Advances in Carbohydrate Chemistry and Biochemistry, Volume 25

Advances in Carbohydrate Chemistry and Biochemistry Volume 25

1902-1969

Advances in Carbohydrate Chemistry and Biochemistry Editor R. STUART TIPSON Assistant Editor DEREK HORTON

Board of Advisors W. W. PIGMAN

S . ROSEMAN

WILLIAMJ. WHELAN

ROY L. WHISTLER

Board of Advisors for the British Commonwealth A. B. FOSTER

SIR EDMUNDHIRST

J. K. N.

JONES

MAURICESTACEY

Volume 25

ACADEMIC PRESS

New York and London

1970

COPYRIGHT 6 1970, BY ACADEMIC PRESS, INC. ALL RIGHTS RESERVED NO PART OF THIS BOOK MAY BE REPRODUCED IN ANY FORM, BY PHOTOSTAT, MICROFILM, RETRIEVAL SYSTEM, OR ANY OTHER MEANS, WITHOUT WRITTEN PERMISSION FROM THE PUBLISHERS.

ACADEMIC PRESS, INC.

111 Fifth Avenue, New York. New York 10003

United Kingdom Edition published by ACADEMIC PRESS, INC. (LONDON) LTD.

Berkeley Square House, London W1X 6BA

LIBRARY OF CONGRESS CATALOG CARD

NUMBER: 45

PRINTED IN THE UNITED STATES OF AMERICA

-

1135 1

CONTENTS LIST OF CONTRIBUTORS .................... PREFACE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

....... .......

ix xi

....

1

Stanley Peat (1902-1969) J . R. TURVEY

Text

......................................................... Gel Chromatography of Carbohydrates SHIRLEYC. CHURMS

I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

11. Types of G e l . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111. Theory of Gel Chromatography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

IV. Application to Carbohydrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. Value of the Technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

13 14 16 31 51

Crystal-structure Data for Simple Carbohydrates and Their Derivatives GERALDSTRAHS Introduction . . . . . . . . . . . . . . . ........................ 53 hydrates , . . General Features of the Crysta Monosaccharides and Derivatives .......................... Disaccharides . . . . . . . ........................ 75 Oligosaccharides .................................. . . . . . . . . . . . . . 77 Antibiotic Substances ........................ 80 Nucleosides and Nucleotides . . . . . . . . . . . . . . . . . . . . . . Miscellaneous Carbohydrates: tamins, and Hydrazones . . . . . 90 Enzyme-Substrate Complexes ............................ 93 .................... 98 Conclusions . . . . . . . . . . . . . . . XI. Addenda. ................................................. 107

I. 11. 111. IV. V. VI. VII. VIII. IX. X.

Oxirane Derivatives of Aldoses NEIL R. WILLIAMS Introduction.. ....................................................... Synthesis.. .......................................................... Reactions ............................................................ Characterization ..................................................... V. Tables of Aldose Oxiranes ............................................

I. 11. 111. IV.

V

109 110 120 170 172

CONTENTS

vi

2. 5.Anhydrides of Sugars and Related Compounds J . DEFAYE 1. Introduction

.........................................................

I1. Methods of Formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

III . heactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV. Utilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V . Table of Properties of 2,5-Anhydrides of Sugars, Alditols, and Aldonic Acids .............................................

181 183 210 215 219

Alditol Anhydrides S. SOLTZRERG

.

......................................................... 229 Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 236 Physical Properties ................................................... 250 Reactions ........................................................... 256 Uses ................................................................. 267 Tables of Properties of the Anhydrides and Their Derivatives . . . . . . . . . . 271

1 Introduction

I1 . 111. IV . V VI .

.

The Sugars of Honey

I . R . SIDDIQUI I . Introduction .......................................................... I1. Honey Monosaccharides .............................................. I11 . Honey Oligosaccharides .............................................. IV. Honey Polysaccharides ............................................... V . Honeydew ...........................................................

285 289 295 306 307

Reactions of Free Sugars with Aqueous Ammonia

M . J . KORT I . Introduction ......................................................... I1. Products Obtained ................................................... I11. Isolation of Products. and Proportions Obtained ....................... IV. Mechanism .......................................................... V. Applications .........................................................

311 314 328 332 349

Synthesis of Nitrogen Heterocycles from Saccharide Derivatives HASSANEL KHADEM I . Introduction

.........................................................

351

I1. Formation of Three-membered Nitrogen Heterocycles . . . . . . . . . . . . . . . . . . .352 I11. Formation of Five-membered Nitrogen Heterocycles ..................... 357

. .

IV Formation of Six-membered Nitrogen Heterocycles ...................... 394 V Formation of Higher-membered Nitrogen Heterocycles ................... 404

CONTENTS

vii

Aspects of the Structure and Metabolism of Glycoproteins R . D . MARSHALLAND A . NEUBERGER I . The Nature and Occurrence of Clycoproteins .......................... I1. Carbohydrate-protein Linkages ........................................ I11. Polypeptide Chains Carrying More than One Type of Carbohydrate-peptide Linkage ....................................... IV. Heterogeneity in Glycoproteins ....................................... V. The Size of the Carbohydrate Moieties in Glycoproteins . . . . . . . . . . . . . . . . VI . Features of the Structure of the Carbohydrate Moieties of Some Clycoproteins ............................................. VII . The Biosynthesis of Glycoproteins .................................... VIII . Some Genetically Determined Diseases in Which Clycoproteins Are Implicated ........................................ IX. Concluding Remarks .................................................

AUTHOR INDEXFOR VOLUME 25 .......................................... SUBJECT INDEX FOR VOLUME 25 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CUMULATIVE AUTHOR INDEX FOR VOLUMES 1-25 ......................... CUMULATIVE SUBJECT INDEX FOR VOLUMES 1-25 .......................... ERRATA...................................................................

407 417 439 443 447 452 467 472 477 479 507 525 533 544

This Page Intentionally Left Blank

LIST OF CONTRIBUTORS Numbers in parentheses indicate the pages on which the Authors' contributions begin.

SHIRLEYC. CHURMS, Council for Scientific and lndustrial Research Carbohydrate Research Unit, Department of Chemistry, University of Cape Town, South Africa (13)

J. DEFAYE,' Institut de Chimie des Substances Naturelles, Centre National de la Recherche Scientijique, 91 Gif-sur-Yuette, France (181) HASSANEL K H A D E M , ~ Institut de Chimie des Suhstances Naturelles, Centre National de la Recherche Scientijique, Paris, France (351)

M . J. KORT, Department of Chemistry, Uniaersity of Natal, Pietermaritzburg, and Sugar Milling Research Institute, Durlian, South Africa (311) R. D. MARSHALL,Department of Chemical Pathology, St. Mary's Hospital Medical School, London, W.2, England (407) A. NEUBERGER,Department of Chemical Pathology, St. Mary's Hospital Medical School, London, W.2,England (407)

I . R. SIDDIQUI,Food Research Institute, Canada Department of Agriculture, Ottawa, Ontario, Canada (285) S. SOLTZBERG, Atlas Chemical Industries, lnc., Wilmington, Delaware (229) GERALDSTRAIIS, Biochemistry Department, New York Medical College, New York, New York (53)

J. R. TURVEY, Department of Chemistry, Unioersity College of North Wales, Bangor, Caernarvonshire, Great Britain (1) NEIL R. WILLIAMS,Chemistry Department, Rirkbeck College, University of London, Malet Street, London W.C.1, England (109)

Present address: Centre de Recherche sur les Macromolecules VBgbtales, C.N.R.S., Domaine Universitaire de Grenoble, 38 St. Martin d'HAres, France. t Permanent address: Faculty of Science, Alexandria University, Alexandria, Egypt.

O

ix

This Page Intentionally Left Blank

PREFACE

With this twenty-fifth Volume, Advances in Carbohydrate Chemist r y and Biochemistry has completed its first quarter-century. This serial publication was initiated with the dual objective of presenting definitive accounts of the status of matured fields and of providing, for areas of high activity, critical evaluations that would serve as guidelines for future research. The past 25 years have seen an acceleration of research unprecedented in the history of science, and the extent to which Advances has usefully fulfilled a need, and yet provided flexibility in accommodating to change, may be judged by the frequency with which many of the older articles are still cited. Over its lifetime, Advances has developed into a permanent source of reference in organized form for practically all of the major subdivisions of knowledge in the field of carbohydrates. It has also stimulated, by means of timely articles on active and controversial areas of research, the exploration of important fields that might otherwise have been neglected or investigated in a more haphazard fashion. The breadth of coverage, as originally conceived and subsequently maintained, has allowed the discussion of carbohydrates from the viewpoints of many specializations. Structural and synthetic organic chemistry have certainly been highly influential, but biochemistry and physical chemistry have been no less important, and the techniques and ideas of agricultural chemistry, analytical chemistry, industrial chemistry and technology, microbiology, pharmacology, and many other disciplines have brought the full breadth of scientific inquiry to bear on this, the largest, class of natural products. In the present era, the idea of interdisciplinary research has become much in vogue, and it is therefore interesting to observe that the original Advisory Board for Advances, in setting a policy of studying a single major class of natural products by a broad spectrum of many classical disciplines, was instrumental in the achievement, over the years, of that very type of cooperation between specialists of different persuasions that is only now becoming properly recognized as an important trend in the future development of science. Bearing in mind the not infrequent asseverations in the past by certain classical specialists (especially, organic chemists) that the study of carboxi

xii

PREFACE

hydrates is a narrow specialization, the tenacity of the Editors and the Advisory Board in adhering to the original interdisciplinary concept is a fitting testimony to the soundness of their ideas in the face of changing opinion. The present volume was in the planning stage at the time of the death of Professor Melville L. Wolfrom, but it is one of the few volumes in the history of this Series not to have received his editorial attention. His influence always reflected the precision and accuracy that he applied in his own writings, and the present Editors will endeavor to meet these criteria. This volume includes an obituary, contributed by J. R. Turvey, of the late Professor Stanley Peat, F.R.S., who was for many years a member of the Board of Advisors of Advances, and who served as Associate Editor for the British Isles for a number of years. The separation of macromolecules by molecular-sieve techniques constitutes a major technical advance, especially for biochemists. The principles of the method and applications in the carbohydrate field are surveyed by S. C. Churms (Rondebosch). The field of X-ray crystal-structure analysis is undergoing rapid evolution because of major advances in methodology; automatic diffractometers and computerized systems for data reduction have advanced the technique to the point where the solution of many simple structures is almost routine, and the chapter by G. Strahs (New York) surveys developments since the article by Jeffrey and Rosenstein in Volume 9 of Advances. The chapter marks a transition between the era of the classical crystallographer, who determined a structure for its own sake, and that of the newer generation of crystallographers concerned with the broader implications of a coordinated plan of attack, where crystallography provides the tool rather than the objective for the study of fine points of the molecular structure of carbohydrates in relation to their conformations and biological roles. The wideranging subject of anhydrides of sugars and their derivatives is treated from the organic chemical viewpoint in three separate chapters in this volume. Because of the extensive literature on sugar anhydrides of various types, it was found impossible to treat developments in this whole area within the confines of one chapter, or even in the three chapters here presented; other aspects remain to be treated in future issues. Oxiranes (epoxides) are discussed by N. R. Williams (London), and ring-forming reactions, of aldoses, that result in the formation of 2,banhydro rings are delineated in the chapter by J. Defaye (Gif-sur-Yvette). The anhydrides of alditols are considered separately by S. Soltzberg (Delaware).

PREFACE

xiii

Modern methods of separation continue to reveal the complexity of mixtures of even simple carbohydrates in various natural sources, as demonstrated in the chapter by 1. R. Siddiqui (Ottawa) on the sugars present in honey. The reactions of sugars with ammonia and amines, a field related to important problems in the food industry, constitute a subject of continuing interest, and are treated by M. J. Kort (Pietermaritzburg). The polyfunctional nature of sugars can be exploited in the synthesis of a multitude of types of heterocyclic derivative, but the uninitiated reader may find the literature confusing and disorganized because of the plethora of structural types possible, even from simple reactions; H. El Khadem (Alexandria) has performed a valuable service by organizing the facts, fictions, and paradoxes in this domain. In the final chapter, R. D. Marshall and A. Neuberger (London) explore the recent developments in our understanding of the structure and metabolism of glycoproteins, an area at the broad frontier of much advancing knowledge in modern biochemistry. The Subject Index was compiled by Dr. L. T. Capell. A well-assorted, international representation of authorship is evident in recent volumes of Advances; the original British-American liaison on which the publication was founded has been substantially expanded to the international level. The present volume includes, in addition to contributions from North America and Great Britain, articles from continental Europe and, coincidentally, three separate chapters by authors based at different points on the African continent.

Kensington, Maryland Columbus, Ohio November, 1970

R. STUART TIPSON DEREKHORTON

This Page Intentionally Left Blank

STANLEY PEAT 1902- 1969

On the 21st of February, 1969, Professor Stanley Peat, D. Sc., F.R.S., suffered a stroke at his home in Bangor, Wale;, and he died 36 hours later without regaining consciousness. Thus ended the career of a scientist who, even as a child, displayed an unquenchable thirst for knowledge and who, all his life, regarded it as a duty to impart that knowledge to others. His parents, Ada and John H. Peat, lived at Bolden, County Durham, but Mrs. Peat was staying with her sister Alice in South Shields when their first child, Stanley, was born in 1902. John Peat was a mining engineer, but, despite his status, the family was rather poor and the care of the infant boy was handed over to his Aunt Alice and her husband, James Gibson, who were childless. Unfortunately, despite their care and attention, the child developed bovine tuberculosis when only a few months old, and had to receive hospital treatment for some time. Possibly as a result of this illness, the child was left with a permanent curvature of the spine. This disability was to have a profound effect on the development of Stanley Peat, since, denied an outlet for his energies in active sports and games, he was thrown back onto his mental resources for amusement and interest. The first consequence of this childhood illness was that Stanley was a rather frail child, requiring much attention from his aunt; this she gave unstintingly, often at the expense of her husband. Both Gibsons were keen members of the Salvation Army, the husband being a musician in the Army Band. The child was thus raised in a household where both Christianity and music were an important part of everyday life, and both were to influence him for the rest of his life. His frailty also resulted in a delayed start to his schooling; he did not attend a formal school until he was eight years old. Fortunately, he was blessed with an inquiring mind and this, coupled with the enforced idleness, resulted in his learning to read and write before he went to school. He became an avid reader, and the local library became an important point of call whenever he went out. From the library he was able to borrow books on travel, science, and a wide 1

2

J. R. TURVEY

variety of other subjects. Two books in particular fired his imagination, one on chemistry and the other on photography. At the age of eight, he started a chemistry set with which to experiment, and, by the age of ten, he was developing and printing his own photographs in an improvized dark-room. This love for experimental chemistry and photography persisted throughout his schooldays and into College, when he still retained his private laboratory and dark-room at home. The lack of formal schooling in his early days had, however, left large gaps in his education which, in spite of his thirst for knowledge, he had been unable to fill by reading alone. Mathematics was a difficult subject to him and was to remain so for the rest of his life. Once at school, Stanley Peat rapidly showed that he was a gifted pupil. He was never happy until he felt that he really understood a subject and, unable to follow active outdoor pursuits, he tended to devote much of his spare time to study and general reading. Not that he was completely debarred from all childish games; he used to play with the local children, and indulged in the usual boyish pranks. At the age of ten, he had returned to live with his parents at Station Road, Walker-on-Tyne, a mining community, and, as he subsequently used to relate with great glee, he was chased by the local policeman for “scrumping” apples from a nearby orchard. One other incident from this period is worth recording, as it throws some light on the boy’s character. His maternal grandfather, then aged 60, was only semi-literate, and young Stanley took it upon himself to help his grandfather to write. Armed with a penny exercise book, a present from Stanley, the grandfather practised his writing under the watchful eye of his grandson. He made such rapid progress, that, within a short time, he was attempting to write an adventure story. This exercise book, with its spidery writing and frequent misspellings, was treasured by Stanley Peat all his life. From Tyne Dock Grammar School, which he attended for some years, Stanley won a Scholarship in 1915 to the Rutherford College Boys’ School in Newcastle. This school had a long tradition of excellence in its teaching of science subjects, and possessed some extremely able and inspiring science masters. Among these was William Carr, the chemistry master, who came to regard Stanley as one of his favorite and most able pupils. Certainly, the young boy spent nearly all his spare time studying, and he was regarded by his schoolfellows as a pleasant but extremely serious young fellow. In 1917, further hospital treatment for his back caused him some pain, but did not prevent him from winning the Form Prize at school. He obtained his School Certificate in 1919, and proceeded to the Sixth Form, where

OBITUARY -STANLEY PEAT

3

he specialized in Chemistry and Physics. In 1921, he was awarded the Higher School Certificate, with distinctions in chemistry and physics; at the same time he won a State Exhibition, an Entrance Exhibition, and an Earl Grey Memorial Scholarship to study at Armstrong College (now the University of Newcastle-upon-Tyne). It is interesting that fellow pupils of Peat at Rutherford College included Professor E. E. Aynsley, W. Charlton, R. Chirnside, and Sir James Taylor. At Armstrong College, Peat read for Honours in Chemistry, with physics and mathematics as ancillary subjects. Once again, he proved to b e an outstanding student, becoming Senior Pemberton Scholar in 1922 and Saville Shaw Medalist in 1923, and winning the Friere-Marreco Medal and Prize in 1924, when he graduated with First Class Honours in Chemistry. In addition to his normal studies at Armstrong College, Peat also found time to pursue a course of lectures and practical work in physiology, a fact that was to influence his subsequent career profoundly. The Professor of Chemistry at that time was W. N. (later, Sir Norman) Haworth who, with E. L. (later, Sir Edmund) Hirst, was carrying on pioneer research work, originated with Sir James Irvine at St. Andrews, on the use of methylation in the study of the ring structures of sugars. Both Peat and Charlton, who had graduated with him, were invited to conduct postgraduate work in Haworth’s research school. While continuing to live with his aunt in South Shields, Peat joined this enthusiastic group of research workers, and rapidly immersed himself in the work on methylation and in the controversies surrounding it. An insight into his character comes from Professor R. Spence, then a junior student and a fellow commuter on the train from South Shields to Newcastle: “ H e was rather senior to me, but h e was completely unpatronizing in his friendship. This readiness to help a younger man was, I am sure, strongly characteristic. He offered his collection of chemicals and apparatus to me when he left Newcastle . . . . Although a loyal co-worker, Stanley could, on occasion, make a pungent comment on the polemics in which Haworth was involved.” In 1925, Haworth was invited to the Chair of Organic Chemistry at Birmingham University, and h e took with him Stanley Peat (as his research assistant), J. Avery, W. Charlton, E. Goodyear, A. Learner, and V. Nicholson. Peat, Charlton, and Learner were in the same lodgings in Birmingham until they were all awarded the Ph. D. degree in 1928. Stanley (or, as h e was more generally known, “Sammy”) Peat did not have many interests outside his work, not that there was much free time after an average 12-hour day in the laboratory. An

4

J. R. TURVEY

occasional hand at bridge and listening to classical music were his chief amusements outside the laboratory. However, while waiting for experiments to “come to the boil,” Peat would indulge in lengthy and fierce arguments on every possible subject with Learner and anyone else who cared to participate. It was usually Peat who took the orthodox viewpoint, which he would argue both logically and forcefully, while Learner upheld the revolutionary view. It is a measure of their characters that, though they argued vociferously they never quarrelled. They found common ground in reading the whole of Ibsen’s and Shaw’s plays, and in inviting lecturers on Psychoanalysis or Comparative Religion to address them on Sundays. It is difficult to assess Peat’s attitude towards religion. The Salvation Army influence of his childhood had been modified in his schooldays by regular attendance at a Methodist chapel with his mother. He had an exceptional knowledge of the Bible, being able to recite long passages from memory and, when older, he took pains to study most of the world’s major religions. In spite of this knowledge, he never became an ardent member of any particular faith; he was too full of intellectual curiosity to accept blindly any dogma or principle that he could not subject to a scientific test. In his later dealings with colleagues and students, however, he was to display many Christian attributes, tolerance and understanding of their viewpoint, and a willingness to temper the wind to the shorn lambs among his undergraduates. In those early days, Stanley Peat’s contribution to carbohydrate chemistry was allied completely with Haworth’s pioneering work on the constitution and ring structures of sugars. His first paper (with Haworth) in 1926 was a revision of the structural formula of D-glucose, in which it was established that the known methyl P-D-glucoside, and, hence, probably D-glucose itself, existed in the pyranoid ring-form. In 1926 also, his name appeared on two papers dealing with the structure of maltose, in which it was unequivocally shown that the linkage was (1-4) between two D-glucose residues, both of which were in the pyranoid ring-form. This was the “classical” type of research which was fast making Haworth’s school the leading center of carbohydrate chemistry in Europe. In yet another paper, the importance of the synthetic approach to the structure of sugar derivatives was underlined by the preparation of D-glucono- and D-mannono1,5-lactones from D-arabinose, by using a cyanohydrin synthesis. Following the award of his Ph. D. degree in 1928, Peat was offered the post of lecturer in Biochemistry in the Physiology Department of the Medical School at Birmingham University. This department

OBITUARY -STANLEY PEAT

5

wished to-expand its teaching on the biochemistry side, and Professor de Burgh Daly had appealed to Haworth to find him a suitable lecturer. Knowing of Peat’s interest in physiology and of his potential as a teacher, Haworth nominated him. Peat threw himself with enthusiasm into the task of gathering material for his lectures to the medical students and of preparing experiments suitable for large classes in very limited laboratory space. To this problem of overcrowding, Peat brought an orderliness and precision that ensured that each student was able to carry out his allotted experiments without hindrance. A few well-chosen words from Peat were enough to quell the most mischievous prankster, thus ensuring the efficient working of the class. Accounts of such incidents were often related with obvious humor by Peat to his friends. To those who did not know him well, he appeared a shy and very earnest young man. Only with his colleagues and the many friends he made in the Founders’ Room at the Edmund Street branch of the University did he show his tremendous sense of humor, his penchant for intellectual argument, and his deep humanity. During this period, h e also taught himself German, and he subsequently made frequent trips to Austria and Germany. On these trips, h e was able to indulge to the full a growing passion for grand opera and classical music. In Birmingham, he started playing golf, using specially shortened clubs, and was frequently to b e seen practising on the University playing fields by the Bristol Road. His greatest enjoyment, however, came from the concerts given by the Birmingham Symphony Orchestra, from the plays at the local “Rep” theatre, and from learning to play the mandolin. Although a heavy teaching-load curtailed his research activities, in 1931 he collaborated in an investigation of a case of phenol poisoning and, with MacGregor, began a series of investigations on the physiological role of histamine in the animal body, subsequently publishing three papers on aspects of this topic. This experience with biological systems and the metabolism of compounds did much to influence his later work. In 1934, Haworth invited Peat to return to the Chemistry Department as a lecturer. Peat thus joined a team which, during the next few years, was to include such eminent workers as E. L. Hirst, J. K. N. Jones, E. G. V. Percival, Fred Smith, M. Stacey, and L. F. Wiggins. I n 1936, Haworth divided the direction of his rapidly growing research school among Peat (plant polysaccharides and amino sugars), Smith (plant gums), and Stacey (polysaccharides of microorganisms); these “compartments” were not, however, “water-tight,” and fruitful collaboration between the groups frequently occurred

6

J. R. TURVEY

and often led to joint publications. Under Haworth’s guidance, Peat began to develop two main lines of research. The first was a continuation and extension of the methylation method for investigating the structure of polysaccharides. Over the next few years, there appeared a steady stream of papers based on the use of this technique that described the constitution of agar, a-amylodextrin, cellulose, dextran, and xylan. Cellulose was the polysaccharide to which he frequently returned; its end-group assay by methylation, the importance of excluding oxygen during its methylation, and the properties of oxycellulose and hydrocellulose were among the aspects that he investigated. The second line of research was a detailed study of the formation and reactions of anhydro sugars. In 1938, with Wiggins, he described the base-catalyzed elimination of the p-tolylsulfonyloxy group from with consequent methyl 3-O-p-toly~sulfonyl-fl-~-glucopyranoside, formation of various methyl anhydrohexosides. The fact that Walden inversion frequently accompanied this type of elimination was stressed, and the opening of the anhydro ring by alkaline reagents, again with inversion, was described. Other p-toluenesulfonic esters were then investigated, and the opening of the anhydro ring by methanolic ammonia to produce aminodeoxy sugars was reported. Peat readily appreciated the importance of these reactions in the synthesis of the rarer sugars, and, particularly, in the preparation of amino sugars. The synthesis of “chitosamine,” and the proof that this sugar is 2-amino-2-deoxy-~-glucose,followed from this work. Also of importance was the recognition, by Peat, Hands, and W. G. M. Jones, that the labile sugar present in agar is 3,6-anhydro-~-galactose. Much of this work and later contributions were elegantly summarized by Peat in an article in Volume 2 of Advances in Carbohydrate Chemistry. During this period, he was also a regular contributor to the Annual Reports of the Chemical Society. At this time, Haworth asked Peat to take over from him the teaching course to first-year students, a course which both regarded as the most important of the syllabus. Peat was desirous of introducing into the teaching the new concepts of “electronic mechanisms” as a help to students in understanding why chemical reactions occur. He designed such a course and, in his meticulously careful way, delivered it with obvious enthusiasm. The students responded to this new approach, and all who attended the course agreed that it was outstanding in its clarity and mode of presentation. This avenue to the teaching of elementary organic chemistry was, with suitable modification, continued by Peat for nearly twenty years in Birmingham and Bangor.

OBITUARY- STANLEY PEAT

7

In the teaching laboratory, he strove to instil into students the importance of neatness, accuracy, and precise observation. An accomplished experimentalist himself, and second only to Fred Smith in his ability to produce crystals from syrups, he would not tolerate untidy or careless work. His private life, which had pursued an even tenor for several years, was also due for a change. During his days in the Medical School at Birmingham University, he had gone to lodge in Station Road, Harborne, with a Mr. and Mrs. Barnes. Mr. Barnes, a dental surgeon, had been forced by ill health to curtail his practice and, in 1937, to retire completely. Stanley Peat bought a house in Edgbaston, and the Barnes family moved in with him. Their daughter Elsie, several years younger than Stanley, would sometimes accompany him on his frequent visits to concerts or the theatre. In the spring of 1939, following a sharp bout of influenza during which Elsie was his faithful nurse, he proposed to her, and they were married in the summer. Peat threw himself wholeheartedly into the business of being a married man; and he planted a garden, but his early efforts were not an outstanding success. To his great joy, a daughter, Gillian, was born to them in 1940, and another daughter, Wendy, in 1942. He derived tremendous pleasure from the company of these children. Unfortunately, the Second World War had brought an end to the happy state of the Chemistry Department. Haworth had agreed to turn much of the research potential in his department over to a study of uranium compounds, a project contributing in no small measure to the production of the atomic bomb. After six months of research on uranium carbonyls, Peat found that he could generate little enthusiasm for the work itself, and none at all for the project to which it was contributing. He persuaded Haworth that he could contribute more to the war effort by carrying out selected research in carbohydrates, with a view to increasing food production. With Haworth’s approval, he took over from M. Stacey a project aimed at producing D-glucose directly from potatoes, and saw this process through to the stage of commercial production. His knowledge of cellulose and its derivatives was called for as a member of the Cellulose and Cordite Panel of the Ministry of Supply, and his interest in agar was used on the Committee charged with finding from among the British seaweeds a suitable substitute for Japanese agar. The eventual collection of seaweeds and their processing to afford a product acceptable to microbiologists, even if somewhat inferior to agar, was a direct result of the work of this Committee. In addition, members of his research team undertook work on Asdic recordings for the Admiralty.

8

J. R. TURVEY

With all this war work progressing, a small group was still able to carry on with the normal carbohydrate research. Papers on starch, D-ghCUrOnOlaCtOne, bacterial levans (with Stacey), and the polysaccharide associated with beta-amylase, all appeared in the Journals. It was in 1940 that a series of papers by C. S. Hanes appeared that were to change completely the ultimate direction of Peat’s research. These papers dealt with the synthesis from D-glucosyl phosphate of a sparingly soluble “starch” by a plant enzyme, phosphorylase. Hanes prepared a batch of this starch and sent it to Haworth for a chemical investigation. With Peat and Heath, Haworth showed that the starch was a short-chain amylose, and that it did not contain any of the branched compound, amylopectin. From that time on, Peat devoted more and more of his time to a study of starch, and particularly to its enzymic synthesis and degradation. A study of the amylolytic degradation of starch was followed in 1944 by the first synthesis of amylopectin, which led directly to the discovery of the Q-enzyme. This work, in which E. J. Bourne played a significant part, added impetus to the search for new methods that would separate whole starch into amylose and amylopectin fractions. One such method, involving the use of thymol or cyclohexanol, was the subject of a patent taken out by Haworth and Peat. A survey of the starches of many plant species followed, and this led to the recognition that certain varieties of pea produce a starch containing an abnormally high proportion of amylose. Many starch derivatives were also prepared, and one in particular, sodium starch glycolate, has since assumed importance as an indicator for iodimetric analyses. In 1946, Peat was made an Associate Editor of a new periodical, Advances in Carbohydrate Chemistry, serving in this capacity until 1954, and he continued as a member of the Board of Advisors for the British Isles until his death. The end of the war brought some changes to the University of Birmingham. Peat became a Reader in Organic Chemistry, and was awarded the D. Sc. degree. The quality of his researches was also recognized in 1948 by his election to Fellowship in the Royal Society, and, shordy afterwards, by his appointment to the Chair of Chemistry at the University College of North Wales, Bangor. He chose two young lecturers to accompany him to Bangor, W. J. Whelan and F. H. Newth, together with two research students. With the enthusiastic co-operation of Whelan and Newth, he soon established a thriving research school despite crowded conditions and lack of finances. With Whelan, further research on starch-metabolizing enzymes was instigated, while Newth concentrated his efforts on anhydro sugars and reaction

OBITUARY - STANLEY PEAT

9

mechanisms of the acetylated glycosyl bromides. The purification and mode of action of several enzymes were studied, and the notable discoveries of R-enzyme, Z-enzyme, and D-enzyme followed. In 1950, Whistler and Durso described the separation of mono-, di-, and tri-saccharides from each other by chromatography on charcoal (activated carbon). This method was immediately tested by Peat and Whelan, and was shown to be outstanding for the resolution of oligosaccharide mixtures, such as are obtained b y enzymic or partial acid hydrolysis of polysaccharides. By 1952, carbon columns were in operation throughout the Bangor chemical laboratories, and the purified oligosaccharides so obtained were being used to elucidate the action patterns of the enzymes. A series of papers on the enzymic synthesis and degradation of starch, the first appearing in 1945 and the last (Part 24) in 1959, represented a major contribution to this field of study and rightly won Peat world renown. Of equal importance, however, was the realization by Peat and Whelan that partial hydrolysis of a polysaccharide with acid, followed by separation on carbon columns and identification of the products, is a supporting method to methylation for investigating polysaccharide structures. The linkages present in the polysaccharide are usually retained intact in the derived oligosaccharides, and, unlike the methylation procedure, the method can give valuable information as to the sequence of linkages in the polysaccharide chain. The method was applied with outstanding success to laminaran, to the glycogen, glucan, and mannan of yeast, and to lichenan, isolichenan, and Aoridean starch. By 1955, Peat was extremely busy, not only with research and teaching but also with the administration of a vigorous department. Wishing to take his share of the burden of other duties, he served for a period as Chairman of the Carbohydrate Nomenclature Committee of The Chemical Society (having already served on the Council and Publications Committee of this Society); also, he had been Dean of the Faculty of Science at Bangor, and was on the College Council. He was a consultant to the Brewing Industries Research Association, and to the Bakers and Flour Millers Research Association. He was also a member of the Scientific Advisory Committee of the Institute of Seaweed Research. When Whelan left Bangor in 1956, it was decided that much of the work on the starch enzymes should be continued by him at the Lister Institute. With J. R. Turvey and P. F. Lloyd (who replaced Whelan), Peat decided to branch into a line of research that he had touched upon previously but now wished to develop more fulIy. The structure of

10

J. R. TURVEY

agar, and, more particularly, the role of sulfuric hemi-esters in certain algal polysaccharides had always intrigued him. With Turvey, a systematic study of the preparation, properties and reactions of sugar sulfates was commenced. This work was coupled with an examination of the galactan sulfates of red algae. Lloyd, who had brought to Bangor an interest in amino sugars and the sulfated polysaccharides of animal tissues, commenced work on some aspects of amino sugar chemistry and on the sulfatase systems present in marine molluscs. The isolation of L-galactose 6-sulfate from an algal polysaccharide, and (with D. A. Rees), from a seaweed of an enzyme system which (as was later established by Rees) could convert galactose 6-sulfate residues in a polysaccharide into 3,6-anhydrogalactose residues, constituted verification of ideas put forward by Peat 15 years previously. In 1959,The Chemical Society recognized Peat’s contribution to the field of investigation of carbohydrate structure and metabolism by inviting him to give the Hugo Miiller Lecture, an invitation that he regarded as a signal honor. He made plans for this lecture, but a heart attack in September put him into hospital. His recovery was very slow, and it was almost two years before he was again able to visit his department regularly. Throughout his convalescence, he encouraged visits to his home in Penrhos by all his staff and research students. Being forbidden to exert himself physically, he took great delight in conversation. His wide knowledge and love of music, of literature (especially Dickens, selected volumes of whose works he read every Christmas Season), and of the history of ancient civilizations provided a constant source of topics for discussion. Once installed in the study of his large house with its beautiful garden, the visitor found it difficult to leave, and hours of delightful conversation ensued. It was during these conversations that he displayed to the full his wonderful sense of humor, his deep interest in his fellow man and, although he claimed to be an agnostic, his essentially Christian outlook on life. When he was again able to return to the department, uncertain health forced him to give up lecturing and to play a less active part in research, both of which he did reluctantly. The supervision of his research students was left entirely to Turvey and Lloyd, but he was always ready to give advice and encouragement whenever these were sought. This fatherly interest in the research work of his department extended to all sections, and he also continued to play an active part in the design of suitable teaching courses for the undergraduates. He tended to view with some caution the new physical methods that were being used by research groups in his department.

OBITUARY - STANLEY PEAT

11

While conceding that much information of a preliminary nature could be obtained by “knob-twiddling” -as he jokingly called it -he stoutly maintained that results obtained by such methods should be treated with caution until reinforced by more tangible evidence. Over the last few years, although still retaining his authority as Head of the Department, he came to rely more and more on the willing co-operation of his colleagues in ensuring the smooth running of the department. Since coming to Bangor he had striven to obtain funds for a new Chemistry Building, and these were granted in 1961. He then gave up all interests outside the department, and devoted much of his energies to the planning and designing of the new building, which came into operation in 1965 and is now one of the showplaces of the College. The research work in his group continued; the investigations on algal polysaccharides and of their degradation by bacterial enzymes were of particular interest to him. Further work on sugar sulfates, an investigation of heparin, and the unravelling of the complexities of molluscan sulfatases were also topics in progress at the time of his death. At home, he preferred relaxing with his books, or selectively viewing television, which provided some consolation because of his love of music and drama. The advent, over the years 1962 to 1964, of four grandchildren was a source of much pride to him, and he derived great pleasure from their company in the ensuing years. He was fascinated by the processes involved in learning in the very young, and would often comment on some observation he had made while watching his grandchildren. In September, 1968, he suffered another heart attack, from which he was still convalescing when he had his fatal illness. Preparations for a Special Issue of the journal Carbohydrate Research in his honor had been made early in 1968, and the issue was in the printers’ hands at the time of his death. Equally tragic was the fact that a Meeting of the Carbohydrate Discussion Group of The Chemical Society had been arranged in his honor in Bangor for Easter, 1969. It was unfortunate that his premature death, a few months before he was due to retire, should have deprived him of the pride and pleasure that these two events would undoubtedly have afforded him. In his address as Dean of the Faculty of Science to the Court of Governors of the College in 1953,Stanley Peat said: “My four predecessors in the Chair of Chemistry-Dobbie, Orton, Simonsen, and E. D. Hughes - were each in their turn elected to the Royal Society, a sufficient indication of the value of their research work and, because of that, of their qualification as university teachers. The influence of

12

J.

R. TURVEY

these men has extended far beyond the boundaries of our College.” The name of Stanley Peat can now, with complete justification, be added to this list of distinguished chemists.

J. R. TURVEY The help of the following people is gratefully acknowledged: Mrs. E. Peat, Dr. J. Avery, Professor E. E. Aynsley, Professor E. J. Bourne, Dr. A. Learner, Professor R. G. S. MacGregor, Mrs. G. Shepherd, Professor R. Spence, Professor M. Stacey, Dr. D. H. Strangeways, and Sir James Taylor.

GEL CHROMATOGRAPHY OF CARBOHYDRATES*

BY SHIRLEYC. CHURMS Council for Scientijic and Industrial Research Carbohydrate Research Unit, Department of Chemistry, University of Cape Town, South Africa I. Introduction.. ........................................................ 11. Types ofGel ........................................................... 111. Theory of Gel Chromatography ......................................... 1. Principles and Definitions ............................................ 2. Separation by Gel Chromatography. .................................. 3. Determination of Molecular Weight. ................................. IV. Application to Carbohydrates ........................................... 1. Sugars ............................................................. 2. Polysaccharides .................................................... 3. Miscellaneous Carbohydrates ........................................ V. Value of the Technique ................................................

13 14 16 16 17 21 31 31 35 47 51

I . INTRODUCTION Gel chromatography (also known as gel filtration, gel-permeation chromatography, or molecular-sieve chromatography) is based on the decreasing permeability of the three-dimensional network of a swollen gel to molecules of increasing size. If a solution containing a mixture of solutes of different molecular sizes is passed through a column packed with a suitable gel, the smaller molecules penetrate farther into the gel pores than do the larger, and they are therefore retained for a longer time on the column. The solutes are thus eluted in the order of decreasing molecular size. Consequently, this procedure affords a rapid, relatively simple method for separating substances that differ in molecular size, or for fractionating polymers, such as polysaccharides, having broad molecular-weight distributions. Because mild conditions are used, the technique is particularly useful for labile biological materials. Also, as the order of elution of a series of similar substances from a gel column is governed largely by mo*This review owes much to the guidance and encouragement of Professor A. M. Stephen, to whom the author expresses her gratitude. The financial support of the South African Council for Scientific and Industrial Research is also gratefully acknowledged.

13

14

SHIRLEY C. CHURMS

lecular weight, gel chromatography provides a means of determining molecular weights of polymers. The technique has been the subject of a detailed and extensive general review.’

11. TYPESOF GEL Gel chromatography originated in 1956 with the work of Lathe and Ruthven,2 who achieved some degree of separation of tri-, di-, and mono-saccharides on a column of potato starch. Owing to the disadvantages attendant on the use of starch in gel chromatography, namely, its high resistance to flow, its instability, and its ill-defined composition, starch was soon superseded by artificially cross-linked dextran gels, the use of which was first reported in 1959 by Porath and F10din.~Dextran gels, commercially available under the name S e p h a d e ~are , ~ ~now widely used in gel chromatography; these gels are hydrophilic. A lipophilic gel, namely, Sephadex LH-20, prepared by alkylation of most of the hydroxyl groups in dextran, is also available. The use of agar in gel chromatography was first reported, in 1961, by Polson.‘ Owing to their more open structure as compared with dextran gels, agar gels are capable of fractionating much larger molecules. However, agar has the disadvantages of ill-defined composition and a strong tendency to adsorb basic substances, owing to the presence of charged groups in the agaropectin component. The agarose component, which is a neutral, linear polymer consisting of alternate residues of D-galactose and 3,6-anhydro-~-galactose,~ has been found superior to native agar as a medium for gel chromatographyBsAgarose gels are now commercially available under the names Sepharose- and Sagarose.’ Another hydrophilic gel that has been successfully used in gel chromatography is the poly(acry1amide) type, first introduced in (1) L. Fischer, in “Laboratory Technique in Biochemistry and Molecular Biology,” P. S. Work and E. Work, eds., Wiley-Interscience, New York, N. Y., 1969,Vol. 1, Part 11, pp. 157-391. (2) G. H. Lathe and C. R. J. Ruthven, Biochem.], 62,665(1956). (3)J. Porath and P. Flodin, Nature, 183,1657(1959). (3a) Pharmacia Fine Chemicals, Uppsala, Sweden. (4)A. Polson, Biochirn. Biophys. Acta, 50,565(1961). (5)C . Araki, Bull. Chem. Soc.Jap.,29,543(1956). (6)S.Hjerten, Biochim. Biophys. Acta, 53,514(1961). (7)Seravac Laboratories, Cape Town, South Africa, and Maidenhead, Berkshire, England.

GEL CHROMATOGRAPHY O F CARBOHYDRATES

15

1962 by Lea and Sehon8 and simultaneously by Hjerten and Mosb a ~ h These .~ gels, prepared by polymerization of acrylamide in the presence of N,N'-methylenebisacrylamide, are available commercially (Bio-Gel P series).ln Gels that swell in organic solvents (for example, polystyrene gels) are required for gel chromatography of hydrophobic macromolecules. Such gels have found wide application in the fractionation of synthetic polymers.11*12 Cross-linked polystyrene containing hydrophilic groups (Aquapakl3) has now become available, permitting the fractionation of water-soluble materials, such as dextrans, on the polystyrene medium.I4 The use of ion-exchange resins to fractionate oligosaccharides15 is possibly another example of molecular-sieve chromatography of hydrophilic solutes on a polystyrene matrix, although, in this instance, other factors may also be involved in the separation. The use of porous glass as a support for molecular-sieve chromatography was suggested by H a l l e P in 1965, and this material, now available in bead form (Bio-Glas'O), has proved useful in the fractionation of solutes of high molecular weight,"*18particularly at high temperat u r e at ~ which ~ ~ organic gels are not suitable. The rigidity and resistance to thermal, chemical, or bacteriological degradation, which are advantages of the use of porous glass as a medium for gel chromatography, are also characteristics of porous silica beads.20 These beads, which have become commercially available ( P o r a ~ i l ' ~are ) , now finding increasing application in gel chromat ~ g r a p h y . ~ ~Certain -*~ polymers [for example, poly(viny1 alcohol)] (8) D . J. Lea and A. H. Sehon, Can.J.Chem., 40,159 (1962). (9) S. Hjerten and R. Mosbach, Anal. Biochem., 3,109 (1962). (10) Bio-Rad Laboratories, Richmond, California, U. S. A. (11) M. F. Vaughan, Nature, 188,55 (1960). (12) J. C. Moore, J . Polym. Sci., Part A, 2,835 (1964). (13) Waters Associates, Framingham, Massachusetts, U. S. A. (14) K. J. Bombaugh, W. A. Dark, and R. N. King,J. Polym. Sci., Part C,21, 131 (1968). (15) S. A. Barker, B. W. Hatt, J. F. Kennedy, and P. J. Somers, Carbohyd. Res., 9, 327 (1969). (16) W. Haller, Nature, 206,693 (1965). (17) S. A. Barker, S. J. Crews, and J. B. Marsters, Symp. Biochem. Eye, Schloss Tutzing (1966). (18) G . Meyerhoff, Angew. Makromol. Chem., 415,268 (1968). (19) J. H. Ross and M. E. Casto, J. Polym. Sci., Part C , 21, 143 (1968). (20) A. J. de Vries, M. LePage, R. Beau, and C. L. Guillemin, Anal. Chem., 39, 935 (1967). (21) M. LePage, R. Beau, and A. J. de Vries,J. Polym. Sci., Part C , 21,119 (1968). (22) S . A. Barker, B. W. Hatt, J. B. Marsters, and P. J. Somers, Carbohyd. Res., 9, 373 (1969). (23) S. A. Barker, B. W. Hatt, and P. J. Somers, Carbohyd. Res., 11,355 (1969).

16

SHIRLEY C. CHURMS

tend to be irreversibly adsorbed on porous silica, but a deactivated form of this support has been developed for use in such worksz4 The most suitable media for gel chromatography of carbohydrates are poly(acry1amide) gels, hydrophilized polystyrene, and porous glass or silica beads. Dextran and agarose gels have the disadvantage of being themselves carbohydrates, so that the effluent solution is liable to contamination with carbohydrate material from the gel. Sephadex and agarose have, nevertheless, been extensively used in this field, but are now gradually being superseded by the newer, non-carbohydrate media. 111. THEORY OF GEL CHROMATOGRAPHY

The theory of gel chromatography has been comprehensively reviewed by DetermannZ5and Altgelt.zsIn the present article, emphasis is laid upon the fundamental aspects most likely to arise in carbohydrate chemistry. 1. Principles and Definitions

In a column packed with a swollen gel, two solvent phases may be distinguished, one within the gel (the stationary phase), and the other outside (the mobile phase). The volume of solvent in the gel is known as the internal volume, designated V,, and the volume of solvent outside the gel particles is the void volume of the column, designated V,. A solute will distribute itself between the two phases to an extent measured by the distribution coefficient Kd, a constant determined by the nature of the solute, the solvent, and the gel, but independent of column ge~metry.~’ The volume of solvent required for eluting the solute from the column in maximum concentration is called the elution volume and is designated Ve; it is equal to the sum of the void volume and the volume of the stationary phase available to the solute, given by & Vi. Hence,

ve= v, + K d - v,,

(1)

from which may be derived the relationship

Kd = w e - V0)lVt (24) K. J. Bombaugh, W. A. Dark, and J. N. Little, Anal. Chem., 41,1337 (1969). (25) H. Determann, “Gel Chromatography” (English Edition), Springer-Verlag New York Inc., New York, N.Y., 1968. (26) K. H. Altgelt, Adoan. Chromatogr., 7,3 (1968). (27) P. Flodin,]. Chromatogr., 5,103 (1961).

GEL CHROMATOGRAPHY OF CARBOHYDRATES

17

If the solute molecules are so large that they are completely excluded from the gel, Kd = 0. In this case, the solute passes through the column entirely in the mobile phase, and its elution volume equals the void volume V,. Small solute molecules that can freely penetrate the gel pores will have a Kd value of 1; here, V, = V, Vi. Between these two extremes lie all solutes that can enter the gel phase to various, limited extents; these will have Kd values lying between 0 and 1, and elution volumes between V, and V, Vi, with both Kd and V, increasing as the molecular size of the solute decreases. The value of Kd cannot exceed unity if the molecular-sieve effect is the sole factor involved, but higher Kd values are found where adsorption effects are superimposed upon the molecular-sieve mechanism. For example, basic substances are adsorbed by ionexchange on gels having ionizable groups (for example, Sephadex ~ ~ , aromatic ~~ and and agar) in solvents of low ionic ~ t r e n g t h , and ~ ~a .mechanism ~~ believed3' heterocyclic compounds are a d s ~ r b e dby to involve the delocalized welectrons present in such compounds.

+

+

2. Separation by Gel Chromatography

The degree of separation of two solutes on any gel depends upon their Kd values on that gel. If one solute has a distribution coefficient K d ' , and the other, K d t ' , their elution volumes on a given column will respectively. The two solutes will, be (v, K d rVi) and (V, K d t t Vi), therefore, emerge from the column separated by a volume V i( K d fthis indicates that separation is improved by increasing Vi, so a column of large volume is desirable. The critical factor is, however, the difference in Kd between the two solutes. This difference depends on the size of the gel pores, which must be of such a size as to differentiate between the appropriate molecular sizes to the greatest possible extent. Optimal separation, on a gel column, of substances having small differences in molecular size is obtained if the gel pores are of a size comparable with that of the substances A gel of given porosity is, therefore, effecbeing chromat~graphed.~~ tive in separating substances within a certain range of molecular size only, the sizes involved decreasing with decreasing size of the gel pores; this range of molecular sizes is known as the fractionation range for the gel.

+

+

&'I);

(28) B. Gelotte,]. Chromatogr., 3,330 (1960). (29) P. Andrews, Nature, 196,36 (1962). (30) K. Sun and A. H. Sehon, C a n . ] . Chem., 43,969 (1965). (31) J.-C. Jamon,]. Chromatog.,28,12(1967). (32) J. Porath, Biochim. E i o p h y s . Acta, 39,193 (1960).

SHIRLEY C. CHURMS

18

Fractionation ranges are conventionally given in terms of molecular weights. The ranges given by manufacturers usually apply to globular proteins. A polysaccharide having a relatively open-chain (random-coiled) structure, such as a dextran, will be larger than a globular protein of the same molecular weight, and, consequently, will be excluded to a greater extent from a gel of a given porosity. The gel will, therefore, be effective in separating dextrans of somewhat lower molecular weight as compared with proteins.33 The fractionation ranges of Sephadex gels, both for protein^^^.^^ and d e x t r a n ~ ? ~ ~~ are well established (see Table I), as are those for S e p h a r o ~ e(see TABLEI Fractionation Ranges of Sephadex Gels Fractionation range (mol. wt.)

Type of gel

Globular p r o t e i n ~ ~ ~ ~ ~ ~ Dextrans36

G-10 G-15 G-25 (2-50 (2-75 G-100 G-150 G-200

700

=s703

s 1,500

< 1,500

1,000-5,000 1,500-30,000 3,000-70,000 4,000- 150,000 5,000-400,000 5,000-800,000

100-5,000 500-10,000 1,000-50,000 1,000-100,000 1,000-150,000 1,000-200,000

Table 11), but, for other gels, only the ranges for proteins have been published by the manufacturers; the ranges for polysaccharides and other substances have to be determined by experiment. For carboTABLEI1 Fractionation Ranges of Sepharose Gels3fi Type of gel -

Agarose (%)

6B 4B 2B

4

6

2

Fractionation range (mol. wt.) Globular proteins

Dextrans

1 X lo4-4 X lo6 4 x 105-20 x 106 3 X 106-40X lo6

1 x 104-1 x 106 3 X 10’-3 X lo6 2 X 106-25X 10‘

(33) P. Andrews, Biochern.J.,91,222 (1964). (34) P. Andrews, Biochern.].,96,595 (1965). (35) K. A. Granath and P. Flodin, Makrornol. Chem., 48,160 (1961). (36) Manufacturers’ literature (1969).

GEL CHROMATOGRAPHY OF CARBOHYDRATES

19

hydrates, the fractionation ranges of the poly(acry1amide) gels BioGel P-300, P-10, and P-4 have been published (see Table III).37 TABLE111 Fractionation Ranges of Some Poly(acry1amide)Gels Type of gel Bio-Gel P-4 P-10 P-300

Fractionation range (mol. wt.) Globular proteins’O 500-4,000 5,000- 17,000

100.000-400.000

Dextrans3’

< 4,000 250-15,000 5.000- 100.000

Values of Kd in gel chromatography are not affected to any great extent by the experimental conditions. Temperature has a very small effect; Selby and Maitland3snoted that a sharp decrease in room temperature caused a slight decrease in the Kd values of certain enzymes on a Sephadex gel, and a corresponding small increase in Kd with increase in temperature from 9 to 60” was observed by Obrink and coworkers39 for fractions of the synthetic “polysucrose” Ficoll on Sephadex. The increase in Kd with increasing temperature has been ascribed39to a decrease in V i that results from shrinking of the gel at higher temperatures; this is possibly due to increased coiling of the dextran chains as a result of changes in the degree of interaction with the solvent. The temperature effect has so far been observed only with dextran gels; whether or not it is universal remains to be ascertained. In some cases, but not in others, the concentration of the sample has been found to have a small effect on the elution volume, and, consequently, on K d . Winzor and Nicho140 noted a slight increase in the elution volumes of certain proteins on Sephadex G-100 as the sample concentration was increased from 1 to 12 mg per ml, but no (weight-average such effect was observed with a dextran of molecular weight) 500,000. The effect observed with the proteins was ascribed to dependence of their migration rates on the concentration. 500,000 with An increase in the elution volume of the dextran of

aw

a,

(37) S. C. Churms and A. M. Stephen, S.Afr. Med.J . , 43,124 (1969). (38) K. Selby and C. C. Maitland, Biochem.]., 94,578 (1965). (39) B. Obrink, T. C. Laurent, and R. Rigler,J. Chromotogr., 31,48 (1967). (40) D. J. Winzor and L. W. Nichol, Biochim. Biophys. Acto, 104,1(1965).

SHIRLEY C. CHURMS

20

increasing concentration (5-20 mg. ml-l) was observed41 in studies using Sephadex G-200. This was attributed to increased void volume, due to shrinkage of the gel resulting from the higher osmotic pressure that accompanies increasing concentration of the non-penetrating solute in the mobile phase. Increase in elution volume with increasing concentration of the sample has also been noted in the chromatogthe dependence raphy of synthetic polymers on polystyrene ge1s,42,43 on concentration being greatest with samples having a broad distribution of molecular weight.42 The elution volumes of dextrans on porous silica (Porasil-D) have also been found to vary with the molecular-weight distribution at high concentrations of the sample;23this has been ascribed to mutual interaction of the solute molecules at high concentrations. The concentration dependence of the elution 10,000, and even of D-glucose, on the volumes of a dextran of poly(acry1amide) gel Bio-Gel P-10 (see Table IV), observed in this

am

TABLEIV Effect of Concentration of Sample on V, and K d of Carbohydrates on Poly(acry1amide) Gels4'

Solute

Gel

Column dimensions (cm)

D-Glucose

Bio-Gel P-10

60 X 1.2

Bio-Gel P-300

60 X 1.2

Bio-Gel P-10

60 X 1.2

Bio-Gel P-300

9OX

Dextran -

M, 10,000 1.5

Sample concentration (mg. m1-I)

1.6 16.5 1.9 19.6 1.6 10.3 18.7 3.8 8.5 19.9

62.7 64.1 70.1 70.1 30.5 31.9 33.7 100.0 100.4 100.6

0.23 0.26 0.30 0.64 0.64 0.64

laboratory,44 can probably be similarly explained. N o such effect was noted with the (more porous) Bio-Gel P-300; less interaction may be expected in gels having large pores. (41) E. Edmond, S. Farquhar, J. R. Dunstone, and A. G. Ogston, Biochem. J., 108, 755 (1968). (42) M. J. R. Cantow, R. S . Porter, and J. F. Johnson, J . Polymer Sci., Part B , 4, 707 ( 1966). (43) K. A. Boni, F. A. Sliemers, and P. B. Stickney, Amer. Chem. SOC., Dio. Polym. Chem., Preprints, 8,446 (1967). (44) S. C. Churms, A. M. Stephen, and P. van der Bij1,J. Chzomatogr.,47,97 (1970).

GEL CHROMATOGRAPHY OF CARBOHYDRATES

21

The degree of separation, or resolution, obtained in any chromatographic process depends not only on the aforementioned factors, which influence the positions of the solute peaks in the resulting chromatograms, but also on the peak widths, which depend on the efficiency of the column. Although Kd values are largely independent of experimental conditions, the column efficiency, which must be high if the narrow peaks necessary for good resolution are to be obtained, is governed by several operating variables, and, therefore, attention to these is as essential in gel chromatography as in other chromatographic processes. The effect of operating conditions on the efficiency of gel chromatography has been investigated by several w ~ r k e r s , ~ ~and , ~ ”it ~has ~ been found that the greatest efficiency is obtained with long columns having a small diameter, gels having small particle size, particles uniformly packed, small volumes of sample, and low rates of flow of the solution. With regard to the last factor, however, it should be noted that an equation of the Van Deemter type48 has been to apply to gel chromatography. The use of very low rates of flow will, therefore, result in a certain amount of peak broadening, due to the effect of longitudinal diffusion of the solute in the mobile phase, although this factor is bel i e ~ e dto~be ~ less significant in gel chromatography than in other chromatographic processes. A n d r e w ~has ~ ~found flow rates in the range of 3 to 10 ml per hr per cm* to be optimal for gel chromatography. The efficiency of gel chromatography is independent of the concentration of the sample,27except where the viscosity of the solution increases rapidly with the concentration; resolution then deteriorates with increasing concentration of the sample, owing to the peak broadening that results from the irregular flow-pattern caused b y the greater viscosity of the s o l ~ t i o n . ~ ~ * ~ ~ 3. Determination of Molecular Weight There is at present no general agreement as to the best method of relating the elution volumes and Kd values in gel chromatography (45) (46) (47) (48)

J. C. Giddings and K. L. Malik, Anal. Chem., 38,997 (1966). W. Heitz and J. CoupekJ. Chrornntogr.,36,290 (1968). J. C. Giddings, Anal. Chem., 40,2143 (1968). J. J. van Deemter, F. J. Zuiderweg, and A. Klinkenberg, Chem. Eng. Sci., 5,

271 (1956). (49) P. Andrews, Brit. Med. Bull., 22,109(1966). (50) J.L. Waters, Amer. Chem. Soc., Dlo. Polym. Chem., Preprints, 6, 1061 (1965).

22

SHIRLEY C. CHURMS

to the molecular weights of the substances being chromatographed; several different relationships have been proposed. An early worker in this field was P ~ r a t h , who, ~ l on the assumption that the gel pores are conical, derived the equation

where k is a constant, characteristic of the gel and the solute, r is the average radius of the gel pores, and a is the effective radius of the solute. For flexible polymers consisting of identical segments, the effective (gyration) radius in solution has been found52to be proportional to the square root of the molecular weight, M. Such a relationship holds, to a good approximation, for d e x t r a n ~ .Here, ~ ~ equation 3 implies a linear relationship between Kd113 and M112for solutes of this type on a given gel. This relationship was observed by Porath, who, to test the equation, used the data of Granath and F 1 0 d i n ~for ~ dextran fractions on Sephadex gels. By using a model of the gel phase in which the gel matrix was assumed to consist of straight, rigid rods that were infinitely long and randomly distributed, Laurent and Killander54derived the equation &, = exp 1 - d (rr + ~ 2 1 , (4) where Z&, is the fraction of the total gel-volume available to the solute, given by (V, - V,)/(V, - V,) [where V, is the total volume of the gel phase (including that of the gel matrix), L is the concentration of the rods in the gel, expressed as cm of rod per cm3, rr is the radius of the rods, and r, is the Stokes radius of the solute]. This equation is based on an equation derived by Ogston55 for the volume available to spherical particles in a system of this type. Assuming a value of rr of 70 nm for dextran gels (a value somewhat larger than the value accepted for the radius of linear dextran chains, since branching and crosslinking effectively thicken the chains), Laurent and Killandef14 determined L for various Sephadex gels by plotting the &, values of a series of solutes of known Stokes radius against r,, and fitting equation 4 to the experimental curves. The constants having been established, the equation was applied to experimental data from the (51)J . Porath, Pure Appl.Chem.,6,233(1963). (52)B.H.Zimm and W. H. Stockmeyer,]. Chem. Phys., 17,1301(1949). (53)K.A.Granath,]. ColZoidSci.,13,308(1958). (54)T.C.Laurent and J . Killander,]. Chromatogr., 14,317(1964). 54,1754(1958). (55)A.G.Ogston, Trans. Faraday SOC.,

GEL CHROMATOGRAPHY OF CARBOHYDRATES

23

literature on the gel chromatography both of proteins and dextrans; it was found to fit, to a good approximation, all of these data. Equation 4 relates K,, to the Stokes radius of the solute molecules, not to their molecular weight. Laurent and Granath56calculated the molecular weights of dextran and of Ficoll fractions from their K,, values on a Sephadex gel by using this equation to calculate a radius equivalent to the Stokes radius for each polysaccharide fraction, and, thence, an apparent diffusion constant for each. The sedimentation constant of each fraction was determined by ultracentrifugation, and this value and the calculated diffusion constant were substituted into the equation of Svedberg and Pedersens7 in order to obtain an apparent molecular weight for the polysaccharide fraction. Siegel and Montys8also recommended this procedure. If, however, the Stokes radius of the solute bears some simple relationship to its molecular weight, the latter may be determined directly from gel chromatography by means of the Laurent-Killander equation.s4 Sims and FolkesS9have pointed out that equation 4 may be more simply expressed in the form (-ln K,,)1/2 = a(p

+

T,)

where a and p are constants characteristic of the gel. Equation 5 predicts a linear relationship between (-ln K,u)1’2and T,, and this has been confirmed experimentally for proteins on various Sagarose gels.60 If the solute molecule is considered to be spherical, as is the case for globular proteins, r, is proportional to the cube root of the molecular weight, M. The linear relationship between (-ln Ka,)112and r, here implies that (-ln K,,)1/2 will also vary linearly with and this has been observed for proteins on Sagarose gels.61 as For dextrans, rs may be regarded as being proportional to already mentioned; therefore, a linear relationship should be observed between (-ln K,,)1’2 and M112.Data obtained by Churms and Stephens2 for the chromatography of dextran fractions on a poly(acry1amide) gel (Bio-Gel P-300) have demonstrated the validity of this relationship. (56) T. C. Laurent and K. A. Granath, Biochim. Biophys. Acta, 136,191 (1967). (57) T. Svedberg and K. 0. Pedersen, “The Ultracentrifuge,” Clarendon Press, Oxford, 1940, p. 5. (58) L. M. Siegel and K. J. Monty, Biochim. Biophys. Acta, 112,346 (1966). (59) A. P. Sims and B. F. Folkes, unpublished work, cited in Sagarose manufacturers’ literature (1966). (60) A. P. Sims, A. H. Bussey, and B. F. Folkes, Ref. 59. (61) Manufacturers’ literature (1966). (62) S. C. Churms and A. M. Stephen, unpublished work.

24

SHIRLEY C . CHURMS

Squires3 used a model of the gel phase in which the volume elements available to solvent within the gel were regarded as a combination of cones, cylinders, and crevices, and derived expressions for the volumes available to a solute of Stokes radius u in these three types of pore. Certain arbitrary assumptions regarding the distribution of solute among the different types of pore gave the following equation:

which simplifies to

where g is a constant characteristic of the gel and the nature of the solute, and T is the average radius of the gel pores. For spherical molecules, a is proportional to M113,and equation 6 may be written as

where C is a constant corresponding to the molecular weight of the smallest spherical solute that is completely excluded from the gel pores. Equation 7 predicts a linear relationship between (V,/V,)113and M113,and this was confirmed by means of experimental data taken from the literature. Squires3 claimed that equation 7 is applicable to dextrans, as well as to proteins. However, in view of the observed proportionality of a and M112(not M113)for dextrans, it is the opinion of the author that, when applied to dextrans, the Squire equation should be modified to

where C represents the molecular weight of the smallest dextran molecule that is completely excluded from the gel. A linear relationship between (Ve/V,)'I3 and M1I2 is predicted by equation 8, and this has been founds2to hold for dextran fractions on the poIy(acry1amide)gel Bio-Gel P-300. The various, theoretically derived relationships between gelchromatographic elution-parameters of macromolecules and their molecular weights are given in Table V. (63) P. G . Squire, Arch. Biochem. Biophys., 107,471 (1964).

GEL CHROMATOGRAPHY OF CARBOHYDRATES

25

TABLEV Theoretically Derived Relationships between Elution Parameters and Molecular Weight Molecular type

Relations hipa

Dextrans Globular proteins Dextrans Globular proteins Dextrans

K d I l 3 = k 1 - k2 MI12 (-ln K,,)1i2= k l k , M1I3 (-ln = kl + k, ML12 (Ve/Vo)113= k , - k, (Ve/V0)"3=k,- k, M'/'

+

References

51 54,59-61 62 63 62

"k, and k, are constants, defined differently in each equation.

A ~ k e r shas ~ ~ interpreted gel chromatography in terms of steric and frictional resistance to the diffusion of the solute in the gel pores, and, on this basis, he has used an equation originally proposed by R e n k i ~ Pfor deriving the following relationship between Kd and the Stokes radius, a, of the solute:

where r is again the effective radius of the gel pores, here assumed to be cylindrical. Values of Kd computed from this equation for solutes of known a on gels of known T were found by A c k e d 4 to agree well with experimental values. An advantage of the Ackers equation is that no arbitrary constants are involved; the only constant is I , which is a property of the gel. The value of T for a given gel can be determined experimentally by measuring the Kd values of solutes of known Stokes radius and substituting the appropriate values of Kd and a into equation 9. With the value of T known, the equation can then be used for determining the Stokes radii of other solutes from their Kd values on the gel. Here, however, no simple correlation with molecular weight is obtainable; in order that the Ackers equation may be used for determining molecular weights, it is necessary that ultracentrifugation data be combined with the data obtained by gel chrornat~graphy.~~ In addition to the various relationships based on theoretical principles, several empirical correlations of gel chromatography data with molecular weights have been reported (see Table VI). A linear correlation between V, and IogM, holding over a molecular-weight (64) G. K. Ackers, Biochemistry, 3,723 (1964). (65) E.M. Renkin, J . Gen. Physiol., 38,225 (1955).

26

SHIRLEY C.

CHURMS

TABLEVI Empirical Correlations between Elution Parameters and Molecular Weight Molecular type

Relationshipa

Proteins, polysaccharides, synthetic polymers Proteins Pol ysaccharides Proteins Dextrans Peptide hormones

V,= k,- k2 IogM

29,33,66-68

& =kl-

30 69 70-72 35 73

k2

IogM

K,, = k,- k, logM VJV, = k,- k, 1ogM V,lV,= k , - k, logM Kd=(kl/logM)- k2

References

"k,and k , are constants that are different in each case. '

range coinciding with the fractionation range for the gel, has been observed in many systems when series of similar solutes have been chromatographed on the same gel co1umn.29,33*66-68 This correlation, which is now widely applied in the determination of molecular weights by gel chromatography, has been observed for protein^,^^*^"^^ polysaccharides,M and synthetic polymed8 on S e p h a d e ~ agar,29 ,~~ p o l y ( a ~ r y l a m i d e ) ,and ~ ~ *po1ystyrene'ja ~~ gels, and may therefore be regarded as universally applicable. A linear correlation between & (or &), and logM under the same conditions has also been observed with several different solute-gel systems, for example, proteins on poly(acry1amide) gels30 and dextran fractions on S e p h a d e ~This .~~ correlation, too, is widely used in molecular-weight determinations. Several ~ o r k e r s have ~ ~ -reported ~~ a linear correlation between VJV, and logM for proteins on Sephadex gels, and Granath and F 1 0 d i n ~ ~ noted a linear relationship between VJV, and log%,, (number-average molecular weight) for dextran fractions on these gels. All of these correlations are with logM; in contrast, Sanfellipo and Surak13found that, on Sephadex G-50, & varies linearly with (logM)-' for the hormonally active proteins and peptides of the anterior pituitary gland. In an attempt to account for these empirical correlations on the (66)D. M.W.Anderson, I.C. M. Dea, S. Rahman, and J. F. Stoddart, Chem. Commun., 145 (1965). (67)A. M.del C.Batlle,]. Chromatogr.,28,82(1967). (68)S.Yamada, S. Kitahara, and Y. Hattori, Kobunshi Kagaku, 23,683(1966). (69)K.A. Granath and B. E. Kvist,]. Chromatogr., 28,69(1967). (70)J. R.Whitaker, Anal. Chem.,35,1950(1963). (71)A. A. Leach and P. C. O'Shea,]. Chromatogr., 17,245(1965). (72)H.Determann and W. Michel,]. Chromatogr., 25,303(1966). (73)P. M.Sanfellipo and J. G . Surak,]. Chromatogr., 13,148(1964).

GEL CHROMATOGRAPHY OF CARBOHYDRATES

27

basis of the various theoretical treatments of gel chromatography, Anderson and S t ~ d d a r t plotted '~ the Kd values computed by A c k e d 4 (by means of equation 9) against the logarithms of the corresponding values of aIr. The Porath treatment"l was also used for calculating values of & for known values of a h ; a value of 1.64, applicable to dextran fractions on Sephadex gels, was used for the constant k of equation 3, and the Kd values thus obtained were also plotted against log a h . Both plots gave sigmoid curves, having extended linear portions, which could be expressed as . .

where k, and k, are constants. For any solute, a is proportional to some fractional power of the molecular weight M, a power that depends on the shape of the molecule but will be the same for any series of similar molecules. If a = M" for a certain type of molecule, then, for any series of solutes of this type on a gel of pore radius r, equation 10 may be written as

=-

k, x logM + k, logr+ k2

=- b logM

+

(11)

C,

where b and c are constants. Equation 11 predicts a linear relationship between & and logM for a series of similar solutes on a given gel over the range of molecular size for which equation 10 holds; this gives theoretical justification to the empirical correlation r e p ~ r t e d . ~ ~ . ~ ~ Anderson and St0dda1-t'~have also justified the correlation between V, and logM by this treatment. From equations 2 and 11,

+ Vi logM + c Vi + V,.

(V, - V,)lVi = -b logM

:. V,

= -b

C.

For a given column, V, and V, are constants, so that this simplifies to V,=-b'logM

+ c',

(12)

where b' and c' are constants. Equation 12 predicts the observed linear relationship between V, and logM. (74) D. M. W. Anderson and J. F. Stoddart, Anal. Chim. Acta, 34,401 (1966).

SHIRLEY C. CHURMS

28

The two other correlations, between logM and V,lV, or VJV,, can also be deduced in this way. The only reported correlation that is not accounted for by this treatment is that observed by Sanfellipo and Kd and (logM)-', for which no theoretical justificaS ~ r a between k~~ tion has as yet been obtained. The work of Anderson and S t ~ d d a r t has ' ~ thus given a sound theoretical basis to the correlations, formerly empirical, of gel-chromatography data with molecular weights, data which have been used with considerable success in the determination of the molecular weights of macromolecules. Of these correlations, the one most universally employed is the simplest; namely, the linear correlation between V, and logM for a series of similar solutes on a given gelcolumn. A number of substances of known molecular weight, or polymer fractions of narrow molecular-weight distribution, with or g,,,are chromatographed, under identical conknown values of ditions, on a column packed with a gel of appropriate fractionationrange, and the elution volumes observed are plotted against logM to give the calibration curve for the column; this should be linear over a range of molecular weights coinciding with the fractionation range of the gel. The substances under examination are then chromatographed on the same column under the same conditions as were used for the calibrating solutes, and their molecular weights are found from the observed elution volumes by reference to the calibration curve. The substances used for calibration must be structurally similar to those being examined, as the relationship between the Stokes radius and the molecular weight should be similar if the calibration is to be valid. For example, if the substances under study are globular proteins, the column must be calibrated with globular proteins; other types, such as glycoproteins, are not suitable for the calibration owing to their different structure.34Well-characterized fractions of dextran are used for caIibration in determinations of the molecular weight of p o l y s a c c h a r i d e ~ . ~ ~ . ~ ~ Values of molecular weight obtained by this method have been found to agree within 5-10% with values determined by classical methods. For example, from chromatography of fractions of the polysaccharide gum from Acacia senegal on Bio-Gel P-300, Anderson and coworkerP obtained Gnvalues of 99,000 and 35,000, respectively; the corresponding values obtained by osmometry were 105,000 and 37,000. Churms and S t e ~ h e n , ~on ' chromatographing the polysaccharide gum from Acacia podalyriaefolia on the same gel, obtained

al,.

GEL CHROMATOGRAPHY OF CARBOHYDRATES

29

a value of 31,500 for M,, the value from sedimentation and diffusion measurement^'^ being 33,500. In order to compare molecular weights determined by use of the simple correlation of V, and logM with the values given by the various theoretical equations, the author has calculated the molecular ~ the elution pattern weights corresponding to 3 peaks ~ b s e r v e d 'in of the gum from Acacia elata on Bio-Gel P-300, and has used, in addition to the V, versus logM correlation, the relationships arising from the P ~ r a t h ,La~rent-Killander,~~ ~~ and Squire63 treatments of gel chromatography. The validity of the linear relationships between M112and the functions Kd113,(- In K,J1/2, and (Ve/V0)1/3, predicted by the Porath equation and the modified versions of the Laurent-Killander and Squire equations, respectively, was tested with data obranging tained by chromatography of dextran fractions of known from 10,000 to 70,000, on the same column (90 X 1.5 cm) packed with the same gel. Plots of these functions against the square root of Mw were found to be linear in all cases. The values of the three functions corresponding to the elution volumes at the peaks in the elution pattern of the A. elata gum were then calculated, and the M , values of the appropriate polysaccharide fractions were estimated by interpolation of these linear plots. These molecular weights were also determined by reference to a V, versus log%, plot for the column, obtained with the same dextran fractions. The results given by the different treatments are compared in Table VII, from which it may be seen that agreement is fairly good, no result deviating from the mean by more than 10%.

zw,

TABLEVII Comparison of Values of Molecular Weight Obtained by Different Treatments M W

V, (ml)

PorathJ'

LaurentKilIandeP

Squirea3

V, versus log7iiw correlation

Mean

82 93 98

31,700 21,900 19,000

32,400 22,500 19,300

28,600 20,200 15,900

29,900 20,000 16,800

30,600 21,200 17,800

(75)G . R. Woolard, Ph.D. Thesis, University of Cape Town (1968). (76)P. I. Bekker, S. C. Chums, A. M. Stephen, and G. R. Woolard, Tetrahedron, 25, 3359 (1969).

30

SHIRLEY C. CHURMS

It may, therefore, be concluded that all of these approaches are valid for galactans of the type i n v e ~ t i g a t e d(and, ~ ~ perhaps, for polysaccharides in general) on poly(acry1amide) gels. The V, OerSus logM correlation has the advantage of simplicity, and it is, therefore, possibly the best method of determining molecular weights by gel chromatography. In the case of polymers having a distribution of molecular weights, such as polysaccharides, any determination of molecular weight can give only an average. The type of average obtained depends upon the experimental method used. Thus, a method that effectively counts the number of molecules present (for example, any method involving colligative properties, such as osmotic pressure) will give the numberaverage molecular weight Be, whereas a weight-average molecular weight is obtained from such methods as light-scattering that depend on the weight of the molecules instead of on their number. Gel chromatography is, however, not an absolute but a relative method of molecular-weight determination; all of the equations purporting to relate gel-chromatographic data to molecular weights or sizes contain constants that can be determined only by calibration of the gel column with similar solutes of known molecular weight or Stokes radius. Therefore, molecular weights obtained in this way depend upon those used in calibration of the column; if M, values are used, the results should be regarded as 2, values, whereas calibration in terms of G,, values gives results that will be closer to 2, for the polymer under examination. That both alternatives are permissible is demonstrated by the linearity, over a wide range of molecular weight, of plots of &” against both GWand %, calculated from results obtained by Granath and KvisP on chromatography of dextran fractions on Sephadex. The fact that the G, values of dextran fractions (from light-~cattering~~) were used by the author in establishing the validity of the Porath equation5I and the modified Laurent-Killandel.s4 and Squires3 equations for dextrans, whereas P ~ r a t h , ~ ~ Laurent and Killander,54 and Squirea3themselves used the results of Granath and F10din~~ (who gave ii?, values) to confirm the applicability of their equations to dextrans, also indicates that either type of average fits these treatments. Thus, gel chromatography may be used to determine either Gwor of a polymolecular substance.

an

(77) Pharmacia, Uppsala, Sweden, technical literature on dextrans (1966).

GEL CHROMATOGRAPHY OF CARBOHYDRATES

31

Iv. APPLICATIONTO CARBOHYDRATES 1. Sugars

a. Monosaccharides. -The behavior of aldoses in gel chromatogby using dexraphy has been thoroughly investigated b y Mar~den,'~ tran gels having a high degree of cross-linking and, consequently, very small pores capable of differentiating between such small solute molecules. Marsden found that the Kd values of these sugars did not decrease monotonically with increasing molecular weight. The value (0.64) found for an aldotetrose, namely, D-erythrose, was slightly higher than that for D-glyceraldehyde (0.61);the latter value may have been lowered as a result of dimerization of the D-glyceraldehyde in solution. The aldopentoses were found to have K d values higher than that of D-erythrose; Marsden attributed this property to the greater flexibility of the pyranose ring of the aldopentoses, as compared with that of the furanose ring in D-erythrose, the rigidity of which might be expected to hinder the entry of D-erythrose molecules into the gel pores. All of the aldohexoses examined exhibited Kd values lower than those of the corresponding aldopentoses, molecular weight being the main factor involved, as both classes tend to adopt the pyranose ring structure. Variations in Kd within the aldopentose and aldohexose series were observed, the Kd values of the aldopentoses increasing from 0.646 for D-arabinose to 0.710 for D-ribose, whereas those of the aldohexoses varied from 0.573 for D-glucose and D-galactose to 0.676 for D-talOSe. This result was attributed to differences in the degree of steric hindrance, resulting from the different spatial arrangements of the hydroxyl groups, to the entry of the various sugar molecules n ~ ~ that, within each series of sugars, into the gel phase. M a r ~ d e noted Kd decreased with decreasing instability rating (Kelly scale79)of the chair conformers of the sugars, that is, with increasing proportion of forms more liable than other conformations to steric hindrance in the gel phase. That this variation in Kd among isomeric sugars is related to conformation is further demonstrated by the absence of such differences in Kd within the corresponding alditol series; all pentitols were found to have a Kd of 0.576, and all hexitols, of 0.547. Despite the presence of anomers, differing in instability rating of their chair conformations, the elution profiles obtained on gel (78) N. V. B. Marsden, Ann. N. Y. Acad. Sci., 125, 428 (1965). (79) R. B. Kelly, Can. /. Chem., 35, 149 (1957).

32

SHIRLEY C. CHURMS

chromatography showed a single, sharp peak for every sugar studied; each sugar was eluted as if it were a single compound. Since mutarotation usually proceeds much more rapidly than elution, the & values determined by Marsden'* are those corresponding to the equilibrium mixtures of the anomers of the various sugars. The enantiomorphs of a given sugar were found to exhibit no appreciable difference in &; for example, the values for D- and L-mannose were 0.626and 0.627,respectively. It appears, therefore, that stereospecific adsorption to the gel, which could possibly result from the presence of the dissymmetric residues of D-glucopyranose in the dextran matrix, is not a significant factor. The & values found for methyl glycosides were lower than those of the corresponding sugars; for example, the values for D-glUCOSe and methyl a-D-glucopyranoside were 0.573 and 0.496,respectively. This behavior may be ascribed to the added bulk of the methyl group. M a r ~ d e n 'did ~ not attempt to separate mixtures of monosaccharides by gel chromatography; the small differences in K d among the different sugars indicate that separation could be achieved only on very large columns by use of the eluants used by this author, namely, deionized water and buffers of ionic strength 0.05(pH 7-8)composed (Tris) and hydrochloric of 2-amino-2-(hydroxymethyl)-1,3-propanediol or acetic acid. A good separation of L-rhamnose, 2-acetamido-2-deoxy-~-glucose, D-glucose, and 2-amino-2-deoxy-D-glucose, eluted in that order, was obtained by ZeleznickRoon a small column packed with the highly cross-linked, dextran gel Sephadex G-25,with 62:15:25 (v/v) butyl alcohol-it4 acetic acid-water as the eluant; however, the gel functioned more as a support for partition chromatography than as a molecular sieve. An almost complete separation of D-glucose from D-ribose on the poly(acry1amide) gel Bio-Gel P-2,with water as the eluant, has been reported by John and coworkers.81A water-jacketed column (127X 1.5 cm) at a temperature of 65" and gel of very fine particle-size (400 mesh) were used. Only under such conditions, where resolution is optimal, does good separation of monosaccharides by gel chromatography become possible.

b. Oligosaccharides. - Mixtures of oligosaccharides differing in (80) L. D. Zeleznick, J . Chromatogr., 14, 139 (1964). (81) M. John, G. TrBnel, and H. Dellweg, J . Chromatogr., 42, 476 (1969).

GEL CHROMATOGRAPHY OF CARBOHYDRATES

33

molecular weight are readily separated by gel chromatography, so that the technique is useful in the isolation of the oligosaccharides produced in degradative studies of polysaccharides. The value of the method was first demonstrated by Flodin and Aspberg,s2who used gel chromatography on Sephadex G-25 in distilled water to separate the sugars produced by acetolysis of cellulose followed by deacetylation of the resulting acetate^.^^,*^ An almost complete separation of the products, from D-glucose up to cellohexaose, was achieved. Similar gel chromatography of the oligosaccharides produced by acetolysis of yeast m a n n a n ~ has ~ ~ *been ~ ~ found to yield elution patterns characteristic of the yeast strains involved.86 For example, whereas the elution pattern given by the degradation products from Saccharomyces cerevisiae mannan revealed the presence of D-mannopentaose, -hexaose, and -heptaose, which, on further acetolysis, yielded di-, tri-, and tetra-saccharide~,~~ the pattern observed for acetolyzates of mannans from Candida strains was indicative of a high proportion of D-mannose, together with acetolysis-stable pentaand hexa-saccharides.86 In this way, the position in the mannan of ,~~ the (1 + 6)-linkages, the ones most susceptible to a c e t o l y s i ~ was indicated. Stewart and Ballous6have suggested the use of this technique in the classification of yeasts. John and coworkersa’ have reported the successful use of the highly cross-linked poly(acry1amide) gel Bio-Gel P-2 to separate the series -( oligosaccharides synthesized by the action of of a - ~ 1-+4)-linked Escherichia coli ML 30 on maltose. D-Glucose and oligosaccharides containing up to 13 D-glucose residues were resolved when the conditions used for separating D-glucose from D-ribose, already described, were employed. On this column, these workers also resolved maltose and isomaltose, and maltotriose and isomaltotriose, the (1+6)linked isomer being eluted first. Isomaltose and isomaltotriose were shown to be absent from the product obtained on incubation, with beta-amylase, of the oligosaccharide mixture formed by the action of amylomaltase in E . coli on maltose; the elution pattern on Bio-Gel P-2 showed that, after incubation for 2 hours, maltotriose, maltose, (82) P. Flodin and K. Aspberg, “Biological Structure and Function,” Academic Press, Inc., New York, N. Y., 1961, Vol. 1, p. 345. (83) E. E. Dickey and M . L. Wolfrom, /. Amer. Chem. Soc., 71, 825 (1949). (84) M . L. Wolfrom and J . C. Dacons,]. Amer. Chem. Soc., 74, 5331 (1952). (85) T. S. Stewart, P. B. Mendershausen, and C. E. Ballou, Biochemistry, 7, 1843 (1968). (86) T. S. Stewart and C. E. Ballou, Biochemistry, 7, 1855 (1968). (87) M . L. Wolfrom, A. Thompson, and C. E. Timberlake, Cereal Chem., 40,82 (1963).

34

SHIRLEY C. CHURMS

and D-glucose were the only degradation products. This further demonstrated the exclusive formation of a-D-( l+.Q)-linkages by action of amylomaltase on maltose.88 The oligosaccharides produced in degradative studies of glycosaminoglycans can also be separated by gel chromatography. For example, Flodin and coworkersss used Sephadex G-25for separating the lower members of the series of oligosaccharides produced by digestion of hyaluronic acid and chondroitin 4-sulfate with testicular hyaluronidase. These workersss succeeded in isolating the di-, tetra-, hexa-, and octa-saccharides by this procedure. The use, instead of distilled water, of an eluant containing sodium chloride in fairly high concentration was found to improve resolution, the best results being achieved with 0.1 M sodium chloride for the digest of hyaluronic acid, and M sodium chloride for the digest of chondroitin 4-sulfate. Eluants of high ionic strength, which have the effect of eliminating any form of electrostatic interaction (such as hydrogen bonding) of the solute molecules with the gel or with one another, are to be recommended in the gel chromatography of carbohydrates in general. In the course of an investigation of the chondroitin 6-sulfate-protein linkage, Helting and Rod&nsofractionated on Sephadex G-25 the oligosaccharides obtained on acid hydrolysis of chondroitin 6sulfate from umbilical cord, with 9:l (v/v) water-ethanol as the eluant. The separation of these oligosaccharides permitted their subsequent identification (by paper chromatography) as 3-O-P-D-galactosyl-4-O-P-D-galactosyl-D-xy~ose, 3-O-~-D-ga~actosyl-D-ga~actose, 4-0~-D-ga~actosyl-D-xy~ose, and 3-O-(~-D-glucosyluronicacid)-D-galactose. with 0.01 M Dietrichsl used gel chromatography on Sephadex G-25, acetic acid as the eluant, to fractionate the products obtained on degradation of heparin by enzymes of adapted Flavobacterium heparinium. The fractions from the gel column were subsequently purified by paper chromatography, and the components were idenditified as the 6-sulfuric ester of 2-deoxy-2-(sulfoamino)-~-glucose, --4)-linked ( to D-glucusaccharides consisting of this molecule a - ~ 1 ronic acid or its 3-sulfate, and tetra- and hexa-saccharides composed of these disaccharide units. Raftery and coworkerss2 have reported the fractionation of the (88) S.A. Barker and E. J. Bourne,]. Chem. Soc., 209 (1952). (89) P. Flodin, J. D . Gregory, and L. RodBn, Anal. Biochem., 8, 424 (1964). (90) T. Helting and L. RodBn, Biochim. Biophys. Acta, 170,301 (1968). (91) C. P. Dietrich, Biochem. I., 108, 647 (1968). (92) M. A. Raftery, T. Rand-Meir, F. W. Dahlquist, S. M. Parsons, C. L. Borders, Jr., R. G. Wolcott, W. Beranek, Jr., and L. Jao, Anal. Biochem., 30,427 (1969).

GEL CHROMATOGRAPHY OF CARBOHYDRATES

35

oligosaccharides obtained on acid hydrolysis of chitin by chromatography on Bio-Gel P-2, with water as the eluant. A good separation of the sugars from 2-acetamido-2-deoxy-~-glucose up to chitohexaose was obtained. Gel chromatography has also been used for fractionating derivatives of these sugars; for example, N-(trifluoroacety1)ated derivatives can be separated on Bio-Gel P-2 in water, and p-nitrophenyl glycosides on the same gel, with 0.1 M sodium chloride as the eluant. Peracetylated derivatives have been separated on Sephadex LH-20 with methanol as the eluant. Dahlquist and Rafteryg3have also used gel chromatography in investigating the action of lysozyme on chitotriose. Chromatography of the product (incubated during 6 hours at 40" at pH 5.0) on Bio-Gel P-2, followed by paper chromatography of the fractions obtained, led to identification of the components present as 2-acetamido-2-deoxy-~glucose and chito-biose, -triose, and -tetraose. It was concluded that transglycosylation (resulting in the formation of the tetraose) occurred, as well as hydrolysis to chitobiose and 2-acetamido-2-deoxy-~-glucose. Naturally occurring oligosaccharides have also been isolated by gel chromatography. Ohman and HygstedP have reported the isolation, from colostrum, of sialic acid oligosaccharides, including diO-sialoyl-lactose [ O-N-acetylneuraminoy1-(2 + 8)-O-(N-acetylneuraminoy1)-(2+ 3)-O-P-D-galactopyranosyl-( 1 + 4)-a-~-glucopyranose], by a procedure involving gel chromatography on Sephadex G-25 with 99: 1 water-butyl alcohol. The lower members (containing 1-10 Dfructose residues) of the fructan series present in dahlia tubers have been separated by Pontisss on Bio-Gel P-2, with distilled water as the eluant. After the fractions corresponding to each peak in the resulting elution curve had been pooled and freeze-dried, paper chromatography showed each product to be a different, pure compound; the Dfructose oligosaccharides were thus completely separated by this method. 2. Polysaccharides a. Dextrans. -The fractionation of dextrans by gel chromatography has been much investigated. Following the earlier work of Flodin and coworker^,^*^^ which demonstrated the feasibility of fractionating dextrans according to molecular weight on Sephadex gels, and established the fractionation ranges of these gels for dextrans, Laurent (93) F. W. Dahlquist and M. A. Raftery, Nature, 213, 625 (1967). (94) R. Ohman and 0. Hygstedt, Anal. Biochem., 23, 391 (1968). (95) H. G. Pontis, Anal. Biochern., 23, 331 (1968).

36

SHIRLEY C. CHURMS

TABLEVIII Separation of Sugars and Derivatives by Gel Chromatography Gel Sephadex G-25

Eluant water

99: 1(vlv) water-butyl alcohol

9: 1(vlv) waterethanol 0.01 M acetic acid 0.1 M sodium chloride

M sodium chloride 62: 15:25 (vlv)butyl alcohol44 acetic acid-water Bio-Gel P-2

water

0.1 M sodium chloride Sephadex LH-20

methanol

Sugars separated

References

oligosaccharides from acetolysis of cellulose oligosaccharides from acetolysis of yeast mannans sialic acid-containing oligosaccharides from colostrum oligosaccharides from acid hydrolysis of chondroitin 6-sulfate oligosaccharides from enzymic degradation of heparin oligosaccharides from hyaluronidase degradation of hyaluronic acid oligosaccharides from hyaluronidase degradation of chondroitin 4-sulfate D-g1ucOSe, L-rhamnose, 2-amino-2-deoxy-D-glucose, 2-acetamido-2-deoxy-~glucose D-ghCOSe from D-ribose; maltose from isomaltose; maltotriose from isomaltotriose; oligosaccharides formed from maltose by amylomaltase fructans (1-10 fructose residues) from dahlia tubers 2-acetamido-2-deoxy-~glucose, chito-biose, -triose, 4etraose 2-acetamido-2-deoxy-Dglucose and oligosaccharides from acid hydrolysis of chitin; N-(trifluoroacety1)ated derivatives of these p-nitrophenyl glycosides of chitin oligosaccharides peracetylated derivatives of chitin olieosaccharides

82 85,86 94

90

91

89

89

80

81

95 93

92

92 92

GEL CHROMATOGRAPHY OF CARBOHYDRATES

37

and Granaths6 fractionated a dextran having a broad distribution of weight into relatively homogenous fractions, ranging in M , from 15,000 to 63,000, by gel chromatography on Sephadex G-200 with 0.2 M sodium chloride as the eluant. Hummel and D. C. Smithy6 investigated the fractionation of dextrans in the molecular-weight range of 30,000 to 500,000 on Sephadex, agar, agarose, and poly(acry1amide) gels, and found that the best fractionation of the dextrans of high molecular weight was achieved on a 7% agar gel, with a 0.2 M Tris hydrochloride buffer (pH 8.0) as the eluant. A 4% agarose gel, with distilled water as the eluant, also proved effective. Sephadex and poly(acry1amide) gels are unsuitable for the fractionation of dextrans of molecular weight above -100,000. Granath and KvisP9 have described a method of determining, by gel chromatography, the molecular-weight distributions of dextrans in the range of 10,000 to 100,000. A column packed with a mixture of two Sephadex gels, G-200 and G-100, in the dry-weight ratio of 1:2 (so that the two occupied equal volumes when swollen) was used by these workers; the eluant was 0.3% sodium chloride. After the column had been calibrated with 17 dextran fractions of known M , and M,, determined by independent methods (light-scattering and end-group analysis, respectively), the dextrans under examination were chromatographed under the same conditions, and the molecular weights corresponding to the observed K,, values were read from the calibration curve. The technique has proved useful in clinical investigations involving the use of d e x t r a n ~ . ~ ~ Dextrans of low molecular weight have also been fractionated by gel chromatography. Bremner and coworkers,97 in preparing (by alkaline degradation) dextrans of low molecular weight suitable for incorporation in the clinically important iron-dextran complex, fractionated their product on Sephadex G-50, with 0J. M sodium chloride as the eluant. Separate fractions, ranging in M , (as determined by osmometry) from 1510 to 4860, were obtained in this way. The newer supports for gel chromatography, such as lyophilized p o l y ~ t y r e n e 'and ~ porous s i l i ~ a , have ~ ~ - ~been ~ successfully applied to the fractionation of dextrans over a wide range of molecular weight. A lipophilic gel, namely, Sephadex LH-20, has found application in studies of partially acetylated dextrans. In describing a modified procedure for the replacement of the 0-acetyl groups in such dextrans by 0-methyl groups prior to acid hydrolysis and identification of the fragments (which indicates the distribution of the substituents in molecular

(96) B. C. W. Hummel and D. C. Smith,J. Chromatogr., 8, 491 (1962). (97) I. Breniner, J. S. G. Cox, and G. F. Moss, Carbohyd. Res., 11, 77 (1969).

38

SHIRLEY C. CHURMS

the dextrans), de Belder and Norrmanesrecommended the use of this gel, with methanol and acetone as eluants, in the isolation of the products obtained after the various steps. In this way, the substituted dextrans are readily separated from reagents and products of low molecular weight.

b. Galactans. - Except for a preliminary investigation by Andrews and Roberts,ss who, in 1962, reported (in a note on the use of agar in gel chromatography) that a larch arabinogalactan and a Strychnos galactan are eluted from an agar column in that order, and both eluted after starch and a galactomannan from clover seed, interest in the gel chromatography of galactans commenced in 1965, when Anderson and coworkers66 described the use of a poly(acry1amide) gel, Bio-Gel P-300, in the determination of the molecular weights of fractions of the arabinogalactan gum from Acacia senegal. On calibration of the gel column with dextran fractions of known the correlation between V, and log il?, was found to be linear over the il?, range of 5,000 to 125,000. Values of Enwithin this range could, therefore, be estimated from elution volumes by reference to this calibration curve.66JooSodium chloride (M)was used as the eluant. This method has been successfully applied by Anderson and coworkers in strucand Araucarido7 gums. Detertural studies of various Acacia100-106 mination of the %, values of the products of hydrolysis and Smith degradation of the gums has permitted certain conclusions to be drawn regarding the distribution of (1+3)- and (1+6)-linkages in the arabinogalactans. The present author and coworkers have used the method of Anderson and coworkers to determine the molecular-weight distributions of the constituents of a number of Acacia and other plant gums, and their degradation product^.^^*^*^^^ The degree of resolution achieved by this procedure is illustrated

En,

(98) A. N. de Belder and B. Norrman, Carbohyd. Res., 8, 1 (1968). ~ (99) P. Andrews and G . P. Roberts, Biochem.J.,84, 1 1 (1962). (100) D. M. W. Anderson and J. F. Stoddart, Carbohyd. Res., 2, 104 (1966). (101) D. M. W. Anderson, Sir Edmund Hirst, and J. F. Stoddart, J . C h e n . Soc. ( C ) , 1959 (1966). (102) D. M. W. Anderson and I. C. M. Dea, Carbohyd. Res., 6, 104 (1968). (103) D. M . W. Anderson, I. C. M. Dea, and R. N. Smith, Carbohyd. Res., 7,320 (1968). (104) D. M. W. Anderson and I. C. M. Dea, Carbohyd. Res., 8,440 (1968). (105) D. M. W. Anderson and I. C. M. Dea, Carbohyd. Res., 8,448 (1968). (106) D. M. W. Anderson, I. C. M . Dea, and Sir Edmund Hirst, Carbohyd. Res., 8,460 (1968). (107) D. M. W. Anderson and A. C. Munro, Carbohyd. Res., 11, 43 (1969).

GEL CHROMATOGRAPHY OF CARBOHYDRATES

39

in Fig. 1, and Fig. 2 demonstrates the reproducibility of the elution patterns obtained. Chromatography of two specimens of A. elata gum collected at different times from the same tree, and of a specimen collected from a different tree in the same locality, gave elution patterns of striking similarity, revealed by comparison of Figs. 2 and 3. These results indicated that the elution pattern given on gel chromatography of a plant gum may be characteristic of the species involved and, therefore, highly significant in chemotaxonomy Products of a Smith degradation of A. elata gum have been fractionated on a preparative scale by gel chromatography on Bio-Gel P-10, with water as the eIuant.62 Anderson and coworkers'08 have reported the successful applica-

.

Volume Lmll

FIG. 1.-Elution Pattern of Material Precipitated by Ethanol from the Mild-Acid Hydrolyzate of Acacia elata Cum. [Bio-Gel P-300, 90 X 1.5 cm column, M sodium chloride as the eluant; flow rate, 3 ml/hr; sample, 2 mg in 1 ml of M sodium chloride.]

(108) D. M . W. Anderson, I. C. M. Dea, and A. C. Munro, Carbohyd. Res., 9,363 (1969).

40

SHIRLEY C. CHURMS

Volume (rnl)

FIG.2.-Elution Pattern of Ethanol-Precipitated A. elata Gum, Collected from Tree 1 in January, 1967. [Bio-Gel P-300,90 x 1.5 cm column, M sodium chloride as the eluant; flow rate, 4 mllhr; samples: (1) 5 mg in 1ml of M sodium chloride, and (2) 3 mg in 1 ml of M sodium chloride.]

tion of an agarose gel, Sepharose 4B, in the fractionation of Acacia gum polysaccharides of molecular weight above the upper limit of in the fractionation range of Bio-Gel P-300. Polysaccharides of the range of -3 x lo5 to 3 x lo6, including many occurring in plant gums, can be fractionated on this gel. Porous glass has also been used in the chromatography of plant gums of high molecular weight. Stoddart and Jonestogestimated the M , for Citrus Zimonia gum by comparison of its elution pattern on Bio-Glas 500 with those given by the gums from Acacia senegal and A. arabica, for which was known. Porous silica is also applicable to the molecular-sieve chromatography of plant-gum polysaccharides. 110

a,

a,

(109) J. F. Stoddart and J. K. N. Jones, Carbohvd. Res., 8, 29 (1968). (110) D. M. W. Anderson, A. Hendrie, and A. C. Munro,]. Chromatogr.,44,178 (1969).

GEL CHROMATOGRAPHY O F CARBOHYDRATES

41

The arabinogalactans occurring in larch wood have been widely investigated by gel chr~matography.~~'-"~ The Sephadex elutionpatterns demonstrate the presence of two components, A and B, in all heartwoods examined. Arabinogalactan A -75,000) preponderates, the proportion of B (M,-14,000) increasing with the age of the wood.112Component B (low molecular weight) is believed to be a product of slow, acid-catalyzed hydrolysis of the material of high molecular weight. Gel chromatography has shown that both A and B are polymolecular, althmgh each has a relatively narrow molecular-weight distribution.'12

(M,

-

M~

I

I00

40

20

10

5

'

I

I

I

I

I

40

1

80

I20

Volume I m l )

FIG. 3.-Elution Pattern of Ethanol-Precipitated A. elata Gum. [Bio-Gel P-300, 90 x 1.5 em column, M sodium chloride as the eluant; flow rate, 4 ml/hr; samples: (1) gum collected from tree 1 in February, 1969 (5 mg in 1 ml of M sodium chloride), (2) gum collected from tree 2 in February, 1969 (3 mg in 1 ml of M sodium chloride.)]

(111) B. V. Ettling and M. F. Adams, T a p p i , 51, 116 (1968). (112) B. W.Simson, W. A. Cote, Jr., and T. E. Timell, Suensk Papperstidn., 71,699 (1968). (113) H. A. Swenson, H. M. Kaustinen, J. J. Bachhuber, and J. A. Carlson, Macromolecules, 2, 142 (1969).

42

SHIRLEY C. CHURMS

c. Other Polysaccharides. -The fractionation of a number of starch dextrins on Sephadex G-75 in distilled water was reported in 1962 by Nordin.l14 The dextrins produced on hydrolysis of glycogens from various sources, and of amylopectin, by pancreatic alpha-amylase, were fractionated b y Heller and Schramm1I5by using Sephadex G-50 with water as the eluant. For the digest obtained from shellfish glycogen, rechromatography of the dextrin fractions first eluted from the column led to the isolation of previously unsuspected dextrins of high molecular weight, having a degree of polymerization (from reducing end-group analysis) ranging from 150 to 330. Such dextrins were not found in the digests from rabbit-liver glycogen, phytoglycogen, and amylopectin. The molecular-size distribution of the dextrins produced on digestion with the enzyme was found to vary considerably from one glycogen to another as a result of the different distributions of branch points in glycogens from different sources, concerning which much information can be obtained from experiments of this type. The molecular-weight distribution of a sample of a glucan, pulM a n , isolated from cultures of the fungus Pullularia pullulans grown in sucrose solutions,'Ifi was determined by Granath and KvisPg by the method developed by these workers for use with dextrans. Here, too, the column was calibrated with dextran fractions, as the shape of the pullulan molecule was believed to be sufficiently similar to that of a dextran to permit use of the same correlation of KUL,with molecular weight for both polysaccharides. However, in determining the molecular-weight distribution of a sample of a fructan, inulin, Granath and Kvistfi9calibrated the column instead of with dextran with fractions of inulin of known Ewand fractions. A different correlation of K,, with molecular weight was observed, indicating that the (1+2)-~-fructofuranosidiclinkages of the inulin chain"' give rise to a relationship between molecular weight and size that differs from that holding for dextrans. Inulin appears to have the more compact structure. Fractionation of the synthetic "polysucrose" Ficoll (manufactured by Pharmacia AB, Uppsala, Sweden) was achieved by Laurent and Granath56by the method already described for dextran. From material having a broad distribution of molecular weight, several fairly homo-

En,

(114) P. Nordin, Arch. Biochem. Biophys., 99, 101 (1962). (115) J. Heller and M . Schramm, Biochim. Biophys. Actn, 81, 96 (1964). (116) H. 0. Bouveng, H. Kiessling, B. Lindberg, and J. McKay, Acta Chem. Scnnd., 16, 615 (1962). (117) E. L. Hirst, D. J. McCilary, and E. C . V. Percival, J. Chem. Sac., 1297 (1950).

GEL CHROMATOGRAPHY O F CARBOHYDRATES

43

geneous fractions, ranging i n m , from 7,000 to 95,000, were obtained. The hemicelluloses that are precipitated on addition of methanol to the neutralized, 10% sodium hydroxide extract of powdered, Norwegian-spruce wood were fractionated by Kringstad and Ellefsen118by gel chromatography on Sephadex. After preliminary fractionation on Sephadex G-100, with 0.5 M sodium sulfate as the eluant, the components of low molecular weight were further separated by rechromatography on Sephadex G-25 in distilled water. The fractions of high molecular weight were subjected both to acid and alkaline hydrolysis, and the hydrolyzates obtained after different periods of time were chromatographed on Sephadex G-100 in 0.05 M sodium sulfate. From the elution patterns obtained, conclusions could be drawn regarding both the structure of the hemicelluloses and the mechanism of the hydrolysis. In an investigation of the polysaccharides present in soil, Stacey and used chromatography on Sephadex G-100 in distilled water to separate these polysaccharides from other components of the soil extracts. The polysaccharides were eluted much earlier than the colored substances present in the soil. Further fractionation of these polysaccharides was achieved by ion-exchange chromatography. Their isolation by this means permitted preliminary characterization; paper chromatography of the acid hydrolyzates revealed the presence of D-glucose, D-galactose, D-mannose, D-xylose, L-arabinose, L-rhamnose, L-fucose, and D-glucuronic acid r e s i d ~ e s . " ~ Barker and coworkers have applied gel chromatography in studies of pneumococcal polysaccharides."' Purification of the type-specific polysaccharide of Pneumococcus Type I1 was effected b y chromatography on Sephadex G-200 in M sodium chloride; in this way, the ribonucleic acid, a persistent impurity in preparations of this polysaccharide, was almost completely removed. The complex formed between the polysaccharide and the nucleic acid is largely dissociated in M sodium chloride, so that the two are free in this solvent and may be separated on the basis of their differing molecular size. Cell-wall lipopolysaccharides of Gram-negative bacteria (Shigetla strains) have also been purified by gel chromatography.122These (118)K. Kringstad and 0. Ellefsen, Papier, 18, 583 (1964). (119)S. A. Barker, P. Finch, M . H . B. Hayes, R. G. Simmonds, and M . Stacey, Nature, 205,68 (1965). (120)S. A. Barker, M. H . B. Hayes, R. G. Simmonds, and M. Stacey, Carbohyd. Res., 5, 13 (1967). (121)S. A. Barker, S. M. Bick, J . S. Brimacombe, and P. J. Somers, Carbohyd. Res., 1, 393 (1966). (122)E.Romanowska, Anal. Biochem., 33,383(1970).

44

SHIRLEY C. CHURMS

polysaccharides are readily separated from smaller contaminants (ribonucleic acids and nonspecific polysaccharides) on the agarose gels Sepharose 2B or 4B. Water has been used as the eluant, but 0.01 M ammonium hydrogen carbonate has proved more effective. After a preliminary, ion-exchange fractionation, *the C polysaccharide from pneumococci was fractionated by Gotschlich and Liu123 by gel chromatography on Sephadex G-200 in M acetic acid. From the elution volumes corresponding to peaks in the elution patterns obtained, the molecular weights of the components were estimated; these were found to range from 65,000 to >200,000. Analysis of the acid hydrolyzates of the fractions revealed the presence of %amino2-deoxy-~-glucose,2-amino-2-deoxy-D-galactose 6-phosphate, muraand its mic acid [2-amino-3-0-(~-l-carboxyethyl)-2-deoxy-~-glucose] 6-phosphate, D-glUtamiC acid, D- and L-alanine, and L-lysine, the proportions differing from one fraction to another. On this basis, it was suggested that the C polysaccharide is a cross-linked aggregate of two or more different types of polymeric chain, variation in the relative proportions of these chains giving rise to the heterogeneity observed. Sialoglycosaminoglycans extracted from rat brain have been purified by Brunngraber and Brown124by a procedure involving gel chromatography on Sephadex G-200 in distilled water; this resulted in the separation of these polysaccharides from u.v.-absorbing impurities. Elution of the sialoglycosaminoglycans commenced at the void volume of the column, whereas most of the impurities emerged much later. The sialoglycosaminoglycans give an elution pattern consisting of a broad, unsymmetrical peak that indicates considerable heterogeneity. Analysis of the fractions of high molecular weight, eluted before the peak, and of the fractions eluted after the peak, revealed that the former contained a higher proportion of sialic acid and a lower proportion of L-fucose than the latter. It was, therefore, suggested that the heterogeneity of the sialoglycosaminoglycans arises from the presence of variable proportions of sialic acid and L-fucose linked to a hexosamine-hexose chain that is common to all of them. The estimation of the molecular weights of glycosaminoglycans by gel chromatography is often complicated by a strong tendency to aggregation, and by changes in molecular shape and size caused by variation in solute-solvent interaction. Brunngraber and BrownlZ4 found that the use of water as the eluant in the gel chromatography (123) E. C. Gotschlich and T.-Y. Liu, J . B i d . Chern., 242, 463 (1967). (124) E. G. Brunngraber and B. D. Brown, Biochern.J.,103, 65 (1967).

GEL CHROMATOGRAPHY OF CARBOHYDRATES

45

of sialoglycosaminoglycans resulted in early elution of these solutes, owing to the formation of aggregates. If, instead of water, a 0.05 M Tris hydrochloride buffer (pH 7.5) that was 0.1 M in potassium chloride was used for elution, the peak elution-volume increased to a value corresponding to a molecular weight of just under 10,000, a value much lower than the apparent molecular weight in water. However, the u.v.-absorbing impurities were not removed, so that, for the main purpose of the experiment, water was the better eluant. The gel-chromatographic behavior of the glycosaminoglycan heparin has also been found to b e markedly dependent on the ionic strength of the eluant. Chromatography of heparin samples on Sephadex G-100 and G-200, with a series of phosphate buffers (pH 6.8) of different concentrations as eluants, revealed a tendency for the elution volume to increase and the peak to broaden with in~ ~the effect persisted at creasing ionic strength of the e 1 ~ a n t . IAs ionic strengths high enough to eliminate any possibility of exclusion of heparin from the gel by reason of repulsion between ionized acidic groups and similar groups in the gel, it was concluded that the lower elution-volumes observed at low ionic strengths could not be entirely due to this factor. The dependence of many properties of heparin (such as viscosity and sedimentation constant) on the ionic strength of the medium suggests that the shape, size, and degree of solvation of the molecules change with the ionic strength. This effect is probably attributable to accompanying changes in solute-solvent interaction, a factor particularly important for heparin, as it contains a large number of sulfate groups per molecule. This behavior makes determination of its molecular weight very difficult; wide discrepancies have been observed between the values of molecular weight estimated from gel chromatographyAz6and those given by other methods. By gel chromatography on Sephadex G-200, with an eluant consisting of 0.15 M sodium chloride mixed with 0.12 M sodium phosphate buffer solution, pH 7.4 (9:1, v/v), heparin of porcine mucosal origin has been shown to be highly polymolecular.A27N o increase in blood anticoagulant activity with increasing molecular size, indicated by the results of an earlier study,'28 was observed when the activity of fractions selected from different parts of the elution curve was determined. The anticoagulant potency of the sample studied was, (125) M. Skalka,]. Chromatogr., 33, 456 (1968). (126) G. B. Sumyk and C. F. Yocum, J. Chromatogr., 35, 101 (1968). (127) M. Hall, C. R. Ricketts, and S. E. Michael, /. Pharm. Pharmacol., 21,626 (1969). (128) C. R. Ricketts, P. L. Walton, and D. R. Bangham, Brit../. Haematol., 12,310 (1966).

46

SHIRLEY C. CHURMS

however, relatively high, and it was suggested that, in a less potent sample, various degrees of alteration of chemical structure may occur. In this case, potency may well vary with molecular weight. Hyaluronic acid from human synovial fluid was isolated by Barker and Young129by gel chromatography on Sephadex G-200 at 4", with M sodium chloride as the eluant. Chromatography was preceded by sterile digestion of the synovial fluid with the proteolytic enzyme pronase, in a Tris hydrochloride buffer (pH 7.9) for 8 hours at 37". The hyaluronic acid, largely excluded from the gel, was eluted in the early fractions, and was well separated from the protein debris (low molecular weight) remaining after the pronase digestion. The same authors130used gel chromatography on agarose for isolating the undegraded hyaluronic acid-protein complex from human synovial fluid. A study by How and Long131of the molecular-weight distribution of hyaluronic acid from normal synovial fluid and of that of patients having rheumatoid diseases has afforded some interesting results. The polymolecularity and average degree of polymerization of the samples of hyaluronic acid were determined by chromatography on agarose gels (1%and 2%) at 2", with 0.01 M phosphate buffer (pH 7.3) in 0.2 M saline as the eluant. The hyaluronic acid in fluid from diseased joints was found to have a lower average degree of polymerization than that in normal synovial fluid. The authors suggested that this may be due either to deficient biosynthesis, or to depolymerization (of the hyaluronic acid) resulting from the action of reducing substances or enzymes in the fluid from patients having rheumatoid disease. The latter possibility is favored by the detection and hyaluronidase activity of 2-acetamido-2-deoxy-/3-~-glucosidase in fluid from arthritic joints. How and Long131used the same method for investigating the effect of drug therapy on the molecular-weight distribution of hyaluronic acid in the synovial fluid of patients having rheumatoid disease. It was found that clinical improvement following the administration of certain drugs was always accompanied by an increase in the average molecular size of the hyaluronic acid, indicating that the symptoms of this group of diseases are related to the degradation of the hyaluronic acid in the synovial fluid. The presence of chondroitin sulfate in the synovial fluid of arthritic (129)S. A. Barker and N. M. Young, Carbohyd. Res., 2, 49 (1966). (130)S. A. Barker and N. M. Young, Carbohyd. Res., 2, 363 (1966). (131)M.J. How and V. J. W. Long, Clin. Chim. Acta, 23, 251 (1969).

47

GEL CHROMATOGRAPHY OF CARBOHYDRATES

patients was demonstrated by Barker and coworkers132by gel chromatography on Sephadex G-200. The glycosaminoglycans in the synovial fluid were isolated in this way after a preliminary ion-exchange fractionation had removed all free protein from the protein-polysaccharide complexes. The use of gel chromatography to determine the molecular-weight distribution of chondroitin sulfate, on a micro scale, has been reported by W a ~ t e s o n .Columns '~~ (60-100 cm x 0.8-3.0mm id.; volume 0.5-7 ml) were packed with Sephadex G-200; a special technique involving the use of vibration was used in packing. Samples containing 100 pg in 50-100 pI were chromatographed on these columns, with M sodium chloride as the eluant. In this way, chondroitin 4-sulfate from and of the various bovine nasal septa was fractionated, the fractions being determined by sedimentation and osmometry, respectively. Linear relationships between K,, and the logarithms of and Gnwere observed over the molecular-weight range both covered (11,500-41,300). The calibration curves thus obtained were used for determining and of another sample of chondroitin 4-sulfate, chromatographed under the same conditions. The values obtained agreed well with those determined by other methods. The molecular-weight distribution of acidic glycosaminoglycans in normal urine, and that of patients having Hurler's syndrome (gargoylism), has been determined by Constantopoulos and coworkers 134 by chromatography on Sephadex G-200, with 0.025 M sodium chloride were as the eluant. Fractions of chondroitin 4-sulfate, of known used for calibration. The high concentrations of heparitin sulfate and chondroitin sulfate B (dermatan sulfate) characteristic of Hurler's syndrome were clearly reflected in the elution patterns and molecular weight averages of the urinary glycosaminoglycans from these patients, which differed significantly from those of normal subjects.

aw an

aw

aw an

aw,

3. Miscellaneous Carbohydrates The phenolic D-glucosides salicin and tremuloidin (2-0-benzoylsalicin) occurring in the bark of Populus tremula have been separated from each other, and from the other carbohydrates present, by gel

(132) S. A. Barker, C. F. Hawkins, and M . Hewins, Ann. Rheumatic Diseuses, 25,209 (1966). (133) A. Wasteson, Biochim. Biophys. Acta, 177, 152 (1969). (134) G. Constantopoulos, A. S. Dekaban, and W. R. Carroll, Anel. Biochem., 3 1 59 ( 1969).

SHIRLEY C. CHURMS

48

TABLEIX Gel Chromatography of Polysaccharides Packing Sephadex G-25 Sephadex G-50

Sephadex G-75

Eluant water water

water

Sephadex G-100 water 0.1 M ammonium hydrogen carbonate 0.5 M sodium sulfate Sephadex G-100 0.3% sodium and G-200 (2:1, chloride dry-weight) Sephadex G-200 water

25 mM sodium chloride 0.15 M sodium chloride-0.12 M phosphate buffer, pH 7.4 (9:1, vlv) 0.2 M sodium chloride M sodium chloride

M acetic acid Bio-Gel P-300

M sodium chloride

A p p1i cation

fractionation of dextrans, M, ) ibid., 819 (1956). (164) E. J. Reist and S. L. Holton, Curboh!/d.Res., 2, 181 (1966).

156

WR'

YCH,

0

NEIL R. WILLIAMS

+

=YCL;;"

X 2,3-Anhydro-P-D- lyxo

YC.cz)"

X 3-subs. p - ~ - ~ a b i ? ~ ~

2-subs. p-~-xylo

(R' = OMe): (Y = OH) X = NH222,163 (3 subs.), SCH,PhlE5(3:2)*, SCN1B6(3:7)"; (Y = OAc) X = Brl@ (2:l)'; (Y = OTr) X = SBzlE7(1:l)'; (Y = OMe) X = NH,IB3 (3 subs); (3:4)' (Y = OCHzPh)X = (R'= OEt): (Y = OH) X = NHZ3' (3 subs.); (Y = tetrahydropyran-2-yloxy)X = NHz34 (3 subs.). (R' =adenine): (Y = O H ) X = O B Z '(3 ~ ~subs.), S E P 9 (3 subs.), SCH,Ph'68 (1:5)', N:%I7O (3 subs.); (Y = H) X = OBz"' (1:2)', SCHzPh172 (l:l)', N31T2(1:3)' (R' =6-(dimethylamino)-2-(methylthio)purine:(Y = O H ) X=NH2173(3 subs.) (R'= uracil): (Y = OH) X = OH174(3:17)', NHz175(3 subs.), Br175R (3 subs.); (Y = OMS) X =OH174(1:4)0,0CHzPh174(3subs., followed by 2,5,-anhydride formation)

YCH, D

0 O

YCH, M

e

-

" V O M e HO

2, 3-Anhydro-a-D-ribo

2-subs. a-o-arabino

+

0

QOMe OH

3-subs.(Y-D-XYIO

(165) G. Casini and L. Goodman,J. Amer. Chem. SOC.,85,2357 (1963);86,1427 (1964). (166) L. Goodman, Chem. Commun., 219 (1968). (167) K. J . Ryan, E. M. Acton, and L. Goodman, J . Org. Chem., 33, 3727 (1968). (167a) J. A. Wright and J. J. Fox, Carbohyd. Res., 13, 297 (1970). (168) W. W. Lee, A. Benitez, L. Goodman, and B. R. Baker J. Amer. Chem. SOC., 82, 2648 (1960). (169) A. P. Martinez, W. W. Lee, and L. Goodman, J . Org. Chem., 31, 3262 (1966). (170) W. W. Lee and A. P. Martinez, unpublished results, cited in Ref. 171. (171) E. J . Reist, D. F. Calkins, and L. Goodman, J. Org. Chem., 32, 2538 (1967). (172) E . J. Reist, V. J. Bartuska, D. F. Calkins, and L. Goodman, J . Org. Chem., 30, 3401 (1965). (173) B. R. Baker and R. E. Schaub, J. Amer. Chem. SOC.,77,5900 (1955). (174) I. L. Doerr, J. F. Codington, and J. J. Fox, J . Org. Chem., 30, 467 (1965). (175) J. F. Codington, R. Fecher, and J. J . Fox, J . Org. Chem., 27, 163 (1962). (175a) M. Mirata, T. Naito, and Y.Nakai, Japan Pat. 6,813,214 (1968);Chem. Abstracts, 70,38051 (1969).

OXIRANE DERIVATIVES OF ALDOSES

157

(Y = OH) X = OMe,75(2 subs.), NH21769177 (3 subs.); NH4SCN+3,5-thietane1"; KCN+ 3,5-1act0ne~~; (Y=OAc) X=BrI7'(5:2)"; (Y =OTs)X=H1"** (2 subs.), SCH,Ph17Hon (2 subs.); (Y = H) X=NH21R"(Zsubs.);(Y =OCH,Ph) X=F1"'

(R' = OMe): (Y= OH) X = H,I6*SEt,'*' NH2176.177; NH,SCN+3,5-thietane17s; (Y= OAc) X = Brl79; (y OTs) X = H,17BooSEtlD8"'., (Y = OCH,Ph) X = F'"; (Y = H) X = NH,.''" (R' = adenine) (Y = OH) X = S E P ; (Y = H) X = SCH2Ph,172 N:) 46(a) 161 46(a) 34(b) 34(b) 34(b)

M W C C

(contintred)

(242) H. Ohle and H. Wilcke, Ber., 71,2316 (1938).

178

NEIL R. WILLIAMS

2,3-Anhydro compound P-D-Lyxofuranoside, methyl 5-0-acetyl5-0-benzyl5-deoxy5-iodo5-0-methyl5-0-p-tolylsulfonyl5-0-trityla-D-Ribofuranoside, ethyl 5-0-acetyl5-0-(tetrahydropyran-2-yl)a-D-Ribofuranoside, methyl 5-0-acetyl5-0-benzyl5-deoxy5-0-(p-nitrobenzoyl)5-O-p-tolylsulfony lp-D-Ribofuranoside, ethyl 5-0-acety15-04tetrahydropyran-2-y1)P-D-Ribofuranoside, methyl 5-0-acetyl5-0-benzyl5-deoxy5-O-(p-nitrobenzoyl)5-0-p-tolylsulfonyl5-0-trityl-

TABLEV (continued) tab, M.P., degrees degrees 74-75

62-64 14-15 76-77 156-157

21-23 134-136 93-95

2-3 98-99 66.5-67 127

-102 -80 -67 -28 -113 -88 -89 -76 13.1 -18.4 -14.3 13 -2.1 -18.1 26 -26

7 -89.5 -108 -99 -109 -112 -90.8 -153.4 -95 -80 -59

Rotation solvent

References

W C E

22 164 167a 46(4 4%) 163(a) 46(4 167 34(W 3403) 34b) 177 177 180a 180 177 178 34b) 34W 3403) 177 177 182 180 177 178 176

W

C C C C W C E M C C W C C W C C M C C C

g-(P-D-PentofuranosyI)adenineDerivatives

Lyxofuranosyl 5-deoxyRibofuranosyl 5-deoxy-

205-206 208-209d 200-203 190d

-23 -3 42

168 171 183 171

W WIP C

6-(Dimethylamino)-2-(methylthio)-9-(p-~-pentofuranosyl)punne Derivatives

Lyxofuranosy 1 Ribofuranosyl 5-0-trityl-

211-212 172-173 213

P

68.7

C

173 186 186

-53.5 23.6

P C

48 48

-43

7-(~-Pentofuranosyl)theophylline Derivatives

a-Lyxofuranosy 1 p-Ribofuranosyl, 5-0-trityl-

204-205 202-203

(continued)

OXIRANE DERIVATIVES O F ALDOSES

179

TABLEV (continued) 2,S-Anhydro compound

M.P., degrees

[ffID, degrees

Rotation solvent

34 3 26

W C W D W

References

l-(p-D-PentofuranosyI)uraci~ Derivatives

Lyxofuranosyl 5-0-benzoyl5-deoxy5-iOdO5-O-(methylsulfonyl)-

139-140 187-190 146-146.5 218-220 177-177.5

-4

16

175 175 175 175 175

TABLEVI 5,6-Anhydrohexofuranoses

5,B-Anhydro compound a-D-Glucofuranose, 1,2-O-isopropylidene3-acetamido-3-deoxy3-0-benzyl3-O-meth yl3-04methylsulfonyl)3-O-p-tol ylsulfonyla-D-Glucofuranose, 1,2-0-(trichloroethy1idene)8-L-lyxo-Hexofuranose, d-deoxy1,2-O-isopropylideneLu-D-xylo-Hexofuranose,3-deoxy1,2-O-isopropylidenep-L-Idofuranose, 1.2-0-isopropylidene3-0-benzyl3-O-(methylsulfonyl)-

M.p.,

degrees 133.5 142-143

80-81 145

27 73-75 96.5-98.5

[ a ] ~ , Rotation

References

degrees

solvent

-26.5 67.5 -51.2 44.8 430 -50.9 -29.8

W C C A C C P

11,243 191 12 188 191 191 244

-27.6

C

204

-15 -25.2 -78.7 -7.4

D C C C

207 12 12 205

(243) K. Freudenberg, H. Toepffer, and C. C. Andersen, Ber., 61,1750 (1928). (244) R. Grewe and G. Rockstroh, Chem. Ber., 86,536 (1953).

This Page Intentionally Left Blank

2,S-ANHYDRIDES OF SUGARS AND RELATED COMPOUNDS”

BY J. D E F A Y E ~ ” lnstitut de Chimie des Substances Naturelles, Centre National de la Recherche Scientifique, 91 Cif-stir-Yvette, France

I. Introduction,. .......................................................... 11. Methods of Formation., ................................................. 1. Deamination of Aminoaldoses with Nitrous Acid ........................ 2. Halogenolysis of Halogenated Derivatives .............................. 3. Intramolecular Displacement of Sulfonates by the Action of a Base . . . . . . . . 4. Solvolysis of Sulfonates ............................................... 5. Miscellaneous Methods. ..................... ................. 111. Reactions. .............................................................. 1. Stability of the Oxolane Ring. .......................................... 2. Reactions Involving a Carbonyl Group .................................. IV. Utilization.. ............................................................ V. Table of Properties of 2,s-Anhydro Derivatives of Sugars, Alditols, and Aldonic Acids ................................

181 183 183 194 198 203 209 210 210 212 215

219

I. INTRODUCTION Almost a century has elapsed since Ledderhose,’ seeking to establish the structure of an amino sugar (“glycosamin”) obtained by hydrolysis of chitin with acid, subjected “glycosamin” to deamination with nitrous acid. He obtained an amorphous, tasteless, nonfermentable, strongly reducing product that he found exceedingly disconcerting, as all his attempts to characterize it led to extensive decomposition. The structure of “glycosamin” (2-amino-2-deoxy-D-glucose) was not established2 until half a century later. As regards the product of deamination with nitrous acid, namely, 2,5-anhydro-~-mannose (chitose), it was not until 1956 that the polemics concerning its struc*Translated from the French by D. Horton. **Presentaddress: Centre de Recherche sur les Macromolecules Vegetales, C.N.R.S., Domaine Universitaire de Grenoble, 38 St. Martin d’HL.res, France.

(1) G . Ledderhose, Z . Physiol. Chem., 4,139 (1880). (2) W. N. Haworth, W. H. G. Lake, and S. Peat,j. Chem. Soc., 271 (1939).

181

182

J. DEFAYE

ture came to an end. Nevertheless, the way was opened for a whole series of brilliant studies that, from Tiemann3+ to Levene7-19by way of Fischer, 6,zo enabled sugar chemistry to contribute to fundamental chemical knowledge of general interest. The chemistry of the 2,5-anhydrides of aldoses subsequently entered a prolonged lull, and Peat's reviewz1of 1946 in this Series does not report on any work later than 1925. The experimental basis of the deamination of amino sugars with nitrous acid was, nevertheless, established. The progress afterwards made in the conformational analysis of sugars made it possible for Shafizadehzzto draw a parallel with the nitrous acid deamination of the aminocyclohexanols, and to rationalize the whole of these results. During the same period of time, the use of chromatography, together with developments in physical methods for the determination of structure, permitted reaction mixtures to be studied more thoroughly. Many investigations revealed the formation of 2,5-anhydrides of aldoses during reactions where such anhydrides were not anticipaied. It is noteworthy that, although the five-membered carbocyclic rings are more strained and possess a higher fi-ee-energy than the sixmembered analogs, the oxolane ring (five-membered, oxygencontaining heterocycles) of the 2,5- or 3,6-anhydrides are common in the aldose and alditol series. The theoretical aspects of ring closure in the sugars has been the subject of two articles in this S e r i e ~ , ~ ~ , ~ ~ and will not be treated further here. (3) F. Tiemann, Ber., 17, 241 (1884). (4) F. Tiemann and R. Haarmann, Ber., 19, 1257 (1886). (5) F. Tiemann, Ber., 27,118 (1894). (6) E. Fischer and F. Tiemann, Ber., 27,138 (1894). (7) P. A. Levene and F. B. LaForge,]. Biol. Chem., 20,433 (1915). ( 8 ) P. A. Levene and F. B. LaForge,]. Biol. Chem., 21,345 (1915). (9) P. A. Leveneand F. B. LaForge,]. Biol. Chem.,21,351(1915). (10) P. A. Levene and F. B. LaForge,]. Biol. Chem.,22,331 (1915). (11) P. A. Levene and G . M. Meyer,J. Biol.Chem.,26,355 (1916). (12) P. A. Levene,]. Biol. Chem., 31,609 (1917). (13) P. A. Levene,]. Biol. Chem., 36,73 (1918). (14) P. A. Levene,]. Biol. Chem., 36,89 (1918). (15) P. A. Levene,]. Biol. Chem., 39,69 (1919). (16) P. A. Levene and E. P. ClarkJ. Biol. Chem., 4 6 , l Q(1921). (17) P. A. Levene, Biochem. Z., 124,36 (1921). (18) P. A. Levene,]. B i d . Chem., 59,135 (1924). (19) P. A. Levene and R. Ulpts,]. Biol. Chem.,64,475 (1925). (20) E. Fischer and E. Andreae, Ber., 36,2587 (1903). (21) S. Peat, Aduan. Carbohyd. Chem., 2,37 (1946). (22) F. Shafizadeh,Advan. Carbohyd. Chem.,13,9 (1958). (23) J. A. Mills, Aduan. Carbohyd. Chem.,10,1(1955).

2,5-ANHYDRIDES OF SUGARS AND RELATED COMPOUNDS

183

The attention focused on facile methods for obtaining heterocycles of this type has led to recent utilization of some of them as intermediates in synthesis, inasmuch as certain ones, such as m u s ~ a r i n e , ~ ~ have been obtained from natural sources, and various others have been shown to have interesting biological proper tie^.^^ The purpose of this article is to attempt a rationalization among the numerous methods for access to 2,5-anhydrides of aldoses; and, from the mass of scattered, and frequently fragmentary, information in the literature, to single out sugar derivatives in this class from the viewpoint of their properties and their utilization. 11. METHODS O F FORMATION

1. Deamination of Aminoaldoses with Nitrous Acid Deamination of aminoaldohexoses with nitrous acid constitutes the oldest route of access to the 2,5-anhydroaldohexoses (see the Introduction). It is only in recent years, however, that the results obtained by this method could be rationalized on a theoretical basis as a consequence of concomitant developments in certain other areas of organic chemistry. The action of nitrous acid on a primary, aliphatic amine is still a complex process, inasmuch as it is necessary to postulate certain steps that are imperfectly understood, but it can nevertheless be considered that the process leads ultimately to a carbonium i ~ n .The ~ ~ , ~ ~ carbonium ion thus formed is termed “hot,” that is to say, it is nonsolvated and chemically activated; the nature of the products resulting from its decay is essentially determined by the conformation of the initial state; the electronic factors in effect at the time of the transition state exert little influen~e.~’ A route through a diazonium-ion intermediate has long been postulated’* (see Scheme 1, path A) from the fact that primary, aromatic amines react with nitrous acid in an acid medium to give diazonium saltsz9that are sufficiently stable to be isolable. However, more-recent (24) S. Wilkinson, Quart. Reo. (London), 15,153 (1961). (25) J. Defaye, P. P. Slonimski, G. Perrodin, and E. Lederer, Compt. Rend., 251, 817(1960). (26) F. C. Whitmore and D. P. Langlois,]. Amer. Chem. Soc., 54,3441 (1932). (27) J. H. Ridd, Quast. Reo. (London), 15,418 (1961). (28) P. Brewster, F. Hiron, E. D. Hughes, C. K. Ingold, and P. A. D. S. Rao, Nature, 166,179 (1950). (29) H. Zollinger, “Azo and Diazo Chemistry,” Interscience Publishers, Inc., New York, N. Y., 1961.

J. DEFAYE

184

work has suggested that, in the aliphatic series, the reaction proceeds by direct decomposition of an intermediate diazohydroxide having the anti configuration30(see Scheme 1, path B).

R-N=N-OH

.. ..

-

..

p.0

R-N=N-O-H

ad..

-

0

R-NEN

..,H + O,H

path A R-NH,

.. ..

H N o z t ~ - NI - ~ = ~ :

-

H

’

.. @

2 0 .

’

+

N,

Ro+

- - t R, .N=N, ’.

HR. *Q-N@

CB.

0

R-N=N-0: I * - .* H

H,O

*.

R ,.N q=N,@

OH

*

p H 2

Scheme 1. Action of nitrous acid on a primary, aliphatic amine

Undoubtedly, the vicinal groups play a fundamental role in the outcome of this reaction, especially with the sugars, where the favored conformation of the molecule at equilibrium is controlled at the outset by groups that determine whether the molecule exists as a cyclic or acyclic structure. The deamination of cyclic and acyclic amino sugar derivatives by nitrous acid will be considered in turn. a. Cyclic Amino Derivatives. (i) 2-Amino-2-deoxyaldohexopyranoses. - The deamination of these compounds by nitrous acid was reviewed in 1958 by Shafizadeh,*‘ who compared the results with the behavior of aminocyclohexanols on deamination with nitrous acid. Various papers devoted to this s u b j e ~ t ~ O show - ~ ~ that the outcome of the deamination of these compounds by nitrous acid depends essentially on the group antiparallel to the amino group. Thus, the deamination by nitrous acid of the isomeric 2-aminocyclohexanols, for which the favored conformation is rigidly fixed by the presence of a bulky substituent at C-4, is highly stereoselective, and M . ChBrest, H. Felkin, J. Sicher, F. SipoS, and M. Tichy, J . Chem. Soc., 2513 (1965). G . E.McCasland,J. Amer. Chen. SOC., 73,2293(1951). W.Klyne, Progr. Stereochern., 1,72(1954). D.Y. Curtin and S. Schmukler,]. Amer. Chem. Soc., 77,1105(1955). W.G.Dauben and K. S. Pitzer, in “Steric Effects in Organic Chemistry,” M. S . Newman, ed., JohnWiley and Sons, Inc., New York, N. Y., 1956,p. 3.

2,5-ANHYDRIDES OF SUGARS AND RELATED COMPOUNDS

185

leads uniquely to the product of rearrangement resulting from attack on the carbonium ion at C-2 by the group trans and antiparallel to the amino group30 (see Scheme 2).

Scheme 2. Deamination by nitrous acid of the epimeric 2-amino-4-tert-butylcyclohexanols (accordingto Ref. 30)

It may, therefore, be expected that the deamination with nitrous acid of those 2-amino-2-deoxyaldopyranoses in which the amino group is oriented equatorially in the most stable conformation will lead to a 2,5-anhydroaldose, with inversion of the configuration at C-2. For example, the deamination of 2-amino-2-deoxy-~-glucose(1) with nitrous acid proceeds principally to 2,5-anhydro-~-mannose'*'~-'~ (2), and the methyl 2-amino-2-deoxy-a- and -P-D-glucopyranosides behave similarly.38The low rate of deamination of the a-Danomer has been attributed to steric hindrance to the approach of the reagent; This difference in rate has been proposed as a method for establishing anomeric configurations in glycosaminoglycans. The structure of 2,5anhydro-D-mannose (for many years known as chitose) was contro-

(35) S. Akiya and T. Osawa, Yakugaku Zasshi, 74,1259 (1954). (36) A. B. Grant, New Zealand/. Sci. Technol., B37,509 (1956). (37) 8.C. Bera, A. B. Foster, and M . Stacey,J. Chem. SOC.,4531 (1956). (38) A. B. Foster, E. F. Martlew, and M. Stacey, Chem. Znd. (London),825 (1953).

J. DEFAYE

186

versial for a long time.39-42In the same way, the deamination of 2amino-2-deoxy-~-ga~actose (3)with nitrous acid Ieads to 2,5-anhydroD - t a l o ~ e ' " ~(4). ~'~~

0

where -Xo = -NP or -N,OH,

The key sequence in the determination of the structures of 2,5anhydro-D-mannose (2) and -D-talose (4) was their reduction to the corresponding 2,5-anhydroalditols (5 and 6) and identification of the asymmetric "dialdehyde" (7) resulting from oxidation of the anhydroalditols with p e r i ~ d a t e . ~ ' . ~ ~ HOCH,

0 HOH,C,H

I(C%OH HO

HOH,C

,c-0-c,

H,CH,OH CHzOH

f

OMe

(39) C. Neuberg, H. Wolff, and W. Neimann, Ber., 35,4009 (1902). (40) W. Armbrecht, Biochern. Z., 95, 108 (1919). (41) L. Zechmeister and G . T6th, Ber., 66, 522 (1933). (42) P. Schorigin and N. N. Makarowa-Semljanskaja,Ber., 68,965 (1935). (43) J. Defaye, Bull. Soc. Chirn. Fr.,999 (1964). (44) E. Vankata Rao, J. G . Buchanan, and J. Baddiley, Biochern.]., 100,801 (1966).

2,5-ANHYDRIDES OF SUGARS AND RELATED COMPOUNDS

187

The nitrous acid deamination of 2-amino-2-deoxy-~-mannose(8),in the favored conformation, the amino group ofwhich is axially attached, leads, in contrast, uniquely to D - g l u ~ o s e ,characterized, '~ after oxidation by nitric acid, as D-glucaric acid (9).This result has also been verified by direct crystallization of the D-glucose and b y assay with Dglucose o x i d a ~ e . ~ ~ ~

t HNO,

CO,H I

HCOH i HOCH I HCOH I

HCOH I CO,H (81

0

(91

where -X* = - N P or -N20H,

It is remarkable that the only compound isolated from this reaction results from attack by solvent on the carbonium ion at C-2. By analogy with the deamination of the 2-aminocyclohexanols by nitrous acid, it might have been expected that a 3-0x0 derivative would have been formed as a result of hydride migration, because the C-3 proton is antiparallel to the leaving group. LeveneI5 reported that heating of 2-amino-2-deoxy-D-mannose (8) in the presence of silver oxide leads to a crystalline, nitrogen-free compound to which he attributed the structure of 2,5-anhydro-~glucose on the basis of its elemental analysis. The possibility of interconversion between the two chair forms C1 (D) + 1C (D), which would bring the amino group at C-2 into equatorial orientation, has been postulated.22Without excluding this possibility, it remains to be proved that the deamination by silver oxide does, indeed, proceed by (44a)D.Horton and K. D. Philips, unpublished results.

188

J. DEFAYE

way of a carbonium ion. The physical properties of this 2,s-anhydride'* are, moreover, unusual (see p. 214). As might be expected, the nitrous acid deamination of methyl 2amino-4,6-O-benzylidene-2-deoxy-a-~-altropyranoside ( 10) hydrochloride leads45,46 uniquely to methyl 2,3-anhydro-4,6-0-benzylidenea-D-allopyranoside (11). The benzylidene group does not play an

essential role in this reaction, unless it is required for stabilization of the conformation of the aminohexose. Although an initial report indicated4' that deamination with nitrous acid of the hydrochloride of ( 12) methyl 2-amino-4,6-0-benzylidene-2-deoxy-~-~-glucopyranoside led to 4,6-O-benzylidene-~-mannopyranose, the results of subsequent work corrected this result, and showed46that the sole product obtained under these conditions is 2,5-anhydro-4,6-0-benzylidene-~mannose (13). The same reaction performed on methyl 2-amino-2deoxy-4,6-O-ethylidene-3-O-methyl-~-glucopyranoside led to a similar r e ~ u l t . ~ "

(45) L. F. Wiggins, Nature, 157,300 (1946). (46) S. Akiya and T. Osawa, Chem. Pharm. Bull. (Tokyo), 7,277 (1959). (47) J. C. Irvine and A. Hynd,]. Chem. SOC., 105,698 (1914). (48) S. Akiya and T. Osawa, Yakugaku Zasshi, 76,1276 (1956).

2,5-ANHYDRIDES OF SUGARS AND RELATED COMPOUNDS

189

(ii) 2-Amino-2-deoxyaldonolactones. -The deamination of Z-amino2-deoxyaldonolactones by nitrous acid was investigated by Leveneg,l4 with promising results, but the theoretical basis for interpreting his observations was not available at that time. Levene found that the nitrous acid deamination of 2-amino-2-deoxyD-mannonolactone (14) gives 2,5-anhydro-~-mannonicacid (16),a result identical with that obtained by similar deamination of the corresponding acid (15). On the other hand, the same reaction performed

H OH H,N o c o

Hop / HCOH I

(15)

HCOH I CqOH

J . DEFAYE

190

on 2-amino-2-deoxy-~-idonolactone (17) led, after oxidation with nitric acid, to 2,5-anhydro-~-gularicacid (18). The inversion of con-

OH

(18) where -Xo

0

= - N P o r -N,OH,

figuration at C-2 evident in the latter example can be explained, assuming a five-membered-ring structure for these lactones, by the fact that the hydroxyl group at C-5 in 17 is, sterically, well placed for attacking a carbonium ion developing at C-2. After deamination with nitrous acid followed by oxidation, 2-amino-2-deoxy-~-idonic acid (19) gives 2,5-anhydro-~-idaricacid (20) (see p. 191). COJi I H,NCH I HYOH HOCH I HCOH I

ChOH

m%

1

OH

b. Acyclic Amino Derivatives. -The outcome of deamination of these compounds by nitrous acid is essentially dependent on the nature of the participating group at C-1. A wide variety of compounds has been obtained to date, and the nature of these is not, as the initial results on the nitrous acid deamination of 2-amino-2-deoxyaldonic acids had previously led investigators to suppose, limited to the class of 2,5-anhydrides of sugars.

2,5-ANHYDRIDES OF SUGARS AND RELATED COMPOUNDS

191

(i) 2-Amino-2-deoxyaldonic Acids. - Deamination of the eight 2-amino-2-deoxy-~-aldohexonic acids with nitrous acid was investigated by Levene and coworkers in an outstanding series of paper^.'."'^,'^ A discussion of these results has already been presented in a Volume in this Series.22 In all known instances, the deamination of 2-amino-2-deoxyaldohexonic acids by nitrous acid takes place with formation of a 2,5anhydro ring and with retention of the configuration at C-2, as illustrated in the sequence 21+22-23.

*H0Cv”’

HCNH, I

HOYH

HNO, +

HOCH I

HCOH

HCOH

HCOH

HCOH

I

I

CH,OH

0

1 0

I

I

CH,OH

HO

(22)

where

The mechanism proposed by Foster49for this deamination postulates the formation of an unstable, intermediate a-lactone (22a),which subsequently undergoes attack by the hydroxyl group on C-5; this process would explain, by a double inversion, the net retention of configuration that is observed. (ii) 2-Amino-2-deoxyalditols. - By analogy with the nitrous acid deamination of 2-amino-2-deoxyaldonic acids, it might be expected that an oxolane ring would result from the deamination of 2-amino-2deoxyalditols with nitrous acid. Although the formation of such a compound had once been envisaged,50 no confirmation of the preliminary claim has been forthcoming. In fact, even though the deamination of l-amino-l-deoxyalditols with nitrous acid certainly leads in the case of the D-mannitols” and D-glucitol derivatives5] to the corresponding 1,4-anhydrides, the same reaction5’ applied to 2-amino-2-deoxy-D-glucitol (24) gives 2-deoxy-~-arabino-hexose (26).Under the same conditions, a mixture (49) A. B. Foster, Clrern. Ind. (London),627 (1955). (50) V. G. Bashford and L. F. Wiggins, Nature, 165,566 (1950). (51) V. G. Bashford and L. F. Wiggins,J. Chern. Soc., 299 (1948). (52) Y. Matsushima, Bull. Chem. Soc.Jnp.,24,144 (1951).

J . DEFAYE

192

CH,OH

HC=O

I

I

HFNH, HOCH I

HYOH

-

HCOH I

CH,OH

HNO,

HO&@ 1

-

1

I HOCH

HCOH

HCOH

HCOH

HCOH

I

I

I

I

c&or

CH,OH

-

1

7% HOCH HCOH I

HCOH I C&OH

(251

(24) where -X@

=

Q 0 -N2 or -",OH,

of 2-amino-2-deoxy-~-arabinitol and 2-amino-2-deoxy-~-ribitol, obtained by reductive cleavage of the (o-nitropheny1)hydrazone of Derythro-pentulose, underwent d e a m i n a t i ~ n ~ to~ give 2-deoxy-Derythro-pentose. A mechanism involving a hydride migration (25) has been proposed by Foster.49 (iii) Dithioacetals of 2-Amino-2-deoxyaldoses. - Bivalent sulfur possesses an electronic structure analogous to that of oxygen. On the other hand, being more polarizable than the oxygen atom, it is more apt to stabilize a carbonium ion by formation of a three-membered ring. In numerous examples, the possible intervention of an episul. ~ ~ participation fonium ion has been pointed out in the l i t e r a t ~ r eThis ~ ~ - ~ ~ is frequently accompanied by migration of the sulfur g r o ~ p to the carbon atom originally carrying the group displaced, and this factor makes prediction of the outcome of the reaction uncertain. In attempting to obtain 2,5-anhydro-~-glucosediethyl dithioacetal, Defaye5*performed the nitrous acid deamination of 2-amino-2-deoxyD-glucose diethyl dithioacetal (27) under weakly acidic conditions (acetic acid and sodium nitrite), and obtained a principal product later shown5' to be ethyl 2-S-ethyl-l,2-dithio-cr-~-mannofuranoside (28). Shortly before, Horton and coworkers60 had reported the forma-

(53) Y. Matsushinia and Y. Imanaga, Bull. Chem. Soc.Jap.,26,506 (1953). (54) For a review, see L. Goodman, Aduati. Carbohyd. Chem., 22, 109 (1967). (55) N. A. Hughes and R. Robson,./. Chem. Soc. ( C ) ,2366 (1966). (56) N . A. Hughes, R. Robson, and S. A. Saeed, Chem. Commun., 1381 (1968). (57) N. A. Hughes and R. Robson, Chem. Commun., 1383 (1968). (58) J. Defaye, Bull. Soc. Chim. Fr., 1101 (1967). (59) J. Defaye, T. Nakamura, D. Horton, and K. D. Philips, Carbohyd. Res., 16, 133 (1971). (60) D. Horton, L. G . Magbanua, and J . M. J. Tronchet, Chem. Ind. (London), 1718 (1966); A. E. El Ashmawy, D . Horton, L. G. Magbanua, and J. M . J. Tronchet, Carhohyd. Res., 6,299 (1968).

2,5-ANHYDRIDES OF SUGARS A N D RELATED COMPOUNDS

193

-

HOCH,

-

I

HOCH

I

HCOH I

-

W

CH,OH

0 S

E

t

(28)

pH 5 . 6

EtS,H,SEt C I HCNHz I HOCH I

HCOH

HNO,

I

H SEt 0,C’ EtS I ‘CH I

HOCH I

HCOH I

HCOH

HCOH

I

I

CH,OH

CH,OH

(27 )

HOCH,

H

O

W

O

H

(29)

tion of 2-S-ethyl-2-thio-~-glucose(29) as the major product from the same dithioacetal (27)when the deamination was performed with aqueous nitrous acid at a more acidic pH; a minor product was identical with compound 28. In both instances, the participation of an ethylthio group in initial stabilization of the carbonium ion is evident. The subsequent course of the reaction nevertheless appears to be strongly influenced by the nature of the medium. In acetic acid solution, there is attack at C-1 by the hydroxyl group on C-4, whereas, in the more-acidic, aqueous medium, it is the solvent that attacks at the same (least hindered) carbon atom, with formation of a hemithioacetal that is rapidly hydrolyzed to give the aldehyde group. To account for

J. DEFAYE

194

the net retention of configuration in the reaction leading to 29, it was ) , which atpostulated that the initial episulfonium ion ( D - ~ u w ~ o in tack by solvent at C-1 would be hindered by the hydroxyl group on C-3, equilibrates with a second episulfonium ion (D-gluco), and the latter is rapidly attacked by solvent at the unhindered, C-1 position.60 Although 2-S-ethyl-2-thio-~-mannoseis readily epimerized to 29 in the presence of base,60asuch epimerization does not occur at the pH levels ~ s e din ~the~deamination * ~ ~ of 27. 2. Halogenolysis of Halogenated Derivatives a. Brominolysis of 2-Deoxy-2-iodo Derivatives in the Pyranose Series.-The action of bromine in the presence of silver acetate on (30)in acetic methyl tri-O-acetyl-2-deoxy-2-iodo-~-~-glucopyranoside acid containing potassium acetate leads6*to 173,4,6-tetra-O-acety1-2,5anhydro-D-mannose methyl hemiacetal (31), obtained as a mixture of the C-1 epimers. The structure of this compound was established by ( a ) conversion into the corresponding dimethyl acetal (32) by acidcatalyzed methanolysis, and (b) obtaining 2,5-anhydro-~-rnannitol (33) upon treatment of the hemiacetal 31 with sodium borohydridg.

Br,, AgOAc

t

KOAc, AcOH

ACO-

A

c

O

W

HC(0Me) (OAc)

OMe

AcO

How

i

MeOH-HC1

J

CH,OH

(33)

AcO

(60a)B. Berrang and D. Horton, Chem. Commun., 1038 (1970). (61) R. U. Lemieux and B. Fraser-Reid, Can. J . Chem., 42, 547 (1964).

(32)

195

2,5-ANHYDRIDES OF SUGARS AND RELATED COMPOUNDS

This reaction has been compared61with the nitrous acid deamination of 2-amino-2-deoxy-~-glucose;attack by a bromonium ion could lead to an intermediate, alkyl-iodonium cation (2-C-I-Br)O, which would decompose through nucleophilic attack by 0-5,which is antiparallel to the leaving group at C-2. This hypothesis would seem justifieds2 by the fact that the same reaction performed on the epimeric halide (manno configuration) (34)leads to a mixture containing 60% of the 3,3,4,6-tetraacetate of methyl 2-bromo-2-deoxy-a-~-urubino-hexopyranosid-~u1ose 3-hydrate (38) and 20% of 1,3,4,6-tetra-O-ace~l-2-O-methyl-~-~-glucopyranose

~ ; c p o h r y A c ; m

OAc

Me0 Me (35) A

c

O

m

OMe (34)

/

(36) (20%)

Br, , AgOAc, KOAc, AcOH

\

(60%)

(36). The mechanism postulated62 for this reaction involves two parallel pathways leading from a common precursor, a carbonium ion at C-2. Stabilization of the carbonium ion by participation of the axial methoxyl group at C-1, to give the intermediate methoxonium ion (35),is followed by attack by solvent at the anomeric position to give the ether (36) obtained. In the alternative route, the initial carbonium ion could subsequently undergo elimination to give an intermediate enol acetate (37), which is then brominated stereoselectively at C-2, to give compound 38 as the major product of the reaction.

b. Action of Lead Tetrafluoride on Pyranoid Glycals.-3,4-Di-Oacetyl-1,5-anhydro-2-deoxy-~(39) and -L-erythro-pent-l-en-itol [D(62) R. U. Lemieux and B. Fraser-Reid, C a n . ] . Chem.,42,539 (1964).

1. DEFAYE

196

and L-arabinal (ribal) diacetates] react63.64with hydrofluoric acid in the presence of lead tetraacetate to give, after deacetylation, 2,5anhydro-l-deoxy-1,l-difluoro-D(40) and -L-ribitol, respectively. The same reagents acting upon 3,4,6-tri-O-acetyl-l,5-anhydro-2-deoxyD-arabino-hex-l-enitol (D-glucal triacetate) (41) a ringcontraction of the same type, to give 2,5-anhydro-l-deoxy-l,l-difluoroD-mannitol (43). Compound 43 is likewise obtained65 by the action of the hydrofluoric acid-lead tetraacetate reagent on ethyl 4,6-diO-acetyl-2,3-dideoxy-a-~-erythro-hex-2-enopyranoside (42), with subsequent deacetylation.

AcO

HO

OH (40)

(39) CH,OAc

2. NaOMe, MeOH HO

f

(43)

?H,OAc

It is known66 that the hydrofluoric acid-lead tetraacetate reagent leads, with unsaturate?. steroids, to cis additions of fluorine. This fact suggests that a cis-difluoro adduct (39b, 42b) is the starting point for the ring contractions observed in the preceding examples. (63) P. W. Kent, J. E. G. Barnett, and K. R. Wood, Tetrahedroii Lett., 1345(1963). (64) P. W. Kent and J. E . G. Barnett, Tetrohedron, S u p p l . 7,69 (1966). (65) K. R. Wood and P. W. Kent,]. Clzein. SOC.(C), 2422 (1967). (66) A. Bowers, P. G. Holton, E. Denot, M. C. Loza, and R. Urquiza, ]. Anier. Chem. Soc., 84, 1050 (1962).This reagent [A. L. Henne and T. P. Waalkes,J. Anier. Chem. Soc., 67, 1639 (1945)l supposedly leads, i n situ, to lead tetrafluoride, an unstable compound that is difficult to isolate.

197

2.5-ANHYDRIDES OF SUGARS AND RELATED COMPOUNDS

Such an intermediate is quite plausible for the D-glucal derivative (41a), which leads to a derivative (43a) of 2,5-anhydro-~-mannitol. The process could be very similar to that already discussed for the brominolysis of 2-deoxy-2-iodo derivatives of D-glucopyranose. The similar rearrangement observed with the "pseudo-glucal" diacetate (42a) could be due to an attack at C-3 by an acetoxyl group, leading to migration of the double bond, with displacement of the ethoxyl group at C-1. From such a reaction scheme, it would, however, be expected that a derivative (44) of 2,5-anhydro-~-arabinitolwould be obtained 1-enit01 from 3,4-di -0-acetyl- 1,5-anhydro-2-deoxy-~-erythro-pent(39a), and not the T i h product observed.

Ace"\; AcO

\

MeONa, MeOH

(39b)

MeOH

"

V

C I HO

H

F

2 (44)

198

J. DEFAYE

3. Intramolecular Displacement of Sulfonates by the Action of a Base

The sulfonic esters of sugars are fundamental intermediates in synthesis, as they permit facile access, through displacement by nucleophiles, to amino sugars, deoxy sugars, halogeno sugars, thio sugars, and anhydro sugars.66aIn the last example, the nucleophilic agent is a hydroxyl group of the same molecule. The reactions of sulfonates are normally of the s N 2 type and proceed with configurational inversion at the carbon atom that originally bore the sulfonyloxy group. Inasmuch as the factors leading to closure of a new oxygen-containing ring are strongly influenced by any ring already present, it is evident that the displacement of sulfonate groups in acyclic derivatives may follow a course quite different from that observed for cyclic derivatives. a. Sulfonie Esters of Acyclic Structures.-The action of one molar equivalent of p-toluenesulfonyl chloride on various dialkyl dithio~ ~ ,D-lyxose,68 ~~ in pyridine solution acetals of D-ribose, D - ~ y l o s e ,and at a temperature below -5”, leads to the corresponding dithioacetals (45, 48, and 51) of 2,5-anhydro-~-ribose,-D-xyIose, and -D-lyXOSe, respectively. The nature of the substituent on sulfur playss7 no particular role in this reaction. The structures of the compounds obtained have been proved by oxidation with lead tetraacetate and, for the D-ribose and D-xylose derivatives, by the isolations7 of a crystalline (p-nitropheny1)hydrazone after cleavage of the dithioacetal group by bromine. Cleavage of the dithioacetal groups from the products, followed by reduction of the resultant carbonyl derivatives (46, 49, 52) with sodium borohydride leads,68with the three compounds (45, 48, and 51), to 1,4-anhydro-~-ribitol(2,5-anhydro-~-ribitol)(47), 1,CanhydroL-xylitol (2,5-anhydro-~-xylitol) (50), and 1,4-anhydro-~-arabinitol (2,5-anhydro-~-lyxitol) (53), identified by comparison with their 1,4-anhydro-~-xylitol,~~ and enantiomorphs, 1,4-anhydro-~-ribitol,~~ 1,4-anhydro-~-arabinitol.’~ (66a)R.S. Tipson, Adoan. Carbohyd. Chem., 8,107 (1953);D . H. Ball and F. W. Parrish, ibid. 23, 233 (1968); 24, 139 (1969). (67) H. Zinner, H. Brandhoff, H. Schmandke, H . Kristen, and R. Haun, Chem. Ber., 92, 3151 (1959). (68) J. Defaye, Bull. Soc. Chim. Fr., 2686 (1964). (69) R. Kuhn and G. Wendt, Chem. Ber., 81, 553 (1948). (70) E. J. Hedgley and H. G. Fletcher, Jr., J. Amer. Chem. Soc., 86, 1576 (1964). (71) R. Barker and H. G. Fletcher, Jr., J . Org. Chem., 26, 4605 (1961).

2,5-ANHYDRIDES OF SUGARS AND RELATED COMPOUNDS

199

HO

HO

(52)

(53)

In sharp contrast to the behavior observed when the three dithioacetals aforementioned are treated with p-toluenesulfonyl chloride in pyridine, dialkyl dithioacetals of D-arabinose, treated under the same conditions, are converted into the corresponding 5-p-toluenesulfonates, generally isolable crystalline in high yielde7' This remarkable difference has been interpreted7*/ on conformational grounds; the D-arabinose dithioacetals are stable in the extended, planar zigzag conformation, whereas the other three examples experience some destabilization in the extended form, because of parallel 1,3-interaction~.~~* Furthermore, the transition state for closure of the 2,s-anhydro ring would be quite strained in the Darabinose series, but not in the other three.?*/ For the reaction leading to the 2,8anhydride, it has been postulated that initial esterification to give the 5-sulfonic ester is followed by intramolecuIar displacement of the 5-substituent by 0-2, either directly or during processing of the reaction mixture in an aqueous medium.67 Another hypothesis, which supposes a dehydration without passage through an intermediate sulfonic ester, has also been advanced.73 (72) H . Zinner, K. Wessely, and H. Kristen, Chem. Ber., 92, 1618 (1959). (72a)J.Defaye and D. Horton, Carbohyd. Res., 14, 128 (1970). (72b)D. Horton and J . D. Wander, Carbohyd. Res., 10,279 (1969). (73) H. Zinner, K. H. Stark, E. Michalzik, and H. Kristen, Chem. Ber., 95, 1391 (1962).

J. DEFAYE

200

b. Sulfonic Esters of Cyclic Structures-The action of sodium methoxide on methyl or ethyl 5-O-p-tolylsulfonyl-a-~-arabinofuranoside (54) leads'4 to the corresponding aIkyl 2,5-anhydro-cu-~-arabinofuranosides (55), the structures of which have been demonstrated by conversion, by way of the aldehyde 56, into 2,5-anhydro-~-arabinitol(57) and comparison of the latter with its enantiomorph, namely, 2,5-anhydro-~-arabinitol.'~

\

H2, Ni

(56)

GHzoH

(57) where R = M e o r Et

Compound 55 is one of the rare examples of a 2,5-anhydro sugar that also possesses a furanoid ring; it has various interesting properties as a result of ring strain in the bicyclo[2.2.1] system, and these are discussed later (see p. 212). A glycosylamine analog in the D series (59), a derivative of uracil, has been p r e ~ a r e d 'by ~ the action of sodium benzyloxide on 1-[2,3anhydro-5-0-( methylsulfonyl)-~-~-lyxofuranosyl]uracil(58). (74) M. Cifonelli, J. A. Cifonelli, R. Montgomery, and F. Smith,]. Arner. Chern. SOC., 77, 121 (1955). (75) A. B. Foster and W. G . Overend, J. Chem. SOC., 680 (1951). (76) I. L. Doerr, J. F. Codington, and J. J. Fox, J. Org. Chern., 30, 467 (1965).

2,S-ANHYDRIDES OF SUGARS AND RELATED COMPOUNDS

201

0

PhC4ONa PhC&OH

PhCH,O

0

I

(59)

I

PhCbO

Compound 59 is substantially more stable toward acid hydrolysis than the alkyl anhydroarabinoside 54, because boiling for three hours under reflux in 0.2 M sulfuric acid is needed in order to liberate the uracil. It should be noted, however, that the acid hydrolysis of pyrimidine nucleosides requires77 conditions considerably more severe than those needed for cleaving the anhydronucleoside 59. The action of sodium methoxide on 1-[5-0-(methylsulfony~)-/3-Dlyxofuranosyl]uracil (60) leads78to 1-(2,5-anhydro-/3-~-lyxofuranosyl)uracil (61); the 3,5-oxetane ring that might equally well have been

M8"cm OH

%

qc&(61)

(77) R. S. Tipson, Advan. Carbohyd. Chem., 1, 193 (1945); J. J. Fox and I. Wempen, ibid., 14, 283 (1959). (77a)B. Capon, Chem. Reu., 69,407 (1969). (78) J . F. Codington, I. L. Doerr, and J. J . Fox, J . Org. Chem., 30, 476 (1965).

202

J. DEFAYE

formed in this reaction is not observed, confirming that formation of an oxolane ring is favored when there are the two possibilities. The conditions for acid hydrolysis of this nucleoside derivative are more or less the same as those for the D-arahinose derivative 59. 2,5-Anhydro-D-lyxose could not, however, he isolated from the hydrolyzate thereof. Inspection of molecular models reveals that 2,5-anhydro-arahinose and -1yxose are the only 2,5-anhydroaldopentoses wherein the simultaneous presence of a 1,li-furanoid ring is sterically possible. c. Displacement of Halide Ions.-The displacement of a halide ion by an anionic atom of oxygen is quite analogous to the displacement of a sulfonate under the same conditions, and two examples of this type of reaction are, accordingly, presented in this Section. The action of (p-nitropheny1)hydrazine on 3,4-di-O-acetyl-2-hromoZdeoxy-D-xylopyranose (62)leads79to two (p-nitrophenyl)hydrazones, namely, 3,4-di-O-acety1-2,5-anhydro-~-xylose (p-nitropheny1)hydra(p-nitropheny1)zone (63) and 3,4-di-O-acety1-2,5-anhydro-~-lyxose hydrazone (64). Simultaneous formation of the two compounds, unex-

(79) A. Gerecs, Magy. Kern. Foly., 68, 211 (1962).

2,5-ANHYDRIDES OF SUGARS AND RELATED COMPOUNDS

203

pected at the outset, can nevertheless be explained if it is considered that displacement of the bromine atom at C-2 by the oxygen atom of the heterocycle 62a leads in the first stage to 3,4-di-O-acetyl-2,5anhydro-aldehydo-D-lyxose(62b) which, under the influence of the basic medium, undergoes epimerization at C-2, by way of the intermediate 62c, before undergoing conversion into the corresponding hydrazones 63 and 64. The 2,5-anhydrides of 2-ketoses constitute a special case, in which the hemiacetalic carbon atom is involved in the anhydride bridge. Such anhydrides more closely resemble the 1,6-anhydroaldoses (glycosans) than the 2,5-anhydroaldoses. An example of a 2,5-anhydride of a 2-hexulose is obtained by the action of base on a D-fructosyl fluoride derivative; treatment of l-O-methyl-/3-D-fructopyranosyl fluoride at 90" with concentrated, aqueous sodium hydroxide leads,Eo in poor yield, to 2,5-anhydro-l-O-methyl-P-D-fructopyranose.

4. Solvolysis of Sulfonates The solvolysis of sulfonic esters of sugars in acid media does not seem to have been greatly exploited. The examples that can be given indicate, nevertheless, that the method has great potentialities for the interconversion of ring systems. From the mechanistic point of view, the reaction can be interpreted, in general, in terms of protonation on an oxygen atom of the ester group, which leads to a displacement of charges on the carbon atom of the carbonyl group thereof, and proceeds further to the formation of a carbonium ion. The action of 1% methanolic hydrogen chloride on 1,2-O-isopropylidene-3,5-di-O-p-tolylsulfonyl-~-xylofuranose (65) leads,E' after boiling for three hours under reflux, to 2,5-anhydro-3-0-ptolylsulfonyl-D-xylose dimethyl acetal (66). The structure of 66 was demonstrated b y conversion into the disulfonate 67; this was prepared independently from a 2,5-anhydro-~-xylosedialkyl dithioa~etal'~ (48) by p-toluenesulfonylation followed by exchange of the acetal group in methanol in the presence of mercury salts. A reaction identical from the standpoint of mechanism is the conversion*' of 3-O-benzyl-1,2-O-isopropylidene-5,6-di-O-p-tolylsulfonyl-D-glucofuranose (68) into 2,5-anhydro-3-0-benzyl-6-O-ptolylsulfonyl-L-idose dimethyl acetal (70) by boiling a solution of 68 in methanol containing 2% of concentrated hydrochloric acid for 40 hours under reflux. These two reactions involve initial removal (80) F. Micheel and E. A. Kleinheidt, Chem. Ber., 98, 1668 (1965). (81) J. Defaye and J. Hildesheim, Tetrahedron Lett., 313 (1968).

J. DEFAYE

204 TsO

-

MeOH, HCl,l%_ 0

Me

MeOH, HC1,2%

1 hour HO

OMe

2.5-ANHYDRIDES OF SUGARS AND RELATED COMPOUNDS

205

of the O-isopropylidene group, with formation, in the case of the D-glucofuranose derivative, of the corresponding methyl D-glucoside (69), which can be isolated as an intermediate after one hour of refluxing. The formation of the dimethyl acetal takes place only as a consequence of the second step, subsequent to, or simultaneous with, the closure of the 2,5-anhydro ring. This solvolysis resembles a substitution of the sN2 type, because, for the D-glucofuranose derivative,R2there is an inversion of configuration at the carbon atom carrying the group displaced. It should be noted that, if the carbonium ion is formed at C-5, the leaving group is sufficiently close to the molecule that the attack can take place only from the opposite side. Apparently related to the two preceding reactions is the action of 2% methanolic hydrogen chloride on 1,2-O-isopropylidene-5-selenoa-D-xylofuranose (72) (obtained from the benzylseleno derivative, 71), which affords83 a mixture of 2,5-anhydro-5-seleno-~-xylose(and -Dlyxose) dimethyl acetal (73). Similarly, the action of the same reagent

Na-NH,

*

O+Me Me

on 1,2-O-isopropylidene-5-thio-D-xylofuranose (74) furnishesM42,5anhydro-5-thio-~-xylose(and -D-lyxose) dimethyl acetal (75).

(82) J. Defaye and V. Ratovelomanana, Curbohyd. Res., in press (1971). (83) T. Van Es and R. L. Whistler, Tetrtihedron, 23, 2849 (1967). (84) B. Nestadt and T. Van Es, Tetrahedron, 24, 1973 (1968).

J. DEFAYE

206

Tq0

MeOH-HCI (2%)

HO

OLOMe

Me

(74)

(75)

A mechanism involving nucleophilic displacement of the hydroxyl group on C-2 by the selenium atom at C-5 has been proposed for the formation of the acetal 73. This mechanism does not, however, account for the presence of two products, epimeric at C-2, in the reaction mixture. It should be noted that 2,5-anhydro-aZdehydo-~-idose (79) had already been obtained85 by the action of aqueous sulfuric acid on 5,6-anhydro-1,2-0-isopropylidene-1~-~-glucofuranose (76); anhydride 79 is formed i n 25% yield, together with D-glucose (80), which is the major product (60%) in the reaction.

L

f

(77)

(85) C . A. Dekker and T.Hashizume, Arch. Biochem. Biophys., 78,348 (1958).

2,8ANHYDRIDES OF SUGARS AND RELATED COMPOUNDS

207

The competition between attack by solvent a t C-6 (intermediate 78) and attack at C-5 by the hydroxyl group on C-2 (intermediate 77) explains these results, which show the extent to which closure of an oxolane ring is favored. In principle, a third (unobserved) possibility exists, namely, closure of a six-membered, oxygen-containing heterocycle as a result of opening of the epoxide at C-6 by the hydroxyl group on C-2. Other examples of the solvolysis of sulfonates to give oxolanes have been reported. Thus, when 2,3-O-benzylidene-tri-O-(methylsulfonyl)D-arabinitol, in a mixture of 9 M acetic acid and 10 M hydrochloric acid is heated for 30 minutes at loo", it gives 2,5-anhydro-l,4-di-O(methylsulfonyl)-D-arabinitol.86 Additional instances of the formation of internal anhydrides (not always well characterized) by acid hydrolysis of various sulfonic esters of alditols have been given by the same authors.86 The formation in low yield of 3,6-anhydro-4,5-0-isopropylidene-~allose dimethyl acetal (82), together with methyl 3-O-p-tolylsulfonylD-glucopyranoside (83) as the main product, by the action of boiling 2% methanolic hydrochloric acid (under reflux for 27 hours) on 1,2:5,6di-O-isopropylidene-3-O-p-tolylsulfonyl-~-~-glucof~~ose (81) has been reported.87

4C-0

Me

TsO "'$-o

Me

HI X

MeOH-HCl(28)

0 Me

(86) S. S. Brown and (7. M. Timmis, J . Chem. SOC., 3656 (1961). (87) R. Ahluwahlia, S. J. Angyal, and M. H. Randall, Carbohyd. Res., 4, 478 (1967).

208

J. DEFAYE

In the cyclitol series, Gorings obtained 2',5-anhydroquinicol (83b) by the action of 50% acetic acid (for 6 hours at 57") on 5-0-p-tolyl-

sulfonyl-epi-quinicol(83a).

P H

o

~

HOAco I&O

~

*" (834

l

O

O

H

(83b)

In another example, the formations9 of racemic 1,rianhydroribitol by the action of dilute hydrochloric acid for 20 hours at 100" on Dribitol 1-phosphate has likewise been assumed to take place by protonation of the oxygen atom of the ester function, followed by an intramolecular substitution by the hydroxyl group on C-4. Formation of this protonated phosphoric ester likewise explains the racemization observed, caused by partial migration of the ester group at C-1 to C-5. Another example of solvolysis of a sulfonate has been described by Buchanan and coworkers;90 it occurs without acid, but utilizes a highly polarizable sulfonic ester group. Solvolysis of methyl 2-0(p-nitrophenylsulfony1)-a-D-glucopyranoside (84) in water in the presence of sodium acetate for 6 hours at 100" leads, after reduction of the reaction mixture by means of sodium borohydride, to a fraction identified as 2,5-anhydro-~-mannitol (85). This reaction, which

where N s

= p-02NC,H4S0,-

(85)

(88)P. A. J . Gorin, Can.]. Chem., 41, 2417 (1963). (89) J. Baddiley, J. G. Buchanan, and B. Carss,]. Chem. SOC., 4058 (1957). (90)P. W. Austin, J. G. Buchanan, and R. M. Saunders,]. Chem. SOC. (C), 372 (1967).

2,5-ANHYDRIDES OF SUGARS AND RELATED COMPOUNDS

209

proceeds by a rearrangement of the Wagner-Meenvein type, resembles the nitrous acid deamination of 2-amino-2-deoxy-D-glucose, which also leads (after reduction of the product) to 85 (see p. 186).

5. Miscellaneous Methods It is conceivable that, starting from a pre-formed, five-membered heterocycle of the tetrahydrofuranol type, or even a furanosyl derivative, a contraction or extension of a side chain, according to the situation, would permit synthesis of a 2,5-anhydride of a sugar. This method has been used effectively on several occasions. a. Oxidative Cleavage of the Side Chain of a 3,8Anhydrohexose. 3,6-Anhydro-~-mannitolis readily obtaineds1 by the action of concentrated hydrochloric acid on D-mannitol. Protection of the hydroxyl groups on C-4 and C-5 as the isopropylidene acetal (86), followed by oxidation with lead tetraacetate, gives a quantitative yield of 2,5-anhydro-3,4-0-isopropylidene-~-arabinose (87).

2,5-Anhydro-~-xylose, characterized as the benzimidazole derivative formed from the corresponding 2,5-anhydro-~-xylonicacid and o-phenylenediamine, has likewise been preparede2 by the action of one molar equivalent of periodic acid on 1,4-anhydro-~-glucitol.

b. Attachment of a Side Chain to a G1ycofuranose.-The action of mercuric cyanide on 2,3,5-tri-O-benzoyl-~-ribofuranosyl bromide (88) in nitromethane affords,s3 in 88% yield, the corresponding nitrile (89) having the p-D configuration; hydrolysis of 89 with hydrochloric acid gives 2,5-anhydro-3,4,6-tri-O-benzoyl-~-allonic acid (90).This elegant synthesis also permits,s4 b y reduction of the nitrile 89, the synthesis of l-amino-2,5-anhydro-1-deoxy-~-al~itol (91), (91) A. B. Foster and W. G . Overend, J. Chem. SOC., 680 (1951). (92) C. F. Huebner and K. P. Link,]. Biol. Chem., 186, 387 (1950). (93) M. Bobek and J. Farkai, Collect. Czech. Chem. Commun., 34, 247 (1969). (94) M . Bobek and J. Farkai, Collect. Czech. Chem. Commun., 34, 1684 (1969).

J. DEFAYE

210

BzOCH,

0

BzO

Hg(CN), MeNO,

(90)

OBz (88)

(89)

characterized as its crystalline salicylidene Schiff base; compound 91 may be used directly as a precursor in the synthesis of analogs of nucleosides (see p. 218). 111. REACTIONS

The reactivity of the 2,5-anhydrides of aldoses is determined by two essential structural features that do not exist in the sugars, namely, the presence of an oxolane ring and of a carbonyl group (most frequently, free) a to the ring-oxygen atom. These two characteristics make the 2,5-anhydroaldoses closer to tetrahydro-2-furaldehyde than to the aldoses, where only in exceptional cases is the carbonyl group not masked by the formation of an intramolecular, five- or sixmembered, hemiacetal ring. 1. Stability of the Oxolane Ring The stability of cyclic ethers toward hydrolytic agents is largely a function of the size of the ring and, hence of the internal strain. Thus, whereas the oxiranes (epoxides) are particularly sensitive to these reagent^,^^.^^ the oxetanes are relatively more table.^' Oxolanes of the tetrahydrofuranol type seem, in particular, little affected by (95)R. E. Parker and N. S. Isaacs, Chem. Rev., 59, 737 (1959). (96)F. H.Newth, Quart. Rev. (London), 13, 30 (1959). (97)N. R. Williams, Adoan. Carbohyd. Chem. Biochem., 25, 109 (1970).

2,5-ANHYDRIDES OF SUGARS AND RELATED COMPOUNDS

211

these reagents; opening of the ring generally takes place only after a step leading to formation of an unsaturated system. This aspect involves consideration of the stability of furan systems to acid; such heterocycles are readily opened by acid.gs Certain authors have presumed that the presence of a carbonyl group (Y to the oxygen atom of the heterocycle renders the ring more sensitive to acid hydrolysis. In fact, however, the action of M hydrochloric acid for three hours at 100" on 2,5-anhydro-3,4-0-isopropylidene-D-arabinose leads75only to hydrolysis of the acetal protectinggroup, without affecting the oxolane ring. There is, however, a certain amount of decomposition under these conditions, because, after the action of methanolic hydrogen chloride on this hydrolyzate, the yield of 2,5-anhydro-~-arabinosedimethyl acetal is only -35%; c01ored, unidentified products constitute the major proportion of the hydrolyzate. F. Smith and coworkers74observed the formation of 2-furaldehyde by the action of 0.1 N acid on 2,5-anhydro-~-arabinose,an observation that has been confirmed by other author^.^^,^^ The formation of furan derivatives by acid-catalyzed dehydration of a 2,s-anhydroaldose can be ascribedg9 to the elimination of a group p to the carbonyl group. It is known that electron-attracting groups strongly reinforce the acidic character of the proton removed by nucleophilic agents in the course of this type of reaction. In contrast, the corresponding anhydroalditols are not affected under the same conditions of treatment.99 Furthermore, 2,5-anhydro-3-deoxy-~-erythro-pentose, which lacks a hydroxyl group p to the carbonyl group, is recoveredloo unchanged after treatment with boiling 0.05 M sulfuric acid for 2.5 hours under reflux. It is probable that a tetrahydrofuranol ring formed between two secondary alcohol groups is more stable toward nucleophilic agents than a similar type of heterocycle where at least one of the original alcohol groups is primary. This property is illustrated by the hydrolysis of 2,5:3,6-dianhydro-~-glucitol in concentrated hydrochloric acid for 8 hours at 100";this treatment cleaves the 3,6-anhydride ring, and 1eads'O' to 2,5-anhydro-6-chloro-6-deoxy-~-glucitol, isolated as its 173,4-triacetate. To date, the only reagent found capable of opening the ring of a (98) F. H. Newth, Adoan. Carbohyd. Chem., 6,83 (1951). (99) J. Defaye and S. D . CBro, Bull. Soc. Chim. Biol., 47, 1767 (1965). (100) J. Defaye and J. Hildesheim, unpublished results. (101) L. Vargha and J. Kuszmann, Carbohyd. Res., 8, 157 (1968).

J. DEFAYE

212

2,5-anhydroaldose is phenyIhydra~ine.~~ This fact has been the cause of a long polemic argument on the structure of 2,5-anhydro-~-mannose, because treatment thereof with phenylhydrazine gives Darabino-hexosulose phenylo~azone,~' not the anticipated phenylhydrazone of 2,5-anhydro-~-mannose. 2. Reactions Involving a Carbonyl Group a. Addition Compounds. - In contrast to the 3,6-anhydroaldohexopyranoses, where, despite a certain amount of strain between the two cyclic systems, it is somewhat unusual for the carbonyl group not to be masked by the formation of an intramolecular hemiacetal, the 2,5-anhydrides of aldoses are only in exceptional cases encountered as hemiacetals of this type. The alkyl 2,5-anhydro-p-~-and -CU-Larabinofuran~sides'~.~~ and l-(2,5-anhydro-~-~-lyxofuranosyl)uraci1~~ represent the sole examples yet known. It is noteworthy that ethyl 2,5-anhydro-a-~-arabinofuranoside in water at room temperature is hydrolyzed completely in 96 hours.74 The action of boiling 1.4% methanolic hydrogen chloride for two (92) hours under reflux on ethyl 2,5-anhydro-a-~-arabinofuranoside affords74a quantitative yield of the corresponding dimethyl acetal(93);

OEt

MeOH-HCI (1.4%) reflux, 2 hours

the reaction well illustrates the high degree of strain that exists in this bicyclic system. The same reagent applied to 2,5-anhydro-~the corresponding dimethyl acetal (95), only. mannose (94) gives36.99

HOCH,

0

MeOH-HC1 (1%)

OC=. reflux, 1 hour

HO

HO

2,5,-ANHYDRIDES OF SUGARS AND RELATED COMPOUNDS

213

N o formation of a methyl glycoside under these conditions has been detected. Inspection of molecular models of the 2,5-anhydroaldohexoses reveals that, when the carbonyl group and the primary alcohol group on C-6 are cis-disposed, as in 2,5-anhydro-aZdehydo-~-glucose (96) for example, the formation of a 1,6-hemiacetal (97) is possible without

HOCH

0

HC=O

HO 0

much strain. This factor is of particular relevance to the “mutarotational” behavior of the 2,s-anhydrides of aldehydo-aldoses. With the exception of 2,5-anhydro-~-glucose,’~,~~ “mutarotation” has been observed for all 2,5-anhydrides of aldehyde-aldoses reported to date (see Table I). Grant36 has attributed the “mutarotation” of 2,s-anhydro-D-mannose to the formation of a 1,4-hemiacetal, although he has shown that the action of methanolic hydrogen chloride on this comTABLEI Mutarotation of Various Free or Partially Substituted 2,5-Anhydro-aZdehydo-aldoses 2,5-Anhydride of D-Mannose

TemperConcen- ature ReferEquilibrium Solvent tration (degrees) ences

[ a l (degrees) ~ Initial +70.8 +33.6 +33.9 -55.8 +8.3 +15.3

1 hour

+21.9 +22.2 +27 -43 +16.6 12.5

D-Talose D-Ribose D-XylOSe D-LyXOSe 3,4,6-Tri-O2.5 hours 14.8 methyl-D-mannoSe +24 w 3,4-0-1~0propylidene-n73 hours -176 arabinose -126 3,4-O-Isopropylidene-D-ribose - 46.8 12 hours - 64.7

+ +

-

HzO H,O H,O H,O H,O H,O

2.6 2.2 0.44 0.86 0.60 3.2

20

25 25 25 25 25

43 68 68 68

CHCI,

2.5

25

99

20

74

27

117

CHCI, ~

35 99

J. DEFAYE

214

pound affords only the corresponding dimethyl acetal. On the other shows hand, 2,5-anhydro-3,4,6-tri-O-rnethyl-aldehydo-~-mannose “mutarotation” in certain solvents (see Table I), and 2,5-anhydro-3,40-isopropylidene-aZdehydo-~-arabinose behaves ~imilarly‘~;these products cannot form internal hemiacetals, 1n.view of these results, the apparent absence of “mutarotation” for 2,5-anhydro-aldehydo-~-glucose, to the impossibility of its forming an internal hemiacetal, remains to be confirmed. Instead of necessarily being the consequence of the formation of intramolecular hemiacetals, it seems likely that the “mutarotation” observed with the 2,5-anhydroaldoses could also be attributable to the formation of hydrates or of unstable hemiacetals with hydroxylated solvents, or even to dimerization. Known examples in related series lend support to these suppositions. It is known, for example, that 1,2-O-isopropylidene-5-~Zdehydo-a-~-xyZo-pentodialdo-1,4-furanose (98) readily forms102a dimer (99). Working with related com-

O=CH

QT

0

O+Me Me O+Me Me (99 )

pounds, Rosenthal and coworkers103have shown that 4,s-di-0-acetyl-

2,6-anhydro-3-deoxy-aldehydo-~-xyloand -D-lyxo-hexose (which, because of the absence of free hydroxyl groups, cannot form dimers of type 99) exist in solution partially in the hydrated form. Horton and coworkerslM have likewise shown that 1,2:3,4-di-Oisopropylidene -a-D-galacto-hexodialdo- 1,s-pyranose exists exclusively as the hydrate in deuterium oxide, and that this compound remains partially hydrated in moist chloroform. As early as 1931, W ~ l f r o mhad ~ ~shown ~ that the formation of hemiacetals from aldehydo sugars in hydroxylic solvents is a common occurrence. (102) R. Schder and H. S . Isbell, J . Amer. Chern. SOC., 79, 3864 (1957). (103) A. Rosenthal, D.Abson, T . D. Field, H. J. Koch, and R. E. J. Mitchell, Can. J . Chern., 45, 1525 (1967). (104) D.Horton, M. Nakadate, and J. M. J. Tronchet, Carbohyd. Res., 7, 56 (1968); D.Horton and J. D. Wander, ibid., 16, in press (1971). (105) M. L.Wolfrom,J.Amer. Chern. SOC., 52,2464 (1930);53,2275 (1931).

2.5-ANHYDRIDES OF SUGARS AND RELATED COMPOUNDS

215

Also relevant to this matter is the fact that the n.m.r. spectra of the in deuterium oxide rarely show106 free 2,5-anhydro-aZdehydo-aldoses the low-field signal characteristic of the free aldehyde group. Such a signal is generally visible when the spectrum (in chloroform-d) of compounds that are partially substituted is recorded, but its intensity is often weak. In 1925, Levene and Ulptsis proposed the possibility that polymerization was responsible for some of the unusual properties of the 2,5-anhydroaldoses. The formation of dimers, hydrates, and hemiacetals with the solvent can likewise be held responsible for the characteristic migration behavior observed for the 2,5-anhydroaldoses with aqueous alcoholic solvents on paper chromatograms.68

b. Dehydration. - The mechanism that leads to rapid transformation, in the presence of acid, of 2,5-anhydro sugars into furan derivatives has been discussed in the previous Section. Although the formation, in low yield, of an oligosaccharide that gives 2,5-anhydroL-idose and D-glucose on hydrolysis with acid had already been noted'*' as a result of the action of an acid on D-glucose or on starch, it seems unlikely that passage through such intermediates constitutes a fundamental pathway in the acid-catalyzed degradation of hexoses and pentoses. For example, the yield of 5-(hydroxymethy1)2-furaldehyde from 2,5-anhydro-~-mannoseis only 12%, whereas, under the same conditions, D-fructose gives 20-25% of the same furan derivative, and the aldohexoses afford the furan derivative in 1% yield.lo8 From present evidence,lo9 it seems that several pathways in each, of which the common stem is a 1,2-enediol precursor, are involved. It is, however, a plausible supposition that one of the minor pathways to 5-(hydroxymethyl)-2-furaldehydeinvolves direct passage by way of a 2,5-anhydro sugar. IV. UTILIZATION Although 2,5-anhydro sugars had, for a long time, been poorly understood, they have now found useful application as intermediates in synthesis, and even as derivatives for the characterization of certain amino sugars. For example, a useful method for the analytical deter(106) J. Defaye, unpublished results. (107) P. Slonimski, J . Defaye, J . Asselineau, and E. Lederer, Compt. Rend., 249, 192 (1959). (108) W. Alberda van Ekenstein and J. J. Blanksma, Ber., 43, 2355 (1910). (109) E. F. L. J. Anet, Aduan. Carbohyd. Chem., 19, 181 (1964).

J. DEFAYE

216

mination of 2-amino-2-deoxy-~-glucose and -D-galactose involves initial deamination with nitrous acid followed by treatment of the resulting 2,5-anhydroaldohexose with indole in an acid medium, to give a colored product having a maximum absorption at 492 nm.Ilo This color reaction is applicable to all of the known 2,s-anhydroaldoses, and appears to be relatively ~ p e c i f i c ; ~ "however, ~" it has been statede5that 2-deoxy-~-erythro-pentosegives a positive reaction in this test. As regards the use of2,5-anhydro sugar derivatives as intermediates in synthesis, the work of Hardegger"' on the synthesis of nonnuscarine (104) is particularly noteworthy. Treatment of the methyl ester (100) of 2-amino-2-deoxy-~-gluconicacid with nitrous acid gave the methyl ester (101) of 2,5-anhydro-~-gluconicacid; the diniethylamide (102) of this acid was readily converted into the corresponding tris(p-toluenesulfonate) (103) which, on reduction with lithium aluminum hydride, afforded normuscarine (104). 0, C '

,OMe I

H,NCH I

HCOH I

HOCH I

HOCH

J HO

TsO

Starting from the appropriate sulfonylated acetals 105, 106, and 107 (from 2,5-anhydro-~-ribose,-D-xylose, and -D-lyXOSe, respectively), the application of a method for elimination of vicinal, secondary disulfonates by reaction with sodium iodide in N,N-dimethylformamide in the presence of zinc112 has permitted113 the preparation of the (110)Z.Dische and E. Borenfreund,]. Biol. C h e n . , 184, 517 (1950). (111)E. Hardegger and F. Lohse, Helo. C h i n . Acto, 40,2383 (1957). (112)R. S . Tipson and A. Cohen, Carbohyd. Res., 1, 338 (1966). (113)J. Defaye and J. Hildesheim, Bull. SOC. Chim. Fr., 940 (1967).

2,5-ANHYDRIDES OF SUGARS AND RELATED COMPOUNDS

217

enantiomorphic R (112) and S (113) forms of tetrahydro-2-furaldehyde114 by way of the corresponding acetals (108 and 109) of the dihydro-4H-2-furaldehydes and those (110 and 111) of the tetrahydro-2-furaldehydes.

TsO

OTs

TsO (107)

NaI, Zn, HCONMe,

0

HC(OR),

0

(114)

(114) J. Defaye, Bull. SOC. Chim. Fr., 2099 (1968).

J . DEFAYE

218

The 2(R)-tetrahydro-2-furaldehyde (112) is reduced1l5 by sodium borohydride to the corresponding alcohol (114). 2,5-Dihydrofurfuryl alcohol can be prepared, starting from 2,5-anhydro-l-O-benzoyl-3,4di-0-p-tolylsulfonyl-D-ribitol, by using the Tipson-Cohen112 method for introducing the 3,4-double bond, with subsequent saponification of the ester."' Finally, the use of 2,5-anhydroalditols in the synthesis of nucleoside analogs that are modified in the base-sugar linkage may be cited. It is known that many synthetic analogs of nucleosides, having structural modifications either in the base or in the sugar, are readily inactivated in vivo by the action of glycosidases. The introduction of a methylene group into the base-sugar linkage, with retention of a furanoid ring, has been p r o p o ~ e d " ~asJ ~a~method of avoiding this enzymic destruction. A new series of nucleoside analogs of such types starting as 115, 116, 117, and 118 have thus been prepared,94J15*117**18 from derivatives of 2,5-anhydroalditols.

(117)

where R = OH, R' = H R = OH; R' = Me R = NH,; R' = H

(118)

where R = OH or NH,

(115) J . Defaye, M. Naumberg, and T. Reyners,]. Heterocycl. Chem., 6,229 (1969). (116) J. CIBophax, J. Hildesheim, and S. D. Gkro, Bull. SOC. Chim. Fr., 4111 (1967). (117) J. Defaye and T. Reyners, Bull. SOC. Chim. Biol., 50, 1625 (1968). (118) J. Defaye and T. Reyners, to be published.

2,B-ANHYDRIDES OF SUGARS AND RELATED COMPOUNDS

219

v. TABLEOF

PROPERTIES OF 2,5-ANHYDEUDES OF SUGARS, ALDITOLS, AND ALDONIC ACIDS

In Table I1 are given the melting point, boiling point, and specific rotation of the 2,5-anhydrides of sugars, alditols, and aldonic acids that have thus far been studied.

E3

TABLEI1

E3 0

Properties of 2,s-Anhydrides of Sugars, Alditols, and Aldonic Acids [ffltl,

Derivative

M.p., "C.

2,5-Anhydro-n-allo-pentitol 181.5-2.5 1-(6-azauracil-5-y1)3,4O-isopropylidenesyrup 119.5-21 1-[2,3,4,5-tetrahydro-3-thioxo-as-triazin-5-2(~-one-~yl 12,5-Anhydro-~-altro-pentitol 182-3.5 1-(6-azauracil-5-yl)240-1.5 3,4-0-isopropylidene233-5 1-[2,3,4,5-tetrahydro-3-thioxo-a~-tri~in-5-(2~)-one-6-y~]2,5-.4nhydroallaric acid 2,5-Anhydroallitol l-amino-l-deoxysyrup 3,4,6-tri-O-benzyl-l-(4-chloro-l,2-dihydropyrimidin-2-one-l-yl)-l-deoxyarnorph. amorph. 3,4,6-tri-O-benzyl-l-deoxy-l-(uracil-l-yl)88.5-9.5 3,4,6-tri-O-henzyl-l-deoxy-l-ureido213-15 1-(cytosin-1-y1)-1-deoxy191-2.5 l-deoxy3,4-O-isopropylidene-l-( uracil-1-yl)118-9 ldeoxy-1-(salicy1ideneamino)146-7 1-deoxy-(1-uracil-1-y1)amorph. l-deoxy-l-ureido2,5-Anhydro-~-allonicacid SYNP 215-7 4,6-diamino-5-aminopyrimidineester arnorph. 3,4,6-tri-O-benzoyl203-4 2,5-Anhydrogalactaric acid 2,5-Anhydro-~-galactonic acid 244 brucine salt

B.p., "C./torr

degrees +48.8

+ 109.2 -122.9 -183.4 0

Rotation solvent

Reference 129 129 129 129 129 129 16 94

-7.6 -1 +16.7 -52.5

94 94 94 94

+9.9 +55.2 +31.3 0

94 94 94 94 93 93 93 10

-9.4

12

-18.1 -42.9 +3.8

* v

5B

2,5-Anhydro-~-glucaricacid 2,5-Anhydro-~-glucaricacid 2,5-Anhydro-~-glucitoI 1,3,4-bi-O-acety1-6-chloro-6-deoxy1,6-di-O-(phenylsulfonyl)1,6-di-O-benzoyl2,5-Anhydro-~-gluconicacid 2,5-Anhydro-~-gluconicacid dimethylamide 3,4,6-tri-O-p-tolylsulfonylmethyl ester 2,5-Anhydro-aZdehydo-~-glucose (?) B,J-Anhydro-~-idaricacid 2,5-Anhydro-~-iditol

1,3,4,6-tetra-O-acetyl1,3:4,6-di-O-benzylidene1,6-dideoxy-1,6-diiodo1-0-p-tolylsulfonyl4,6-O-isopropylidene6-0-trityl1,6-di-O-p-tolylsulfonyl-

2,5-Anhydro-aldehydo-~-idose 3-O-benzyl-6-O-p-tolylsulfonyl-, dimethyl acetal 1-(2,2-diphenylhydrazone)

163 163 56-9 130-1 115-2OiO.25 109-9.5 137-8 135-7 amorph. 144 172 141-2 d. 175 d. 240 226 117-8 111-3 116.5-8.0

-19.1 -29

124-40/0.003 147-8 110-1 109-10 146 98- 100 syrup 146 148-9.2 151 amorph. SYNP 154-6

+38.53 -38.79 +24.4 -53 -4.4 +21 +3.2 +1.2 +70.3

-96 -93.3 t-9.7 + 12.6 +6.2 -13.2 +32.3 +97.1 + 105.7 +3.8 +8.1 + 16.3 +6.6 -6.6 -9 +11 20.4 +20

9 9 120 121 101 121 122 101 18 111 111 111 111

15 9.14

124 123 120 123 124 126 82 82,125 106 125 124,126 85 82 85 82 85 (continued)

N

E3

TABLEI1 (continued)

KJ

to bID1

Derivative

2,5:4,6-Dianhydro-~-idose &O-benzyl-, dimethyl acetal 2,SAnhydro-~mannaricacid 2,5hhydro-~-mannitol 1,3,4,6-tetra-O-acetylldeoxy- 1,l-difluoro6deoxy-6-iodo-3,4-di-O-p-tolylsulfonyl3,4,6t1i-O-methyl3,4,6-t~i-O-p-tolylsulfonyl1,3:4,6-di-O-methylene1,6-di-O-p-tolylsulfonyl1,6-di-O-trityl2,5-Anhydro-~mannonicacid

ethyl ester methyl ester 3,4,6-tri-O-methylmethyl ester 2,5-Anhydro-mmannose

1,3,4,6-tetra-O-acetyl-, (methyl “a”-hemiacetal) 1,3,4,6-tetra-O-acetyl-, (methyl “B”-hemiacetal) 4,6-O-benzylidene1-(2-benzy1-2-phenylhydrazone) diethyl dithioacetal dimethyl acetal

M.p., “C. 52-3 184-5 100-01 syrup syrup 85

B.p., “C./torr 132-5/0.1

degrees +41 +46.1 +58.2 +27.3; +26.4 +31.6

60-64/0.02 95 120 133.5 149 amorph. 106 syrup SYNP amorph. amorph. amorph. SYNP syrup

145-7 144-5 85 SYNP SYNP

+16 -59.1 r+ 6 +9.6 +60.7 +44.5 +38.3 +46.7

125/0.1 100/0.1

+69.9 +55.3 +33&1 +70.8++21.9 +33.6++22.2 +49.4 +46.3 -22.5 -22.4 +50.2 +60.4 +30.9

Rotation solvent

Reference

CHCI, HzO HzO CHCl, HzO

82 5,6,15 37,120 61 65 65 65 CHCl, 65 p-dioxane 127 35 C&N C&N 35 HzO 6 5%HCI in H,O 18 36 EtOH 36 HzO 36 HzO 36 H,O 19 H,O 35 H,O 99 CHC1, 61 CHCl, 61 Me,CO 46 Me2C0 47 36 EtOH Me&O 36 HzO 36

?

*

-I -4

M

3,4,6-tri-O-acetyl3,4,6-tri-O-methyl3,4,6-tri-O-methylsemicarbazone (p-nitropheny1)hydrazone

(2,4-dinitrophenyl)hydrazone 1-(2,2-diphenylhydrazone)

syrup 148 185 175 144-5

90/0.1 l2OlO.l

144-5 2,5-Anhydro-~talitol tetra-0-acetyl2,5-Anhydro-~-talonicacid, brucine s a l t 2,5-Anhydro-~-talose

112-3

t37.5 +7.1 +49.8 +30 +28.3 +44.5

MeOH EtOH MeOH

+7.4 +72

+ 193 +15.2 +121 +115 85-100/0.001 65-7510.135

-81.7 - 167 - 141

137 SYNP amorph. syrup syrup syrup

CHC13 HzO

-12.4 +20 +33.9++27

218 amorph. amorph.

(2,5-Anhydro-l,3,6-trideoxy-~-ribo-hexitol-l-yl)-t~imethyl179-80 ammonium chloride [L-(+)-Muscarinechloride] 2,5-Anhydro-P-~-arabinofuranose 194-5 l-deoxy-1-(5,6-dihydrouracil-l-yl)d. 260-2 1-deoxy-1-(uracil-1-y1)249-50 3-0-benzoyl241-3 3-0-benzyld. 196-8 3-O-(methylsulfonyl)2,5-Anhydro-a-~-arabinofuranoside ethyl methyl %carbanilate 3-0-methyl2,5-Anhydro-~-arabinose dimethyl acetal 3,4-O-isopropylidene-

110-15/0.005 160-5/0.3

+41.7 +39.8

115-20/0.01 118-24/0.03

-114 -22 -44.4 -131.5 - 126-+- 176 12

+

HzO HzO Hz0

61 36 36 36 36 36 35 42 43 43 12 19 43

H20

111

EtOH MeOH HZO

H20 HzO MeZCO p-dioxane Me,CO

76 76 76 76 76

0.1 M NaOH in HzO HzO Me&O

74

Hz0 HzO MeOH C.5& C6H6

H@

74 74 74 91 91 91 74 74

(continued)

Ls

tQ w

TABLEI1 (continued) [CIID,

Derivative

1-(2-benzyl-2-phenylhydrazone) dimethyl a c e d 2,s-Anh ydro-~arabinitol 3-O-acetyl-1,4-di-O-(methylsulfonyl)1,3,4-tri-O-benzoyl1,4-di-O-(methylsulfonyl)1,3,4-tri-O-p-tolylsulfonyl2,5-Anhydro-~-arabini to1

1,3,4-triazido-l,3,4-trideoxy1,3,4-tri-O-p-tolylsulfonyl2,5-Anhydro-~-arabinonic acid, methyl ester 2,5-Anhydro-~lyxitol 1,3,4-tri-O-benzoyI1,3,4-tri-O-(p-nitrobenzoy1)2,5-Anhydro-~-lyxitoI 4-azido-l-O-benzoyl-4-deoxy-3-O-p-tolylsulfonyl1,4-diazido-l,4-dideoxy-3-O-p-tolylsulfonyl1,3,4-tri-O-(p-nitrobenzoyl)1-(2,5-Anhydro-p-~-lyxofuranosyl)uracil 3-04methy1sulfonyl)2,S-Anh ydro-D-lyxose diisobutyl dithioacetal 3,4-di-O-p-tolylsulfonyldibutyl acetal 3,4-di-O-p-tolylsulfonyl2,5-Anhydro-~-ribitol 1-amino-1-deoxy-, hydrochloride

M.p., "C.

B.p., 'C./torr

degrees

Rotation solvent

115-20/0.005 125-35/0.24 syrup 75-6 117-8 153-4 128-9 115-25/0.09 120-2510.05 128 syrup 14S/0.0006 syrup 80- 1

Reference

+30.5 - 1.4 20.5 -0.5

MeOH HjO H,O

-36.1 +21 +27.4 +0.2 +34.7 -27.4

CHCI, CiHiN CHCl , H,O CHCI, CHCl,

74 74 74 128 86 128 86 74 74 130 74

t25.3 +77.5 -8.5

MeOH CH,CI, CHCl

74 68 128 68

-38

CHCl,

129-30

67-8 liquid 80-2 220-3 d. 190-1 amorph. 100-01 98

+85.1 +234 + 163 +15.3++12.5 +46.8 +35

syrup 98-9 164-8

-67 -79

CHCI, H,O p-dioxane H20 CHC1, CHCl

130 130 71 78 78 68 68 114

H,O H,O

114 68 119

w

z

?i

P

.e

m

l-azido-l-deoxy-3,4-O-isopropylidene62-510.005 l-azido-l-deoxy-3,4-di-O-p-tolylsulfonyl69-70 l-O-benzoyl-3,4-di-O-p-tolylsulfonyl98-9 l-O-benzyl-3,4-O-isopropylidene125-710.02 l-O-benzyl-3,4-di-O-p-tolylsulfonyl87-88 1-(6-arninopurin-9-y~)-l-deoxy-, hydrochloride 205-206 l-(cytosin-l-yl)-l-deoxy263-265 1-deoxy-1,l-difluoro75 l-deoxy-l-(2,4-dinitrophenyl)amino-3,4-O-isopropylidene-144,5-145,5 l-deoxy-l-(2,4-dini trophenyl)amino-3,4-di-O-p-tolylsulfonyl- 75 l-deoxy-l-iodo-3,4-O-isopropylidene6510.03 l-(N-acetylcytosin-l-yl)-l-deoxy-3,4-O-isopropylidene203-205 1-(5-amino-4-chloropyrimidin-6-yl)-l-deoxy-3,4-0-

isopropylidene1-(6-aminopurin-9-y1)l-deoxy-3,4-O-isopropylidene1-deoxy-1-( purin-6-ylamino)-3,4-O-isopropylidenel-deoxy-3,4-O-isopropylidene-l-(uracil-I-yl)l-deoxy-l-(purin-6-ylamino)1-deoxy-1-(thymin-1-yl)1-deoxy-1-(uracil-1-yl)-

3,4-0-isopropylidene1-0-p-tolylsulfon yl3,4-di-O-p-tolylsulfonyl1,3,4-tri-O-j~-tolylsulfonyl-

2,5-Anhydro-~-ribitoI 1,3,4-tri-O-benzoyl1-deoxy-1, l-difluoro2,5-Anhydro-~-ribose dibutyl acetal,3,4-di-O-p-tolylsulfonyl diisobutyl dithioacetal 3,4-O-isopropylidene-

145-146 185-187 244-245 179,5-180 230 178-180 198-199

-46 - 104 - 74 -28.6 -38 -51 -78.4 -26 -63.6 -82 -23.1 - 134 -98 -98

-85.3 -78

-67.5 -58.5 -69.6

60-510.005 62-63 118 125-126 99-100 73-75 73-4 amorph. amorph.

-40.1 -21.3 -53.5 -45.6 +66.7 +lo7 +25 -55.8+-43 -95.4

CHCI, CHCI, CHCI, CHCI, CHCl, H,O

KO

MeOH CHCI, CHCI, CHCI, MeOH

CHCI, MeOH MeOH CHCls H,O H,O HzO

MeOH CHCI, CHCI, CHCI:, H,O

CHCI, MeOH HzO

H,O

syrup

89-90 120-125/0.02

-85.6 -51.9

MeOH

CHCI,

119 130 116 116 116 117 118 64 119 130 118 118 117 117 119 118 119 118 118 119 119 116 116 69,71 128 64 68 67 114 67 117 (continued)

!2

TABLEI1 (continued) Derivative 3,4-di-O-p-tolylsulfonyl3,4-O-isopropylidene(p-nitropheny1)hydrazone dipropyl dithioacetal 2,5-Anhydro-~xylitol 1,3,4-tri-O-acetyl-

3,4-di-O-acetyl-1-0-trityl2,5-Anhydro-~-xylitol 1,3,4-tri-O-acetyl2,5-Anhydro-~xylose 3,4-di-O-acetyl-,dimethyl acetal dibenzyl dithioacetal 3,4-bis(phenylurethan) diisobutyl dithioacetal 3,4-di-O-p-tolylsulfonyldimethyl acetal, 3-0-p-tolylsulfonyl3,4-di-O-p-tolylsulfonyl(p-nitropheny1)hydrazone 2,5-Anhydro-3-deoxy-~-erythro-pentitol 1,4-di-O-acetyl2,5-Anhydro-3-deoxy-~-erythro-pentose diethyl dithioacetal 2,5-Anhydro-4-deoxy-~-erythro-pentitoI 1,3-di-O-benzoyl1,3-di-O-(p-chlorobenzoy1)-

M.p., "C.

90- 1 140 67-70

153

amorph. amorph. 98 168 94 76-8 76-8 86-7 196

amorph.

47-8 60-61

B.p., 'C./torr

blu, degrees

-50 -46.8+-64.7 50-5lO.02 +6.5++3.2 -91.4 10 150-55/0.2 -45.5 90/0.02 -39.3 90/0.02 +6 -11.2 k1 160-70/0.02 90-98/0.04 +31.5 k1.5 +9.0 +8.3++16.6 110-15/0.15 - 16 +159.6 +9.3 f11.5 +23.5 +53.7 ffi4.5 +145-++93.3 +1 *I 8W0.02 -40 55-60/0.02 +13.5++15 -t+8 - 15 90/0.01 105- 15/0.02 t48.1 -59.4 +68.1

+

Rotation solvent

CHCI, CHCI, MeOH MeOH HzO CHCI, CHzCIz CsH, HzO CH,CI, HzO HzO CHCI, CHCI, C,H,N CHCl, CHCln CHCI, CHCI, MeOH MeOH CHCI,

HzO CHCI, H20 CHzCI, CH,CI,

Reference 113 117 67 67 68 68 68 68 70 70 67 68 131 67 67 67 113 81 81,113 67 100 100 100

100 128 128 128

? ?1

t;M

1,3di-O-( p-methylbenzoy1)1,3-di-O-(p-nitrobenzoyl)Optically active di- and tetra-hydrofurfury1 derivatives formed from 2,5-anhydropentoses (2R)-2,5-Dihydro-2-fancarboxaldehyde dibutyl acetal diisobutyl dithioacetal (2s)-2,5-Dihydro-2-furancarboxaldehyde dibutyl acetal diisobutyl dithioacetal (2R)-2,5-Dihydro-2-frylmethanol O-benzoylO-benzyl(2R)-Tetrahydro-2-furancarboxaldehyde dibutyl acetal (2S)-Tetrahydro-2-furancarboxaldehyde dibutyl acetal (2R)-Tetrahydro-2-furylmethan01 O-benzoylldeoxy-l-iodoldeoxy-1-(uracil-1-y1)1-0-p-tolylsulfonylDianhydrides 2,5:3,6-Dianhydro-~-glucitol 4-0-acetyl-1-0-tritylldeoxyl-deoxy-l-iodo4-0-(rnethylsulfony1)ldeoxy-1-iodo-4-O-p-tol ylsulfonyll-deoxy-4-O-(methylsuIfonyl)1-deoxy4-0-p-tolylsulfonyl-

1,4-di-O-(methylsulfony1)1,4-di-O-p-tolylsulfonyl-

CH,Clz CHzClz

128 128

t97.8 +95.5

CHCl, CHCI, CHCI,

114 113 116

118 syrup

-123.3 -95.3 170 f101.2 +67 -3.4 +9.1 f4.2 - 10.7 -14.8 -23 19 -52.5 -15.5

CHCl, CHCl, MeOH CHCI, CHCl, MeOH CHCI, MeOH CHC1, CHCl, CHCI, CHCl, CHCl, CHCl,

114 114 116 116 116 114 114 114 114 115,132 116 115 115 115

117-9 143-5 81-2 111-3 113-4 92-3 85-6 123-5 123-4

+94.4 +30.7 107.9 +61.1 +46.7 +69.3 f60.1 +50.2 +37.7

HzO CHCl, Hz0 CHCl, CHCl, CHCI, CHCl, CHCl, CHCl,

101 101 101 101 101 101 101 101 101

+63.5 +65

63-5 137-8

90-95/0.1 90-95/0.02 103-05/0.01 90-95/0.05 90-95/0.005 120/40 78-80/0.01 80-8210.01 75-80/0.02 60-5/0.05 75-80/0.02 60-6510.005 85/18 80-8210.05 35-40/0.05

+123.5

+

+

+

rA

c Q

k v)

?-

z

U

E3

El

4

228

J. DEFAYE

(119) J. CIBophax, J. Defaye, and S. D. CBro, Bull. Soc. Chim. Fr., 104 (1967). (120) J. W. LeMaistre, personal communication (1965). (121) J. M. Sugihara and D. L. Schmidt,/. Org. Chem., 26, 4612 (1961). (122) R. C. Hockett, M. Zief, and R. M. Goepp, Jr., J. Amer. Chem. Soc., 68,935 (1946). (123) L. Vargha, T. Puskas, and E. Nagy, /. Amer. Chem. SOC., 70, 261 (1948). (124) L. Vargha and E. Kasztreiner, Chem. Ber., 93, 1608 (1960). (125) L. Vargha, Ber., 68, 1377 (1935). (126) L. Vargha and T. Puskcis, Ber., 76, 859 (1943). (127) A. K. Mitra and P. Karrer, Helo. Chim. Actu, 38, 1 (1955). (128) A. K. Bhattacharya, R. K. Ness, and H. G . Fletcher, Jr., /. Org. Chem., 28, 428 ( 1963). (129) M. Bobek, J. Farkag, and F. Sorm, Collect. Czech. Chem. Commnn., 34, 1673 (1969). (130) J. ClBophax, J. Hildesheim, R. E. Williams, and S. D. GCro, Bull. S O C . Chim. Fr., 1415 (1968). (131) J . Defaye and M. Naumberg, unpublished results. (132) The enantiomorphic tetrahydro-2-furylmethanols and certain of their derivatives can also be obtained by resolution of the racemic forms, or by cyclization of a 1,2,5-pentanetriol derivative. For relevant references, see Ref. 115.

ALDITOL ANHYDRIDES BY S. SOLTZBERG Atlos Chemical Industries, Inc., Wilmington, Delaware

I. Introduction.. . . .

.................................................

229

...........

1. Industrial.. .......................................................... 267 268 2. Biological.. .......................................................... VI. Tables of Properties of the Anhydrides and Their Derivatives. ............... 270

I. INTRODUCTION

An article on the anhydrides of polyhydric alcohols appeared in Volume 5 of this Series,' and it is the purpose of the present article to bring the subject up to date. Wiggins' article' was limited to the anhydrides of the pentitols and hexitols; the present article will include the anhydrides of tetritols and of alditols higher than the hexitols. However, anhydrides having the three-membered (oxirane) ring will not be considered, as they are discussed elsewhere in this Volume.la 2,5-Anhydrides of aldoses are also treated in this Volume.lb (1) L. F. Wiggins, Aduan. Carbohyd. Chem., 5, 191 (1950). (la) N. R. Williams, This Volume, p. 109. (lb) J. Defaye, This Volume, p. 181.

229

230

S. SOLTZBERG

In a broad sense, little that is new has been added to the procedures for synthesizing the anhydroalditols and their derivatives. However, three reactions, which will be discussed at the appropriate places, are of general interest. These are ( a )the oxidation of isolated hydroxyl groups to ketone groups in the presence of platinum oxide, (b) the isomerization of 1,4:3,6-dianhydrohexitols,and (c) the synthesis of C-aryl and -alkyl derivatives of anhydroalditols (C-glycosyl compounds, the so-called “C-glycosides”). A fourth very interesting reaction, of limited application, is formation, from 1,4:3,6-dianhydro-~glucitol (“isosorbide”) sulfonates, of 1,4:2,5:3,6-trianhydro-~-mannitol. As many derivatives as possible of the anhydroalditols are listed in Tables, with references to their synthesis.

11. SYNTHESIS

1. Anhydrotetritols 1,4-Anhydroerythritol (cis-3,4-tetrahydrofurandiol) has been obtained in 94.2% yield by dehydration of erythritol in the presence of an anionic resin.2 A mixture containing mainly 1,4-anhydro-o~threitol (truns-3,4-tetrahydrofurandiol),with some 1,4-anhydroerythritol, was formed by refluxing 2-butene-174-diol for 20 hours with hydrogen peroxide in the presence of such oxidizing acids as nitric or peroxysulfuric acid and salts of arsenic, bismuth, mercury, or tin;3 the yield of 174-anhydrothreitolranged from 60 to 84%, depending on the cation. Similarly, 2,5-dihydroxytetrahydrofuranwas converted by hydrogen peroxide in 85% formic acid into 3,4-tetrahydrofurandiol in high yield,4 but no mention was made of the distribution of isomers. Hartman and R. Barkefl synthesized 174-anhydroerythritol(3)by (1).The reaction the saponification of 1-0-p-tolylsulfonyl-D-erythritol apparently proceeded by direct reaction of the 4-hydroxyl group with the 1-p-tolylsulfonoxy group, and not through the intermediate formation of the l,e-anhydride, which would have afforded the epimer, namely, 1,4-anhydro-~-threitol(2). Similarly, 1-O-p-tolylsulfonyl-Dthreitol gave 1,4-anhydro-~-threitol(2). (2)F.H.Otey and C. L.Mehltretter,J. Org. Chem., 26,1673(1961). (3)Badische Anilin und Soda-Fabrik Akt.-Ges., Ger. Pat. 833,963 (Mar. 13, 1952); Chem. Abstracts, 52,10201(1958). (4)Badische Anilin und Soda-Fabrik Akt.-Ges., Ger. Pat. 855,861(Nov. 17,1952); Chem. Abstracts, 52,9215 (1958). (5)F.C.Hartman and R. Barker, J. Org. Chern., 28, 1004 (1963).

ALDITOL ANHYDRIDES H.,COTs I HCOH I HCOH I H&OH

231

4 p HO (2)

HO

OH (3)

However, saponification of 2-O-p-tolylsulfonyl-~-erythritol or of 2-O-p-tolylsulfonyl-D-threitol did not furnish any 1,4-anhydro compounds, but only the respective tetritol. Erythritol and L-threitol have also been dehydrated by means of 50% sulfuric acid at 120" to the corresponding 1,4-anh~drides.~ On heating erythritol in the presence of a small proportion of a xylenesulfonic acid at 130- 150" under vacuum, an 89%yield of 1,4-anhydroDL-erythritol was obtained.' 2. Anhydropentitols a. 1,4-Anhydropentitols. - The synthesis of anhydropentitols by acid-catalyzed dehydration of pentitols was discussed by Wiggins,' but the only pentitol mentioned was xylitol. A solution of ribitol in 2 M hydrochloric acid was heated for 24 hours at 100" to yield an by analogy with anhydroribitol assumed to be 1,4-anhydro-~~-ribitol~ the anhydridation of xylit01.~ It was also shown that, under the conditions applied to ribitol, L-arabinitol and D-xylitol afford negligible proportions of anhydrides.8 That the anhydro product from ribitol is a derivative of 1,4-tetrahydrofuran was established'O from the results of periodate oxidation (consumption of 1 mole of periodate per mole, with no formation of (6) H. Klosterman and F. Smith, J . Amer. Chem. SOC.,74, 5336 (1952). (7) C. M . Himel and L. 0. Edmonds, U. S. Pat. 2,572,566 (1951); Chem. Abstructs, 46, 6157 (1952). (8) J. Baddiley, J. G . Buchanan, B. Carss, and A. P. Mathias,]. Chem. SOC.,4583 (1956). (9) J . F. Carson and W. D . Maclay, J . Amer. Chem. SOC., 67, 1808 (1945). (10) J. Baddiley, J . G . Buchanan, and B. Carss,]. Chem. SOC..4058 (1957).

S. SOLTZBERG

232

formic acid) and the correspondence of the infrared (i.r.) spectrum (with minor variations) to that of an authentic sample of 1,4-anhydroD-ribitol. MacDonald and coworkers" corroborated the fact that L-arabinitol is resistant to anhydridation under conditions that, applied to ribitol, gave an almost quantitative yield of 1,4-anhydro-~~-ribitol. The resistance of L-arabinitol and D-xylitol to anhydride formation under relatively mild conditions may be due to greater nonbonded interactions than with ribitol. Although this situation is not readily apparent from the stereo structures shown, inspection of models reveals that the relatively bulky hydroxymethyl group on C-5 is sterically unfavorable to the 3-hydroxyl group of 6 (eclipsed arrangement); and for D-arabinitol(5), the hydroxymethyl and the 2-hydroxyl group are opposed transannularly. It would, therefore, appear that the nonbonded interaction of the cis-hydroxyl groups of 4 is weaker than those of 5 and 6, thus permitting more facile anhydride formation.12

OH

HO

HO

OH

Ribitol

D-Arabinitol

Xylitol

(41

(5)

(6)

In contrast to ribitol and xylitol (which form a DL mixture on anhydride formation because of their molecular symmetry), Darabinitol may, in principle, on heating with acid, give two different anhydrides, namely, a 1,4-anhydroarabinitol or a 2,5-anhydroarabinitol (1,4-anhydro-lyxitol), as the following reaction sequence illustrates.

(1 1) D. L. MacDonald, J. D. Crum, and R. Barker, J . Amer. Chem. SOC., 80,3379 (1958). (12) Baddiley and coworkers* speculated that the threitols would be more resistant to anhydride formation than erythritol, because their secondary hydroxyl groups are trans-disposed. It is of interest that Klosterman and Smith: using somewhat more vigorous conditions, had found earlier, that, for anhydride formation, a lower temperature and shorter heating period could be used for erythritol than for threitol, thus lending support to this speculation.

ALDITOL ANHYDRIDES

233

However, if the reasons already given for the differences in ease of (8) anhydridation of the pentitols are valid, 1,4-anhydro-~-arabinitol should be the chief product, because, in 7, not only would there be a transannular, nonbonded interaction between the hydroxymethyl group and the 2-hydroxyl group, but the (bulky) hydroxymethyl group and the 3-hydroxyl group would be in eclipse. H,YOH HOCH I

HYOH H,COH (7)

Baddiley and coworker^'^ further found Lat, under the same conditions (loo0 and 2 M hydrochloric acid), allitol (9) and D-altritol (D-talitol) (10) are converted into anhydrides in high yield, as shown CH,OH I HCOH I

ym HOCH I

HCOH I HCOH

HCOH I HCOH

HCOH

HCOH

I

I

C&OH (9)

I

I

CH,OH

(10)

by paper chromatography of the products. These hexitols, like ribitol, possess three contiguous, cis-hydroxyl groups. The hexitol anhydrides were not isolated, however, and their structures were not established, but it was supposed that they were the 174-anhydrides.DGlucitol, D-mannitol, and L-iditol were also stated to yield anhydrides to various extent^.'^ l74-Anhydropentito1s have also been synthesized by the reductive desulfurization of l-thioglycofuranosides.l Syrupy 1,4-anhydro-~arabinitol, characterized as the tris( p-nitrobenzoate), and crystalline 1,4-anhydro-~-ribitol'~ have been obtained by reduction of the cor-

(13) J . Baddiley, J. G. Buchanan, and B. Carss,J. Chem. Soc., 4138 (1957). (14) (a) R. Barker and H. G, Fletcher, Jr.,J.Org. Chenr., 26,4605 (1961).(b) F. Weyyand and F. Wirth, Chem. Ber., 85, 1000 (1952).

S . SOLTZBERG

234

responding acylated pentofuranosyl bromides with lithium aluminum hydride, a procedure introduced by Hudson and coworkers15during the synthesis of 1,5-anhydroalditols. On heating pentitols with acid under more drastic conditions, derivatives of 1,4-anhydrides may be produced. Thus, on treating molten xylitol with dry hydrogen chloride at loo", or in concentrated hydrochloric acid at 106-108",a moderate yield of syrupy 1,Panhydro5-chloro-5-deoxy-~~-xylitol was obtained.16 Several 2,5-anhydropentitols have been synthesized from suitably protected pentoses by standard procedures (see also, Ref. lb). Thus, ethyl 5-O-p-tolylsulfonyl-c~-~-arabinofuranoside (11) was converted" into the corresponding 2,5-anhydride ( 12) by treatment with methanolic sodium methoxide. The product was hydrolyzed, and the resulting 2,5-anhydro-~-arabinosewas hydrogenated in the presence of Raney nickel to give 2,5-anhydro-~-arabinitol(13) as a syrup.

("9"

uoEt

OH HO H,COTs (11)

OH (12)

(13)

The enantiomorph, also a syrup, was synthesized from 3,6-anhydro4,5-O-isopropylidene-~-mannitol'~ by periodate oxidation to the corresponding D-arabinose derivative, followed by reduction in the presence of Raney nicke1.l' 2,5-Anhydro-~-xylitol,2,5-anhydro-~ribitol, and 2,5-anhydro-~-lyxitol(1,4-anhydro-~-arabinitol)'~*~~ were (15) R. K. Ness, H. G. Fletcher, Jr., and C. S. Hudson,]. Amer. Chem. SOC.,72, 4547 (1950). (16) S. N. Danilov and V. F. Kazimirova, Sb. Statei Obshch. Khim., 2, 1646 (1953); Chem. Abstracts, 49, 6840 (1955). (17) M. Cifonelli, J. A. Cifonelli, R. Montgomery, and F. Smith,]. Amer. Chem. SOC., 77, 121 (1955). (18) A. B. Foster and W. G. Overend,]. Chem. SOC., 680 (1951). (19) J. Defaye, Bull. SOC.Chim. Fr., 2686 (1964). (20) DefayelS called attention to the fact that the 2,5-anhydro-tri-O-(p-nitrobenzoyl)-~lyxitol that he prepared is the enantiomorph of Barker and Fletcher's 1,4-anhydrotri-O-(p-nitrobenzoyl)-~-arabinitol.'~'~' This is one of the relatively few instances on record where alditol anhydrides derived, without Walden inversion, from two configurationally different sugars have been shown experimentally to possess the same, although enantiomorphic, configuration.

235

ALDITOL ANHYDRIDES

obtained by reduction of the corresponding 2,5-anhydro-~-pentoses with sodium borohydride. The xylitol and arabinitol derivatives were syrups. The anhydro-D-arabinitol was characterized as the 3,4-diO-acetyl-l-O-trityl derivative. 2,5-Anhydro-l-deoxy-l,l-difluoro-~-ribitol~~ (15) was, by an unusual ring-contraction, obtained from di-O-acetyl-D-arabinal (3,4-di-0acetyl-1,5-anhydro-~-en~th~o-pent-l-enitol) (14) by way of the following reaction sequence.

AcO

AcO

(14)

OAc

HO

OH (15)

Also, di-O-acetyl-L-arabinal was converted into the corresponding anhydro-L-ribitol derivative (see also, pp. 195- 197). The configuration of the products was established by periodate oxidation of the D compound, and reduction of the resulting dialdehyde to an optically active diol whose sign and magnitude of optical t2.5"in water) correspond to that of the diol ( [a]iO t-6.8" rotation ( [a]iO in water) obtained by similar treatment of methyl a-L-arabinopyranoside. The bis(p-nitrobenzoates) of the two diols exhibited similar relationships. Hence, it appears likely that the configuration at C-2 of the anhydride is the same as that at C-1 of the L-arabinoside. The cis relationship of the hydroxyl groups was retained from the arabinal started with. The formation of oxetane (1,3-anhydro) rings has received but was oblittle attention. 1,3-Anhydro-2,4-0-methylene-~~-xylitol tained'l from 2,4-0-methylene-l-O-p-to~y~su~fony~-D~-xy~ito~. The 1,3-ring was established by elimination of the possibility of other products, through test reactions and formation of derivatives.22 No work on the synthesis of 1,5-anhydropentitols appears to have been reported since the earlier article.' Only two dianhydropentitols have been reported. One was ob(21) (a) P. W. Kent, J . E. G. Bamett, and K. R. Wood, Tetrahedron Lett., 1345 (1963). (b) K. R. Wood and P. W. Kent,J. Chem. Soc. (C), 2422 (1967). (22) R. M. Hann, N . K. Richtmyer, H. W. Diehl, and C. S. Hudson, J. Amer. Chem. Soc., 72, 561 (1950).

236

S. SOLTZBERG

tained, as a syrup, by refluxing a solution of 174-anhydro-5-chloro-5deoxy-DL-xylitol'6 in dry acetone with dry sodium hydroxide, or by the action of sodium methoxide in acetone at room temperature; it was supposed that the product was 1,4:2,5-dianhydro-~~-xylitol (16). Wiggins suggested that a dianhydro-xylitol obtained by GrandelZZa by heating xylitol with methanedisulfonic acid was the 1,4:2,5dianhydride (16). The second dianhydro-xylitol was obtainedz3by methylenation of 1,4-anhydro-~~-xylito1, with paraformaldehyde and concentrated hydrochloric acid, to give the 3,5-methylene acetal; this was methylated to the 2-methyl ether, which was hydrolyzed, and the product mono-p-toluenesulfonylated to the 5-p-toluenesulfonic ester. On treatment with aqueous sodium hydroxide at 40-50", the last gave syrupy 1,4:3,5-dianhydro-2-O-methyl-~~-xylitol (17, D form).

The crucial methylenation step in this synthesis undoubtedly gave the 3,5-methylene acetal, not the 2,3-acetal (or the highly improbable 2,5-methylene acetal), because, otherwise, a 6-O-methyl-2(or 3)-0p-tolylsulfonyl compound would have been obtained which could only have formed an epoxide (oxirane). T h e oxirane ring might have been opened under the conditions of the saponification; the product would not then have been a dianhydride. 3. Anhydrohexitols a. Anhydrohexitols. - Synthesis of 1,4-anhydrohexitols by new procedures has not been reported. However, 1,4-anhydro-~-mannitol (224 F. Crandel, U. S. Pat. 2,375,915 (1945), characterized, by boiling point only, the dianhydro-xylitol that he prepared (b.p. 170"/5 torr). As xylitol is symmetrical, an anhydride ring would b e formed first at C-l,C-4 (= C-5,C-2). Examination of Fisher-Hirschfelder models reveals that a second tetrahydrofuran ring would b e extremely difficult to form unless inversion had occurred at some point. This is a possibility (see Ref. 82). T h e same steiic considerations apply to the compound mentioned in Ref. 16. (23) G. E. Ustyuzhanin, E. M. Kogan, N. S. Tikhomirova-Sidorova, and S. N. Danilov,

Zh. Obshch. Khim., 33,3622 (1963); Chem. Abstracts, 59, 5246 (1963).

ALDITOL ANHYDRIDES

237

was isolatedz4 as one of the products formed on heating D-mannitol with sodium pyrosulfite solution at 130-180". The reduction of per-0-acylaldopyranosyl halides has been shown to constitute a definitive synthesis of 1,5-anhydrohexitols. Thus, Hudson and coworkers'5 reduced tetra-O-acety~-a!-D-glucopyranosyl bromide, and tri-0-acetyl-a-L-rhamnopyranosyl bromide with lithium aluminum hydride; deacetylation of the respective product gave 1,5-anhydro-~-glucitol (polygalitol) and 1,5-anhydro-~-rhamnitol. cox or^^^ used this method for synthesizing 1,5-anhydro-~-allitolfrom D-allOSe. He prepared the enantiomorph (18) by reduction of 2,3,4tri-0-benzoyl-D-ribopyranosyl cyanide, followed by saponification, reduction of the product with lithium aluminum hydride (with simultaneous deacetylation), and deamination.25 2,6-Anhydro-~-aIlitol (19) is 1,5-anhydro-~-allitol(IS).

HO

OH

Because of an analogous relationship, 1,9anhydro-~-mannitol (styracitol) was obtained by hydrogenation of 2,6-anhydro-~-mannose.26Also, 1,5-anhydro-~-gulitol,obtained by reduction of tetra-0acetyl-D-gulopyranosyl bromide with lithium aluminum hydride, and its enantiomorph, obtained2' by similar reduction of tetra-o-benzoylP-D-fructopyranosyl bromide are 2,6-anhydro-~- and -D-glUCitOl, respectively. It should be noted that reduction of the tetra-o-acylketohexopyranosyl halide results in inversion at C-2; hence, the D-glucitol derivative is formed (the 1,5-anhydro-~-mannitol derivative was obtained in very low yield). Because a solution of the bromide in ether was dropped into a solution of the reductant, the aluminum bromide or aluminum bromide etherate formed in the initial reduction may have anomerized the succeeding portions of the D-fructose derivative. With per-0-acylaldopyranosyl halides, epimerization at C-1 does not affect the configuration of the resulting anhydroalditol. (24) B. Lindberg and 0. Theander, Suensk Papperstidn., 65, 509 (1962); Chem. Abstracts, 58, 9773 (1963). (25) B. Coxon, Tetrahedron, 22, 2281 (1966). (26) F. Micheel, W. Neier, and T. Riedel, Chem. Ber., 100, 2401 (1967). (27) R. K. Ness and C. S. Hudson, J . Amer. Chem. Soc., 75, 2619 (1953).

238

S. SOLTZBERG

1,5-Anhydro-D-altritol was obtained2s in low yield by reduction of tetra-0-benzoyl-D-altropyranosyl bromide with lithium aluminum hydride, in contrast to the high yields of anhydroalditols generally obtained by this technique. It is possible that most of the bromide was decomposed prior to reduction, as a very low yield of crude intermediate was obtained at this stage. A different approach,28starting with 1,5-anhydro-~-glucitol,afforded 1,5-anhydro-~-altritolin satisfactory overall yield. 1,5-Anhydr0-4,6-O-benzylidene-2,3-di-Op-tolylsulfonyl-D-glucitol was transformed in methanolic sodium methoxide into the corresponding 2,3-anhydride whose configuration was not established; this was treated with hot, aqueous alkali to give 1,5-anhydro-4,6-0-benzylidene-~-altritol, which was hydrolyzed to the anhydroalditol. The physical constants of the product differed from those of 1,5-anhydro-~-glucitolor -D-mannitol, and hence, it had the D-altritol configuration. A different reductive approach was employed by Zervas and Zioudrou,29 who catalytically reduced tetra-0-acetyl-a-D-glucopyranosyl bromide in the presence of palladium to obtain 1,5-anhydroD-glucitol. It would be of interest to determine whether this or a similar reduction of tetra-0-benzoyl-P-D-fructopyranosyl bromide would give the D-mannitol anhydride. Synthesis of 1,5-anhydroalditols by use of a compound already having a pyranoid ring was accomplished by Rice and Inatome.so They reduced tetra-0-acetyl-a-D-glucopyranosyl nitrate and tri-0acetyl-P-L-arabinopyranosyl nitrate with sodium borohydride, and obtained the corresponding 1,Sanhydroalditol acetates; saponification afforded the corresponding 1,5-anhydroalditols. As the acetylated glycosyl halides had been used for preparing the nitrates, the procedure has no advantage over reduction of the halide. Inasmuch as it was shown that sodium borohydride does not reduce nitric esters of primary or secondary alcohol groups, nitrated sugars are not suitable starting materials. In a different synthesis of 1,5-anhydrohexitols from a compound having a pyranoid ring, Lehmann and Friebolinsl treated 1,Sanhydro2-deoxy-~-arabino-hex-l-enitol (D-glucal) (20) with a-toluenethiol in the presence of light, and obtained 1,5-anhydro-2-S-benzyl-2-thio-~mannitol (21) and the epimeric anhydro-D-glucitol in equal amounts. (28) E. Zissis and N. K. Richtmyer, I . Amer. Chem. SOC., 77, 5154 (1955). (29) L. Zervas and C. Zioudrou, J. Chem. Soc., 214 (1956). (30) F. A. H. Rice and M. Inatome, J. Amer. Chem. SOC., 80, 4709 (1958). (31) (a) J. Lehmann, Carbohyd. Res., 2, 486 (1966); (b) J. Lehmann and H. Friebolin, ibid., 2, 499 (1966).

239

ALDITOL ANHYDRIDES

(21)

(20)

where Bzl = PhCH,.

The configuration at C-2 was established by comparison with the optical activities of 1,5-anhydro-~-mannitoland -D-ghcitOl and the proton magnetic resonance (p.m.r.) spectra of their acetate^.^"^' This reaction is useful for the preparation of C-2 substituted derivatives, but the configuration at C-2 of the products must then be established. T r i - O - a c e t y l - 1 , 5 - a n h y d r o - 2 - d e o x y - ~ - a r a b i o (tri-0l acetyl-D-glucal) was the starting point for the synthesis of 1,5-anhydro-2-chloro-2-deoxy-~-glucitol or -D-mannitol by way of the 1,2dichloride (22), which was reduced:j2b y means of lithium aluminum hydride to the 2-chloro-2-deoxy derivative (23).That the chlorine AcOCH, AcO

AcOCH,

-

..

A

c

O

Y

o

\

A

C T “ c1

atom was situated on C-2 was inferred by analogy with the comparable reduction of peracetylated glycosyl halides (see p. 237) and from its low reactivity toward sodium hydrogen carbonate or silver nitrate, as compared with that expected for an a-chloro ether. Because the distilled product, obtained in good yield, was a viscous syrup, it may have consisted of a mixture of the C-2 epimers. Anhydroalditol compounds of a rather unusual type are the Cglycosyl compounds (erroneously called “C-glycosides”). As pointed out by Treibs,”%these are not glycosides, but they are anhydroof approaches has been employed for synthesis of a l d i t o l ~A. ~variety ~ (32) C. D. Hurd and H. Jenkins, Carbohyd. Res., 2, 240 (1966). (33) W. Treibs, Nuturwissenschuften, 48, 378 (1961); Chem. Abstracts, 55, 24705 (1961). (34) See Ref. 37, footnote 2. It is now suggested that these compounds be named similarly to the alditols higher than the hexitols, because an asymmetric carbon atom has been introduced at C-1 of the anhydride.

240

S. SOLTZBERG

this class of compound. Hurd and Holysz35treated tetra-O-acety1-aD-glucopyranosyl chloride or tetra-0-acetyl-a-D-mannopyranosyl bromide with such metallo-organic compounds as phenyllithium and butyllithium. The reaction between the former glycopyranosyl chloride and phenyllithium gave three products, which were separated by chromatography on alumina. Two of the compounds were shown to be the tetraacetates of the known36P-D-glucopyranosyl- (24) and a-D-glucopyranosyl-benzene (25). The third product has been in(26). ferred to be 1,5-anhydro-2-C-phenyl-~-glucito1~~

HO HO

HO

HO

HO

Ph

The formation of the tetraacetate of the a-D anomer (25a) (no inversion) has been postulated to take place by way of a four-center transition state:

It was presumed that the third product (26) results from elimination of the elements of hydrogen chloride, in addition to deacetylation, by the strong base, to give a 2-hydroxy-~-glucal (1,5-anhydro-~arabino-hex-1-enitol) anion (27) which then adds phenyllithium

(27)

(35)C.D.Hurd and R. P. Holysz, ]. Amer. Chem. SOC., 72, 1735 (1950). (36)W.A. Bonner and J . M. Craig,J. Amer. Chem. SOC., 72, 3480 (1950). (37)C.D.Hurd and H. T. Miles,]. Org. Chem., 29,2976(1964).

ALDITOL ANHYDRIDES

241

across the double bond. As proof of this theory, 2-hydroxy-~-glucal tetraacetate was treated with phenyllithium, and the mixture was processed as for the reaction of tetra-O-aCetyl-cu-D-glUCOpyranOSyl chloride. It was demonstrated, both by paper chromatography of the deacetylated, partially purified, crude material and by gas-liquid chromatography of the (fully acetylated) material, that the principal component of the mixture corresponded in behavior to pure 1,5anhydro-2-C-phenyl-~-glucitol. This configuration,,instead of the D-manno form, is based on the assumption that the tertiary hydroxyl group of the hypothetical Dmanno derivative, being in a favorable, trans-coplanar arrangement with the axial hydrogen atom on C-1, should be dehydrated when it is acetylated under forcing conditions. On the other hand, the tertiary hydroxyl group of the D-gluco configuration is not trans-coplanar with any neighboring C-H bond. Inasmuch as dehydration did not occur on forced acetylation (acetic anhydride and sodium acetate at 140"), it was decided the compound had the D-ghco configuration. Mild acetylation at loo", or in acetic anhydride-pyridine at 25", yielded only a triacetate (not a tetraa~etate).3~ The 1,s-anhydride, namely, P-D-glucopyranosylbenzene (24), was formed by the nucleophilic-displacement type of reaction expected. The reaction of tetra-O-acetyl-a-D-mannopyranosylbromide with phenyllithium likewise gave a mixture; from it, only tetra-O-acety1-aD-mannopyran~sylbenzene~~ was isolated crystalline. The two other products have, apparently, not been further investigated. A mixture of products was obtained from the reaction of butyllithium with tetra0-acetyl-a-D-ghcopyranosyl bromide; from it, a crystalline tetra-0acetyl-D-glucopyranosylbutanewas isolated, and identified as the a anomer, because of its identity (by mixed melting point) with the broproduct from the reaction of tetra-O-acetyl-a-D-glucopyranosyl mide with the butyl Grignard reagent. The use of the Grignard reaction to synthesize 1-C-substituted carbohydrate derivatives from peracetylated glycosyl halides was . ~ ~reaction gives more introduced in 1945 b y Hurd and B ~ n n e rThis homogenr-ms products in much better yields than the organolithium reaction. However, the product isomeric at C-1 is generally produced simultaneously to a greater or lesser extent. Only sporadic use was made of the reaction until the late 1950's, when Zhdanov and coworkers undertook the synthesis of numerous (38)C.D.Hurd and R. P. Holysz,J. Amer. Chem. SOC.,72, 1732 (1950). (39)C.D.Hurd and W. A. Bonner,]. Amer. Chem. Soc., 67, 1972 (1945).

242

S. SOLTZBERG

l-C derivatives embracing a variety of substituted phenyl compounds and such unsaturated acyclic groups as ally1 and vinyl.40 That it is not necessary to employ substituted aryl compounds in order to obtain a substituted product was demonstrated by Gerecs and Windholz,4' who nitrated tetra-O-acety~-~-D-glucopyranosylbenzene with cuprous nitrate in acetic anhydride to give a mixture (-4: 1)of the corresponding o- and p-nitro derivatives. D-Glucopyranosylation of an acetylene compound was achieved by treating tetra-0-acetyl-a-D-glucopyranosyl bromide with 2-phenylethynylmagnesium bromide.42 The product, namely, tetra-o-acetylp-~-glucopyranosyl-( 2-phenylethyne), was reduced to the corresponding phenethyl derivative, which was obtained directly by use of phenethylmagnesium bromide. The p-D configuration was inferred from the levorotation of the crystalline products. The mother liquors were dextrorotatory, and were assumed to contain the a - anomers. ~ The tetra-O-acetyl-(phenylethyne) derivative could be crystallized in both the anhydrous and the monohydrated form, depending on the solvent used. From 95% ethanol containing ethyl ether, the deacetylated product was also obtained as a solvate, but it was not determined whether it contained water or alcohol of solvation. An attempt was made to extend the reaction to sodium acetylide, but only a small amount of an unidentified, crystalline, carbohydrate derivative (40) (a) Yu. A. Zhdanov and G. N. Dorofeenko, Dokl. Akad. Nauk SSSR, 112,433(1957); (b) ibid., 113, 601 (1957); Chem. Abstracts, 51, 13767, 14561 (1957). (c) Yu. A. Zhdanov, G. A. Korol'chenko, and S. I. Uvarova, ibid., 122, 811 (1958); Chem. Abstracts, 53, 4183 (1959). (d) Yu. A. Zhdanov, G. A. Korol'chenko, and L. A. Kubasskaya, ibid., 128, 1185 (1959); Chem. Abstracts, 54,8644 (1960). (e) Yu. A.

Zhdanov, G. A. Korol'chenko, L. A. Kubasskaya, and R. M. Krivoruchko, ibid., 129, 1049 (1959); Chem. Abstracts, 54, 8641 (1960). (f) Yu. A. Zhdanov, G. N. Dorofeenko, and L. E. Zhivoglazova, ibid., 117, 990 (1957); Chem. Abstracts, 52, 8056 (1958). (9)Yu. A. Zhdanov, G. A. Korol'chenko, G. N. Dorofeenko, and G. V. Bogdanova, ibid., 152, 102 (1963); Chem. Abstracts, 59, 15373 (1963). (h) V. I. Komilov, S. S. Doroshenko, V. I. Tikhonov, and Yu. A. Zhdanov, Materialy Vses. Konf. Probl. Khim Obmen Uglevodou, 3rd, Moscow, 1963,85 (1965). (i) Yu. A. Zhdanov and V. I. Komilov, Izo. Vysshikh Uchebn. Zavedenii Khim. Khim. Teknol., 9, 65 (1966); Chem. Abstracts, 65, 13806 (1966). fj)Yu. A. Zhdanov and G. A. Korol'chenko, Dokl. Akad. Nauk SSSR, 139, 1363 (1961); Chem. Abstracts, 56, 531 (1962). (k) Yu. A. Zhdanov, G . A. Korol'chenko, G. N. Dorofeenko, and G. V. Bogdanova, ibid., 152, 102 (1963); Chem. Abstracts, 59,15373 (1963). (1) Yu. A. Zhdanov, Uglevody i Uglevodn. Obmen v Zhiootn. i Rust. Organizmakh Materialy Konf., Moscow, 1958,53 (1959); Chem. Abstracts, 54, 19503 (1960). (41) A. Gerecs and M . Windholz, Acta Chim. Acad. Sci. Hung., 13, 231 (1957); Chem. Abstracts, 52, 11778 (1958). (42) R. Zelinski and R. E. Meyer, J . Org. Chem., 23, 810 (1958).

ALDITOL ANHYDRIDES

243

having an analysis corresponding to the composition calculated for C16H2009 was obtained. A reaction that appears to be of fairly broad application, and that yields a mixture of compounds containing some 1-C-substituted anhydroalditol, consists in treating a free reducing sugar, an oligomer, or a polysaccharide with an aromatic compound in the presence of liquid hydrogen fluoride. The procedure was first reported by Linn.43However, the structure of the anhydride obtained was not fully elucidated; nor were the structures of the compounds obtained from the acidcatalyzed dehydration of l-deoxy-l,l-bis(3,4-dimethylphenyl)-~g l u ~ i t o l , whose ~~ configuration was not rigorously established. It has been found that numerous plants contain 1-C-substituted anhydroalditols. For example, barbaloin, a major constituent of commercial aloin, is l0-(a-~-glucopyranosyl)-1,8-dihydroxy-3-(hydroxymeth~1)anthrone.~~ Numerous other 1-C-substituted anhydroalditols have been isolated from a variety of plant sources by various investigators. The antibiotic substances ~ h o w d o m y c i n ,formycin ~~ and formycin B,47348and l a u r u ~ i n *have ~ been shown to be 1,4-anhydroribitol-l-y1 compounds. It was proved that laurusin and formycin B are the same compound.48 A new synthesis of 2,5-anhydro-~-iditolwas achieved49by reduction of 2,5-anhydro-~-idosewith sodium borohydride. The product was identified as its crystalline 1,6-bis(p-toluenesulfonate), which was already known.' The synthesis of an alditol having a 4-membered (oxetane) ring was first reported by Ustyuzhanin and coworkers,50who prepared 1,3anhydro-5,6-di-O-methyl-2,4-0-methylene-~-glucitol by saponification of the 1-p-toluenesulfonate of the corresponding derivative of D(43)C.B. Linn, Am. Chem. Soc., Div. Petrol. Chem., Preprints, 2, No. 3, 173 (1957). (44)J. Heerema, G.N. Bollenback, and C. B. Linn, Am. Chem. SOC. Div. Petrol Chem., Preprints, 2, No.4,185 (1957);1.Amer. Chem. SOC., 80, 5555 (1958). (45)(a) J. E. Hay and L. J. Haynes,]. Chem. SOC., 3141 (1956).(b) R.A. Barnes and W. Holfeld, Chem. Znd. (London), 873 (1956). (46)K. R. Darnall, L. B. Townsend, and R. K. Robins, Proc. Nut. Acad. Sci. U . S . , 57, 548 (1967). (47)G . Koyama, K. Maeda, and H. Umezawa, Tetrahedron Lett., 597 (1966). (48)R. K. Robins, L. B. Townsend, F. Cassidy, J. F. Gerster, A. F. Lewis, and R. L. Miller, J . HeterocycL Chem., 3, 110 (1966). (49)C. A. Dekker and T. Hasizume, Arch. Biochem. Biophys., 78, 348 (1958). (50)G. E. Ustyuzhanin, N. S. Tikhomirova-Sidorova, and S. N. Danilov, Zh. Obshch. Khim., 33, 453 (1963);Chem. Abstracts, 59,5247 (1963).

S. SOLTZBERG

244

glucitol. Later, Haslam and R a d f ~ r dwhile , ~ ~ investigating the synthesis of 1,5-anhydro-2,4-0-benzylidene-~-glucitol, prepared 5,6-anhydro-2,4-O-benzylidene-l-O-p-tolylsulfonyl-~-glucitol. On saponification, instead of the 5,6-anhydro-2,4-0-benzylidene-~-glucitol expected, de-p-toluenesulfonyloxylation occurred, affording a crystalline anhydride (28) that did not agree in physical properties with the monoanhydromonobenzylidenehexitols known. The possibility that it was a 1,5-anhydro-~-iditolcompound was excluded on the basis of the following evidence. The nuclear magnetic resonance (n.m.r.) spectrum of the ethers obtained by treatment of the 5,6-anhydride with an alcohol and a base, and of their monoacetates, indicated that the free hydroxyl group in the ether was attached to a methine carbon atom (not to a methylene group); this showed that the 5,6-anhydro ring had opened normally. Hence, the reactions must have proceeded as shown. &COTS I HCO I \ HOCH CHPh I / HCO I

r (I F \

0 HCO

NaOH_LCH I

/

CHPh

HCO I HCOH I H,COH

lX;i;:,

@ ,:

AcSO

OCH CHPh I /

HCO I HCOAc I

GCOAc

(28)

E: I

\

-

CKOTs I

HCO I

\

OCH CHPh ROHsbase HOCH CHPh I / I / HCO HCO

I HCOH I H,COR

HkOH

I H,COTs

where R = Me, Et, or PhC&.

It is probable that the two 1,Sanhydro compounds just described are anhydrides of the same alditol. The compound of Ustyuzhanin and coworkerss0 is unequivocally a 1,3-anhydro-~-glucitolderivative, because the hydroxyl groups at C-2,4,5, and 6 were protected prior to the ring formation. Positive identity of the two would have been achieved by converting the monomethyl ether obtained by Haslam (51)(a) E.Haslam and T. Radford, Chem. Commun., 631 (1965).(b) E.Haslam and T. Radford, Carbohyd. Res., 2, 301 (1966).

ALDITOL ANHYDRIDES

245

and Radford5' into the diether. Moreover, periodate oxidation of 28 could have provided incontrovertible chemical proof, through the formation of formaldehyde, that ring closure could only have taken place between C-1 and C-3. Oxepane (seven-membered) rings (1,6-anhydrohe~itols)~* have been prepared from the 3,4-isopropylidene acetals of D-mannitol, Dglucitol, and L-iditol, by way of alkaline hydrolysis of the corresponding 1,2:5,6-dianhydrides. The ring structures of the products were established through periodate oxidation and lead tetraacetate oxidation; the requisite amount of formic acid was produced, and 3 equivalents of lead tetraacetate were consumed. No inversions at any of the asymmetric centers were involved in the reactions conducted, so the oxepanes had retained the configurations of the starting hexitols. In the synthesis of 1,6-anhydro-~-ghcitol,at least, a 2,6-anhydride is a b y p r ~ d u c tIts .~~ structure was assigned on the basis of its i.r. spectrum, in comparison with those of the 1,6-and 1,5-anhydro-~-glucitols, and the fact that no diglycolic acid was obtained after periodate oxidation followed by oxidation with hypobromite. b. Dianhydrohexitols. -A dianhydro-D-glucitol having 1,5 and 3,6 rings was synthesized by S. B. Baker;54he saponified 2,4-0-methylene-1-0-p-tolylsulfonyl-D-glucitol with dilute aqueous alkali, to form a monoanhydride that did not consume lead tetraacetate. This result eliminated the possibility of a 1,3-anhydro ring, but left open the question whether the compound was a 1,5-or 176-anhydride. Presumptive evidence that the compound contained a primary hydroxyl group was presented by the formation of a trityl ether. The compound was then p-toluenesulfonylated to a mono-p-toluenesulfonate, and this was saponified to the 2,4-O-methylene derivative of the postulated dianhydro-D-glucitol. The structure was conclusively established by starting with 1,5-anhydro-~-glucitol (polygalitol), mono-p-toluenesulfonylating, acetylating, and then saponifying the diester. A dianhydride was obtained that afforded a methylene acetal identical with that obtained on starting with 2,4-0-methylene-1-0-ptolylsulfonyl-D-glucitol. It is noteworthy that saponification of Baker's first compound afforded a 1,5-anhydride, whereas saponification of the corresponding 5,6-anhydro-2,4-0-benzylidene derivative furnished 1,3-anhydro-2,4O-benzylidene-~-glucitol.~' (52) L. Vargha and E. Kasztreiner, Chem. Ber., 93, 1608 (1960). (53) P. Sohar, L. Vargha, and E. Kasztreiner, Tetmhedron, 20,647 (1964). (54)S . B. Baker, Can.J.Chem., 32, 628 (1954).

246

S. SOLTZBERG

Vargha and KuszmannJ5 obtained the compound long known as “p-mannide”5g together with a small proportion of material believed (from its chromatographic behavior) to be 1,4:3,6-dianhydro-~-mannito1 (isomannide) on treating either 1,6-dichloro-l,6-dideoxy-~-mannito1 or 1,6-di-O-(methy~sulfony~)-~-mannito~ with methanolic sodium methoxide. A 2,5:3,6-dianhydro structure was suggested for “p-mannide” because, on tritylation, it gave a monotrityl derivative (indicative of the presence of one primary hydroxyI group); also, its bis(ptoluenesulfonate) gave only a monoiodo product on treatment with sodium iodide in acetone, indicative of the presence of one primary p-tolylsulfonyloxy group. No periodate was consumed by the “pmannide,” suggesting that contiguous hydroxyl groups were absent. The absence of oxirane and oxetane rings was indicated by the resistance of the compound to hot dilute acid or to alkali, There remained, therefore, only the 2,5:3,6-dianhydro structure that satisfied the evidence. The compound is the same as that obtained if ring closure is 1,4:2,5. It was shown to possess the D-gluco configuration as follows: treatment with fuming hydrochloric acid at 100” opened the 3,6 (1,4) ring, because displacement of the chlorine in the resulting monochloromonodeoxy anhydride by benzoyloxy, followed by deacylation and conversion into a dibenzoate, yielded a compound identical with 2,5-anhydro-l,6-di-O-benzoyl-~-glucitol. Of all the anhydrohexitols currently known, the structure of two dianhydrohexitols (reported by Wiggins5’) remain unresolved. The two dianhydrides were obtained on heating D-mannitol with hydrochloric acid; one had m.p. 118” and [a]D-34, and the second was isolated only as its dimethanesulfonate, m.p. 113-114”. The apparent enantiomorph of the first was formed on treating 1,6-dichloro-1,6dideoxy-D-mannitolwith sodium methoxide. Wiggins5Tureported that the physical constants of a product that he had obtained from 1,2:5,6-di-0-isopropylidene-3,4-di-O-p-tolylsulfonyl-D-mannitol by deacetalation with 70% acetic acid followed by acetylation and saponification with sodium methoxide were those of 1,4:3,6-dianhydro-~-iditol. Tipson and CohenJ7*have presented evi(55) L. Vargha and J. Kuszmann, Corbohvd. Res., 8, 157 (1968). (56) A. Siwoloboff, Ann., 233, 368 (1886). (57) L. F. Wiggins, Nature, 164,672 (1949). (57a) L. F. WigginsJ. Chem. SOC., 1403 (1947). (5%) R. S. Tipson and A. Cohen, Corbohyd. Res., 7,232 (1968).

ALDITOL ANHYDRIDES

247

dence, based on a study of the saponification of the 3,4-dimethanesulfonate and 3,4-bis(p-toluenesulfonate)of D-mannitol, that Wiggins’ product must have been grossly impure. They likewise pointed out that the physical properties reported by Wiggins, as compared to those of the authentic enantiomorph, and the yields of ester derivatives, also showed that the product had been a mixture.

c. Trianhydrohexitols. -Only one example of a trianhydroalditol is thus far known. Cope and Shen5*obtained 1,4:2,5:3,6-trianhydroD-mannitOl on boiling a solution of 1,4:3,6-dianhydro-2,5-di-O-ptolylsulfonyl-D-glucitol in ethanol containing sodium ethoxide under reflux. As the endo-p-tolylsulfonyl group (on 0-5) is protected by the neighboring ring, they postulated a trans-p-toluenesulfonylation resulting from an sN2s displacement of the exo-p-tolylsulfonyl group, with transfer of the endo-p-tolylsulfonyl group to the e m position. The resulting endo-alkoxide group could then bring about a normal sN2 displacement of p-toluenesulfonate. The reasonableness of this postulated mechanism was demonstrated by the fact that 2-0-acetyl1,4:3,6-dianhydro-5-O-p-tolylsulfonyl-~-glucitol~~ (erroneously assumed by them to be the 5-0-acetyl-2-0-p-tolylsulfonyl isomer) gave the same product. Hence, trans-p-toluenesulfonylation and the same intermediate were involved. Lemieux and M c I n n e ~ pointed ~ ~ ~ ) out the error in the assumption, and, by means of i.r. spectroscopy, unequivocally established the structure of the mono-p-toluenesulfonic and mixed esters, and verified conversion of the 2-O-acetyl-5-O-ptolylsulfonyl compound into 1,4:2,5:3,6-trianhydro-~-mannitol. Jackson and H a y ~ a r dhad ~ ~previously ~) pointed out that Cope and Shen’s assumption that the exo-hydroxyl group in 174:3,6-dianhydroD-glucitol is preferentially sulfonylated prior to conversion into the mixed ester is incorrect. They based their conclusions on studies of the rate of replacement of p-tolylsulfonyloxy group by iodide in the three dianhydro stereoisomers. However, their proof of structure was questioned by Lemieux and M c I n n e ~as~ being ~ ~ ) equivocal, because of possible formation of an acetoxonium intermediate prior to attack by the iodide ion.

(58) A. C. Cope and T. Y. Shen,]. Amer. Chem. SOC., 78,6912 (1956). (59) (a) R. U. Lemieux and A. G . McInnes, C a n . ] .Chem., 38,136 (1960).(b) M. Jackson and L. D. Hayward, ibid., 37, 1048 (1959).

248

S. SOLTZBERG

4. Anhydrides of Higher Alditols

a. Anhydroheptitols. - Sowden and FischeFI obtained a 2,6-anhydro-5,7-O-benzylidene-l-deoxy-l-nitroheptitol as a byproduct in the condensation of 4,6-O-benzylidene-~-glucopyranosewith nitromethane; this was hydroIyzed to the free anhydroheptitol. The configuration at C-2 (C-1of the original D-glucopyranose) has not yet been established. By analogy to the cyclization of 1-deoxy-1-nitro-D-mannit01,'~~ it would be expected that the compound should possess the D-glycero-D-ido configuration. Rosenthal and coworkers have applied the hydroformylation technique to D-glucal t r i a ~ e t a t e D-galactal ,~~~ t r i a ~ e t a t e 2-hydroxy,~~ D-glucal tetracetate, and 2-hydroxy-~-galactalt e t r a ~ e t a t e(Hydroxy.~~ methy1)ation occurred at C-1 in all cases. [After deacetylation, there were obtained 2,6-anhydro-3-deoxy-~-gZuco-heptitol and 2,6-anhydro3-deoxy-~-manno-heptitol,~~ 2,6-anhydro-3-deoxy-~-gaZacto-heptitol and 2,6-anhydro-3-deoxy-~-tuZo-heptitol(2,6-anhydro-S-deoxy-~~Ztro-heptitol),~~ and 2,6-anhydro-~-gZycero-~-guZo-heptitol as the principal product from 2-hydroxy-~-glucal,and 2,6-anhydro-~-gZyceruL-manno-heptitol as the principal product from 2-hydroxy-~-galactal.5w The latter compounds were shown to be identical with those described by Coxon and F l e t ~ h e r , ~which ~ ~ * ~ were ~' prepared by h y d r ~ g e n a t i o nof ~ ~2,3,4,6-tetra-O-acetyl-/3-~-glucopyranosy1~~ and -P-D-galactopyranosyl cyanides, respectively, and subsequent deamination by nitrous acid. Hough and Shute5* heated 1-deoxy-1-nitro-D-glycero-L-mannoheptitol in water, and obtained the corresponding 2,6-anhydride (28a) as the chief product; this behavior parallels that of l-deoxy-lnitro-D-mannit01.'~~Hough and Shute5* also obtained evidence for the probable presence of 2,6-anhydro-l-deoxy-l-nitro-~-gZyceru-~gZuco-heptitol and 2,5-anhydro-l-deoxy-l-nitro-~-gZycero-~-munnoand -L-gluco-heptitol. Sowden and Oftedah1132have suggested that such anhydridation occurs by way of an intermediate l-nitroald-l(59a)J. C. Sowden and H. 0. L. Fischer, U.S. Pat. 2,480,785(1949);Chem. Abstracts, 44,656 (1950). (59b)A. Rosenthal and H. J. Koch, Can.J.Chem., 43,1375(1965). (5%) A. Rosenthal and D. Abson, Can.]. Chem.,43,1985(1965). (59d) A. Rosenthal, Carbohyd. Res., 3, 112 (1966). (59e)B.Coxon and H. G. Fletcher, JrJ. Amer. Chem.Soc., 85,2637(1963). (590 B.Coxon and H. G . Fletcher, Jr.,J. Amer. Chem. Soc., 86,922(1964). (59g)L. Hough and S. H. Shute, 1. Chem. Soc., 4633 (1962). (59h)A. Camerman, H.J. Koch, A. Rosenthal, and J. Trotter, Can. J . Chem., 42,2630 (1964).

ALDITOL ANHYDRIDES

249

enitol, as shown in the following reaction sequence (according to Hough and Shute). HCNO,

%?NO, HCOH I HCOH I HOCH

HOCH

HOCH

HOCH

II

HO

HC I

HCOH I

I

I

C%NO2

I

I

HCOH

HCOH I H,COH

I

H,COH

2,B-Anhydro-1-deoxy- l-nitroD-,qlycero -L- manno-heptitol

b. Anhydrooctitols. - 1,5-Anhydro-~-erythro-~-ga~acto-octitols appear to be the only members of this group thus far reported. These anhydro-octitols have been prepared by the Raney nickel desulfurization of derivatives of the lincomycin group of antibiotics. Thus, 6acetamido- 1,5-anhydro-6,8-dideoxy-3,4-O-isopropylidene-~-eryt~roD-gazacto-octitol (28b)59iJ(“N-acetyl-3,4-O-isopropylidenelincosaminol”) was obtained from the corresponding 1-thiol, 1,5-anhydro-6,8dideoxy-l\r-(4-propyl-~-hygroyl)-~-eryth~o-~-gaZacto-octitol~~~ (anhydrolincomycitol) (2&) from the corresponding 1-S-methyl-1-thio Me I

pHrQ Me

HOCH I AcNHCH

Me

I

Q

‘CNHCH

Pr

OH

(28b)

OH (28~)

derivative, and the 7-0-methyl derivatives9’of 28b (“N-acetyl-3,4-0isopropylidenecelestoraminol”) from a celesticetin degradation product. (59i) H. Hoeksema, US. Pat. 3,255,175(1966);Chem. Abstracts, 65, 17039 (1966). (59j) W. Schroeder, B. Bannister, and H. Hoeksema, J . Arner. Chem. SOC., 89, 2448 (1967).

(59k)R. R. Herr and G. Slomp,J. Amer. Chem. SOC., 89,2444 (1967). (591) H. Hoeksema, J. Arner. Chern. SOC., 90, 755 (1968).

250

S. SOLTZBERG

111. PHYSICAL PROPERTIES 1. Infrared Spectra

ani -2-OThe infrared spectra of 1,4-anhydro-3,5-0-methy~nemethyl-DL-xylitol have been studied.s0 The 2-methyl ether was obtained by converting 1,4-anhydro-3,5-0-methylene-~~-xylitol into its monomethyl ether, and then hydrolyzing off the methylene group. A methyl ether prepared from the known 1,4-anhydro-3,5-0-isopropylidene-2-O-methyl-~~-xylitol proved to be identical with this compound, thus establishing at the same time that the methylene group in the known acetal is attached to 0-3 and 0-5 of 1,4-anhydro-~~xylitol. The methylene group, having a 1,3-dioxolane structure, was characterized by an absorption band at about 2800 cm-’. It was e ~ t a b l i s h e dthat ~ ~ the ring structures of anhydroalditols of unknown structure could be inferred solely on the basis of the multiplicity of the C-H stretching-vibration frequencies, without interpreting the bands below 1600 cm-’. The anhydroalditols examined were 1,5-, 1,6-, and 2,6-anhydro-~-glucitoland 1,6-anhydro-~iditol, and partially deuterated forms of the first two and the last. The following wavenumbers between 3000 and 2960 cm-’ (LiF) were assigned to the C-OH bonds on the ring: for 1,5-anhydro-~-g~ucitol, a strong band at 2985 cm-’ was attributed to the two axial C-OH bonds “above” the plane of the ring, and a weaker band at 2998 cm-’, to the single axial C-OH “below” the plane. It was claimed that the hydroxymethyl group in this compound was responsible for the bands at 2940 and 2885 cm-’. In 1,6-anhydro-~-glucitol,three of the four axial C-OH groups are spatially similar, and a strong absorption at 2975 cm-’ was assigned to them, the fourth (in the opposite direction) was assumed to cause the band at 2990 cm-’. 1,6-Anhydro-~-iditolhas four axial groups in 2 pairs, with absorptions of equal intensities; the bands at 2998 cm-I and 2960 cm-I were attributed to them, the former being assigned to the pair “above” the plane. The band due to the C-H part of the CH(0H) grouping is very weak in all of the compounds. For 2,6anhydro-D-glucitol, the assignments were 2992 cm-’ for the axial hydroxyl group “above” the plane, 2980 cm-’ for the equatorial hydroxyl group “above” the plane, and 2970 cm-’ for the equatorial hydroxyl group “below” the plane. All of the bands were of medium intensity. The asymmetrical and symmetrical stretching-frequencies of the (60) A. N. Anikeeva, E. I. Pokrovskii, L. G . Revel’skaya, E. F. Fedorova, and S. N. Danilov,J. Gen. Chem. USSR (Engl. Transl.), 36, 203 (1966).

ALDITOL ANHYDRIDES

251

-CH-(0-) grouping were weak, and ranged from 2920 to 2928 cm-I for the former, and 2860 to 2862 cm-' for the latter. The assignment of structure to 2,6-anhydro-D-glucitol (previously unknown) through its i.r. spectrum was based on the following reasoning. ( a ) There are four strong bands for CH2groups: -CH,-(OH), at 2938 and 2880 cm-'; -CH2-(O-),at 2920 and 2860 cm-'. As may be seen, the compound has a primary hydroxyl group as well as a primary \ ether group. (b) There are three ,CH-(OH) bands, thus excluding the presence of a ring of less than six members. ( c ) There is a weak shoulder on the band at 2938 cm-', presumably due to coalescence of the weak secondary ether grouping with the strong absorption band of the hydroxymethyl group at the same wavenumber. ( d ) The spectrum is not identical with that of 1,5-anhydro-~-glucitol.(e) A molecular model of the compound shows that it has one axial and two quasi\ equatorial ,CH-(OH) groups. (Stretching and out-of-plane deformation vibrations for OH and OD were also given.) Barker and Stephens6' studied the i.r. spectra of 3,6-anhydro-~and glucitol, 1,4-anhydro-~-mannitol,1,4:3,6-dianhydro-~-glucitol, 1,4:3,6-dianhydro-2,5-di-O-methyl-~-mannitol, and grouped the absorption bands into four characteristic types: Type A, a symmetrical, ring-breathing frequency of the single hydrofuranol ring (924 -1-13cm-', strong); Type B, a rocking vibration of the methylene group in the ring (879 &7 cm-'); Type C, at 858 +7 cm-I; and Type D, at 799 217 cm-' (tentatively assigned to the C-H deformation mode of a hydrogen atom attached to the ether carbon-atom of the hydrofuranol ring). The frequencies observed (cm-I) are summarized in Table I. TABLEI Frequencies (cm-') of Absorption Bands of 3,6-Anhydro-D-glucitol (l), 1,4-Anhydro-~-mannitol(2),1,4;3,6-Dianhydro-2,5-di-O-methyl-~-mannitol (3), and 1,4:3,6-Dianhydro-~-glucitol (4) Compound

A

B

C

D

1 2

933s 929s

886s 880s

781m 782s

3

923s 915s

880vs

867s 868s 860s 842m 846s

922m

884m

4

865111

816vs 814m 802m 830s 733s

"C-0stretching mode of methyl ethers. (61) S. A. Barker and R. Stephens, J . Chem. SOC., 4550 (1954).

Other bands

789m 753m

981s 995vs 949vs 966s 981vs 9440 976s 951m

252

S. SOLTZBERG

For dianhydroalditols having two hydrofuranol rings fused together, Type A and D absorptions appear as doublets, in conformity with the spectra of other sugar derivatives having fused hydrofuranol rings. However, for the dianhydro-D-mannitol, the bands at 923 and 816 cm-’ are single peaks, because of the identical nature of the anhydro rings. The i.r. spectrum of 1,4:3,6-dianhydro-~-glucitol dinitrate (isosorbide dinitrate) was recorded by Sammul and coworkers.62Hayward and coworkers6Yfound that the symmetrical stretching frequencies (us) for the NO, bands were characteristic for the endo (1282 *1 cm-’) and exo (1274 -+Icm-*) configurations in the mono- and di-nitric esters of the 1,4:3,6-dianhydrides (isohexides) of D-glucitol, D-mannitol, and L-iditol. The us (NO,) (1642-1647 cm-’) was not significantly affected by the configuration at the carbon atom to which the nitrate group was attached. Likewise, v(0N) (841-846 cm-I) was relatively constant. On the other hand, u(0H) showed configurational influence, the band for the exo-hydroxyl group appearing at 3640 cm-I, whereas the band for endo attachment was located at 3595 cm-’ for the dianhydro-Dmannitol and at 3565 cm-’ for the dianhydro-D-glucitol (OH-5). The assignments for the endo (hydrogen-bonded to the ring-oxygen atom of the second ring) and exo (free OH) hydroxyl groups are in agreement with the findings of Brimacombe and coworker^,^^ who found the OH stretching frequency to be at 3540 cm-’ for 1,4:3,6dianhydro-D-mannitol (endo OH) and at 3624 cm-’ (exo OH) and 3540 cm-’ (endo OH) for 1,4:3,6-dianhydro-~-glucitol. Included in the study were the i.r. spectra of 1,4-anhydro-2-deoxy-~-threoand -L-erythro-pentitols and 1,4-anhydro-~-threitoland -erythritol. The i.r. bands observed were consistently lower for the hydrogen-bonded OH as compared to the non-bonded or “free” OH. 2. Nuclear Magnetic Resonance Spectra

Proton coupling-constants were determined65 for the tri-O-acetyl derivatives of bis(ethylsulfony1)-a-D-lyxopyranosylmethane [2,6anhydro-l,l-bis(ethylsulfonyl)-ldeoxy-~-galactitol] (29) and bis(eth(62) 0. R. Sammul, W. L. Brannon, and A. L. Hayden, J . Ass. Ofic.Agr. Chem., 47, 918 (1964). (63) L. D. Hayward, D. J. Livingstone, M. Jackson, and V. M. Csizmadia, Can. J . Chem., 45, 2191 (1967). (64) J. S. Brimacombe, A. B. Foster, M. Stacey, and D. H. Whiffen, Tetrahedron, 4, 351 (1958). (65) L. D. Hall, L. Hough, K. A. McLauchlan, and K. Pachler, Chern. Znd. (London), 1465 (1962).

ALDITOL ANHYDRIDES

253

yIsu1fonyl)-a-D-arabinopyranosylmethane[2,6-anhydro-I,l-bis(ethylsu~fony~)-l-deoxy-D-glucitol] (30). Based on the coupling constants found (see Table 11), the angles calculated (by using the equation TAE~LE I1 Proton Coupling-constants (Hz) for Compound 29"

Jm.,

Anele (decrees)

11.2

1.7

~~

10.4 180 180

Estimated Ideal geometry

13.4

12.3

~~

3.0 54 60

3.3 52 60

"In chloroform, 3 mole percent.

] =J,cos28- 0.28 Hz and values of J,, = 9.26 Hz for angles between 0

and 90" andJ, = 10.35Hz for angles between 90 and 180")were almost in agreement with those required for the paired hydrogen atoms of a chair conformation, namely, 1C (D). 5

6

4

3

OAc

AcO

a

0

The acetoxyl-proton resonances were also in accordance with the (D) conformation, namely, one equatorial and two axial acetoxyl resonances for 29, and the converse for 30 (see Table 111), since methyl groups of axially oriented acetoxyl substituents of a pyranoid ring generally resonate at lower field than those of similar, equatorially attached groups.

1C

TABLEI11 Proton Chemical-shifts" (7-values) Compound

29 30

Acetoxyl-proton resonances

7.98 8.46

7.84 8.62

7.84 7.88

'In chloroform, 3 mole percent.

Proton magnetic resonance spectra of the three 1,4:3,6-dianhydrohexitols and their diacetates and dimethanesulfonates were used in

254

S. SOLTZBERG

1 , 4 :3,6-DianhydroL- iditol

1 , 4 :3,B-DianhydroD-glUCitOl

1 , 4 :3,B-DianhydroD-mannibl

determining the shapes of the oxolane ringss6 Proton couplings throughout the series were found to be constant, thus indicating that one conformation was common to all three compounds. The intramolecular hydrogen-bonding with the ether oxygen atom of the other and 174:3,6-dianring that occurs in 1,4:3,6-dianhydro-~-mannitol hydro-D-glucitol as a result of the presence of endo hydroxyl groups has little or no effect on the shape of the ring. Long-range couplings were identified, and it was noted that the coupling of the endohydroxyl proton to the proton on the same carbon atom was almost twice that for compounds having an exo-hydroxyl group. The n.m.r. spectrums7 of dianhydro-xylitolZ3and its methyl ether confirmed the chemical evidence that it has a 4-membered and a 5membered ring, and hence is 1,4:3,5-dianhydro-xylitol. This conclusion was further supported by the i.r. spectrum of the methyl ether, which has bands at 980 cm-' (oxetane ring) and at 1050 cm-I (oxolane ring).67

3. Circular Dichroism Spectra Circular dichroism of the nitrato (ONO) chromophore of the mono- and di-nitric esters of 174:3,6-dianhydrohexitolswas found to consist of two dichroic bands, one weak and positive, at about 265 nm, and a stronger band at -228 nm, which was positive for the endo-(R)-nitrato chromophore and negative for the em-(S)-nitrato dinitrate, which group.88 However, for 1,4:3,6-dianhydro-~-glucitol has both an endo and an exo nitrato group, both bands were positive, (+ 7,260) for the band at 225 nm was about half of the but [t$J,, algebraic sum of this band for the di-endo (+ 18,400) and di-exo (-4,460) compounds. (66)F. J. Hopton and G. H. S . Thomas, Can.J. Chem., 47, 2395 (1969). (67) G . E. Ustyuzhanin, A. I. Kol'tsov, N. S. Tikhomirova-Sidorova, and S. N. Danilov, Zh. Obschch. Khirn., 34, 3805 (1964); Chem. Abstracts, 62, 11892 (1965). (68) L. D. Hayward and S. Claesson, Chem. Commun., 302 (1967).

ALDITOL ANHYDRIDES

255

4. Crystal S h c t u r e The crystal and molecular structures of 1,4:3,6-dianhydro-2-0(p-bromophenylsulfony1)-(em)-D-glucitol 5-nitrate(endo) were reported by Camerman, Camerman, and Trotter.6gThe bond distances and valency angles were found to be normal, with an intramolecular apparent attraction between an oxygen atom of the NO, and the ring-oxygen atom of the second ring; the distance was found to be 0.29 nm.

5. Surface Activity By using light-scattering measurements, Becher'O observed a configurational effect influencing the micellar behavior of the four isohexide monostearates dissolved in benzene. He found that micelles are formed only when the molecule contains an endo stearic ester group (see Table IV). TABLEIV Micellar Behavior of Monostearates of Dianhydrohexitols in Benzene Solution Monostearate of

Critical micelle conc. (gldl)

Micellar molecular weight

Aggregation number

none

-

-

0.51

14,300

35

0.35

9,000

22

1,4:3,6-Dianhydro-~-glucitol (em) 1,4:3,6-Dianhydro-~-glucitol (endo) 1,4:3,6-Dianhydro-~-mannitol (endo) 1,4:3,6-Dianhydro-~-iditol (ex01

none

-

-

It is possible that a fused, oxygen-containing ring-system can behave as the center for micelle formation. For the dianhydro-D-mannito1 monostearate (endo), the stearate group is out of the way, and cannot interfere with agglomeration of the isohexide moieties. On the other hand, for the dianhydro-D-iditol monostearate ( e m ) , the stearate group can cover up the isohexide portion, and thus inhibit micelle formation. Hence, for dianhydro-D-glucitol monostearate, (69)A. Camerman, N. Camerman, and J. Trotter, Acta Cwstallogr., 19,449 (1965); Chem. Abstracts, 63, 10793 (1965). (70)P.Becher, 1.Phys. Chem., 64, 1221 (1960).

256

S. SOLTZBERG

the 2-ester ( e m ) does not form micelles, whereas the 5-ester (endo) does. Surface activity and micellar behavior in aqueous solutions of a number of D-glucopyranosyl alkyl-substituted benzenes and alkanes were s t ~ d i e d . 'It ~ was found that D-glucopyranosylto~uenedoes not form micelles, although it is surface-active. The higher alkyl derivatives (ethyl and butyl) show micelle formation. Moreover, D-glucopyranosylhexane does not form micelles, whereas the higher Dglucopyranosylalkanes (for example, the derivatives of octane and dodecane) form micelles at regularly decreasing concentration; these occur in the same order as with the D-glucopyranosyl derivatives of the alkylbenzenes. T o k i ~ a 'investigated ~ the surface tensions of a homologous series of D-glycopyranosyl derivatives of alkylbenzenes, and developed a relationship between the surface tension and the concentration. Likewise, an equation of state was evolved for the monolayers at relatively high pressures. It was observed that the free energy of absorption changes by about 650-700 cal/mole for each additional methylene group in the chain, except for methyl to ethyl, where the change was 380 cal. mole-'.

IV. REACTIONS 1. Ring Opening a. Anhydrohexitols.-Ring opening of an anhydroalditol was and was discussed first described for 1,4:3,6-dianhydro-~-mannitol by Wiggins.' Foster and coworker^'^ treated 1,4:3,6-dianhydro-~glucitol with an excess of boron trichloride in the presence of dichloromethane at room temperature, and, after removal of borate and unreacted boron trichloride, obtained a mixture of products that was treated with benzaldehyde. The major portion of the crude, solid benzylidene acetal produced proved to be 2,4-O-benzylidene-l,6dichloro-1,6-dideoxy-~-glucitol, and there was a small proportion of the corresponding 2,4:3,5-dibenzylidene acetal. Also, a small proportion of other (unidentified)benzylidene-containing substances was isolated. The yield of benzylidene derivatives of 1,6-dichloro-1,6(71) E. Hutchinson, V. E. SheaEer, and F. Tokiwa, J . Phys. Chem., 68, 2818 (1964). (72) F. Tokiwa, Bull. Chem. Soc.Jap., 38,751 (1965); Chem. Abstracts, 63,3646 (1965). (73) M. A. Bukhari, A. B. Foster, and J. M. Webber, Carbohyd. Res., 1, 474 (1966).

ALDITOL ANHYDRIDES

257

dideoxy-D-glucitol was high, compared to that obtained74by heating the starting dianhydride with fuming hydrochloric acid in a sealed tube at 110". Institoris and coworkers, in a patent issued to Chinoin from Gyogys~er,'~prepared 1,6-dibromo-l,6-dideoxy-~-mannitol 1,4:3,6-dianhydro-~-mannitol, reportedly in 91% yield, by using a methanolic solution of concentrated aqueous hydrobromic acid as the reagent. The ring opening of 1,4:3,6-dianhydro-~-iditol does not appear to have been attempted, although it should occur analogously to that of the other dianhydrides mentioned. Cope and Shen58 observed that, when a solution of 1,4:2,5:3,6trianhydro-D-mannitol in concentrated hydrochloric acid is heated on a steam-bath, the 1,4- and 3,6-rings are opened, affording 2,5-anhydrol,6-dichloro-l,6-dideoxy-~-mannito1 in good yield. It had earlier been mentioned55that treatment of 2,5:3,6-dianhydroD-glucitol with fuming hydrochloric acid at 100" causes opening of the 3,6-oxolane ring. The marked tendency for ring opening between the oxygen atom and the methylene (terminal) carbon atom is consistent with the known behavior of ethers and of oxolane (tetrahydrofuran) derivatives.

b. Anhydropentitols. - 1,4-Anhydro-5-deoxy-5-fluoro-~~-xylito~ was obtained in about 30 % yield by heating 1,4:3,5-dianhydro-~~-xylitol with potassium fluoride in aqueous diethylene glycol.7sThe product was identical with that obtained by heating 1,4-anhydro-5-chloro-5deoxy-DL-xylitol with powdered potassium fluoride in diethylene glycol. On heating 1,4-anhydro-xylitol with dry hydrogen chloride77 at 105", a crystalline "xylityl chloride," presumably l-chloro-l-deoxyxylitol, was formed, but, on the basis of results of derivatization, it probably retained the anhydride structure. Moreover, a subsequent report7* indicated that, under apparently the same conditions, 1,4anhydro-xylitol affords 1,4-anhydro-5-chloroS-deoxy-~~-xylitol. (74) R. Montgomery and L. F. Wiggins, J. Chem. Soc.. 237 (1948). (75) Chinoin Gyogyszer es Vegyeszeti Termeker Gyara Rt., Hung. Pat. 149,870 (1962); Chem. Abstracts, 60, 10776 (1963). (76) S. N. Danilov and E. Ya. Afanas' eva, Zh. Obshch. Khim., 36, 1406 (1%); Chem. Abstracts, 66, 18827 (1967). (77) S. N. Danilov, A. N. Anikeeva, N. S. Tikhomirova-Sidorova,and A. N. Shirshova, Zhur. Obshch. Khim., 27, 2434 (1957); Chem. Abstracts, 52, 7162 (1958). (78) S. N. Danilov, N. S. Sidorova, A. N. Anikeeva, Yu. A. Bol'shukhina, G. M. Zarubinskii, T. I. Orlova, and G . E. Ustyuzhanin, Sintez i Suoistoa Monomerou, Akad. Nauk S S S R , 1962,247 (1964); Chem. Abstracts, 62,9218 (1965).

258

S. SOLTZBERG

Danilov and coworker^'^ treated 1,4:3,5-dianhydro-~~-xylitol with a number of primary and secondary amines. The 3,s-anhydro ring was opened, to give the corresponding 5-amino-5-deoxy compounds. Hedgley and FletcheflO found that 1,4-anhydro-n-ribitol and 1,4anhydroerythitol are comparatively stable to liquid hydrogen fluoride at 18". 2. Isomerization

a. Acid-catalyzed Isomerization. - In an extensive investigation of the behavior of acylated inositols and 1,4- and 1,s-anhydroalditols in liquid hydrogen fluoride, Hedgley and Fletcher have showneothat a feature common to all instances in which inversions occur is the presence of a cis-trans sequence of three ester groups; inversion occurs at the middle carbon atom of the sequence. A mechanism involving intermediates of the charged, cyclic ortho-ester type was postulated. Hence, 1,4-anhydroerythritol diacetate does not isomerize to 1,4anhydrothreitol,80 although ring opening with formation of threitol and erythritol occurs. On the other hand, 1,4-anhydro-~-ribitol tribenzoate yields a complex mixture containing lY4-anhydro-ribitol, 1,4-anhydro-lyxitol (2,5-anhydro-arabinitol), and, possibly, 1,4-anhydro-arabinitol (as shown by the results of paper electrophoresis) and at least one pentitol, namely, DL-arabinitol. 1,4-Anhydro-~arabinitol triacetate afforded a mixture which, by paper electrophoresis, was found to contain 1,4-anhydro-lyxitol, 1,4-anhydro-ribitol, and 1,4-anhydro-arabinitol in one fraction, and arabinitol, ribitol, and xylitol in the second fraction. On the other hand, 1,4-anhydroD-XylitOl triacetate gave, solely, 1,4-anhydro-~-ribitol. Although these anhydropentitols lack the cis-trans triple-ester arrangement, the following respective intermediates from 1,Canhydrowere tri-0-benzoyl-D-ribitol and tri-O-acetyl-l,4-anhydro-~-xylitol proposed, in order to account for the epimerizations that occur.

(79)(a) S. N. Danilov, N. S. Tikhornirova-Sidorova, G. E. Ustyuzhanin, and G. A. Efirnova, Zh. Obshch. Khim., 32, 3614 (1962);Chem. Abstracts, 59,5246 (1963). (b) ibid., 32, 3617 (1962);Chem. Abstracts, 59, 5246 (1963).(c) G.A. Efimova, G . E. Ustyuzhanin, N. S. Tikhomirova-Sidorova, and S. N. Danilov, ibid., 33, 1429 (1963);Chem. Abstracts, 59, 12893 (1963).(d) S. N. Danilov, N. S. Tikhomirova-Sidorova. G. E. Ustyuzhanin, and C. A. Efirnova, USSR Pat. 162,522(1964); Chem. Abstracts, 61,9574(1964). (80) E. J. Hedgley and H. G. Fletcher, Jr., /. Amer. Chem. SOC.,86, 1576 (1964).

ALDITOL ANHYDRIDES

259

-0-C-Ph

HO,

,Me I

For 1,4-anhydro-xylitol, ring opening was postulated to occur as follows: H,COH I HCOAc

I

ACOCH,

I

OAc

HCO

I

H,COAc

(31)

Intermediate 31 can then rearrange to the isomeric pentitol carbonium intermediates, through attack by a neighboring acetyl group. These intermediates then cyclize to the epimeric 1,4anhydropentitols as a result of attack by the 1-hydroxyl group, or are hydrolyzed to the corresponding pentitols. An analogous ring-opening was proposed for tri-0-benzoyl-D-ribitol, in order to account for the 2,5-anhydroarabinitol and DL-arabinitol formed. 1,4-Anhydro-~-glucitoltetraacetate was envisaged as opening to form the intermediate 32 which, through a series of rearrangements and cyclizations, yields 1,4-anhydro-~-glucitol,1,4-anhydro-~-man-

260

S. SOLTZBERC H&OH 1

HCOAc

nitol, an unidentified product that is neither 1,4-anhydro-~-iditol nor 1,4-anhydrogalactitol, and galactitol, D-glucitol, L-iditol, and Dmannitol. In contrast, the derivatives of galactitol and D-mannitol that contain the requisite cis-trans arrangement within the ring gave only simple mixtures of the unchanged anhydride and one epimer, namely, 1,5-anhydro-~-gulitol (2,5-anhydro-~-glucitol) and 1,5-anhydro-~altritol, respectively, in accordance with predictions.81 On the other and tri-O-acetyl-1,5-anhyhand, tetra-O-acetyl-1,5-anhydro-~-glucitol dro-D-arabinitol, which lack the cis-trans arrangement are not epimerized.sl Although allitol, galactitol, and talitol, which lack a trans arrangement of the 3- and 4-hydroxyl groups, are difficult to anhydridize to the 1,4:3,6-dianhydrides, Hartmanns2 found that, on heating these hexitols or their 1,4-anhydrides at 110 to -185" with a catalytic amount of an acid (such as sulfuric acid or p-toluenesulfonic acid), 1,4:3,6dianhydrides were formed that were not dianhydrides of the starting hexitol. In this way, galactitol was transformed into 1,4:3,6-dianhydroDL-gluCitO1, inversion of configuration occurring at either C-3 or C-4.

b. Isomerization by Hydrogenation Catalysts.- The isomerization of a 1,4:3,6-dianhydrohexitolwas first reported by Fletcher and or -D-manG ~ e p pwho , ~ ~treated either 1,4:3,6-dianhydro-~-glucitol nitol with Raney nickel at 140" in the absence of added hydrogen, and then at 190-200" in the presence of hydrogen. From both compounds, the product was found to contain 1,4:3,6-dianhydro-~-iditol, isolated as the dibenzoate. Wright and Brandners4 have shown the reversible interconversion of aqueous solutions of these three dianhydrides in the presence of a nickel-on-kieselguhr catalyst under hydrogenating conditions. At the steady state, the product contains 57% of 1,4:3,6-dianhydro-~(81) E. J. Hedgley and H. G . Fletcher, Jr., J . Amer. Chem. Soc., 85, 1615 (1963). (82) L.A. Hartmann (to Atlas Chemical Industries, Inc.), US. Pat. 3,454,603 (1969). (83) H. C. Fletcher, Jr., and R. M. Coepp, Jr., J. Amer. Chem. Soc., 68, 939 (1946). (84) L. W. Wright and J. D. Brandner,]. Org. Chem., 29,2979 (1964).

ALDITOL ANHYDRIDES

26 1

iditol, 36% of the corresponding D-glUcitOl dianhydride, and 7% of the D-mannitol dianhydride. In the presence of 3% copper-promoted, nickel-on-kieselguhr catalyst, the respective yields weres5 58.9, 35.5, and 5.7%.

3. Selective Behavior a. Catalytic Oxidation. -When aqueous solutions of the three 1,4:3,6-dianhydrohexitolswere treated, at or above room temperature, with Adams’ catalyst and oxygen, only the endo-hydroxyl groups gave 174:3,6-dianhydrowere oxidized; 1,4:3,6-dianhydro-~-mannitol D-fructose and 1,4:3,6-dianhydro-~-threo-2,5-hexodiulose, 1,4:3,6dianhydro-D-glucitol gave 1,4:3,6-dianhydro-~-sorbose, and 1,4:3,6dianhydro-L-iditol was u n ~ h a n g e d . ~ ~ . ~ ~ For 174-anhydridesof the hexitols, oxidation of hydroxyl groups on C-2 or C-3 is determined by whether they are cis or trans, and whether the side chain is cis or trans to the 3-hydroxyl group on the oxolane ring and can form a hemiacetal ring on oxidation of the primary hydroxyl group.88 The final oxidation products from 1,4-anhydro-~glucitol (33) were 3,6-anhydro-~-gulono-1,4-lactone (34) and the hydrated form (35) of 3,6-anhydro-xyZo-~-hexulosono-1,4-lactone. Similarly, 1,4-anhydro-~-iditol (36) gave 3,6-anhydro-~-idono-1,4lactone (37) and 35. 1,4-Anhydro-~~-galactitol (38) gave 3,6-anhydro-

HoLQ H,COH

(33)

-4 (34)

OH

0

OH

H,COH I

DL-galactonic acid (39). 1,4-Anhydro-~~-talitol(3,6-anhydro-~~altritol) (40) gave 2-O-(carboxymethyl)-D~-threaric acid (41) and 1,4anhydro-D-mannitol(42)gave 20-(carboxymethy1)erythraricacid (43). (85) L. W. Wrightand J. D. Brandner, U.S. Pat. 3,023,233 (1962). (86) K. Heyns, W. P. Trautwein, and H. Paulsen, Chem. Ber., 96,3195 (1963). (87) Atlas Chemical Industries, Inc., French Pat. 1,425,204 (Jan. 28, 1966). (88) E. Alpers, Dissertation, University of Hamburg (1967).

262

S . SOLTZBERG H,COH

CO,H

HO

HO

I

(38)

H,COH I HCOH 0

(39)

CO,H 1

COzH

I

HCOCH,CO,H HO,C

H,COH

CO,H

C0,H

I

CO,H

CO,H I

HCOCH$O,H HO,C (42)

CO,H

I CO,H

(43)

The lY4-anhydrotetritolswere oxidized to diglycolic acid. However, 1,4-anhydrothreitol, in which the hydroxyl groups are trans, was much more resistant than 174-anhydroerythritolto attack. The following order of oxidizability or selectivity was formulated: primary OH = cis OH on the oxolane ring > trans OH. If the resulting anhydroaldonic acid can lactonize, the secondary hydroxyl group on the lactonic ring is oxidized.

b. Selective Acylation.-Of the anhydroalditols, lY4:3,6-dianhydroD-glUCitOl provides an interesting instance of selective acylation. (The two other dianhydrohexitol isomers either have both hydroxyl groups endo, or both exo, so that differentiation would not be expected.) In the D-glucitol dianhydride, the 2-hydroxyl group is exo, and the 5-hydroxyl group is endo and, apparently, blocked sterically by the second ring. Therefore, it had been assumed by Cope and Shen5*that the 2-hydroxyl group would be the more readily esterified. However, Jackson and Hayward5O'*)showed, by comparison of the rates of reaction of the mono-p-toluenesulfonates of the isomers with iodide, that the 5-hydroxyl group is probably more readily p-toluenesulfonylated. That this is correct was firmly established by Lemieux and M c I n n e ~ , ~who ~ ( ~proved, ) by i.r. spectroscopy, that the

ALDITOL ANHYDRIDES

263

mono-p-toluenesulfonate obtained in 11.7%yield had an intramolecular hydrogen bond, whereas, the mono-p-toluenesulfonate obtained in 45.4% yield showed no evidence of hydrogen bonding. Because only the endo hydroxyl group is favorably situated for hydrogen bonding, the first ester must be the 2-ester, and the second, the 5-ester. Selective acetylation of the 5-hydroxyl group was observed by Buck and who also showed that the presence of hydrogen chloride strongly influences the yields of isomers. At room temperature, on acetylation with acetic anhydride-pyridine, the 2-acetate was obtained in somewhat higher yield (27.6-28.6%) than the 5-acetate (16.9-17.1%). However, when acetylation was performed in the presence of pyridine hydrochloride, the yield of the 5-acetate was 42.9%, and that of the 2-acetate was 12%. These yields are in remarkable agreement with those of Lemieux and M c I n n e ~ ~for ~ ' ~the ' monosulfonylation reaction. As the latter reaction was conducted with ptoluenesulfonyl chloride, pyridine hydrochloride was formed, and it seems likely that, in some as-yet-unexplained manner, the presence of a strong acid augments the hydrogen-bonding. This effect is further illustrated in the partial esterificationgOof the dianhydro-Dglucitol with p-phenylazobenzoyl chloride in pyridine at 37". A 36% yield of the 5-ester and 12% of the 2-ester were obtained. A less obvious instance is the acylation of 1,5-anhydro-4,6-0benzylidene-D-glucitol. The 1,5-anhydride has a chair conformation, and the 2- and 3-hydroxyl groups are equatorially attached trans to each other, and are, apparently, essentially equivalentg1However, the 2-benzoate and 2-p-toluenesulfonate were obtained almost exclusively on using the acyl chlorides in pyridine, although in the former acylation, an appreciable proportion of the dibenzoate was formed. In order to explain his observations on the conversion into the corresponding 2,3-anhydride7 Newthsl suggested a conformational shift to a boat conformation. However, should a conformational shift occur early during the esterification (as a result of the formation of a small proportion of hydrogen chloride), the hydroxyl groups could become trans-axial, and the 2-hydroxyl group would be in a position to form a hydrogen bond with one of the acetal oxygen atoms, almost analogously to the situation with 1,4:3,6-dianhydro-~-glucitol. (89) K. W. Buck, J. M. Duxbury, A. B. Foster, A. R. Perry, and J. M. Webher, Carbohyd. Res., 2,122 (1966). (90)K. W. Buck, A. B. Foster, A. R. Perry, and J. M. Webber,]. Chern. Soc.,4171(1963). (91) F. H. Newth, ]. Chern. SOC.,2717 (1959).

264

S. SOLTZBERG

c. Reaction of Sulfonyloxy Groups with Nucleophiles. - It is well known that isolated, secondary sulfonyloxy groups on sugars react with difficulty with nucleophilic reagents.92Hence, it was incorrectly assumed that the same would apply to alditols, and that the sulfonic esters of the 1,4:3,6-dianhydrohexitolswould be resistant to displacement. However, it has been found that a considerable difference in the ease of reaction exists between the endo- and em-sulfonic esters of the 1,4:3,6-dianhydrohexitol isomers; this was discussed by Wiggins.' In the early work, no attention was paid to the fact that Walden inversion occurs, and the compounds formed were incorrectly named. Cope and Sheng3called attention to this serious error, and corrected the configurations that had been given for many of the compounds. Those corrected by Cope and Shen,g3 and some additional compounds, are listed in Table V. TABLEV Revised Configurations of Derivatives of Anhydroalditols

New structure

Old configuration

References

1,4:3,6-Dianhydro-5chlaro-5-deoxy-

D-glucitol

16,93

L-iditol 1,4:3,6-Dianhydro-5-deoxy-5-iodo-2-O-ptolylsulfonyl-L-iditol

D-glucitol

93

1,4:3,6-Dianhydro-2,5-dichloro-2,5dideoxy-L-iditol 1,4:3,6-Dianhydr0-2-chloro-2-deoxy-5-0-

93

(methylsulfonyl)-D-glucitol

93

1,4: 3,6-Dianhydro-2-chloro-2-deoxy-5-O(phenylcarbamoy1)-D-glucitol

D-mannitol

93

D-mannitol

93

D-mannitol D-mannitol

93 93

xylitol

79b

xylitol

79b

xylitol

79b

1,4:3,6-Dianhydro-2,5-dideoxy-2,5diiode L-iditol

2,5-Diamino-l,4:3,6-dianhydro-2,5dideoxy-L-iditol

1,4:3,6-Dianhydro-2,5-dithio-~-iditol 1,4:3,5-Dianhydro-2-anilino2-deoxy-~~-arabinitol?

1,4-Anhydro-2,5-dideoxy-2,5dipiperidino-DL-arabinitol?

1,4-Anhydro-2,5dideoxy-2,5-bis(diethylamino)-DL-arabinitol?

1,4-Anhydro-2-(butylamino)-2-deoxy3,5-O-methylene-~~-arabinitol? 1,4-Anhydro-2-anilino-2-deoxy3,5-O-methylene-DL-arabinitol?

xylitol

145

xylitol

145

(92) R. S. Tipson, Adoan. Carbohyd. Chen., 8, 107 (1953). (93) A. C. Cope and T. Y. Shen,]. Amer. Chem. SOC., 78,3177 (1956).

ALDITOL ANHYDRIDES

265

The three 1,4:3,6-dianhydrohexitolsshow the shielding effect of the cis-fused oxolane ring on their di-0-sulfonyl derivatives. As C-2 and C-5 of 1,4:3,6-dianhydro-~-mannitol 2,5-di-p-toluenesulfonate and 2,5-dimethanesulfonate are open to attack from the exo direction, they react readily with such nucleophilic reagents as benacetate in acetone,93t h i ~ a c e t a t e , ~ ~ zoate in N,N-dimethylf~rmamide,~~ and iodide in acetic anhyphthalimide in N,N-dimethyIf~rmamide,~~ dride.59(b1 Inversion occurs at these positions and the L-iditol configuration results, as pointed out in the several examples in Table V. However, hydroxide ion may merely saponify,92with retention of c~nfiguration.~~ 2,5di-p-toluenesulfonate, generFor 1,4:3,6-dianhydro-~-glucitol ally only the 5-p-tolylsulfonyloxy group (endo) is displaced, with formation of the L-ido onf figuration;^^.^ but it may lose both sulfonyloxy groups under more vigorous conditions, affording 1,4:2,5:3,6trianhydro-~-mannitol~* as already mentioned. as The 2,5-bis(p-toluenesulfonate) of 1,4:3,6-dianhydro-~-iditol, would be expected, is quite resistant to nucleophilic attack and can be recovered unchanged in good yield.93,96,97 However, reaction occurs to some e ~ t e n t , ~ ~but ( * ) the , ~ ~resulting product or products have not as yet been isolated. Under very vigorous conditions, the 2-p-tolylsulfonyloxy group reacts, and of 1,4:3,6-dianhydro-2,5-di-O-p-tolylsulfonyl-~-glucitol unsaturated products r e s ~ l t . ~ " ~ ~ d. Formation and Behavior of Acetals. -The benzylidenation of 1,4-anhydroerythritol has been monitored by n.m.r. s p e c t r o s c ~ p y . ~ ~ The initial product was postulated to have the phenyl group endo to the fused-ring system. It was found that, at equilibrium, the endo and exo forms are present in approximately equal amounts. For 1,4anhydro-3,5-O-benzylidene-~-mannitol in N,N-dimethylformamide containing p-toluenesulfonic acid, the low-field benzyl proton signal (T 4.35) changed to the high-field signal (T 4.54) of the known endo-2,3-O-benzylidene isomer.gRHowever, on prolonged treatment, a low-field signal again appeared, but, this time, it was that of the exo-2,3-O-benzylidene isomer. The latter was isolated by column chromatography on silica gel; it had m.p. 111-112", [aID-40" (in (94) D. H. Buss, L. D. H9I1, and L. Hough, I . Chem. Soc., 1616 (1965). (95) P. Bladon and L. N. Owen, J . Chem. SOC.,585 (1950). (96) N. K. Matheson and S. J. Angya1,J. Chern. Soc., 1133 (1952). (97) L. F. Wiggins and D. J. C. Wood, J . Chern. Soc., 180 (1951). (98) F. S. Al-Jeboury, N. Baggett, A. B. Foster, and J. M. Webber, Chern. Comrnun., 222 (1965).

266

S. SOLTZBERG

water), as compared with m.p. 94-96", [a],, -88" (in water) for the 2,30-benzylidene isomer previously known. As the zeta1 group in 2,5-O-methylene-~-mannitol is relatively resistant to hydrolysis, S. B. Bakerss studied the stabilities of 1,4:3,6dianhydro-2,5-O-methylene-~-mannitol and -D-iditol relative to that of the D-mannitol acetal. He synthesized 1,4:3,6-dianhydro-2,5-0methylene-D-mannitol by cyclizing 2,5-O-methylene-l,6-di-O-p-tolylsulfonyl-D-mannitol or by methylenation of 1,4:3,6dianhydro-~-mannitol. The D-iditol isomer was prepared by cyclizing 1,6-di-O-benzoyl2,5-0-methylene-3,4-di-O-p-tolylsulfonyl-~-mannito~. 2,5-O-Methylene-~-mannitolis quite stable to hydrolysis by dilute acids, or to acetolysis by acetic anhydride-sulfuric acid. It was found that its 1,4:3,6dianhydride could be hydrolyzed merely by treatment with hot water. The dianhydro-D-iditol derivative is stable to hot water and hot mM hydrochloric acid, but not to hot 10 mM acid. Inspection of molecular models fails to reveal any reason to expect the behavior observed for these two dianhydro-0-methylene isomers, as it would appear that a greater strain should be present in the bonds of the methyIene bridge in the exo (D-iditol) isomer than in the endo (D-mannitol)isomer. It is possible that the oxygen atoms of the oxolane rings strongly repel the acetal oxygen atoms of the dianhydro-0methylene-D-mannitol, favoring rupture of the methylene bonds; or, because of the proximity of the oxygen atoms to each other, a proton could have a longer residence-time in the vicinity, thus destabilizing the methylene group. Indeed, it is probably more noteworthy that 1,4:3,6-dianhydro-2,5-O-methylene-~-iditol is incapable of existence at all, unless a considerable adjustment of the fused oxolane rings should occur.

e. Miscellaneous.- On treatment of 1,4:3,6-dianhydro-2,5-dideoxy2,5-diiodo-~-iditolwith silver nitrate in dry acetonitrile, racemization occurs, with formation of the corresponding 2,5-dinitrates of 1,4:3,6dianhydro-D-glucitol, -D-mannitol, and -L-iditol in poor yield (although 70% of the theoretical yield of silver iodide was isolated). The low yield was not attributable to destruction of nitric ester, because 1,4:3,6-dianhydro-~-glucitol dinitrate was recovered in almost 90% yield when it was subjected to the same conditions.loOThis test also served to demonstrate that racemization had occurred during the formation of the nitric ester, and, hence, a carbonium ion intermediate was postulated. Different reactivities of the dinitrates of the three dianhydro(99)S. B. Baker, Can. /. Chem., 31, 821 (1953). (100)L. D.Hayward, M. Jackson, and I. G . Csizmadia, Can./. Chem. 43, 1656 (1965).

ALDITOL ANHYDRIDES

267

hexitols toward hot pyridine was observed.'"' Arrhenius activation energies of 21,24, and 41 kcal.mole-' were found for the dianhydro-Liditol, -D-ghcitol, and -D-mannitol dinitrates, respectively, indicating that the exo nitrato group is the more readily affected. 1,4:3,6-Dianhydro-~-mannitol was directly converted,lo2 in quantitative yield, into 1,4:3,6-dianhydro-2,5-dideoxy-2,5-diiodo-~-iditol on treatment with triphenyl phosphite methiodide in dry benzene at 20".

v. USES

1. Industrial As mentioned by Wiggins,' the principal industrial use of the hexitol anhydrides (prepared almost exclusively from D-glucitol and D-mannitol) is in the manufacture of nonionic surfactants. The anhydrides are generally formed in situ during esterification with a variety of saturated and unsaturated fatty acids. However, the anhydrides, frequently in the form of poly(oxythy1ene) adducts, are mentioned in a wide range of patents, of which the following selection is only a small sample. On monoacylation with lauric acid and treatment of the product with ethylene oxide, anhydro-D-glucitols afford an antistatic agent for photographic film.ln3 1,4:3,6-Dianhydro-~-glucitol was one of a number of diols used in preparing polymeric phosphites (esters) useful as flame-proofing agents.lo4 Water-soluble glycidyl with ethers were prepared by treating 1,4:3,6-dianhydro-~-glucitol epichlorohydrin and 50%, aqueous sodium hydroxide solution.1o5 Poly(oxypropy1ene) ethers of 1,4-anhydro-~-glucito1or -D-mannitol have been used in the preparation of nonfriable polyurethan foams.lM 1,4:3,6-Dianhydro-~-glucitol, 1,4-anhydro-D-glucito~,or 1,5-anhydroD-mannitol are used in modifying thermosetting melamine resins.lo7 (101) M . Jackson and L. D. Hayward, Can.]. Chem., 38,496 (1960). (102) (a) N. K. Kochetkov and A. I. Usov, Tetrahedron, 19,973 (1963).(b) N. K. Kochetkov and A. I . Usov, lzv. Akad. Nauk S S S R , Otd. Khim. Nauk, 1042 (1962);Chem. Abstracts, 57,15213 (1962). (103) Adox Fotowerke, Ger. Pat. 1,134,586 (Aug. 9, 1962); Chem. Abstracts, 57, 12006 (1962). (104) L. Friedman and H. Could, U.S. Pat. 3,053,878(Sept. 11,1962);Chem.Abstracts, 58, 3354 (1963). (105) J . G. Morrison, U.S. Pat. 3,041,300 (June 26, 1962); Chem. Abstructs, 57, 8739 (1962). (106) (a) Takeda Chemical Industries, Ltd., Brit. Pat. 989,144 (April 14, 1965); Chem. Abstracts, 63, 1963 (1965). (b) Takeda Chemical Industries, Japan Pat. 16,198 (Sept. 12, 1966); Chem. Abstracts, 66, 18605 (1967). (107) (a) J. D. Larkin and G. M. Grudus, U.S. Pat. 3,194,719 (July 13, 1965). (b) G. M. Grudus and J. D. Larkin, U.S. Pat. 3,194,720. (c) G. M. Grudus and J. D. Larkin, U.S. Pat. 3,194,723.

268

S. SOLTZBERG

Clear lacquers that dry to tack-free films in four hours were obtained by use of unsaturated polyesters of D-glucitol anhydrides having two or more free hydroxyl groups.l0*Linear polymers were obtained from the acetate and the acetals of l74-anhydro-xy1itoI monomethacryits diethers, and its diesters may late.loS1,4:3,6-Dianhydro-~-glucitol, be used for stabilizing polypropylene against oxidation and discoloraand -di-0-ethtion by sunlight.'1° 1,4:3,6-Dianhydro-2,5-di-O-methylyl-D-glucitol are useful stabilizers for solutions of tetracycline and related antibiotic substances."' A so-called "liquid, universal shortening composition" is prepared from a mixture of fatty acid esters of 1,4:3,6-dianhydro- and 1,4-anhydro-~-glucitoland their poly(oxythylene) derivatives with vegetable oils.112 2. Biological 1,4:3,6-Dianhydro-~-glucitol dinitrate has found important medicinal use as a coronary vasodilator, and a considerable volume of literature has developed from studies thereof; of this, the references given here constitute only a part.'13 In one investigation, the pharmacological behavior of the three isomeric dianhydrohexitol dinitrates was c ~ m p a r e d . " ~ 'It~ ' was shown that, although glycerol trinitrate dinihas a slightly more rapid action, 1,4:3,6-dianhydro-~-glucitol trate has a longer-lasting effe~t."~(~'-'~' (108) Howards of Ilford, Ltd., Brit. Pat. 927.786 (June 6, 1963); Chem. Abstracts, 59, 6534 (1963). (109) A. N. Anikeeva and S. N. Danilov, U.S.S.R. Pat. 175,660 (Oct. 9, 1965); Chem. Abstracts, 64,6786 (1966). (110) P. E. Oberdorfer, Jr., U.S. Pat. 2,967,169 (Jan. 3, 1961). (111) R. A. Nash and B. E. Haeger, U.S. Pat. 3,219,529 (Nov. 23, 1965). (112) A. S. Geisler, U.S. Pat. 3,184,575 (May 25, 1965). (113) (a) J. E. Halliday and S. C. Clark,]. Pharm. Pharmacol., 17,309 (1965).(b) W. H. Bunn, Jr., and A. N. Cremos, Angiology, 14, 48 (1963); Chem. Abstracts 58, 11876 (1963). (c) A. L. Smith, Angiology, 13, 425 (1962); Chem. Abstracts, 58, 2768 (1963). (d) D. A. Sherber and I. J. Gelb, Angiology, 12, 244 (1961); Chem. Abstracts, 55, 18991 (1961). (e) J. P. Buckley, M. D. G. Aceto, and W. J. Kinnard, Angiology, 12, 259 (1961); Chem. Abstracts, 55, 18991 (1961). (f) A. J . Dietz, Jr., Biochem. Pharmacol., 16, 2447 (1967); (g) V. N. Dzyak, L. T. Furs, and B. N. Bezborod'ko, Farmakol. i Toksikol., 26, 47 (1963); Chem. Abstracts, 59, 4439 (1963). (h) J. C. Krantz, Jr., G. G. Lu, F. K. Bell, and H. F. Cascorbi, Biochem. Pharmacol., 11, 1095 (1962).(i) E. Kimura, K. Ushiyama, T. Yamazaki, K. Yoshida, N. Kojima, and T. Kanie, Proc. Asian-Pacific Congr. Cardiol. 3rd Kyoto, 1, 745 (1964); Chem. Abstracts, 65, 1265 (1966).(j) V. V. Buyanov, Famakol. i Toksikol., 30, 30 (1967); Chem. Abstracts, 66, 84327 (1967). (k) H. P. Cjuchta and R. F. Gautieri, J . Pharm. Sci., 52, 974 (1963). (1) A. Skoda, G. G. Rowe, W. C. Lowe, and C. W. Crumpton, Amer. J. Med. Sci., 246, 584 (1963). (m) K. Ogawa and S. Gudbjarnason, Arch. Int. Pharrnacodyn. Ther., 172, 172 (1968); Chem. Abstracts, 68, 113158 (1968).

ALDITOL ANHYDRIDES

269

Interest has developed in the use of 1,4:3,6-dianhydro-~-glucitol as an orally administered, osmotic diuretic. It has been examined in the lab~ratory,"~ and ~linically"~ in patients having cirrhosis of the liver. It was found to be similar in effectiveness to intravenous Dmannit01."~Its use in diuretic compositions has been patented.ll6 The dianhydro-D-glucitol lessens cerebrospinal fluid pressure and brain mass when administered orally to dogs;"' this effect is accompanied by osmotic diuresis. Intravenous injection likewise produces a drop in the cerebrospinal fluid pressure, but this is followed b y a greater "rebound." have shown that orally administered 1,4: Becker and 3,6-dianhydro-~-glucitol, at doses of 1.5 to 2 g/kg, effectively lessens the intraocular pressure in patients having cataracts, and at doses of 0.5 to 2 g/kg for glaucoma. Only minimal side-effects were noted in two of nineteen patients. were 1,5-Anhydro-~-glucitol and 1,5-anhydro-6-deoxy-~-glucitol used in a study of the mechanism of the active transport of sugars.118 1,5-Anhydro-~-glucitol,D-glucose, D-galactose, and 6-deoxy-~-glucose were found mutually to inhibit the active transport of each of the other compounds, indicating that a single transport-mechanism is involved for all of In connection with the transport of sugars, the Michaelis constant, the relative maximum rate with respect to D-glucose, and the phosphorylation coefficient were found to respectively, for the hexokinase in rat-epidibe 50, 0.5, and 3 x dymal, adipose-tissue homogenate~."~ The competitive inhibition of aldolase by a number of structural analogs of D-frUCtOSe 176-diphosphate has been investigated. Among these analogs were 1,4-anhydro-~~-ribitol 5-phosphate7 1,4anhydro-DL-xylitol 5-phosphate, 1,4-anhydro-~-arabinitol5-phosphate, 2,5-anhydro-~-mannitol 176-diphosphate, and 2,5-anhydroD-glUCitOl l,6-diphosphate. Their respective, enzyme-inhibitor, (114)(a) J. F.Treon, L. E. Gongwer, and W. H. C. Rueggeberg, Proc. Soc. Erp. Biol. Med., 119, 39 (1965).(b) J. H.Shinaberger, J. W. Coburn, R. C. Reba, K. G. Barry, and L. C. Clayton,]. Phormacol. Erp. Ther., 158,460(1967). (115)0.Gagnon, P. M. Gertman, and F. L. Iber, Amer. ]. Med. Sci., 254, 284 (1967). (116)Atlas Chemical Industries, Inc., Brit. Pat. 1,067,298(May 3,1967). (117)B. L. Wise, J . L. Mathis, and J. H. Wright,j. Neumsurg., 25,183 (1966). (117a)B.Becker, A. E. Kolker, and T. Krupin, Arch. Ophthalmol., 78, 147 (1967). (118)(a)R. K. Crane and P. Mandelstam, Biochim. Biophys. Acta, 45, 460 (1960). (b)R. K. Crane, ibid., 45,477(1960).(c)I. Bihler, K. A. Hawkins, and R. K. Crane, ibid., 59,94 (1963).(d) R. A. Ferrari, P. Mandelstam, and R. K. Crane, Arch. Biochem. Biophys., 80,372(1959). 86,166 , (1963). (119)A. Hernandez and A. Sois, Biochem.I. (120)F.C. Hartman and R. Barker, Biochemistry, 4, 1068 (1965).

S. SOLTZBERG

270

(K,)were found to be 4.6 X 3.8 X 1.3 x 3.0 x and 1.3 x respectively. It was concluded that the binding to the enzyme is due principally to the phosphate dissociation constants

groups. The contribution of the hydroxyl groups is not significant. Constituents of certain nonionic surfactants (SpanlZoQ 40, Span 60, Span 65, and TweenlZ0“65) were found to cause a flocculation reaction with serums containing C-reactive protein.121‘Q’ These surfactants are partial esters of fatty acids with a mixture of anhydro-D-glucitols. 1,4:3,6-Dianhydro-~-glucitol and 1,4-anhydro-~-glucitolmonostearates were found to be highly The presence of a free hydroxyl group was found to be a requisite for binding. The relatively low toxicity of the partial esters of the hexitol anhydrides makes them desirable emulsifiers for injectable medications. “D-Mannide” mono-oleate was used in a patented, single-injection, vaccine compositionlzZand in an adjuvant vaccine.Iz3 1,4:3,6-Dianhydro-~-mannitol, its 2,5-dimethyl ether, and 1,4:3,6dianhydro-2,5-dichloro-2,5-dideoxy-~-iditol were found to be inactive as adjuvants for pyrethrins in fly sprays.12*

VI. TABLES OF PROPERTIES OF THE ANHYDRIDES AND THEIRDERIVATIVES Tables VI-XI show the physical properties of the alditol anhydrides and certain of their derivatives.

(12Oa)Registeredby Atlas Chemical Industries, Inc. (121) (a) I. M . Tuomioia, P. Kajanne, and R. Junnila, Ann. Med. E x p . Biol. Fenniae (Helsinki), 39,29 (1961). (b) ibid., 39,35 (1961). (122) Wright-Fleming Institute of Microbiology, St. Mary’s, Brit. Pat. 1,081,796 (Aug. 31,1967);Chem. Abstracts, 68,35362 (1968). (123) A. F. Woodhour and T. B. Stim, U.S. Pat. 3,149,036 (Sept. 15, 1964); Chem. Abstracts, 61,13136 (1964). (124) R. W. Ken, Aust. Commonwealth Sci. Znd. Res. Organization, Bull. No. 261, 32 (1951); Chem. Abstracts, 46,2227 (1952).

TABLEVI Anhydrotehitols and Their Derivatives

M.P., Compound 1,4Anhydro-D-threitol 2,3-di-O~p-nitrobenzoyl)1,4Anhydro-~-threitol

2,3-di-O-(p-nitrobenzoyl)-

degrees 191-2 60-1 63-4 191-2

160-5"/0.17tom 144/2-3 torr

1,4AnhydIO-DL~rythntOl 2,3-sulfite 2,3-O-benzylidene2,3-di-O-(p-nitrobenzoyl)-

B.P., degrees

106-8

160-5a/0.05tom

173-4

"Bath temperature. Temperature, 24".Temperature, 20".

[alo, degrees

Rotation solvent

-115.0 -5 -4

CHCl, HzO HzO

nD

References 5 6 64 6

1.437V 1.4767'

6 64 133 134 6

$

z

$ > z

2U

EFl

10

TABLE VII

-4 10

Anhydropentitols and Their Derivatives" Compound

M.P., degrees

B.P., degrees 14516mtorr

2,3,5-hi-O-(p-nitrobenzoyl)2,3,5-tri-O-benzyl1,4Anhydro-~-arabinitol 2,3,5-tri-O-p-nitrobenzoyl-

1,4-Anhydro-~~-arabinitol* tri-0-acetyltri-0-benzoyl1,4-Anhydro-~~-ribitol 2,3,5-tri-O-benzoyl-

1,4-Anhydro-~-ribitoI 1,4-Anhydro-D~-xylitol 3,5-0-methylene-2-0methacryloyl3,5-0-benzylidene-2-0methacryloyl3,5-O-isopropylidene-2-0methacryloyl

80-1 S Y W

80-2 115-16 122-3 144-5 74-5 76.5-7 113-14 116-17 100-1 99 98-9

la],, degrees

Rotation solvent

25.3

MeOH

-85 +0.6

CHCI, CHCl,

85.1

CHCl,

nD

References

1.4917 1.4868

19 19 19 142 14a 146 146 146 10 11 10 11 14a 14b 19

77

126

138

126

57

126

v,

E 4

N W

M

T1

2,3-di-O-benzoyl-5-chloro114-15 5-deoxy3,5-0ethylidene-2-0-p86 tolylsulfonyl5-chIoro-S-deox y3,5-Oethylidene96-7 3,s-0-methylene82-3 2-0-methyl-3,5-0-methylene- 48 1,4-Anhydro-~-xylito1 1,4-Anhydro-~-xylitol (2,5-anhydro-~-xylitol) 2,3-di-O-acetyl-5-0-trityL 153 104-5 2,4-O-methylene2,4-0-methylene-5-082-3 p-tolylsulfonyl5-deox) 5-iodo-2,4-097-8 methylene5-deoxy-2,4-0-methylene64-5 3,5-Anhydro-~-xylitoI l-deoxy-2,4-O-methylene84-5 2,5-Anhydro-~-arabinitol ( 1,4-anhydro-~-lyxitol) 75-6 3,4-O-isopropylidene1,5-Anhydro-~-arabinitol 50-2 2,3,4-tri-O-acetyll,S-Anhydro-%deoxv-D-threo-pentitol 68 3,4-di-O-acetyl1,5-Anhydro-2-deoxy-~erythro-pentitol 3,4-di-O-acetyl-

16 125 16 125 23 23 80

160-1/3 tom

160-70/0.02 tom 150-5/0.2 ton

-11.2k1 10

HzO

1.4957

19 19 22 22 22 22

H20

127

-40.5

H20

17

73.6 -29.6 -38

CHCI, HzO CHCI,

30 64 64

H D Hi0

64

28.6

102/0.5 torr

l20/0.2-0.3 tom 86-9010.2 tom

64 75

64 ~~

(continued)

to

TMLE VII (continued)

M.P., Compound

degrees

1,5-Anhydro-2-deoxy-~erythro-pentitol 42-4 3,4-di-O-benzoyl89-91 3,4di-O-acetyl3,4-di-O-(p-nitrobenzoyl)117-19 1,s-Anhydro-DL-ribitol tri-0-acetyl133 1,4:3,5-Dianhydro-D~-xylitol 2-0-methyl2-0-methyl-3-0-p-tolylsulfonyl- 77

B.P., degrees

4

A

degrees

Rotation solvent

-51.4 -65.1 -47.3 -97.1

H# CHCl, CHC13 CHCl,

[ab

nD

References 128 128 128 128 136

5112 tom

1.4542

23 23

“For additional compounds, see Ref. 1 and J. Defaye, This Volume, p. 181. q h i s appears to be a case of isomerization of xylitol. The original report was unavailable, and so was not discussed under “Isomerization.”

vl v)

n

F 4 N

TABLEVIII Monoanhydrohexitols and Their Derivatives" Compound 1,5-Anhydro-D-allitol(2,6-anhydro-~-allitol) 1,5-Anhydro-~-allitol 6-amino-6-deoxy-, hydrochloride

M.P., degrees

B.P., degrees

+ 34.0

151-2 151-2 67-9

- 33.4 - 23 - 18.4

6-acetamido-tri-0-acety l-6-deoxy-

3,6-Anhydro-~-altritol( 1,4-anhydro-~-talitol) 1,2:4,5-di-O-isopropylidene4,5-O-isopropylidene1,5-Anhydro-~altritoI(2,6-anhydro-~talitol) 4,6-O-benzylidene2,3,4,6-tetra-O-benzoyl1,4-Anhydro-~-galactitol

tetra-0-benzoyl3,6Anhydro-~-galactitol I-deoxy1-deoxy-tri-0-p-tolyIsulfonyl1,5-Anhydro-~galactitol 4,6-benzeneboronate 1,5-Anhydro-2-deoxy-~-lyro-hexitol 4,6-benzeneboronate 2,6Anhydro-~-glucitol( 1,5-anhydro-~-gulitol) tetra-0-acetyl4,5-di-O-benzoyl-1,3-0-benzylidene1,3-O-ben~ylidene-~

72-610.1 tom 77-9 127-9 125-6 176-7 95-6

-37 -43

28.4 - 22.7 - 6.8

- 35.2 - 18.0

99-1Olb 87-9Zb

41.7

92.5-3 73

Rotation solvent

69

114-15 amorph. 115-8 156-SlO.03 tom

n,

References

HzO HZO H*O CHCI,

25 25 25 25

CHC1, CHCl, H@ CHCl, CHCl, EtOH H20 CHC1,

135 135 28 28 28 138 138 138 138

140 140

32.2

141-2

177-80 173-6 164-7

[&,

degrees

p-dioxane

139

60 pdioxane 7.8 20.2 EtOH 15.4 H*O 20.38 CHCI, 3.6 50.2 CHC1, -6.8 k0.3 EK3H

139 27 53 53 27 27 (continued)

c 2

? $

2 e E

2.t

4 u1

TABLE VIII (continued)

Compound

4,5-di-O-benzoyl-l,3-O-benzylidene2,6-Anhydro-~~-glucitol l,%O-benzylidene4,Mi-0-benzoyl- 1,3-O-benzylidene1,5-Anhydro-L-glucitoI (2,6-anhydro-~-gulitoI) 6-deoxy-&nitro1,5-hhydro-D-glucito1(2,6-anhydro-~-gulitol; polygalitol) hepta-O-acetyl-4-O-~-D-glucopyrnnosyl-

M.P.,

B.P.,

degrees

degrees

amorph. 168-9 176-8 177-80 136-41 164-5 141-2 115-6; 135-6O

194 172 Z-S-benzyl-2-thi0-~ 167-8 104-5 3,4,6-tri-O-acetyl-ZS-benzyl-Zthio134.5-5.5 2-C-phenyl161.5-2 3,4,6-tri-O-acetyl-2-C-phenyl142-3 tetra-0-acetyl-2-C-phenyl3,4-di-0-acetyl-2-C-phenyl-6-O-p-tolylsulfonyl- 164-5 (dec.) 176.5-7.5 tri-O-acetyl-2-C-phenyl-6-O-p-tolylsulfonyl177-8 3,4di-O-acetyl-6-deoxy-6-iodo-2-C-phenyl154-5 SO-benzyl101-2 hi-0-acetyl-3-0-benzyl144-5 tri-O-acetyl8-O-p-tolyIs~IfonyI143.5-4.5 135-5.5 2,4-O-methylene176-7 4,6-benzeneboronate 4-O-p-D-glUCopyEi1IOSyl-

N

2 degrees

Rotation solvent

-7.920.2 5.8 k0.4

HzO EtOH

27 27

-3.1 40.2

CHCI,

27

- 15.2

EtOH CHCI, HZO HzO

27 27 132 132

4.6 29.5 30.5 2% 13.9 49.3 - 2.3 58.8 21.7 48.6 34.6 15.0

CHCI, HzO MeOH CHCI, EtOH CHCl, CHCI, CHCI, CHCI, CHC1, p-dioxane pdioxane

29 29 31a 31b 37 38 38 37 37 37 51b 51b 51b

62.2 - 24.2 -80"

CHCl, H& p-dioxane

54

[a],,

e0.0 -0.0 -40.2

n,

References

54 131

v)

E 4 N W

m

8

Weoxytri-0-acetyl-6-deoxy1,3-Anhydro-~-glucitoI 5,6-di-O-methyl-2,4-0-methylene2,40-benzylidene5,6-di-O-acety1-2,4-O-benzylidene1.6-Anhydro-sglucitol ~3,4,5-di-O-benzyIidene-d l,&Anhydro-L-iditol 3,4O-isopropylidenetetra-0-acetyl1,S-Anhydm-D-mannitol(styracitol) tetra-o-methyl3,Pdi-O-meth y l3,kli-O-methyl-2,6di-O-( p-nitrobenzoy1)2-S-benzyl-2-thio-c 3,4,6-tri-O-acetyI-2-S-benzy1-2-thio1,6Anhydro-~-rnannitol 2,&Anhydro-~mannitol(1,5-anhydro-~-mannitoI) ldeoxy-l-nitrotri-0-acetyl- ldeoxy- 1 - n i b 1-amino-1-deoxy-,oxalate I-amino-ldeoxy-, oxalate, monohydrate 2,5-Anhydro-~mannitoI 1,tidichloro-1,tidideoxy1,6-dichloro-1,6-dideoxy-3,4di-O-(methyIsulfonyl)3,4di-O-acetyl-1,6-dichloro-l,6dideoxy3,4di-O-acetyl-l,6-dideoxy-

149-50 119-21 9819 109-10 121-3 81-2 137-8 163-4 129-31 124-5 69-70

29 7.7 - 1 - 55.9

MeOH CHCl,

11.0 - 20.0 - 62.0 49.1 52.5 44.1

p-dioxane H,O CHCI, H*O H2O CHCI,

80-80.5 146.5-7.5 151-1.5 79-80

- 40.2 - 16.3 - 37.6 - 12.9 14.5 34

136-7

-51.1

EtOH CHCI, CHCI, MeOH MeOH CHCI, HZO

loo c/2 tom

170-1 77-8 124-5 128-31 100-01 87.6-88 98.2-99.2

137 137 51 50 51b 51b 52 52 52 52 52 1.4521

- 52.5 -69 - 42.6 -39.5 +58.2 14.2

15.2/0.2 tom 1lo/12 torr

1.4750

15.1

MeOH

141 141 141 31a 31b 31b 52 132 132 132 132 14% 58 58 58

58 (continued)

N

TABLEVIII (continued) M.P., Compound

degrees

B.P., degrees

41

[a],,,

degrees

Rotation solvent

nD

References

3,6-Anhydro-~-mannitol(1,4-anhydro-~-mannitol) 4,5-O-isopropylidene83-4 143 2-0-acetyl-4,5-0-isopropylidene-l-O-p-tolylsulfonyl- 95-6 -23.2 MeOH 143 l-bromo-l-deoxy-4,5-O-isopropylidene115-2oC/0.01torr -42.1 MeOH 1.4970 143 l-deoxy-4,5-O-isopropylidene78-82c/0.01ton -62.9 MeOH 1.4554 143 l-deoxy-4,5-O-isopropylidene-2-O-p-tolylsulfonyl70-1 -70.1 CHCI, 143 l-deoxy-l-iodo4,5-U-isopropylidene75-6 -44.6 CHCI, 143 l-amino-l-deoxy-4,5-O-isopropylidene-36.9 EtOH 1.4610 143 1,5-Anhydro-6-deoxy-~-mannitol (1,5-anhydro-~-rhamnitol) 123-4 83.8 HZO 15 tri-0-acetyl61-2 48.1 CHCl, 15 hi-0-benzoyl169-70 279 CHCl, 15 "For additional compounds, see Ref. 1 and J. Defaye, This Volume, p. 181. *Dimorphic. =Corrected configuration; see Ref. 31(b). dRing positions not determined. Bath temperature.

?

g cl

N

m

P

Q

TABLEIX Di- and Tri-anhydrohexitols and Their Derivatives"

Compound 1,5:3,6-Dianhydro-~-glucitol 2,4-O-methylene2,5:3,6-Dianhydro-~-glucitol (p-mannide) 1,4:3,6-Dianhydro-~-glucitol 2-0-acetyl-5-O-p-toly lsulfonyl-

5-0-acetyl-2-0-p-tolylsulfonyl5-nitrate 5-nitrate, 2-O-p-toly lsulfonyl2-O-(p-bromophenylsulfonyl)-, 5-nitrate di-0-nicotinoyldi-0-benzyl-

M.p.9 degrees 150-2 79.5-80 117-19 118-19

oil

52.3 74.5-5.5 75.0-6.0 143-5 155-60/0.01tom 200-10/0.01 tom 195-200/0.01ton

I

91.5-2 54.5-6 44-4.5 72.5-3.0 55.5-7 83.5-4.5

[a],,

degrees

Rotation solvent

n,

4.1 - 44.9 94.4 94

64-5 65.5-6 64-5

bis-0-(dipropylphosphinite) bis-0-(dipropylphosphonite) bis-0-(dipropylthiophosphonite)

2-0-stearoyl5-0-stearoyl1,4:3$-Dianhydro-~-iditol 2-0-acetyl-5-deoxy-5-iodo2-O-nitro-5-O-p-tolylsulfonyl2-0-nitro2(5)-O-stearoyl1.4:3.6-Dianhvdro-~-mannitol Mononicotinate

B.P.,

degrees

50.5 79.2 77.9 83.6

29.8 75.7 26.9 47.6 64.3 35.12 50.2 100.5

References 54 54 55 1

CHCI, CHCI, CHCl, CHCI,

CHCl, CHC1,

CHCI, CHCl, CHCl,

CHCl,

1.5544 1.4910 1.4750 1.5225

58 59a 59b 59a 63 63 63 129 130 144 144 144 147 147

9

z

r 0 9

z

213 $

59b 63 63 147

_

113.5- 14.5

110

CHCI,

129 (continued)

r; (D

tQ 0

TABLEIX (continued)

Compound dinicotinab

%nitrate 2-0-nitro-90-p-tolylsulfonylbis-0-(dipropylphosphinite)

m.p., degrees

bis-0-(dipropylthiophosphonite)

2(5)-O-stearoyl1,4:2,5:3,6-Trianhydro-~mannitol

b, degrees

[ah

57.5

131-2 69.0-70.5 119-20 I55-60/0.01 tom 190-5/0.01 tom

bis-0-(dipropylphosphonite)

W

64-5 57-9 66.5-7.2 68-8.6

"For additional compounds, see Ref. 1 and J. Defaye, This Volume, p. 181.

Rotation solvent CHC1,

1.4910 1.4780

42.5 79.8 101.3 128.4

n,

HA)

References 129 63 63 144 144 144 147 58 59a

(" vl

0 r 4

N

m !a

m

Q

ALDITOL ANHYDRIDES

281

TABLEX Anhydroheptitols ~

Compound 2,6-Anhydroheptito1 l-deoxy-l-nitro5,7-O-benzylidene2,6-Anhydro-~-glyceroL-manno-heptitol penta-0-acetyll-amino-l-deoxy1-amino-1-deoxy-, hydrochloride l-acetamido-3,4,5,7tetra-0-acetylI-deoxyI-amino-1-deoxy-, p-toluenesulfonate l-deoxy-l-nitro3,4,5,7-tetra-0acetyl-l-deoxyl-nitro2,6-Anhydro-~-glyceroD-gulo-heptitol penta-0-acetyll-amino-l-deoxyl-acetamido-3,4,5,7tetra-0-acety ll-deoxy2,6-Anhydro-3-deoxy-~mnnno-heptitol tetra-o-acetyl4,5,7-tri-O-acetyI1-0-p-tolylsulfonyl2,6-Anhydro-3-deoxy-~gluco-heptitol tetra-0-acety l4,5,7-tri-O-acetyIl-0-p-tolylsulfonyl4,5,7-tri-O-acetyl1-0-( p-bromopheny1)sulfonyl-* 2,6-Anhydro-3-deoxy-~galacto-heptitol tetra-0-(p-nitrobenzoy1)-

M.P., degrees 177-7.5 211-12

[&lo, degrees

Rotation solvent

8.2 -35

HZO MeOH

59a 59a

References

121-22" 55-57 189-91(dec.)

32.6" 14.2 30.0

HzO CHCI, Hz0

59d,f 59f 59f

21 1

31.3

HzO

59f

CHCI,

59f

153-54 141-43 198.5-9.5

102-3 204-5 203-5 89 91-2 164-65

1.3 19.7 36.5

18.2 0.0 20.3 optically inactive 0.0 -0.2

59g 59g CHCI, HzO

59g 59e 59d

CHCl,

-6.7

HzO

59e 59d 59e

120

4.5

CHCll

59e

152-153 123-124

60 28

HzO Me,CO

59b 59b

30

CHCI,

59b

137-138 79-80

-1 2

Hz0 Me,CO

59b 59b

117-118

-5

CHCI,

59b

104

-10

CHCl,

59h

158-159 210-211

24 -23

H1 0 CHCla

59c 59c

oil

(continued)

282

S. SOLTZBERG TABLEX (continued) M.p.9

Compound 2,6-Anhydro-3-deoxy-~talo-heptitol 4,5-O-isopropylidene2,5-Anhydro-~-glyceroL-gluco-heptitol l-deoxy-l-nitro-

[a]D,

degrees

degrees

168 104-105

68 12

155-157

-0.5

Rotation solvent

References

HzO CHCI,

59c 59c

59g

“Hemihydrate. bCrystal structure has been determined by X-ray analysis.5* TABLE XI Anhydro-octitol ~~~~

~

~

degrees

Rotation solvent

References

220-35

70

5070EtOH

59ij

206-10”

not measurable

H,O

59k

198- 205

71

50% EtOH

59k

149-50

65

5070EtOH

59k

M.P.,

Compound 1,5-Anhydro-~-erythroD-gulacto-octitol 6-acetamido-6,8dideoxy-3,4-0isopropylidene6-amino-6,8-dideoxyN-(propylhygroyl)-, hydrochloride 6-acetamido-6,8dideoxy-3,4-0isopropylidene-7-0methyl6-acetamido-6,8dideoxy-3,4-0isopropylidene-2,7-di-Omethyl~~

“Hemihydrate.

degrees

[a]D,

ALDITOL ANHYDRIDES

283

(125)A. N. Anikeeva and S. N. Danilov, Zh. Obshch. Khim., 34,2532 (1964);Chem. Abstracts, 61,16139(1964). (126)A. N. Anikeeva and S. N. Danilov, Zh. Obshch. Kham., 34, 1063 (1964);Chem. Abstracts, 61,1929 (1964). (127)E.Zissis and N. K. Richtmyer,]. Amer. Chem. So&;75,129 (1953). (128)A. K. Bhatracharya, R. K. Ness, and H. G. Fletcher, Jr., J . Org. Chem., 28,428 (1963). (129)Aspro-Nicholas, Ltd., French Pat. M2318 (Mar. 2, 1964);Chem. Abstracts, 61, 715 (1964). (130)R. Allerton and H. G. Fletcher, Jr.,]. Amer. Chem. Soc., 76,1757(1954). (131)R.J. Ferrier,]. Chem. Soc., 2325 (1961). (132)J . C.Sowden and M.L. Oftedahl, 1.Org. Chem., 26,1974(1961). (133)J. S. Brimacombe, A. B. Foster, E. B. Hancock, W. G. Overend, and M. Stacey, J. Chem. Soc., 201 (1960). (134)N. Baggett, K.W. Buck, A. B. Foster, and J. M.Webber,]. Chem. Soc., 3401 (1965). (135)J. S. Brimacombe, M. E. Evans, A. B. Foster, and J. M.Webber, J . Chem. SOC., 2735 (1964). (136)S. Tejima, T. Maki, and M. Akagi, Chem. Pharm. Bull. (Tokyo), 12,528(1964). (137)M. Akagi, S. Tejima, and M. Haga, Chem. Pharm. Bull. (Tokyo), 11, 58 (1963). (138)R. K. Ness, H. G. Fletcher, Jr., and C. S. Hudson,]. Amer. Chem. Soc., 73,3742 (1951). (139)R.J. Ferrier, A. J. Hannaford, W. G. Overend, and B. C. Smith, Carbohyd. Res., 1,38(1965). (140)S. Akiya and A. Hamada, Yakugaku Zasshi, 78, 119 (1958);Chem. Abstracts, 52, 10892(1958). (141)E.D. M.Eades, D. H. Ball, and L. Long, Jr.,J. Org. Chem., 30,3949(1965). (142)Y. Rabinsohn and H. G. Fletcher, Jr.,]. Org. Chem., 32,3452(1967). (142a)B. C.Bera, A. B. Foster, and M.Stacey,J. Chern. Soc., 4531 (1956). (143)A. B. Foster and W. G. Overend,]. Chem. Soc., 1132 (1951). (144)K. A. Petrov, E. E. Nifant’ev, A. A. Shchegolev, and N. A. Knudyntsev, Zh. Obshch. Khim., 32,3074(1962);Chem. Abstracts, 58,11456(1963). (145)A. N. Anikeeva, T. L. Orlova, and S. N. Danilov, Zh. Obshch. Khim.,31, 3544 (1961);Chem. Abstracts, 57,9934(1962). (146)K. Anno, Nippon Nogei Kagaku Kaishi, 22, 145 (1949);Chem. Abstracts, 46, 3501 (1952). (147)L.Hartmann, Atlas Chemical Industries, Inc., unpublished data.

This Page Intentionally Left Blank

THE SUGARS OF HONEY BY I. R. SIDDIQUI Food Research Institute, Canada Department of Agriculture, O t t a w a , Ontario, Canada

I. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

285

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289 ...................... ................................ 293 2. Granulation . . . . . . . . . . . . . . . . . . . . 111. Honey Oligosaccharides . . . . . . .............. . . . . . . . . . .295 1. Composition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295 2. Origin . . . . . . . .......................... 298 IV. Honey Polysaccharides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306 V. Honeydew . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307

I. INTRODUCTION “And thy Lord inspired the bee, saying: choose thou habitations in the hills and in the trees and in that which they thatch; then eat of all fruits, and follow the ways of thy Lord, made smooth (for thee). There cometh forth from their bellies a drink of diverse hues, wherein is healing for mankind.”’ This is by no means the oldest account of honey, but it is undoubtedly an authentic one, put forth 1,400 years ago. The oldest evidence of the association of man with honey dates back to the Stone Age, about 15,000 years ago, by the discovery of a prehistoric, Spanish rock-painting. From prehistoric days to the dawn of recorded history, as honey was the only concentrated sweetener available to man to satisfy his many needs, it was destined to occupy an exalted position in his art, legend, literature, medicine, mythology, and religion. The latest treatise on the biology of the honeybee2 has an entire volume devoted to history, ethnology, and folklore. It is full of amusing and little-known details, and is undoubtedly a very

(1) M. M . Pickthall, “The Meaning of the Glorious Koran,” Allen and Unwin, London, 1953, p. 199. (2) “Treatise on the Biology of the Honeybee,” R. Chauvin, ed., Masson and Co., Pans, 1968. 285

286

I. R. SIDDIQUI

interesting extension of the classical works of B e ~ s l e rB, ~~ d e n h e i m e r , ~ F r a ~ e rand , ~ Ransome.6 If the ancients regarded honey as an elixir of life, the folk medicine7-*of our time has proclaimed it to be a universal cure. Medical science has substantiated some of these claims, and has enumerated The - ~ aforementioned ~ treatise on other medicinal uses of h ~ n e y . ~ the biology of the honeybee2 has a considerable portion of Volume 3 devoted to the therapeutic aspects of hive products, including honey. The use of honey in athletic nutrition is also documented. Lloyd Percival of Sports College in Canada, from the results of a four-year study, has no hesitation in recommending honey for athletes and those interested in maintaining a high level of energy.29White lists a number of other sports30wherein honey has been used with beneficial results. There are constant reports in the bee literature advo(3) J. G. Bessler, “Geschichte der Bienenzucht,” Kohlhammer, Stuttgart, 1886. (4)F. S. Bodenheimer, “Materialen zur Geschichte der Entomologie bis Linne,” Junk, La Haye, 1928. (5)J. G. Frazer, “Totemism and Exogamy,” Macmillan, London, 1910. (6)H. M. Ransome, “The Sacred Bee,” Allen and Unwin, London, 1937. (7) D. C. Jarvis, “Folk Medicine,” Fawcett Publications, Inc., Greenwich, Conn., 1958. (8) D. C. Jarvis, “Arthritis and Folk Medicine,” Fawcett Publications, Inc., Greenwich, Conn., 1960. (9)R. Moreaux, Rev. Pathol. Comparke H y g . Gkn.,51,59(1951). (10)F. J. Pothmann, Z. Hyg. Znfektionskrankh., 130,468(1950). (11)A.C. Mikhailov, Pchelooodstoo, 2,117(1950). (12)W.Blechschmidt, Med. Monatsschr., 4,506(1950). (13)P. Beckmann, Deut. Med. Wochschr.,75,426(1950). (14)E.Hiller, Med. Monatsschr., 5,626(1951). (15)R. Chauvin, Reu. Fr. Apicult., 11,619(1951). (16)H. Stadler, Deut. Z. Verdaungs. Stoffwechselkrankh.,12,108(1952). (17)G. K.Osaulko, Vestn.Oftalmol., 32,25(1953). (18)K.Fochem, Strahlentherapie, 93,466(1954). (19)0. Martensen-Larsen, Brit. Med.].,4885,468(1954). (20)M.Pestel, Proc. Med. Gaz. Apicult., 56,63(1955). (21)M.W.Bulman, Middlesex Hosp.].,55,63(1955). (22) N.P. Ioirish, Pchelouodstuo, 33,50(1956). (23) K.Von Am, Praxis (Bern), 46,717 (1957). (24)N. Rubin, A. R. Gennaro, C. N. Sideri, and A. Osol, Amer. J. Pharm., 131, 246 ( 1959). (25)F. Alison and R. Narbouton, Nutr. Dieta, 3,33(1959). (26)V. I. Maksimenko, Pchelouodstuo, 37,49(1960). (27)F. Baungarten, Deut. Bienenwirtsch., 11,137(1960). (28)R. Lyonnet and R. Berthelier, Lyon Med., 38,505(1961). (29)L. Percival, Amer. BeeJ.,95,390(1955). (30) J. W.White, Jr., in “The Hive and the Honey Bee,” R. A. Grout, ed., Dadant and Sons, Inc., Hamilton, Illinois, 1963,p. 369.

THE SUGARS OF HONEY

287

cating honey for diabetics; such claims are unfounded. It is, however, possible that honey is a lesser evil to the diabetic than table sugar (sucrose), but surely, because on the average its content of D-glucose is -3070,the use of honey by diabetics cannot be freely advocated. The U. S. Federal Food and Drug Act of 1906 defined honey as “The nectar and saccharine exudation of plants, gathered, modified and stored in the comb by honey bees (Apis mellifera and Apis dorsata) is levorotatory; contains not more than 25% water, not more than 0.25% ash and not more than 8% sucrose.” According to White30 and Feinberg,31 these figures are unrealistic; the values for moisture and sucrose are too high, and the recommended values for ash are far too low. The Canadian Food and Drug Act and Regulations state that honey should be derived entirely from the nectar of flowers and other sweet exudations of plants by the work of the bees, and should not contain more than 20% of moisture, 8% of sucrose, and 0.25% of ash, and should contain not less than 60% of invert sugar. In most European countries, honey is defined in similar terms. However, certain quality factors considered by Europeans, especially the Germans, of importance in the marketing of honey are the levels of the enzymes invertase and diastase, and of 5-(hydroxymethyl)-2furaldehyde. White has discussed these requirements in relation to suggested standards for the Codex A l i m e n t a r i u ~ The . ~ ~ German insistence on these requirements is outlined in a volume of A p i a ~ t a . ~ ~ The Codex Alimentarius Commission, in its Draft Provisional Standard34 (following definition and description, and subsidiary definitions and designations) has laid down certain compositional criteria. It is obvious that all definitions of honey so far discussed deal with two chemically and physically distinct commodities: namely, nectar honey and honeydew honey. A comparison between the two shows that honeydew honey is lower in D-fruCtOSe and D-glucose and higher in pH, oligosaccharides, acidic components, ash, and nitrogen than nectar honey. A distinct feature of honeydew honey is the trisaccharide melezitose, which has been identified in the exudate manna of the Douglas fir; the European larch, and the North American Jack or scrub pine. According to the late Professor C. S. Hudson, Turkestan manna contains 20-38% of melezitose, and Douglas-fir manna, 5070 of melezitose. At one time, melezitose formed 20-30% of the total (31) B. Feinberg, Amer. BeeJ.,91,471 (1951). (32) J. W. White, Jr., Amer. Bee]., 107, 374 (1967). (33) H. Duisberg, Apiactn, 3,26 (1967). (34) Joint FAO/WHO Food Standards Program, Codex Alimentarius Commission Draft Provisional Standardfor Honey, March, 1968.

288

I. R. SIDDIQUI

sugars of Maryland and Pennsylvania honeys that had been collected during a period of drought when nectar was not readily a ~ a i l a b l e . ~ ~ . ~ Regarding the origin of melezitose, it has been stated that the sugar arises by the action of an aphid enzyme on plant-sap su~rose.~’ It is therefore obvious that bees could not be regarded as the sole producers of honeydew honey. For this reason, and because nectar honey is a commodity that is traded internationally, it will be advisable to treat it as an entity separate from honeydew honey. Further distinctions between floral or nectar honeys, as set forth in the Codex Standard, are based on mixed floral and unifloral sources of nectars, and are too sophisticated to be practical. Such limits as 65 5% for the invert sugar and similar limits for other variables, to embrace all or most nectar honeys, would constitute better criteria. The terms “reducing sugar as invert sugar” and “apparent sucrose,” used in the Codex draft, are ambiguous and misleading. The characteristic to be given should be either (a) total reducing sugar, which includes D-glucose, D-fructose, and the reducing oligosaccharides, or (b) invert sugar, which represents merely D-glucose and D-fruCtOSe. Similarly, “apparent sucrose” should be modified to “nonreducing oligosaccharides,” in view of the fact that the Codex methods currently in use measure (a) total reducing sugar and (b) nonreducing oligosaccharides (by difference, after hydrolysis). If such values as percentage of invert sugar, true sucrose, and other reducing and nonreducing oligosaccharides, some of which are present in honey in proportions greater than s u c r o ~ e ,are ~ ~ to * ~be ~ measured, resort will have to be made to chromatographic methods. In this connection, adoption of a paper-chromatographic procedure, followed by quantitative elution and colorimetric determination of the components, will have a better chance of wide acceptance than a more sophisticated, elaborate, and expensive gas-phase chromatographic procedure. Most of us know that bees are busy little creatures, but how busy and persistent they are is apparent from the estimate that a worker bee gathers, from one flower, a minimum of 10 micrograms of nectar and brings back after one flight a total of 50 milligrams, yielding approximately 20 milligrams of honey; to produce a kilogram, the bees would have to fly approximately 250,000 miles, or the equivalent _+

(35) C. S. Hudson and S. F. Sherwood,]. Amer. Chem. SOC., 40,1456 (1918). (36) C . S. Hudson and S. F. Sherwood,j. Amer. Chem. SOC., 42,116 (1920). (37) J. S. D. Bacon and B. Dickinson, Biochem.J.,66,289 (1957). (38) I. R. Siddiqui and B. Furgala,.!. Apicult. Res., 6, 139 (1967). (39) I. R. Siddiqui and B. Furgala,J.ApicuEt. Res., 7,51(1968).

THE SUGARS OF HONEY

289

of a flight to the It is also estimated that the present worldproduction of honey is 500,000 tons, produced by about 5 million beekeepers from 40-50 million colonies comprised of 1.5 billion bees .41 11. HONEYMONOSACCHARIDES

1. Composition and Analysis In the production of honey, the foraging bees carry the nectar in their honey sac, mix it with enzyme-rich secretions from their hypopharyngeal glands, carry it to the hive, and pass it over to the house bees. The house bees, in turn, transmit it among themselves, and carry out the process of ripening by mixing it with further amounts of glandular secretions and by removing water. The process of ripening is continued until the raw material loses about 50% of its water content; subsequent ripening takes place automatically in the cells of the comb, in the stream of dry air that constitutes the ventilation system of the hive. When a moisture content of 20% is reached, the cells containing the honey are capped by the bees.42a43 During the whole process of ripening, and afterwards in storage, honey undergoes very intricate enzymic reactions about which very little is known. However, two fundamental steps in the manufacture of honey are the removal of water and the inversion of the sucrose contained in the nectar. Honey, therefore, consists of a concentrated solution of two monosaccharides, D-glUCOSe and D-fructose, and it is correct to state that these sugars constitute, in the free and the combined form, -95% of the honey solids. The ratio of D-glucose to D-fructose is characteristic of certain types of honey. In mixed, floral honeys, they are present in almost equal proportions. On the other hand, unifloral honeys contain appreciably more D-fructose than D-glucose. Examples of honeys rich in D-fructose are Robinia, Salvia, Tupelo, and Sweet chestnut; honeys rich in D-glucose are rarer, examples being those from dandeA monosaccharide derivative that occurs in honey to lion and rape.44*45 (40) I. Khalifman, “Bees,” Foreign Languages Publishing House, Moscow (1953). (41) E. Crane, in Ref. 30, p. 18. (42) 0. W. Park,]. Econ. Entomol., 26, 188 (1933). (43) J . F. Reinhardt,J. Econ. Entomol., 32,654 (1939). (44) A. Maurizio, Bee World,40,275 (1959). (45) J . W. White, Jr., M . L. Riethof, M. H. Subers, and I. Kushnir, US.Dept. Agr. Tech. Bull. No. 1261, (1962).

290

I. R. SIDDIQUI

an appreciable extent is D-gluconic acid; inorganic and organic acids, including amino acids, are also present, and D-gluconic acid con~ . ~formation ' in honey is stitutes 70-90% of the organic a ~ i d s . ~ Its ascribed to the presence of a D-gluCOSe oxidase that produces Dgluconic acid and hydrogen peroxide from D-glUCOSe.4s Inhibine, the antibacterial agent reported in honey, is now recognized to be hydrogen peroxide. The presence of traces of D-galaCtOSe was reported in one instance,49but, in the author's opinion, this was merely an attempt to justify the frequent reports of the occurrence of raffinose in honey. Analysis of honey on this continent dates back to 1892, when 500 commercial samples of honey were analyzed.50 The analytical methods used were, with certain modifications, employed over the years by several worker^,^'-^^ until a critical study of the methods of sugar analysis was made by White and coworkers.55 Following the development of carbon-Celite column chromatogr a p h ~ and , ~ ~its application to sugar analysis of honey by White and collaborators, a new procedure for the analysis of honey emerged and was termed the selective-adsorption method. Application of this technique resulted in the recognition of the presence in honey of reducing disaccharides and higher oligosaccharides, and more accurate values for the content of D-gluCOSe and D-fructose were ~ b t a i n e d . ~ ~This + l method was also used in Canada,6z Chile,63and (46)E. E.Stinson, M. H. Subers, J. Petty, and J. W. White, Jr., Arch. Biochem. Biophys., 89,6(1960). (47)S. Maeda, A. Mukai, N. Kosugi, and Y. Okada, J . Food Sci. Technol. (Japan), 9, 270 (1962). (48)J. W. White, Jr., M. H. Subers, and A. I. Schepartz, Biochim. Biophys. Acta, 73, 57 (1963). (49)S. Goldschmidt and H. Burkert, Z. Physiol. Chem., 300,188(1955). (50)H.W. Wiley, U . S. Diu. Chem. Bull., 13,pt. 6,744(1892). (51)C.A. Browne, U . S . Bur. Chem. Bull., llO(1908). (52)R. E.Lothrope and R. L. Holmes, lnd. Eng. Chem.,Anal. Ed., 3,334(1931). (53)J. E. Eckert and H. W. Allinger, Caltf.Agr. E x p . Sta. Bull., 631 (1939). (54)J. A. Elegood and L. Fischer, Food Res., 5,559(1940). (55)J. W. White, Jr., C. Ricciuti, and J. Maher, J . Assoc. Ofic.Agr. Chemists, 35, 859 (1952). (56)R.L.Whistler and D. F. Durso,J. Amer. Chem. SOC.,72,677(1950). (57)J.W. White, Jr., and J. Maher,]. Assoc. Ofic. Agr. Chemists, 37,466(1954). (58)J. W. White, Jr., and J. Maher,]. Assoc. Ofic. Agr. Chemists, 37,478(1954). (59)J. W. White, Jr.,J. Assoc. Ofic. Agr. Chemists, 40,326(1957). (60)J. W. White, Jr.,J.Assoc.Ofic. Agr. Chemists, 42,341(1959). (61)R. A. Osborn, M. Oakley, and K. L. Milstead, J . Assoc. Ofic. Agr. Chemists, 42, 26 (1959). (62)G. H. Austin, Proc. Intern. Congr. Entomol. IOth, Montreal, 1956,4,1001(1958). (63)D.A. Bravar, An. Tac. Quim. Farmac., 9,149(1958).

THE SUGARS OF HONEY

291

South The average composition of 490 samples of honey, with respect to sugar constituents and moisture and their range of values, as determined by White and collaborator^,^^ is given in Table I. The values given in Table I for maltose and sucrose do not TABLEI Average Composition of Honey from 490 Samples45 Component

Percentage

Standard deviation

Range

Moisture D-Fructose D-Glucose Sucrose Maltose Higher sugars Undetermined

17.2 38.19 31.28 1.31 7.31 1.5 3.1

1.46 2.07 3.03 0.95 2.09 1.03 1.97

13.4 -22.9 27.25-44.26 22.03-40.75 0.25- 7.57 2.74-15.98 0.13-13.2 0.0 -13.2

represent the true content of these sugars. Sucrose values are interfered with by other ketose disaccharides, including turanose, and maltose values are subject to interference by other reducing disaccharides, such as kojibiose and n i g e r o ~ e . ~ ~ In recent years, it has become fashionable to determine the composition of honey, and in the current literature are many reports dealing with the compositional aspects of honey from various parts of the w ~ r l d . ~ It ~ -is, ' ~however, regrettable that application of paper chromatography, which has played an outstanding role in the development of carbohydrate chemistry, has not yet been fully explored in (64) R. H. Anderson and L. S. Perold, S. Afr.1. Agr. Sci., 365 (1964). (65) C. Pelimon and H. Baculinschi, An. Inst. Cercet. Zooteh. (Bucharest), 13, 621 (1955). (66) A. K. Mallick, J . Indian Inst. Chem., 30, 171 (1958). (67) Y. Arai, K. Akiyama, S. Sakai, M. Doguchi, N. Suzuki, and S. Ogawa, l a p . B e e l . , 13, 101 (1960). (68) T. Watanabe, Y. Motomura, and K. Aso, TohokuJ.Agr. Res., 12,187 (1961). (69) J. Pourtallier, Bull. Apic. Inform. Doc. Sci. Tech.,5,138 (1962). (70) C. H. W. Flechtmann, C. F. Caldas Felho, E. Amaral, and J. D. P. Arzolla, Bol. Ind. Animal. (Sao Paulo), 21,65 (1963). (71) V. G. Chudakov, Pchelooodstoo, 40,18 (1963). (72) S. Aoyagi, K. Fudeya, and S. Takeshima, Tamagawa Daigaku Nogakutsu Kenkyu Hokoku, 8, No. 7,181 (1968). (73) Y. Okimoto, K. Toshida, M . Yamazaki, and H. Fujioka, Tamagawa Daigaku Nogakutsu Kenkyu Hokoku, 8, No. 7,184 (1968).

292

I. R. SIDDIQUI

connection with the quantitative analysis of sugars of honey. In this laboratory, the analysis of a large number of honey samples from various parts of Canada was undertaken, in order to standardize Canadian honey composition.73aThe sugars were separated by paper chromatography, quantitatively eluted, and determined spectrophotometrically by the Nelson-Somogyi m e t h ~ d . ? ~The - ~ ~average composition of 95 Canadian honeys is given in Table 11. TABLEI1 Average Composition of Honey from 95 Samples73o Component Moisture D-Fructose D-Glucose Oligosaccharides Undetermined

Percentage

Range

17.9 37.1 33.7 7.4 3.9

15.0-21.8 3 1.1-4 1.4 28.5-40.7 2.2-15.2 0.0-10.8

Of the 95 samples examined, 4.2% had monosaccharides below 65%; 32.6%, between 65 and 70%; and 63.296, above 70%. It is not intended to imply that this method of analysis is the only one worthy

of consideration; other colorimetric methods in conjunction with paper chromatography will be as useful. What is implied is that simple, paper-chromatographic separations, followed by colorimetric determinations, provide results equally as good as, if not better than, those afforded by other available methods. Moreover, the chromatographic methods are easy to handle and less time-consuming, and do not require complicated instrumentation. However, paper chromatography has limitations; for example, its lack of resolution between oligosaccharides (such as sucrose from turanose, maltose from other disaccharides, and melezitose from other trisaccharides) does not permit determination of important individual components. Such problems can be resolved by the application of paper electrophoresis, thin-layer chromatography, or gas-liquid chromatography. The use of thin-layer chromatography for the determination of sugars in honey has been p r ~ p o s e d , and ~ ' the potential value of gas(73a)I. R. Siddiqui and associates, unpublished results. (74) N. Nelson,]. B i d . Chem., 153,375 (1944). (75) M . Somogyi,]. Biol.Chem., 160,61(1945). (76) M . Somogyi,]. Biol.Chem., 195,19(1952). (77) J. Pourtallier, Bull. Apic. Inform. Doc. Sci. Tech., 7,197 (1964).

THE SUGARS OF HONEY

293

liquid chromatography in the quantitative analysis of the sugars of honey has been demonstrated. Experiments testing the reproducibility of the results were conducted with one sample of honeydew honey from a coniferous tree. In four independent determinations, the following results were obtained: D-fructose, 36.7; 37.0; 36.8; and 37.2%; and D-glucose, 27.8, 28.0, 27.5, and 28.0%. The sugars of honey were transformed into their volatile per(trimethylsily1) derivatives, and these were injected into a gas chromatograph. The chromatography was conducted with a column (10 ft. X 0.125 in.) of 4% SE 52 on Chromosorb W (100-120 mesh) at a nitrogen flow-rate of 30 ml/min, with temperature programming. Good separations of D-glucose from D-fructose were obtained. The retention times for the D-fructose and P-D-glucose derivatives, with reference to that of a-D-glUCOSe, were 0.73 and 1.51, re~pectively.'~ The resolution of oligosaccharides is, however, far from optimal and is, indeed, a challenging problem because of their diversity and complexity. Perhaps, further experimentation, with different column packings and manipulation of conditions of analysis, would lead to a solution of the problem. The present challenge is to determine, quantitatively, most, if not all, of the sugars in honeys, and not merely to determine the content of D-glucose and D-frUCtOSe b y yet another method.

2. Granulation Granulation of honey has been a problem to the honey producer for a long time. A fully crystalline pack is acceptable, but a clear, liquid pack is more appealing to the eye of the consumer and lends itself more readily to packaging in the little containers that have become popular in recent years. Partially crystalline packs containing much larger crystals constitute a real problem; such packs sometimes have to be redissolved and repacked, resulting in additional costs. The quest for methods for keeping honey in the liquid state for reasonable periods of time has demanded the attention of honey specialists for a long time. Several indexes have been proposed for describing the tendency of honey to granulate, in order to provide a basis for blending honeys for packing in the liquid form. Among these, the L/D ratio (where L denotes levulose and D, dextrose) was, for a long time, considered a criterion of the crystallization (78) J. Pourtallier, Z . Bienenforsch.,9,217(1968).

294

I . R. SIDDIQUI

potential of honey; high values are indicative of liquid, or slowgranulating, honey. From a study of the solubility of pure D-glucose in the presence of D-fructose, sucrose, and a combination of the two, Jackson and Silsbee79proposed two indices, namely, the supersaturation coefficients and the (D-W)/Lratio, where W denotes water. Later, on the basis of a comparison of L/D ratios and supersaturation coefficients, Austin62 concluded that the supersaturation coefficient is a more logical index, but that it is not so good as the D/W ratio; he pointed out that the latter was simpler to determine, since it could be calculated from only two variables. He also suggested that, when honeys are compared by this method, their composition should be calculated on the basis of equivalent moisture-content. White and have shown that DlW ratios, not adjusted to a common moisture-content, show the most closely related relationship to the granulating tendency than does any other index. DlW ratios of 1.7 or lower indicate nongranulating honey, and values of 2.1 and higher predict occurrence of rapid granulation. In the writer’s laboratory, results obtained from the analysis of a large number of samples of Canadian honey by paper chromatography were expressed in these ratios; namely, L/D, (D-W)/L, and DlW, and these were compared with the granulation rate observed for each sample during 6 months by the method of White and coworkers.45 For the Canadian honeys thus examined, there was no discernible relationship between the rate of granulation and any of the indexes used. The purpose of a granulating index is to relate the composition of honey to its granulating tendency, in order that such behavior may be predicted; such predictions were, of course, not possible, because the factor actually involved is the presence or absence of appropriate crystal nuclei. In conformity with this conclusion are the results on the crystallization of honey published in a recent bookeofrom the Ministry of Agriculture in Greece. In relating crystallization to chemical composition, the authpr proposed yet another index. This was, however, a nongranulating index B-D/D, where B represents degrees Brix and D represents the content of D-glucose. As could have been predicted, the indexes L/D, (D-W)/L, and D/W were not useful in this connection. (79) R. F. Jackson and C. G. Silsbee, Nut. Bur. Stand. (U.S . ) Technol. Pap., No. 259, 18,277 (1924). (80) M. I. Kodoyne, “The Crystallization of Honey,” Ministry of Agriculture, Athens, Greece (1962).

THE SUGARS OF HONEY

295

111. HONEYOLIGOSACCHARIDES 1. Composition

For a long time, honey was considered to be a mixture of D-glucose, D-fructose, and sucrose. Although the presence of maltose in some honeys was recognized 45 years ago,81it remained for Van VoorstR2 and Hurd and coworkersR3to show that it is probably a component of all honeys. Recognition of the presence of other oligosaccharides (for systematic names, see Table 111) was only a matter of time, and was made possible by the development of paper chromatography and its application to sugar analysis. By using paper chromatography, KeupS4demonstrated the presence of as many as 15 components, corresponding, among others, to maltose, turanose, isomaltose, erlose, kestose, raffinose, and melibiose. After fractionation of honey by carbon-column chromatography and paper chromatography, Aso and coworkersE5detected 22 components, 15 of which were classified as ketoses. Other reports dealing with the paper-chromatographic identification of honey oligosaccharides are those of Goldschmidt and B ~ r k e r t , 4 ~ Po~rtallier,~ Flechtmann ~ and coworkers,70 and Curylo.H6Attempts to identify the oligosaccharides of honey by methods other than chromatography were limited by the difficulties encountered in the isolation of these compounds in the pure form. An attempt to isolate and identify some of the disaccharides and their octaacetates by infrared spectroscopy was made by White and HobanR7by comparison of the spectra with those of authentic samples of disaccharides and their derived octaacetates. Fractionation of honey by carbon-Celite chromatography, and analysis by gas-liquid chromatography of the per(trimethylsilyl) derivatives of the fractions eluted with water, and 2.5%, 5%, and 10% aqueous ethanol, resulted in the detection of a glucose, a fructose, sucrose, turanose, maltulose, leucrose, kojibiose, nigerose, maltose, isomaltose, melezitose, and two undetermined components.s8 However, (81) E. Elser, Mitt.Geb. Lebensmittelunters. Hyg., 25, 92 (1924). (82) F. T. Van Voorst, Chem. Weekbl.,38,538 (1941). (83) C. D. Hurd, D. T. Englis, W. A. Bonner, and M . A. Rogers, J . Amer. Chem. SOC., 66,2015 (1944). (84) N. Keup, Inst. Grand-Ducal Luxembourg, Sect. Sci. Nut. Phys. Math. Arch., 24, 91 (1957). (85) K. Aso, T. Watanabe, and K. Yamau, Hakko Kogaku Zasshi, 36,39 (1958). (86) J . Curylo, Pszczel. Zeszyty Nauk., 6,1(1962). (87) J. W. White, Jr., and N . Hoban, Arch. Biochem. Biophys., 80,386 (1959). (88) J . Matsuyama, G. Fudeya, T. Ishida, and T. Echigo, Tamagawa Daigaku Nogakubu Kenkyu Hokoku, 8, No. 7, 188 (1968).

I. R. SIDDIQUI

296

TABLEI11 Glossary of Relevant Oligosaccharides Trivial name Centose Erlose Gentiobiose Isomaltose Isomaltotriose Isomaltotetraose

Isomaltopentaose

Isomaltulose Isopanose I-Kestose Kojibiose Laminarabiose Leucrose Maltose Maltotriose Maltulose Melibiose Nigerose Panose Raffinose Sucrose Theanderose a$-Trehalose Turanose (I

n

a

‘No recognized trivial name.

Systematic name O-a-D-GlUCOpyranOSyl-(1+4)-0-[a-~glucopyranosyl-(1+2)1-D-glucopyranose O-CZ-D-GI ucopyranosyl-( 1+ 4)-a-~-glucopyranosyl j3-D-fructofuranoside O-P-D-Glucopyranosyl-(1+ 6)-D-glucopyranose O-a-D-Clucopyranosyl-(1+ 6)-D-glucopyranose 0-a-D-Glucopyranosyl-(1+ 6)-O-a-D-glucopyranosy~(1+6)-D-glucopyranose O-a-D-Glucopyranosyl-(1+ 6)-O-a-D-glucopyranosyl(1 +6)-O-a-~-glucopyranosyl-( 1 6)-~glucopyranose O-a-D-Ghcopyranosyl-(1+ 6)-O-a-D-glucopyranosyl(1+6)-O-a-~-glucopyranosyl-(1 4 6 ) - 0 - a - ~ gIucopyranosyl-(1-+6)-~-glucopyranose O-a-D-Glucopyranosyl-(1+ 6)-D-fmctofuranose 0-a-D-Glucopyranosyl-(1+ 4)-O-a-D-glucopyranosyl(1+6)-D-glucopyranose O-cY-D-Glucopyranosy1-(1+2)-j3-D-fructofuranosyl(1+2) j3-D-fructofuranoside O-a-D-Glucopyranosyl-(1+ 2)-D-glucopyranose O-j3-D-Glucopyranosyl-(1+ 3)-D-glucopyranose O-a-D-Glucopyranosyl-(1+5)-D-fructopyranose O-a-D-Ghcopyranosyl-(1+4)-~-glucopyranose O-a-D-Glucopyranosyl-(1+ 4)-O-a-D-glucopyranosyl(1+4)-D-glucopyranose 0-a-D-Glucopyranosyl-(1+4)-D-fructose 0-a-D-Galactopyranosyl-( 1+ 6)-D-glucopyranose O-a-D-GlucopyranosyI-(1+ 3)-D-glUCOpyrWOSe 0-a-D-Glucopyranosyl-(1+ 6)-O-a-D-glucopyranosyl(1-+4)-D-glucopyranose 0-a-D-Galaclopyranosyl-(1+6)-0-a-~glucopyranosyl-(1+2) j3-D-fructofuranoside a-D-Glucopyranosyl j3-D-fructofuranoside 0-a-D-Glucopyranosyl-(1+ 6)-a-D-glucopyranosyl j3-D-fructofuranoside a-D-Glucopyranosyl P-D-glucopyranoside O-a-D-Glucopyranosyl-(1 3)-D-fructose O-a-D-Glucopyranosyl-(1+ 6)-O-a-~-glucopyranosyl(1-B 8)-D-glucopyranose O-B-D-Glucopyranosyl-(1+ 6)-O-a-D-glucopyranosyl(1+4)-D-g~ucopyranose 1-O-a-bglucopyranosyl-D-fructose -+

THE SUGARS OF HONEY

297

the first unequivocal identification of four disaccharides, namely, kojibiose, nigerose, maltose, and isomaltose, in honey -as their crys~ ~ work talline P-octaacetates -was made by Watanabe and A s o . This was made possible by the development of Magnesol-Celite column chromatography by Wolfrom and coworkers.90 Following these reports, a large-scale fractionation was undertaken by Siddiqui and F ~ r g a l a ~on * , a~ carbon-Celite ~ column; they used an enriched oligosaccharide fraction, isolated from honey by using a batch operation employing carbon-Celite as the adsorbent. The fractions eluted with 2.515% aqueous ethanol were further separated by paper chromatography, occasionally by paper electroph~resis,~' and, frequently, by thin-layer c h r o m a t ~ g r a p h y after , ~ ~ acetylation. These fractionation procedures resulted in the characterization of at least 24 oligosaccharides. The oligosaccharides maltose, kojibiose, isomaltose, nigerose, a$-trehalose, gentiobiose, laminarabiose, melezitose, O-a-D-glucopyranosyl-( l+6)-O-a-D-g~ucopyranosyl-(1-4)D-glucopyranose, and maltotriose were identified as crystalline P-octa- and hendeca-acetates. Sucrose, turanose, l-kestose, and panose were identified as the crystalline sugars, maltulose as the crystalline phenylosazone, and isomaltotriose as its crystalline P-hendecabenzoate. Erlose and theanderose were identified by the isolation and characterization of the disaccharides derived by partial hydrolysis with acid or an enzyme. Isopanose, isomaltotetraose, and isomaltopentaose were identified b y their specific optical rotations and behavior on hydrolysis with acid or an enzyme. The disaccharides isomahlose and l-O-a-D-glucOpyranOSyl-D-fructose were tentatively identified. Evidence was also presented for the presence in honey of two new trisaccharides, namely centoseS3{O-a-D-glucopyranosyl(1+4)-O-[a-D-glucopyranosyl-( 1-+2)]-D-glucopyranose} and 0-p-Dglucopyranosyl-(1+6)-O-a-~-glucopyranosyl-(i ~ ~ ) - D - g ~ u c o p y r a n o s e . A synthesis of centose s3a was presented during the preparation of this review. The approximate proportions of these sugars in the oligosaccharide fraction (3.65%) of the honey were: maltose, 29.4; kojibiose, 8.2; turanose, 4.7; isomaltose, 4.4; sucrose, 3.9; ketose band (mixture (89) T. Watanabe and K. Aso, TohokuJ.Agr. Res., 11,109 (1960). (90) W. H. McNeely, W. W. Binkley, and M. L. Wolfrom,]. Amer. Chem. Soc., 67,527 (1945). (91) A. €3. Foster,]. Chem. SOC., 982 (1953). (92) M . E. Tate and C. T. Bishop, Can.].Chem.,40,1043 (1962). (93) I. R. Siddiqui and B. Furgala, Carbohyd. Res., 6,250 (1968). (93a)B. H. Koeppen, Carhohyd. Res., 13,417 (1970).

298

I. R. SIDDIQUI

of at least three ketoses, including maltulose and isomaltulose), 3.1; nigerose, 1.7; a,p-trehalose, 1.1;gentiobiose, 0.4; laminarabiose, 0.09; erlose, 4.5; theanderose, 2.7; panose, 2.5; maltotriose, 1.9; 1-kestose, 0.9; isomaltotriose, 0.6; isomaltotetraose, 0.33; melezitose, 0.3; isopanose, 0.24; isomaltopentaose, 0.16; centose, 0.05%; O - ~ - D glucopyranosyl-(1+6)-O-a-D-glucopyranosyl-( 1+3)-D-glucopyranose, minute proportion; and O-p-D-glucopyranosyl-(1+6)-O-a-D-glucopyranosyl-(1+4)-D-glucopyranose, minute proportion. This extensive analysis also revealed certain features of composition that were at variance with the results of other workers. Although Watanabe and Asos9 detected leucrose in their sample of honey by paper electrophoresis, its presence could not be confirmed by Siddiqui and F ~ r g a l a . On ~ ” ~the ~ other hand, although the former workers searched for trehalose, they failed to detect its presence. Another point of difference from the results in the literature was uncovered in the reported presence of raffinose in a variety of honeydew and nectar honeys.49.70.77.84*H6 It was found that 0-a-D-glucopyran0syl-(1+6)-a-D-glucopyranosyl p-D-fructofuranoside, assigned the trivial name theanderose, had been confused with raffinose owing to the identical behavior of the two sugars in paper chromatography and toward ketose spray-reagents. This clarification further demonstrated that paper-chromatographic methods alone are insufficient to warrant such conclusions. The confusion would not have arisen had an attempt been made to confirm the supposed presence of raffinose by paper electrophoresis, since the M G values (0.18 for theanderose and 0.31 for raffinose) are sufficiently different to have clarified the situation. 2. Origin

The answer to the biochemical origin of honey oligosaccharides lies in recent developments in the field of sugar and enzyme chemistry. It is obvious that, in the formation of these saccharides, both trans-D-glucosylation and trans-D-fructosylation occur. In the former process, D-glucopyranosyl groups are transferred from a D-glucopyranosyl donor to an acceptor molecule, which may be a mono-, oligo-, or poIy-saccharide. In the latter process, D-fructofuranosyl groups are transferred by a D-fructofuranosyl-transferring enzyme to other sugars, giving D-fructose-containing oligosaccharides. These transfer reactions probably occur through the formation of a Dfructofuranosyl-enzyme or D-g~ucopyranosy~-enzyme complex, but there is at present no direct evidence for the formation of such

299

THE SUGARS OF HONEY

complexes. A simple scheme for trans-D-fructosylation has been postulated as follows.94 D-Fruf-D-Gp

+E

+

~-Frnf-E D-Fruf-D-Gp D-Fruf-E

Fruf-E

+ D-Gp

D-Fruf-Fruf-D-Gp

+E

+ D-Fruf-D-Fruf-D-Gp

+

D-Fruf-D-Fruf-D-Fruf-D-Gp E, and so on,

where D-Fruf= D-fructofuranosyl, E =the enzyme, and D-Gp= D-glucopyranosyl. Trans-D-glucosylation can similarly be depicted by the following scheme:

+

D-GP-D-G~ E e D-Gp+ D-GP-E

+ D-GP-D-GP-D-G~ + D-GP-E * D-GP-D-GP-D-GP-D-G~+ E, and so on. +

E~ D-GP-D-G~ D-GP-E e D - G ~ - D - G ~ - D - G

Trans-D-glucosylation in the monosaccharide series has been known for a long time,95 but, in the oligosaccharide series, it was first described by Pigman and Blair.96.97The ability of carbohydrases to catalyze both the hydrolysis and the synthesis of oligosaccharides is now well established. It has been reported that preferential trans-D-glucosylation to primary alcoholic groups on aldose-containing substrates is brought about by many a-D-glucosidases from molds.98 Although maltose is reported to be present in some nectars,99 it is unlikely that the sugar occurs in proportions sufficient to account for the percentage found in honey. In conformity with this conclusion, certain evidence indicates that it is also formed in the bee stomach by tranS-D-ghCOsylation. White and Maherloo have demonstrated the presence in “honey invertase” of a trans-D-glucosylase capable of synthesizing maltose and isomaltose from D-glucose. Studies have been reported on the trans-D-glucosylating action of the tropical yeast Schizo(94) E. H. Fischer, J. Kohtes, and J . Fellig, Helo. Chim. Acta, 34, 1132 (1951). (95) J. Rabate, Compt. Rend., 204,153 (1937). (96) W. W. Pigman,]. Res.Not. Bur. Stand.,33, lOS(1944). (97) M. C. Blair and W. W. Pigman, Arch. Biochem. Biophys., 48,17 (1954). (98) J. Edelman,Aduan. Enzymol., 17,189 (1956). (99) C . R. Wykes, New Phytologist, 51,210 (1952). (100) J. W. White, Jr., and J. Maher, Arch. Biochem. Biophys., 42,360 (1953).

300

I. R. SIDDIQUI

saccharomyces pombe on a variety of substrates, including maltose,10'.'02 ~ - g l u c o s e , 'and ~ ~ a starch-syrup substrate.lM From these investigations, it follows that D-glucose and several members of the a - ~ 1+4) -( homologous series give rise to such oligosaccharides as kojibiose, nigerose, isomaltose, panose, and isomaltotriose. From the work of Peat and coworkers,lo5it is known that small proportions of nigerose are produced by the action of an enzyme preparation of Aspergillus niger on concentrated solutions of D-glucose. However, Aspergillus oryzae enzyme yields a substantially greater proportion of nigerose from a mixture of D-glucose and maltose.'06 Isomaltose, panose, and isomaltotriose have also been synthesized from maltose by various enzyme preparations from mold^.'^^-'^^ However, panose was first crystallized by Pan and coworkersl'O as a product of trans-D-fructosylation of maltose b y a culture filtrate of Aspergillis niger. Yasumura"' found maltulose and turanose in the mixture obtained by action of a brewers' yeast extract on sucrose; these were obviously formed by a transfer of &-D-glUCOSyl groups to 0 - 3 and 0-4of D-fructose. Avigad'I2 has reported the isolation of a number of oligosaccharides, including ~-O-a-D-g~ucopyranosy~-Dfructose, isomaltulose, and maltulose, as products of trans-D-glucosylation of D-fructose moieties by a hybrid-yeast a-D-glucosidase. A trehalose is an important constituent of the blood sugar of ins e c t ~ , " ~ , "including ~ the honey bee,'lS where it ha.s been reported to occur as the a,a form. By paper chromatography, it was shown'ls that the blood of pollen-collecting bees is very rich in a trehalose, and that the level of D-glucose and D-fructose is almost nil. On the other hand, the blood of the nectar-collecting bees is comparatively rich in D-glucose and D-fructose. However, Maurizio116has found (101)K.Shibasaki and K. Aso, TohokuJ.Agr.Res., 5,131(1954). (102)K.Shibasaki, TohokuJ.Agr. Res., 6,47(1955). (103)K.Shibasaki, Tohoku]. Agr. Res., 6,171(1955). (104)K.Shibasaki, Mem. Publ. Fac. Agr., Tohoku Uniu., 26 (1958). (105)S.Peat, W. J. Whelan, and K. S. Hinson, Chem. lnd. (London), 385 (1955). (106)J. H. Pazur, T. Budovich, and C. L. Tipton,]. Amer. Chem. Soc., 79,625(1957). (107)J . H.Pazur and D. French,]. Biol. Chem., 196,265(1952). (108)S.A. Barker and T. R. Carrington,]. Chem. Soc., 3588 (1953). (109) K. V. Grii, A. Nagabhushanam, V. K. Nigam, and B. Bellavadi, Science, 121, 898 (1955). (110)S.C. Pan, L. W. Nicholson, and P. Kolacho,J. Amer. Chem. Soc., 73,2547(1951). (111)A. Yasumura, Seikagaku, 26,200(1954). (112)G. Avigad, Biochem.J.,73,587(1959). (113) G. R.Wyatt and G. F. Kalf,J. Gen. Physiol., 40,833(1957). ., (1959). (114)R.Geigg, M. Huber, D . Weismann, and G. R. Wyatt, Acta T T O ~16,255 (115)D.R.Evans and V. G. Dethier,]. Inst. Physiol., 1,3(1957). (116)A. Maurizio,]. Inst. Physiol., 11,745(1965).

THE SUGARS OF HONEY

301

that a trehalose is a constituent of the blood sugar of all three types of bee, namely, workers, drones, and queens. a#-Trehalose was known only as a synthetic preparation until its presence, together ~ with the a,a isomer, was detected in saki and in koji e ~ t r a c t . ”These two sugars have also been reported to occur in Royal Jelly.”* a,aTrehalose has been synthesized enzymically by Peat and and P,P-trehalose has been isolated from an almond-emulsin digest of D-glUCOSe after incubation for 5 weeks. Theanderose, together with panose and other oligosaccharides, has been synthesized by the action of a cell-free extract of Aspergillus niger 152 on a mixture of sucrose and maltose. The analog erlose, namely 0-a-D-glucopyranosyl-(1+4)-a-~-glucopyranosyl p-D-fructofuranoside, was obtained by White and Maher1Ig by treating sucrose with honey invertase. Of the honey trisaccharides, 1-kestose [O-a-D-glucopyranosy1(1+2)-O-P-D-fructofuranosyl-( 1+ 2) p-~-fructofuranoside] has been the most widely investigated. It has been synthesized (sometimes accompanied by other kestoses) by the action of trans-D-fructosylases from various sources on sucrose: by the action of takadiastase,lZ0 yeast invertase120J22 and various fungal enzyme^.'^^-'^^ Barker and were the first to obtain it crystalline. The oligosaccharide has also been prepared by the action of enzymes from sugar-beet leaves and from other higher plants.lZ6 It has been identified as a component of the soft xylem of aspen and of the oligosaccharides of maple sap.lZ8 The absence of appreciable proportions of melezitose from nectar honeys suggests that it is not formed during elaboration of the nectar in plants or in the ripening of honey by the bees; hence, it is logical to assume that the small proportions of this sugar present in floral honey may have originated as a result of interference by aphids. On the other hand, honeydew honeys contain a high proportion of this sugar, but

(117) K. Matsuda, TohokuJ . Agr. Res., 6,271 (1956). (118) I. R. Siddiqui and B. Furgala,]. Apicult. Res., 4,89 (1965). (119) J. W. White, Jr.. and J. Maher,].Amer. Chem. SOC.,75,1259 (1953). (120) J. S . D. Bacon and D. J . Bell,]. Chem. Soc., 2528 (1953). (121) J. S. D. Bacon, Biochem.]., 57,320 (1954). (122) D. Gross, P. H. Blanchard, and D. J. Bell,]. Chem. SOC.,1727 (1954). (123) J. H. Pazur,j. Biol. Chem., 199,217 (1952). (124) S. A. Barker and T. R. Carrington,]. Chem. SOC., 3588 (1953). (125) S. A. Barker, E. J. Bourne, and T. R. Carrington,]. Chem. SOC.,2125 (1954). (126) P. J. Allen and J. S. D. Bacon, Biochem.]., 63,200 (1956). (127) J. B. Pridham, Biochem. J . , 76, 13 (1960). (128) S. Haq and G. A. Adams, Can.]. Chem., 39,1165 (1961).

302

I. R. SIDDIQUI

there is a divergence of views as to its origin. Hudson129considered that melezitose is a constituent of the sap of various species of plants, whereas others believe that insects produce it. Bacon and D i ~ k i n s o n ~ ~ have discussed, in considerable detail, the origin of melezitose and erlose in honeydew honeys, and have shown that these sugars originate by the interaction of certain enzymes from aphids with plantsap sucrose. The trisaccharide O-cY-D-glucopyranosy1-(1+6)-O-a-D-glucopyranoSyl-(1- 3)-D-glUCOpyranOSehas been obtained by the action of potato T-enzyme on nigerose. The same trisaccharide is probably also produced, among others, by the transfer of D-glucopyranosyl groups from maltose to D-glucose in the presence of enzymes from Aspergillus oryzae.lW Other examples and explanations of such reactions are to be found ~ ~ Jother ~ ~ examples of the biosynin review^^^^^^^' and b o ~ k s . ’ Still thesis of some of these oligosaccharides, including isopanose, are those brought about by bacterial and algal As regards the occurrence of p-D-linked disaccharides in the absence of a p-D-linked substrate, one is tempted to conclude that these oligosaccharides are synthesized by the enzymic reversion of D-glucose by a P-D-glucosidase, As White and Maherloohave found that their honey-invertase preparation had no /3-D-glucosidase activity, it would appear that these sugars are carried into the hive as constituents of nectar. In vitro syntheses with enzymes almost invariably result in poor yields, but use of such syntheses is essential when chemical syntheses have not yet proved feasible. Another advantage of enzymic processes lies in their similarity to the processes occurring in Nature. It is implicit that similar enzymes or similar enzymic reactions are operative in the formation of honey, but, unfortunately, our knowl(129) C. S. Hudson, Adoan. Carbohyd. Chem.,2,1(1946). (130) S . A. Barker and E. J . Bourne, Quart. Reo. (London), 1,56 (1953). (131) R. A. Dedonder, Ann. Reo. Biochem., 30,347 (1961). (132) R. W. Bailey, “Oligosaccharides,” Macmillan, New York, N.Y., 1965. (133) J. Stanek, M. Cerny, and J. Paclk, “The Oligosaccharides,” Academic Press Inc., New York, N.Y., 1965. (134) M . Killey, R. Dimler, and J . E. Huskey,]. Amer. Chem. Soc., 77,3315 (1955). (135) S. A. Barker, E. J . Bourne, P. M. Grant, and M . Stacey, Nature, 178,1221 (1956). (136) D. S. Feingold, G. Avigad, and S. Hestrin, Biochem.]., 64,351 (1956). (137) K. Aso, K. Shibasaki, and M. Nakamura, Nature, 182,1303 (1958). (138) E. J . Bourne, D. H. Hutson, and H. Weigel, Biochem.].,79,549(1961). (139) J. H. Pazur and T. AndoJ. Biol. Chem.,235,297 (1960). (140) W. A. M. Duncan and D. J. Manners, Biochem.]., 69,343 (1958).

THE SUGARS OF HONEY

303

edge of the enzyme chemistry of honey is still inadequate. Honey has been shown to contain an acid phosphatase and a phosphorylase141J42; and an enzyme (D-glucose oxidase) that produces acid has been reported. However, the most important enzymes in honey are invertase and diastase. Invertase brings about the inversion of sucrose in nectar, and has been recognized to be an a-D-glucosidase: the end products it produces from sucrose differ from those formed by yeast invertase. Enzymic synthesis from action of invertase on SUcrose gives rise to six oligosaccharides, five of which differ from those produced by yeast invertase; the structure of only one of these has This invertase preparation appears to be as yet been a mixture of enzymes that contains a little D-fructosylase and maltase activity. Under the circumstances, the most favored reaction with honey invertase would be resynthesis of oligosaccharides from sucrose by transfer of D-glucosyl and D-fructosyl groups. In this connection, it should be emphasized that, although the honey invertase would be expected to be specific with regard to the donor molecule and the resulting glycosidic linkage, its specificity towards the acceptor molecule will vary considerably; hence, further experimentation, with other acceptor oligosaccharides, should lead to very interesting results. It is highly probable that theanderose would be the product of one such reaction. Further fractionation of crude a-D-glucosidase of honey by ionexchange chromatography, gel filtration, and starch-gel electrophoresis has shown that it contains 7-18 components (isozymes) having143 a molecular weight of -51,000. The origin and function of diastase are little underin honey, supposedly arising chiefly from the bee,144 stood. The same is true of phosphorylase. Phosphorylases, which transfer the D-glucopyranosyl group from D-glucopyranosyl phosphate to a monosaccharide acceptor (and thus give rise to a disaccharide), have been found in many bacteria, and several disaccharides have The mode of synthesis of oligobeen synthesized by their use.145-148 saccharides in honey is subject to further complications in view of the presence of osmophilic (sugar-tolerant) yeasts, said to become active (141) W. Zalewski, Pszczel. Zeszyty Nauk., 9, l(1965). (142) F. Gunther and 0.Burckhart, Deut. Lebensm. Rundschau, 63,41(1967). (143) J. W. White, Jr., and I . Kushnir,]. Apicult. Res., 6,69 (1967). (144) R. Ammon, Biochem. Z., 319,295 (1949). (145) W. Z. Hassid and M. Doudoroff, Aduan. Enzymol., 10,123 (1950). (146) C. J. Sih, N. M . Nelson, and R. H. McBee, Science, 126,1116 (1957). (147) E. W. Putman, C. Fithing-Litt, and W. Z. Hassid,]. Amer. Chem. Soc., 77,4351 ( 1955). (148) Z. Selinger and M. Schramm,/. Biol. Chem., 236,2183 (1961).

304

I. R. SIDDIQUI

with an increase in the moisture content and the granulation of the honey. White has discussed these aspects, and has made recommendations for arresting or minimizing these effects. Of the osmophilic yeasts, four have been shown to be Schwanniomyces occidentilis, Saccharomyces torulosus, Saccharomyces bisporus, and Zygosaccharomyces japonicus; Nematospora ashbya gossypii has been tentatively identified, and a new species is reported that needs further i d e n t i f i ~ a t i o n . ' ~ ~ Logical questions in connection with the origin of oligosaccharides in honey are the composition of the raw material (or nectar) with respect to these constituents, and the extent of alterations in the composition caused by the nectar enzymes and those incorporated by the bees. Occurrence of both trans-D-glucosylation and trans-Dfructosylation has been established, the former in the nectar from flowers of Robinia pseudoacacia, and the latter, in the extra-floral nectar of Impatiens but very little is known concerning the nature and composition of sugars present in nectars, and the work has so far been of a qualitative or routine quantitative nature.100*'s2-15S The principal reason for this deficiency is the difficulty encountered in obtaining sufficient quantities of the material. For this reason, information is only available on the three major sugar constituents of nectars, namely, D-glucose, D-fructose, and sucrose (and their relative proportion, in some sources). Indications have also been obtained that small proportions of other sugars are present, but these have not yet been properly identified. Maurizio has reviewed the results of some of these investigations in It is apparent that our knowledge of the oligosaccharide composition of nectars is elementary, and the same is true of the changes caused by the bee enzymes. Some information regarding these effects is based on paper-chromatographic analysis of mixtures of sugars obtained from in vitro experiments (149) S. Aoyagi and C. Oryu, Tamagawa Daigaku Nogakubu Kenkyu Hokoku, 8, No. 7, 203 ( 1968). (150) M. H. Zimmermann, Ber. Schweiz. Botan. Ges., 63,402 (1953). (151) M. H. Zimmermann, Experientia, 10,145 (1954). (152) R. Beutler, Z. Vergleich. Physiol., 12,72 (1930). (153) A. Maurizio, Ann. Abeille, 4,291 (1959). (154) M. S. Percival, New Phytologist, 60,235 (1961). (155) B. Furgala, T. A. Gochnauer, and F. G. Holdaway, Bee World, 39,203 (1958). (156) D. Wanic and L. Mostowska, Zesz. Nauk. Wyzszej Szkoly Rolniczej Olsztynie, 17,54311964). (157) E. A. Mikhailova, Zzu. Tomsk. Otd. Vses. Botan. Obshchestua, 5,121 (1864). (158) A. G. R. Nair, S. Nagarajan, and S. S. Subramanian, Cum. Sci., 33, 401 (1964). (159) A. Maurizio, Bee World, 43,66 (1962).

THE SUGARS OF HONEY

305

using crude extracts from the hypopharyngeal glands and the mid-gut of the honey bee; these are the organs where sugar-inverting enzymes have been found.160The effects of these preparations on sucrose, maltose, a,a-trehalose, and other disaccharides have been examined and discussed by M a ~ r i z i o . ' ~ It ~ -is, ' ~ however, ~ difficult to draw any definite conclusions, except that enzymic changes occur and that these sugars are not only hydrolyzed but also give rise to higher oligosaccharides. It was found that the enzymes from the two sources, the hypopharyngeal gland and the mid-gut, differ in their speed of action on these substrates, and produce different end-products. Differences in the ease with which summer and winter bees inverted sugars were noted, and their age, nutrition, physiological condition, and race were found to affect the efficiency of the enzymes. Goldschmidt and BurkerP have attempted to explain the origin of a number of the sugars in the various honeys from the results obtained by paper chromatography of nectar honey, honeydew honey, and the honey derived by feeding sucrose syrup to bees. It was concluded that such sugars as 1-kestose and melezitose, which occur in honeydew honey in considerable proportions (but not to an appreciable extent in nectar honey), are apparently carried to the hive from the plant. On the other hand, such sugars as maltose, erlose, isomaltose, and isomaltotriose are formed as enzymic, secondary products formed by trans-D-glucosylation in the body of the bee. The complexity of the operations involved in the formation of honey is indicated by the fact that those processes that start in the nectaries (the plant glands that secrete nectar) continue in the secreted nectar, are modified in the bee stomach, and continue in the hive during ripening and then in storage. White and coworkers164have shown that chemical changes in composition and biochemical activity occur, even when honey is stored at 26 k3". During two years of such storage, there was a conversion of about 9% per year of monosaccharides into oligosaccharides. Free D-glucose disappeared more rapidly than free D-fructose, so the D-fructose/D-glucose ratio increased markedly. It may therefore be concluded that the biochemical origin of oligosaccharides in honey is still at the speculative stage. Thorough knowledge of nectar composition, including sugar components and enzymes involved, and in vitro studies using nectar, bee, and honey (160) E. Kralky, 2. Wiss. Zool. Abt. A, 139,120 (1931). (161) A. Maurizio, Ann. Abeille, 5,215 (1962). (162) A. Maurizio, Ann. Abeille, 8,113 (1965). (163) A. Maurizio, Ann. Abeille, 8,167 (1965). (164) J. W. White, Jr., M. L.Riethof, and I. Kushner, J . Food Sci., 26, 63 (1961).

306

I. R. SIDDIQUI

enzymes under properly controlled conditions are needed for explaining on a more realistic basis the enzymic build-up of oligosaccharides.

IV. HONEYPOLYSACCHARIDES Despite the voluminous literature on honey, its colloidal constituents have remained essentially uninvestigated. The gummy material from the filters of honey-processing plants contains a large proportion of a polymer associated with undesirable turbidity, caramelization, and crystallization tendencies of honey. The presence of such constituents has been recognized for many years; they have been reported to consist of gummy and colloidal aggregates of proteins, waxes, and polysaccharides, and have been thought to affect the properties of honey rather adversely. Their removal has been shown to result in the elimination of turbidity and the production of brilliant clarity in honey. After removal of these constituents, which amount to 0.2-1% of light to dark honeys, the caramelization and crystallization tendency of honey is also greatly l e ~ s e n e d . ' ~ A ~-'~~ more definitive report on the nature of these colloidal constituents appeared'68 in 1953. On the basis of electrophoretic and sedimentation analysis of buckwheat honey, the presence of three polymeric constituents was claimed; the two major components were proteins, and the minor component appeared to be a polysaccharide. was made, by the Following these reports, a detailed author, of the polysaccharide isolated from honey collected at the Central Experimental Farm, Ottawa. The crude polymer constituted 0.2% of the honey, and provided two polysaccharide fractions that amounted to only about 0.002%of the honey. On hydrolysis, the impure fraction gave six monosaccharide components, and, on this basis, it appeared to be a mixture of polysaccharides. However, the purified fraction yielded only three sugars, namely, D-mannose, Larabinose, and D-galactose (ratios 1.0:2.04:4.04) and appeared to be homogeneous on the basis of boundary electrophoresis and sedimentation analysis. Methylation and hydrolysis of the methylated product afforded seven methylated fragments which were fully characterized. From the methylation data, it was concluded that an average unit of the polysaccharide is made up of 22-23 sugar residues consisting of (165)R.E.Lothrope and H. S. Paine, Amer. BeeJ.,71,28 (1931). (166)H.E.Lothrope and H. S. Paine, Amer. Bee]., 72,444 (1932). (167)H.S.Paine and R. E. Lothrope, Amer. Bee]., 73,53(1933). (168)T.C.Helvey, Food Res., 18,197(1953). (169)I. R.Siddiqui, Can.J.Chem., 43,421(1965).

THE SUGARS OF HONEY

307

6 terminal, nonreducing end-groups comprising 3 residues of Dmannose, 2 of D-galactose, and 1 of L-arabinose; 5 to 6 residues of Dgalactose are involved in branching at 0-3 and 0-6. The remaining 11nonterminal units consist of six (1+6)-linked D-galactose residues, and one (1+3)-linked and four (1-5)- or (1+4)-linked (or both) Larabinose residues. Although the molecular weight of the polysaccharide was not determined, its rapid rate of diffusion on ultracentrifugation, and the fact that synthetic boundary-cells had to be used, indicated that the molecular weight was less than 10,000. The molecular weight of buckwheat-honey polysaccharide is 9,000. It was not possible to formulate a unique structure for the polysaccharide from the data obtained. Further studies, including fragmentation analysis, are needed for determination of its detailed structure and its triheteropolymer nature. Other polysaccharides were undoubtedly present, but their isolation and fractionation in quantities sufficient to permit detailed studies have been major obstacles to their structural elucidation. Regarding the origin of these polysaccharides, no information is available. However, there are several ways in which these compounds could originate. They could, for example, be components of the pollen grains present in honey, or, alternatively, could result from the fermentation of honey by sugar-tolerant yeasts. An interesting speculation is that these polysaccharides may be products of the detoxifying agents or enzymes in the bee, since Dgalactose, D-mannose, and L-arabinose are reputed to be bee poisons. V. HONEYDEW The importance of honeydew as raw material for honey is increasing in many European countries as a consequence of changed agricultural practices which have imposed restrictions on bee pastures. Both kinds of raw material, nectar and honeydew, have a common origin in the phloem saps of plants, but the two reach the bees through different routes; whereas nectar is drawn from flower nectaries, honeydew passes through insects, and, in the process, undergoes characteristic and noteworthy changes that result in the production of the so called honeydew honey or forest honey. It is appropriate to mention that, in certain regions of central Europe, honeydew honey is more highly prized than floral honey, although, on the North American continent, it is, from the esthetic point of view, considered inferior to floral honey. The insects associated with the formation of honeydew honey are the homopterous insects, such as

I. R. SIDDIQUI

308

plant lice (aphids) and scale insects that thrive on various parts of the host plant. In the past, there has been a divergence of opinion concerning the origin of honeydews (and mannas); however, the prevalent view is that insects are always responsible for formation of h ~ n e y d e w . "Such ~ insects pierce the plant tissues and suck the sap into their alimentary tract and, after modification by aphid enzymes, the unwanted constituents of the sap are excreted in the form of droplets that collect on the leaves and other parts of the plants. The plants associated with these deposits are such trees as ash, beech, cedar, elm, fir, hickory, linden, maple, oak, poplar, spruce, tulip, willow, and fruit trees.30 Bees collect the mannas or honeydews from these sources when nectar is not readily available, and process it to a product which, as already mentioned, is darker in color, lower in D-glucose and D-fructose, and higher in pH, oligosaccharides, total free acid, ash, and nitrogen than the original honeydew. The average composition of honeydew honey with respect to carbohydrates and moisture, based on 14 samples, is given in Table IV. TABLEIV Average Composition of Honeydew Honeys and Range of Values for 14 Samples, According to White30 Component, % Moisture D-Fructose DGlucose Sucrose Maltose Higher sugars Undetermined

Average

Standard deviation

16.3 31.80 26.08 0.80 8.80 4.70 10.1

1.74 4.16 3.04 0.22 2.51 1.01 4.91

Range

12.2 -18.2 23.91-38.12 19.23-31.86 0.44- 1.14 5.11-12.48 1.28-11.5 2.7 -22.4

A characteristic feature of honeydew honey is the presence of the trisaccharides rnelezitose and erlose; the presence of the former was recognized35 as early as 1918, and the latter was found"' in 1954. The mode of formation of these trisaccharides, as pointed out earlier (see p. 302), has been discussed at length by Bacon and Di~kinson.~' However, for the present, it suffices to state that, based on the fore(170)G. Tanret, Bull. SOC. Chim. Fr., 27,56(1920). (171)H.E.Gray and G.Frenkel, Science, 118,304(1954).

THE SUGARS OF HONEY

309

going evidence, White30 believed that there are at least two types of honeydew honey; namely, the melezitose type, which may granulate rapidly (frequently in the comb itself), and the erlose type, which does not granulate. To this list must be added yet a different type of honeydew honey; this supposedly arises from the phloem sap secreted from wounds on a plant, without passage through the body of insects; according to some beekeepers, bees often collect such sap.

This Page Intentionally Left Blank

REACTIONS OF FREE SUGARS WITH AQUEOUS AMMONIA* BY M. J. KORT Deparfment of Chemistry, University of Natal, Pietermaritzburg, and Sugar Milling Research Institute, Durban, South Africa

I. Introduction.. ..........................................................

11. Products Obtained.. .................................................... 111. Isolation of Products, and Proportions Obtained. .........................

IV. Mechanism.. ........................................................... 1. Lobry de Bruyn-Alberda van Ekenstein Transformation ................. 2. Formation of Glycosylamines and Aminodeoxy Sugars . . . . . . . . . . . . . . . . . . 3. Fragmentation Mechanism ............................................ 4. Saccharinic Acids. ........................... ..................... 5. Reaction of Dicarbonyl Compounds with Ammo ..................... 6. Fission of Reducing Sugars in Alkaline Solution.. ...................... 7. Imidazole Formation.. ............................................... 8. Recombination of Sugar Fractions.. ................................... V. Applications ................................... .....................

311 312 328 332 332 333 340 341 344 345 347 348 349

I. INTRODUCTION It is well known that sugars react with aqueous ammonia to produce heterocyclic compounds in low yield. The products of the reaction of the mono- and di-saccharides with concentrated, aqueous solutions of ammonia are dependent on three factors: (I) the length of time during which the reaction proceeds, (2) the temperature of the reaction, and (3) the catalyst used. This matter has been briefly but comprehensively reviewed.' The present article is more detailed; it covers the literature to the end of July 1970, and has been prepared for the use of chemists who may not be specialists in the carbohydrate field, as well as of those who are. * The author thanks Professor D. A. Sutton for suggesting that this subject be re-

viewed. (1) M. R. Grimmett, Reti. Pure Appl. Chern., 15,101(1965).

311

TABLEI Reaction of Ammonia with Various Sugars at Low Temperature (Short Reaction Time, with No Catalyst) Reaction

Sugar

D-GluCOSe ( 11

Temp. (depees)

Time (hours)

37

-

D-arabinose (2)

37

48

-

-

Dpsicose (5) di-mglucopyranosylamine (6) 2-amino-2deoxy-m glucose (8)

38

mFructose (3)

38

48

48

References

Products dark, high polymer

D-fructose (3)

imidazoles

Dfructose (3)

Dglucopy8 imidazole ranosylamine (7) compounds

D-mannose (4)

2

mmannose (4)

3 4

5 5

2-amino-2deoxy-D-

glucose (8) Lactose (4-0-8-Dgalactopyranosy l-D glucopyranose)

37

48

=galactose (9)

lactulose

Dlyxose (10)

resinous matter

s

?

Dtagatose (11)

3

R

g

4

TABLEI (continued) Maltose

37

48

D-arabinose (2)

D-fructose (3)

(4-0-a-DglucopyranOSyl-D-glUCOpyranose)

maltulose (4-0-a-nglucopyran-

4-0-a-~glucopyranOSYl-Dpsicose

Dglucose (1)

imidazoles

3

>

n

2

D-mannose (4)

5

OSyl-D-

Melibiose (6-0-w~galactopyranosyl-Dglucopyranose)

37

48

fructose) 6-O-cx-~galactopyranosyl-p-Dmannopyranose pentosecontaining disaccharide

w

m

-l

w Dgalactose (9)

imidazoles

Dtagatose (11)

melibiulose (6-0-a-~galactopyranosyl-Dfructose)

4(5)-methylimidazole

6

m m

2Q

* ZI

3 X

>

0 C

m

0

5

(2) L. Hough, J. K. N. Jones, and E. L. Richards, Chem. Ind. (London),545 (1954). (3)L. Hough, J. K. N. Jones, and E. L. Richards,J. Chem. SOC.,2005 (1953). (4) M. KGmoto, Nippon Nogei Kagaku Kaishi, 36, 305, 310 (1962);Chem. Abstracts, 59,15361d,f (1963). (5) K. Heyns and W. Koch, 2. Natulforsch., B , 7,486 (1952). (6) L. Hough, J. K. N. Jones, and E. L. Richards,/. Chem. SOC.,295 (1954).

?i 5

M. J. KORT

314

11. PRODUCTSOBTAINED When sugars are treated with aqueous ammonia for a short time at low temperature in the absence of a catalyst, the reaction is arrested before heterocyclic compounds can be formed in appreciable proportion, and the products are mainly epimerization products of the sugars, probably formed by way of their 2,3-enediols. These epimerization products are summarized in Table I which shows the reactions of D-glUCOSe, D-fructose, lactose, maltose, and melibiose with aqueous ammonia for a short time at low temperature. A dark-colored, high polymer is also formed in some instances (the browning reaction). In the ammoniacal solution, the monosaccharides are epimerized; the disaccharides are epimerized and, in addition, may be hydrolyzed to monosaccharides that can also be epimerized; hence, the variety of products obtained may be considerable. With a prolonged reaction-time, the reaction is more complex and passes beyond the epimerization stage. Many products are formed; from a single sugar, Hough and coworkers' obtained at least 15 components, and K6moto8 separated 8 compounds by paper chromatography. Many substituted imidazoles and pyrazines have been isolated and identified. Compounds isolated by various workers are shown in Table I1 (which includes compounds 12 through 58). The products H

c=o

I HCOH

I HOC H I HCOH I HCOH I C &OH (1)

H

c=o HOCH I

HOCH I

HCOH I HCOH I C &OH (4)

H

c=o I

HOCH I HCOH I

HCOH I C H,OH (2)

CH,OH I

CH,OH

I

c=o I

HOC H I

HCOH I HCOH I CH,OH (3)

7

c=o I

HCOH I HCOH

HCOH

HCOH

HCO

I

I

CH,OH

(5)

(6)

(7) L. Hough, J. K.N. Jones, and E. L. RichardsJ. Chem. Sac., 3854 (1952). (8) M. KiSmoto, Nippon Nogei Kagaku Kaishi, 36, 407 (1962); Chem. Abstracts, 61, 12227g(1964).

REACTIONS OF FREE SUGARS WITH AQUEOUS AMMONIA H

H

c=o

c=o

I

I

HCNH, I HOCH

HOCH I

HCOH I

HOCH

I

HCOH

I

HCOH I HCOH

I

HCO I

315

HOCH I HCOH I

I

CH,OH

CH,OH

CQOH

(7)

(8)

(9)

C€&.OH I

H

c=o

c=o

I HOCH I HOCH I HCOH I CH,OH

I

HOCH I HOCH I HCOH I CH,OH 111)

(10)

TABLE I1

Compounds Isolated from the Reaction of Free Sugars with Aqueous Ammonia Cornpound number

Compound name Miscellaneous "Galactosimine," C,HlaNO,.C,H ,N,O,.4 H,O-Zn(OH), Nitrogen-containing disaccharide, CiJL&"io(V Oxalic acid D-Xylosylamine

12

13 14 15

Imidazoles

16 (continued)

M. J. KORT

316

TABLEI1 (continued) Compound number

Compound name SubstituentsO at

17 18

C-2 CHSCOH

c-4 Me R

c-5 H 2-acetyl-4(5)-methylH 4(5)-(~-erythro-2,3-dihydroxybuty1)-

OH -CHzC-CH,OH H H Me H Me Me Me H H H Et H H Et Et Me H -CHzCHzOH H H H H -CH,OH -CH,OH H H Me H -CHzOH R' H -CH,OH H

4(5)-(~-glycero-2,3-dihydroxy-

30 31 32 33 34 35 36 37

H H H Me H H H H

H -COzH -CONH, H Me R" R' R"'

H H H H H H H H

Pr0PYl)2,4(5)-dimethyl4,5-dimethyl2-ethyl4(5)-ethyl4(5)-ethyl-5(4)-methyl4(5)-(2-hydroxyethyl)2-(hydroxymethy1)4(5)-(hydroxymethyl)2-(hydroxymethyl)-4(5)-methyl2-(hydroxymethyl-4(5)-(D-lyxotetrahydroxybuty1)4(5)-imidazole -4(5)-carboxylic acid -4(5)-formamide 2-methyl4(5)-methyl4(5)-(~-arabino-tetrahydroxybutyl)4(5)-(~-lyxo-tetrahydroxybutyl)4(5)-(~-erythro-2,3,4-trihydroxy-

38 39 40 41

H H H Me

R"" R""' R""" Me

H H H Me

19

20 21 22 23 24

25 26 27 28 29

buty1)-

4(5)-(D-erytho-trihydroxypropyl)4(5)-(~-erythro-trihydroxypropyl)4(5)-(~-threo-trihydroxypropyl)2,4,5-trimethyl-

Pyrazines Substituentsaat c-2

42 43 44 45

R" Me Me Me

C-3 H Me H H

C-5

R" H Me H

C-6 H H H Me

2,5-bis(~-arabino-tetrahydroxybutyl)2,S-dimethyl2,5-dimethyl2,6-dimethyl(continued)

REACTIONS OF FREE SUGARS WITH AQUEOUS AMMONIA

317

TABLEI1 (continued) Compound number

Compound name

46 47 48 49

-CH,OH -CH,OH Me Me

H H H H

H H R”

H H H H

50

Me

H

H

R“

51

Me

R“”

R”

H

52 53

H R”

H H

H R”’

H H

Me

2-(hydroxymethy1)2-(hydroxymethyl)-5-methyl2-methyl2-methyl-5-(~-arubino-tetrahydroxybuty1)2-methy~-6-(~-arabino-tetrahydroxybuty1)2-methyl-5-(~-arabino-tetrahydroxybutyl)3(D-erythro-trihydroxypropyl) pyrazine 2-(~-arabino-tetrahydroxybuty~)-5-(D-

erythro-2,3,4-trihydroxybutyl)-

$x H

H

Piperazines

Substituents at

c-2 54

Me

C-3 H

C-5 Me

C-6 H 2,5-dimethyl-

Pyridines

55 56 57

Pyridine 2-Methylpyridine (a-picoline) C,H,NO probably a (hydroxymethy1)pyridine C,H,,NO, probably 4-acetyl-5,lidihydro-3-methyl-2-pyridinol

58

H H

H H OH

H OHOH

“where R = -CH,-C-C-CH,; R’ = -C-C-C--CH,OH; R“ =-C-C-C--CH,OH;

OHH H OHOH OHOHH OHOH OHOH H H R”‘ = -CH,-C--C-CH,OH; R”“ = -C-C-CH,OH; R””‘ = -C-C-CH,OH; and H H H H OHOH H OH R””” = -C-C--CH,OH. OHH

M. J. KORT

318

obtained on prolonged reaction of sugars with aqueous ammonia at low temperature in the absence of a catalyst are summarized in Table 111. TABLE111 Reaction of Ammonia with Various Sugars at Low Temperature (Long Reaction Time with No Catalyst) Reaction sugar

Temp. Time Products Refer(degrees) (days) 1 6 7 8 13 15 27 30 34 35 42 49 50 51 NH, ences

2-Amino-2-deoxy- room D-glucose D-Fructose 20-23 room D-Fructose" room D-Glucose 37 20 20-23 D-Glucose* 20

0 0

9

0 0 0 0 0 0 0

10 11 11

0

180 42 60

0 0

0 0

60 14 16 42 7

0

7

0 0

12 10 13

00

0 0

0 0 0 0 0 0

0

0

0 0 0 (+ epimerization

products) D-Glucosylamine' D-Xylose

20 20

10 0 0 7

0 0

0

(+unidentified heterocycles)

12 13 ~

"Plus oxygen. bVapors of 25% aqueous ammonia. cWater present, but no ammonia.

At elevated temperatures, sugars react more quickly than at lower temperatures with aqueous ammonia, to give, mainly, substituted imidazoles and pyrazines, summarized in Table IV, and a catalyst is not needed. At the higher temperatures, there is an increase in the methyl-pyrazine and -imidazole fractions and a decrease in the (hydroxymethy1)imidazole and (tetrahydroxybuty1)pyrazinefractions, indicating occurrence of thermal cleavage of the side chains.1° (9) M. I. Taha,J . Chem. SOC.,2468 (1961). (10) I. Jezo and I. LuzAk, Chem. Zuesti, 20, 586 (1966); Chem. Abstracts, 65, 1866913 (1966). (11) j. Parrod, Ann. Chim. (Paris), 19, 205 (1933); Bull. Soc. Chim. Fr., 51, 1424 (1932); 53,196 (1933). (12) M. KGmoto, Nippon Nogei Kagaku Kaishi, 36, 403 (1962); Chem. Abstracts, 61, 12227d (1964). (13) M. S. Dudkin, N. G . Shkantova, and A. F. Yatsuk, Zh. Prikl. Khim., 41,385 (1968); Chem. Abstracts, 69,19441r (1968).

TABLEIV Reaction of Ammonia with Various Sugars at High Temperature with No Catalyst Reaction sugar D-Glucose

D-Ghcosylamine" Invert sugar Molassesb

Temp. Time Products (degrees) (hours) 1 3 6 7 8 13 20 21 27 34 35 42 44 45 46 47 48 49 50 51 52 54 55 56 NH3 References

100 100 100 100 120 120 120

Molasses' Sucrose

1 35 40 1 2 2

000 0 0 0 0

-

140 220 120-260 110

16 16 18 >5

0

0

000

0

0 0

2

-

0 0

0 0 0

0 0 0 0 0 0 0 0 0 0 0

0 0 0 0

0 0 0 (+lOimidazole spots) (+substituted pyrazines and imidazoles) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

12 14 14,15 12 16 17.18 19 19a

10 10

20 21

"Water present, but no ammonia. 60 lb. in.?. 'Molasses itself, no treatment with aqueous ammonia; the necessary ammonia must have arisen during the processing of the sugar beet (or cane). (14) P. Brandes and C. Stoehr,]. Prakt. Chem., 121,54,481(1896). (15)C . Tanret, Compt. Rend., 100, 1540 (1885);Bull. SOC. Chirn. Fr., 44, 102 (1885). (16)B. K.Davison and L. F. Wiggins, Chem. Ind. (London), 982 (1956). (17)L. F. Wiggins, Proc. Congr. Intern. SOC. Sugar-Cane Technologists, gth, India, 2,525(1956). (18)L.F.Wiggins and W. S . Wise, Intern. Sugar]., 57,435(1955). (19)L. F.Wiggins and W. S . Wise, Chem. Ind. (London), 656 (1955). (19a)1. Jezo, Listy Cukrooar., 82,300(1966);Chem.Abstracts, 66,66989~ (1967). (20)I. Jezo,Chem. Zuesti, 17,126(1963);Chem. Abstracts, 60,4139b(1964). (21)M. R.Grimmett, R. Hodges, and E. L. Richards, Aust.]. Chem., 21,505(1968).

9

0 C

m

0

c,

M. J. KORT

320

Use of a catalyst in the reaction of sugars with ammonia does not appear to alter the reaction qualitatively. The reaction still requires a longer time at room temperature (see Table V) than at higher temperatures (see Table VI), and the products of the reaction at low temperature and long reaction-time (see Table V) do not appear to TABLEV Reaction of Ammonia with Various Sugars for a Long Time, at Room Temperature in the Presence of a Catalyst Reaction Sugar

L-Arabinose D-Fructose

D-Galactose D-Glucose D-Glucose" Lactose Maltose D-Mannose L-Rhamnose* D-Sorbose Sucrose L-Xylose

Time (months)

Catalyst

Products Refer12 14 20 24 27 28 30 34 35 57 58 ences

1.5 4 1 2

2 2 4 1.5 1.5 1.5 4 3 6,18 6,18 4 6 12 4 24 4

0 0 0

0

0

0

0

? 11 0 11 0 23 0 23 0 22 0 24,25 0 23 0 0 26 0 23 0 23 0 23 0 0 23 0 0 0 0 0 27,28 0 23 no imidazole compound 23 0 23

0

0

22 23 11 11

OHOHH H H "Plus acetaldehyde. *L-Rhamnose = H , C - C - C - - C - C - - C = O . H H OHOH

(22) K. Inouye, Ber., 40,1890 (1907). (23) A. Windaus, Ber., 40,799 (1907). (24) A. Windaus and F. Knoop, Ber., 38,1166 (1905). (25) K. K. Koessler and M. T. HankeJ. Biol. Chem., 39,497 (1919). (26) A. Windaus, Ber., 39,3886 (1906). (27) A. Windaus and A. Ullrich, 2. Physiol. Chem.,92,276 (1914). (28) A. Windaus and W. Langenbeck, Ber., 55,3706 (1922).

TABLEVI Reaction of Ammonia with Various Sugars for a Short Time in the Presence of a Catalyst

sugar

Temp. Time (degrees) (hours) Catalyst

D-Arabinose” 95 0.5 L-Arabinosea 95 0.5 Cane and beet 200-275 18 molasses Cellobiose 200-275 18 Cellulose 200-275 18 “Dialdehyde” 200-275 18 of starch D-Fructose 60 4 1 D-Fructose“ 100 D-Galactose room 96 “Galactosimine” (12) 100 2 D-Glucose 40 %,48 42 120 60 6 100 4 D-Glucose” 40 24 60 1 plus 70 5 D-GlUCOSe‘ 40 24 Hydrolyzed starch 200-275 18 60 6 Invert sugar Lactose 200-275 18 L-Rhamnosed 44-45 130

Products

Refer-

12 18 19 20 21 23 24 25 27 28 30 34 37 38 39 40 43 44 45 48 54 ences

0

Cu(OAc), Cu(OAc), (NH&PO,

0

(NH&PO, (NH&PO, (NHJSPO,

0 0

0

0

0 0 0 0

0 0 0 0

0 0

0 0

0 0 0 0 0 0 0 0

0

0

29 29 30 30 30 30

v,

C

Q

Zn(OH), Cu(OAc), Zn(OH), 0

0

-

Zn(OH), Zn(OH), Zn(OH), Zn(OH), Zn(OH), ZnC03 Zn(OH), (NH,),PO, Zn(OH), (NHJ3P04 Zn(OH),

31 32 23 23 31 8,33-36 31 31 31 37

0

P

O

0

0

0 0 0 0 0 0

(+2unidentified)

0 0

0 0 0

0

0

0 0 0

0

0 0

0 0 0 0 0

0 0 0 0

0 0 0 0 (+ 2 unidentified)

31 30 31 30 38,39

(continued)

b

w

E

TABLEVI (continued)

0

E

Reaction Sugar

D-Ribose" Starch of various origins Sucrose D-XyIOSe"

Temp. Time (degrees) (hours) Catalyst

Products Refer12 18 19 20 21 23 24 25 27 28 30 34 37 38 39 40 43 44 45 48 54 ences

95 200-275

0.5 18

CU(OAC)~ (NH,),PO,

120-160

18

(NHJaPO, ZnO

160 95

6,183 0.5 CU(OAC)~

0 0

0

0

0 0 0 0 0

29 30

0 0

0

0

0 0 0 0

20

0 0

0

0

0 0 0 0

20 29

0

was obtained. "Plus acetaldehyde. aPlus formaldehyde. *A racemic mixture, namely 4(5)-(~~-glycero-2,3-dihydroxypropyl)imidazole, OHOHH H H Paper chromatograms of the products obtained in the reaction of L-rhamnose and L-fucose dL-Rhamnose= H3CC-C-C-C-C=O. H H OHOH (6-deoxy-~-galactose)with aqueous ammonia are identi~al.~* e(NH4)2Mo04,(NH4)3P04,NH,VO,, (NH,),WO,, Co,O,, Cu(NH,),SO,, CuO, NiO, Na2A102+ SOz,ZnC12,and ZnO. B. N. Ames, H. K. Mitchell, and M. B. Mitchel1,J. Amer. Chem. SOC., 75, 1015 (1953). I. Jeio and I. LuGk, C k m . Zuesti, 17,255 (1963);Chem. Abstracts, 60,4139e (1964). K. Bemhauer, Z . Physiol. Chem., 183,67 (1929). R. Weidenhagen, R. Hemnann, and H. Wegner, Ber., 70, 570 (1937). M. Emoto, Nippon Nogei Kagaku Kaishi, 36, 461 (1962); Chem. Abstracts, 60, 2332d (1964). M. Gmoto, Nippon Nogei Kagaku Kaishi, 36,464, 546 (1962); Chem. Abstracts, SO, 2332qh (1964). F. Fujii and M. Ksmoto, Nippon Nogei Kagaku Kaishi, 39, 114 (1965);Chem. Abstracts, 63, 14953e (1965);M. KGmoto, ibid., 36, 541 (1962); Chem. Abstracts, 60,2332g (1964); M. Gmoto, S. Fujii, and H. Tsuchida, Hyogo Noka Daigaku Kenkyu Hokoku, 5, 124 (1962); Chem. Abstracts, 60, 2249d (1964). S. Fujii, H. Tsuchida, and M. Gmoto, Agr. Biol. Chem. (Tokyo), 30, 73 (1966). R. W. Liggett and H. L. Hoffman, Jr. (to Atlas Chemical Industries, Inc.), U.S. Pat. 3,030,376 (1962); Chem. Abstracts, 57, 9859g (1962). H. Tsuchida and M. KGmoto, Agr. Biol. Chem. (Tokyo), 31, 185 (1967). M. KGmoto and H. Tsuchida, Agr. Biol. Chem. (Tokyo), 32,983 (1968).

5 ?

3

REACTIONS O F FREE SUGARS WITH AQUEOUS AMMONIA

323

TABLEVII Reaction of Ammonia with Various Sugars with Oxygen Bubbled Through the Solution at Room Temperature in the Presence of a Catalyst Reaction Sugar

L-Arabinose D-Fructose

Time (days) 30 15 30 30

60 60 D-Fructose" b b

D-Galactose D-Glucose

20 60 15 C

Invert sugaf' D-Mannose L-Rhamnose L-Xylose

15 30 30

Catalyst Cu(OH)* Cu(OH)* MethyleneBlue NH,HCO, + Cu(OH)* (=CuCO, + NH,) Fe2(S04)3, MnS04, FeSO, Ca(OH)* CUCO, cuco3 Cu(OAc), Cu(OH)* Cu(OH)* CU(OH)~ Cu(0H)z CUCO$ CU(OH)~ CU(OH)~ CU(OH)~

Products Refer14 27 29 30 31 32 33 34 35 36 38 ences 0 0 0

0 0 0

0 0 0

11 11,40,41 11,42 11,43

0 0

0

11 0

0

?(+at least 1 unidentified)

0 0 0 0 0 0

0 0 0 0 0

0 0 0

0 0

0 0 0

0 0 0

0 0

11 44

45 45,46 1 1,47 11 11,40 48 49 11,50 11 11

uPlus formaldehyde. *2.5hr at 100"."3years (at room temperature). d2hr on hot-water bath, plus 6-8hr at room temperature.

(40)J. Parrod, Compt. Rend., 192,1136 (1931);P. Girard and J. Parrod, Ann. Physiol. Physicochim. Biol.,7,295(1931). (41)P. Girard and J. Parrod, Compt. Rend., 190,328(1930). (42)J.Parrod, Compt.Rend., 195,285(1932). (43)J. Parrod and Y.Garreau, Compt. Rend., 195,1110(1932). (44)W.J. Darby, H. B. Lewis, and J. R. Totter,J. Amer. Chem. SOC., 64,463(1942). (45)J. R.Totter and W. J. Darby, Org. Syn., 24,64(1944). (46)R. I. Meltzer, A. D. Lewis, F. H. McMillan, J. D. Genzer, F. Leonard, and J. A. King,]. Amer. Pharm. Assoc., Sci. Ed., 42,594(1953). (47)Y.Garreau and J. Parrod, Compt. Rend., 194,657(1932). (48)A. Windaus and A. Ullrich, Z. Physiol. Chem.,90,366(1914). (49)L. P. Kulev and R. N. Gireva, Zh. Prikl. Khim., 30, 811 (1957);J . Appl. Chem. USSR, 30,858(1957). (50)J. Parrod and Y. Garreau, Compt. Rend., 193,890(1931).

w

TABLEVIII

to

A

Reaction of a-Dicarbonyl or a-Hydroxycarbonyl Compounds with Ammonia in the Presence of Formaldehyde Reaction Time

Temp. Reactant 2,SButanedione (CHa-CO-CO-CH,) 3-Deoxy-D-glyceropentosulose (59) 3,6-Dideoxy-~-erythrohexosulose (60) 1,4-Dihydroxy-2-butanone (CH,OH-CO-CH,-CH,OH) 1,3-Dihydroxy-2-propanone (CH ,OH-CO-C H 2 0 H ) Glycolaldehyde (CHXOH-CHO) Glyoxal (CHO-CHO) D-aruhino-Hexosulose (61) 2-Oxobutanal (CH:$-CH 2-C0-CHO) Pyruvaldehyde (CH:,-CO-CHO)

Products

(degrees) (hours)(months)

Catalyst

0 120 room

12 8 24

-

room

30

-

55 ?

0.5 ?

room 40 95 100

short ?

? 0 room

? 12 24

room room 40

100

6 2.4

6

24 2

23

Zn(OH), Zn(0Hh Zn(OH)2 Zn(OH).L

25

27 30 34

35 41

0 0

0 0

0

References 51 51,52 36 38

0

CU(OAC)~ cuco, Zn(OH), Zn(OH), C u (0Ac) Cu(OAc), Zn(OH),!

immediate

18 19 21

0 0

53 0 0

54

31 55 55

0 0

0

56

0 0

0 0 0 0

g

33

11,40 39 54 31 31 31

g

REACTIONS O F FREE SUGARS WITH AQUEOUS AMMONIA

325

H

H

c=o I c=o I FHP HCOH

I

CH,OH (59)

H

c=o I c=o

c=o c=o I

7H2 HOCH I HOCH I c H3 (60)

H

c=o I c=o

I HOCH I HCOH

HCOH

HCOH

HCOH

I

I

CH,OH

(61)

I

y

2

I I

C H,OH

(62)

differ significantly from those under similar conditions in the absence of a catalyst (see Table 111). Similarly, the products of the reaction at higher temperature and shorter reaction-times (see Table VI) do not differ significantly from those under the same conditions in the absence of a catalyst (see Table IV). In Table VII are listed further reactions of sugars with aqueous ammonia in the presence of a catalyst at room temperature, but, in these cases, oxygen was passed through the solutions. The imidazoles formed in the reaction of aqueous ammonia with other a-hydroxycarbonyl compounds, for example, the triose DLglyceraldehyde, and such a-dicarbonyl compounds as 3-deoxy-~glycero-pentosulose (59), and the 3,6-dideoxy-~-erythro-,D-arahino-, and 3-deoxy-~-erythro-hexosu~oses (60, 61, and 62), respectively, are summarized in Table VIII for reactions in which formaldehyde was added, and in Table IX for reactions in which it was not added. (51) R. G. Fargherand F. L. Pyman,]. Chem. Soc., 115,217 (1919). (52) A. Windaus, Eer., 42,758 (1909). (53) C. F. Huebner,J.Amer. Chem. Soc., 73,4667 (195 1). (54) B. J. Sjnllema and A. J. H. Kam, Rec. Trac. Chim., 36, 180 (1916). (55) R. Weidenhagen and R. Herrmann, Ber., 68,1953 (1935). (56) R. Behrend and J. Schmitz, Ann., 277,310 (1893).

TABLEIX

Reaction of a-Dicarbonyl or a-Hydroxycarbonyl Compounds with Ammonia in the Absence of Formaldehyde Reaction Time

Temp.

Catalyst Products

Reactant 2,3Butanedione

3-Deoxy-~glyceraldehyde OH CH&--CHO H 3Deoxy-D-eythrohexosulose (62) 1,3-Dihydroxy-2propanone DL-Glyceraldehyde (CHZOH-CHOH-CHO) Glycolaldehyde

(degrees) (hours)(weeks) andreagent cold plus 95 100

Refer14 16 17 20 22 23 24 25 26 27 28 30 33 34 37 41 ences

-

0

57

z

.

8

3

4

1 1

-

room

21

-

room 40

24

2

CU(OH)~,O~ 0 Zn(OH), +CH3CH0

37

8

-

room

8

-

0

0

0

0

39

0 0

0

0 0

0

0

11 31

Ob

58

‘I

0

0

21

59

Glyoxal

60-70 room 0 ?

Hydroxypymvaldehyde (CH 2 0 H-CO-CHO) Pymvaldehyde

room room

-

48

48

Zn(OH)*+ CH,CHO NH,OAc HOAc (no NH,) + CHZOH-CHO

+

CHZOH-CHO

60 61 62 63

0 0

0

-

5

room

0 0

CH,CH,CHO

6 ?

0 0

CH,CHO (noNH,) CH,CHO.NH,

12

19 40

95

-

short ? 24 1

0 0

63 59

0 0

0 0

6 4 31

0

8

0

g +c,

2 3 g

’II

!a

m

m vl

c

’3 +0

0

58

“Possibly not found because of the small quantities used. bAlso identified: 1,3-dihydroxy-%propanone,DL-gluCOSe, DL-fructose, mannose, DL-arabinose, DL-lyxose, and DL-XylOSe (and, possibly, DL-ribose).

DL-

3:

s+ m 0 C vl

(57)H. von Pechmann, Ber., 21,1411(1888). (58)M.R. Grimmett and E. L. Richards, Aust.]. Chem., 17,1379(1964). (59) M. R. Grimmett and E. L. Richards, Aust.]. Chem., 18,1855(1965). (60) H. Debus, Ann., 107,199(1858). (61)G . Wyss, Ber., 10,1365(1877). (62)B. Radziszewski, Ber., 15,2706(1882). (63) B. Radziszewski, Ber., 16,487(1883). (64)M.R.Grimmettand E. L. Richards,]. Chem. SOC., 3751 (1965).

Ez

0

$ 0 E3

4

328

M. J . KORT

111. ISOLATION OF PRODUCTS, AND PROPORTIONS OBTAINED From the Tables of the products obtained from the reaction of various sugars with ammonia, it may be seen that the results often appear to be irreproducible from one worker to another; this is because, in some experiments, only certain products were being investigated or specifically synthesized. Often, other compounds were pres~ ' ~spraying -~~*~~*~~*~~ ent, as shown by paper c h r o m a t ~ g r a p h y , ~ - ~ ~ ~with with diazotized sulfanilic acid (the Pauly reagent),s5 with or without ammoniacal silver nitrate,66but all of them were not identified, and some could not be obtained crystalline." In the earlier work, too, the yields might have been too low for presence of the compounds to have been observed (before the advent of paper and thin-layer and thin-layer21.59*64 chrochromatography). Papel.8-'0.21.33-36,38,39,58.59,64 matography have been used as an aid in identifying the various products formed. Chromatography4 on ion-exchange resins applied to the reaction mixture, and cellulose-column chromatography of the brown syrup or after obtained after concentrating the reaction mixture,s*7,g,21.36*38,39 passing the reaction mixture through a column of an ion-exchange resin and then c o n ~ e n t r a t i n g ,have ~ ~ * ~been used for separating the components of reaction mixtures. Column chromatography on alumina . ~ ~ , ~the~ brown syrup was extracted has also been ~ ~ e dAlternatively, with ethefl*20*30 or ethanol,6' and, after removal of the solvent, the residue was distilled to yield the imidazoles8-6'or a pyrazine fraction and an imidazole f r a c t i ~ n . The ~~.~ still ~ residue8 could be separated into the constituent imidazoles by fractional recrystallization of their picrates. Fractional recrystallization of the picrates has also been pyrazine~,"-'~ -'~*~~*~~ used by other workers to isolate i m i d a ~ o l e s , ~ ~ ' ~ the reaction mixtures. The percentages of the and p i c ~ l i n e 'from ~ pyrazine and imidazole fractions obtained at various temperatures and with various proportions of catalyst^,^^*^^ and with different catalysts and times of reaction,20 have been given in detail. With sucrose,Po the overall yield of both fractions increased with time, indefinitely, 18 hours being the most convenient length of time. In the absence of a catalyst, 220" was the optimal temperature (16.8% of pyrazine fraction and 8.3% of imidazole fraction). The most effective catalysts were ammonium phosphate (0.625%in the reaction mixture gave the highest yields) and zinc oxide, with yields at the (65) H. Pauly,2. Physiol. Chem., 42,508(1904);44,159(1905). (66)S.M.Parbidge,l3iochem.J..42,238(1948).

REACTIONS OF FREE SUGARS WITH AQUEOUS AMMONIA

329

optimal temperature (200") of 12.60 and 10.20% of pyrazine and 7.20 and 2.20% of imidazole fractions, respectively. With wheat starch,305% was the optimal concentration of the ammonium phosphate catalyst; and the optimal temperature for beet and cane molasses, lactose, cellobiose, starch syrup, dextrin, and the "dialdehyde" of starch was 225", and for cellulose and wheat, corn, and potato starch, it was 250". The percentages of the individual imidazoles and pyrazines obtained from the imidazole and pyrazine fractions from the various carbohydrate^^^.^^ were given in detail. Only one example, sucrose, is shown in Table X, in which the quantities of the products obtained in some of the reactions detailed in Tables I and 111-IX are summarized. The yields of some of the products obtained are given in Table XI for those experiments in which the amounts of the insoluble copper or zinc salts of the imidazoles were determined. Brandes and StoehrI4 (see Table IV) added potassium hydroxide to each reaction mixture before distilling it, and added sodium hydroxide to the distillate, liberating an oil which was separated and purified, and from which pyridine and four pyrazines were separated by means of their mercury and gold salts. In the various experiments in which a zinc hydroxide catalyst was used (see Tables V, VI, VIII, and IX), the zinc salts of the imidazoles precipitated from the reaction mixtures, and, after they had been or in filtered off, they were suspended either in warm waters*1'*24,36--39 and the zinc was precipitated as sulfide by bubacetic bling hydrogen sulfide through the suspension. The zinc sulfide was filtered off, the clear filtrate was evaporated to a syrup, and the base was either (a) extracted directly into chloroform and the product distilled after removal of the chl0roform,3~or (b) extracted into acetone,2s into ether,22 or into ether after addition of potassium Removal of the ether yielded an oil from which the various imidazoles were separated by fractional recrystallization of their oxalates23-27~31*54 or p i ~ r a t e s . ~ Cellulose-column ~,~~,~' chromatography of the ether extract also separated the constituent imidazo1es.36,38,39 The free imidazoles could be liberated from their oxalates24,25*54 by adding potassium carbonate, extracting with ether, and evaporating the extract; and from their picrateP by adding sulfuric acid, filtering off the picric acid (which was completely removed from the solution by washing with ether), concentrating the solution, adding potassium carbonate, extracting with ether, evaporating the extract, and distilling the residue to give a pure product. Similarly, where copper salts were used as the catalyst (see Tables -27331s4

0 0 0

TABLEX Yields of the Products Obtained Reactant

16

%Amino-Zdeoxy&glucose L-Arabinose 2,3-Butanedione D-Fructose

20

21

27

28

Weight (g) of product, based on 100 g of reactant 30 31 34 39 41 42 44 45 48 3.0

49

50

51

54

55

1.05 3.67

9

5

29 13.2

19.6 25.528.2

D-Glucose

0.05 6.71

0.12

0.2 0.1

1.o

0.2

4 Glyoxal Invert sugar

References

1.5

2.75 16.423.7"

7

8

8

s

14 48 61 49

Pyruvaldehyde

75.8b

54 64

Sucrose

12.4 4.71

"Assulfate.bAs oxalate.

0.38 0.15 0.31

2.65 0.88 9.95

0.81

m

REACTIONS OF FREE SUGARS WITH AQUEOUS AMMONIA

331

TABLEXI

Yields of Products Obtained, Where the Amount of the Copper or Zinc Salt Was Determined Reactant L-Arabinose

Cu"

ReferWeight (9)of product, based on 100 g of reactant Zn" 1 4 20 27 29 30 33 34 35 36 ences

105

4.4

11 22 23 11

31

12.5 21 1,3-Dihydroxy- 150 2-propanone 1,3-Dihydroxy2-propanone* 155 D-Fructose

0.1

36

0.1

12

8 3

15

31 11

2.5

26 20 D-Galactose

38.9

77.8 180

0.4 15

5

0.5

10

2.8

24.2 D-Glucose

1.5

145

24

1

26 20 31

10 9.4c

40

D-Ghcose* Invert sugar Lactose Maltose D-Mannose

Pyruvaldeh ydeb L-Rhamnose 10 D-Sorbose L-Xylose

31

38.7 24 16 0.8,1.4 4.8,7.9 80

8

4

10

5

3

26 33.3 0.1 I .5

3

1.7 21

8.5 9.6c

1.9

33 32 26

142

11 23 31 32 11 22 23 11 23 24 26

7.3

37 31 31 23 23 11 23 31 11 23 27 23 11 23

"Cu = Insoluble copper salt of the imidazoles; Zn = insoluble zinc salt of the imidazoles. *Plusacetaldehyde. 'As oxalate.

V to IX), the copper complexes of the imidazoles also precipitated from the reaction mixture, and were filtered off. Oxalic acid and 4(5)-imidazole (30)were determined in the f~ltrate.".~~*~' The complex was suspended in hot water,11~29~32~40*4'~43~47*50~53~55 dilute sulfuric a ~ i d , ~or * ,dilute ~ ~ hydrochloric the copper was removed as the sulfide, with hydrogen sulfide or sodium sulfide,49and the excess of hydrogen sulfide was removed with lead a ~ e t a t e . " * ~The ~ * clear ~'~~~

332

M.J. KORT

solution obtained after filtering off the precipitated copper sulfide was further purified with decolorizing carbon32*49*53 and evaporated in oucuo to give the i m i d a z ~ l e , ~or~the * ~product ~ * ~ ~ in the residue was then extracted into c h l o r o f ~ r m ,or ~ ~the residue was distilled55 to afford the pure imidazole. In addition, ion-exchange chromatography has been used for isolating the i m i d a z ~ l e s .Alternatively, ~~ the imidazoles were isolated from the copper-free solution as the picrates,11~40~41*44~47*so which were separated ( a ) by fractional recrystallization, (b) as the p h o s p h o t u n g ~ t a t e ~ (after ~ * ~ ~removal , ~ ~ * ~ ~of ammonia by extraction with ether48),or (c) by concentration of the solution in the presence of barium hydroxide and removal of the barium ions with sulfuric The free base was liberated from the picrate by adding sulfuric acid, removing the picric acid liberated by filtration and ether extraction, neutralizing the sulfuric acid with barium carbonate, filtering, and concentrating the liquid to a small volume, whereupon the free base crystallized out on Alternatively, for isolating the free base, a mixture of a solution of the picrate with potassium carbonate was evaporated to dryness, and the free base was extracted into hot acetone, from which it crystallized on To obtain the free base from the phosphotungstate, the salt was suspended in water, barium hydroxide solution was added, and the suspension was filtered. After addition of carbon dioxide to remove the barium, and filtration, the filtrate was evaporated, and the residue was recrystallized from a l ~ o h o l " , [for ~ ~ ,4(5)~~ (hydroxymethy1)imidazole (27)] or from aqueous acetone4s [for imidazole-4(5)-carboxylicacid (31)l. Use has been made by Grimmett and R i c h a r d ~of~ a~ quantitative, ,~~ colorimetric method for determining imidazole derivative^.^^ IV. MECHANISM

1. Lobry de Bruyn- Alberda van Ekenstein Transformation

The products obtained by the reaction of sugars with aqueous ammonia for a short time at low temperature, in the absence of catalysts (see Table I) are simply those obtained by the action of alkali on the sugars. The reaction is known as the Lobry de BruynAlberda van Ekenstein transformations7 after the chemists who dis(67)c.A. Lobry de Bruyn, Rec. Trao. Chim., 14, 150 (1895);c.A. Lobry de Bruyn and W. Alberda van Ekenstein, ibid., 14, 195 (1895);15,92 (1896);16,241,245, 256, 264 (1897);18, 147 (1899);J. C. Speck, Jr., Aduan. Casbohyd. Chern., 13, 63 (1958).

REACTIONS OF FREE SUGARS WITH AQUEOUS AMMONIA

333

covered this rearrangement, which occurs when a reducing sugar is dissolved in water containing an alkaline catalyst; the transformation is illustrated in Scheme 1 for D-glucose. H C=Q

H cCOH

rl

11

H-COH

0

C-0-H I

I

HOCH

HOCH I

I

R

-

R

ChOH I

n

C=O 17 HOC-H I R

CH,OH I

COH II

HOC I

R

o-Fructose

o-Glucose

(3)

(1)

H CH,OH

C=O

I

c=o

HOCH

I HCOH I R

I

HOCH I

R

o-Peicose

D-Mannose

(5)

(4)

I HCOH I

where R = HCOH I

CH,OH

. Scheme I

The reaction of the sugars with ammonia was stopped before condensation with ammonia could occu13 and only the epimeric aldoses and the corresponding ketose were ~ b t a i n e dHowever, .~ on prolonged reaction, i m i d a z o l e ~ ~or- ~other ~ ~ condensation products4a5[namely, di-D-glucopyranosylamine (6), D-glucopyranosylamine (7), and 2amino-2-deoxy-D-glucose (8)] were also isolated (see Table I). 2. Formation of Glycosylamines and Aminodeoxy Sugars There are two theories as to the mechanism of the formation of imidazole and pyrazine compounds, namely, ( a ) the prior formation of glycosylamines and aminodeoxy sugars, and ( b )the fragmentation mechanism. The similarity of the products obtained from the mono- and the di-saccharides supports the hypothesis of Je5020*68 and Jeio and Luihk30 that the di- and poly-saccharides are first hydrolyzed to the monosaccharides. Although most glycosidic linkages are stable to (68)I.Jeio,Listy Cukroun~.,82,259(1966);Chem.Abstracts, 67,3201k(1967).

M. J. KORT

334 hydrolysis Sucrose

__ __t

+

o-Glucose

o-Fructose (3)

HO H NH, I l l R-C-C-C-H I

H

l

HO R-C-

l

OHOH D-

Amadori

H

I I

OH

Fructosylamine Heyns

(see Schemes 4 and 5)

HO

I I

NH, C-CH,OH

I

H

0-Glucosylamine

R-C-

I

HO I

H

H

l

l

R- C -C- C =O

C -CH,NH, I1 O

I

H

1 -Amino- l-deoxy-

1

NH,

2- Amino- 2- deoxy~-glucoiie

D- f ructose

HO I R-C-C=O

HN

‘CH-C-R

I

-+

2-Amino- 2-deoxy D-glucose

+

2 -Amino- 2-deoxy D- glucose

1- Amino- 1-deoxy o-fructose

+

2. dehydration

..

A

m

e

H

r ization

- H,O

-

F

H P OH

where R = -c-L-CH,OH. I I H H

Formation of Pyrazines

Scheme 2

I I OH

2- Amino- 2-deoxy o- glucose

the double bond

REACTIONS OF FREE SUGARS WITH AQUEOUS AMMONIA

335

alkali at 25- lOO", Lindberg and coworkers69investigated the alkaline hydrolysis of some glycosides at 170". Hydrolysis certainly occurred, although often to only a small extent. The reaction of sucrose with ammonia was conducted at a high temperature, and even should it only be hydrolyzed to a small extent, this would be sufficient to support the mechanism of J e i o and Lui&ka3O The hypothesisz0,68for the next step is the formation of 2-amino-2-deoxy-D-glucose from Dfructose and ammonia, without prior fragmentation of the sugar molecules.I0 Condensation of this product with a second molecule or with a molecule of l-amino-l-deoxyof 2-amino-2-deoxy-~-glucose D-fructose (similarly formed from the D-glucose produced and ammonia) would give the substituted pyrazines,20,68as illustrated in Scheme 2, and the substituted imidazoles,68as shown in Scheme 3. The Amadori rearrangemenPo of D-glucopyranosylamine to 1amino-1-deoxy-D-fructopyranose, and the Heyns rearrangement7' of D-fructopyranosylamine to 2-amino-2-deoxy-D-glucopyranose,are shown in Scheme 4. These mechanisms for the Amadori and Heyns rearrangements are, in effect, two alternatives for the same type of transformation, even though the former involves protonation of the nitrogen atom at C-1, and the latter, protonation of the oxygen atom at C-6. Indeed, the mechanism of the Amadori rearrangement has been explained on the basis of N - p r o t o n a t i ~ nand ~ ~ of O-protonat i ~ n Hodge70 . ~ ~ prefers the former. Although the mechanisms shown in Scheme 4 are acid-catalyzed, the Amadori rearrangement occurs also in alkaline media. The mechanisms proposed for the Amadori and Heyns rearrangements in alkaline solution are shown in Scheme 5. Jeio further proposedz0 that the methylpyrazines are formed by thermal detachment of the side chains (see Scheme 6); these hagments could then combine with ammonia to form the imidazoles. From Scheme 2, it may be seen that D-glucosylamine and D-fructosylamine are also considered to be intermediates in the formation of heterocyclic compounds. This intermediary formation of l-amino1-deoxy-D-fructose and D-glucosylamine from D-glucose and am-

(69)B. Lindberg, Suensk Papperstidn., 59, 531 (1956);E. Dryselius, B. Lindberg, and 0. Theander, Acta Chem. Scand., 11, 663 (1957);12, 340 (1958);J. Janson 14,2051(1960). and B. Lindberg, ibid., 13,138(1959); (70)J . E. Hodge, Adoan. Carbohyd. Chem., 10,169(1955). (71)K. Heyns, H. Paulsen, R. Eichstedt, and M. Rolle, Chem. Ber., 90, 2039 (1957); K. Heyns and K.-H. Meinecke, ibid., 86,1453(1953). (72)R. Kuhn and F.Weygand, Ber., 70,769(1937);F.Weygand, ibid., 73, 1259 (1940); H.S.Isbell, Ann. Reu. Biochem., 12,205(1943). (73)A. Gottschalk, Biochemj.,52,455(1952).

M. J. KORT

336

Sucrose

I

hydrolysis

INH,

D-Glucose

t

1

R----T

D-

2- Amino- 2-deoxy- D- glucose

H

H0 H H OH H/NHe\ I I LCH-C-C-R

I

n HC=O U .

H

I

I

Cf,:

I-\/> HI OH I cOH R-CH ,CH-C-C-R

1

OHH

HocH,-A-iiH I Ll HO H

HN? ' H

H

I

where R = -C-C-CH,OH 1 1 HO OH

Fructosylamine

i

rearrangement

1- Amino-1-deoxy-D-fructose

I

(see Scheme 2)

+

D-Glucosylamine

H

D-Fructose

,

Formation of Imidazoles Scheme 3

I

I

HO

H

"7 Hi[g

REACTIONS OF FREE SUGARS WITH AQUEOUS AMMONIA

337

Amadori Rearrangement

0

cfi".

CH H-COH f-1 I

+Ho

HCOH I

HOCH

HOCH

~

I

HCOH I HCO

-

HCOH

1

HCOH I HCOH

H&Oq I

I

CH,OH

1

CH,OH

CH,OH

a-D - Glucopyranosylamine

I

HYOH

HFOH

HToH J

HCOH/

I

HOCH I

HYOH HCOH

.. CH,Od-H I

C%O

1-Amino- 1-deoxy

-

I

CH,OH

- (Y

D-fructopyranose

Heyns Rearrangement

-

HO~H I I

I CH,O (Y-D-

CH,-0

i

HCOH I

HCOH I

CH,OH

H

/lH@

Fructopyranosylamine

HCNH, I HOYH

"PO

HCO I

CH20H

Z-Amino-2-deoxy-ao-glucopyranose

Scheme 4

HCO-H 10

CH,OH

M. J. KORT

338

"1

Amadorl Rearrangement

HOCH

r-

HCNH, H-C-0 /I f-El I alkali c. HOCH I

HCOH

HCOH

HCO

HCO I CH,OH

I

I

CH,OH

CH,NH, I

c=o

-

I HOCH

1

HCOH

n

H&O@ I CH,OH

a-D-Glucopyranosylamlne

CH,NH, I

c=o I

-

1

HCOH

HOCH HCOH I HCOH

I

C H,O

C H,OH

1-Amino- 1 -deoxy- a-0-f ructopyranose

Heyns Rearrangement H

c=o

H ~ - C 1H- J -- C N ~

HCOH HCOH

-

I HOCH I HCOH I HCOH I f CH,O

alkali

CH,O

I

HCNH, I

HOCH I

__c

HCOH I HCOH &,O@

a-u-Fructopyranosylamlne H

c=o

HCOH

-

HCO I

CH,OH

I HCNH, I HOCH I

HCOH I HCOH I

CH,OH

2-Amino-2-deoxy- a-D-glucopyranose

Possible Mechanisms of Amadori and Heyns Rearrangements in Alkali Scheme 5

REACTIONS OF FREE SUGARS WITH AQUEOUS AMMONIA H O*C-

+

I

H

339

OH

CI -CH,OH 1

H

Scheme 6

monia, and of 2-amino-2-deoxy-~-glucoseand D-fructosylamine from D-fructose and ammonia, has some experimental support. (a) The isolation of 2-amino-2-deoxy-~-glucose from the reaction of D-glucose and D-fructose with a m m ~ n i aat ~ low ~ ’ ~temperature (see Tables I and 111). ( b ) The isolation of D-glucosylamine from the reaction of Dglucose with ammonia at low temperature4*I2(see Tables I and 111) and at high temperature’* (see Table IV). ( c ) 2-Amino-S-deoxy-~glucose and ammonia at low temperature for 6 months yieldedg three of the substituted pyrazines (42, 50, and 51) (see Table 111) obtained from sucrose and ammonia at high temperaturelo for a shorter time (see Table IV). The high temperature would decrease the time of the reaction with sucrose, and would be necessary in order that the sucrose would first be hydrolyzed. ( d )2,5-Bis(~-arabino-tetrahydroxybuty1)pyrazine (42) has been synthesized by self-condensation of two molecular proportions of 1-amino-l-deoxy-~-fructose~~ or of 2-amino-2-deoxy-~-glucose.~ The pyrazines 50 (Ref. 9), 51 (Ref. 9) (see Table III), and 53 (Ref. 74) were also obtained from the selfIn addition, a similar condensation of 2-amino-2-deoxy-~-g~ucose. disubstituted pyrazine, namely, 2,5-bis(~-erythro-2,3,4-trihydroxybutyl)pyrazine, has been synthesized from two molecular proportions Indeed, the formation of 2,5of 3-deoxy-~-ribo-hexosylamine.~~ bis(D-arabino-tetrahydroxybuty1)pyrazine (42) (D-frUCtOSaZine)from 1-amino-1-deoxy-D-fructose(“isog~ucosamine”),76from 2-amino-2deoxy-D-glucose (“D-g~u~osamine”),~~ and from D-fructose plus ammonia78 (which form 2-amino-2-deoxy-~-glucose)~has been known for a long time. Although the early workers could not prove the formation of 2-amino-2-deoxy-~-g~ucose from the reaction of D-fruCtOSe with a m m ~ n i a , Lobry ~ ~ - ~ de ~ Bruyn assumed its presence. Heyns and Koch5 have, however, since obtained 2-amino-2-deoxy-~glucose from D-fructose plus ammonia, and have proposed a mecha(74) R. Kuhn, G. Kruger, H. J . Haas, and A. Seeliger, Ann., 644,122 (1961). (75) F. Micheel, S. Degener, and I. Dijong, Ann., 701, 233 (1967). (76) K. Maurer and B. Schiedt, Ber., 68,2187 (1935). (77) C. A. Lobry de Bruyn and W. Alberda van Ekenstein, Rec. Truu. Chim., 18, 77 (1898); C. A. Lobry de Bruyn, Ber., 28, 3082 (1895); 31, 2476 (1898); R. Breuer, ibid., 31, 2193 (1898); K. Stolte, Beitr. Chem. PhysioE., 11, 19 (1908). (78) C. A. Lobry de Bruyn, Rec. Trau. Chim., 18, 72 (1898).

340

M. J. KORT

nism for the reaction. By analogy, it is therefore possible that l-aminol-deoxy-D-fructose might be formed in the reaction of D-glucose with ammonia. In contradiction to this intermediary formation of the amines, K?5moto'2 found that D-glucosylamine partially decomposes at 20", and at 100", to D-glUCOSe and ammonia (see Tables I11 and IV). He maintained that D-glucosylamine, the condensation product of Dglucose with ammonia, does not, therefore, appear to be the primary intermediate product in the browning reaction between D-glucose and aqueous ammonia. However, the D-glucosylamine was only partially decomposed, and K6moto estimated the velocity of browning by measuring the absorption at 400 and 580 nm. Hence, his evidence does not negate the possible presence of D-glucosylamine as an intermediate in the formation of the pyrazines and imidazoles. 3. Fragmentation Mechanism

For imidazole formation, Windaus and Kn0opz4proposed that a molecule of D-glucose decomposes to give two D-glyceraldehyde molecules, and that the D-glyceraldehyde could either decompose further to formaldehyde, or form pyruvaldehyde by dehydration followed by migration of a hydrogen atom. The fragmentation products from L-rhamnose were also d i s c ~ s s e d . ~P~a, ~r r*~ d " 'extended ~~ this hypothesis to D-fructose, which would fragment to D-glyceraldehyde plus 1,3-dihydroxy-2-propanone. He then proposed occurrence of oxidation of the free sugars to a-ketoaldehydes; for example, of D-glucose (l),D-mannose (4, and D-fructose (3)to D-arabino-hexosulose (61),D-galactose (9) to D-lyxo-hexosulose, and D-glyceraldehyde and 173-dihydroxy-2-propanone to hydroxypyruvaldehyde. D-Glyceraldehyde could further fragment to formaldehyde and glyoxal, which could be oxidized to the oxalic acid that he isolated in his reactions. KGmoto also proposed the conversion of decomposition products of D-glucose into unstable a-diketones and aldehyde^.^,^^,^^ The aketoaldehydes could then condense with ammonia and the formaldehyde to give the various imidazoles shown in Scheme 7. In support of this mechanism, imidazoles are commonly prepared from adicarbonyl compounds (a-diketones and a-ketoaldehydes) and ahydroxycarbonyl compounds, with or without a catalyst, some examples of which are listed in Tables VIII and IX. The results of further investigations strongly support this formation of a-diketones and aldehydes. Before this work is considered, it is pertinent to discuss the action of alkali on sugars. As has already

REACTIONS O F FREE SUGARS WITH AQUEOUS AMMONIA

341

been noted, the action of ammonia on reducing sugars initially parallels the action of alkalis, which cause epimerization (the Lobry de Bruyn-Alberda van Ekenstein transformation).

44

H-N

.

0" g

8 0

sE 5

424

R. D. MARSHALL AND A. NEUBERGER

(98) M. Satake, T. Okuyama, K. Ishihara, and K. Schmid, Biochem. J., 95,749 (1965). (99) R. W. Jeanloz, in Ref. 6, Chapter 11, Section 4B. (100) R. Bourillon, R. Got, and D. Meyer, Biochim. Biophys. Acta, 83,178 (1964). (101) T. Yamauchi, M. Makino, and I. Yamashima, J . Biochem. (Tokyo), 64,683 (1968). (102) A. Caputo, A. Floridi, and M. L. Marcante, Biochim. Biophys. Acta, 181, 446 (1969). (103) K. Ishihara and K. Schmid, Biochemistry, 6,112 (1967). (104) L. Mester, E. Moczar, G. Medgyesi, and K. Laki, Compt. Rend., 256,3210(1963). (105) M. A. Cynkin and R. H. Haschmeyer, Fed. Proc., 23,273 (1964). (106) R. H. Haschmeyer, M. A. Cynkin, L.-C. Han, and M. Trindle, Biochemistry, 5 , 3443 (1966). (107) L. Mester, E. Moczar, and L. Szabo, Compt. Rend., 265,877 (1967). (108) S. Iwanaga, B. Blomback, N. J. Grondahl, B. Hessel, and P. Wallen, Biochim. Biophys. Acta, 160,280(1968). (109) J. R. Clampand F. W. Putnam,J. B i d . Chem., 239,3233(1964). (110) N. Duquesne, M. Monsigny, and J. Montreuil, Compt. Rend., 262,2536(1966). (111) G. M. Edelman, B. A. Cunningham, W. E. Gall, P. D. Gottlieb, U. Rutishauser, and M. J. Waxdal, Pmc. Nut. Acad. Sci. U.S., 63, 78 (1969). (112) C. Nolan and E. L. Smith,J. Biol. Chem.,237,446(1962). (113) N. Duquesne, M. Monsigny, and J. Montreuil, Compt. Rend., Ser. D, 261, 1430 (1965). (114) R. L. Hill, R. Delaney, R. E. Fellows, and H. E. Lebovitz, Proc. Nat. Acad. Sci. U.S., 56,1762 (1966). (115) F. Melchers, Biochemistry, 8,938 (1969). (116) A. B. Edmundson, F. A. Sheber, K. R. Ely, N. B. Simonds, N. K. Hutson,and J. L. Rossiter, Arch. Biochem. Biophys., 127,725(1968). (117) G. Spik and J. Montreuil, Intern. Symp. IV. Chromatographie EZectrophorese, Presses Acad. Europbenes, Bruxelles, 1968,p. 385. (117a)J.Williams and I. Graham, personal communication. (118) A. B. Rawitch, T. Liao, and J. G. Pierce, Biochim. Biophys. Acta, 160,360(1968). (119) P. G. Johansen, R. D. Marshall, and A. Neuberger, Biochem.J.,78,518 (1961). (120) Y. C. Lee and R. Montgomery, Arch. Biochem. Biophys., 97,9 (1962). (121) L. W. Cunningham, R. W. Clouse, and J. D. Ford, Biochim. Biophys. Acta, 78, 379 (1963). (122) E. D. Kaverzneva, F. V. Shmakova, and A. P. Andrejeva, Biokhimiya, 32, 964 (1967). (123) M. Monsigny, A. Adam-Chosson, and J. Montreuil, Bull. Soc. Chim. Biol.,50, 857 (1968). (124) R. Montgomery and Y . C . Wu, J . B i d . Chem., 238,3547(1963). (125) M. Tanaka, Yakuguku Zasshi, 81,1470(1961). (126) R. J. DeLange, Fed. Proc., 28,343 (1969). (127) J. Williams, Biochem.]., 108,57 (1968). (127a)T.C. Elleman and J. Williams, Biochem.]., 116,515(1970). (128) J. Z. Angustyniak and W. G. Martin, Cun.J.Biochem., 46,983 (1968). (129) T. H.Plummer and C. H.W. HirsJ. B i d . Chem., 239,2530 (1964). (130) T. H. Plummer, A. Tarentino, and F. Maley,]. Biol.Chem., 243,5158(1968). (131) R. L. Jackson, V. N. Reinhold, and C. H. W. Hirs, Fed. Proc., 27,529 (1968). (132) C. H. W. Hirs, personal communication. (133) B. J. Catley, S. Moore, and W. H. Stein,]. B i d . Chem., 244,933(1969). (134) Y. Yasuda, N. Takahashi, and T . Murachi, Biochemistry, 9,25 (1970). (135) M. Anai, T. Ikenaka, and Y. Matsushima, J . Biochem. (Tokyo), 59,57 (1966).

STRUCTURE AND METABOLISM OF GLYCOPROTEINS

425

Studies on the sequence of amino acids in the neighborhood of that L-asparagine residue to which the 2-acetamido-2-deoxy-~-glucosy1 moiety is attached (see Table VI) has revealed that no single, favored, primary structure exists in this region, with an important exception: in the great majority of compounds thus far studied, X-Q-YThr (or Ser) occurs, where Q represents the glycosylated residue, and X and Y are amino acid residues. I t was suggested that the sequence in the apoprotein may be a necessary condition for glycosylation to occur,136but the finding of sequences of this type in proteins not glycosylated suggests that this is not a sufficient condition. Thus, the overall stereochemistry and other factors must also be considered. Examples of such non-glycosylated proteins that contain a sequence of this type include human137calcitonin M, a penicillinase from Staphylococcus aureus in which it occurs twice (residues 186-188 and 241-243),13"nucleases from the same bacteria (residues 118-120),139 and guinea-pig immunoglobulin K-chain, close to the C-terminal end.140It may be mentioned that an earlier report to the effect that a sequence of this type occurs as residues number 102-104 of ox chymotrypsin was erroneous, and that residue 102 is not L-asparagine, but L-aspartic acid.14' Other non-glycosylated sequences of this type have been de~cribed.'~'" It has been suggested'42 that the nature of the side chain of residue Y may influence the size of the carbohydrate moiety found at that point, but further data are needed before this hypothesis can be considered firmly established.

(136) A. Neuberger and R. D . Marshall, in "Symposium on Foods-Carbohydrates and their Roles," H. W. Schultze, R. F. Cain, and R. W. Wrotstad, eds., Avi Publishing Co.,Westport, Connecticut, 1969, p. 115. (137) P. Sieber, M. Brugger, B. Kamber, B. Riniker, and W. Rittel, Helu. Chim. Acta, 51,2057 (1968). (138) R. P. Ambler and R. J. Meadway, Nature, 222,24 (1969). (139) C. L. Cusumano, H. Taniuchi, and C. B. Anfinsen, J . B i d . Chern., 243, 4769 (1968). (140) M. E. Lamm and B. Lisowska-Bernstein, Nature, 220,712 (1968). (141a)L.T. Hunt and M . 0. Dayhoff, Biochern. Biophys. Res. Cornmun., 39,757 (1970). (141) D. M. Blow, J. J. Birktoft, and B. S. Hartley, Nature, 221,337 (1969). (142) C. H. W. Hirs, R. L. Jackson, and I. Kabasawa, Abstracts Papers Amer. Chern. Soc. Meeting, 158, CARB 048 (1969).

426

R. D. MARSHALL AND A. NEUBERGER

2. Linkages in which O-2-Acetamido-2-deoxy-a-~-galactopyranosyl is Linked to L-Serine or L-Threonine Residues Structures of this type (2) occur in a number of glycoproteins. HO

H O I

1

HO,C -C-C-R I I H,N H 0 -@-Acetamido- 2-deoxy - (Y -Dgalactopyranosy1)-L-serine (R = H) or -L-threonine (R = Me)

(2)

Through studies with purified enzymes, it has been established that this type of linkage is of the a-Danomeric c o n f i g ~ r a t i o n . ' ~ ~In -'~~" this Section are included certain glycoproteins for which identification of the sugar residue involved in the carbohydrate-peptide linkage has not yet been unambiguously made. Recognition of the amino acid, and, more particularly, of the sugar involved in the linkage, would clearly be accomplished most satisfactorily were the sugar-amino acid compound constituting this part of the structure to be isolated from various proteins and characterized by the techniques of organic chemistry. Sometimes, this approach is difficult, because of the inability of proteolytic enzymes to hydrolyze the glycoprotein; this is particularly encountered with those molecules in which the extent of substitution by sugars is very high, as in the blood-group substances from ovarian cysts,145in which every second amino acid residue is an 0-glycosylated a-amino-phydroxy acid. With other glycoproteins, this problem does not arise, and it may be noted that the half-life of the glycosidic bond in O-aD-glucosyl-L-serine in 0.4 M hydrochloric acid at 100" is146about 6 hours; this indicates that the glycoside has a stability to acid some(143) B. Weissmann and D. F. Hinrichsen, Biochemistry, 8,2034 (1969). (144) E. Buddecke, H. Schauer, E. Werries, and A. Gottschalk, Biockem. B i o p h y s . Res. Cotnmrin., 34, 517 (1969). (144a)A. S. R. Donald, J . M. Creeth, W. T. J. Morgan, and W. M . Watkins, Biochem. J., 115,125 (1969). (145) W. M. Watkins, i n Ref. 6, Chapter 11, Section 7. (146) V. A. Derevitskaya, M. G. Vatina, and N. K. Kochetkov, Cnrbohyd. Res., 3, 377 (1967).

STRUCTURE AND METABOLISM OF GLYCOPROTEINS

427

what greater than the P-aspartylglycosylamine linkage (see p. 421). Moreover, if the sugar involved in the linkage is an N-acetylhexosamine, it might b e expected that, during hydrolysis with acid, a relatively much more stable 0-(aminodeoxyhexosy1)-L-serinewould be produced to some extent. Thus, when 0-(2-acetamido-2-deoxy@-D-glucosy1)-L-serinewas treated with 1.5 M hydrochloric acid for 90 minutes at loo", a compound having the electrophoretic mobility expected for the deacetylated derivative was formed. Similar results were obtained with the corresponding derivative of ~ - t h r e o n i n e . ' ~It' has been customary to design experiments based on the effects of alkali, in order that the amino acid and the sugar involved in the linkage might be recognized. Usually, the aim is to bring about a @-eliminationof the carbohydrate moiety. a. The P-Elimination Reaction. -One of the highly characteristic features of a compound having a structure of the type depicted in 2 is the facility with which it may undergo @elimination under alkaline conditions, and this type of reaction can be considered in general mechanistic t e r r n ~ . ' T~h~e ~reactions '~~ of glycoproteins here considered require the presence of a base, indicating that the reaction is a bimolecular elimination and that it may be represented as follows.

Transition state

where X is the 0-( 2-acetamido-2-deoxy-~-galactosyl)group, R is H or CH, (for L-serine or L-threonine, respectively), and R' and R" are the -CO- and -NH- parts of the peptide groups involved in bonding to the remainder of the chain. It should be emphasized that, in the notation generally used to , carbon atoms are the reverse describe this type of reaction, the LY and B of those usually employed in amino acid chemistry. The effects, in general, of the groups X, R, R', and R" were con(147) M. Monsigny, M . Buchet, and J. Montreuil, Intern. Syrnp. ZV ChrornatogruphieElectrophorese, Presses Acad. EuropCenes, Bmxelles, 1968, p. 361. (148) W. Hanhart and C. K. Ingold, j . Chern. SOC.,997 (1927). (149) M. L. Dhar, E. D. Hughes, C. K. Ingold, A. M. N. Mandour, G. A. Maw, and L. I. Woolf,J. Chem. SOC., 2093 (1948).

428

R. D. MARSHALL AND A. NEUBERGER

sidered in detail by Ingold and his colleagues.lS0The group X must always be a strongly electron-attracting group. Reaction is facilitated through operation of the inductive effect if R’ and R” are strongly electron-withdrawing, and because -NHR’ and -COR” are not alkyl groups, they cannot affect the reaction through hyperconjugation. The presence of the methyl group (R) in derivatives of L-threonine may, through hyperconjugation, have a stabilizing effect on the transition state, but the extent to which this is operative cannot in general be predicted.ls1 b. Side Reactions that may Occur When a Glycoprotein or Glycopeptide is Subjected to Conditions Suitable for &Elimination. When the conditions that result in occurrence of this reaction are applied to glycoproteins, a number of other reactions of relevance may occur. In the first place, N-deacetylation of the 2-acetamido-2deoxy-D-galactosyl group (X) may occur to some extent, and the resultant group, X’,has a lessened electron-withdrawing ability as compared with that of X. The effect of this change in structure on the rate of p-elimination has not yet been investigated. The effect of splitting the peptide bond -COR” will be replacement of a powerful, electron-attracting group by a powerfully electron-donating, negatively charged carboxyl ion, so that reaction will then occur much more slowly. Splitting of the peptide bond -NHR’- will, likewise, be expected to result in a diminution in the rate of reaction, but the effect of splitting this peptide bond would be expected to be less than in the aforementioned case, because the uncharged amino group thereby produced is still slightly electron-withdrawing.1s2 Moreover, the uncharged amino group may stabilize the transition state through the operation of a mesomeric effect. The extent to which these two opposing effects affect the rate of reaction cannot be predicted. c. The Effects of Alkali on Model Compounds of This Type. - The effects of substituents on the rate of p-elimination of model compounds have been examined q~antitative1y.l~~ During 24 hours at 37”, at pH 11, N-( benzyloxycarbony1)-0-P-D-glucopyranosyl-L-serine methylamide reacts to the extent of 95%, whereas, under the same (150) C. K. Ingold, “Structure and Mechanism in Organic Chemistry,” G. Bell & Sons, Ltd., London, 1963, Chapter 8; 2nd. Edition, 1969, Chapter 9. (151) J. W. Baker, “Hyperconjugation,” Oxford University Press, Oxford, 1952, Chapter 6 . (152) M. J. S. Dewar, “Electronic Theory of Organic Chemistry,” Oxford University Press, Oxford, 1949, p. 52.

STRUCTURE AND METABOLISM OF GLYCOPROTEINS

429

conditions, the corresponding free acid, namely, 0-P-D-glucopyranosyl-L-serine is completely stable. Similarly, O-P-D-glUCOpyranOSylL-serine methylamide is undegraded under these conditions. Somewhat surprising is the finding that, under the same conditions, 0P-D-glucopyranosyl-N-glycyl-L-serine methylamide is stable. Moreover, it was found that the free acid derivative already mentioned is not degraded by treatment with 0.1 M sodium hydroxide during 8 days at 37". Similar results were obtained with the corresponding derivatives of D-galactose. Under somewhat different conditions, namely, 0.1 M sodium hydroxide plus 0.3 M sodium borohydride during 24 hours at 20°, the 0-(2-acetamido-2-deoxy-~-glucosyl) derivatives of L-serine and L-threonine are stable, whereas their amides are almost completely ruptured by occurrence of 8-elimination.'53 Synthetic compounds in which the hydroxyl group of L-serine is glycosidically coupled with derivatives of D-glUCOSe,'46'154,155*155a D-gala~tose,'~ 2~ - amino , ~ ~ ~- 2 , ~- deoxy ~ ~ ~ - D - glucose,155bJ56*157 D - xy1ose,'55~158~159 lactose, and c e l l ~ b i o s e , 'as ~ ~well ~ as ones in which L-threonin? is linked to D-xylose,160D-glucose,'61or 2-amino-2-deoxyD-glUCOSe,'62have been described. d. Recognition of the Amino Acid Involved in the Carbohydratepeptide Bond. -Treatment with alkali of glycoproteins that contain glycosidic linkages involving L-serine or L-threonine, or both, leads to the formation, from 0-glycosylated L-serine and L-threonine residues, respectively, of 2-aminoacrylic and 2-aminocrotonic acid residues within the chain. The difference in the number of residues of L-serine or L-threonine, or both, in acid hydrolyzates of the glycoprotein, both before and after treatment with alkali, is, in general, (153) J. Montreuil, M . Monsigny, and M. Buchet, Compt. Rend., Ser. D , 264, 2068 (1967). (154) J. K. N. Jones, M. B. Perry, B. Shelton, and D. Walton, C a n . ] . Chem., 39, 1005 (1961). (155) F. Kum and S. Roseman, Biochemistry, 5,3061 (1966). (155a)K.Kum, Carbohyd. Res., 11,269 (1969). (155b)E.Rude and M. Meyer-Delius, Carbohyd. Res., 8,219 (1968). (156) F. Micheel and H. Kochling, Chem. Ber., 93,2372 (1960). (157) J . R. Vercellotti and A. E. Luetzow,]. Org. Chem.,31,825 (1966). (158) B. Lindberg and B. Silvander, Acta Chem. Scand., 19,530 (1965). (159) K. Brendel and E. A. Davidson, Carbohyd. Res., 2,42 (1966). ~ (160) M. Higham, P. W. Kent, and P. Fisher, Biochem.]., 1 0 8 , 4 7 (1968). (161) V. A. Derevitskaya, E. M. Klinov, and N. K. Kochetkov, Carbohyd. Res., 7, 7 (1968). (162) J. R. Vercellotti, R. Femandez, and C. J. Chang, Carbohyd. Res., 5,97 (1967).

430

R. D. MARSHALL AND A. NEUBERGER

approximately equal to the number of carbohydrate moieties that were formerly linked through these residues and that are released from the polypeptide chains. However, this reaction may not always be sufficiently specific for glycosylated L-serine or L-threonine moieties. It has been found that, under conditions often employed, namely, 0.5 M sodium hydroxide for 96 hours at 4", there is often little or no decomposition of nonglycosylated L-serine or L-threonine residues, as judged from the results of control experiments in which edestinIti3 or orosomucoid1"J65 were employed. However, it has been reported that P-hydroxy a-amino acid residues in the protein derived from sheep submaxillary-gland glycoprotein by removal of the sugar residues are unstable under alkaline conditions.'66 Treatment of lysozyme at pH 12.8 (ionic strength 1.6)for 45 minutes at 70" led to a loss of somewhat less than 15%of the L-serine and L-threonine contents of the protein.'"' We believe that data concerned with alkaline conversion of L-serine and L-threonine residues into the corresponding unsaturated derivatives must be interpreted with caution. It would not be altogether surprising were certain unsubstituted P-hydroxy a-amino acid residues, in their protein environment, to undergo this reaction with facility, Decrease in the proportion of L-serine or L-threonine, or both, when pig gastric blood-group substance,lti4 casein,16Eor a mucin derived from colloid carcinoma of the breast44 were subjected to alkaline conditions may suggest that they all contain linkages of this type. The increase found in the proportion of glycine in the lastmentioned glycoprotein when it was treated with alkali and the product was hydrolyzed with acid was interpreted as being due to a second possible pathway for breakdown of 2-aminocrotonic acid residues, namely a retroaldol reaction.1Bg Peptides that contain 2-aminoacrylic acid and 2-aminocrotonic acid groups absorb in the ultraviolet region, the former at a wavelength maximum of 240 nm, and the latter with a more generalized absorption. At 240 nm, the molar absorptivity is the same (4,200) for the two (163)B. Anderson, P.Hoffman, and K. Meyer, Biochim. Biophys. Acta, 74,309(1963). (164)B. Anderson, N. Seno, P. Sampson, J. G . Riley, P. Hoffman, and K. Meyer,]. Biol.Chem., 239,PC 2716 (1964). (165)B. Anderson, P.Hoffman, and K. Meyer,]. Blol. Chem.,240,156(1965). (166)J. E. McGuire and S. Roseman,]. Biol. Chem., 242,3745(1967). (167)S. Harbon, G. Herman, B. Rossignol, P. JollBs, and H. Clauser, Biochem. Biophys. Res. Commun., 17,37(1964). (168)F.H. Malpress and M. Seid-Akhaven, Biochem.]., 101,764(1966). (169)J. B. Adams, Biochem.]., 97,345(1965).

TABLEVII The Effects on the Contents of Serine and Threonine of Some Glycopeptides and Glycoproteins on Treatment with Alkali in Presence of Borohydridea Alkaline conditions Type of material

Molarity Molarity Glycoprotein

Glycopeptide yA myeloma protein rabbit yC globulin lactotxansfemn yA lactoglobulin Glycoprotein ox submaxillarygland glycoprotein pig submaxillarygland glycoprotein earthworm-cuticle collagen

Temp.

Time

of OH@ of NaBH, (degrees) (hr.) 0.1

0.3

0.5

-

0.5 0.1 0.1 0.1 0.1 0.5 0.1 0.1 0.05

0.3 0.3 0.3 0.3 0.3

-

0.3 0.3 0.15

Reduced further with

Decrease inb Increase inb

Refer-

Ser

ences

Thr

270

-

2

20

-

1.0

20

I70 70 70 216 216 96 6 6 24

-

1.2 3.9 3.5 218 170 243 394 373 162

4

20 20 5 5 4 45 45 30

0.15 2.9 sodium-liquid ammonia 2.3 357 NaBH,-palladiumchloride 317 H2-Adams’catalyst 462 NaBH,-palladiumblack 683 NaBH,-palladium chloride 553 51

Ala

But

0.3

172 20

0.13 1.75 0.63 3.4 2.9 311 32 147 311 92 90 192 567 483 258

117 173 171 174 164 175 79 ~~

“The 2-aminoacrylic and 2-aminocrotonic acid residues initially produced by alkali are reduced to L-alanine and 2-aminobutyric acid (But) residues, respectively. Use of a further reducing agent is indicated. T h e results are expressed as decrease or increase in moles of amino acid for the glycopeptides, and as pmoles per g for the glycoproteins.

432

R. D. MARSHALL AND A. NEUBERGER

types of derivative,I7O so that, in some instances, quantitative measurements may be used for determining the number of unsaturated residues formed. L-Cysteine and L-cystine residues might further complicate the interpretation of the data, as could also the presence of phosphorylated p-hydroxy a-amino acid residues. Addition reactions across the double bonds of the 2-aminoacrylic and 2-aminocrotonic acid residues formed often permit comparisons of the amounts of L-serine or L-threonine, or both, that are decomposed by alkaline conditions, and estimates of the quantities of the addition products to be made. Pigman and coworkers171introduced the use of sodium borohydride to reduce the double bonds formed during alkaline treatment, and this reagent is highly effective in converting 2-aminoacryloyl into L-alanyl residues within the modified polypeptide chain (see Table VII). However, the corresponding derivative of L-threonine is less readily reduced by this reagent, and further reduction with sodium borohydride in alkali containing palladium chloride leads to higher, but still not quantitative, yields of 2-aminobutyric acid moieties (see Table VII). Related studies with the A and B substances from ovarian cysts, as well as with pig gastric-mucin (A H) substance showed that over 80% of the Lserine and L-threonine was decomposed on treatment with sodium borohydride (0.26 M ) in alkali (0.2 M ) during one week at room temperature. The extent of formation of 2-aminobutyric acid was small. It was further shown that alkaline decomposition of L-serine and Lthreonine residues proceeds to a greater extent in the presence than in the absence of sodium b~rohydride."~."~ L-Cysteic acid residues were produced within the peptide chain when submaxillary-gland glycoprotein from sheep was treated with alkali (pH 9.0), during 24 hours at room temperature in the presence of sulfite (0.1 M).The yield was -55% on the basis of the L-serine decomposed, and the corresponding derivative from L-threonine was not detected.IB7

+

(170)V. E.Price and J. P. Greenstein, Arch. Biochem. Biophys., 18,383(1948). (171)K. Tanaka, M.Bertolini, and W. Pigman, Biochem. Biophys. Res. Commun., 16, 404 (1964). (172)G . Dawson and J. R. Clamp, Blochem../.,107,341(1968). (173)J. Descamps, M.Monsigny, and J. Montreuil, Conpt. Rend., Ser. D, 266, 1775 (1968). (174)K.Tanaka and W. Pigman,./. Blot. Chem.,240, PC 1487(1965). (175)N.Payza, S. Rizvi, and W. Pigman, Arch. Btochem. Btophys., 129,68(1969). (176)E. A. Kabat, E. W. Bassett, K. Pryzwansky, K. 0. Lloyd, M. E. Kaplan, and E. J. Layug, Biocherndstry,4,1632(1965). (177)D. M. CarlsonJ. Blol. Chem.,243,616(lQ68).

STRUCTURE AND METABOLISM OF GLYCOPROTEINS

433

e. Recognition of the Sugar Involved in the Carbohydrate-peptide Bond.-It has been shown that alkaline conditions, namely 0.5 M sodium hydroxide, during 48 hours at 22", cause elimination of 2acetamido-2-deoxy-~-galactose from glycopeptides prepared from submaxillary-gland glycoprotein of sheep. The glycopeptides used residues as the contained single 2-acetamido-2-deoxy-~-galactose carbohydrate moieties, and the number of these residues (which behaved colorimetrically as free 2-acetamido-2-deoxy-~-galactose after the reaction) was equal to 90% of the number of L-serine and L-threonine residues decomposed b y the treatment. The difference (10%)was found to be equivalent to the amount of free 2-acetamido-2deoxy-D-galactose which, in control experiments, was converted into under compounds not estimatable as 2-acetamido-2-deoxy-~-galactose these condition^.'^^ However, in general, when the carbohydrate attached to the glycopeptide or glycoprotein under examination has a more complicated structure, very considerable decomposition of the sugar moiety originally involved in the carbohydrate-peptide linkage may occur under the conditions used to split the latter. Thus, when blood-group substances from ovarian cysts were treated with alkaline borohydride or alkaline borodeuteride under the conditions frequently employed to cause P-elimination from glycoproteins, namely, 0.2 M alkali with 0.26 M sodium borohydride or sodium borohydride-d4, the elimination reaction was succeeded by extensive alkaline degradation of the oligosaccharide chains concomitant with reduction, to the corresponding alditols, of the reducing-sugar moieties thereby e x p o ~ e d . ' ' ~ Similar losses were observed when pig submaxillary-gland glycoprotein and the M- and N-active substances from human erythrocytes were treatedIs0 under similar conditions, namely, 0.02 M sodium hydroxide plus 0.4 M sodium borohydride for 16 hours at 25", or 0.2 M sodium hydroxide plus 0.2 M sodium borohydride for 48 hours at 25". Alkaline degradation of an oligosaccharide leads to loss of the sugar originally in the reducing position, because the reaction may be expected to proceed by p-elimination. If the vicinal sugar is linked to 0 - 3 of the reducing-sugar residue, the reaction occurs with much greater facility than when 0-4or 0-6is involved. A linkage involving 0 - 2 is stable under alkaline conditions. Studies have been made with oligosaccharides of the type encountered in glycoproteins.'s' For the blood(178) R. Carubelli, V. P. Bhavanandan, and A. Gottschafk, Biochim. Biophys. Acta, 101,67 (1965). (179) K. 0. Lloyd, E. A. Kabat, and E. Liciero, Biochemistry, 7,2976 (1968). ( 1 8 0 ) P. Weberand R. J. Winzler, Arch. Biockem. Biophys., 129,534 (1969). (181) A. Neuberger and R. D. Marshall, in Ref. 6, p. 262.

434

R. D. MARSHALL AND A. NEUBERGER

group substances already mentioned, it is considered that 2-acetamido2-deoxy-D-galactose residues are linked to L-serine and L-threonine residues within the polypeptide chain, and that some of these sugar residues are substituted at 0 - 3 (see p. 452). It is not at present possible to interpret unambiguously the finding that part of the 2-amino-2deoxy-D-glucose originally present in the material resembling keratan sulfate that is a constituent of chick allantoic fluid is decomposed when the material is treated with alkali.64 The occurrence of 2-acetamido-2-deoxy-D-galactose as part of the carbohydrate-peptide linkage in several proteins is based on acceptable evidence. In the submaxillary-gland glycoprotein from sheep and ox, the carbohydrate moieties consist of 2-acetamido-2deoxy-6-O-sialyl-~-galactosylresidues. The action of neuraminidase results in the production of modified glycoproteins in which 2-acetamido-2-deoxy-~-galactoseis by far the major sugar component. This was also the only amino sugar in a glycopeptide isolated from rabbit y G globulin.20 Treatment of blood-group B substance from ovariancyst fluid with a solution of 25 mM sulfuric acid in acetic acid at 60" for 24 hours led to the production of a modified glycoprotein that had an amino acid composition closely similar to that of the parent material, but from which sugars other than 2-acetamido-2-deoxy-~-galactose had been largely eliminated.'44n

3. 0-P-D-Xylopyranosyl-L-serine as a Carbohydrate-peptide Linking Moiety Linkages of this type (see 3) that occur within glycoproteins may

0-CH,-C-CO,H

I

H

0-$-o-Xylopyranosyl-L- serine (3)

also undergo p-elimination under alkaline conditions. In acid, however, the rate of cleavage may be expected to be of the same order as that for 0-P-D-glucopyranosyl-L-serine, which is reported'46 to be quite stable in solution at p H 1.5 during 24 hours at 100". O-P-DXylopyranosyl-L-serine has been isolated from acid hydrolyzates (pH 1.55), after 3 hours at loo", of commercial heparin containing L-serine as the only amino acid present in significant proportion. The

STRUCTURE AND METABOLISM OF GLYCOPROTEINS

435

preparations of heparin used were presumably from heparin proteoglycans that had been subjected to proteolysis prior to isolation. The nature of the compound isolated was established by hydrolysis with 4 M hydrochloric acid during 3 hours at loo", followed by chromatographic identification of both a xylose and a serine. That the sugar was D-xylose was proved with the use of D-xylose isomerase (Dxylose aldo-keto-isomerase, E.C. 5.3.1.5).183The quantity of material obtained accounted for 13% of the original L-serine, in addition to a further 135%of an 0-D-galactosyl-D-xylosyl-L-serine and 17% of the free amino acid. O-D-Xy1osyl-L-serine was isolated from acid hydrolyzates (pH 1.5) obtained, after 4 hours at loo", from the serine-oligosaccharides obtained from enzymic hydrolyzates of "chondromucoprotein," consisting mainly of chondroitin 4-sulfate, from nasal septa.184 Isolation methods have been employed for demonstrating that, in umbilical cord, chondroitin 6-sulfate is linked to a protein chain by the same type of bond.IK5It had earlier been established that chondroitin 6-sulfate from shark cartilage undergoes a considerable loss in L-serine (35 pmoles per gram) when it is treated with 0.5 M sodium hydroxide for 19 hours at 0 or 24". About half of this L-serine is converted into L-alanine by hydrogenation in the presence of Adams' catalyst. le5 Moreover, D-xylose occurs in glycopeptides isolated from enzymic digests of chondroitin 6-sulfate-protein complexes from a variety of sources.Is6 Comparable experiments were performed with chondroitin 4-sulfate, with similar results. Xylitol was identifiedIx7 by gas-liquid chromatography after chondroitin 4-sulfate from pig costal-cartilage had been subjected to the action of 0.2 M potassium hydroxide and sodium borohydride during 20 hours at 4".Xylitol was also identifiedIs8 as its pentaacetate after treatment with 0.5 M sodium hydroxide and 0.25 M sodium borohydride during 75 minutes at room temperature. O-@-D-Xylosyl-L-serinewas also isolated from the cartilage,Is8as well as from chick-embryo ~ a r t i 1 a g e . l ~ ~ (182) E. R. B. Graham and A. Cottschalk, Biochim. Biophys. Acta, 38,513 (1960). (183) U. Lindahl and L. Rod&]. Biol. Chem., 240,2821 (1965). (184) L. Rod& and U. Lindahl,]. Biol. Chem., 240,606 (1965). (185) T. Helting and L. Rod&, Biochim. Biophys. Acta, 170,301 (1968). (186) M. Schmidt, A. Dmochowski, and B. Wierzbowska, Biochim. Biophys. Acta, 117,258 (1966). (187) N. Katsuraand E. A. Davidson, Biochim. Biophys. Acta, 121,120 (1966). (188) E. E. Grebner, C. W. Hall, and E. F. Neufeld, Arch. Biochem. Biophys., 116, 39 1 (1966). (189) H. C. Robinson, A. Telser, and A. Dorfinan, Proc. Nut. Acad. Sci. U . S., 56, 1859 (1966).

436

€3. D. MARSHALL AND A. NEUBERGER

Comparison of the optical rotatory dispersion exhibited by 0-p-Dxylopyranosyl-L-serine isolated from glycosaminoglycans with those of the two synthetic anomers revealed that the natural material has Moreover, the O-D-XylOSyl-L-Serine the p-D anomeric configurati~n.'~~ present in normal urine also has the p-Dconfiguration, as had earlier been suggested.190The origin of the urinary compound is unknown, but it may be a catabolite of certain glycoproteins. Although this type of linkage is frequently labile under alkaline conditions, it is known that extensive digestion with papain of "chondromucoprotein" may lead to the formation of chondroitin 4-sulfate in which L-serine is the major if not the only amino acid present. Not surprisingly, under the circumstances, the carbohydrate-amino acid bond is not readily split'O by 0.15 M sodium hydroxide during 96 hours at 20" or1"' by 0.2 M potassium hydroxide during 20 hours at 4". At one linkage of this type in the chondroitin 4-sulfate from pig c~stal-cartilage,'~~ the sequence of amino acids is reported to be -Glu-Gly-Ser-Gly-, where Ser is the glycosylated residue. including The presence of D-XylOSe in several other glycoprotein~,~~ an alpha-amylase (see p. 442) and pineapple-stem b r ~ m e l a i n has '~~ been reported, but its mode of attachment is not yet known. 4. 5-O-/3-D-Galactopyranosyloxy-L-lysine as a Carbohydrate-peptide Linking Moiety This structural entity (see 4) occurs at the junction of the oligosacHO

H

H

5- 0 - p - D - Galactopyranosyloxy-L-lysine (4)

charide units to the polypeptide chain in both the citrate-soluble and citrate-insoluble collagen'g1 from guinea-pig skin, in the lens capsule (190) F. Tominaga, K. Oka, and H. Yoshida,]. Riochem. (Tokyo), 57,717 (1965). (191) L. W. Cunningham and J. D. Ford,]. B i d Chem.,243,2390 (1968).

STRUCTURE AND METABOLISM OF GLYCOPROTEINS

437

of the eye,lgla in the soluble stroma of calf cornea,lg2 and in the basement membrane of ox-kidney g l o m e r u l ~ s . ' Compounds ~~ in which a disaccharide is linked through the hydroxyl group of 5hydroxy-L-lysine have been isolated from the hydrolysis, with 2 M sodium hydroxide during 16-20 hours at QO-lOF, of glycopeptides isolated from these proteins, or, for the basement membrane, from the protein itself. Not unexpectedly, this type of linkage is stable under alkaline conditions. The material obtained from basement membrane was further degraded by 0.05 M sulfuric acid during 28 hours at 100" and 5-O-~-~-galactopyranosyloxy-~-lysine was isolated.lg3 The occurrence of the positively charged amino group in the aglycon under acidic conditions has the effect of stabilizing the glycosidic linkage to a considerable degree. lg3 This factor has been investigated quantitatively by comparing the rate of acid hydrolysis of ethyl P-D-glucopyranoside with that of 2-aminoethyl p-D-glucop y r a n 0 ~ i d e . The l ~ ~ presence, in the aglycon, of the positively charged group attached to the carbon atom in the P position to the anomeric oxygen atom lowers the rate of hydrolysis of the glycosidic bond by a factor much less than that for glycosides of 2-amino-2-deoxy-D-glucose (see Table VIII). The most probable explanation for this difference TABLE VIII The Rates of Hydrolysis of Some Glycopyranosides in 2 M Hydrochloric Acid at 80" Glycopyranoside Ethyl P-D-glucopyranoside 2-Aminoethyl P-D-glucopyranoside Methyl 2-amino-2-deoxy-a-D-glucopyranoside"

k (min-') 19.2X 4.3 x lo-:' 0.09 x lO+J

References

195 194 196

"In 2.5M hydrochloric acid.

lies in the fact that the doubly charged, positive, oxonium-ion intermediate produced in the hydrolysis of 2-aminoethyl P-D-glucopyranoside undergoes cleavage in such a way as to cause separation of the two positive charges (see Fig. la). On the other hand, the oxonium (191a)R.G. Spiro and S. Fukushi,]. B i d . Chem.,244,2049(1969). (192)E. Moczar, L.Robert, and M. Moczar, Eur0p.J. Biochem., 6,213(1968). (193)R.G. Spiro,]. Biol. Chem., 242,4813(1967). (194)E. R. B. Graham and A. Neuberger,]. Chem. SOC. ( C ) ,1638(1968). (195)W.G. Overend, C. W. Rees, and J. S. Sequeira,]. Chem. SOC.,3429 (1962). (196)R. G . C. Moggridge and A. Neuberger,J. Chem. SOC.,745 (1938).

438

R. D. MARSHALL AND A. NEUBERCER

FIG.1.-Reaction Mechanism for the Acid Hydrolysis oE(a) 2-Aminoethyl p-~-Glucopyranoside and (b) Methyl 2-Amino-2-deoxy-~-~-g~ucopyranoside.

ion produced when methyl 2-amino-2-deoxy-~-glucopyranoside is protonated is likely to undergo conversion into the carbonium ion with difficulty, because this reaction involves a still closer approach of two positive charges to each other (see Fig. lb). An alternative explanation is based on the assumption that the protonation required as the preliminary step in cleavage of the glycosidic bond occurs on the ring-oxygen atom.197If, in 2-aminoethyl P-D-glucopyranoside, a di-cation having positive charges on the amino group and on the ring-oxygen atom is cleaved, repulsion between the two charges will be greatly lessened. In contrast, with the a shift of the proton from ethyl 2-amino-Zdeoxy-~-gIucopyranosides, the anomeric oxygen atom to the ring-oxygen atom would not cause the distances between the two positive charges to be altered to the same extent. The two explanations are not mutually exclusive. were Isomeric forms of 5-O-~-D-galactopyranosyloxy-L-lysine isolated after hydrolysis, with 2 M sodium hydroxide during 16 hours at go", of skin collagen,198and it was suggested that racemization of the amino acid moiety had occurred during the hydrolysis. In the natural compound, the anomeric configuration is believed to be p-D, because a-D-galactosidase is without effect on the 5-O-P-D-galaC(197)C . A. Bunton, T. A. Lewis, D. R. Llewellyn, and C. A. Vernon, J. Chem. SOC., 4419 (1955). (198) L. W. Cunningham, J. D. Ford, and J. P. Segrest, J. B i d . Chem., 242,2570 (1967).

STRUCTURE AND METABOLISM OF GLYCOPROTEINS

439

topyranosyloxy-L-lysine isolated from basement membrane, either before or after N-acetylation. Moreover, although the compound is not split by P-D-galactosidase, its N-acetylated derivative is, albeit very slowly.1s3 The sequence of amino acids in the vicinity of at least some of the glycosylated 5-hydroxy-~-lysineresidues in the collagen from guineapig skin was shown1ss to be the following, relatively rare structure,

-Gly-Met-Hyl-Gly-His-Arg-, where Hyl represents the amino acid that is glycosylated. This sequence is not the only, or even the necessary, sequence for glycosylation, because glycopeptides from the basement membrane do not contain this sequence. It has been shown that one molecule of these glycopeptides contains about two molecular proportions of glycine and one of L-glutamic acid, in addition to 5-hydroxy-L-lysine.*OOBoth the citrate-soluble and the citrate-insoluble collagen from guinea-pig skin contain -19 5-hydroxy-~-lysineresidues per 3 X lo5daltons, and about one-third of these are g l y c o ~ y l a t e d . ' ~ ~

111. POLYPEPTIDECHAINSCARRYING MORE THANONE TYPEOF CARBOHYDRATE-PEPTIDE LINKAGE

The possibility of categorizing glycoproteins according to the nature of the carbohydrate-peptide bond contained has been sugg e ~ t e dTo . ~ some ~ extent, this method is possible, but, as was pointed out, it does not result in a rigid classification, for examples are known in which more than one type of carbohydrate moiety is attached to a given polypeptide chain. In studies conducted by Partridge and Elsden,68 it was found that " chondromucoprotein" prepared from ox nasal-septa without the use of alkali or proteolytic enzymes gave rise, on treatment with 0.5 M sodium hydroxide during 24 hours at 25", to protein-free chondroitin sulfate and a heterogeneous, glycoprotein fraction whose carbohydrate composition suggested that it was keratan sulfate. The carbohydrate-peptide linkage binding the latter to protein appears to be more resistant to alkali than that joining chondroitin sulfates to the polypeptide chain (see p. 435). It is, therefore, likely that chondroitin sulfate and keratan sulfate may be attached to the same polypeptide chain by different types of linkage, as indicated also by the (199) W. T. Butler and L. W. Cunningham,J. Biol. Chem.,241,3882 (1966). (200) R. C . Spiro, j . B i d . Chem., 242,1923 (1967).

440

R. D . MARSHALL AND A. NEUBERGER

work of Gregory and R o d h Z o 1Human knee-joint cartilage also contains a complex composed of protein attached to keratan sulfate and chondroitin sulfate; it has been reported that Li-hydroxy-~-proline occurs in the glycopeptides obtained from papain digests of this material.z0zIt should be emphasized, however, that this compound is not always found, for chondroitin 4-sulfate-protein complexes extracted from pig laryngeal-cartilage are free from keratan sulfate.203 The absence of the latter from these glycoprotein preparations might result from the polypeptide chain's not containing the necessary marker-sequence (see p. 425) that may well be required so that keratan sulfate can become attached to the protein. The polypeptide chain to which chondroitin sulfate is attached is not always the same.70~165~zo3--205 There is probably a variety of polypeptide chains to which chondroitin sulfates, keratan sulfate, and smaller oligosaccharides may be attached. In some complexes, the nature of the recipient chain for the sugar moieties, or the distribution of the relevant sugar transferases, is such that more than one type of carbohydrate moiety becomes attached. 2-Acetamido-2-deoxy-~-galactose residues are attached to the polypeptide chain that is part of the structure of the keratan sulfate prepared from Pan-Protease digests of old human-rib cartilage, and these residues are linked to P-hydroxyl groups of L-serine and Lthreonine residues. Some, at least, of the 2-acetamido-Z-deoxy-~galactose residues form branch points, probablyzo5with substituents at 0-3 and 0-6. Other types of keratan sulfate have been shown also to contain alkali-labile linkages in which L-serine and L-threonine p a r t i ~ i p a t e . It ' ~ is, ~ however, possible that the disaccharide repeatingunit (which is that structure comprising the major component usually identified as keratan sulfate) is not attached to this 2-acetamido-2deoxy-D-galactose, and L-serine or L-threonine residues (or both) either directly or indirectly, because the major carbohydrate component is not split from the keratan sulfate isolated from chick allantoic-fluid when this material is treated with 0.5 M sodium hydroxide for 48 hours at room temperature. All of the samples of keratan sulfate used contained various proportions of sialic acid and L(201)J. D.Gregory and L. Rodbn, Biochem. Biophys. Res. Commun.,5,430(1961). (202)H.Greiling and H. W. Stuhlsatz, 2.Physiol. Chem.,350,449(1969). (203)H.Muir and S. Jacobs, Biochem.]., 103,367(1967). (204)A. A. Castellani, B. Bonferoni, S . Ronchi, G. Ferri, and M. Malcovati, Ital. J . Biochem., 11,187(1962). (205)C. P.Tsiganos and H. Muir, Biochem.J,113,885(1969). (206)B.A.Bray, R. Lieberman, and K. MeyerJ. BioE. Chem.,242,3373(1967).

STRUCTURE AND METABOLISM OF GLYCOPROTEINS

441

f u c o ~ e . These ~ ~ ‘ constituents may occur as part of carbohydrate moieties that involve at least some of the 2-acetamido-2-deoxy-~-galactose residues in structures analogous to those present in blood-group substances. D-Mannose, also, was found in keratan sulfate isolated from human rib208 and kneezo9 cartilage. Some preparations from rib cartilage have chondroitin 6-sulfate linked, by way of D-xylose, to the same polypeptide hai in.^^^,^^^ By the use of somewhat different techniques, a keratan sulfate that is largely free from L-fucose and sialic acid may be isolated from human costal cartilage, and the protein and carbohydrate entities of such preparations cannot be cleaved”’ with 0.5 M potassium hydroxide during 48 hours at 25”. Ox-corneal keratan sulfate is also attached to the protein core by linkages stable under alkaline conditions207 and the linkage is believed212 to involve L-asparagine and 2-acetamido-2-deoxy-~-glucose, although definite proof is still lacking. It is, therefore, reasonable to infer that cartilage contains glycoproteins in which there are residues of ( a ) L-asparagine that are substituted with 2-acetamido-2-deoxy-~-glucoseand that, to some extent, may be part of the keratan sulfate side-chain(s),( b )L-serine substituted by D-XylOSe attached to chondroitin sulfate polysaccharide moieties, and (c) L-serine and L-threonine to which are attached 2-acetamido-2deoxy-D-galactose residues that form a part of small oligosaccharide moieties; the antigenic cross-reactivity of keratan sulfate and bloodgroup substances may be relevant here.213More work is required before a general hypothesis can be put forward with confidence. In interpreting some of the earlier data, there are difficulties partly attributable to the fact that proteolytic digestion had preceded isolation of the material; in these studies, different proteolytic enzymes were used. The finding that, when the protein-polysaccharide complex (mainly chondroitin 4-sulfate) from pig rib-cartilage was treated with 0.2 M potassium hydroxide for 20 hours at 4”,only about 80% of the carbohydrate moieties were removed by &elimination, was interpreted as (207) N. Seno, K. Meyer, B. Anderson, and P. Hoffman,J. Biol. Chem., 240,1005 (1965). (208) V. P. Bhavanandan and K. Meyer (1967), quoted in Reference 206. (209) H. Greiling, in “Aktuelle Probleme des Rheumatismus,” H. Riissler and R. Heister, eds., F. K. Schattauer-Verlag, Stuttgart and New York, 1969, p. 53. (210) K. Meyer, P. Hoffman, and A. Linker, Science, 128,896 (1958). (211) M. B. Mathews and J. A. Cifonelli,J. Biol. Chem., 240,4140 (1965). (212) H. Greiling, H. W. Stuhlsatz, R. Kisters, and L. Plagemann, Ahstr. 5th Meeting Fed. Europ. Biochem. Soc., Prague, 1968, p. 155. (213) 0.Rosen, P. Hoffman, and K. Meyer, Fed. Proc., 19,147 (1960).

442

R. D. MARSHALL AND A. NEUBERGER

evidence that the remainder were bound to the polypeptide chain by linkages of a type different from the glycosidic ones involving Lserine and L-threonine. The inability to detect D-XylOSe or D-galactose in the fraction stable to alkali supports the view that a second type of linkage may be present.73 Several other examples of more clearly defined glycoproteins are known in which there is more than one type of carbohydrate-peptide bond. In rabbit yG globulin, an L-asparagine residue in the Fc region of the heavy chains is substituted by 2-acetamido-2-deoxy-~-glucose (see Table VI), and an L-threonine residue is glycosylated with 2-acetamido-2-deoxy-~-galactose in the hinge region of about 35% of the total number of heavy chains.20These two types of linkage occur also in yA (Bra) myeloma protein172and in human, chorionic gonadot r ~ p i n , ~but, " in these examples, L-serine provides the site of attachment of 2-acetamido-2-deoxy-~-galactose.A "heavy-chain disease" protein (Cra),215l a ~ t o t r a n s f e r r i n ,yA ~~~ l a c t ~ g l o b u l i n ,and ~ ~ ~an oxaorta glycoprotein216are yet further compounds in which these two types of linkage are present. alpha-Amylase from Aspergillus oryzae (EC 3.1.1.1)was found to have an amino acid sequence around a carbohydrate moiety

-Ser-Glu-Asp-Gly-(Ala,Thr)-, and to the L-serine residue was attached an oligosaccharide that contained 8 molecules of D-mannose and one of D-XylOSe per molecule.217 Although the whole molecule contained about two molecules of hexosamine, the amino sugar(s) was (were) not present in the glycopeptides isolated. Later studies with other preparations of alphaamylase revealed the presence of another linkage, involving 2acetamido-2-deoxy-~-glucose and ~ - a s p a r a g i n e . 'Interpretation ~~ of these data is complicated by the realization that the carbohydrate composition of preparations of enzyme probably varies from one batch to another.217u Collagen and basement membrane may contain several types of carbohydrate-peptide linkage. They clearly contain linkages in which 5-hydroxy-~-lysine is involved, as already discussed (see p. 436). From the preparations studied were also isolated glycopeplinking tides in which an asparagine-2-acetamido-2-deoxy-~-glucose moiety was p r e ~ e n t . ' ~ ' * ~ ~ (214) 0.P. Bahl,]. B i d . Chem., 244,575 (1969). (215) J. R. Clamp, G. Dawson, and E. C. Franklin, Biochem.]., 110,385 (1968). (216) B. Radhakrishnamurthy and G. S. Berenson,]. Biol. Chem., 241,2106 (1966). (217) A. Tsugita and S. Akabori,J.Biochem. (Tokyo),46,695 (1959). (217a)J. F. McKelvy and Y. C. Lee, Arch. Biochem. Biophys., 132,W (1969).

STRUCTURE AND METABOLISM OF GLYCOPROTEINS

443

Whether, in these examples, both types of linkage occur in one polypeptide chain is at present undecided, but the finding, in normal urine, of glycopeptides that, on acid hydrolysis, give rise to 4-hydroxyL-proline and other amino acids, as well as to D-galactose, 2-amino-2deoxy-D-glucose, 2-amino-2-deoxy-~-galactose, and smaller proportions of L-fucose,218might suggest that collagen contains types of carbohydrate-peptide linkage other than that involving 5-hydroxy-Llysine, assuming that these glycopeptides are catabolites of collagen or basement membranes, or both.

IV. HETEROGENEITY IN GLYCOPROTEINS Much of the early work was performed on glycopeptides that, as is now realized, were heterogeneous with regard to their carbohydrate moieties. Such heterogeneity is not surprising in view of the mechanism of biosynthesis of the prosthetic groups, and there is the further complication that our knowledge of the catabolism of these substances is still scanty. Heterogeneity of this kind is probably widespread, and it is likely that the separation of glycopeptides containing carbohydrate moieties that are structurally very closely related may require highly refined techniques. Differences in the structure of glycoproteins can manifest themselves in a number of ways. Perhaps the first point to be emphasized is that the nature of the polypeptide chain to which oligosaccharide moieties become attached is not always the same. This is a situation that we have already discussed in connection with certain of the glycosaminoglycans (see p. 440). Differences between proteins occur with respect to the attachment of carbohydrate; a given amino acid may be present either unsubstituted or glycosylated. Ribonucleases A and B appear to differ solely by virtue of the latter having an oligosaccharide moiety attached through L-asparagine residue number 34 from the N-terminal end.lZ9 Only about 35% of the heavy chains of rabbit yG immunoglobulin occur as a form in which an L-threonine residue is substituted by a glycosy1 group,2oSuch examples are not likely to be the only ones. Analogous situations are known wherein a potential amino acid acceptor for entities other than carbohydrate in a protein does not appear to be present, in at least some molecules, in a modified form. Thus, 5hydroxy-L-lysine and L-lysine residues occupy equivalent positions in the amino acid sequence219in the a, chains of collagen, as do L(218)A.Bourillon and J. L. Vernay, Biochim. Biophys. Acta, 117,319(1966). (219)E.J. Miller, J. M. Lane, and K. A. Piez, Biochemistry, 8,30(1969).

444

R. D. MARSHALL AND A. NEUBERGER

proline and hydroxy-L-proline.220The same applies to L-lysine and 6-N-acetyl-~-lysinein calf-thymus histoneZz1and L-serine and 0phosphono-L-serine in egg albumin and other proteins.222The finding that collagens can be glycosylated, by uridine 5’-(D-galaCtOpyranOSyl pyrophosphate) and an enzyme isolated from rat kidney, may well be an indication that a similar situation exists with regard to the glycosylation of the 5-hydroxy-~-lysineresidues of collagen.223The significance of these findings, either in chemical or biological terms, is not as yet understood. Another type of variation is, perhaps, best illustrated by the results of studies with various types of collagen.224The carbohydrate moiety occurs, in part, as 2-O-a-D-glUCOpyranOSyl-o-~-D-galaCtOpyranosyl and, in part, as O-P-D-galaCtOpyranOSyl groups, that is, both with and without addition of the terminal D-glucopyranosyl groups. Similar variations in structure occur in corneal g l y c o p r o t e i n ~ . ~ ~ ~ It is not unreasonable to suppose that, at least in part, these two (closely related) types of carbohydrate moiety are attached to 5hydroxy-L-lysine residues that occur in the same relative position in the peptide chain. In other glycoproteins, a given carbohydrate moiety may, in some molecules, be present in molecules in which one or more terminal sialic acid residues are present, and, in others, in which no sialic acid occurs at all. As examples of this situation , ~ ~ ~46, ~~ and a may be mentioned a mouse Bence-Jones p r ~ t e i n MOPC yA myeloma ~ r 0 t e i n . lIt~ ~is not unlikely that ox pancreatic ribonucleasess B, C, and D are further examples. Variations of this type in the structure of glycoproteins have been described as peripheral heterogeneity.172 Analogous variations in the structure of various oligosaccharides in other glycoproteins almost certainly occur; these include f e t ~ i n transferrin,226 ,~~~ o r o s o r n ~ c o i d ,/3,-glycoprotein,22s ~~~ and zinc az-glycoprotein.229The extents of their heterogeneity, as revealed by electrophoresis in poly(acry1amide) gels, were in all (220) P. Bomstein, J. Biol. Chem., 242,2572 (1967). (221) R. J. DeLange, D. M . Famborough, E. L. Smith, and J. Bonner, J. Biol. Chem., 244,319 (1969). (222) G. E. Perlmann, “Phosphorus Metabolism,” W. D. McElroy and B. Glass, eds., Johns Hopkins Press, Baltimore, Md., 1952, Vol. 2, p. 167. (223) R. G. Spiro and M. J. Spiro, Fed. Proc., 27,345 (1968). (224) R. G. Spiro,]. Bid. Chem.,244,602 (1969). (225) Y. Oshiro and E. H. Eylar, Arch. Biochem. Biophys., 127,476 (1968). (226) S.-H.Chen and H. E. Eldon, Genetics, 56,425 (1967). (227) K. Schmid, J. P. Binette, S. Kamiyama, V. Pfister, and S. Takahashi, Biochemistry, 1,959 (1962). (228) H. E. Schultze, K. Heide, and H. Haupt, Naturwissenschu~fer1,48,719(1961). (229) K. Schmid and S. Takahashi,Nature, 203,407 (1962).

STRUCTURE AND METABOLISM OF GLYCOPROTEINS

445

instances markedly decreased after removal, by treatment with neuraminidase, of their sialic acid residues. A similar structural variation may be present in human-serum alkaline p h o s p h a t a ~ e . ~ ~ ~ Related examples in which oligosaccharides are present in glycoproteins, either with or without terminal glycosyl groups (other than the charged sialic acid residues), are known. Differences of this kind occur in the blood-group substances (see p. 452). The presence of oligosaccharide moieties, attached at a specific position on the polypeptide chain of a yA myeloma protein, has been described; some contained nonreducing, terminal 2-acetamido-2-deoxy-~-glucoseand L-fucose residues, and some did There are conflicting reports concerning the structure of the oligosaccharide prosthetic group of ox residues ribonuclease B. None of the 2-acetamido-2-deoxy-~-g~ucose in a glycopeptide isolated from the protein could be oxidized with periodate,130 although other workersz3’ reported that 2-acetamido-2deoxy-D-glucose can be removed from a nonreducing, terminal position of other preparations of ribonuclease B by use of a highly (2-acetamido-2-deoxypurified 2-acetamido-2-deoxy-~-~-glucosidase @-D-ghcoside2-acetamido-2-deoxy-~-~-glucohydrolase, E.C. 3.2.1.30) from Phaseolus vulgaris. It seems possible that the preparations of ribonuclease B differed with respect to the absence or presence of a nonreducing, terminal 2-acetamido-2-deoxy-~-glucoseresidue. In addition to heterogeneity of this type, which is primarily concerned with the nonreducing, terminal sugar residue(s), there may be a more complicated variation in the structure of the carbohydrate moieties of glycoproteins. This is best illustrated by considering the single oligosaccharide unit that occurs in hen’s-egg albumin. Although a number of physical tests failed to reveal any heterogeneity in the carbohydrate moiety of glycopeptide isolated from this protein,z3zthe glycopeptide could be fractionated by chromatography on DowexdO X2 in buffers of low ionic strength. The fractions thus obtained had various contents of D-mannOSe and 2-amino-2-deoxy-~glucose.z33-z3sIt was further shown that the differences found were (230) J. C. Robinson and J. E. Pierce, Nature, 204,472 (1964). (231) 0. P. Bahl and K. M. L. Agrawal,]. BioE. Chem.,243,98 (1969). (232) A. Neuberger and R. D. Marshall, in Ref. 6, p. 306. (233) L. W. Cunningham, J. D. Ford, and J. M. Rainey, Biochim. Biophys. Acta, 101, 233 (1965). (234) V. D. Bhoyroo and R. D. Marshall, Biochem.].,9 7 , l l p (1965). (235) L. W. Cunningham, in “Biochemistry of Glycoproteins and Related Substances,” E. Rossi and E. Stoll, eds., S. Karger, New York, 1968, Part 11, p. 141. (236) C. C. Huang and R. Montgomery, Biochem. Biophys. Res. Commun., 37, 94 (1969).

446

R. D. MARSHALL AND A. NEUBERGER

unlikely to be due to genetic variation.z35Treatment of similar material with a molecular sieve, Sephadex G-25, also gave rise to a number of glycopeptide fractions containing differing proportions of sugars.237 A number of other proteins are now known to exhibit heterogeneity with respect to their carbohydrate content; these include pig panyG immunoglobulin, creatic ribonuclea~e,~ rabbit19 and human109*238 and the c e r u l o p l a ~ r n i n ,2-acetamido-2-deoxy-j?-~-glucosidase,~~~ ~~~ blood-group substances from ovarian ~ y s t s . ' ~ ~ * ~ ~ ' In summary, a polypeptide chain may have one or more oligosaccharide units attached to it, although, in some, only a portion of an amino acid residue that is a potential acceptor-site for a sugar actually becomes glycosylated. The size of the prosthetic group attached at different points along the polypeptide chain may differ radically, as in calf thyroglob~lin,2~~ which is reported to contain nine carbohydrate units consisting of five residues of D-mannose and one of 2-amino-2deoxy-D-glucose, and 14 very much larger units, per molecule of protein (1 mole = 670,000 g). Structural features of this type have also been shown to occur in ox-aorta glycoprotein218and o v o m ~ c o i d . ~ ~ ~ This type of structural feature was described as central heterogeneity,172a term that might better be reserved for such situations as that already described for egg albumin, in which structural variations occur within a carbohydrate moiety attached at a specific position of the polypeptide chain. It is not unreasonable to suppose that the last-mentioned form of variation in the structure of glycoproteins may be widespread, and, as already discussed, more-sensitive fractionation procedures must probably be applied to these materials. It was found that glycoprotein prepared from sheep submaxillarygland182,244 contains, as prosthetic groups, almost exclusively a large number of disaccharide units, namely, 2-acetamido-2-deoxy-6-0sia~y~-a-D-ga~actosy1 groups. This result might suggest that substitution by sialic acid residues is, in all cases, complete in z)iz)o,but it is possible that the properties of those molecules in which this reaction is incomplete would be sufficiently different as to lead to their removal during purification procedures. Other variations in structure may also (237) G. A. Levvy, J. Conchie, and A. J. Hay, Biochim. Biophys. Acta, 130,150 (1966). (238) J. L. Fahey and A. P. Horbett, J. Biol. Chem., 234,2645 (1959). (239) G. A. Jamieson,J. Biol. Chem.,240,2019(1965). (240) D. Robinson and J. L. Stirling, Biochem.J.,107,321 (1968). (241) W. M. Watkins, Science, 152,172 (1966). (242) R. G . Spiro,J. Biol. Chern.,240,1603 (1965). (243) R. Montgomery and Y.-C. Wu,J. Biol. Chem., 238,3547 (1963). (244) V. L. N. Murty and M. I. Horowitz, Carbohyd. Res., 6,266 (1968).

STRUCTURE AND METABOLISM OF GLYCOPROTEINS

441

occur, in view of the presence of other types of sugars, even in those molecules in which the disaccharide units are all complete, and may be responsible, at least in part, for the heterogeneity found by immunological methods in similar protein preparations from cattle.245 The analogous protein from pig has a more complicated structure, the oligosaccharide groups varying from a single 2-acetamido-2-deoxy-~galactose residue to a pentasaccharide. Products, obtained by reduction in alkali, that are believed to correspond with the carbohydrate moieties existing in the protein will be described (see p. 458).

V. THESIZE OF

THE

CARBOHYDRATE MOIETIES IN GLYCOPROTEINS

A number of measurements have been made of the molecular weights of glycopeptides isolated from glycoproteins. Often, the determinations may have been performed on material that was not homogeneous. To illustrate the sort of problems that may arise, the studies of Clamp and Putnam1O9may be considered. The sedimenting patterns obtained with glycopeptides isolated from human yG globulin appeared to indicate that the substance was, from the point of view of size, fairly homogeneous. However, application of the mathematical methods of Trautman and C r a m p t ~ to n ~the ~ ~data led to the deduction that t w o species, having molecular weights of 1170 and 3150, respectively, were present. On the other hand, the related procedure of Klainer and K e g e l e ~appeared ~~~ to indicate that a species of only one size was present and that this had a molecular weight of 2,200. More-varied analytical procedures for sedimentation data obtained with other glycopeptides might indicate wider heterogeneity with regard to size than has thus far been reported. Despite all the problems of heterogeneity with respect to the carbohydrate moieties mentioned, certain statements about structural features can be made. At a given position in the polypeptide chain, the prosthetic group may differ widely from one such group to another. It may consist of a structure as simple as a single sugar residue, as in pig submaxillary-gland glycoprotein, or a disaccharide, as in submaxillary-gland glycoproteins or collagen, or the moiety may consist of -12 to 15 sugar residues. Some of the relevant data are summarized in Table IX, where the methods often employed for de(245)M. I. Horowitz, L. Martinez, and V. L. N. Murty, Biochirn. Biophys. Acta, 83, 305 ( 1964). (246) R. Trautman and C. F. Crampton,j. Amer. Chem. SOC., 81,4036 (1959). (247) S . M. Klainer and G. Kegeles, j . Phys. Chern., 59,952 (1955).

TABLEIX Molecular Weights of Glycopeptides Isolated from Various Glycoproteins Glycomotein" Fetuin (3) Orosomucoid (11)*=

Guinea-pig a,-acid glycoprotein

Human yG globulin

yA Myeloma protein (4) & Globulin Ceruloplasmin (10) Sheep ICSH Human chorionic gonadotropin (2) Calf thyroglobulin (23)

Mol.wt. 4300 4500 2800 2540 2840 2600 1800 4120 4260 1805 2100 3070 2700 2450 3Ooob 22w 28W 3305 1950 3160 3300 4300 1250 4100

Method employed sediment. equil. sediment. rate and viscosity sediment. rate vapor phase osmometry trinitrobenzenesulfony1 subst. sediment. equil. sediment. equil.

Relevant data uW

(determined) 0.664 ml/g

[qlup (determined) 5.83 ml/g

(assumed) 0.664 ml/g u (calculated)0.614 ml/g -

u (assumed) u (assumed) 0.644 mllg

)sediment. equil. -

sediment. equil. gel filtration '(Sephadex G 25) sediment. equil. sediment. rate Archibald method freezing-pt. osmometry sediment. equil. Archibald method dinitrophenyl substitution gel filtration (Sephadex G 50) sediment. rate

u (assumed) 0.664 ml/g

-

u (assumed) 0.7 ml/g

Total amino acid Referpresent (moles) ences contains 77% of carbohydrate

248

2.0 2.1 2.1 1.07 1.04 6.8 6.5 n.d. 1.4 3.7 1.4 1.4 5 3.16 1

249 250 251 252

)2% 21 109 255

-

u (assumed) 0.664 ml/g

v (determined) 0.78 ml/g

-

u (calculated)

0.634 ml/g

n.d. 3.22 9 9.7 13.1 2.2

253 239 257 214 242

“Egg white” (mainly ovomucoid) Ovomucoid (3)

Egg albumin (1)

Urinary T-H protein Ox aorta glycoprotein Deoxycholate soluble fraction of rat-liver microsomes Earthworm-cuticle collagen Soybean hemagglutinin (1) Chondroitin 4sulfate’* (ox nasal-septa)

Chondroitin 4-sulfate (pig laryngeal cartilage)

-2000

diffusion rate

258 -

2500 2840 2580 3000 2800 1880 1580 1570 1500 1560 -4000 2300 1500 2950 2360

di- and tri-Dgalactosides 4600 1500 15300 18300 21700 29800 41300 150008*h.‘

sediment. equil.

v

(assumed) 0.664 ml/g

gel filtration (Sephadex G 25) phenylthiohydantoin subst. dinitrophenyl substitution gel filtration (Sephadex G 25) electrothermal micromethod benzyloxycarbonyl substitution isopiestic method thermoelectric osmometry analysis and rate of dialysis

1.09 1.09 2 1 1 1 1

analytical data analytical data gel filtration (Biogel P-2) -

u (calculated)0.615 ml/g

sedimentation rate and sedimentation equil.

254 243 254 259 260 261 262 263

n.d. 6.5 5.7 7.1

28

nonee

265

264

v, 4

c: 5e n

2 P

$ U

g R

4

9m

8 E ~

1

0

analytical data, based on Dxylose/2-amino-2-deoxy-Dgalactose ratio and on the (continued)

$ a

TABLEI X (continued) ~~~

Glycoprotein" Chondroitin 4-sulfate (pig costalcartilage) Dermatan sulfate (pig skin)

Mol.wt.

Method employed

1430OgAJ

pelimination reaction sedimentation equilib.

27WhJ

sedimentation equilib.

Relevant data (assumed) 0.57 ml/gz69

~

~

Total amino acid Referpresent (moles) ences

0

205 268

-

u (determined) 0.57 ml/g

6% of amino acids 269

"The number of carbohydrate residues per molecule of glycoprotein is indicated in parentheses. bThe material used in this study was heterogeneous with respect to size, and only an order of magnitude is indicated. "See the comments on p. 447. dGel-filtration on Sephadex G-25 and G-50 showed that glycopeptides from a yA myeloma protein had molecular weights of 1,900 and 2,600, respectively. Glycopeptides from a yM globulin had= molecular weights of 1500 and 1730. 'The materials isolated were prepared by treating the protein with aIkaline (0.05M sodium hydroxide) borohydride (0.15 M) for 24 hours at 30", but the oligosaccharides are likely to exist in the protein in units of about the size given.28sThe saccharides are composed of o l - ~ - ( l +2)-linked D-galactose residueszGa'The values are for various fractions isolated by Scott's procedure.46 The unfractionated material hadzwM , 25,300 and M . 20,800. Weightaverage molecular weights of 21,700 and 21,100 were found for materials prepared from digests with (0.5 M sodium hydroxide for 19 hours at 4") and papain, respectively.267'As used here, the term chondroitin sulfate refers to the carbohydrate chain only (not to its complex with protein). "The carbohydrate chains are heterogeneous with respect to molecular size. 'A number-average value. 'A weightaverage value.

?

z

m

C

STRUCTURE AND METABOLISM OF CLYCOPROTEINS

451

termining the molecular weights are also described. At present, it is difficult to set a precise upper limit to the size of a carbohydrate unit, but, in general, the molecular weight would appear not to exceed -3,000 in those moieties in which there is not, so far as we know, a repeating pattern. However, in soybean hemagglutinin, the value would appear3s to be of the order of -4,500,and the single carbohydrate moiety in this glycoprotein consists of approximately 25 residues of a mannose and 5 of 2-amino-2-deoxy-~-g~ucose. The carbohydrate chains of the chondroitin sulfate type would seem to have rather larger molecular weights, up to values of the order of 40,000 daltons. Here again, there is considerable heterogeneity with respect to size (see Table IX). The size of the carbohydrate chains of keratan sulfate, having repeating units of (1+3)-0fi-D-galactopyranosyl-( 1+ 4)-2-acetamido-2-deoxy-fi-~-glucopyranosyl 6-sulfate (see p. 459),is unknown. Ox-corneal keratan sulfate was separatedZ7Ointo a number of fractions having molecular weights of 4,100 (F = 0.55 ml/g), 8,800(0.53ml/g), 10,000 (0.51ml/g), 12,500 (248) R. G. Spiro, J. Biol. Chem.,237,646 (1962). (249) N. Ui and 0.Tarutami,J. Biochem. (Tokyo), 51,370 (1962). (250) P. V. Wagh, I. Bomstein, and R. J. Winzler,]. Biol. Chem.,244,658 (1969). (251) S. Kamiyama and K. Schmid, Biochim. Biophys. Acta, 58,80 (1962). (252) E. H. Eylar, Biochem. Biophys. Aes. Commun., 8,195 (1962). (253) W. E. Marshall and J.Porath,J. Biol.Chem.,240,209 (1965). (254) W. L. Cunningham and J. L. Simkin, Biochem. J., 99,434 (1966). (255) G. Dawson and J. R. Clamp, Biochem. Biophys. Res. Commun., 26,349 (1967). (256) T. Bhatti and J . R. Clamp, Biochim. Biophys. Acta, 170,206 (1968). (257) H. Papkoff, Biochim. Biophys. Acta, 78,384 (1963). (258) P. A. Levene and A. Rothen,J. Biol.Chem., 84,63 (1929). (259) E. D. Kaverzneva and V. P. Bogdanov, Biokhimiya, 27,273 (1962). (260) A. P. Fletcher, R. D. Marshall, and A. Neuberger, Biochim. Biophys. Acta, 71, 505 (1963). (261) P. G. Johansen, R. D. Marshall, and A. Neuberger, Biochem.J.,78,518 (1961). ~ (262) Y. C. Lee, Y.-C. Wu, and R. Montgomery, Blochem.]., 9 1 , 9 (1964). (263) A. Gottschalk, in “Chemistry and Biology of Mucopolysaccharides,” G. E. W. Wolstenholme and M. O’Connor, eds., J. & A. Churchill, Ltd., London, 1958, p. 287. (264) Y.-T. Li, S.-C. Li, and M. R.Shetlar,]. Biol. Chem.,243,656 (1968). (265) Y. C. Lee and D. LangJ. Biol.Chem., 243,677 (1968). (265a)L.Muir and Y. C. LeeJ. Biol. Chem., 244,2343 (1969). (266) A. Wasteson, Biochim. Biophys. Acta, 177, 152 (1969). (267) M. Luscombe and C. F. Phelps, Biochem. J . , 103,103 (1967). (268) E. Marler and E. A. Davidson, Proc. Nut. Acad. Sci. U . S., 54,648 (1965). (269) C. Tanford, E. Marler, E. Jury, and E. A. Davidson, /. Biol. Chem., 239, 3034 (1964). (270) T. C. Laurent and A. Anseth, Exp. Eye Res., 1,99 (1961).

452

R. D. MARSHALL AND A. NEUBERGER

(0.49 ml/g), 13,700 (0.52 ml/g), 17,300 (0.49 ml/g) and 19,100 daltons (0.47 ml/g). Some or all of these fractions may have contained more than one carbohydrate chain linked together by short lengths of peptide, but it is clear that the smallest chain of keratan sulfate has a molecular weight that does not exceed -4,000. Keratan sulfates from other sources have number-average molecular weights of 11,500 (from bull-shark cartilage), 10,000 (human cartilage), 8,900 (calf cornea), and 9,700 (chicken cornea).211 The factors responsible for limiting the size of the carbohydrate moieties are as yet unknown. Speculations may be made as to whether, in the prosthetic group itself, certain structural features, such as terminal fucosyl, sialyl, a-D-galactosyl or 2-acetamido-2-deoxy-cu-~-galactosyl residues may, in appropriate cases, play a role here. However, kinetic factors, and the distribution and specificities of the various activated sugar transferases during passage of the nascent protein through the cysternae of the endoplasmic reticulum, may be of prime importance.

VI. FEATURES OF THE STRUCTURE OF THE CARBOHYDRATE MOIETIES OF SOME GLYCOPROTEINS Many methods are available for determination of the structure of carbohydrate moieties. One method involves periodate oxidation, succeeded by reduction of the hemialdal groups formed. The resulting polyalcohols are hydrolyzed by acid, and, from the nature of the products, it may be possible to deduce a structure for the carbohydrate under considerati~n.~'~ However, problems of interpretation are likely to arise if the prosthetic group is markedly heterogeneous. It is possible that, so far, the most valuable data concerned with the structure of the carbohydrate moieties (including the anomeric configuration of glycosyl groups) of many glycoproteins have been obtained by techniques involving partial hydrolysis.

1. Blood-group Substances from Ovarian Cysts Some of the results determined in this way with blood-group substances are described in Table X. A number of other oligosaccharides

(271) M. B. Mathews and J . A. Cifonelli,]. Biol.Clzem., 240,4140 (1965). (272) M. Abdel-Akher, J. K. Hamilton, R. Montgomery, and F. Smith, J . Amer. Chem. SOC.,74,4970 (1952).

T A ~ LX E Oligosaccharides Isolated from Partial Hydrolyzates of Certain Blood-group Substances Oligosaccharide isolated 1" P-D-Gd-( 1

-+

Source ovarian cyst Leo substance

3)-D-GNAc 4

t

2

p-D-Gd-( 1

3

~-L-Fuc-(1

4

-+

-+

Hydrolysis conditions triethylamine (2.5%) in 50% aqueous methanol; 60"; 18hr

References 273

2 2 C

n 4

C

1 a-L-Fuc 4)-D-GNAc-

2

m triethylamine (2.5%) in 50% aqueous methanol; 60"; 18hr triethylamine (2.5%) in 50% ovarian cyst H substance aqueous methanol; 60"; 18 hr ovarian cyst A and B substances triethylamine (2.5%) in 50% aqueous methanol; 60"; 18hr 0.2 M NaOH; 0.25 M NaBH,; ovarian cyst A,B and H room temp.; 7 days substances triethylamine (2.5%) in 50% ovarian cyst H substance aqueous methanol; 60"; 18hr ovarian cyst A and B substances triethylamine (2.5%) in 50% aqueous methanol; 60"; 18hr triethylamine (2.5%) in 50% ovarian cyst A substance aqueous methanol; 60"; 18hr ovarian cyst Lea substance

2)-D-Gal

~ - L - F u c - -+ ( ~ 2)-~-Gal-(1 4)-D-GNAc -+

1-+ 4)-D-CNAc 5b a-D-CalNAc-(1 -+ 3)-p-~-Gd-(

2

T

a-L-FUC 6C.d a-D-Gd-( 1-+ 3)-P-D-Gd-(1 4)-D-GNAc 2 -+

t

ovarian cyst B substance

triethylamine (2.5%) in 50% aqueous methanol; 60";18 hr

273 274 275 276 274 275 275

275

1

~-L-Fuc (continued)

TABLEX (continued) Source

Oligosaccharide isolated 7

1 + 3)-D-Gd p-~-Gal-(l-+~)-P-D-GNAc-( 2 4

t

8

ovarian cyst Lebsubstance

T

1 a-L-Fuc a-L-Fuc P-~-Gal-(l-+ 3)+3-~-GNAc-(l + 6)-~-GalacitolNAc ovarian cyst Lea substance 3

Hydrolysis conditions water-soluble, alkaline resin, pH 8.5; 100"; 20 min

+

References 277

0.2 M NaOD 0.25 M NaBD, in D,O; room temp.; 7 days

179

0.2 M NaOD +0.25 M NaBD, in D,O; room temp.; 7 days

179

t

1 p-~-Gal 9

p-~-Gal-(l+ ~)-P-D-GNAc 1

ovarian cyst Lea substance

.1

6 Galactitol 3

?

t

C

z

m

1 10

p-~-Gal-(l-+ 3)-P-D-GNAc D-GdacitolNAc

11 p-~-Gal-(1 --* 4)-P-D-GNAc-(1 + 3)-/3-D-Gd3

ovarian cyst Lea substance ovarian cyst Lea substance

0.2 M NaOD -k 0.25 M NaBD, in D20;room temp.; 7 days water-soluble, alkaline resin; pR 8.6; loo"

278

t

1 a-L-Fuc 12

p-~-Gal-(1-+ ~)-P-D-GNAC-( 1+ 6)-p-D-Gal

ovarian cyst Lea substance

water-soluble, alkaline resin; pH 8.6; 100"

278

15

a-~-GalNAc-(l+ 3)-P-~-Gal-(1+ 3)-&GNAc a-D-GalNAc-(1+ 3)-P-~-Gal-(1 + 4)-D-GNAc a-D-Gal-(1 + 3)-/3-D-Gal-(1 + 3)-D-GNAc

ovarian cyst A substance ovarian cyst A substance ovarian cyst B substance

16

a-D-Gal-(1 + 3)-p-D-Gal-(1+ ~)-D-GNAc

ovarian cyst B substance

17

p-D-Gal-(l

13 14

18

19 20

+ 3)-@-D-CNAc-(l+ 3)-D-Gd

ovarian cyst A,B,H, and Lea substances p-D-Gd-( 1 + 4)-P-D-GNAc-(1 + 3)-D-Gal ovarian cyst A,B,H,Lea substances 1 -+ 3)-p-D-Gal-(1 + 3)-~-GalNAc ovarian cyst A,B,H,Lea ~-D-G-NAc-( substances a - ~ - G a l N A c1(-+ 3)-P-D-Gal-(1 3)-P-~-GaladitolNAcpig submaxillary gland 2 6 glycoprotein -+

t

21

+

t

2 CX-L-FUC ff -D-sidyl 1+ 2)-P-D-Gal-(1-+4)-P-D-GalactitolNAc CY-L-FUC-( 1

0.1 M HCl; 100"; 30 min 279,280 0.1 M HCI; 100"; 30min 200 0.04 N poly(styrenesu1fonic acid); 281 87" 0.04N poly(styrenesu1fonic acid); 281 87" 0.04 N poly(styrenesu1fonic acid); 282 90";4 hr 0.02 N (polystyrenesulfonicacid); 282 90"; 4 hr 0.02 N (polystyrenesulfonicacid); 282 90"; 4 hr 177 0.05 M KOH 1.0 M NaBH,; 45"; 15h r

v)

2

5

2 n z

* U

5

*m 4

? pig submaxillary gland glycoprotein

0.1 M NaOH +0.3 M NaBH,; 28"; 120 hr

283

"A related trisaccharide probably also occurs at the nonreducing terminal of some of the oligosaccharide side-chains of Lea substance, in which the positions of the two terminal sugar residues are interchanged; this structure is almost devoid of activity as a determinant bOligosaccharides of a clearly g r ~ u p . ' ~This g compound may be related to the tetramer isolated by Marr and coworkers (see item 11).278 related type have been isolated from hydrolyzates of A substance (0.2 M NaOH at room temperature for 7 days) in the presence of 0.25 M NaBH,: they were blocked at their reducing end by a hex-3-ene-1,2,5,6-tetrol. Both mono-L-fucosyl and di-L-fucosyl oligosaccharides were described, the latter having a second L-fucose residue attached at the 2-acetamido-2-deoxy-D-glucose residue, but not276at 04. COligosaccharides analogous to those described in footnote b were isolated from B substance.276dOligosaccharides analogous to those described in footnote b were isolated from H substance,*" but there was no nonreducing, terminal a-D-galactosyl group.

-

F 30

5fl z

v)

456

R. D. MARSHALL AND A. NEUBERGER

that could be related biogenetically to those described therein have also been isolated. Thus, such structures as P-D-GNAc-(1 --* 6)-~-Gal 3

t

P-D-GNAc-~ or even simpler related ones, had been isolated some years earlier by Y o ~ i z a w a ~from * ~ hydrazinolyzates of blood-group substance A obtained from pig gastric mucus; they may have arisen from a region similar to that from which the oligosaccharide shown as item 9 in Table X was derived. Alkaline hydrolysis is particularly valuable in obtaining oligosaccharides, because the compounds isolated must have arisen from preliminary cleavage of the carbohydrate-peptide bond, followed by breakdown, to a greater or lesser extent, from the reducing, terminal end thereby formed. The oligosaccharides isolated must, therefore, have occurred at nonreducing, terminal positions in the carbohydrate moieties of the glycoprotein or glycopeptide. The oligosaccharide isolated may, in some instances, have a modified, reducing terminus, because of the action of the alkali. Some of the work with blood-group substances is of particular interest in this regard. First, oligosaccharide 1 (see Table X) represents that structure shown by serological tests285to be largely responsible for Lea antigenic activity. Both this structure and the disaccharide N-acetyl-

(273) V. T. Rege, T. J. Painter, W. M. Watkins, and W. T. J. Morgan, Nature, 204, 740 (1964). (274) V. T. Rege, T. J. Painter, W. M. Watkins, and W. T. J. Morgan, Nature, 203, 360 (1964). (275) T. J. Painter, W. M. Watkins, and W. T. J. Morgan, Nature, 206,594 (1965). (276) K. 0.Lloyd, E. A. Kabat, E. J. Layug, and F. Gruezo, Biochemistry, 5,1489 (1966). (277) A. M. S. Marr, A. S . R. Donald, W. M. Watkins, and W. T. J. Morgan, Nature, 215, 1345 (1967). (278) A. M. S. Marr, A. S. R. Donald, and W. T. J. Morgan, Biochem.]., 110,789 (1968). (279) G. SchifFman, E. A. Kabat, and S. Leskowitz, J. Amer. Chem. Soc., 84.73 (1962). (280) I. A. F. L. Cheese and W. T. J. Morgan, Nature, 191,149 (1961). (281) T. J. Painter, W. T. J. Morgan, and W. M. Watkins, Nature, 199,282 (1963). (282) V. P. Rege, T. J. Painter,-W. M. Watkins, and W. T. J. Morgan, Nature, 200,532 11963). (283) k. L. katzman and E. H.Eylar, Biochem. Biophys. Res. Commun., 23,769 (1966). (284) Z. YosizawaJ. Blochem. (Tokyo), 51,145 (1962). (285) W. T. Watkins and W. T. J. Morgan, Nature, 180,1038 (1967).

STRUCTURE AND METABOLISM OF GLYCOPROTEINS

457

lactosamine (item 2) occur in nonreducing, terminal positions in the various carbohydrate moieties of Lea substance. Clearly, the carbohydrate moieties are independently formed, and neither of them is likely to be the precursor of the other. The second of these structures is believed to be the antigenic determinant responsible for interaction of the macromolecule with antibody produced in the horse against the capsular polysaccharide of Type XIV pneumococcus.285,286 Probably, the N-acetyl-lactosamine is usually linked to D-galactose, and this sugar may have different degrees of substitution and different positions of the substituents. Substituents on 0-4of D-galactose may occur, but direct evidence for such substitution is 1 a ~ k i n g .Other l~~ structures are also situated at the nonreducing, terminal ends of some of the carbohydrate side-chains (see, for example, items 8-12, footnote a of Table X). One of these (item 8) has been suggested as being derived from the region through which linkage to the polypeptide chain occurs, although decisive evidence is not available.179 It has been found that some of the carbohydrate moieties of A and B substances do not contain the 2-acetamido-2-deoxy-a-~-galactose or a-D-galactose residues, respectively, that are essential features of these antigenically determinant sites (see items 3 and 4). Furthermore, it is clear that, unless a very high degree of branching occurs, the antigenic determinants for A and B activity (items 5 and 6) cannot be present on the same carbohydrate moieties; both activities are probably present in the same molecule. The Leb antigenic determinant (item 7) has structural features in common with both H (items 3 and 4)and Lea determinants. The juxtaposition of the two L-fucose residues in the Leb-active carbohydrate moieties sufficiently affects the general stereochemistry of the group to make it a completely new antigenic determinant, quite distinct from those for either H or Lea substance. The Leb characteristic of blood-group substance seems always to occur with one or more other such activities as A, B, and H,2x7so that it is likely that there is considerable variation in the structure of various oligosaccharides in the same molecule. Other oligosaccharides obtained from ovarian-cyst blood-group substances indicate that these macromolecules have many features in common (items 17-19) between the various types; but, as they have been isolated after partial hydrolysis with acid, it is, in general, not possible, from this information alone, to assign a position to them within the carbohydrate moieties. Those items represented by 13, (286) P. Z. Allen and E. A. Kabat,J. Imtnunol., 82,340(1959). (287) W. T. Watkins, Science, 152, 172 (1966).

458

R. D . MARSHALL A N D A. NEUBERCER

14 (from A substance) and those by 15,16 (from B substance) are likely to arise from reducing, terminal regions of carbohydrate chains of A and B substances, respectively: numbers 13 and 15 are of the structural type designated type I, in which the 2-acetamido-2-deoxyD-glucose, as the reducing terminal sugar, is linked by 0-3, and numbers 14 and 16 are of type 11, in which 0 - 4 of this corresponding sugar is substituted.282Structural features of this type may form in the macromolecular part of such branched moieties as that represented by item 9. Also, it is not yet known whether such structures as those represented by 13-16 exist as such, as part of the macromolecule (that is, without any terminal L-fucose residues). Oligosaccharides of the type represented by items 5 and 6 would be expected to lose L-fucose residues relatively easily under acidic conditions.288 Serological and enzymic techniques involving various glycosidases have been used for determining certain structural features of the secreted blood-group substances. Reviews of these data are available,241*285 and so these aspects will not be discussed. This very brief survey of some of the extensive and elegant studies made over many years, mainly by Morgan, Watkins, Kabat, and Yosizawa and their colleagues, has emphasized certain features of the chemistry of the secreted blood-group substances. The carbohydrate moieties in a given substance may differ quite markedly from one another, both in the nature of the nonreducing terminal residues and in the structures of the “internal” parts of these prosthetic groups. The sizes of the moieties probably differ, possibly quite extensively. If an attempt is made to construct a model in which are incorporated all of the various oligosaccharides listed in Table X, the resultant prosthetic groups would have to consist of more than 20 sugar residues, a somewhat unlikely possibility. It is more reasonable to suppose that these glycoproteins have a number of types of somewhat smaller carbohydrate moieties, in which certain variations of structure occur. It would be desirable to examine homogeneous preparations of glycopeptides from these sources, but the resistance of the glycoproteins to proteolytic digestion”’ makes this a difficult task at present. 2. Pig Submaxillary-gland Glycoprotein

A protein obtained from the submaxillary gland of pigs and having blood-group A activity was found to contain, as part of its carbo(288) A. Neubergerand R. D. Marshall, in Ref. 6, Chapter8

STRUCTURE AND METABOLISM OF GLYCOPROTEINS

459

hydrate moieties, the structure whose reduction product is given as item 20 in Table X. Closely related, but reduced, “incomplete” oligosaccharides in which (a)the terminal 2-acetamido-2-deoxy-cr-Dgalactosyl group and ( b )the nonreducing, terminal trisaccharide were missing from this structure have been found, as well as all of the analogous substances lacking a N-glycoloylneuraminic acid residue. In addition, a similar type of reduced oligosaccharide was isolated from the same source by Katzman and EylarZs3;in it, the @-D-galaCtOSe residue was linked to 0-4 of 2-acetamido-2-deoxy-~-galactitol(item 21). Some prosthetic groups consisted of monomeric 2-acetamido-2deoxy-D-galactose residue^.'^' The structures of these oligosaccharides seem to be simpler than some of those that occur in bloodgroup substances from ovarian cysts.

3. Glycosaminoglycans

The structures of the carbohydrate chains of the various glycosaminoglycans have been described in considerable detail,289-291a and therefore only a few limited aspects concerned with these glycoproteins will be discussed. Keratan sulfate has usually been obtained by treatment of extracts of cornea and of cartilage with proteolytic enzymes, followed by treatment with hyaluronidase and application of various fractionation procedures. The resulting material usually contains up to 5-1070of amino acids; these are probably liberated by hydrolysis of a polypeptide backbone to which carbohydrate branches may be linked. The term keratan sulfate is often applied to the carbohydrate moiety composed of a disaccharide repeating unit, namely, (1-3)-O-p-D-galactopyranosyl-(1~4)-2-acetamido-2-deoxy-~-~-glucopyranose, in some of which, both sugars are sulfated at C-6 (but far from completely in each type of sugar moiety). Other sugars are also often found in the substance isolated from various sources; these include sialic acid, L-fucose, and 2-amino-2-deoxy-~-galactosein relatively small proportions.207From methylation studies, it seems likely that L-fucose, as usual, occurs in a nonreducing, terminal position and, moreover, that, in keratan sul(289) H. Muir, Intern. Reo. Connect. Tissue Res., 4,lO (1964). (290) J. S. Brimacombe and J. M. Webber, “Mucopolysaccharides,” Elsevier Publishing Co., Amsterdam, 1964. (291) S. Schiller, Ann. Rev. Physiol., 28,137 (1966). (291a)C. Quintarelli, “Chemical Physiology of Mucopolysaccharides,”Little, Brown & Co., Boston (1968).

460

R. D . MARSHALL AND A. NEUBERGER

fate from rib cartilage, some D-galactose is likewise in a terminal, nonreducing p o s i t i ~ n .Keratan ~ ~ ~ * sulfate ~ ~ ~ from the latter source is relaand the extent to tively richer in 2-amin0-2-deoxy-D-galactose,~"~ which this sugar, as well as L-fucose and some of the D-galactose, occurs as carbohydrate moieties independent of the repeating disaccharide unit that makes up other prosthetic groups has not yet been fully established. As mentioned earlier (see p. 440), such units could be responsible for the cross-reactions with anti-blood-group sera that have been described.213The recognition of D-mannose (which is not released from keratan sulfate by treatment with alkali) indicates that there are further, unknown, structural features.292It has been suggested209that this sugar occurs in such a position that the side chains of keratan sulfate of human, knee-joint cartilage have the general structure [Gal- GNAc In- Man- GalNAc- Ser, but further studies on this aspect are desirable (see p. 440). The antigenic activity of the keratan sulfate-like, protein complex from chick allantoic fluid results mainly from the presence of relatively large proportions of prosthetic groups (hexasaccharides) that differ greatly from that usually described as keratan sulfate.293This complex is immunologically identical with a component of influenzavirus envelope.294 Chondroitin sulfate, heparin, and heparitin sulfate chains are linked to the polypeptide backbone by the following linear sequence of sugars:

1--* 3)-/3-D-Ga1-(1 + 3)-/3-D-Gal-(1+-4)-/3-~-Xyl-L-Ser, @-D-GUA-( where the anomeric carbon atom of D-xylose is involved in the carbohydrate-peptide bond. This structure was demonstrated for chondroitin 6-sulfate prepared from umbilical cord,'85 for chondroitin 4-sulfate from ox nasal septa,295for dermatan sulfate from pig skin2Mand for heparin from pig intestinal m u c ~ s a . ~The ~ ' sugar sequence in this It is * *of~interest ~~ region is probably identical in heparitin ~ u l f a t e . ~ @ (292) V. P. Bhavanandan and K. Meyer, J . Biol. Chem., 243,1052 (1968). (293) K. Meyer, N. Seno, and B. Anderson, Rheumatismus (Darmstadt),36,13 (1965). (294) W. G. Laver and R. G. Wehster, ViroEogy,30,104 (1966). (295) L. Roden and G. Armand,]. B i d . Chem., 244,65 (1966). (296) L. A. Fransson, Biochim. Biophys. Acta, 156,311 (1968). (297) U. Lindahl, Biochim. Biophys. Acta, 130,368 (1966). (298) J. Knecht, J. A. Cifonelli, and A. Dorfman,J. Biol. Chem., 242,4652 (1967). (1963). (299) S. Jacobs and H. Muir, Biochem.J.,87,37~

STRUCTURE AND METABOLISM O F GLYCOPROTEINS

461

TABLEXI The Main Structural Features of the Carbohydrate Chains of Glycosaminogl ycans Carbohydrate chain

Main disaccharide repeating unit

References

4404

I

Dermatan sulfate"

a-( 1 + 4)-p-~-IdoUA-( 1 + J)-D-GalNAc-

289- 29 l a

4-SO4

I Chondroitin 4-sulfate*

+-(l

+ ~)-P-D-GUA-(~ + 3)-~-GalNAc-

289-291a

640,

I

-p-(1 4 ~)-/~-D-GUA-( 1 + 3)-~-GalNAc-

289-291a

Heparind

-p-(1+ 4)-P-D-GUA-(l -+ 4)-D-GNSO,

297

Heparitin sulfate'

D-GUA-D-GNSO, D-GUA-D-GNAc

298

6-sulfatec

aIn dermatan sulfate from shark skin, there is about a 40% excess of sulfate groups over 2-acetamido-2-deoxy-~-galactose residues."o The small proportions of D-glucuronic acid (D-GUA) present in dermatan sulfate occur in regions of the carbohydrate chains both near to, and far from, the carbohydrate-peptide linkage region. The isolation of the tetrasaccharide so4

so4

I

I

GUA-CalNAc-IdoUA-GalNAc from dermatan sulfate from pig skin reveals that the carbohydrate chains are heterogeneous in at least some preparation^.^"' "Cartilage of squid and of horse-shoe crab contains chondroitin 4-sulfate, sulfated at 0 - 4 of 2acetamido-2-deoxy-D-galactoseresidues to the extent of about 75% only. Overall, there are about 1.5 molar equivalents of sulfate per mole of 2-acetamido-2-deoxyD-galactose."u' 'From shark skin may be obtained various fractions containing 1.0to 1.3 molar equivalents of sulfate per mole of 2-acetamido-2-deoxy-~-galactose.~~~ dThe 2amino-2-deoxy-D-glucose residues (in the repeating sequence of this sugar containing uronic acid in the region of the linkage region) is N-acetylated, not N - ~ u l f a t e d . ~ ~ ~ Small proportions of L-iduronic acid are also reported to be present,28' as well as some unsubstituted hexosamine residues.3a3As well as the N-sulfate groups, there are, on average, a further 1.5 sulfate groups per disaccharide repeating nit;^"^,'^ these are mainly on 0 - 2 of about one-third of the D-glucuronic acid residues and on 0-6 of 2-amin0-2-deoxy-D-glucose.~" Probably, branch points also occur in heparin from lungs.3os 'An alternating sequence of 2-acetamido-2-deoxy-D-g~ucose-uronic acid and N-sulfated 2-amino-2-deoxy-D-glucose-uronicacid moieties is not present; long sequences of alternate D-glucuronic acid and 2-acetamido-2-deoxy-~glucoseresidues are present.*% The material probably comprises a family of closely related compounds, and may be separated into fractions having various proportions of total sulfate, Nsulfate, and acetyl groups.3q

462

R. D. MARSHALL AND A. NEUBERGER

that, even in dermatan sulfate, in which the preponderant uronic acid is L-iduronic acid, it is D-glucuronic acid that occurs near the carbohydrate-peptide linkage region. Furthermore, the sugar attached by its anomeric carbon atom to the D-glucuronic acid residue in the linkage region of heparin is not the N-sulfate of 2-amino-2-deoxy-DThus, the main strucglucose, but 2-acetamido-2-deoxy-~-glucose.~~~ tural features of the carbohydrate chains of the glycosaminoglycans (see Table XI) do not extend to the carbohydrate-protein linkage area. The occurrence of heterogeneity in them, as exemplified by the small proportions of D-glucuronic acid (instead of L-iduronic acid) in dermatan sulfate, is a further aspect of the heterogeneity of glycoproteins . The carbohydrate chains of chondroitin sulfate are not antigenic, although antibodies can be raised in rabbits and guinea pigs against pig chondromucoprotein.307The nature of the site(s) against which antibodies are produced is as yet unknown.

4. Some Uses of Glycosidases

A large number of studies have been concerned with the structure of the carbohydrate moieties of glycoproteins, but these will not be listed, partly because of limitations of space. Also, there are discrepancies in some of the data obtained by different workers and in their interpretation. Such discrepancies may, at least in part, arise from problems of heterogeneity, as discussed in various parts of this Chapter. There are, however, a number of characteristics common to a large number of these prosthetic groups. Firstly, many of them contain p-D-galactose attached to 0 - 4 of 2-acetamido-2-deoxy-~-glucose: N-acetyl-lactosamine is a structural feature of blood-group substances

(300)N. Seno and K. Meyer, Biochim. Eiophys. Acta, 78,258(1963). (301)L.A. Fransson and L. Rod&, J . Biol. Chen., 242,4170(1967). (302)M. B. Mathews, J. Duh, and D. Person, Nature, 193,378(1962). (303)G.J. Durant, H. R. Hendrickson, and R. Montgomery, Arch. Eiochem. Eiophys., 99,418(1962). (304) I. Danishefsky, H. Steiner, A. Bella, and A. Friedlander, J . Biol. Chem., 244, 1741 (1969). (305)J. A.Cifonelli, Fed. Proc., 24,354(1965). (306) J. A.Cifonelli and A. Dorfman,]. Eiol. Chern.,235,3283(1960).

STRUCTURE AND METABOLISM OF GLYCOPROTEINS

463

(see item 2, Table X), and of keratan sulfate (see p. 459), and it has been isolated after partial, acid hydrolysis of a number of other glycoproteins (see Ref. 181). Secondly, frequently branched structures are present in the carbohydrate entities (see Ref. 315).Finally, the effect of purified glycosidases indicates that there are a number of common features. When present, sialic acid is always in a nonreducing, terminal position. In certain other natural products, such as colominic sialic acid has been found to be further substituted, but only by sialic acid, and, in view of their occurrence in oligosaccharides found in colostrum,30git would not be surprising should such structures occur in some glycoproteins. The use of neuraminidase purified from a number of sources leads to the release from many glycoproteins of the major part of the sialic acid, without any changes in the protein (see Table XII). However,

TABLEXI1 Release of Sialic Acid from Certain Glycoproteins by Use of Neuraminidase

G1ycoprotein

Mol. Wt.

Source of neuraminidase

Sialic acid (moleslmole of protein) Total

Released

References

1 Human, chorionic

27,000

Vibrio cholerae

8

8

310

gonadotropin 2 Fetuin 3 a,-Acidglycoprotein

45,000 40,000

Vibrio cholerae Diplococcus pneumoniae Vibrio cholerae Vibrio cholerae

13 14

13 14

248 311

16.9%

680

312 182

76

61

39

6

4

313

4 Lensubstance 5 Submaxillary-gland 1,000,000 glycoprotein (sheep) 6 Stomach-carcinoma glycoprotein 7 a,-Acute-phase globulins

270,000 45,000

Clostridium pelfringens Vibrio cholerae

17.9% 800

(307) G . Loewi and H. Muir, Immunology, 9,119(1965). (308)G.T.Barry and W. F. Goebel, Nature, 179,206(1957). (309)R. Kuhn and A. Gauhe, Chem. Ber. 98,395(1965). (310)0.P.Bahl,]. Biol. Chem., 244,567(1969). (311)R.C. Hughes and R. W. Jeanloz, Biochemistry, 3,1535(1964). (312)A. Pusztai and W. T. J. Morgan, Biochem.]., 78,135(1961). (313)A. H. Gordonand L. N. Louis, Biochem.J.,113,481(1969).

464

R. D. MARSHALL AND A. NEUBERGER

structures may occur in which, although sialic acid is in a nonreducing, terminal position, it is nevertheless not split off by neuraminidase. An example is sialylganglio-N-biose 11: P-D-GalNAc-(1+ 4)-D-Gal 3

t

2 a-D-sialyl The 2-acetamido-2-deoxy-~-galactose residue has to be removed from this compound before the sialyl linkage can become susceptible to n e ~ r a m i n i d a s e Sialic . ~ ~ ~ acid may also be released from glycoproteins under conditions of very mild acidity, very few other changes occurring in the glycoproteins, apart from the loss of small proportions of ~-fucose.~~~ Sialic acid occurs naturally as both the N-acetylated and the N glycoloylated form, the proportion in a given environment being different from one species to anotherS3lsOxidation of the acetyl group to the glycoloyl group occurs after the formation of 2-acetamido-2deoxy-~-glucose.~~~ In those proteins containing L-fucose, it would seem that this sugar is, probably always, also in nonreducing terminal positions.315This may not, however, be a general rule, because it would appear that only -75% of the L-fucose in A and B substances from ovarian cysts is released318after prolonged heating in M acetic acid at 100”.This result could indicate that L-fucose is not in a nonreducing, terminal position, so that cleavage of a second glycosidic linkage would be necessary in order that free L-fucose might be released, but this is not the only explanation possible for these findings. Thus, if, in some carbohydrate moieties, L-fucose were glycosidically attached to 0-3 of an N-acetylhexosamine residue, degradation of the moiety would probably involve, in part, a deacetylation of this “aglycon.” The L-fucosidic bond of this product would be more resistant to hydrolysis than that in the original entity, because of the presence of the positively charged

(314)R. Kuhn and Wiegandt, 2.Naturforsch., B , 18,541(1963). (315)See Ref. 6. (316)E. Martensson, A. Raol, and L. Svennerholm, Btochlm. Biophys. Acta, 30, 124 (1956). (317)R. Shauer, H.J. Schoop, and H. Faillard, Z. Phystol. Chem., 349.645 (1968);350, 155 (IeSQ). (318)R. A. Gibbons and W. T. J. Morgan, Blochem.J., 57,283(1954).

STRUCTURE AND METABOLISM OF GLYCOPROTEINS

465

amino group on the 2-amino-2-deoxyhexosyl group. The finding that, either before or after release of sialic acid, the single molecule of L-fucose is removed from a molecule of human, chorionic gonadotropin by use of the a-L-fucosidase from Aspergillus nigeP9 shows that this sugar residue is a nonreducing terminus. D-Galactose tends to be the next innermost sugar in nonreducing, terminal positions. Thus, about 70-80% of the D-galactose residues of sialic acid-free f e t ~ i n ,or~ of~ a~ glycopeptide , ~ ~ ~ isolated from it,248 were released by the action of P-D-galactosidase. Similar results were TABLEXI11 Release of 2-Acetamido-2-deoxy-~-glucose from Glycoproteins and Glycopeptides (GP) by Use of 2-Acetamido-2-deoxy-~-~-glucosidase 2-Acetamido-2-deoxyD-gluCOSe (moles/ mole of protein) Glycoprotein or GP

M.W.

Source of enzyme

1 Egg albumin GP 1,500 pig epididymis 2 Human IgG-GP 2,300 pig epididymis 3 Human, chorionic 27,000* Proteus vulgaris gonadotropin Proteus vulgaris 4a HCG-GP, 4,300 b 5a HCG-GP, 3,300 Proteus uulgaris b 6 al-Acid 40,000 Diptococcus pneumoniae Glycoprotein 7 Fetuin-GP 4,300 pig epididymis 8 Ovomucoid 28,000 Turbo cornutus

Total Released

References

1" 3.1 8'

319-321 109 214

24

O.gd 4.0" 1.7d 3.8' 6.2"

214 214 214 214 311

15 24

9.3d 3

323 324

3 7.7 11

5.6 4.9

"2-Acetamido-2-deoxy-~-glucosidase released about 1.5 moles of 2-acetamido-2deoxy-D-glucose from an amount of an egg-albumin glycopeptide containing 4 moles of this s ~ g a yResults . ~ ~ ~ obtained by Huang and Montgomery by use of enzymesz3' d o not agree with' those of other They rep0I-P" that all residues of 2-acetamido-2-deoxy-~-glucose but one are released by 2-acetamido-2-deoxy-~-glucosidase. T h i s is the molecular weight of the reduced, alkylated "The substance was free from sialic acid; it had been pretreated with P-D-galactosidase. No 2-acetamido-2deoxy-D-glucose was released, unless the sialic acid was first removed. dBy direct action on the glycopeptide (which was free from sialic acid).

(319) 0. P. Bahl and K. M. L. Agrawal,]. B i d . Chem., 244,2970 (1969). (320) H. H. Kaufman and R. D. Marshall, Abstr. 6th Intern. Congr. Biochem., New York, 2,92 (1964). (321) J. R. Clamp and L. Hough, Biochem.]., 94,502 (1965). (322) J. Conchie, A. J. Hay, I. Strachan, and G. A. Levvy, Biochern.]., 115, 717 (1969). (323) R. G. Spiro, Methods Enzymol., 8,26 (1966). (324) T. Muramatsu,]. Biochem. (Tokyo), 64,521 (1968).

R. D. MARSHALL AND A. NEUBERGER

466

obtained with orosomucoid311*319 and human, chorionic gonadotropin.319D-Galactose is not released at all, unless sialic acid has previously been removed. 2-Acetamido-2-deoxy-~-~-glucosidase acts directly on a number of glycoproteins and glycopeptides (see items 1,2,and 7 of Table XIII). Sequential use of enzymes may give results of considerable value. was allowed to act Thus, if 2-acetamido-2-deoxy-~-~-glucosidase directly on two glycopeptides prepared from human, chorionic gonadotropin (HCG) from which sialic acid had been removed, amounts of 0.9 mole (HCG-GP,, item 4a) and 1.7 moles (HCG-GP,, item 5a) of 2-acetamido-2-deoxy-~-g~ucose,respectively, were released per mole. However, if this reaction was preceded by the action of P-D-galactosidase (P-D-galactosidegalactohydrolase, E.C. 3.2.1.23), considerably larger amounts were released (see items 4b and Sb, respectively). Glycoproteins or glycopeptides are frequently treated with a - ~ mannosidases; (a-D-mannoside mannohydrolases, E.C. 3.2.1.24); D-mannose is often released (see Table XIV). Prior treatment with other enzymes is often found necessary (see items 5 and 6). TABLEXIV Release of D-Mannose from Glycoproteins and Glycopeptides (GP) by Use of a-D-Mannosidase D-Mannose (moles/ mole of protein) ReferGlycoprotein or GP

1 Eggalbumin 2 Egg albumin GP

3 Ovomucoid 4 Human, chorionic gonadotropin 5 HCG-GP, 6 HCG-GP, 7 a,-Acid glycoprotein

M. W.

45,000 1,500

Source of enzyme

Jack-bean meal Jack-bean meal Charonia lampas liver Turbo cortunus liver 28,000 Jack-bean meal Aspergillus 27,000b niger 4,300 Proteus vulgaris 3,300 Proteus vulgaris 40,000 Jack-bean meal

Total 5 5 5

Released 1.1" 3.3 1.3

ences 325 325 326

5

3.3

327

7.3

0.7

325

8.0 5.1

1.8' 4.1c,d 4 . 9 ~ 2.gC

214 214 ~214 325

5.4 12

"Another group of workers has reported that 2.5 moles of D-mannose per mole are split off by use of this enzyme,3" and another reported that all of the D-mannose is removed. bThis is the molecular weight of the reduced, (carboxymethy1)ated proteir~.~'O CAfterremoval of sialic acid, followed by treatment with p-D-galactosidase and 2-acetamido-2-deoxy-~-~-glucosidase. dD-Mannose was not released unless the substrate had been pretreated with p-Drgalactosidase and 2-acetamido-2-deoxy-p-~glucosidase.

STRUCTURE AND METABOLISM OF GLYCOPROTEINS

467

The glycosidic linkages in which 2-acetamido-2-deoxy-~-galactose participates are frequently a - ~ These . linkages occur in the nonreducing, terminal residues of A substance (see Table X) and in the carbohydrate-peptide linkage of at least some g l y c ~ p r o t e i n s . ' ~ ~ * ' ~ ~ Although D-galactose is usually @-linked, it is a-D-linked in B substance (see Table X). Full interpretation of data of this type must await further understanding of the specificities of the enzymes involved, of the nature of the residual oligosaccharide, and of the problem of possible heterogeneity discussed earlier. An advantage of enzymic methods is that the anomeric configuration may be revealed.

VII. THE BIOSYNTHESIS

OF

GLYCOPROTEINS

1. The Carbohydrate-peptide Linkages Certain parts of ihe structure of glycoproteins are rigidly controlled; it may therefore be deduced that a fairly high degree of specificity is involved in their biosynthesis. It would appear that an L-asparagine moiety in a polypeptide chain may be an acceptor of only one type of sugar, namely, 2-acetamido-2-deoxy-~-glucose, provided, in general, that the acetamido sugar is followed, in the p position on the C-terminal side, by a P-hydroxy-a-amino acid residue (see p. 425). As the acquisition of the sugar residue appears to depend on the presence of an amino acid (L-serine or L-threonine) that is incorporated into the peptide chain after the L-asparagine moiety, it is reasonable to suppose that the sugar residue concerned is added, as a post-ribosomal event, to a predetermined sequence of amino acids in the polypeptide chain. Direct experimental verification of this inference is at present lacking. On the other hand, there is evidence for direct, enzymic addition of 2-acetamido-2-deoxy-D-galactoseresidues to L-serine and Lthreonine residues of the polypeptide chains that are probable precursors of the submaxillary-gland glycoproteins. T h e enzymes, and sheep,16fihave a largely particulate in the glands from high degree of specificity, not absolute with regard to the receptor, as discussed later (see p. 468). The activated 2-acetamido-2-deoxy-

(325) Y.-T.Li,]. Biol. Chern.,241, lOlO(1966). (326) T. Murarnatsu,]. Biochem. (Tokyo),62,487 (1967). (327) T. Muraniatsu and F. Egami,]. Biochem. (Tokyo),62,700 (1967). (328) A. Hagopian and E. H. Eylar, Arch. Biochem. Biophys., 129,515 (1969)

468

R. D. MARSHALL A N D A. NEUBERGER

D-galactose is specific, and is not transferred, by the enzyme, from pyrophosphate) uridine 5'-(2-acetamido-2-deoxy-~-galactopyranosyl to other sugar molecules, regardless of whether they are monosaccharides or exist as integral parts of glycoproteins. Known receptors are the ox and sheep submaxillary-gland glycoproteins from which the disaccharide prosthetic groups have been removed. Peptides derived from pronase digests of these modified proteins do not function as receptors, and treatment of these proteins with trypsin leads to products that can be glycosylated, although much less readily. Most of the natural proteins do not function as acceptors, although at least one is known to, namely, a basic protein, encephalitogen, having a molecular weight of 16,400, isolated from the myelin of spinal cord; it is normally devoid of carbohydrate, but, presumably, some of the L-threonine residues have the environment requisite to occurrence of g l y c o ~ y l a t i o n .Enzymes ~~~ of this type presumably operate first on those polypeptide chains that are destined to become glycoproteins. Probably, enzymes that add 2-acetamido-2deoxy-D-galactose residues to existing carbohydrate moieties belong to a different class. linkage of collagen is formed The ~-galactose-5-hydroxy-~-lysine after production of the polypeptide chain, and, indeed, as a step subsequent to the conversion of L-lysine into 5-hydroxy-~-lysine residues.330 An enzyme involved in the biosynthesis of this type of linkage was identified in skin of embryonic guinea-pig and was purified 160-fold. The nature of the receptor is highly specific, only 5-hydroxy-~-lysineacting in this capacity, preferably when this amino acid is combined in collagen.331 Enzymic transfer of D-XylOSe from uridine 5'-(D-xylopyranosyl-'"C pyrophosphate) to L-serine residues of endogenous protein acceptors from ( a ) a cell tumor of the mouse188and ( b )chick-embryo cartilagelS9 occurs in cell-free extracts of both of these tissues, in the absence of biosynthesis of protein. The enzyme preparations employed were from the supernatant liquor, although activity was also present in the insoluble fractions. In these two types of tissue, the acceptors are heparin and chondroitin sulfate, respectively, but the presence of other D-xylose-containing glycoproteins in ascites fluid from (329) A. Hagopian and E. H . Eylar, Arch. Biochem. Biophys., 126,785 (1968). (330) J. Rosenbloom, N . Blumenkrantz, and D. J . Prokop, Biochem. Biophys. Res. Conmuti., 31,792 (1968). (331) H. B. Bosmann and E. H. Eylar, Biochem. Biophys. Res. Commun., 33, 340 (1968).

STRUCTURE AND METABOLISM OF GLYCOPROTEINS

469

cancer patients332 may indicate that other acceptors for D-xylose also exist. 2. Addition of Other Sugar Residues

The formation of the carbohydrate moieties of glycoproteins occurs by enzymic transfer of single sugar residues from glycosyl esters of nucleotides to nonreducing, terminal positions of the growing prosthetic groups. Certain oligosaccharide esters of nucleotides occur n a t ~ r a l l y , and ~ ~ ,it~ is ~ ~not impossible that these may be intermediates in the formation of glycoproteins; however, there is, at present, no proof of this hypothesis. The possibility must also be considered that the branch points that occur in the oligosaccharide moieties of a number of glycoproteins are formed by enzymes having activities analogous to that of amylo-( 1 -+ 4)+( 1 + 6)-trans-D-glucosidase (EC 2.4.1.18). Activated sugar-transferases prepared from the particulate fraction of chick embryo were used in a study of the formation of the linkage region (see p. 460) of chondroitin s ~ l f a t e . T~h~e ~ enzyme ,~~~ preparation transfers D-galactose from uridine 5’-(D-galactopyranosyl pyrophosphate) to D-xylose or u-xylosides, 0-D-xylosyl-L-serine, or endogenous acceptor, so that, in all of these reactions, D-galactose becomes attached to 0 - 4 of D-xylose. In the second stage of the reaction, D-galactose becomes attached to such substrates as DGal-( 1-+4)-D-Xyl and D-Gal-(1+4)-D-Xyl-L-Ser, as well as to endogenous acceptors; this step of the reaction appears to have fairly exacting requirements as regards the structure of the acceptor molec ~ l e The . ~ ~ enzyme ~ involved in attaching the D-glucuronic acid residue to 0 - 3 of the D-galactose in the neighborhood of the carbohydrate-peptide linking moiety of chondroitin 4- and 6-sulfate is probably different from that involved in the formation of later stages of the carbohydrate chains.335 A large number of glycosyl ester nucleotide transferases are known to exist; these include those involving cytidylic acid-sialic acid336*337 (332) K. Sugimoto, Tohoku J. E x p . Med., 64,271 (1964). (333) Y. Nakanishi, S. Shimizu, N. Takahashi, M. Sugiyama, and S. Suzuki, 1. H i d . C h e m . ,242, 967 (1967). (334) T. Helting and L. RodCn, ./. B i d . Cheni., 244, 2790 (1969). (335) T. Helting and L. RodCn, J . B i d . C h i n . , 244, 2799 (1969). (336) S. Roseman, “Biochemistry of Glycoproteins and Related Substances,” E. Rossi and E. Stoll, eds., S. Karger, Basel and New York, 1968, Part 11, p. 244. (337) M . J . Spiro and R. G. Spire,]. B i d . Chem., 243,6520 (1968).

470

R. D. MARSHALL AND A. NEUBERGER

as well as uridine 5‘-(D-glucopyranosyl pyrophosphate) for formation of the carbohydrate moieties of collagen.”8 Transferases acting upon uridine 5’-(D-galactopyranosyl pyrophosphate) are also known; some of these may result in the formation of P-D-galactosyl groups,339 and others, in the production of a-D-gahCtoSyl groups”O in carbohydrate moieties. Enzymes that catalyze the transfer of 2-acetamido-2deoxy-D-galactose from uridine 5’-(2-acetamido-2-deoxy-~-galactopyranosyl pyrophosphate) to precursors of the blood-group A type of carbohydrate moieties (see Table X) occur in milkJ4‘ and submaxillary glands::’4zof A and AB individuals, as well as in pig gastric mucosa.““:’ Transferase activities acting with guanosine 5’-(~-fucosyl pyrophosphate) occur in human milk, that forming the ( 1 + Z)-Iinkage to D-galactose being absent from that of those individuals lacking A, B, or H blood-group, specific antigens,“44and that forming the (1+ 4)-linkage to 2-acetamido-2-deoxy-D-g~ucosebeing absent from that of those lacking Lewis determinant^."^ All of these enzymes have, with regard to the acceptor molecules, fairly rigid requirements as to specificity.

3. Incorporation of Sulfate into Carbohydrate Moieties The incorporation of sulfate into the carbohydrate moieties of glycoproteins occurs after incorporation of the sugar moieties, and thus, here again, is a post-ribosomal event. The process is, therefore, not directly under control of the genome. Sulfation of the sugar moieties of chondroitins occurs by a sequence of steps, the first of which involves formation of 3‘-O-phosphono5’-adenylyl hydrogen sulfate, the sulfate d ~ n o r . ” ~This J ~ ~enzyme is present in a number of the tissues in which these glycoproteins (338) H. B. Bosmann and E. H. Eylar, Biochem. Biophys. Res. Commun.,30,89 (1968). (339) M. J. Spiro and R. G . Spiro,}. Biol. Chem., 243,6529 (1968). (340) A. Kobata, E. F. Grollman, and V. Ginsburg, Biochem. Biophys. Res. Commun., 32,273 (1968). (341) A. Kobata, E. F. Grollman, and V. Ginsburg, Arch. Biochem. Biophys., 124, 609 (1968). (342) V. M. Hearn, Z. G . Smith, and W. M. Watkins, Biochem.}., 109,315 (1968). (343) H . Tuppy and W. L. Staudenbauer, Nature, 210,316 (1966). (344) L. Shen, E. Grollman, and V. Ginsburg, Proc. Nat. Acad. Sci. U . S., 59,224 (1968). (345) E. Grollman, A. Kobata, and V. Ginsburg, Fed. Proc., 27,345 (1968). (346) P. W. Robbins and F. Lipmann,J. Amer. Chem. Soc., 78,2652 (1956). (347) S. Suzuki and J. L. Strominger,}. Biol. Chem., 235,257 (1960).

STRUCTURE AND METABOLISM OF GLYCOPROTEINS

471

are sulfated, including chick-embryo ~ a r t i l a g e ~ ~and ~ - ~the ~ Oisthmus of the hen As already mentioned (see p. 470),it is probable that sulfation of the carbohydrate chains occurs, to a considerable extent at least (if not wholly), after the polymerization has occurred. This probability was suggested by Meyer and c o ~ o r k e r s , ~because ~ ’ , ~ ~ ~of their finding of chondroitin in ox cornea that was free from, or very low (