subjected to s i l y l ether analysis indicating the number of peaks and their degree of separation. Primary trlmethyl s i l y l ether protons resonate at higher field than secondary trimethyl s i l y l ether protons. Barring shielding effects, the re-sonance signal for a trimethyl s i l y l group w i l l reflect electronegativity of the hydroxyl to which i t is attached. The wide diversity of signal value can be seen in almost complete separation of signal for eight different s i l y l ethers present in an unresolved collection of two arabinose TMS ether peaks. This p.m.r. technique provides a non-destructive check on component purity as eluted from the gas chromatograph. hi FIGURE 16: -COMPOUNDS SUBJECTED TO P.M.R. [100 MHZ] ANALYSIS OF TRIMETHYLSILYL PROTONS TRIMETHYL- TRIMETHYL-SILYL • COMPOUND SILYL RM.R. SPECTRUM k& FIGURE 16: » COMPOUNDS SUBJECTED TO P.M.R. [100 MHZ] ANALYSIS OF TRIMETHYLSILYL PROTONS [Continued] TRIMETHYL-SILYL COMPOUND ARABINOSE Opeoks) TRIMETHYL-SILYL RM.R. SPECTRUM TMS GLUCOSE (peok I) 2 mg. a or/S-FURANO GLUCOSE (peak 2) a-D-PYRANO . GLUCOSE (peak 3) /9-D- PYRANO METHYL 2,3-DI-O-METHYL-o-D-GLUCOPYRANOSIDE METHYL fi -D-GLUCOPYRANOSIDE k9 Obviously there w i l l be situations where overlap occurs between t r i -methyl s i l y l ether proton resonances. Perfect overlap w i l l occur in cases such as erythritol where there are two identical primary and two identical secondary hydroxyl positions. This results in only two peaks of equal in-tensity. Normally some hint of overlap w i l l be visible. For example, two primary peaks of glycerol overlap but they have, as a result, twice the signal intensity of the secondary TMS ether. In many cases where two compounds pro-duce peaks that overlap, such overlapping w i l l not be absolute making i t obvious that there are two contributing components. In cases where some signals are well resolved, integration of component peak area w i l l indicate percentage composition. This technique for examination of s i l y l methyl proton resonances provides a useful tool in structural elucidation and detection of component mixtures eluted as a single gas chromatographic peak. SIGNAL ENHANCEMENT An additional benefit can arise from examination of p.m.r. spectra of TMS ethers. A nine-fold increase in number of protons occurs on silylation of each hydroxyl position. In effect, TMS ethers act as built-in signal accum-ulators. The resonance peak for TMS ether protons is very sharp. The signal undergoes only weak proton-proton interactions because of the oxygen bridge between carbon and silicon. Normally, proton resonances are interacting, forming multiplets of lower intensity. TMS resonances are unaffected by hydrogen bonding which can broaden hydroxyl proton spectra of alcohols making 50 them even more difficult to observe than normal proton resonances. Jackman [56] discusses various factors exerting influence on the appearance of ethanol's hydroxyl proton. It can be broadened, split into a multiplet or increased in signal size by the presence of water. TMS protons give a sharp signal on the low side of tetra methyl silane. The signal is nearly ten times as large as the hydroxyl proton signal would have been had i t been present. This signal enhancing effect allows for ex-amination of very small samples of TMS ethers. The number and structure of hydroxyIs may be determined. A small peak occasionally occurs in glucose TMS ether gas chromatograms. It represents only about one percent of total peak area. Although i t was thought to be a furanoside anomer i t could not be ruled out that i t was due to under silylation or an anhydro form of glucose. Gas chromatography of a large amount of TMS glucose resulted in collection of only two milligrams of this compound. Examination of the p.m.r. spectra of TMS protons indicated a compound containing five hydroxyIs. This ruled out under silylation or an-hydro sugars as possible causes of this component peak. The G.L.C. retention time of this component agrees with no other common sugar. These points strongly indicate the minor component to be o( or /6 TMS glucofuranoside. This infor-mation regarding a minor component was obtained on two milligrams of sample. P.m.r. spectra for the remaining protons of the glucose molecule were com-pletely obscured by spectrometer noise. It w i l l be possible using TMS proton spectra to obtain number and nature of hydroxyIs on very small samples isolated by gas chromatography. Purity of single peaks may also be evaluated. 51 TRIMETHYLSILYLATION OF GALACTURONIC ACID During studies of lemon gum a method for estimation of aldobiouronic/^ acids was investigated. Work has been done on estimation of fully methylated aldobiouronic acids and on silylated methyl esters of aldobiouronic acids [57]• It was hoped that a procedure could be developed for separation of fully s i l y -lated aldobiouronic acids. A reagent was selected which would react with alcohol and acid substituents to form corresponding TMS ethers and esters. The reagent chosen was bis-trimethylsilyl-acetamide [BTMSA] which was known to react with amino acids to form s i l y l esters. Galacturonic acid was chosen as a model compound containing the necessary alcoholic and acidic functions. Silylations were carried out with BTMSA in various solvents; pyridine, dimethyl formamide, dimethyl sulfoxide, hexane and dioxane. Subsequent gas chromatography [SF-96 liquid phase] indicated four components. Two components were well resolved, two overlapped to form a central pair of peaks. Since under silylation is known from our own and other work [58] to produce extra peaks, a check was carried out on the eluted peaks. Possible lactone form-ation had to be ruled out as well, since lactones are known to have shorter retention times than parent acids. Other liquid phases were evaluated for gas chromatography of the silylated galacturonic acids. None proved more efficient than standard SF-96 liquid phase used in earlier studies. Preparative gas chromatography was convenient with this liquid phase because of a high maximum operating temperature. Small samples of the four galacturonic acid peaks were isolated and analysed by p.m.r. spectrometry. This preliminary examination indicated that each component con-tained five trimethyl s i l y l groups. This fact eliminated lactones or under s i l y l -ation as causes for any of the four components isolated. For complete p^m.r. spectral analysis large scale sample separations •were carried out. Analysis of the p.m.r. spectrum for ring protons should confirm furano and pyranoside forms for the four peaks. The well separated f i r s t and last peak were isolated in a quantity large enough to confirm their structures by p.m.r. spectroscopy. Large scale separation of central peak components was extremely difficult. In addition degradation was observed due to prolonged sample collection. An experiment was carried out on an easily isolated peak to determine the exact cause of degradation. A sample of peak four of TMS galacturonic acid was treated with a small amount of methanol. The TMS proton region of the p.m.r. spectra was monitored prior to and after addition of methanol. The spectra revealed a rapid loss of TMS ester. It is apparent that although precautions were being taken to keep the samples dry, loss of TMS ester was taking place. Samples of peak one of TMS galacturonic acid were dissolved in dry chloroform. Spectra run at intervals over twenty-four hours underwent changes. The changes in peak height were indicative of changes in substitu-tion of the galacturonic acid isomer isolated. These latter changes were relatively small in comparison to the complete loss of s i l y l ester observed when methanol was added. Losses, however, were large enough to confuse in-terpretation of spectra in an unknown compound.J It was possible to completely assign structures to components one and four of TMS galacturonic acid by interpretation of their p.m.r. spectra. Component one was an o<-furanoside and component four used the /3 -pyranoside of galacturonic acid. Component two was isolated in sufficient quantity to indicate that a l l couplings between protons were very small. This finding is consistent with a /S- furanoside structure where a l l ring protons can have ~ 90 52 For complete p.m.r. spectral analysis large scale sample separations •were carried out. Analysis of the p.m.r. spectrum for ring protons should confirm furano and pyranoside forms for the four peaks. The well separated f i r s t and last peak were isolated in a quantity large enough to confirm their structures by p.m.r. spectroscopy. Large scale separation of central peak components was extremely difficult. In addition degradation was observed due to prolonged sample collection. An experiment was carried out on an easily isolated peak to determine the exact cause of degradation. A sample of peak four of TMS galacturonic acid was treated with a small amount of methanol. The TMS proton region of the p.m.r. spectra was monitored prior to and after addition of methanol. The spectra revealed a rapid loss of TMS ester. It is apparent that although precautions were being taken to keep the samples dry, loss of TMS ester was taking place. Samples of peak one of TMS galacturonic acid were dissolved in dry chloroform. Spectra run at intervals over twenty-four hours underwent changes. The changes in peak height were indicative of changes in substitu-tion of the galacturonic acid isomer isolated. These latter changes were relatively small in comparison to the complete loss of s i l y l ester observed when methanol was added. Losses, however, were large enough to confuse in-terpretation of spectra in an unknown compound. It was possible to completely assign structures to components one and four of TMS galacturonic acid by interpretation of their p.m.r. spectra. Component one was an c<-furanoside and component four used the /3 -pyranoside of galacturonic acid. - Component two was isolated in sufficient quantity to indicate that a l l couplings between protons were very small. This finding is consistent with a ft- furanoside structure where a l l ring protons have small 53 dihedral angles. Sample purity, however, was insufficient, for assignment of ring protons. Low yield and peak overlap made impossible the large scale isolation of pure component three. The anomeric proton of this third fraction had a coupling constant of 4.3 Hz. The TMS proton spectra indicated a mixture of at least two components. The coupling constant 4.3 Hz was consistent with oC-galactopyranosyluronic acid [59]' One furanoside and one pyranoside form of galacturonic acid had been isolated and characterized. Since i t is known that anomeric configuration has l i t t l e bearing on mass spectral fragmentation, confirmation of the structure of the two central peak components was determined by mass spectrometry. A problem known to exist in interpretation of mass spectra of s i l y l ethers is the tendency of samples on ionization to undergo ring contraction and expansion. As a result, only small intensity differences exist in the m/e 191 and 204 mass peaks. These are sufficient, however, to confirm that the i n i t i a l two com-pounds isolated were TMS galactofuranosyluronic acid esters and the latter two peaks pyranoside forms. A mixture of aldobiouronic acids obtained from lemon gum was silylated with BTMSA. Although p.m.r. spectral evidence indicated formation of TMS esters in the aldobiouronic acid mixture, no product could be detected on subsequent gas chromatography. It appears that s i l y l esters are not sufficiently stable at high temperatures to allow separation of disaccharides. It may be possible, however, that part of the problem is associated with the gas chrom-atograph used in these experiments. Although BTMSA silylation results in formation of TMS ethers and esters, i t does not appear to be a useful reagent for analysis of uronic acids. Com-parison of obtained results with other methods of silylation of uronic acids [6o], indicates a marked change in equilibrium concentration for the various isomers. Other methods [60] indicate an isomer distribution more like that of the parent sugar galactose. Galactose has only a small amount of furanose present at equilibrium. FIGURE 17 PROTON MAGNETIC RESONANCE SPECTRA FOR THE TRIMETHYLSILYL PROTONS OF SILYLATED GALACTURONIC ACID TMS GALACTURONATE 100 MHz TMS ETHERS GLC PEAK. I a- FURANO GLC PEAK 2 P - FURANO PEAK 3 at - PYRANO GLC PEAK 4 ^-PYRANO TMS ESTERS u . V J • I I I - I I I L . TMS • • • • • • • • - i • • i i I 40 36 32 28 24 20 16 12 8 4 0 * METHANOL 5% IN CHCI3 REMOVES TMS ESTERS IN LESS THAN 30 MIN. ^ r 300 200 100 0 CrS TMS «-D-GALACT0FURANUR0NATE - GLC PEAK I -rV 5.01 4.51 "PPM 4101 ; 3.51 ; I FIGURE 18: P.M.R. SPECTRUM OF TRBETHYLSILYLATED GALACTURONIC ACID [G.L.C. PEAK 1] V.1 ON FIGURE 19: P.M.R. SPECTRUM OF TRE-IETHYXSILYLATED GALACTURONIC ACID [G.L.C. PEAK 2] 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 380 400 420 440 460 VJi 0 0 FIGURE 21: MASS SPECTRUM OF THE TRBLETHYLSILYLATION PRODUCT OF GALACTURONIC ACID [G.L.C. PEAK 1 ] 100 75 50 25 ~.:z~: 7 3 '-: TMS " " G A L A C T U R O N A T E '~fi-f zz z--'z:. •.•Ev E... -Z 1Z . : :: . zrz:i :zz ~ _ r: " .){-_:: ~y - . . _ _..'J\J 4 Zzztzz: Z'zZiFl zzzz zzizz: •. E-EE- rz t>L<J P t A K C ''• E.*E--- 1 — — 1 _ SZJzZZZZZ ^TTTTi "•" 7,"'—TTT^ T : E i > 'i~'[zlzz Z—ZZZZZZiZZZ r: ,zz". zz zz- ' S r ' i —• — • ^ r — t . r — t-- - — 1 — . — — ; ; ir-zzT zzz - — Ei — - — ZZZZZZZZ zzzzz: -: . zz-zELzizzzz zzzz. ~-FzTz'. zzzrl E E E E ; -- r — i 1 T _ . ! ; \ n r ~ E E = ZZZZZZZZZV—ZZ ~z z. zzzzz. • ... — ~z\ r_z:.\i:~. ™ "1 "7T_TE'." r~~ T .. 1 . - i - — , — , , —-T—.— -_— .zizzzzzrzzz .— • , , • * ' — 217 •: : .. ^ — - - : - 2 5 -~:~Z.iZZT.~ZLZZ ^— r r- : : -r ~ -'— ..zzzzzz f _ . . . ElEEEf: rzzzzzrzzxzzr-zf 1 i. ; - -zzzzzizz ;-.-K-:r= !. — r r — i - — i — _ Ef=r£i~ zZztzzzz j_ zt. _ ! : E j • : : • :•• zz-zzj-z — ~ <~ zz_~~ : zz-zztz-zzzzzzir = - --ZLi—\-.zzr - - { — 4 - - 1 - :| r\- • \ zzi E ^ - i ^ ^ ^ \zz::zzzz E l E r l E : .--:r- .--•fr— »=t ~ zz z^zzZEizzTE • T — — ' \.. -Z —{_— 0 BO I " : ' : 480 500 520 .5 ' JO 560 , • -"^2 L !—: ] ~£~-z-zzzzz ZZZZZZZ _4 — ~-zzzzzzzzzzzzzzz i —}._ : j . ! L - f z\zzzzzzzzzzzztzz "T . i — — 1 • : 1 — X3 Z ZZZ , .—1 tZzl EE " ;  : z\A ,,— ^ 1_ —2« — ->2" ~ Z - — i . 19 'ZZ"ZLZZ '_7_~ ™ " i EEEE zz .•^  - Jr - . -V ~ E " - ' I Q | " ' . • i zzzzzzzzz- •• " L-z - i : — - — j — •. —zz ~-V—- zz . •: \zzzz}- . _ ,-„ : zz -~---t— 1 " ::".::^ zr zzzzzz: Z- ::z. .r':\: • [ ... ' ; E : ' .E ^ • E f ^ E E E . 204 i ' ; — i f : • 1 i —jzzzzzz_ „l! •1 i d ' J ' 1 il H i i l i 1 n it i 1, • Jk . 1 A In1 \ Ml!! II I II 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 380 400 420 440 460 VJ, VO FIGURE 22: MASS SPECTRUM OF THE TRD-ffiTHYXSILYLATION PRODUCT OF GALACTURONIC ACID [G.L.C. PEAK 2] FIGURE 23: MASS SPECTRUM OF THE TRE-LETHYXSILYIiATION PRODUCT OF GALACTURONIC ACID [G.L.C. PEAK 3] FIGURE 2k: MASS SPECTRUM OF THE TRIMETHYLSILYLATION PRODUCT OF GALACTURONIC ACID [G.L.C. PEAK k] 62 METHYLATION OF LEMON GUM Methylation has been extensively employed in structural analysis of poly-saccharides. Free hydroxyIs in a polymer w i l l be converted to methyl ethers during methylation. Hydrolysis of a methylated polymer w i l l yield carbohydrate fragments characteristic of the original polymer's linkage. The main problems associated with methylation of polysaccharides are yield and degree of substitution. Methylations using the classical Haworth [6l ] and Purdie [62] procedures are often long, involving many repeated re-actions. Haworth methylations do not normally achieve a high degree of sub-stitution on a single reaction. In addition polysaccharides are in strongly basic solutions [20-30$] for prolonged periods of time. Purdie methylations are normally carried out on partially methylated polysaccharides. Silver oxide used in this methylation adsorbs polymer resulting in poor yields. This ad-sorption appears to be more pronounced with poorly methylated starting materials. Table XIV gives literature yields and information on polysaccharides methylated by these procedures. Losses incurred on complete methylation cast some doubt on the quality of the methylation analysis. When large losses occur i t is impossible to know i f some structurally significant portion of the polymer has been missed. Many other methylation reactions have been proposed in recent years. Table XV li s t s some of these with references to their use. Complete methylation of a polymer is necessary to avoid erroneous in-terpretation of the original structure [63]* Each unblocked hydroxyl detected in the methylated polysaccharide hydrolysate appears as a linkage point in the 63 TABLE XIV YIELDS AND DEGREES OF METHYLATION OF VARIOUS POLYSACCHARIDES Polysaccharide $ Yield . Methylated Polymer Methoxyl Found $ Ref. Acacia arabica Gum 56 38.2 [64] Armillaria mellea fruit body heterofcalactan 6l 41.6 [65] Amylose polyalcohol 65 • 44 .3 [66] Arthrobacter viscosus extracellular polysaccharide 60 44.7 [67] Tamarind kernal polysaccharide 54 39.5 [68] Aeodes orbitosa polysaccharide [desulfated] 26 35.0 [69] Aldotriouronic acids from Sapote Gum 1 45 N.R. [70] Aldotriouronic acids from Sapote Gum 2 13 N.R. 11 Aldotriouronic acids from Sapote Gum 3 25 N.R. H Synthetic araban 78 39.8 [71] A-Based on starting material weight over methylated polymer weight a no : result would be 135-145$. loss TABLE XV RECENTLY DEVELOPED METHYLATION PROCEDURES I Methyl iodide with silver oxide a] b]. in dimethylformamide [DMF] in dimethylsulfoxide [DMSO] [72] [75] II Methyl iodide with barium oxide a] or barium hydroxide b] in DMF in DMSO [74] [73] III Dimethylsulfate with barium a] oxide or barium hydroxide . b] in DMF in DMF and DMSO [75] [75] IV Dimethylsulfate with powdered sodium hydroxide in DMSO [76] V Methyl Iodide with sodium a] hydride b] ether type solvents DMF [77] [77] 64 original polymer. Since molecules such as galactose are known to be difficult to methylate at position four, errors result in proposed structures. An ex-ample of undermethylation being interpreted as a more complicated structure is outlined by Sandford and Conrad [78]. The method chosen to provide optimum yield and degree of methylation was the sodium hydride, DMSO and methyl iodide procedure of Hakomori [79]. Some workers have shown degradation of polysaccharides on use of these reagents. It was felt, however, that a polymer more closely resembling original material would be obtained by this method. Degradations which have been reported have occurred in uronic acid portions of carbohydrate polymers. Hakomori methyl-ations were shown to cause O-glucuronides to undergo elimination reactions yielding 4,5-dehydrouronic acids [54 ]. Conversely, Anderson and Cree [80] ex-amined methylation of some Acacia polysaccharides using this procedure. They looked specifically for condensation of ester groups with methylsulphinyl carbanion to give a sulfoxide or ,6 -elimination to unsaturated acids. No evidence was found for either of these reactions. The slight reaction found with benzyl glucuronide may have been due to unfavourable reaction conditions or simply a higher concentration of uronic acid. Identification of degradation product may be easier in a more concentrated system. Benefits derived from use of this methylation procedure outweighed any possible minor degradation of uronic acid. One point must be emphasized in the use of Hakomori methylations. In a meth-ylation of an acidic polysaccharide, uronic acids exist as sodium salts which limit any degradation to unsaturated uronate [8l]« Methyl esters are known to undergo elimination more readily. Uronic acid sodium salts are converted to 65 methyl esters during the course of methylation providing an opportunity for elimination. It appears from methylation results that some uronic acid salts remain after methylation and excess sodium hydride has been destroyed. Infra-red spectra for the methylated polysaccharides show the presence of carboxylate. Should a polymer undergo depolymerization due to formation of 4,5 unsaturation then reduction employing lithium borohydride in DMSO would eliminate the cause of the problem. Hakomori methylation produces virtually complete methylation in a single reaction. Several possible methods of analysis are available for the methylated sugars present on hydrolysis. The preferred techniques involve use of gas liquid chromatography. Several methods of sample preparation for G.L.C. are possible. Depolymerization of methylated polysaccharide may be achieved by methanolysis. Dependent on the sugars present, the samples may be analysed directly or further reacted to acetates or s i l y l ethers. Acid hydrolysis give free sugars which may be reduced by sodium borohydride to alditols. The samples may be made volatile by acetylation or silylation. A common procedure is simple methanolysis, since volatile derivatives are obtained for most partially methylated sugars. Although G.L.C. retention time offers a good indication of the nature of a product i t is by no means a positive identification. Further tests must be carried out to provide positive identification. The classical approach has been to isolate the components and prepare crystalline derivatives. This method of analysis can be extremely difficult. Contaminating products can be completely missed because of purification during derivative preparation. The need for a more rapid and smaller scale identification has prompted many workers to employ mass spectrometry. Good mass spectra may generally be obtained on samples which pass through a gas chromatograph. Two procedures 66 have been employed before mass spectral analysis. Methanolysis products have been silylated for greater thermal stability. This method [82] provides very precise identification of components. A drawback., however, is the formation of different ring and anommeric forms. For example, 2,3,6 tri-O-methyl-D-galactose will form at least three different derivatives, fairly large amounts of both pyranosides and smaller amounts of a furanoside. These extra peaks can aid and/or hinder analysis of complex mixtures. Lindberg [83] has developed an alternate procedure employing partially methylated polyol acetates. Methylated polysaccharides are hydrolysed with 72$ sulfuric acid, reduced with sodium borohydride, acetylated and analysed by gas chromatography. Mass spectral fragmentation is dependent on methyl ether substitution of the original polyol and provides for very easy interpretation of structure. Resolution of methylated polyol acetates is less than for cor-responding methyl glycosides. This lack of resolution can introduce com-plications during analysis of complex mixtures. 67 MASS SPECTROMETRY OF METHYLATED POLYOL ACETATES The following is a brief outline of controlling factors in mass spec-trometry of partially methylated alditol acetates as outlined by Lindberg [83]. The base peak in almost a l l determined spectra was m/e 43 [CR^CO*]. It was stated that only peaks of intensity, greater than 10$ of base peak were in-cluded. It should be noted that spectra of isomeric alditols having the same substitution pattern are not sufficiently different to allow for distinction between them. Thus information w i l l sometimes be lost when converting into alditols. For example, mass spectra from 2,3 and 3,k di-0-methyl pentoses w i l l be identical. Reduction with deuterioborohydride would eliminate this difficulty [83]. The proposal is that intense peaks [> 10$ base peak] of the mass spectrum and retention time of the component on G.L.C. in combination with sugar composition of the original polymer w i l l , in most cases, give sufficient evidence for an unambiguous characterization of the methylated sugar. Chizhov et a l . [84] have determined that primary fragments from alditol acetates arise through fission between two carbon atoms of the carbon chain. Either fragment can carry the positive charge. Secondary fragments are formed from primary fragments by elimination of acetic acid [62] and/or ketene [42]. Intensity of primary fragments increases with decreasing mass number. Chizhov et a l . [84] have shown that intensities of primary fragments given by alditol acetates are much lower than intensities of corresponding peaks from partially methylated alditol acetates. Fission takes place between carbon .atoms in partial structures V and VI in preference to structure VII. Positive ions are stabilized by methoxyl groups. Primary fragments of high intensity 68 containing two vicinal methoxyl groups are not observed, except when they can be formed by fission between two other vicinal methoxyl groups, indicating that structure V is cleaved in preference to structure VI. CH3OCH HCOCH3 HC-OCH3 HC-OCCH3 0 HC-O-C-CH3 HC-O-C-CH3 li. 0 V VI VII The secondary fragments observed in methylated alditol acetates are der-ived from primary fragments by single or consecutive elimination of acetic acid [60], ketene [42], methanol [32] or formaldehyde [30]. The lowest molecular weight primary fragment, m/e 45 [A], is produced by substances having a methoxyl group at C-l. No other polyol produces this pri-mary peak. A peak [m/e 45] of low intensity [<5$] nay be seen in mass spectra of other substances and is probably a secondary fragment. Primary fragment D, m/e, 117, is obtained when C-l is acetylated and C-2 methylated. An analogous primary fragment m/e 131 [E] is obtained from a 6-deoxy hexitol methylated at C-4 but not at C-5« CH2=0-CH3 - + HC=0CH3 . H2C-OCCH3 II 0 HC=0-CH3 HC-OC-CH3 II 0 CH3 D 69 Lindberg [83] reports that two primary fragments have m/e l 6 l , Fi and Fa- Of these, Fx is obtained in low intensity from alditols methylated at positions 2 and 3. If k is also methylated then i t becomes a prominent frag-ment. Fragment Fa, obtained from alditols methylated at positions 1 and 3> is always of high intensity. The secondary fragments m/e 129 and 101 are ob-tained from Fi and Fa by loss of methanol and acetic acid respectively. Further loss of ketene from m/e 129 gives m/e 87. Loss of formaldehyde from m/e 101 may give m/e 71. Primary fragment m/e 175 [G] is obtained from 6-deoxyhexitols methylated in positions 2, 3 and h. HC=0CH3 I HC-OCH3 H2C-OC-CH3 II 0 Fi + HC=0CH3 0 HC-OCCH3 H2C-OCH3 Fa + HC =0(3*3 HC-OCH3 I HC-OCOCH3 CH3 G + HC=0CH3 HC-OCOCH3 HaC-0C0CH3 H + HC=0CH3 I HC-OCOCH3 HC-OCOCH3 •I CH3 I Primary fragment m/e 189 [H] is given by alditols methylated at position 3 but not at positions 1 or 2. Secondary fragment m/e 129 may be formed by elimination of acetic acid from H. Loss of formaldehyde or ketene from m/e 129 gives respectively m/e 99 or 87. M/e 203 [I] is an analogous primary frag-ment obtained from a 6-deoxy hexitol methylated at position 3 but not h or 5* Three primary fragments of m/e 205 way be expected [Ki, Kg and K 3]. The fir s t , derived from alditols methylated at C -2 , 3 and k, is not observed in high intensity. Fragment Ka is observed from alditols methylated at positions 1, 2, h and 55 K3 from alditols methylated at positions 1, 3> h and 5« In either case i f position 5 is n°"t methylated, the corresponding peak appears with only low intensity. Several secondary fragments may be obtained from Ki, K 2 and K3. Loss of acetic acid from K3 to produce m/e 145 appears to be an especially important secondary fragment. 70 + HC=0CH3 HC-och3 HC -OCH3 H 0 C - O C O C H 3 + HC=0CH3 . HC-OCOCH3 I HC-OCH3 I H2C-OCH3 + HC=0CH3 I HC-OCH3 I HC-OCOCH3 H2C-OCH3 HC=0CH3 I HC-OCOCH3 I HC-OCOCH3 HaC-0CH3 HC=0CH3 . I HC-OCOCH3 I HC-OCOCH3 I K2C-OC0CH3 Ki K3 M Three primary fragments of m/e 233 may be expected. One of these [L] is obtained from alditols methylated at C-l and 4. A prominent secondary peak [m/e 113] is most probably derived from L by loss of two molecules of acetic acid. The primary peak m/e 2 6 l [M] is the highest mass number of relatively high intensity. Alditols methylated at 3 and acetylated at k, 5 and 6 produce this peak. Lindberg et a l [83 ] have used formation of these primary and secondary fragments to analyse partially methylated alditol acetates. Tables of fragments expected under these rules are presented in Lindberg's i n i t i a l paper on mass spectrometry. Knowledge of the partially methylated polyols has aided their analysis of polysaccharide structure. The Lindberg procedure for analysis was chosen because i t offered an inate simplicity not present in other analytical schemes. Lemon gum has been analysed by methanolysis [ l ]. It was felt that this sort of complex polysaccharide would give a good evaluation of reduction - acetylation as a method for analysis of methylated polysaccharides. 71 Four polysaccharides were methylated by the Hakomori method and analysed by Lindberghs mass spectrometry - gas chromatography procedure. The four -were: 1. Lemon gum 2. Lemon gum degraded with l/lQN H2SO4 3. Lemon gum periodate oxidized and borohydride reduced [before hydrolysis] k. Lemon gum periodate oxidized and borohydride reduced [after 0.5N H2S04 hydrolysis] METHYLATION ANALYSIS OF LEMON GUM Lemon gum was methylated using the Hakomori method. Extreme care was. taken to avoid interfering side reactions. After methylation, uronic acids were reduced with lithium aluminum hydride. Reduction avoided problems assoc-iated with formation of aldobiouronic acids on hydrolysis. Since lemon gum was fully methylated before reduction, no methylation was carried out to block the six positions in the newly formed neutral sugar. A l l steps in the reactions outlined produced high yields. After reactions to form acetylated-methylated polyols, analysis was carried out on the gas chroma to graph. Two columns were evaluated, one of ECNSS-M, the other butanediol succinate. Resolution of the mixture was virtually identical, with butanediol succinate being slightly superior. Samples were collected for analysis on the mass spectrometer since access to a coupled GLC-mass spectrometer was not possible. The collected components were analysed by thin layer chromatography [T.L.C.]. T.L.C. revealed that single peaks on the gas chromatograms were, in fact, at least two components. Using thin layer chromatography, samples of the components could be separated. The samples were then reanalysed by GLC and T.L.C. to ensure their purity. 72 32 MIN. FIGURE 25: GAS CHROMATOGRAM OF LEMON GUM AFTER METHYLATION, LITHIUM ALUMINUM HYDRIDE REDUCTION, HYDROLYSIS, BOROHYDRIDE REDUCTION AND ACETYLATION [140-195°C at 3 °M n - 3 O O O OO-o o ° 8 CO O F R O N T O *10 *9 *a -7 *6 "5 M X 3 " 2 *1 FIGURE 26: THIN LAYER CHROMATOGRAM OF THE TEN FRACTIONS OF LEMON GUM METHYLATED POLYOL ACETATES 73 During the isolation and purification of these components, a technique was developed for handling microgram quantities of compounds collected from the gas chromatograph. This procedure for micro manipulation is discussed in the appendix to the thesis. Mass spectra were obtained on a l l significant components isolated by gas-liquid chromatography. Analysis of the spectra indicated that although simple rules had been set down by Lindberg, intensity of some component peaks was altered. This may be a function of the means of sample introduction. Samples to be analysed were inserted into the spectrometer via the probe. It was not possible to identify components using the simple rules proposed by Lindberg. It was, however, possible to identify the majority by careful consideration of fragmentations. Care was taken with the spectra to ensure no background contribution which would have lead to misleading peak intensities. Two checks were made on the spectra. A background spectrum was obtained and subtracted from the component, spectra. Analysis of the spectra in their order of production confirmed that major peaks were not contributing to subsequent spectra. The spectra are in-cluded in the appendix and numbered to indicate both order of elution and order of mass spectra determination. Although Lindberg and co-workers utilized a coupled gas chromatograph-mass spectrometer, there are advantages in using direct sample collection. The presence of overlapping peaks is a clear example of need for further separations. Peak broadening also occurs in the gas chromatograph-mass spectrometer separator chamber. The method of micro manipulation outlined in the appendix has proven extremely valuable in purification and sample analysis. 74 Variation in peak intensity for some mass values has created a problem. Peaks which were expected to be of intensity greater than 10$ were, in fact, less than this. In addition peaks at lower m/e values had higher intensity than predicted. Availability of complete spectra would have aided considerably interpretation of structure. '. > Quantitative estimation of the identified components was not possible due to lack of thermal conductivity molar response factors for partially methylated alditol acetates. A relation should exist between subsitution pattern and molar response factor for a given sugar polyol. A relatively large number of molar response factors are available for flame ionization equipment but none are presently available for thermal conductivity detectors. METHYLATED POLYOLS FROM LEMON GUM Peak 1: Component one on the gas chromatograph was unidentified. The retention time was less than that reported for the most volatile sugar alditol polyol. The increased volatility of this component was likely due to elimin-ation or anhydride formation. Hydrolytic etherification has been reported for partially methylated polyols and sorbitol [85]. Strong peaks at m/e 117 and 101 indicated the molecule was terminated by an acetyl group adjacent to two methoxyls. A strong m/e 88 peak was not consistent with any reported peak. Strong even numbered fragments do not often occur in mass spectra. Peak 2: The spectrum for this component does not correspond to any of the known methylated polyol acetates. Its structure does, in fact, appear to be very different from that of a polyol. None of the characteristic primary 75 fragments except m/e 45 can be seen. Although some masses correspond to secondary fragments, i t is difficult to determine their origin. In addition a peak m/e 74 is almost as intense as the base peak m/e 43* Peak 3'» T.L.C. indicated that peak three contained two components. It was confirmed experimentally that 2, 3, 5-toi-O-methyl-L-arabinitol and 2, 3, 4-tri-O-methyl-L-rhamnitol acetates overlap as a single peak. Two components are clearly seen in the mass spectrum. This ability to distinguish two com-ponents as a result of knowledge about sugar composition and chromatographic behaviour could be used to determine percentage composition. Only standard spectra of the two components are required. Peak 4: This component, present in only small amounts, is 2 , 3> 4 - t r i -O-methyl-Lj-arabinitol acetate. With this component i t became clear that the level of 10$ of base peak for included peaks was not going to be applicable. In addition to m/e values of 43, 101, 117 and l 6 l required for identification, m/e values 45, 5$ and 87 were a l l above 10$ of base peak. Peak 5s This peak contains two components, which, being very minor con-stituents, were impossible to separate. Some information was available from the mass spectrum of this peak. Additional information may be deduced from G.L.C. retention time. Since 2, 3, 4, 6-tetra-0-methyl-D-galactitol acetate comes immediately after these two components their retention times must be less than its value. The only two applicable compounds are 2, 5*and 3> 5~ di-O-methyl-L-arabinitol acetates. The mass spectrum also indicated the pres-ence of these two components. Peaks associated with 3> 5-di-Q-methyl-L-arabinitol are of greater intensity and correspond in strength to the second 76 peak of peak 5« Positive identification of the smaller component as 2,5-di-O-methyl-L-arabinitol is clouded by strong signals from 3/5-di-Q-methyl-L-arabinitol. Identification is based on m/e 233 and 113 peaks. Reduced in-tensity of higher mass values increases the difficulty of identification •without pure standard spectra. Peak 6: This peak contained two components as shown by T.L.C. Peak 6 was collected and separated into its components by preparative T.L.C. Reiso-lation of the constituent parts by gas chromatography indicated no con-tamination from peaks 5 or T- Analytical T.L.C. confirmed the clean sep-aration of peak 6 alditols. The two compounds were designated 6T and 6B on the basis of position on ascending T.L.C. Peak 6T; 2,3-di-O-methyl-L-arabinitol acetate is the major component of peak 6. This molecule is an example of a component that could be derived from two sugars [2,3-or 3,k-&±-O-methyl-L-arabinose]. Both sugars reduce to polyols which fragment in an identical manner. Signal strength for higher m/e peaks was reduced in intensity. Peak 6B: This constituent proved compounds present as minor constituents in a gas chromatographic peak can be isolated and characterized. Peak 6B is 2,$>h,6-tetra-O-methyl-D-galactitol acetate. Except for reduced intensities for expected peaks the mass spectrum is very clear. A standard sample of this compound coincided in retention time with peak 6. This retention time con-firmation aided identification of preceeding chromatographic peaks. Peak 7: Like peak 6 this fraction contained two components. They were separated, numbered and confirmed pure in the same manner as components in peak 6. 77 Peak 7T: Analysis of this compound's mass spectrum showed i t to he a 2,3,4-tri-O-methyl-hexitol. Since a l l hexitols substituted in a like manner give similar fragmentation patterns i t was necessary to consider G.L.C. re-tention times. Of the two possible 2,3,4-tri-O-methyl-hexitols in reduced lemon gum, glucitol has the shorter retention time. Therefore, peak 7T is 2,3,4-tri-O-methyl-D-glucitol acetate. Peak 7B; This peak was thought originally to be only one compound. On detailed examination of the mass spectrum i t appeared to consist of two com-ponents. No one compound could produce the necessary fragments to account for a l l peaks occurring in the mass spectrum. The strong m/e 233 peak must arise from 2,3,6-tri-O-methyl-D-galactitol acetate. Only two other compounds produce this peak and they require a strong m/e 189 which was not present. M/e l 6 l was of such intensity that i t had to be considered significant. Only 2,4,6-tri-O-methyl-D-galactitol acetate had a fragmentation pattern compatible with the peaks present. A l l other compounds which give rise to m/e l 6 l have peaks at m/e 145 or 189, or are chromatographically unacceptable. Component 7B consists of a mixture of 2,3,6-and 2,4,6-tri-O-methyl-D-galactitol. Peak 8: T.L.C. indicated that this peak was a single component. The mass spectrum was virtually the same as that for 7T, indicating an identical methyl substitution. A longer retention time was consistent with the structure pro-posed, 2,3,4-tri-O-methyl-ri-galactitol acetate. Peak 9: There were at least two components in this gas chromatograph peak. The small quantity of these constituents made purification by T.L.C. impossible. One component appeared to be 2,3-di-O-methyl-D-glucitol acetate. The gas chromatogram retention time was identical with an authentic sample. The second component in this peak may be 2,3-di-O-methyl-D-galactitol. The 78 mass spectrum of this peak would not distinguish between these two compounds. Some peak intensities, however, appear to be too intense, for example m/e 85 and 127. These fragments should arise from a 3-0-methyl hexitol, however, the G.L.C. retention times for such compounds are too great. No other polyol accounts for these mass fragments. Peak 10: This peak typifies the difficulties that may be encountered when trying to interpret mass spectra of methylated polyol acetates. Peaks for mass values which should have been greater than 10$ of the base peak did not reach this value. Peaks with intensities less than 10$ of base peak can then not be completely rejected. Thus the mass peak m/e 233 i n component 10 which would not normally be considered must be evaluated. Complete spectra present a better picture of relative intensities to be expected. Also, a preliminary table such as given by Lindberg et a l [83] is a definite aid. If the table listed values of peak intensity relative to the base peak, an easier inter-pretation of spectra might be possible. Peak 10 is 2,4-di-O-methyl-D-galactitol although values of certain fragments do not exceed the required 10$ value. The component compounds from fully methylated lemon gum are listed in Table XVI, together with peak area for the components. TABLE XVI METHYLATION ANALYSIS OF LEMON GUM Compound Area $ Unknown . 1 0.51 Unknown 2 0.70 2,3,5-Me3-ARAB 7.02 2,3,lj.4te3-HHM 3 J 2,3,4-Me3-ARAB k 1.40 2,5-Me2-ARAB 5 1.83 3,5-Me2-ARAB 5 3.62 2,3-Me2-ARAB 611 21.7 2,3,4,6-Me4-GAL 6B. 2,3,4-Me3-GLUC .. 7T " 28.5 2,3,6-Me3-GAL 7B 2,4,6-Me3-GAL 7B. 2,3,4-Me3-GAL 8 3.96 2,3-Me2-GLUC 9 " 3.16 Unknown 9 . 1 . . . •  • 2,4-Me2-GAL 10 - 27.6 8o METHYLATED DEGRADED LEMON GUM When lemon gum was hydrolysed with l/lON sulfuric acid a polymer was obtained which contained essentially no L-arabinose. Methylation of this product gave information about the galactan framework of the plant gum. In addition, i t provided further proof regarding structures of some partially methylated galactitols derived from lemon gum. The acid degraded polymer was methylated using the Hakomori method [79]. Hydrolysis was carried out by Lindberg's method [5 ], uronic acids being removed by ion exchange resin. The acids were not examined further, since they represented only a small amount of acidic material. Neutral methylated sugars were reduced and analysed in a manner similar to that described for lemon gum. There were four principle peaks in the gas chromatogram. FIGURE 27: GAS CHROMATOGRAM OF DEGRADED LEMON GUM .METHYLATED POLYOL ACETATES [140-205°C at 3°/min.] log I/ION H2S04 LEMON GUM 4 8 12 16 ~20~ 24 28 32 MIN. The peaks were analysed by mass spectrometry with the following results. — 1 — 4 Peak 1: The most volatile of the major components exhibited a mass spectrum characteristic of 2,3,4,6-tetra-O-methylrD-galactitol acetate. Again, intensity of some peaks of lower mass was greater than the 10$ cut off suggested 8 i by Lindberg. Two such peaks are m/e 55 and 57* The spectrum, however, was consistent with fragmentation of 2,3,4,6-tetra-O-methyl-D-galactitol acetate in a l l other respects, clearly indicating the value of the method. Peak 2: This gas chromatographic peak corresponds to 7B from lemon gum. It contains two inseparable components, 2,3,6-and 2,4,6-tri-O-methyl-D-galactitol. The exact molar balance between these two components w i l l have to await ac-curate intensity determinations for the two pure compounds. Peak 5; The component contained in this peak was 2,3,4-tri-O-methyl-D-galactitol, corresponding to peak 8 in the gas chromatogram of methylated lemon gum. Peak hi The component of this peak was 2,4-di-O-methyl-D-galactitol acetate and corresponds to peak 10 of lemon gum. TABLE XVII METHYLATION ANALYSIS OF DEGRADED LEMON GUM Compound G.L.C. Peak Area i» I n i t i a l k peaks 1.09 2,3,4,6-Me4-GAL 1 27.8 2,3,6-Me3-GAL 16.6 2,k,6-Me3-GAL 2 j 2,3,4-Me3-GAL 3 34.0 ' 2,4-Me2-GAL k 20.5 82 METHYLATION ANALYSIS OF LEMON GUM AFTER PERIODATE OXIDATION AND BOROHYDRIDE REDUCTION Methylation of periodate degraded lemon gum should generate information concerning free hydroxyl positions of the remaining intact polysaccharide. Total hydrolysis of the methylated polymer •would produce sugars from lemon gum which contained no vicinal bydroxyls. Gas chromatographic evidence in-dicated that this was achieved to a limited extent. However, mass spectra of isolated peaks were very poor. Peak 1; This component displaying chromatographic characteristics of 2,3,5-tri-O-methyl-L-arabinitol probably resulted from incomplete periodate oxidation and/or basic hydrolysis during methylation. In addition, some mild acid hydrolysis may have occurred during isolation of the borohydride reduced polymer. It was not possible to collect this peak because of sample volatility and the small quantity present. Peak 2: Like peak 1 this component was present in very low concentration in the mixture of methylated polyol acetates. Retention time was similar to that of 2,3,4-tri-O-methyl-L-arabinitol acetate. The mass spectrum was similar to peak k of lemon gum but additional mass peaks indicated the small sample collected was impure. Peak 5? Like the previous two peaks this was a very minor peak in the gas chromatogram. At most i t represented only 3 percent of total sample. A figure of 1.5 mole percent is probably more realistic. The G.L.C. retention time corresponds to that for 2,5-and 3>5-di-Q-methyl-L-arabinitol acetate. The mass spectrum can be assigned to these compounds, however, the intensities make absolute assignment di f f i c u l t . • 83 Peak hi The fir s t of the major peaks, this component had the gas chro-matographic characteristics and mass spectrum of 2,3-di-0-methyl-L-arabinitol acetate. Since this product contains vicinal hydroxyIs, blocking groups may have been removed by elimination reactions during the strongly basic methylation. It would appear that a small amount of tetra-O-methyl-D-galactitol may be present in this sample. Peak 5: This peak consisted of two components on T.L.C. Attempted sep-aration of the small quantities available was unsuccessful. Since this peak cor-responded in R^, to peak 7 of lemon gum, one component should have been 2 ,4 ,6 -tri -0-methyl-D-galactitol acetate. Peaks resulting from this product may be seen in the mass spectrum, however, there are other minor mass peaks consistent with the other consituents of peak 7 of lemon gum, 2,3,6-tri -0-methyl-D-galactitol and 2,3,4-tri-O-methyl-D-gulcitol. Peak 6: The major component isolated from this oxidized gum was 2,4-di-Q-methyl-2-galactitol acetate, accounting for two thirds of the isolated methylated polyols. This information provided further proof that the gal-actan framework of this polymer is highly branched with many side chains. Analysis of a periodate oxidized borohydride reduced polysaccharide should be an extremely useful analytical tool. It appears necessary that more study be made of parameters involved in possible hydrolysis during methylation. Lemon gum may not undergo complete oxidation. Alginate ' has shown this dif-ficulty. This underoxidation may account for the presence of fully methylated sugars in the hydrolysate after methylation. FIGURE 28 GAS CHROMATOGRAM OF LEMON GUM SMITH POLYALCOHOL METHYLATED POLYOL ACETATES [l40-205°C at 3 % i n . ] 4 1 1 1 1 1 1 1 I 1 i I I 1 r — — T 4 8 12 16 20 24 28 32 MIN. TABLE XVIII METHYLATION ANALYSIS OF LEMON GUM SMITH POLYALCOHOL Compound ' * ' Area # 2,3,5-Me3-ARAB . 1 1.03 2,3,4-Me3-ARAB 2 1.94 2,5-Me2-ARAB 3 0.37 3,5-Me2-ARAB 3 3«10 2,3-Me2-ARAB 4" 12.9 2,3,4,6-Me4-GAL 4 . 2,3,6-Me3-GAL 5 " 18.4 2,4,6-Me3-GAL 2,3,4-Me3-GLUC 5 . 2,4-Me2-GAL 6 62.3 85 METHYLATION OF LEMON GUM SI Lemon gum SI was obtained from lemon gum after periodate oxidation, boro-hydride reduction and mild acid hydrolysis. The reduced and acetylated hy-drolysate from methylated lemon gum SI gave a gas chromatogram much like that of lemon gum with some notable exceptions. Peak 1: The structure of this component was 2,3,5-tri-O-methyl-L-arabinitol. The mass spectrum was similar to that from peak 3 of lemon gum with the minor contribution of 2,3,4-tri-0-methyl-L-rhamni tol removed. Peak 2: Corresponding to peak 6 of lemon gum in R^, this peak contained only one component. That component was identified as 2,3,4,6-tetra-Q-methyl-D-galactitol acetate from its mass spectrum. Peak 3: Corresponding to peak 7 of methylated lemon gum this peak has no tri-O-methyl glucitol constituent. It contained, therefore, only 2,4,6-and 2,3,6-tri-O-methyl-D-galactitol. Peak 4: Corresponding to peak 8 of lemon gum this peak was mainly 2,3>4-tri-O-methyl-D-galacitol. The presence of this component indicates the im-portance of the 1—*"6 linkage in lemon gum. The increased signal strength for m/e 45 may be due to a trace of 2,6-di-0-methyl-D-galaeitol acetate. This compound has the correct mobility and might result from undermethylation. Peak 5: Like peak 9 of lemon gum this was a rather odd peak. Its mass spectrum fragmentation pattern indicated the presence of 3-Q-methyl-D-gal-actitol acetate. The spectrum, however, lacks m/e 189 and i t may be that this component both here and in lemon gum is a specific degradation product capable of fragmenting like 3-0-methyl-D-galactitol. This peak also appears to contain 23-di-O-methyl glucitol acetate. This compound may indicate some underoxidation during periodate cleavage. 86 Peak 6; The final peak is that of 2,4-di-Q-methyl-D-galactitol. FIGURE 29 GAS CHROMATOGRAM OF LEMON GUM SI METHYLATED POLYOL ACETATES [UO-205°C at 3°/min.] LEMON GUM ; S \ ' 4 TABLE XIX METHYLATION ANALYSIS OF LEMON GUM SI Compound G.L.C• Peak Area $ 2,3,5-Me3-ARAB 1 4.69 3,5-Me2-ARAB Between 1 and 2 1.86 2,3A>6-Me4-GAL 21 15.5 2,3-Me2-ARAB 2 . 2,3,6-Me3-GAL 5 ] 15.1 2,4,6-Me3-GAL 3 . 2,3,VMe3-GAL k 29.7 ' 2,3-Me2-GLUC 5 ' 3.4 Unknown 5 . 2,4-Me2-GAL 6 29.0 87 The present work on lemon gum is the fi r s t example of analysis of a plant gum by the Lindberg et a l [83] method of mass spectrometry [partially methylated alditol acetates]. The complexity of the lemon gum mixture analysed was such that alternate separations were required on TLC. Micro manipulation, as out-lined in the appendex, of less than milligram quantities has facilitated the purification and mass spectrometry of the isolated methylated alditol acetates. Hakomori methylation [79] has given high yields and complete methylation im-proving the reliability of methylation analysis. The combination of Hakomori methylation and methylated alditol acetate mass spectrometry has reduced the quantity of polysaccharide needed for methylation analysis to the milligram range. The ability to manipulate these sub-milligram quantities of methylated hydrolysis products from polysaccharides w i l l be of immense value to workers faced with complex mixtures or mass spectrometers without a coupled gas chromato-graph. Methylation analysis of lemon gum and degraded lemon gum are in good agreement with results obtained and the structure proposed by Jones and Stoddart. Analysis of Smith degraded lemon gum is also in good agreement with one minor exception. In the previous analysis of lemon gum there was a substantial and increasing amount of 3>5-di-0-methyl-L-arabinose in the Smith degraded gum. The increase has not been confirmed by this work - an unexpected result in view of the generally milder hydrolysis conditions employed in the preparation of the Smith degraded lemon gum. 88 EXPERIMENTAL GENERAL METHODS Paper chromatography was carried out on Whatman No. 1 and 3 paper by the descending method in the following solvent systems: a] Ethyl acetate-acetic acid-formic acid-water [18:3:1:4] . b] Ethyl acetate-pyridine-water [8:2:2] top layer. c] Butan-l-ol-pyridine-water [6 :4 :3] . d] Butanone-water azeotrope. Thin layer chromatography was carried out on alumina, and s i l i c a gel [with and without calcium sulfate binder] by the ascending method in the following solvent systems: e] Butanone-water azeotrope. f] Benzene-methanol [96:4]. g] Ethyl ether-toluene [2 :1] . Sephadex gel permeation chromatography was performed in a Sephadex column K25/45 fitted with flow adaptors using Sephadex G-15 fine. Prior to packing the column, G-15 was swollen in water for 16 hours. A constant pressure head was maintained by use of a Mariotte flask. Operation of the column was in the ascending mode using water with 0.5$ chloroform as eluting solvent. Samples applied to the column through the flow adaptors using a hypodermic syringe barrel with luer lock adaptor were collected on elution by a; timed fraction collector. Reducing sugars on paper chromatograms were detected with either p-anisidine hydrochloride reagent [86] or alkaline silver nitrate dip [87]. Non-reducing carbohydrates and polyhydric alcohols were detected with either the latter dip 89 or alkaline periodate spray reagent [88]. The rate of movement of substances on paper chromatograras is quoted relative to the solvent front or an internal standard such as galactose. Visualization of T.L.C. was accomplished by charring [sulfuric acid spray]. When samples were to be isolated, silicone grease was spread in thin ribbons on a clean glass plate. When pressed to the T.L.C. plate, a small amount of adsorbant was transferred, then charred in the normal manner. Melting and boiling points are uncorrected. Unless otherwise stated optical rotations were measured at 21 + 3°C. Infrared absorption spectra were recorded using a Perkin-Elmer model 257 or 21. Samples were examined as KBr pellets, chloroform or carbontetra-chloride solutions [5-10$ W/V] in sodium chloride cells, or as thin films on sodium chloride plates. Nuclear magnetic resonance spectra were determined on a Varian HA-100 or A-60 spectrometer. Mass spectra were determined on an Atlas AEI MS9 spectrometer. Gas liquid chromatography was carried out on an F and M model 720 dual column instrument fitted with thermal conductivity detectors. Polyhydric alcohols and sugars were analysed on two columns [8 f t . x 0.25 in. coiled copper] packed with equal weights [to within 20 mg] of 20$ SF 96 on 60-80 mesh Diatoport S. The columns were held isothermally at 130° for 6 rain, and then programmed at 3° per minute to hold at 220°. During analyses for sugars only, the programme was started at 190° and immediately programmed at 2° per minute to hold at 220°. Thevinjection port was 270°, the detector block 295° and the helium flow 88 ml per minute [6.8 sec. for 10 mis]. 90 For simple systems of the above compounds other columns may be used. Thus s i l y l derivatives of ethylene glycol, glycerol and methyl /3 -D-gluco-pyranoside may be separated on a 2 f t . x 0.25 in. column of SE30. This column was run isothermally at 70° until the ethylene glycol emerged, then programmed at 10° per minute to hold at 250°C Sugars were dissolved in water containing a small amount of chloroform [0.5$ W/V] and allowed to come to equilibrium. Aliquots were removed and con-centrated to dryness at 40° on a rotary evaporator. Solutions of the poly-hydric alcohols in anhydrous pyridine were added to give the test solutions reported in Tables II - VI. These pyridine solutions [10 parts] were immediately silylated with hexamethyldisilazane [k parts] and trimethylsilyl chloride [2 parts] and after 5 minutes shaking, portions were injected directly. The values quoted in the Tables represent the mean of at least five deter-minations . For the determination of molar response factors separate solutions of the individual compounds with butane-1,4-diol were prepared. In the case of the sugars these runs also served to determine relative concentrations of different forms at equilibrium. In preliminary experiments i t was ascertained that none of the compounds involved were lost on passage through Amberlite 1R 120 and Duolite A-4 resins nor on concentration of their aqueous solutions. Methylated polyol acetates obtained from methylated polysaccharides were examined on columns with ECNSS-M and butanediol succinate liquid phases, the latter giving a slightly improved separation. Methylated polyol acetates were analysed on a stainless steel column [6 f t . x 0«25 in.] with butanediol 91 succinate [10$ W/W] on Diatoport S [6O-8O mesh]. Column temperature was programmed immediately from 140-205°C at 3° per minute using the standard flow rate and other settings. Peak areas were measured in i t i a l l y with a Disc integrator with suitable correction for baseline drift. In later work an Infotronics digital integrator was used. Silylation of galacturonic acid was carried out in a variety of solvents using bis-trimethylsilyl acetamide as the silylating reagent [89]. Acetylated carbohydrates were prepared by two methods. Some samples were dissolved in anhydrous pyridine and an excess of acetic anhydride added. The samples were allowed to stand 16 hours, then heated to 60°C for 1 hour. The solvents were removed by rotary reduced pressure evaporation. Alternatively, samples were dissolved in pyridine and an excess of acetic anhydride added. The samples were sealed in test tubes and heated at 100°C for 1 hour. Work up was similar to the above case. Unless otherwise stated hydrolyses were carried out by either of the following methods. A sample [5-20 mg] was dissolved in sulfuric acid [UN] [5-10 ml] and hydrolysed in a sealed tube at 100° for 8 - l 6 hours. Alternatively a sample [5-20 mg] was solubilized with 72$ sulfuric acid [1 ml] for 1-2 hours at room temperature, then diluted to 7«2$ with distilled water and hydrolysed at 100° in a sealed tube for k hours. After cooling the hydrolysates were neutralized partially with barium hydroxide, then completely with a small amount of barium carbonate. Alternatively complete neutralization was achieved using a slurry of barium carbonate. The precipitated barium sulfate together with 92 the residual barium carbonate -was removed by centrifugation. The. supernatant -was then passed through a small bed of Amberlite TJR120 [H + form] ion exchange resin to remove any cations present in solution. The column of Amberlite was often placed in series with a column of Duolite A-4 [OH - form] ion exchange resin to remove sugar acids or other acidic material. The eluant was con-centrated to a syrup by evaporation under reduced pressure [ca. 15 mm]. Reduction of free sugars was carried out on the eluant from the ion ex-change column [used to remove barium ions s t i l l in the hydrolysate]. A large molar excess [3-5 moles] of sodium borohydride was added to the eluant. After 24 hours excess borohydride was destroyed by addition of acetic acid. Sodium ions were removed using a column of Amberlite IR 120 [H + form] ion exchange resin. The solution was evaporated to dryness under reduced pressure, the resulting syrup being repeatedly dissolved in methanol and evaporated to dry-ness. The repeated evaporation served the dual purpose of removing the boric acid and traces of acetic acid remaining in the reduced sugar sample. Reductions with lithium aluminum hydride were carried out on methylated polysaccharides and oligosaccharides in the following manner. To a 1$ [W/V] suspension of lithium aluminum hydride in T.H.F. was added slowly a 2$ [W/V] solution of methylated carbohydrate in anhydrous T.H.F. The reaction mixture was stirred during the addition and for 16 hours afterwards. Excess hydride was destroyed by cautious addition of water to the vigorously stirred sus-pension. The tetrahydrofuran solution was filtered and the insoluble residue extracted continuously with chloroform. Filtrate and extract were combined and evaporated to dryness. Moving boundary electrophoresis was carried out by Dr. K. Hunt on a Perkin-Elmer electrophoresis apparatus Model 238 using 2 ml Tiselius cells. 93 Samples were prepared for analysis by dialysis of a solution of sample [1-3$ W/V] in a barbituate buffer [veronal] against the same buffer solution. This latter solution was used in the buffer vessels. During the electrophoresis the power output was adjusted to approximately 2 watts. Dialysis of samples was carried out in viscose tubing against running water when only the polymer was desired or against a static volume when col-lecting dialysable materials. Dialysis of DMSO solutions resulted in extensive loses unless water was added to dilute the DMSO before dialysis. The formation of methyl ethers was carried out by one or a combination of two of the following methods. These were: I] Hakamori's method [79] II] Purdie's method [62] III] Kuhn's method [72] A generalized account of these methods as applied to the methylation of poly, oligo and monosaccharides is given below. I] HAKOMORI METHOD [79] The methylation procedure was essentially that described by Hakomori [79] wherein the methylsulfinyl anion [90] was used to generate the polysaccharide alkoxide prior to the addition of methyl iodide. Routinely, 1 g. samples of polysaccharide were methylated. The procedure described is for this amount. The methylsulfinyl anion was prepared as follows. Into a dry round bottom flask [250 ml] containing a magnetic stirring bar was weighed 1.5 g. of sodium hydride [55$> coated with mineral o i l ] . The sodium hydride was washed three times by stirring with 50 ml portions of anhydrous petroleum ether [3O-6O0] and 94 decanting the wash. After the third wash the flask was fitted with a serum cap. The residual petroleum ether was removed by successive evacuations with a vacuum pump through an l8-gauge hypodermic needle inserted through the serum cap. After each evacuation, the flask was f i l l e d with nitrogen [dried by passage through sulfuric acid and sodium hydroxide]. The flask f i l l e d with nitrogen was then placed in a nitrogen f i l l e d drybox. The serum cap was removed and 15 ml of an-hydrous dimethyl sulfoxide was transferred into the flask. DMSO was dried with and distilled from calcium hydride under reduced pressure then stored over dried molecular sieve [Linde, type 4A]. The flask was stoppered with a mercury gas trap and removed from the drybox. The flask was placed in a force draft oven at 60°C and swirled occasionally until the solution became clear and green and the evolution of hydrogen gas had ceased [ca. 60 minutes]. For generation of the polysaccharide alkoxide, polysaccharide was f i r s t freeze dried and then dried overnight at 60° under vacuum. Dried material [1 g.] was added to 50 ml of dried dimethyl sulfoxide in a 250 ml round bottom 2 necked flask containing a magnetic stirrer. The flask was tightly stoppered with a polyethylene stopper and removed from the drybox. The sample was stirred until a l l the polysaccharide had dissolved. Polysaccharide and methylsulfinyl anion solutions were placed in the drybox. A 50$ excess over the number of equivalents of base required for hydroxyl plus carbonyl was added to the polysaccharide. The number of equivalents required was calculated on the basis of the poly-saccharide structure. Upon the addition of the anion [ ~ 12 ml] to the poly-saccharide a gel formed. The flask was stoppered with a serum cap and ther-mometer, before removal from the drybox. The gel, on stirring, liquified and the reaction mixture appeared homogeneous. The reaction was allowed to run for 4 hours, a time indicated as minimum for complete alkoxide formation [78]' 95 In the methylation reaction the polysaccharide alkoxide solution was cooled to less than 20° in an ice bath. Methyl iodide [5 ml] was added to the stirred solution with a hypodermic syringe at such a rate that the temperature did not rise above 25° [ ~ 10 minutes]. The amount of methyl iodide added was not crit i c a l as long as i t was in molar excess of the base. Shortly after addition of methyl iodide, heat evolution ceased, the solution became clear and the viscosity was markedly reduced. At this stage reaction was complete. The reaction mixture was poured into water, then dialysed overnight against running water. The methylated polysaccharide partially precipitated by water came completely out of solution on dialysis. The methylated polymer was washed from the dialysis tubing and dissolved in chloroform. The chloroform solution was dried with anhydrous sodium sulfate and evaporated to dryness. The methyl-ated polysaccharide was fractionated using chloroform and petroleum ether [30-60°] extractions. The major fractions were dried in vacuo and analysed by IR and microanalysis. II] PURDIE METHOD [62] The Purdie method was generally applied only to partially, methylated poly-saccharides. The partially methylated polysaccharide was dissolved in methyl iodide or a methyl iodide/acetone mixture. The solution was vigorously stirred and boiled under reflux in the presence of Drierite. Silver oxide was added in portions over a period of 6-l8 hours. The silver oxide was removed by filtration through a Celite pad and was continuously extracted in a Soxhlet with boiling . chloroform. If acetone was a component of the original mixture the silver oxide was washed with acetone or methanol instead of chloroform. The combined filtrate and washings were concentrated to dryness and the material was examined by means of an IR. -96 III] KUHN METHOD [72] This method was employed with partially methylated polysaccharides. The partially methylated polysaccharide was dissolved in a mixture of redistilled dimethyl formamide and redistilled methyl iodide. The solution was stirred with Drierite. Silver oxide was added in small portions over two hours and stirring was continued for a further 16 hours. The solid material was removed by centrifugation and was washed with portions of dimethyl formamide and chlor-oform. The centrifugate and the washings were poured into a solution of sodium cyanide [1$ W/V] and this solution was extracted with chloroform. Extracts were concentrated to remove chloroform and residual dimethyl formamide was re-moved by continual co-distillation with water. In some experiments Drierite was omitted, in others the methylated polymer was isolated by dialysis. Distillation of dimethyl formamide was intentionally carried out without stringent precautions for the exclusion of moisture. Often traces of silver salts remain in the methylated product after Kuhn and Purdie methylations. These salts were sometimes successfully removed by redissolving the methylated polysaccharide in chloroform and filtering the solution through a Celite pad. METHYLATION OF OLIGO AND MONOSACCHARIDES PURDIE AND KUHN METHODS The methods used were similar to those described for the polysaccharides with the following modifications. During Kuhn methylations a second portion of silver oxide was added after 4 hours of stirring. In addition dialysis was not applicable in these cases. 97 HAKOMORI METHOD The method described for polysaccharides was followed for mono and oligosac-charides with the following exceptions and changes. Some of the lower molecular weight samples could not be freeze dried. As well the dialysis step was of necessity omitted. The samples were isolated by extracting with chloroform the water-dimethyl sulfoxide solution used for dialysis in the polysaccharide case. After three extractions or continuous extraction the chloroform was washed with distilled water to remove dimethyl sulfoxide. The products were evaporated to a syrup and residual dimethyl sulfoxide was removed by high vacuum distillation of a part or a l l of the sample. PERIODATE OXIDATION Analytical periodate oxidations were carried out in the following manner. Duplicate samples of the ash-free polysaccharide [•—50 mg accurately weighed] were dissolved in water and the pH was adjusted to 7- The solutions were transferred to standard flasks [100 ml] fitted with ground glass stoppers and sodium metaperiodate solution [25 ml] was added. This solution was obtained by dilution of a O.JM sodium metaperiodate solution [10 ml diluted to 100 ml]. The polysaccharide solution was made up to 100 ml and stored in the dark at 0° or room temperature. Aliquots [5 nil] were removed at intervals for the estimation of periodate consumption. Duplicate solutions which contained sodium meta-periodate and water but no polysaccharide were made up in the same manner to act as reagent blanks. 98 ESTIMATION OF PERIODATE [91] An aliquot [5 ml] was added to a solution containing a phosphate buffer at pH 7 [25 ml] and 20$ potassium iodide solution [10 ml]. The liberated iodine was titrated against standard sodium thio sulfate [COIN] using starch as the indicator. The phosphate buffer contained disodium hydrogen phosphate [5*68 g.] and potassium dihydrogen phosphate [3*62 g.] in one l i t e r of distilled water. SIMULTANEOUS ESTIMATION OF POLYHYDRIC ALCOHOLS AND SUGARS The experimental work outlined here was part of a group project involving Reid, Jensen and Gibney [92]. The estimation of erythritol with threitol, the computer analysis of these results and examinations of the effects of glycolaldehyde were mainly the work of Jensen. Columns employing the following liquid phases were evaluated for their potential use in the separation of galactitol, glucitol and mannitol acetates: Apiezon J, K, L, M, N Butane 1,4-diol Succinate Carbowax 2CM FAPP Neopentyl Glycol Sebacate Silicone DC QF-1 Silicone GE SE30, 52, SF96, XE-60 Versamide 900 None of these columns would produce the required separation. The literature has since reported two systems which w i l l achieve this separation. However, the columns when used on thermal conductivity have resulted in only limited 99 success. If large samples are injected the columns overload and the separation is lost [93]. A certain advantage is apparent when galactose is a minor com-ponent of the mixture. Columns were examined for the separation of equilibrium mixtures of neutral sugars using the silylation technique of Sweeley et a l [15], The columns ex-amined were: Silicone GE SF-96 Silicone GE SE-30 Silicone GE SE-52 Dow Corning Silicon High Vac Grease The column having the clearest resolution was that reported in this work [SF -96], It has been shown in the discussion that a scheme can be devised where, knowing the equilibrium values of the components, the total analysis of a sample of the five common neutral sugars may be evaluated. The column length, the programme rate and the flow rate were chosen experimentally to give optimum splitting between the two tetritols, erythritol and threitol. Samples of the following polyhydric alcohols were purified by vacuum dis-tillation; ethylene glycol, glycerol and butane - 1 , 4-diol. D-Threitol. and erythritol were recrystallized to literature melting points. The five neutral sugars were recrystallized and analysed by paper chromatography and gas liquid chromatography for impurities. Pyridine refluxed over sodium hydroxide pellets was distilled. A narrow.boiling range fraction was stored over sodium hydroxide pellets. Hexamethyldisilazane and trimethylchlorosiline were used as obtained from Penninsular Chem Research. Analysis of mixtures of internal standard and each of the five neutral sugars was carried out in the following manner. The sugars were dissolved in water containing 0-5$ W/V chloroform and allowed to equilibrate. A period of 100 2k hours appeared to be ample with no further change being detectable in the equilibrium ratio determined. The sample evaporation to dryness was c r i t i c a l to avoid any change in equilibrium. The samples were evaporated under vacuum on a rotary film evaporator until the sample just became dry, a further minute of evaporation was allowed and then the sample was removed. Immediately a sample of pyridine containing the internal standard was pipetted into the evaporation flask. The samples were swirled and hexamethydisilazane then t r i -methylchlorosilane were added. Any variation from this procedure will result in alteration of the equilibrium mixture. This was especially pronounced when a component which crystallized easily such as xylose was predominant. The equilibrium ratios found for pyridine are similar to those for water [22] resulting in l i t t l e change in the equilibrium mixture i f these conditions are rigidly followed. Analysis of single compounds and internal standard butane -1,4-diol yielded data reported in the discussion for molar response factors and equilibrium con-centrations for various anomers. Sample size was varied with respect to in-ternal standard over a wide range to confirm that detector response was linear with sample size. A l l other parameters were held constant to minimize variation in molar response factors. Sufficient data were evaluated to provide molar response factors and equilibrium anomer ratios [at least three samples were run in duplicate for peaks of the polyols, five samples were run in duplicate for sugars]. Samples were then prepared of the various synthetic mixtures [expected to result from periodate oxidations,of plant polysaccharides]. Samples were analysed with three runs in each sample determination. The data have been presented in the discussion. 101 Data were also obtained using erythritol as the internal standard and employing the shorter programme from 190-220° as a method for the analysis of crude hemicellulose hydrolysates as reported in the discussion. Accumulated experimental data indicate that for sugar analysis the following limiting percentages may be determined. . Xylose 0.5$ Arabinose 0.5$ Mannose 0.5$ Galactose 1.0$ Glucose 0.5$ The value expressed is normally a practical limit in general hydrolyses. For a clearly detected sugar i t may be considerably lower, in the order of 0.01$. Galactose is notably higher because only a single peak containing 30$ of the total peak area is visible in hemicellulose hydrolysates. The final check on this method was the analysis of three samples prepared independently and analysed by the proposed method. The results were in good agreement with expected values. PURIFICATION OF LEMON GUM Lemon gum*[crude] was dissolved in water and filtered through several layers of gauze followed by sintered glass. The solution was poured in a thin stream into a five volume excess of alcohol acidified with acetic acid. This procedure, however, left the gum with a high ash so the gum was redissolved and purified by two alternate routes which yielded products which were identical. A solution of gum was acidified with hydrochloric acid * The sample of lemon gum used in these experiments was received from the late Dr. E. Anderson. 1G2 and precipitated into 5 volumes of alcohol. This product showed only a trace of ash. A second portion of gum solution was deionized by passage through Amberlite IR 120 [H+]. The solutions were immediately frozen and freeze dried. Optical rotation of the products gave identical values of [ Oi ]^  +19.2°. The polymers were hydrolysed and analysed by G.L.C, for results see discussion. A sample of the freeze dried gum was analysed by Tiselius moving boundary electrophoresis and showed only a single band. The gum was also chromato-graphed on DEAE cellulose [2.5 x 50 cm] using water [150 mis] followed by 0.05M NaH2P04 [150 mis] increasing by equal steps to O.25M NaH2P04. The gum was eluted during 0.10M NaH2P04 irrigation. The gum tailed somewhat on the column and as the ionic strength was increased more carbohydrate was eluted from the column. The bulk of the gum was isolated in a volume of 50 ml, tubes 72 to 82. G.L.C analysis indicated that the product isolated had the same neutral sugar analysis as the starting material and an identical optical rotation of [ OC ] +19.3°• The neutralization equivalent of the ash free gum was determined by t i t r -ation with 0.100N NaOH using a pH meter. The equivalent weight was calculated to be 793 g-OPTIMUM HYDROLYSIS OF LEMON GUM - PERIODATE DEGRADED Analytical periodate oxidation of lemon gum produced results in close agreement with those reported earlier for this gum [ 1 ]. Lemon gum [20 g.] dissolved in water [1000 mis], was oxidized with sodium metaperiodate [29 g.] for one week at 5°C. Dialysis was used to removed low molecular weight materials after destruction of excess periodate with ethylene 103 glycol. The polyaldehyde present in the dialysate was reduced with sodium borohydride [5 &•] for 48 hours. An excess of acetic acid was added to destroy unreacted borohydride and aid in break down of borate complexes. Dialysis was again used to remove low molecular weight materials, yield 12.5 g-Samples of the polyalcohol [20 mg] were dissolved in water [5 ml] and the acid concentration was adjusted so that the final volume [10 ml] had the f o l -lowing sulfuric acid concentrations; 0.05, 0 . 1 , 0 .2 , 0.5 and 1.0N. Hydrolysis was continued for 24 hours at room temperature. The acid was neutralized with barium carbonate. The residual polysaccharide was isolated by precipitation and analysed on the gas chromatograph. The 0.5N acid solution, as indicated in the discussion, produced the most satisfactory hydrolysis, as determined by residual glycerol [see page 22 ]. A sample of polyalcohol [200 mg] was dissolved in water and the normality of the acid and volume were adjusted to 0.5N and 20 mis. Samples [2 ml] were removed from the reaction mixture at the following intervals: 1, 2, k, 8, 16 and 32 hours. Precipitation of the sample taken was carried out by direct dilution with alcohol [8 ml]. Immediate centrifugation and work up minimized further hydrolysis. In addition to the samples taken above, a sample of precipitated polysac-charide [32 hours] was dialysed for 2k hours against running water. Neutralization of a sample of supernatant from the 32 hour hydrolysis precipitation provided a fraction containing material lost during hydrolysis. Gas chromatographic analysis was carried out on hydrolysates of fractions isolated including a sample of starting material and a sample of unhydrolysed 104 soluble material at 32 hours. The data from these hydrolyses are presented in Table X [see page 26]. Lemon gum polyalcohol [9'17 g«] was hydrolysed for the required l 6 hours •with 0.5N H2SO4 [400 ml]. The solution -was neutralized •with barium carbonate. Yield of precipitated polysaccharide designated lemon gum SI on addition of k volumes of alcohol, 3*51 g- The supernatant fraction was evaporated to a syrup under reduced pressure. To this thick syrup was added 30 mis of methanol and 30 n^ ls of acetone [2 times] to extract the very low molecular weight fraction. The fractions were centrifuged to remove the insoluble higher mole-cular weight material. On evaporation of the 120 mis of solution 3*89 g« of product was obtained. The insoluble material was dissolved in water and freeze dried, yield 1.06 g. Lemon gum SI [100 mg] was oxidized with an excess of periodate until the oxidation levelled off [88 hours]. Periodate consumption was 0.40 moles per anhydro sugar based on galactose and arabinose analysis. A larger scale oxidation of 2.30 g. of polysaccharide yielded 1.22 g. of poly alcohol. Smith degradations carried out as before indicated that an optimum yield of polysaccharide was obtained using the same hydrolysis conditions as outlined for the fi r s t periodate oxidized polysaccharide, namely 0.5N H2SO4 for l 6 hours. Hydrolysis under these conditions of 2XIO4/BH4 lemon gum yielded a polysaccharide containing only a limited amount of arabinose in comparison to galactose as outlined in the discussion. AW? SMITH FRAGMENTS Direct gas chromatography of the soluble portions from the original periodate oxidized gum programmed to elevated temperature on SF-96 and SE-52 [10$ on 80-100 Diatoport S, 8' x l / V and SE-30 [20$ on 60-80 Chromasorb W, 2' x l/4"] showed only one component in sufficient quantity to isolate and identify. The column temperature [275°C] was such that maltotriose was eluted [SE-52 - 80 min., SE-30 - 30 min.] [94]. Isolation of a small sample and hydrolysis yielded galactose and glycerol. A second sample was isolated, the s i l y l groups were removed by methanol/water and the sample was methylated by the Hakomori procedure. Hydrolysis, reduction and acetylation produced a partially methylated alditol acetate with the same mobility as in authentic sample of 2,3,4 ,6-tetra -0- methyl-D-galactitol acetate. Analysis of the mass spectrum confirmed this assignment. THE REDUCTION OF ACIDIC POLYSACCHARIDES ESTERIFICATION Lemon gum [1.00 g.] which had been deionized by passage through Amberlite . IR-120 [H+-form] and freeze dried was dissolved in water [50 mis]. Ethylene oxide [12.5 g-] was bubbled into the polysaccharide solution cooled in an ice bath. The vessel was sealed and stored at room temperature. During the next fourteen days the sample was cooled daily in an ice bath and the pH of the solution was measured. The solution pH rose from an i n i t i a l value of 3-2 "to slightly over 7 at the end of the fourteenth day. The gum solution was di-alysed against running water for twenty-four hours and freeze dried, yield 0.93 B-106 Analysis of the degree of esterification was carried out using acid hy-drolysis with IN H 2S0 4 at 100°C for 20 hours. The quantity of ethylene glycol and what is presumed to be diethylene glycol were in excess of the molar re-quirements for complete esterification. REDUCTION OF ETHYLENE GLYCOL ESTER OF LEMON GUM Lemon gum ester [1.00 g.] was dissolved in dry dimethyl sulfoxide [dried as for Hakomori methylations] [50 ml.]. To this solution was added lithium borohydride [0.5 g « 3 - Such.a large molar quantity was required for interaction with free hydroxyl functions in the polysaccharide. The solution formed a gel during the f i r s t hours of reduction. The gel was relatively stable and persisted throughout the reduction. The gel could be broken down partly by swirling the mixture. The reaction was maintained anhydrous and hydrogen gas was vented through a mercury trap. The reduction was monitored by daily analysis of aliquots. A l l samples were worked up in the same manner. Acetic acid was added slowly to the DMSO solution to destroy the excess borohydride and to break down borate complexes. The samples were dialysed then freeze dried, yield O.83 g. Material [1.00 g.] from a large scale esterification and reduction was subjected to a second esterification to re-esterify any acids which were sapon-ified. Salts of uronic acids were converted to free acids by ion exchange chromatography as in the previous reduction. Quantities of ethylene oxide and duration of reaction were as before, yield O.96 g. This freeze dried ester was reduced by the method indicated previously, yield 0.90 g. 107 The samples were analysed by gas chromatography and the results are reported In the discussion. SIGNAL ENHANCEMENT Compounds which were examined by silylation of free hydroxyls [Figure 16] were synthesised by known pathways or purified by preparative gas chromato-graphy as reported in the discussion. Pure starting materials were silylated and purified by both gas chromatography and simple removal of silylating reagents by evaporation under reduced pressure. TRIMETHYL SILYLATION OF GALACTURONIC ACID A commercial sample of D-galacturonic acid was silylated using each of the following solvents: pyridine, dimethyl formamide, dimethyl sulfoxide, hexane and dioxane, by warming and triturating the sample in the solvent with added bis-trimethylsilyl acetamide [BTMSA]. There were significant variations in the percentages of sample silylated. Those solvents which dissolved the galacturonic acid showed better yields of silylated product. In addition, silylation was carried out in pure BTMSA. The percentage of anomers present when these solutions were analysed on the gas chroma,tograph varied widely. The best general yield arid the most reproducible result was obtained with pure BTMSA. 108 TABLE XX ANOMERIC RATIOS FOR BTMSA SILYLATION OF GALACTURONIC ACID IN DIFFERENT SOLVENTS Solvent Peak 1 < Peak 2 and 3 < fi Peak 4 $ BTMSA 49 -5. 24.3 26.1 DMF 52-1 18.1 28.8 Dioxane 52.5 43.3 4.2 Pyridine 18.8 33.4 47.8 DMSO 36.2 33.6 30.2 Hexane 45.O 30-9 24.1 METHYLATION ANALYSIS Lemon gum [1 g.] was methylated by the Hakomori procedure described earlier, yield 0.90 g. A second methylation of this methylated material [0.70 g.] was carried out and the DMSO solution -was poured directly into a dialysis sack without dilution by water. The weight of recovered lemon gum was only 0.020 g. A second methylation [5-0 g.] was carried out, yield 6.198 g. This sample was fractionated using chloroform in 30-45° petroleum. The extracted fraction insoluble in 20$ CHCI3 but soluble in 30$ CHCI3 contained 6.011 g. of the total sample. Analysis: OMe 39»46$. A small sample [1.0 g.] of methylated lemon gum was acidified with anhydrous hydrogen chloride in chloroform. A small amount of solid material [NaCl] was centrifuged off and the polysaccharide was pre-cipitated with 30-45° petroleum. The polysaccharide was dried under high vacuum over sodium hydroxide. This methylated polysaccharide was given a Purdie methyl-ation in order to complete the methylation of the uronic acids. Analysis: OMe .40. 109 This polymer, although s t i l l showing traces of uronic acid as the salt form-, was reduced with lithium aluminum hydride to give a product showing only a trace of carboxyl, yield 90$. The following polymers [100 mg] were also methylated by the Hakomori procedure: a] Lemon gum degraded with l/lON H2SO4, yield 105 mg. b] Lemon gum periodate .oxidized and borohydride reduced before hydrolysis, yield 82 mg. c] Lemon gum periodate oxidized and borohydride reduced after O.5M H2SO4 hydrolysis, yield, 95 ™S« None of these polymers was reduced with lithium aluminum hydride. A l l four of these methylated polymers were analysed by the method outlined in the introduction to the experimental section. The following monosaccharides were prepared by known routes to facilitate the peak identification on the gas chromatograph: 2,3^4,6 -tetra-0-methyl-D-galactose 2,3>4-tri-0-methyl-D-rhamnose 2, J>, 5 -tri-Q-methyl-L-arabinose 2,3 - d i-0-me t by 1-D - glu co s e 2,3>4-tri-0-methyl-L-arabinose 110 APPENDIX I ESTIMATION OF DEGREE OF POLYMERIZATION During the course of evaluating G.L.C. as a method for the estimation of polyhydric alcohols and sugars, i t became apparent that these procedures could be applied to an estimation of D.P. of oligosaccharides and small polyaschar-ides. Reduction of the reducing end group -with sodium borohydride produces on hydrolysis the terminal component which may be estimated with respect to the remainder of the non reduced carbohydrate. Work by Reid, Rowe, Rowe and Dutton [95] has shown that a D.P. of 150 may be estimated in this manner. Previous experience with hydrolysis of solutions containing polyols in-dicated a need to examine closely the hydrolysis of reduced oligosaccharides. Certain factors must be taken into consideration before D.P. may be estimated by the method of Reid et a l . The compound[s] being reduced must be known. Also within limits complete reduction must occur. The complexity of the sugar mix-ture on hydrolysis must be such that the polyol produced on reduction is clearly resolved on G.L.C. from the remaining anomeric carbohydrate forms. Only with very low D.P. oligosaccharides can analysis be successfully carried out when overlap occurs. It is known that sugars decompose during the hydrolysis re-action. Previous work in this laboratory has indicated that dehydration of polyols to anhydropolyols can also occur [85 ]. A series of maltosaccharides were obtained by enzymatic hydrolysis of starch with cX-a.mylase. Isolation of the higher D.P. maltosaccharides [D.P.> 4] in a pure state was not possible although paper and Sephadex chromatography were employed in the attempted purifications. Prior to examination of the re-duced maltosaccharides the effects of hydrolysis on glucitol were investigated. I l l Prolonged acid hydrolysis confirmed thatdianhydro-D-glucitol was formed as the major dehydration product. This was in agreement with previous hydrolysis findings [85]. The more volatile component in the gas chromatogram was con-firmed to be 1 ,4 ,3,6 dianhydro-D-glucitol by mass spectrometry and comparison-with an authentic sample. FIGURE 30 GAS CHROMATOGRAM OF GLUCITOL HYDROLYSED WITH 1.0N H2S04 FOR 24 HOURS Standard analysis conditions but programmed at 10°/rnin. 20 30 4 0 min. During these hydrolysis studies i t was discovered and has now been reported independently by another Laboratory [58], that undersilylation can occur. The undersilylated product is eluted before the fully silylated derivative. The silylation of hexitols is much slower than for lower molecular weight polyols or the parent sugars. Glucitol in the presence of excess reagent requires two hours of reaction before the under silylated component becomes insignificant. Further work is being carried out to confirm the structure of these under silylated hexitols. 112 FIGURE 31 GAS CHROMATOGRAM OF GLYCITOL SHOWING UNDERSILYLATION AFTER 10 MIN. OF REACTION In order to obtain the maximum D.P. accuracy using reduction, hydrolysis and G.L.C, careful consideration must be taken of carbohydrate degradation and loss of polyol by dehydration. EXPERIMENTAL A series of glucose oligosaccharides was isolated by enzymatic hydrolysis .of starch [95]. I n i t i a l purification was achieved using a stubby charcoal column [10 cm in diameter x 2.5 cm deep]. The glucose and maltose were displaced from the column by gradually increasing the alcohol to water ratio in the elution solvent. The alcohol percentage was increased to 50$ and the higher D.P. oligosaccharides -were eluted as a group. These higher oligosaccharides were spotted on Whatman No. 2 f i l t e r paper and chromatographed in solvent C. 113 Components were detected by development of guide strips. The bands of oligo-saccharides were eluted with water and the sample was isolated in an easily handled form by freeze drying. Samples of the major components were rechromato-graphed on paper and Sephadex. Only in the case of the t r i - and tetrasaccharide • were relatively pure compounds isolated. Gas chromatographic analysis of the reduced oligosaccharides did not produce accurate [+ 5$] D»P» values because of trace quantities of higher oligosaccharides present as contaminants. Samples of glucitol were hydrolysed and analysed by gas chromatography. The presence of l,k; 3,6-dianhydro-D-glucitol was confirmed by comparison of mass spectra obtained from an authentic sample and one isolated by gas chromato-graphy. 114 APPENDIX II ENZYMATIC HYDROLYSIS OF LEMON GUM Enzymatic hydrolysis has been employed successfully in the glycolytic cleavage of certain classes of polymers. Notable is the extensive work which enzymes has been discovered and utilized in preparing the maltosaccharides for the earlier work on D.P. and amylo 1,6-glucosidase, the enzyme which debranches 1,6 linkages. The enzymatic hydrolysis of cellulose and some of its derivatives is well known. The fact that ruminants convert cellulose containing plants into useful food sources, microorganisms destroy cellulosic fabrics, &n<3. "the host of wood rots decompose wood, has also focussed attention on nature's abun-dant cellulase activity. The food industry employs pectinase preparations for clarifying wines, fruit juices and liquefaction of pectin containing gels. Xylans from wood and plant sources have been hydrolysed with pectinase and xylanase preparations to yield a product containing xylose and a series of oligosaccharides [ 9 8 ]• As a result of the success encountered with these enzymatic degradations an examination was carried out on the effects of various general enzyme preparations on lemon gum. Table XXE lists the enzyme prepara-tions employed in this survey. has been carried out on various starches [97]« -A wide variety of different TABLE XXI ENZYME PREPARATIONS EXAMINED FOR ACTIVITY ON LEMON GUM d -Amylase [salivary] Pectinol 41-P concentrate [Rohm and Haas] Pectinase [BDH] Hemicellulase [BDH] HP-150 [Rohm and Haas] Cellulase [BDH] Cellulase 36 [Rohm and Haas] A -Glucosidate [BDH] Xylanase [BDH] BDH - British Drug House 115 Although some of the preparations listed in Table XXI -were not expected to show activity they were examined because of possible activity of minor impurities. Two of the preparations [Hemicellulase and HP-150] were expected to show the highest activity. This was especially true of the latter material since i t is designed to reduce viscosity in plant gum solutions. Assay of the o< -amylases, pectinases, jQ -glucosidase and xylanase with lemon gum showed no activity. The enzymes present in these preparations were active as was evidenced by formation of maltosaccharides and a series of xylose oligossacharides from birch xylan [98]. The remaining enzymes could not be assayed by dialysis [99] due to high cellulase activity.' The dialysis tubing was attacked, resulting in release of glucose oligosaccharides and rupture of /the tubing. Activity^of these preparations on lemon gum was examined after the substrates had been in contact for two days. The enzyme was inactivated by a short heat treatment. There were several minor low molecular weight com-ponents in the solutions. They were, however, attributable to contamination of the enzyme preparation or autohydrolysis of the lemon gum substrate. Lack of activity with these latter preparations was unexpected. A sample of HP-150 was fractionated with ammonium sulfate into a series of precipitates. Assayed against lemon gum these fractions were s t i l l not active. Fractionation was also carried out on Pentinol 4 IP. Ko activity was found in the isolated fractions. The only oligosaccharides present on dialysis of fractionated HP-150 or Pectinol 4lP treated lemon gum were attributable to autohydrolysis. . This fact was verified by running blanks containing only lemon gum and distilled water. During these experiments a parallel set was being conducted on mesquite gum by Rowe [100]. Equally negative results were obtained. In addition, i t was 116 discovered that HP-150 had transferase activity capable of forming arabinose oligosaccharides. Since, on hydrolysis of either lemon or me3quite gums large amounts of arabinose are expected, any oligosaccharide isolated -which contained arabinose would be suspect. The lack of activity against lemon gum and the presence of transferase activity would appear to rule out convenient hydrolysis of lemon gum. When the role of the transferase activity is more fully known [100] i t may be possible to utilize HP-150 for the hydrolysis of oligosaccharides obtained by hydrolysis of lemon gum. Careful examination wil l be necessary to ensure no other transferase activity [for example galactose transferase] exists. The closing discussion is speculation regarding low activity against a polymer such as lemon gum. Lemon gum is a highly branched and extremely com-plicated molecule containing many different subunits. The gum is exuded on damage to the tree bark. To be effective as a barrier the exudate must not be readily attacked by organisms. In an evolutionary sense those plants which exuded a barrier which was difficult to degrade would have had the best chance of survival. The inability of any of these enzyme preparations to degrade lemon gum may be a reflection of this evolutionary protection. An interesting experi-ment would be the examination of micro-organism growth on media containing lemon gum. When samples of gum have become contaminated in our laboratory, after a short growth period further growth seems to stop. This growth stoppage may reflect a lack of nutrients such as nitrogen or complete utilization of the carbohydrate available as a result of autohydrolysis. It is possible also, that a pH change or formation of a metabolite may be factors which are limiting growth. Experiments designed to test these factors may indicate whether there is limiting growth rate or ability to utilize lemon gum as a carbon source. 117 EXPERIMENTAL ENZYMATIC HYDROLYSIS OF LEMON CUM The enzyme preparations listed in Table XXI were screened for activity on lemon gum by the following procedure. A small sample of ash free gum [ ^ 1 . 0 g.] was dissolved in water [50 ml] and a sample [5 ml] was pipetted into solutions of the various enzyme preparations. The samples were allowed to interact for KQ hours, then excess enzyme and polysaccharide were precip-itated with alcohol. The reactions were carried out in centrifuge tubes [50 ml] to facilitate work up. A second solution of lemon gum was adjusted to pH 7 with sodium bicarbonate and a similar series of analyses repeated. Paper chromatograms were spotted with samples from the concentrated super-natants of the two experiments, with a blank from a pure gum solution and a blank of each enzyme. The chromatograms showed no spots which could not be attributed to either autohydrolysis of the lemon gum or contaminating carbo-hydrate from the enzyme preparations. Samples of the enzymes and lemon gum were also examined by dialysis [99L since i t was possible that a certain concentration of oligosaccharide could inhibit further enzymatic hydrolysis. The results as before were negative except with lemon gum acid which showed autohydrolysis products. Four enzyme preparations were not amenable to dialysis as a means of oligosaccharide iso-lation. Hemicellulase, HP-150 and the two cellulase preparations show such high cellulase activity that the viscose tubing, even though forming a hetero-geneous reaction'site, was weakened-to the point that i t ruptured. In addition new compounds formed by enzymatic hydrolysis of viscose tubing were released into the solution. 118 These four preparations were examined by a second dilute reaction solution and the enzyme was inactivated by heating. Alcohol precipitation and concen-tration of the whole solution were both used to prepare samples for paper chromatographic examination. The results for the dialysis and the latter experiments were as before negative showing no increase in oligosaccharides which could be attributed to enzymatic hydrolysis. FRACTIONATION OF PECTINASE, CELLULASE AND HP-150 Pectinol 41-P concentrate, Cellulase and HP-150 were a l l fractionated using ammonium sulfate by the following general scheme. Enzyme preparation [10 g.] was dissolved in water [200 ml] containing sodium bicarbonate [4 g.]. Additions of ammonium sulfate were made to the solution in approximately 10 g. lots. When a precipitate formed i t was centrifuged off and further additions of ammonium sulfate were made. No attempt was made to determine quantitative data on various fractions. Rather a general qualitative result was determined. PECTINOL 41-P CONCENTRATE A significant portion of enzyme preparation was insoluble in water because of f i l l e r added to the preparation. The prepared enzyme fractions were reacted with dilute solutions [1-2$] of lemon gum. A l l samples gave results similar to an autohydrolysed gum sample. PECTINOL 41-P FRACTIONATION 119 Ammonium Sulfate [g.] Molish on Enzyme Samples Retaining Molish Color Amount ofPpt 0 + — large ho ++ ++++ small 50 ++ — small 60 ++ — fair 70 ++ ++ fair 80 ++ — fair 90 ++++ ++++ fair 100 ++++ +++ fair 110 +++++ +++++ fair Final Solution +++ CELLULASE 36 A slightly cruder fractionation was obtained on this enzyme. Ammonium Sulfate [g.] Amount of Ppt Molish Test 0 large [ f i l l e r ] + 60 small +++ 120 fair +10 120 [over night] small ++++ Supernatant ++ Examination of activity against lemon gum was similar to that for Pectinol l+l-P and showed no specific increase in components isolated over autohydrolysis. 120 HP-150 A similar crude fractionation was obtained for this enzyme preparation and no activity increase was noted in any of the five fractions examined. Extensive purifications of this preparation were carried out by Rowe [100] but because of the transferase activity noted, work was not carried further. 1 2 1 APPENDIX III CARBON-CARBON BOND CLEAVAGE, DIALYSIS AND DESTRUCTIVE PURIFICATION AN UNDERGRADUATE ORGANIC EXPERIMENT Carbohydrates are among the most abundant natural products. They are widely distributed in both the plant and. animal kingdoms. Their variety is vast, ranging from cellulose and starch to glycogen, deoxyribonucleic acids, streptomycin and blood group glycoproteins. Carbohydrates have unique and varied biological activity, ranging from stores of potential energy in animals to supporting tissues in plants. Bacterial polysaccharides frequently carry the dominant immunological character of the bacterial cell and, in the case of pathogenic species, may form essential ingredients of vaccines. Experiments in undergraduate laboratories with carbohydrate compounds have been very limited-. Recently a procedure for conversion of polymeric glucose into erythritol was developed.. Since this carbohydrate experiment involved a variety of techniques and reactions not often utilized in undergraduate lab-oritories i t was felt to be of possible general interest. This experiment, as described below, can be used to illustrate hetero-geneous and hemogeneous reactions, oxidative carbon-carbon bond cleavage, di-alysis, reduction, destructive purification, decolorization, crystallization, melting point and chromatography. This experiment allows for a development of varied pathways to the final product. Through the use of prepared samples the duration of the experiment may be shortened from one involving several lab periods to only two lab sessions. The quantity used in the oxidations may be large or sufficient for only chromato-graphic identification of the final product. Table XXII below li s t s the various reagents and purification steps involved in the experiment. Not a l l pathways are possible,therefore, i f desired a student may be allowed to determine a feasible pathway. 122 TABLE XXII VARIABLE STEPS IN PREPARATION OF ERYTHRITOL Polymer Oxidant Purification Reduction Purification Isolation Cellulose NaI04 Dialysis NaBH4 Dialysis Precipitation Soluble starch Insoluble starch HT04 Filtration Filtration Freeze drying Not necessary a] filtered off Glycogen b] next reaction on solution Hydrolysis Neutralization Decolorization Characterization Time and acid strength may be varied Ion Exchange Ba[OH] - BaC03 Short charcoal column Repeated Crystallizations Melting point -mixed melting point Chroma to graphy - gas liquid - paper - thin layer As a result of the -wide variety of pathways different points may be i l l u s -trated with the experiment. The carbon-carbon bond cleavage may be carried out as a homogeneous or heterogeneous reaction depending on the solubility of the • glucose polymer. Depending on the homogeneity of the reaction, products from the oxidation and reduction may be purified by dialysis or filtration. As an alternate, soluble polymers may be purified by precipitation i f periodic acid is the oxidant. Barium hydroxide w i l l precipitate iodate ions remaining after the oxidation. 123 The following experimental procedure outlines the necessary details to convert one of the polymers into erythritol. The quantities and time may be altered by the use of prepared samples and method of characterization. Where prepared samples are used students should carry out the reactions, but pool their sample at each stage in order to supply subsequent laboratory periods with prepared samples. Alternate reaction pathways differ from the experiment below as noted. Starch [l6.2 g.] was dissolved by heating in distilled water [400 ml]. Sodium metaperiodate [23.4 g.] was added to the cooled solution. The reaction was allowed to stand in the dark for 2k hours [Note 1] at room temperature. At the end of this time the reaction was stopped by the addition of ethylene glycol. The solution was poured into a piece of cellulose tubing double knotted at one end. The air was expelled from the tubing and the top end was double knotted. The tubing was completely immersed in running water for 2k hours [Note 2]. Sodium borohydride [4.0 g.] [Note 3] was added to the dialysate and after 2k hours the excess reducing agent was destroyed by the addition of a small amount of acetic acid [5-0 ml]. Dialysis was employed to remove the sodium borate [Note k] and the solution was concentrated to a small volume [approx-imately 100 ml]. The solution was made 2N with sulfuric acid and re fluxed for 3 hours. The acid was neutralized with slightly less than an equivalent amount of barium hydroxide. The slight excess of acid was neutralized with a small amount of barium carbonate. The precipitate was removed by suction filtration through a pad of diatomaceous earth. The filtrate was concentrated on a rotary evaporator to dryness. The resulting syrup was dissolved in boiling methanol and the solution filtered to remove traces of. inorganic salts. The solution was 124 diluted with an equal volume of water and passed through a stubby charcoal column [101] [Note 5']. A wash solution of the same solvents removed small amounts of erythritol from the column. The column eluate was concentrated to dryness and the clear syrup was dissolved in a small amount of methanol. On cooling erythritol crystallized from-the solution. Yield 5.Q g' m.p.t. 121.5. NOTE 1: ' Reaction rates with soluble or insoluble polymers may be reduced to 3 hours [102], NOTE 2: Insoluble polymers when oxidized may be purified by filtration rather than dialysis. Washes with 50 ml. of water and intermediate filtration are necessary until essentially free of iodate, usually involving about 6 washes. NOTE 3 : It has been recommended that a l 6 fold excess of borohydride. be used when reducing oxycelluloses [103]. For this reaction however, a small amount of unreduced product w i l l not effect the results. It is recom-mended as well that insoluble polymers be shaken or stirred during reduction to prevent material being li f t e d out of the reduction solution. NOTE 4: Insoluble polyalcohols are filtered out, washed with water, steeped for 30 minutes in 10$ acetic acid, and washed several more times with water. NOTE Purification may be achieved by repeated crystallizations but the colored matter produced during hydrolysis is difficult to remove. 125 CHARACTERIZATION The crystalline erythritol may be characterized by melting point and mixed melting point. The effectiveness of the oxidation may be investigated by chrom-atographic means. Examination of the hydrolysis solution wi l l show up to three components; glucose from unoxidized polymer, erythritol,and from terminal glucose residues, glycerol. The amount of glycerol present is an indication of the degree of branching in the polymer [953* The presence of these three components may be.shown on paper [10^3 or thin layer chromatograms [1053* Gas liquid chromato-graphy can be employed to obtain quantitative results regarding the effectiveness of the oxidation and the degree of branching of the starting polymer [95]. 126 APPENDIX IV LEMON GUM: DEGRADATION WITH O.IN SULFURIC ACID Further proof of the galactan branch on branch structure of lemon gum may be seen on analysis of the residual polysaccharide isolated after hydro-lysis Of lemon gum with O.IN sulfuric acid. The residual polysaccharide con-tains essentially ho L-arabinose units. This polymer contained 98.2 mole percent galactose as neutral sugar. This material was further analysed by methylation as reported earlier. EXPERIMENTAL Lemon gum [20 g.] was dissolved in water. The gum solution was adjusted to the desired O.UT sulfuric acid strength and a final volume of 250 mis. The solution was refluxed for 12 hours. The residual galactan was obtained on precipitation of the neutralized solution with k volumes [1000 ml] of alcohol, • yield 8.20 g. Analysis of the galactan showed only a trace of L-arabinose [D-galactose 98.2 mole # ] . ' ' ' si AUT0HYDR0LYSIS OF LEMON GUM Lemon gum contains a large amount of uronic acid [approximately one residue in five]. Free acids were obtained when the gum was acidified or deionized with ionexchange resin prior to precipitation. The pH of the gum acid is nearly three and when a solution is heated extensive hydrolysis occurs. Oligosaccharides were isolated from autohydrolysis and from 0.1N sulfuric acid hydrolysis. The components from both hydrolysis procedures were identical in their chromatographic behaviour and composition. Close comparison of the oligosaccharides isolated 127 by these tvio procedures indicated they were identical chromatographically in several solvents. Table XXIII is a l i s t of the oligosaccharides Isolated. Included in this.table are two neutral D-galactose containing oligosaccharides obtained when the 0.1N sulfuric acid residual polysaccharide was subjected to stronger acid hydrolysis. A l l oligosaccharides had been isolated previously [ l and ref. therein]. Identification was achieved by comparison with known mobility on paper chromatograms. The pure oligosaccharides were isolated from paper chromatograms and hydrolysed to indicate the component sugars. Oligo-saccharides containing uronic acids were reduced in DMSO with lithium boro-hydride prior to hydrolysis. TABLE XXIII OLIGOSACCHARIDES ISOLATED ON HYDROLYSIS OF LEMON GUM. 1] 3- 0-yS -D-galactopyranosyl-D-galactose 6-0-/3 -D-galactopyranosyl-D-galactose 4- Q-[4-O-methyl- o< -D-glucopyranosyluronic acid]-L-arabinose 2] 3] 4] 0-[4-0-methyl- oc-D-glucopyranosyluronic acid]-[l —> 4 ] -0 -& ? -L-arabinopyranosyl-[1—>5]-L-arabino s e k_o-[4-0-methyl- c<-D-glucopyranosyluronic acid]-D-galactose 6-0-[ fS> -D-glucopyranosyluronic acid]-D-galatose 6] 128 EXPERIMENTAL The pH of a lemon gum solution [10 g. in 80 ml of water] was measured as 3«17 for a sample of Amberlite IR-120 deionized and freeze dried gum. Deion-ized lemon gum [l6 g.] was dissolved in water [200 ml]. The solution was placed in a dialysis tube previously inserted in a column [ 9 9 ] fitted with a side arm to allow intake of liquid and a bottom outlet to remove the d i -alysate silution as i t passed by the bag. The unit was wound with 8 feet of heating tape and the temperature was maintained at 70-8o°C [measured on a thermometer inserted between the heating tape and the column]. After one week of dialysis the solution in the tube had expanded sufficiently to prevent liquid from passing around the dialysis sack. A small amount of liquid was removed from the dialysis sack and dialysis was continued for a further ten days. The polysaccharide removed earlier and this final solution were pre-cipitated into alcohol and a total of 11.90 g. of polysaccharide was recovered. The collection of solutions from the autohydrolysis had been made in four separate fractions. Paper chromatography indicated that a l l four fractions had a similar composition. Fractions 2, 3 and 4 were combined and fraction 1 was further divided into an acid and a neutral fraction by absorption on Duolite A-4 [OH] and subsequent elution with 10$ acetic acid. 1 OLIGOSACCHARIDES Oligosaccharides isolated from the 0.1N sulfuric acid and the autohydrolysis were chromatographed on paper and on inspection proved to be virtually ident-ical . Only minor variation could be detected in the concentrations of some of the components isolated. 129 Samples of the oligosaccharides present were isolated by elution from paper chromatograms. Since a l l compounds isolated were known,data are present in Table XXIV. TABLE XXIV NEUTRAL AND ACIDIC COMPONENTS OBTAINED ON HYDROLYSIS OF LEMON GUM OLIGOSACCHARIDES Neutral Components RGAL Solvent [A] Components on Hydrolysis Rhamnose 2.15 Rhamnose Arabinose 1.30 Arabinose Galactose 1.00 Galactose 1 0.53 Galactose 2 0.37 Galactose Acid GAL" Components. Reduced Components Components Solvent [B] On Hydrolysis on Hydrolysis 3 1.13 arabinose arabinitol 4-0-Me-Glucose 4 O.69 arabinose arabinose, arabinitol 4-0-Me-Glucose 5 0.75 galactose galactitol, 4-0-Me-Glucose 6 0.23 galactose galactitol, glucose 130 APPENDIX V A TECHNIQUE FOR HANDLING MICRO SAMPLES COLLECTED FROM THE GAS CHROMATOGRAPH A problem exists when samples are collected in capillary tubes on their elution from the column of a gas chromatograph. If these samples are to be used for TLC or Mass Spectra they may be dissolved in a solvent which can then be evaporated away on the TLC plate or on the probe for the Mass Spectro-meter. However, for micro analysis, IR and often with the Mass Spectra i t is imperative to avoid contamination of the sample with any solvent. To avoid this problem and s t i l l allow 25 jug samples to be handled requires a tech-nique which avoids losses due to adhesion to loops, wires or other transferring systerns. A system has been developed which can handle small quantities with l i t t l e or no loss. The technique involves manipulation of the sample in a glass collecting tube in the same way i t was collected, heat and, flowing gas. If a sample can be handled on the gas chromatograph without degradation i t may be handled by this technique. If the sample is reasonably stable i t may be manipulated using only heat since, as the sample is heated its viscosity is reduced and the sample w i l l flow as the tube is drawn through the furnace. A gentle flow of dry nitrogen or gravity may be used to increase the movement' of the'sample for less stable compounds, minimizing any degradation. In reality a portion of the gas chromatograph can be taken to the analyst or mass spectroscopist. In addition, purity of collected samples can be determined by analysis on the gas chromatograph, TLC or IR spectrometer. In order to handle 25 ju. g samples i t is necessary to reduce the bore of the melting point capillary to the size of a vacuum leak capillary. This can be done on a l l samples since the capillary action helps draw the 131 compound to the tip of the melting point tube. Positioning in the fine capillary-facilitates removal of the sample to the tip of the mass spectrometer probe, for spotting on a TLC or IR plate. In Mass Spectroscopy, since only o. JAZ sample is required ample sample for examination by other techniques is s t i l l available. This technique of moving the condensed sample in a melting point capillary by the use of a thermal push overcomes the problems faced by workers using one of the simplest and yet s t i l l efficient collection techniques. The capillary tube acts as an air condenser and sample container. Samples may be stored by simply sealing both ends in an oxygen -natural gas flame. Should the sample spread in the tube on long standing, slow insertion of the tube in the heater w i l l allow the sample to flow and collect in a small band. Even i f the sample should crystallize, its melting in the oven wil l cause i t to flow in the tube. The melting point capillary should be fire polished at both ends for clean insertion through the holed septum in the exit port of the gas chromatograph. By inserting the tube in this manner a l l the carrier gas and compound wil l pass through the tube and maximum collection w i l l be achieved. The fine capillary must be drawn after the collection because the extremely fine bore desired on the end of the tube would restrict gas flow and increase the pressure drop across the column. The minimum sample that can be handled this way was weighed at 25yug and represented an approximate one inch high peak on an F and M 720 gas chromatograph at maximum sensitivity. The upper limit of size on the col-lection of a peak, even when the tube has been blown slightly would appear to be about 7-10 mg. A sample larger than this is simply blown through the tube. 132 FIGURE 32 COLLECTION OF GAS CHROMATOGRAPHY SAMPLES IN GLASS CAPILLARIES AND OVEN DESIGN FOR THEIR MICROMANIPULATION SWAGELOCK NUT SEPTUM SAMPLE •EXJT PORT CAPILLARY TUBE VARJAC POWER UNIT SPIRAL. WOUND CORE If gas is required to help move the sample, a lecture bottle of pure nitrogen may be mounted on the board with the small oven. Extreme care must be used when using gas as i t is very easy to blow the sample out of the micro capillary tip. A large opening in the gas line, such that when open there is no gas flow out the exit tube and into the capillary is an excellent safety valve and avoids loss of sample due to blow through. This opening was closed with an oversized rubber stopper, since i t was easily removed when the sample neared the tip of the drawn tube. A small section of rubber tubing can be used as a bulb to force out sample. The tubing can be placed on the capillary without forcing sample out. A finger over the tubing end wi l l then move as much of the sample out of the tube as desired. Utilizing this collection 133 FIGURE 33 OVEN AND GAS SUPPLY FOR MICROMANIPULATION technique and method of removal,it has been possible to analyse a single run on the gas chromatograph of a complex mixture of partially methylated alditol acetates. Individual components varied in size as much as 100:1. This method of micromanipulation w i l l have wide application to the col-lection of gas chromatography samples for the entire range of organic compounds This technique w i l l permit analysis of gas charomatograph runs much in the way one would with a GLC - Mass Spectrometer combination. 134 APPENDIX VI COMPUTER SEARCHING THE CURRENT CARBOHYDRATE LITERATURE The information explosion has affected a l l chemists. As journals mul-tiply and papers increase i t has become increasingly difficult to cover a l l the journals which contain important references. Services such as "Chemical Titles," published by the American Chemical Society, have eased the need to search for journals among different library branches, finding on occasion that they are missing, borrowed or have not arrived. It i s , however, a prodigious task to review a l l the probable journals and scan the key words to ensure that one has not missed interesting or pertinent references. It is now possible, however, to utilize one of the new computerized Chemical Title search systems to retrieve a comprehensive l i s t of titles of papers of interest to carbohy-drate chemists. Not only the laborious work of searching but also of writing is immeasurably reduced since the computer print out provides the subscriber with a printed reference sheet for each t i t l e retrieved. The programme de-veloped here is designed to cover carbohydrate chemistry from synthetic to polysaccharides. Two restrictions were eventually imposed on the developed programme because of interests and economy of terms. The nucleotides were re-moved from the search profile as well as certain specific names. The names were included i n i t i a l l y but later removed when i t was clear that there were few retrievals that could not have been picked up by checking the key words of Chemical Titles. The extra possible profile words were more valuable in limiting retrievals in more difficult areas. The profile development will be outlined in order that individuals with specific or broader interests who may wish to develop their own search profiles w i l l be able to improve and shorten their profile development period. 135 The search profile was designed to retrieve references to carbohydrate chemistry using the CAN/SDI^" Project for.computerized searching of Chemical Titles. Since nomenclature for the field of carbohydrates is systematic, i t is possible to use these systematic endings and structural words to retrieve a l l references to carbohydrates. Profile words used in the search are given a letter code. Initial lists of terms were prepared. These were shortened by truncation and utilization of common letter groups. Truncation allows retrieval of words using letter groups preceeded or followed by other letter groups. Truncation Retrieval • 1. OX* ox, oxen, oxide 2. *0X ox, box, fox, Xerox 3* OX ox 4. *0X* a l l the above, plus , hydroxy, foxes, boxes, dioxide, etc. The l i s t of words of desired retrievals should be scanned for common letter groupings, for example: Cerebr o side s Gangli o side s ' ' ' • Furan o side s Pyran o side s Glyc o side Incorrect truncation w i l l result in excess retrieval. •KSIDE-* besides the terms above picks up side, consider, residence, prus-side, sidewise, inside and sidechain, however, *OSIDE* picks up none of these. National Science Library, National Research Council of Canada 136 TABLE XXV COMPUTER PRINT OUT OF PROFILE ONE jus * •G9 P 526 T A * ALD* PROF I L E OUITON. G G C * C E L L * "D""*ER YTHR*~" E-_*FUC*_ F *HEX* G *KE T* I T T I ~J T T T "IT T r r H *P ENT* 1 * R I B * J * SORB* K *TAL* 'L~"*TETR*' M * THRE* I T T T T T N *XYL* 0 *ALTR* _ _Q_*FRUCT_*__ R * GAL AC T* S *GUL* I T r j T T T *LYX* U *MANN* V *RHAMN* W_*GLUC* X *GLYC* Y CARBOHYDR* T T T T T T T "T" T "T" T AT PULYUL* AG_*SACCHAR* "AH"*ST ARCH*~ A1_*SUGAR*_ "AJ *AR I C * AK *ASE* - A X ~ * T X J R X N U ' * " AM * I T O L * "AN ~*L ACTuNE* AC) *0N1C* ~AP~*OS AMINE"*" AO *OSE* 1 : ATTVpyRA-jTO* T AS * S I O E * -y Ar~*~U'R ON * : T AU AGAR* T AV ALGTN* T AW CARRAGEENAN* ~T A X C H T T T N * T AY C H O N D R U I T I N * - j AZ DE'X TR* ""' T BA_1NULIN*_ _ _ _ "T BB K E R A T A N * T BC L E V A N * T T nr.-' T r T Z E X U U A l h * AA GUM* _ '_ A8 MUC* AC_*MYCIN* AO. *NUCLEOT I D * " AE POLY HYDROXY T BD NEURAMi N* T BE PECT* T BF SIALIC*" ' I BG TEICHO'IC* T "" BH ACETOLYSIS T BI EPGXID* i T "T "~ T "EOF E02 ~BTJ RTTTiYirAI i U N * BR MUIARUT* . ~BL~ KLEBS ICLL A : BM- PER IOOAT.t - Qq- FA" | "B—Ny^XATfAK—A TT 99 ( G j P - V ) E03 I E04 T "E0'5T" E06 T E07~T" E08 T 99 99" 99 "99" 99 w is X|X " ( Y U - A T ) _ ( A J ] A L I A M j A O j A P J A R j A T ) "A"Q'i~A"Q : : : AS | AS h09 i yy TAUTA1/-BM j 137 Another example: Muc ilage Muc ic Muc o polysaccharide MUC * however, picks up as well mucosa, mucosal, mucous, mucoid, muconic, mucor; however, MUCI* and MUCO are completely selective. The f i r s t search profile was over-truncated and the pairing of most terms was not sufficient to eliminate many unwanted references. Thus:-Profile Word Unwanted Retrieval *ALD*- aldrin, diels-alder *AMYL *• carbamylation XCELL* cells, sub-cellular KHEX-* hexane, hexanol *PENT*. pentoxide *RIB* distribution, ribosomyl, ribosomes *SORB* adsorbed •*TAL* metal, skeletal, congential, digitalis, orbital, crystalline, etc. Removal of the prefix truncation on a large number of the root words was at-tempted with some degree of concern since i f the root were in a compound name, such as trimethyl-D-ribose, the computer might view this as an unbroken word and no retrieval would take place. Fortunately this was not the case and a l l roots preceeded by a hyphen were printed out. The following additional changes were made in the preparation of the second profile. AO *0NIC* from E 06 because of terms ending in "onic" electronic, anionic, embryonic, paraconic, carbonic, etc. AC * MYCIN*- because of too many references to drug use. TABLE XXVI 1J8 COMPUTER PRINT OUT OP PROFILE TWO U 5 26" " " P R O F I L E DUTTON» "G G S T A ALD* T ' B AMYL * T AF PCLYOL* T. T : C 0 CELLULOSE I T ERYTHR* ~ "j T " " * AG AH *SACCHARI* *STARCH* T E FUC* T A l •SUGAR* T " F HEX* ~ ~ "7 "~ , .T •• AJ •ARIC* T G KET* T AK *ASE* T "" H PENT* . , T "AL " * FUR ANO* T I RIB* T AM *ITOL* ( •LACTONE* T J S O R B * " T~~ AN T K T AL* ) T AO *ONIC* T L' TETR* " • • • • •- : T AP *OS A M I N *. T M THRE* T AQ *OSE* T ' N XYL* T "AR *PYRANO* T 0 *ALTR* T AS *S IDE* T ' " P ARAR* ~ : _ _ T - AT •URONIC T Q •FRUCT* : . T • AU AGAR* T R *GALACT* " ' T . "AV ALGIN* T S GUL* T AW CARRAGEENAN* T " T *LYX* T AX C H IT I N * T U MANN* T AY CKONDKO IT IN* T V * RH AMN* " " T ~ AZ DEXTR* T w *GLUC* T BA INUL IN* T X •GLYCO* ' T BB KERATAN* T Y CARBCHYDR* T BC LEVAN* ... _ EXUDATE* ' T BO NEURAKIN* T AA GUI"* ' T BE PECT* T AB MUC* T B F S I AL IC* T AC GLYCAN* T BG TEICHOIC* T "" AD *NUCLEOTID* T AE POLY HYDROXY .1 T BH ACETOLYSIS 1 T BI EPCXID* T BJ" METHYLAT ION* T BK MUTAROT* . ! T BL KLE6S I E L L A ' ' ' T BM PER IODATE EO l T 99 "(AlBlD-NJCCAJIAk-AT)"™ E02 T 99 ( G l P - V ) E0 3" "T 99" / w I w EC4 T 99 X 1 X EC5 T 99 (Y | Z-AC) E06 T 99 . (AOIAE-AI) E07 T 99 ~" ( A J ] AL i AM 1 API AR | At ) EO£ T 99 AQ | AO EC 9" T" 99" AS1AS E 10 T 99 , (AUIAV-BM) E l l T 99 , C)C 139 X *GLYC* and replacement with * GLYCOL and GLYCAN* because of references to glycine, hypoglycemic, etc. AT HURON* to *URONIC because of neuron, pleuroncipes. AG *SACCHAR-* to AG *SACCHARI* avoids saccharomyces. AP -tfOSAMINE* to AP *0S AMIN* CT word splitting. C *CELL* to cellulose to many words with CELL EQUATIONS EOl Remove C *CELL* to new equation E l l E05 Made into two equations E05 and E06 to many retrievals E06 - » E07 Remove term AO because of terms inding in onicj electronic, anionic, embryonic, paraconic, carbonic, etc. PROFILE NUMBER 2 EVALUATION EOl S t i l l too many titles which contain, by chance, two of the required terms, i.e. THRE, ASE Triplet State effects in dye lasers at threshold. HEX, OSE Binding energy and compressibility of body centered cubic and close packed hexagonal sodium. Therefore at the risk of losing a few references [2] EOl was split into two search equations. The equations were selected so that the [3] EOl contained a l l those root words which were found to commonly occur in other TABLE XXVII COMPUTER PRINT OUT OF PROFILE THREE T ~ U 526 - - - - PROFILE DOTT'O'NT, T. G S T A ALD* 6 ERYIHK*. 1 AF *SACCHAR1* T C HEX* ! T AG *STARCH* T"' " " D KET* T " "AH * SUGAR* T E PENT* T A l * A R I C * "T " F R I R * " ~ " r AJ' *ASE* " "~ ~ T G TETR* . T AK *FURANO* T H'THRb* • i AL * I T l . ) L * T I AMYL* 1 T AM *L AC TONE* " "T J FUC* i T "AN~*ONTC* T K SORB* ! ] AO *OS AM IN * ~ T'~ L"TAL* "' • i T AP *OSE T M XYL* : T AQ *nSFS " T N * A L 1 K * ; 1 AR" *PYR A N O * " ' T 0 ARAB* : T AS * S I O E * ~- T P *FRUCT* , . . . T . . . . . . AT *URONIC ! T Q *GALACT* ! T .AU ADIPOSE* : T ~ R GUL* T AV CLOSE* ; • T S LYX* T AW DOSE* T T MANN* 1 AX MANN ILH ; T U RHAMN* : T AY S I D E * "AZ" METABOL I SM T~"'" ... j... V * G L U C * ~ " T ~ ; T W *GLYCO* T BA *ENE* i t " ~ X T. ARBOHYDR* """"", 1 BB GLUCONEOGENESTS T Y EXUDATE* T BC, GLYCOLYSIS T L GLYLAN* ' ' i BD MUCOSA T AA GUM* ! T BE CELLULOSE* - - - -j- ~A8 MUC * . ' " " ~ ~" ' " i . e . Bh CKBKA T AC *NUCLEOTID* j . T BG PHOSPH* T ~AD" POLYHYDROXY T AE POLYOL j i _ EOl'T" "99 (A J B'-H )'£ ( A I | A K J AL I "AO | A P |~A"Q fA'R\ A T}-»("AU ] AV j "A W | A I) i - E02 T 99 ( I | J - M ) & ( A I | A J - A T J M A U | A V l A W l A Y | A Z ) E"03~T" 99" (N|U-U)MAX|AZ|Bti) E04 T 99 BE | BE "" "E05 T" "99 ""' V-.IAZI BB 1 BG) E06 T 99 W M A Z l B A I B C ) F07 "T -go. (X|Y-AB)-.(AZ | 0 n ) E03 T 99 <AC| AD-AH)--(AZ) 1 E- 0-q- T- 99 AK|AL1 AO|AR|AT E l O T 99 ( AP | AQ ) -»( AU | AV | AW ) E11'"'T "99 AS-AY ~~ " E12 T 99 BF|BF I k l fields of chemistry. The following very common endings were removed: ase , lactone , onic , side In addition certain not terms were added to remove some commonly occurring words. adipose, close, dose, metabolism The remainder of [2] EOl was the same as before with the addition of the not words above and the word AY SIDE* . Equation [2] E02 formed [3] E03 with the not terms MANNICH, METABOLISM and PHOSPHX added. A number of not terms were added to the rest of the equations in order to eliminate commonly occurring non-carbohydrate retrievals, i.e. *ENEK was added as a not term for the equation [3] E06 to remove terms such as poly-ethylene glycol. In order to reduce the profile to 60 terms and to incorporate the required not terms i t was necessary to delete the terms AU to BG of profile No. 2. As a check on this completeness of the retrieval of references the coden for the Journal Carbohydrate Research has been given as the last equation in the profile in order to pick up any titles not previously printed in order to see why they have not been included. PROFILE NUMBER 3 The following changes were made in profile No. 3 to produce the final profile. 142 Profile No. 3 Term AB MUC* AC * NUCLEOTIDE AS *SIDE* AY SIDE * AZ METABOLISM BB GLUCONEOGENESIS BC GLYCOLYSIS BD MUCOSA Profile No. 4 Term AB MUCO AC MUCI AS *OSIDE* AY GLUCAGON AZ METABOLI* BB INSULIN BC GLYCOLY-K-BD CORTICO* BH TRANSPORT Reason Too many non-carbohydrate words Removal of nucleotides ref. from profile MUCI and MUCO replace MUC Improved truncation Replaced because of improved AS term New not term Better truncation now will pick out metabolic as well Term AB * ENE K covers this term New not term Better truncation [glycolytic] No longer needed [AB] New not term New not term The equations for the search expressions were modified to include the new not terms. Amylase references have been eliminated as well as most references to carbohydrate phosphates. The equations are, however, open to general refer-ences such as sugar phosphates. 143 TABLE XXVIII COMPUTER PRINT OUT OF PROFILE FOUR U "52"6 " •. 0 9 0 6 6 9 PROFILE DUTTON, G G S T A A L P * . T 8 ERYTHR* t AF * S A C C H A R T * _._ T C_HEX* T AG_*STARCH*_ "T , " ~D' KET* " | f " AH '* SUGAR* T E__PENT* _ J T A l * A R.IC* - T "F R I B * " - f* A J ~ * A S E * T G TETR* ' T AK .* FUR ANO* _ T H THRE* T AL * I T O L * T I_ AMYL* • T AM *LACTONE* f " j "F(JC*" ' f AN'"*'ONIC* T K _ S n R R _ * I. T AO_*OS AMIN* "J" ' L TAL* ' " ~ f "AP"' *OS'E T M_XY J L * : T AO *OSF.S T N *ALTR* T AS *OSin " E ' * T 0_ARAB* J _ T A R_*PYRANO* _ T - ~ P ~ * " F R U C'f* " " ~J " A T • U R O N I C * " T Q__*G AL ACT* A U _ A D I P0S F _ _ f " " R GUL* : ~T AV CLOSE* T S L Y X * _ . ' T AW DOSE*  T T M A N N * T AX MANNICH _ T U Rl-lAMN* T _ AY_GLIJCAGON_ _ T " ' "V *G L L T C * " ." ; f ~AZ~ META B O L I * T W_*GLYCO* | T __BA_*ENF*_ f X C ARBOHYDR* "\ "f BB INS1JL I N T Y EXUDATE* • j T BC GLYC . O L Y * T 7. GLYCAN* T BD CORTICO* T " AA GUM* ; | T . BE CELLULOS* ' "f A B "Mucn " . " ' I C BF" CRBRA T A C _ M U C _ I * ' T 8G PHGSPH* y - - AD POLY HYDROXY • •" i T AE POLYOL | ' T BH TRANSPORT* ' , ' E O l T 99 ( A 1 B - I ) G ( A K | A L I AO 1AP1AQ| ARI A S ] A T ) ^ ( A U I A V ) A W 1 A Z 1 B G ) EO? T 99 ( J 1 K-M ) & ( A I I AJ-AT )-. { AU I A V I AW I AZ I BG ) E 03 T 99 ( N I O-UJM A_X I AZ I B G ) _ .... " E 0 4 " t ~ ' 9 9 B E IBE E 0 5_ T 99 y> ( AZ I 80 I BG I AY 1 8A l_BD I B_H) t  "EQ6 T 99" AZ | B A I DC I BB I R G) E07 T 99 (X|Y-AC)~»(AZ|BB| BH) E08 T 99 I AD | AE—AH )-»(AZ|BB|BH) E09 T 99 I AK I AL J AO | AR | AT ) M AZJ BG |_BHJ ¥.10~f~99 (AP | AQ ) -»( AU 1 AV 1 AW I AZ 1 BB | B G) E l l T__9_9 AS->(AZ 1 BB 1 BG I BH ) E l ? " ! 99 BF'I BF TABLE XXIX. EVALUATION SEARCH PROFILES PROFILE NO. 1 Search Expression C.T. No. 11, 1969 C.T. No. 12, 1969 Yes No Total Yes No Total 1 12 87 18 81 99* 2 8 46 54 10 52 62 3 15 34 49 6 24 30 4 16 65 10 35 45 5 50 69 99j 29 53 82 6 5 94 99* 6 78 84 7 12 62 74 13 39 52 8 4 13 17 3 10 15 9 7 11 18 5 15 20 Total 109 481 590 100 387 $ of Total 18.5 81.5 20.5 79.5 Maximum number of retrievals 99• PROFILE NO. 2 Search C.T. No. 13, 1969 C.T. No. 14, 1969 Expression Yes No Total Yes No Total 1 13 61 74 23 72 95 2 9 10 19 18 11 29 3 8 22 30 19 33 52 4 15 19 34 11 25 36 5 12 7 19 12 16 28 6 22 20 42 18 26 44 7 2 7 9 5 7 12 8 11 42 53 15 35 50 9 7 15 22 6 15 21 10 6 17 23 7 19 26 Total 105 220 325 134 259 393 $ of Total 32.3 67.7 34.1 65.9 TABLE XXIX [Cont'd] PROFILE NO. 2 Search Expression C.T. No. 15, 1969 C.T. No. 16, 1969 Yes No Total Yes No Total 1 20 79 99* 16 65 81 2 13 10 . 23 7 7 14 3 6 ' 26 32 15 29 44 4 12 29 41 10 15 25 5 9 16 25 6 7 15 6 23 44 67 16 21 37 7 6 21 27 1 8 9. 8 11 60 71 4 34 38 9 5 18 23 2 17 19 10 9 13 7 14 21 Total 109 312 421 84 217 301 $ of Total 25-9 74.1 27.9 72.1 PROFILE NO. 3 Search C.T. No. 17, 1969 C.T. No. 18, 1969 C.T. No. 19, 1969 Expression Yes No Total Yes No Total Yes No. Total 1 16 4 20 13 7 20 8 7 15 2 6 4 10 1 2 3 1 9 10 3 7 5 12 5 9. 14 4 5 • 9 4 12 11 23 6 11 17 15 11 26 5 8 20 28 7 - 21 28 3 18 21 6 11 16 27 12 14 26 7 8 15 7 10 9 . 19 10 2 12 5 5 10 8 13 30 43 9 18 27 17 34 51 9 2 4 6 1 6 7 2 1 3 10 4 16 20 2 11 13 3 13 16 11 5 16 21 2 13 15 l 9 • 10 Total 94 135 229 68 114 182 66 120 186 $ of Total 41.0 59.0 37.4 62.6 35.5 64.5 TABLE XXIX [Cont'd] PROFILE NO. 4 C.T. No. 20, 1969 C.T. No. 21, 1969 Search Expression Yes No No But Valid Total Yes No • No But Valid Total 1 14 1 4 19 9 0 6 ' 15 2 4 2 0 6 1 5 2 6 3 13 1 0 14 9 1 11 21 4 10 0 7 • 17 7 0 8 15 5 8 0 10 18 6 0 17 23 6 8 2 9 19 8 5 15 26 7 10 0 2 12 7 0 7 14 8 10 3 9 22 28 0 23 51 9 2 3 1 6 1 0 1 2 10 4 3 2 9 5 7 2 14 11 2 0 2 4 2 0 10 12 12 5 0 1 6 0 0 0 0 Total 90 14 48 152 83 14 102 199 $ of Total 59.2 9.2 31.6 41.7 7-0 51-5 C.T. No. 22, 1969 C.T. No. 23, 1969 Search Expression Yes No No But Valid Total Yes No No But Valid Total 1 11 0 6 17 8 0 4 12 2 3 0 • 1 4 2 2 0 4 5 10 2 7 19 13 0 2 15 4 10 0 12 22 9 0 9 18 5 9 0 17 26 14 0 14 28 6 10 2 8 20 6 5 9 18 7 3 2 8 13 3 0 0 5-8 13 2 12 27 15 3 12 30 9 2 1 2 5 1 2 2 5 10 2 6 9 17 6 3 8 17 11 1 0 4 5 7 0 9 16 12 3 0 0 3 0 0 0 0 Total 77 15 86 178 84 13 62 166 $ of Total 43.3 8.4 48.3 50.6 7.8 41.6 Ikl The final search profile shows extremely good reference retrieval. The references which were not useful retrievals have been broken down into two sub-groups, references retrieved because of carbohydrate words and references re-trieved because of other factors. Many of the non-useful references are legitimate retrievals but outside our fields of interest. The retrieval has been based on the presence of a carbohydrate word in the t i t l e of the article. One example for each search expression is listed below. E01 HEX, OSE E02 SORB, ITOL E03 ARAB E04 CELLULOS E05 E06 E07 E08 E09 E10 E l l GLUC GLYCO CARBOHYDR SUGAR ITOL OSE OS IDE ... A Genetic Reappraisal of Hexose Transport by Kidney and Intestine ... Effects of Ethanol, Sorbitol and Thyroid Hormones L-Arabinose Binding Protein ... Chromatography ... on Columns of Benzoylated di ethyl-amino ethyl cellulose Effect of Serotonin on Glucose ... in Rabbit Brain Catabolism of Plasma Glyco Proteins ... ... Chickens Fed Carbohydrate Free Diets ... Effects on Sucrose Yield of Sugar Beets ... Effect of Myo Inositol on the Prevention ... Nature of Lactose Fermenting Salmonella, ... Radio Denaturation of DNA Deoxy Nucleosides. The other terms which are non-useful are normally less than 10$ of the retrievals and f a l l into one of the following groups of retrievals. 1] Accidental combination of two words XYL, ASE Xylene, bases AID, FURANO Diels-Alder, furanophane 148 2] Truncation retrievals GALACT Galactic SACCHARI Saccharin OSE Rose, purpose, those, etc. 3] Specific words found in other fields of chemistry OS, AMIN Di ethyl nitros amine GLYCO Glycolate POLY, HYDROXY Poly hydroxy phenylenes It is clear that the final profile w i l l retrieve a highly significant portion of the carbohydrate literature present in any issue of Chemical Titles. The profile may be modified as well to be more specific and to retrieve ref-erences missed by altering the profile words and search equations. If at f i r s t glance the number of non-useful retrievals seems large i t should be borne in mind that i t is a matter of moments to peruse the print out and discard those not wanted. In practice a high percentage of marginal material can be tolerated before the mechanical retrieval system ceases to be competitive with a manual search of the literature. The computer print out has in effect carried out a preliminary screening and "short listed" possible references for closer attention. Furthermore, each entry of the print out is on a separate [perforated] sheet of paper which may serve as the basis of a personal f i l i n g system. The time saved in avoiding the necessity of trans-cribing references of interest should also be considered when assessing the merits of the method. The use of the journal coden for Carbohydrate Research allows a check on the percentage of papers which would not be retrieved by the profile. Of over two hundred references in Carbohydrate Research only 8.6$ would not have been 149 retrieved by the profile. A portion of these were intentional exclusions by not terms and a few references would have been obtained by examination of Chemical Titles keyword lists for the twenty or so profile words eliminated to include the not terms and shorten the profile [compare profiles No. 1 and No. 4]. If these references were included only 5'4$ would not have been re-trieved. Examination of several t i t l e pages of Carbohydrate Research wil l in-dicate the type of references which wil l not be retrieved by the search profile. In general, i t w i l l be seen that the titles often give no clue to their car-bohydrate nature. In terms of this evaluation the profile proposed is 95$ effective in retrieving desired references from journal t i t l e pages. The same profile has also been used to search Chemical Abstract Condensates with good results although the percentage of undesirable retrievals is higher. This situation could now easily be improved by the addition of a profile word PATENT as a not term. This possibility was unavailable at the time profile 4 was developed. 150 APPENDIX VII LEMON GUM : 0.1 Li in ! i; i ! H o i l l liiL . , , . J T i o l •o ' "T g 1 lO ro * . I. I I i'r' •': !o sit: 0) li w :iJ! r I Li mi in IT'IS~: ; . i : o p i , !1TI lio-r jllifpii X :• 1 . . . HI ' i l i ' t l ' . — • ! ! o o o in "F 1 CO CI 0) •? 1 LT\ 1 -4-O •> • \ o w OJ OJ 3 P4 o X 50 100 150 200 m/e FIGURE 36: greater amount LEMON GUM PEAK 5 x 2,5-Me2-ARAB o 3,5-Me2-ARAB 153 50 100 150 . 200 m/e FIGURE 38: LEMON GUM PEAK 6T 2,3-Me2-ARAB § 156 -III I •Mil I: i ! i : llii . to jfil,, .1.1 ...!.. Ml M l ! !M h i Ml !Mi M tii- -J_MJ ,41-1 IM LliUl cvj • O). i.o>l ! .031 M II:: I i M-MM CM O O O O w o o o m o 3 § i? 8 f a -=r s 3 K\ w w •» 3 CL. CVJ o 158 .1 50 100 150 200 m/e FIGURE 43 LEMON GUM PEAK 10 l/lON H2S04 LEMON GUM PEAK 4 . SMITH POLYALCOHOL LEMON GUM PEAK 6 LEMON GUM SI PEAK 6 [DIFFERENT SAMPLES SHOTTED ONLY MINOR INTENSITY CHANGES] 2,4-Me2-GAL H 162 1 i l lll:.r, I! 'j • ill ;ir !i:i i Hi -H !i Am. . I. ! I Hi to. .lijil I . i ! i i ill ; i i Ah ~i • ill!, ii'i si--t;T| Hi! 4j. . i I 'III .a liiitii i.L f rrtt Till LUfl 8 CM iliii L;3 iiniiip o N!! H1! [ill o o i|'0>., Mil o m I:: • I :..I-:i. I,,' I : 'l.l: g O 3 i co QJ 8 w ^ rl < IA ^> w rH (X, CM O FIGURE Vf: SMITH POLYALCOHOL LEMON GUM PEAK 2 2,3,4-Me3-ARAB 100 43 ox i - - - - j 1.'. ':X-.:-\":-S T^ ?;T:4 r-r{->";zr- ;!;=r.t'.-. |l:'-":~4 i t'l-Tti'.f^ -il-r1 R.I. 50 200 m/e FIGURE 48: SMITH POLYALCOHOL LEMON GUM PEAK 3 2,5-Me2-ARAI5 [greater amount] 3,5-Me2-ARAB 100 200 m/e FIGURE 49: SMITH POLYALCOHOL LEMON GUM-PEAK 4 o 2,3,4,6-Me4-GAL x 2,3,-Me2-ARAB $ R.I. 50 43ox 1 - Z.Z zizz - -. T^7l 7777^! - -77: • . ". . - , .7. 77 .... 7I7-~7~_™7~ : - ._ • 17-I-I §7 -77r-r 7777.4:4^~l:T""v--f ZZ : - .. : : *. 1 ,-Z... " ^ 7 7 7 -.-77^  • -• zzz'.. jr-T-7j'f ] : : \ M z l z z s z z z 'z _- ; -~"zErLL ''•zTzi-lz-i.r - ——t — - — -. — — — -. „ — J — ^ — , — rE-.f:r—irr-rr 4inzTT-f:7r777 — i | . _ -ZZZTJ.ZZZZ <—'.*.-. - 1 7I~77_ - ZZZZ r-7i: I — . — . r H .-TTTT zz: ^ zzzzzzzzz. 'z-zzzzrrr "zi : r —"7 -77U77777 777. 7; LZZJlZZZ - - r — ZZTZZZ, .7177771 7771 1 . 1 ~rr~-J-7-- r~ "1— r—j - - r r . " • r Z"- JZZZ.ZZZZ- zzzzZ.zz.zJ^z .: _.'7 7_ ... . J _ ~|7~ "t\.^ ~. 7_f r — j--; - ; --• 1 ~t r f " • ". 1 — : . * i - ^. - j * . . . .. . ; —' —-r—r . : i • * , . _.. „ , : _ _ . ~! „ 7771 • —'- -P" i 4 1-.-' I ' , f"7.. ,'77 „*l*^  : \7 :. "7 " ! 71777  - ~f -r*-1— -77.77±~.7777 —•— r ~ 7 7 77 . • , , , — j . : • ._ : . - 777^7X71727^  7~ . F  _ 7777777X77; .J-_}"_ u,- :- - j- -,- - T r t r r r : - - _ r •777r_1.7 zz. - — — i ' . 1 I „ r i J — . 1 i — i _—... , ~ i ~ "7—77 ' L_ .— — -f — -~-— • - — — —r— — — .. f,. • 1 z _ • ; 7ZT " i - ' i • —_—i.—1_„„ - ; — —1____„_.. ,.„-. j „ „ i . .777.777 177:^777777^7 i— —r~: .77777 l~ "-T - ! L . I 1 ';; 77^  ~ . , • j, • • •,' ; l- - - { . _ .. |... „; ... r « f ,• • ( • "T . j _L:  r • i . i ' ? 1 1 \ . i •. - 1 I 1 - ~ t T ;; , i ; 1 b~7m ~ LvX ...... —.—r.— --j —p - t- — • ;. i 'I i' • • • ' i - , - - :  -t—^ ' " „ — i —; { - - i „I_-L4. — — —i—.—^_ T — :—i • :—f 1 .. — y — , f. . ^ „ L R . „ — . • ^ — -f 1—1  • -—• -™ -r- -— fTTT™ 1771777 ;—. 17777777:.-7771 - — ; — r r r t~*~ zzzzzzz 777777 . . . . . . _j. .... i 7 77 X777 L - ^ E - - : ;z~rt7:7:r|""_H f- * tz~:~~>~r~~ U7777j"—^  •r±rxr.:- i ' . : .' T7..7-—U;— r - f — 77 7777J j. ^ — — ( _ . — , r . • . i • 7 17  — r i f ~ • - -,. .zzzz -!-r- . , • - .+ - F . z h -zzz^.ii -.- ^  ^ TTTTT^—rlr_7.„> -• ' 1  •7T7":?',7" .', . 77 - r " - , - ' — T™ \ . . i • i ~~ " ' ' f " ^ „.— — i i •;—**• ~^-rr— - \ f y~—'• "i ~ . . ' f _ .—:—i— 1 „ . zz I • —_ i _1 [ f . vzz • ——I — i ' 1 1 , 1 . . J . , — — : j i i i —. _ j • .' r77_ i • • ! 45 ox 7.-77 zi r ~~ ' • ! • i ; • i zz _: - z • h- H 7 r t v 1 . j • • • r . - . • • ; I zz ' r—~"'Xn " l?9r t " 1 T r : f _ i . z~ • • - I .1 7J1777_ 1 ' ! . 1 • . .„i :- ' . 1 . , • I" -J. zz — _ i — . _ .^ i " !• : • -z — — — ; • : 87ox 7.77 Z.ZZZ .... j • i • ( i i _l _;_ IZ _ _: u. :: - 1 ; ! - . — i - i ~ ' : — — j — t77"7 "1 — lOlox "77 1  L : j. , _4_. ^^^^^i . v t77__ ll .zzzz ~7717 , . .7 „ ~ r . 77 .7. ..777 . Z—Zi:~ II —' ' ~? — — • — — - . — - i , _f , 1 , : — ™ If -_ - L — -| - -777—. _- ^ & A 77. ~7_ -—~ 77T7 | j ; _ 4. — 7~7777X7~' i _ . ^ j! 7 . 7 ; _ ; ~ T: I 77  77-;:-777 . 7_. _7_ • . -~_777777J r " . J 1610 : . ' ' - -- 189 r r r - t — 233x II -L _ . 7  ™ r — | j :^ , - . ; 1 _ • -1 ll 1 || l l 1 1 l 1 l l II 1 " 1 1 •' .--7 7. -7777-7-77—7 -7777 7-7r7.7-_77.7-7 50 100 150 200 m/e FIGURE 50: SMITH POLYALCOHOL LEMON GUM PEAK 5 2,3,4-Me3-GLU x 2,3,6-Me3-GAL ^ 0 2,4,6-Me3-GAL o £3 50 100 150 200 m/e FIGURE 51: LEMON GUM SI PEAK 1 2,3,5-Me3-ARAB H ON 50 100 150 200 m/e FIGURE 52: LEMON GUM SI PEAK 2 2,3,4,6-Me4-GAL H 8 50 100 . 1 5 0 200 m/e FIGURE 53: LEMON GUM SI PEAK 3 o 2,4,6-Me3-GAL x 2,3,6-Me3-GAL 50 100 150 200 m/e FIGURE 54: LEMON GUM SI PEAK 4 2,3,4-Me3-GAL H O 172 BIBLIOGRAPHY 1. J.F. Stoddart and J.K.N. Jones, Carbohydrate Res., 8, 29 [1968]. 2. J.M. Bobbitt, Adv. in Carbohydrate Chem., 11, 1 [1956]. 3. J.R. Dyer, "Methods of Biochemical Analysis," Interscience Pub., New York, Vol. 2, 111 [1956]. 4. G.G.S. Dutton, K.B. Gibney and P.E. Ried, Can. J. Chem., Vj[, 2494 [1969. 5« H.O. Bouveng and B. Lindberg, Adv. in Carbohydrate Chem., 15_, 53 [ i960]. 6. I.J. Goldstein, G.W. Hay, B.A. Lewis and F. Smith, Abstract Papers, Am. Chem. Soc, 135. 3D [1959]. 7. I.J. Goldstein, G.W. Hay, B.A. Lewis and F. Smith, Methods in Carbohydrate Chem., Ed. R.L. Whistler, Academic Press, New York, Vol. 361 [1965]. 8. F. Smith and R. Montgomery, "The Chemistry of Plant Gums and Mucilages," Reinhold Publishing Corp., New York, N.Y., p. 194 [1959]. 9. G.G.S. Dutton and A.M. Unrau, J. Chromatog., 36, 283 [1968]. 10. F. Eisenburg and A.H. Bolden, Carbohydrate Res., 349 [1967]. 11. G.G.S. Dutton and A.M. Unrau, Can. J. Chem., *_3_, 1738 [1965]. 12. J.K. Hamilton and F. Smith, J. Am. Chem. Soc, J8, 5907 [1956]. 13. A.M. Unrau and F. Smith, Chem. Ind. [London], 330 [1957]-14. N. Dubois, J.K. Hamilton, K.A. Gilles, P.A. Rebers and F. Smith, Anal. Chem., 28, 350 [1956]. 15. R. Bentley, C.C. Sweeley, M. Makita and W.W. Wells, J. Am. Chem. Soc, 8_5, 2497 [1963]. 16. T.E. Timell, Adv. in Carbohydrate Chem., 20, 410 [1965]. 17. J.C. Bartlet and D.M. Smith, Can. J. Chem., _3§, 2057 [I960]. 18. G. Wulff, J. Chromatog., 18, 285 [1965]. 19. R.J. Alexander and J.T. Garbutt, Anal. Chem., 303 [1965]. 20. J.S. Sawardeker and J.H. Sloheker,. Anal. Chem., _3j[, 9^ 5 [1965]. 21. H.E. Brower, J.E. Jeffery and M.W. Folsom, Anal. Chem. ;3_3, 362 [1966]. 22. P.O. Bethge, C. Holmstrdm and S. Juhlin, Svensk Papperstid., 62, 60 [1966]. 173 23. Y. Halpern, Y. Houminer and S. Patal, Analyst, %2, 714 [1967]. 24. R. Gelin and F. Godot, Bull. Soc. Chem. France, 3096 [1966]. 25. W.A. Dietz, J. Gas. Chromatog., 2> 68 [1967]. 26. A. De Grandchamp-Chaudun, Compt. Rend., 262C, l44l [1966]. 27. C.T. Bishop and F.P. Cooper, Can. J. Chem., 793 [I960]. 28. M. Zinbo and T.E. Timell, Svensk Papperstid., 68, 647 [1965]. 29. D. Gardiner, Carbohydrate Res., 2, 234 [1966]. 30. G.G.S. Dutton and G.D. Jensen, unpublished results. 31. C.T. Bishop, Adv. in Carbohydrate Chem., 12, 95 [19653• 32. G.G.S. Button and A.M. Unrau, Can. J. Chem., 42, 2048 [1955]. 33. D.A. Rees and J.W.B. Samuel, J. Chem. Soc, 2295 [1967]. 34. G.G.S. Dutton and A.M. Unrau, Carbohydrate Res., 1, 116 [1965]. 35. G.G.S. Dutton and P.E. Reid, Abstracts IUPAC Meeting, Stockholm, p. 14 [1966]. 36. P.E. Reid, B. Donaldson, D.W. Secret and B. Brandford, J. Chromatog., 47, 199 [1970]. 37. E.V. White, J. Am. Chem. Soc, 62, 715 [1947]. 38. T.E. Timell, Chem. Ind. [London], 1208 [1963]. 39. J. Kenner, Chem. Ind., [London] 727 [19553* 40. D.A. Rees and J.W.B. Samuel, Chem. Ind. [London], 2008 [1965]. 41. F. Smith and A.M. Stephen, Tetrahedron Letters, No. 7> 17 [i960]. 42. H.C. Brown and B.C. Subba Rao, J. Org. Chem., 22, 1135 [1957]. 43. E.L. Hirst, E. Percival and J.K. Wold, J. Chem. Soc, 1493 [1964]. 44. R.J. Ross andN.S. Thompson, Tappi, 48, 376 [1965]. 45. J.H. Manning and J.W. Green, J. Chem. Soc. [C], 2357 [1967]. 46. G.R. Pettit and T.R. Kasturi, J. Org. Chem., 26, 4557 [196l]. 47. C.T. Bishop and V. Zitko, Can. J. Chem., 44, 1275 [1966]. 48. R.S. Aries and H. Schneider, "Encyclopedia of Chemical Technology," Ed. R.E. Kirk and D.F. Othmer, Interscience, New York, Vol. p. 906 [1950]. 174 49. S.A. Barker, P.J. Somers and M. Stacey, Carbohydrate Res., ^ , 26l [1967]. 50. G.O. Aspinall and A. Cartas -Rodriguez, J. Chem. Soc., 4020 [1958]. 51. B. Vollmert,. Makromol. Chem., 110 [1950]. 52. G.O. Aspinall and J.P. McKenna, Carbohydrate Res., J, 442 [1968], 53. E. Schenker, "Newer Methods of Preparative Organic Chemistry," Ed. W. Foerst Academic Press, New York, Vol. 4, p. 198 [1968], 54. T. Imanari and Z. Tamura, Chem. Pharm. Bull. [Japan], 1^ , 1677 [1967]. 55. J.N. BeMiller, G.V. Kuraari, Abstract Papers, Am. Chem. Soc, l 6 l , CARB. 9 [1971]. 56. L.M. Jackman, "Applications of Nuclear Magnetic Resonance Spectroscopy in Organic Chemistry, " p. 20, Pergamon Press, Oxford [1959]* 57. 0. Raunhardt, H.W.H. Schmidt andN. Neukom, Helv. Chim. Acta, 50, 1267 [1967]. 58. J.M. Oades, J. Chromatog., 28, 246 [1967]. 59. B. Capon and D. Thacker, Proc. Chem. Soc, 369 [1964]. 60. R.C. Wiley, M. Tavakoli and M.D. Moore, Proc. Amer. Soc. Hort. Sci., 89, 34 [1966]. 61. W.N. Haworth, J. Chem. Soc, 107• 8 [1915]. 62. T. Purdie and J.C. Irvine, J. Chem. Soc, 83, 1021 [1903]. 63. K. Wallenfels, G. Bechtler, R. Kuhn, H. Trischmann and H. Egge, Angew. Chem., Internat. Edit. 2, 515 [1963]. 64. D.M.W. Anderson, E. Hirst and J.F. Stoddart, J. Chem. Soc. [C], 1476 [1967]. 65. R.N. Fraser and B. Lindberg, Carbohydrate Res., 4, 12 [1967]. 66. A. Misaki and F. Smith, Carbohydrate Res., 4, 109 [1967]. 67. I.R. Siddique, Carbohydrate Res., 4, 277 [1967]. 68. H.C. Srivastava and P.P. Singh, Carbohydrate Res., 4, 326 [1967]. 69. J.R. Nunn and H. Parolis, Carbohydrate Res., 6, 1 [1968], 70. R.D. Lambert, E.E. Dickey and N.S. Thompson, Carbohydrate Res., 6, 43 [1968] 71. G.G.S. Dutton and A.M. Unrau, Can. J. Chem., 43_, 924 [1965]. 175 72. R. Kuhn, H. Trischmann and I. Low, Angew. Chem., 6 j , 32 [1955]. 73« G. Bechtler, Ph.D. Thesis, Universitat Freiburg. [1962], 74. R. Kuhn and H. Trischmann, Chem. Ber., £4, 2258 [1961]. 75. R. Kuhn and H. Trischmann, Chem. Ber., <?6, 284 [1963]. 76. H.C. Srivastava, P.P. Singh, S.N. Harshe and K. Virk, Tetrahedron Letters, 493 [1964]. 77. J.S. Brimacombe, D.B. Jones, M. Stacey and J.J. Willard, Carbohydrate Res., 2, 167 [1966]. 78. P.A. Sandford and H.E. Conrad, Biochemistry, j?, 1508 [1966]. 79. S. Hakomori, J. Biochem. [Tokyo], 205 [1964]. 80. D.M.W. Anderson and G.M. Cree, Carbohydrate Res., 2, 162 [1966]. 81. B. Lindberg, H. Bjorndal, C.G. Hellerquist and S. Svensson, Angew. Chemie., Int. Ed., % 610 [1970]. 82. G. Pettersson, 0 . Samuelson, K. Anjou and E. von Sydow, Acta Chem. Scand., 21, 1251 [1967]. 83. H. Bjorndal, B. Lindberg and S. Svensson, Carbohydrate Res., 433 [1967]. 84. O.S. Chizhov, L.S. Golovkina and N.S. Wulfson, Izv. Akad. Nauk. SSS R, Ser. Khim., 1915 [1966]. 85. K.B. Gibney, M.Sc. Thesis, University of British Columbia [1967]. 86. L. Hough, J.K.N. Jones and W..H. Wadman, J. Chem. Soc, 1702 [1950]. 87. W.E. Trevelyan, D.P. Procter and J.S. Harrison, Nature, 166, 444 [1950]. 88. R.U. Lemieux and H.F. Bauer, Anal. Chem., 26, 920 [1954]. 89. A.E. Pierce, "Handbook of Silylation," Pierce Chemical, Rockford, Illinois, p. 11 [1970]. 90. E.J. Corey and M. Chaykovsky, J. Am. Chem. Soc, 8 j , 1345 [1965]. 91. C-. Neurnuller and E. Vasseur, Arkiv. Kemi, 2, 235 [1953]. 92. G.G.S. Dutton, K.B. Gibney, G.D. Jensen and P.E. Reid, J. Chromatog., ^ 6 , 152 [1968]. 93.. J.N.C. Whyte, K. Hunt and R. Walker, personal " communication. 176 94. K. M. Brobst and C.E. Lott, Cereal Chem., 43_, 35 [1966]. 95. G.G.S. Dutton, P.E. Reid, J.J.M. Rowe and K.L. Rowe, J. Chromatog., _4j/ 195 [1970]. 96. W.J. Whelan, "Methods in Carbohydrate Chemistry," Ed. R.L. Whistler, Academic Press, New York, Vol. 4, 252 [1964]. 97. D.J. Manners, Adv. in Carbohydrate Chem., 1£, 371 [1962]. 98. T.E. Timell, Svensk Papperstidn., 65., 435 [1965]. 99- 0 . Perila and C.T. Bishop, Can. J. Chem., jg, 815 [196l] . 100. G.G.S. Dutton and J.J.M. Rowe, unpublished results. 101. M.L. Wolfrom and A. Thompson, "Methods in Carbohydrate Chemistry," Ed. R.L. Whistler, Academic Press, New York, Vol. 1, p. 369 [1962]. 102. CL. Mehltretter, "Methods in Carbohydrate Chemistry," Ed. R.L. Whistler, Academic Press, New York, Vol. 4, p. 317 [1964]. 103. T.P. Nevell, "Methods in Carbohydrate Chemistry," Ed. R.L. Whistler, Academic Press, New York, Vol. j>, P' [1963]' 104. I.M. Hais and K. Macek, "Paper Chromatography," Academic Press, New York, p. 312 [1963]. 105. E. Stahl and U. Kaltenbach, "Dunnschicht Chromatographic," Springer Verlag, Berlin, p. 478 [1962]. 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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


Related Items