Advances in Carbohydrate Chemistry and Biochemistry, Volume 47

Advances in Carbohydrate Chemistry and Biochemistry Volume 47 This Page Intentionally Left Blank Advances in Carboh...

Author: R. Stuart Tipson

40 downloads 1821 Views 18MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

Advances in Carbohydrate Chemistry and Biochemistry Volume 47

This Page Intentionally Left Blank

Advances in Carbohydrate Chemistry and Biochemistry Editors R. STUART TIPSON DEREK HORTON Board of Advisors LAURENS ANDERSON J. GRANTBUCHANAN GUYG. S. DUTTON STEPHEN J. ANGYAL HANSH. BAER BENGTLINDBERG CLINTON E. BALLOU HANSPAULSEN JOHNS. BRIMACOMBE NATHAN SHARON ROYL. WHISTLER

Volume 47

ACADEMIC PRESS, INC. Harcourt Brace Jovanovich, Publishers

San Diego New York Berkeley Boston London Sydney Tokyo Toronto

This book is printed on acid-free paper. @

COPYRIGHT 0 1989 BY ACADEMIC PRESS, INC. All Rights Reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher.

ACADEMIC PRESS, INC. San Diego, California 92101 United Kingdom Edition published by ACADEMIC PRESS LIMITED 24-28 Oval Road, London NWI 7DX

LIBRARY OF CONGRESS CATALOG CARD NUMBER:

ISBN

0-12-007247-5

(ak. paper)

PRINTED IN THE UNITED STATES OF AMERICA

89 90 91 92

9 8 7 6 5

4

3 2

1

45-1 1351

CONTENTS PREFACE................................................................

vii

Complexes of Metal Cations with Carbohydrates in Solution STEPHEN J . ANGYAL 1. 11. 111. IV . V.

lntroduction ........................................................ Detection and Characterization of the Complexes.......................... The Structure of Cation-Carbohydrate Complexes ........................ Applications ........................................................ Biological Implications................................................

1

3 12 35 42

Anomeric and Exo-anomeric Effects in Carbohydrate Chemistry

IGOR TVAROSKA AND TOMASBLEHA Introduction ........................................................ 45 Definition of the Anomeric and Exo-anomeric Effects ...................... 47 Experimental Data on These Effects ..................................... 59 Molecular Orbital Calculations of the Anomeric Effect...................... 75 V . The Anomeric and Exo-anomeric Effects in Potential-Function Calculations . . . 103 Vl . Nature of the Anomeric Effect.......................................... 106 VII . Role of the Anomeric Effect in the Reactivity of Carbohydrates .............. 116 1. I1. Ill . IV .

IF-Nuclear Magnetic Resonance-Spectral Studies of the Interactions of Metal Ions with Carbohydrates: Use of Relaxation Probes

KILIANDILLAND R. DOUGLAS CARTER Introduction ........................................................ General Considerations Concerning Carbohydrate Structure ................. General Considerations Concerning Metal Ions ........................... Uses of Mn2+and Gd3+............................................... V . Gd3+and Mn2+Interactions with Carbohydrates .......................... Vl . Conclusions......................................................... 1. 11. Ill . IV.

125 127 128 135 137 165

Application of Anhydrous Hydrogen Fluoride for the Structural Analysis of Polysaccharides EVGENY V. VINOGRADOV. AND ANDREW J . MORT YURIYA . KNIREL. 167 1. Introduction ........................................................ 11. Mechanism of Reaction between Hydrogen Fluoride and Carbohydrates....... 168

V

CONTENTS

vi

111. Techniques for Performing Solvolysis....................................

IV . V. VI . VII .

Preparation of Monosaccharides........................................ Preparation of Oligosaccharides ........................................ Other Applications of Hydrogen Fluoride ................................ Conclusion .........................................................

.

173 174 180 200 202

.

The Thermal Decomposition of Carbohydrates Part I The Decomposition of Mono-. Di.. and Oligo-saccharides PIOTR

TOMASIK. MIECZYSLAW PAEASIASKI. AND STANISLAW WIEJAK

I. Introduction .................................................... I1. Caramel and Caramelization ....................................... 111. The Pyrolysis of Sugars ...........................................

.

.... ....

....

203 204 270

.

The Thermal Decomposition of Carbohydrates Part I1 The Decomposition a Star-

PIOTRTOMASIK. STANISLAW WIEJAK.AND MIECZYSEAW PAEASI~SKI IV. Introduction ........................................................ V . Dextrins and Dextrinization ........................................... VI . The Pyrolysis of Starch................................................

279 281 335

The Macrostructure of Mucus Glycoproteins in Solution STEPHEN E. HARDING

I. Introduction ........................................................ I1. Composition ........................................................ 111. Primary and Secondary Structure: The Basic Unit of the Mucus Glycoprotein . . IV . Tertiary Structure: Assembly of Basic Units............................... V . The Gross Conformation of Mucus Glycoproteins in Solution ............... VI . Mucin Heterogeneity ................................................. VII. SummaryandProspects ...............................................

345 347 349 352 370 314 380

AUTHORINDEX...........................................................

383

........................................................... SUBJECTINDEX

411

PREFACE The first chapter in Volume 47 is an authoritative account by S. J. Angyal (Kensington, N.S.W., Australia) on complexes of metal cations with sugars in solution. The potential uses of such complexes to control the behavior of sugars are not as well recognized as they might be, and Angyal‘s account will be of great value for analytical chemists seeking enhanced separation methodology, for spectroscopists studying molecular structure, and for synthetic workers. A comprehensive survey of the anomeric and exo-anomericeffects in carbohydrate chemistry is contributed by I. TvaroSka and T. Bleha (Bratislava, Czechoslovakia).The stereochemicalinfluenceof these effectson the reactions of sugars and on conformational behavior about the glycosidic linkage has far-reaching significancebut, as the article points out, our theoretical understanding has not yet been perfected to a fully integrated and quantitatively predictive basis. The contribution provided by K. Dill and R. D. Carter treats a specialized aspect of metal ion-carbohydrate interaction, devoted to the use of I3C-n.m.r. spectroscopy with Gd3+and Mn2+in particular as shift reagents and relaxation probes to study the behavior of biological carbohydratesthat normally interact with Ca2+and Mg2+.A newer technique that will certainly take its place as a standard tool in polysaccharide structural analysis is the selective depolymerization by anhydrous hydrogen fluoride, here surveyed by Yu. A. Knirel and E. V. Vinogradov (Moscow, U.S.S.R.) and A. J. Mort (Stillwater, OK). The thermal decomposition of sugars and oligosaccharides to produce caramels for use in foods, and the thermal modification of starch to manufacture “dextrins” for use as adhesives, constitute traditional technological arts of considerable commercial significance,although much of the specific chemistry involved in these processes, as well as in the higher temperature pyrolytic breakdown, remains poorly understood by contemporary standards. In two parts, P. Tomasik, S. Wiejak, and M. Palasidski (Poland) bring together a vast amount of information not hitherto readily available to workers in the field. It may be hoped that their efforts will now stimulate further work using modem techniques to furnish precise structural characterization for these materials, which would help in the rational control of manufacturing processes and in the precise definition of the products. The last chapter, written by S. E. Harding (Nottingham, United Kingdom), describes and discusses the macrostructure of mucus glycoproteins, complex polyelectrolyteswhose behavior in solution is governed by aspects of secondary and tertiary structure that control their interactionsin biological systems. Kensington, Maryland Columbus, Ohio August 1989

R. STUART TIPSON DEREK HORTON vii

This Page Intentionally Left Blank

ADVANCES IN CARBOHYDRATE CHEMISTRY AND BIOCHEMISTRY. VOL . 47

COMPLEXES OF METAL CATIONS WITH CARBOHYDRATES IN SOLUTION BY STEPHENJ . ANGYAL

.

School of Chemistry. University of New South Wales. Kensington. N.S. W. 2033 Australia I . Introduction ......................................................... I1 . Detection and Characterization of the Complexes ........................... 1 . Paper Electrophoresis ............................................... 2 . Thin-layer Ligand-exchange Chromatography ........................... 3. Nuclear Magnetic Resonance Spectroscopy ............................. 4 . Other Methods .................................................... 111. The Structure of Cation -Carbohydrate Complexes ......................... I . Complexing Sites................................................... 2 . Cations ........................................................... 3. The Structures of Crystalline Complexes ............................... 4 . Stability Constants ................................................. 5 . Alduronic Acids ................................................... IV . Applications ......................................................... 1 . Electrophoresis and Thin-layer Chromatography......................... 2 . Preparative Separations on Columns of Cation-exchange Resins ............ 3. High-performance Liquid Chromatography ............................. 4 . Nuclear Magnetic Resonance Spectroscopy ............................. 5 . Synthetic Applications .............................................. V . Biological Implications ................................................

1

3 3 5 5 12 12 12 19

22 26 31 35 35 36 40 40 41 42

I . INTRODUCTION

Complex-formation between salts and carbohydrates is not a new subject. Crystalline adducts of sugars with inorganic salts have been studied since 1825. when an adduct ofD-glucose with sodium chloride was first described. mainly for the purpose of isolating and purifying sucrose and glucose . These studies were not particularly rewarding: in a paper on the adduct of sucrose with sodium chloride. presented to The Chemical Society in 187 1. (1) F. Calloud. Mkm . Soc. Acad . Suvoie. 1 (1825) 34; J . Pharm., 1 1 (1825) 562-564 .

1

Copyright 0 1989 by Academic Res. Inc. AII rights of reproduction in any form reserved .

2

STEPHEN J. ANGYAL

Gill2commented thus: “Having obtained some positive results, I beg to lay them before the Society, for though of little interest in themselves, their publication may prevent the necessity ofsome other worker wastingso much time on an ungrateful subject.” It was in 1971, just 100 years later, that there appeared the first paper to discuss the relationship between structureand complex-formationof neutral carbohydrates with cation^.^ Most of the developments during the interven.~ ing 100 years were covered in an article by Rendleman in this S e r i e ~The present discussion is an updating of Rendleman’s chapter, and it covers the renaissanceof interest in metal- sugar complexes. The information given by Rendleman will not be repeated here unless it is required for the understanding of subsequent developments. The scope of this chapter is more restricted than that of Rendleman’s. His dealt extensively with crystalline adducts of salts and sugars, many of which were then known, although their structureshad not yet been determined. In aqueous solution, metal cations are coordinated to water molecules. Hydroxyl groups of carbohydrates can also coordinate to cations, but single hydroxyl groups cannot compete with the solvent; only a suitably arranged combination of two or three hydroxyl groups will result in significant complex-formation. There is, of course, always a certain amount of coordination between any sugar and any cation, but, in many cases, the stability constant of the complex is so small (<0.1 M-l) that complex-formation is not detectable. The expression “does not form a complex,” as used in this chapter, means that there is very little (< 1%) of the complex present in equilibrium. In the crystalline state, however, there being no, or only a few, solvent molecules present, hydroxyl groups of any carbohydrate may coordinate to any cation. Crystalline adducts of sucrose with numerous salts, in stoichiometric proportions, are known; and yet, in aqueous solution, there is no detectable complex-formationbetween sucrose and any cation. In this chapter, these crystalline adducts will not be regarded as complexes and will not be treated, except where their structure throws light on the nature of complexes formed in solution. Rendleman’s chaptefl also reviewed compounds formed from carbohydrates and metal oxides or hydroxides, that is, adducts formed in alkaline solution. Andrasko and Fo&n5 have shown, in the case of D-riboseand Na+, that complex-formation is unaffected by the pH if it lies between 1 and 10, but increases steeply at higher pH values. We shall not discuss these com(2) C. H. Gill, J. Chem. Soc., 24 (1871) 269-275. (3) S. J. Angyal and K. P. Davies, Chem. Commun., (1971) 500-501. (4) J. A. Rendleman, Jr., Adv. Curbohydr.Chem., 21 (1966) 209-271. (5) J. Andrasko and S. F o d n , Biochem. Biophys. Res. Commun., 52 (1973) 233-239.

COMPLEXES OF METAL CATIONS WITH CARBOHYDRATES

3

plexes formed in alkaline solution;they are actually salts formed by dissociation of one or more hydrogen ions from the carbohydrates, and in the past two decades, there has not been much interest in this type of compound. Nor shall we deal with complexes (actually esters) formed with such metalcontaining anions as molybdate or tungstate. This article, therefore, deals with complexes (briefly designated as “metal -sugar complexes”) that are formed, in detectable proportions, in aqueous solution from cations and un-ionized carbohydrates. As an exception, we shall briefly discuss (in Section 111,s) complexes of cations with aldonic and alduronic acids, because of their structural similarities to metal - sugar complexes. At the time when Rendleman4wrote his article, there was still some doubt about the existence of metal - sugar complexesin solution, although actually there already existed clear evidence for them, mainly from electrophoretic migration of sugars in solutions containing cations.6-8Carbohydrates are very soluble in anhydrous methanol containing sufficientcalcium chloride: undoubtedly owing to complex-formation. Nothing was known, however, about the stericand structural requirementsfor complex-formation, nor had any stability constants been determined. The best understanding of complex-formationwas shown, a generation earlier, by Isbell,lo who demonstrated that a-D-gulopyranose,but not P-D-gulopyranose, forms a complex with calcium chloride in solution. From his data, the stability constant ofthe complex can be calculated to be 3.7 M-l. There are now several methods whereby the occurrence of complexformation between a cation and a carbohydrate can be recognized. These methods will be treated first, before the structure of the complexesand their stability are discussed. 11. DETECTION AND CHARACTERIZATION OF THE COMPLEXES

1. Paper Electrophoresis Complex-formation between cations and carbohydrates can be readily detected by paper electrophoresis.” In 1959, Frahn and Mills6 published data on the paper-electrophoretic separation of carbohydrates in which, besides several complexinganions, they also utilized basic lead acetate as the (6) (7) (8) (9) (10) (11)

J. L. Frahn and J. A. Mills, Aust. J. Chem., 12 (1959) 65-89. J. A. Mills, Biochem. Biophys. Rex Commun., 6 (1961-62)418-421. J. A. Rendleman, Jr.,J. Org. Chem.. 31 (1966) 1839-1845. K. B. Domovs and E. H. Freund, J. Dairy Sci.,43 (1960) 1216-1223. H. S. Isbell, Bur. Stand. J. Rex, 5 (1930) 741 -755. H. Weigel, Adv. Carbohydr. Chem., 18 (1963) 61-97.

4

STEPHEN J. ANGYAL

supporting electrolyte. All of the sugars, alditols, and cyclitols that they tested migrated towards the cathode, showing that they were coordinated to the cation. The rate of migration gave an approximate measure of the relative extent of coordination. Paper electrophoresisis, even in this age of high-performanceliquid chromatography (h.p.l.c.), a useful method of analysis. The equipment required is inexpensive,and the technique is simple. It lends itself well to comparison of the coordinatingability of many carbohydrates with any one cation, or of that of several cations with any one carbohydrate. The behavior of Pbz+ proved to be typical of that ofcomplexingcations; this behavior was not fully understood by Frahn and Mills,6 but they made the observation that the mobility of alditols increases with increasing number of threo hydroxyl groups.6 A few years later, in a short but important note, Mills' compared the electrophoretic mobilities of some representative polyols with several cations: Caz+,Sr2+,and Ba2+were found to be the most effective. Amongst the polyols, cis-inositol(1) had the greatest mobility with all cations; hence, in subsequent studies, it was adopted as the standard. HO

?H

I

1

Bourne and coworkers'* studied the electrophoretic behavior of many polyols in copper acetate and basic copper acetate solutions as electrolytes. Although the sugars were found to be rather unresponsive, all alditolstested showed considerable mobility, which was interpreted as being due to complex-formation with Cu2+ions. Actually, the preponderant ionic species in those solutions are [ C U ~ ( O H ) ~and ] ~ +CuOAc+; the behavior of copper acetate with carbohydrates will be discussed in Section III,2. The electrophoretic mobilities of a large number of carbohydrates have been determined in buffered calcium acetate s~lution.'~ The method is very sensitive for the detection of complex-formation: the smallest detectable mobility (Mi0.0 1, compared to that of cis-inositol) represents less than 1% of complex-formation. All carbohydrates that have two vicinal hydroxyl groups on a six-membered ring or on an acyclic chain, or cis on a fivemembered ring, show some mobility; only derivatives lacking vicinal hy(12) E. J. Bourne, F. Searle, and H. Weigel, Curbohydr. Res., 16 (1971) 185-187. (13) S. J. Angyal and J. A. Mills, Aust. J. Chem., 32 (1979) 1993-2001.

COMPLEXES OF METAL CATIONS WITH CARBOHYDRATES

5

droxyl groups, or having them trans on a five-membered ring, remain stationary. D-G~UCOW, usually regarded as non-complexing, has M i0.02. 2. Thin-layer Ligand-exchange Chromatography This method makes use of commercially available, thin-layer plates coated with a cation-exchangeresin in the sodium form; the sodium ions can readily be exchanged for other cationsby immersing the plate in an appropriate salt solution. Cupric, calcium, and lanthanum acetates have been ~ s e d ’ ~to 9 ’introduce ~ cations that complex readily. The plate is then developed with water as the irrigant. Compounds that form complexes with the cations are retained and have low R F values; those that complex weakly are found near the solvent front ( R F -0.9) and are not readily distinguished from those that do not form a complex. This is a rapid and simple method for detecting complex-formation, unless it is very weak, and to estimate its relative strength.

3. Nuclear Magnetic Resonance Spectroscopy Complex-formation causes changes of the chemical shifts in the n.m.r. spectra of sugars and polyols. These changes show that complexeshave been formed, and may lead to determination of their structures and of their stability constants. Establishment of the equilibrium between polyols and cations in aqueous solution is fast on the n.m.r. time-scale. The rate offormation ofthe complex from D-glucitol and Ca2+ has been found16 to be lo8 M-I * s-’ at 40”. Hence, separate spectra for the polyol and its complex are not observed, but only an average spectrum. Further addition of a complexing cation causes further shifts of the signals; complete conversion into the complex is, however, not achieved at concentrations that allow the recording of n.m.r. spectra. Hence, the “limitingshifts” cannot be directly determined;moreover, at high concentrations,weaker complexingwith other hydroxyl groups causes additional small shifts. When recording the shift changes, it is necessary, therefore, to indicate the concentration of each component, or that of one component in addition to the molar ratio. A more satisfactory way, if the spectra have been recorded at several molar ratios, is to report the initial slope of the shift versus the molar ratio (shift equivalents).” (14) (15) (16) (17)

J. Briggs, P. Finch, C. Matulevicz, and H. Weigel, Carbohydr.Rex, 97 (1981) 181 - 186. S. J. Angyal and J. A. Mills, Aust. J. Chem., 38 (1985) 1279- 1285. J. K. Beattie and M. T. Kelso, Aust. J. Chem., 34 (1981) 2563-2568. S. J. Angyal, D. Greeves, L. Littlemore, and V. A. Pickles, Aust. J. Chem., 29 (1976) 123 1 - 1237.

6

STEPHEN J. ANGYAL

Muellitol(2) presents the only reported instance of a polyol that shows a slow interchange between the complex and its components, probably owing to steric hindrance by the bulky sidechains.18With lanthanides,but not with uni- and di-valent cations, separate lines are found in the spectrum for the complexed and the uncomplexed molecules, even at 99". (Slow exchange was postulated for the complex of Cr3+with cis-ino~itol,'~ but the evidence presented is not conclusive.) OH

2

Diamagnetic cations cause small, but readily interpreted, changes in the n.m.r. spectra; those induced by paramagnetic cations are much larger, but not so easily evaluated. The two types of cation will be discussed separately. a. Complexes with Diamagnetic Cations.-On formation of a complex between a polyol and a diamagnetic cation, all ofthe signalsin the 'H-n.m.r. spectrum of the polyol shift downfield, to variousextents. This "diamagnetic shift" is caused by inductive effects and by a direct field effect. According to Buckingham,20the change in the screening constant (An) of a proton due to polarization of the C-H bonding electrons when subjected to an electric field effect E is A0=-2X

1O-l2E,- 10-'*E2,

where E, is the component of E in the direction ofthe C- H bond. According to this formula, the effect can be positive or negative, but instances of a negative effect have not so far been found amongst polyols. A change of conformation accompanying the formation of the complex will also cause changes in the chemical shifts (for instance, for a-D-lyxopyranose and methyl /3-D-ribopyranosidez'). These changes could be larger than those

(18) S. J. Angyal, D. Greeves, and L. Littlemore, Aust. J. Chem., 38 (1985) 1561- 1566. ( 19) R. D. Carter and K. Dill, Inorg. Chim. Acra, I25 ( 1986) ~9- L 1 1. (20) A. D. Buckingham, Can. J. Chem., 38 (1960) 300-307. (21) S. J. Angya1,Adv. Chem. Ser., 117(1973) 106-120.

COMPLEXES OF METAL CATIONS WITH CARBOHYDRATES

7

TABLEI Induced Shifts" (p.p.m.) of Proton Signals of Some Cyclitols on Addition of LaCI, Solution H-1

H-2

H-3

H-4

H-5

Hd

Me

epi-Inositol(4) 0.17 3-0-methyl- 0.19 2-C-methyl- 0.23

0.27 0.22

0.55 0.43

0.27 0.22 0.29

0.17 0.19 0.18

0.07 0.08 0.12

0.18 0.16

Cyclitol

a

0.58

Limiting shifts calculated from the experimental data, with K = 10 M-I.

caused by the electric charge of the cation, and can also be negative (for example, for 2-deo~y-2-C-methyl-epi-inositol~~). When the complexingsite consistsof a sequence of an axial (a), an equatorial (e)and an axial hydroxyl group (as in 3), the C-H bond on the central carbon atom will be at a small angle to the cation-H vector; hence, E, is large. Typically, this proton shows the greatest induced shift. The a,e,a site was first identified by consideration of such calcium-induced shift^.^ Examples of shifts induced by La3+are shown in Table I. epi-Inositol(4)has been used exten~ively~~ for study of induced shifts, because it complexes well and has a simple spectrum in which every proton signal is readily identified, no matter where it is shifted to. The presence of an 0-methyl or C-methylgroup at the complexing site does not substantially alter the induced shifts (in contrast to those caused by the contact effect; see later). The calciuminduced shifts are about half as large; univalent cations cause very small shift-changes.

H

3

L

Similar shift-changes were found" in pyranoses that have an a,e,a arrangement of three oxygen atoms: for example, in the spectrum of methyl a-D-allopyranoside,the H-2 signal shifts to the greatest extent, followed by those of H-3 and H-1; in that of 1,6-anhydro-P-~-allopyranose, the H-3 (22) S. J. Angyal, D. Greeves, and V. A. Pickles, Carbohydr.Res., 35 (1974) 165- 173. (23) S. J. Angyal and R. J. Hickman, Ausr. J. Chem., 28 (1975) 1279-1287.

STEPHEN J. ANGYAL

8

signal shifts the most, and then those of H-4 and H-2; and in that of 1,6-anhydro-a-D-talopyranose, that of H-2 the most, and then those of H-3 and H- 1. When complexing occurs with three syn-axial hydroxyl groups, it would be expected that the (equatorial) hydrogen atoms vicinal to those three hydroxyl groups would be the most affected. It was found22that, in the spectrum of cis-inositol (l),calcium ions cause the axial-proton signals to shift downfield somewhat more than the equatorial ones;24there are here, however, three a,e,a sites, in addition to the triaxial site. MgZ+ions, which do not complex at the a,e,a sites (see Section III,l), cause the signals of the equatorial protons to shift twice as much as those of the axial ones. These diamagnetic shifts are useful in determining the complexing site. For example, in the spectrum of cyclohexane-1,2,3,4,5/0-pentol (54, on addition of Ca2+,the signal of H-3 shifts strongly d~wnfield;~indicating complexingat 0-2,O-3,O-4. However, addition of Mgz+shiftsthe signals of H- 1, H-3, and H-5 downfield, showing that complexing occurs at the three axial hydroxyl groups of the alternative chair form (5b). The signal of the methyl group in the spectrum of 2deoxy-2-methyl-epi-inositol (6) shifts upfreld on addition of either Ca2+or Mg2+;this shows that ring inversion occurs, to provide a,e,a and triaxial complexing sites.22

AOH HO

HO

HO

e

5a

OH

H o @

5b HO

HO &OH

6

By cooling, and careful control of the pH of, aqueous solutions, it is possible to obtain well-resolved hydroxyl-proton resonances, and these also (24) Because each hydroxyl group has to pass between two other hydroxyl groups during ring inversion of cis-inositol,this inversion is slow on the n.m.r. time-scale, and the coalescence Hence, temperature of the two proton signals is well above ambient temperat~re.2~.~~ although the two chair forms of cis-inositolare equivalent,the signalsof the equatorial and the axial protons are separately observed. (25) S . Brownstein, Can. J. Chem., 40 (1962) 870-874.

COMPLEXES OF METAL CATIONS WITH CARBOHYDRATES

9

are displaced downfield on complex-formation with calcium ions.26The hydroxyl-proton signals of D-ribose have been studied in detail, and, from these data, the stability constants of the complexes of its various forms have been determined. The signals in the ')C-n.m.r. spectra of polyols are also shifted on complex-formation; for example, the limiting shifts of epi-inositol (4) on addition of lanthanum chloride are2': C-1,5, -0.9; C-2,4,0; C-3, 1.3; and C-6, -0.9 p.p.m.

+

b. Complexes with Paramagnetic Cations.- Addition of paramagnetic cations to solutionsof complexingsugars causes shiftsin the n.m.r. spectrum that are much larger than those caused by diamagnetic ions, and are more complex to interpret. Some results are shown in Table 11. The lanthanideinduced shifts (1.i.s.) are the combined results of three different mechanisms: diamagnetic, pseudocontact, and contact interactions.28The diamagnetic shifts are caused by the electric field; the pseudocontact effect is transmitted through space, and is strongest at the closest nuclei; and the contact effect is transmitted through bonds. Correction can be made for the diamagnetic contribution to the induced shift by deductingthe value of the shifts induced by the (diamagnetic) La3+ion. The pseudocontact and contact shifts can be separated by mathematical2*or g r a p h i ~ aprocedures l ~ ~ ~ ~ ~that are based on the differing proportions of contact and pseudocontact shifts induced by each lanthanide ion. The contact and pseudocontact shifts have opposite signs (except for Tb3+,Dy3+,and Ho3+). TABLEI1 Lanthanide-induced Shifts" (p.p.m.) in the Spectra of Two C y ~ l i t o l s ~ ~ Cyclitol

Cation

H-1,5

H-2,4

H-3

H-6

Me

epi-Inositol(4)

PrN NdN Eu3+ Yb3+ Eu"

- 1.05

- 2.05 -1.10

2.05 6.1 -7.2 -4.8 0.15

- 3.5 -1.65 1.65 4.45 -0.2

0.45

3-0-methyl-

0.3 -0.35 -0.25 -0.05

1.6 -2.05 0.55

Limiting shifts calculated from the expimental data, with K = 10 M-'.

(26) M. C. R. Symons, J. A. Benbow, and H. Pelmore, J. Chem. Soc., Furuduy Trans. I , 78 (1982) 3671 -3677; 80 (1984) 1999-2016. (27) S. J. Angyal, L. Littlemore, and P. A. J. Gorin, Aust. J. Chem.. 38 (1985) 41 1-418. (28) For reviews, see F. Inagaki and T. Miyazawa, Progr. Nucl. Mugn. Reson. Spectosc., 14 (1981) 67-111; J. A. Peters and A. P. G . Kieboom, Recl. Truv. Chim. Pup-Bus, 102 (1983) 381 -392. (29) C. N. Reilly, B. W. Good, and J. F. Desreux, Anal. Chem., 47 (1975) 21 10-21 16. (30) J. Reuben, J. Am. Chem. Soc.. 99 (1977) 1765- 1768. (31) S. J. Angyal and D. Greeves, Aust. J. Chem., 29 (1976) 1223-1230.

10

STEPHEN J. ANGYAL

Only the pseudocontact shifts give exact information on the location of the cation, and, therefore, on the structure of the complex. According to an equation developed by McConnell and Robertson,32the lanthanide-induced shift, Av, is Av = C (3 cosz 8 - l)/r3, where r is the length of the vector joining the cation to the observed nucleus, and 8 is the angle between this vector and the principal magnetic axis. The 1.i.s.thus allow calculation ofthe position ofthe cation. Becausethe direction of the principal magnetic axis has also to be determined, at least five independent items of data are required for the calculation, and complexes of simple sugars with cations seldom provide so many data. Few such calculations have been performed. Grasdalen and coworkers33determined the geometry of the complex formed from lanthanide ions and methyl (Y-Dgulopyranoside (7):the cation was found to be situated at approximately equal distances (- 220 pm) from 0-1,O-2, and 0-3. An earlier attempt%to perform this calculation, without separating the contact from the pseudocontact shifts, gave unreasonable lanthanide- oxygen distances. Kieboom and coworkers3scarried out similar calculations based on the shifts caused by Yb3+in the spectrum of D-glUCitOk satisfactory results were obtained, showing coordination to 0-2, 0-3, and 0-4, although the contact interactions were disregarded. Yb3+gives a smaller contact :pseudocontact ratio than any other lanthanide ion. HO CHzOH

HO

OMe

7

There is no established, quantitativecorrelation between the contact shifts and the structure of the complex. However, it was observed that, for triden(32) H. M. McConnell and R. E. Robertson, J. Chem. Phys., 29 (1958) 1361 - 1365. (33) H. Grasdalen,T. Anthonsen, 0.Harbitz, B. Larsen, and0. Smidsred,Acta Chem.Scund., Ser. A , 32 (1978) 31-39. (34) H. Grasdalen, T. Anthonsen, B. Larsen, and 0.Smidsred, Acta Chem. Scund., Ser. B, 29 (1975) 17-21. (35) A. P. G . Kieboom, A. Sinnema, J. M. van der Toorn, and H. van Bekkum, R e d Trav. Chim. Pays-Bus, 96 (1977) 35-37.

COMPLEXES OF METAL CATIONS WITH CARBOHYDRATES

11

tate complexes of polyols, contact shift is greatest for those protons that lie along a planar zigzag path from the cation; even five bonds away, the contact shift can be detected if this geometry prevail^.'^,^' This relationship may, in many cases, point to the location of the cation. It is not valid for the 13C-n.m.r. signals, which tend to have alternate positive and negative induced shifts along the bonds from the cation. In the cation complexes of polyhydroxy compounds, the contact shiftsare strong,31particularly with Eu3+,which has the highest ratio of contact to pseudocontact shifts. Often, even a cursory glance at the induced shifts will indicate the site of complex-formation (see Table 11). However, when coordination occurs at an ether or acetal oxygen atom, rather than at a hydroxyl group, there is very little contact shift;in the glycosides, the anomeric proton shows no contact shift.l 7 Apparently, the contact-shift mechanism requires at least partial overlap of the orbitals of the cation and the oxygen atom; this is more easily achieved when the oxygen atom is bonded to a hydrogen atom, rather than to two carbon atoms. With polyols, the procedure of separating contact from pseudocontact ~ h i f t sometimes s ~ ~ ~ ~proves unsatisfactory: some of the higher lanthanides give discordant r e s ~ l t s .It~appears ~ , ~ ~ that complexesof these higher lanthanides, that have smaller ionic radii, have different interatomic distances and valency angles, and probably different coordination numbers, even if they have the same structures. Theoretical studies confirmed these difference^.'^ It is advisable, therefore, to disregard the higher lanthanides in this proced~re.~~ Structural information may also be obtained from n.m.r. spectra by using paramagnetic cations, such as Gd3+and Mn2+,that do not cause changes in the chemical shifts but affect the relaxation times of the molecules, thereby causing broadening of the n.m.r. signals. The effect of these cations on the I3C-n.m.r. spectra of carbohydrates has been discussed elsewhere in this Volume.37 Because they provide so much more information in their n.m.r. spectra, lanthanide complexes have been used, with great success, as probes for the study of biologically important Ca2+ c~rnplexes.~~ The assumption was made that the complexing site for the two complexes is the same, and the strength of complexing approximately equal. This approach should be applied to uncharged carbohydrateswith some caution: comparisonofthe t.1.c. RFvalues for Ca2+and La3+showed substantial discrepancies, particularly for alditols and glyco~ides.'~ (36) J. A. Peters, J. Mugo. Reson., 68 (1986) 240-251. (37) K. Dill and R. D. Carter, Adv. Curbohydr. Chem. Biochem.. 47 (1989) 125- 166. (38) R. J. P. Williams, Q. Rev. Chem. Soc., 24 (1970) 331-365.

12

STEPHEN J. ANGYAL

4. Other Methods

Complex-formation with a cation does not, in itself, affect the optical rotatory power of a carbohydrate.I0 However, complex-formation is often accompanied by a change ofconformation that causes a change in the optical rotation. For example, the rotation of D-glucitol and, to a lesser extent, of D-mannitol is affected39by the presence of cations, in the order Na+ < Mg2+< Zn2+< Ba2+< Sr2+C Ca2+(see Section II1,l). The optical rotation of methyl P-D-ribopyranosideand PD-lyxopyranoside would, undoubtedly, be substantially changed by complex-formation, but this experiment has, apparently, not been reported. Therefore, a change of optical rotation on addition of a salt can be regarded as proof of complex-formation,but lack of such change does not necessarily indicate that no complex is formed. Complex-formation has been shown to occur between some methyl glycosides and Na+, Ca2+,and La3+ions by a microcalorimetric technique,@ and the enthalpies of interaction have been detem~ined.~' This method also gives information on the stoichiometry of the complex. Osmotic coefficients, determined by the isopiestic comparison method, of systems containing an electrolyte and a non-electrolyte have been shown to offer simple and sensitive means for the study of weak complexing interactions in aqueous solution.42 The infrared spectra of cation - sugar complexes have been extensively in~estigated.~~ However, these studies have not provided any useful correlation between the spectra and the structure, configuration, or strength of the complexes. Complex-formation between Eu3+and some methyl glycosides has also been studied by luminescence excitation spectroscopy." 111. THESTRUCTURE OF CATIONCARBOHYDRATE COMPLEXES

1. Complexing Sites The proposal that strong complexingoccurs between cations and a contiguous a,e,a sequence of hydroxyl groups in carbohydrates was first based simply on the observation that many compounds having such an arrangement complex well, and many others lacking such an arrangement complex (39) J. G. Dawber and G. E. Hardy, J. Chem. Soc., Faruday Trans. I , 80 (1984) 2467-2478. (40) A. V e d a and H. Ldnnberg, Acta Chem. Scund., Ser. A, 35 (1981) 123- 126. (41) A. Vesala, H. Lonnberg, R. Kippi, and J. Arpalahti, Curbohydr.Res., 102 (1982) 312315. (42) A. Vesala and R. U p p i , Finn. Chem. Left.,13 (1986) 27-32. (43) H. A. Tajmir-Riahi, Biophys. Chem.. 23 (1986) 223-228, and references cited therein. (44) A. V e d a and R. Kippi, Polyhedron, 4 (1985) 1047- 1950.

COMPLEXES OF METAL CATIONS WITH CARBOHYDRATES

13

poorly, as shown by electrophoretic mobilities3 The proposed structure was later confirmed by X-ray crystal structure determinations (see Section III,3) and n.m.r. studies (see Section 11,3). This is the most favorable arrangement for complex-formation (apart from the rare, 1,3,5-triaxial arrangement). Similar to this arrangement is the cis,cis-1,2,3-triol grouping on a fivemembered ring, also confirmed by X-ray and n.m.r. studies; in a twist conformation, the three oxygen atoms are in a geometrical arrangement similar to the a,e,a arrangement.45The three oxygen atoms in the complex are then almost equidistant from the cation, and that distanceis favorablefor complex-formation with larger cations; the smaller ones, however, form only weak complexes. A compound may not possess one of these complexing sites in its most stable conformation but only in a less favored one; it may then form a complex after having changed into the latter. An example is P-D-ribopyranIn aqueous solution, this sugar is a conformationalmixture of 75% ofthe4C,(8a) and 25%ofthe 'C4(8b) forms, as shown by 'H-n.m.r. spectroscopy. In 1.27 Mcalcium chloride solution, over 60%ofthe pyranose is in the 'C4form, most of it as the complex with Ca2+.Because thep-pyranose has to change into a higher-energy form in order to react with the cation, it will form a complex to a lesser extent than does the a anomer, which has an a,e,a sequence in its preponderant conformation.

-

OH

"

Y HO

8a

O

H

f + OH

HO

8b

If the difference in free energy between the non-complexing and the complexing conformation is large, complex-formation will not take place, as shown by the examples of myo-inositol and cis-cyclohexane-l,3,5-tnol. Complex-formation then does not supply sufficient free energy for the conformational change to take place. All of the alditols form complexeswith cations, but the extent of complexformation varies considerably,the Mi values4' of hexitols ranging from 0.24 (45) S . J. Angyal, Aust. J. Chem., 25 (1972) 1957- 1966. (46) R. E. Lenkinski and J. Reuben, J. Am. Chem. SOC.,98 (1976) 3089-3094. (47) R,refers to mobility on a cation-exchange,t.1.c.plate in the Ca2+form15;M ivaluesgive the electrophoreticmobility relative to cis-inositol in a 0.2 M solution of Ca(OAc), in 0.2 M aqueous acetic acid.l'

STEPHEN J . ANGYAL

14

for iditol to 0.09 for allito1.48By rotation around carbon- carbon bonds, any alditol can assume a (+)-gauche-(-)-gauche arrangement of three oxygen atoms (which is analogousto the a,e,aarrangement in six-membered rings), but the energy required for this process will depend on its c o n f i g ~ r a t i o n . ~ ~ , ~ ~ When three consecutive carbon atoms of an alditol have the threo-threo configuration, the complexingconformation (9)has no unfavorable interaction other than that between the two terminal oxygen atoms; the arrangement is favorable for complex-formation. An erythro-threo configuration produces a complexingarrangement (10) in which there is a gauche interaction between two segments of the carbon chain; this is somewhat less favorable. An erythro-erythro sequence produces a complexingarrangement (11) in which there is a 1,3-parallel interaction between two segments of the carbon chain; this is an unfavorable conformation, and erythro-erythroconfigurations do not give rise to significant proportions of complexes. Similar considerationsapply to the terminal hydroxyl groups when they are involved in c ~ m p l e x - f o r m a t i o nThe . ~ ~more ~ ~ ~ threo pairs of hydroxyl groups there are in the alditol, the stronger will be its complexes. The site of complexformation is indicated by the n.m.r. spectra, which also show that the conformation has to change from the preponderant one before complex formation can occur (9- l l). If one of the three hydroxyl groups is replaced by a methoxyl group, complexingstill occurs, but it is somewhat weaker (Section 111,4). If all three hydroxyl groups are methylated, complex formation becomes negligible.22 I HOFH HFOH

I HFOH HOCH HCOH I

HO

HCOH I C

9

HO

10

HCOH HYOH HFOH I HO

I

C

kc C

11

Electrophoretic mobility shows that even two consecutive hydroxyl groups can form a complex, although a very weak one, if they can approach (48) S. J. Angyal, D. Greeves, and J. A. Mills, Aust. J. Chem., 27 (1974) 1447- 1456. (49) A. P. G. Kieboom,T. Spoormaker,A. Sinnema, J. M. van der Toorn, and H. van Bekkum, Red. Trav. Chim. Pays-Bas, 94 (1975) 53-59.

COMPLEXES OF METAL CATIONS WITH CARBOHYDRATES

15

each other sufficiently.Both cis- and trans-diols on six-membered rings Will form weak complexes, but only cis diols on five-membered rings.13 Although complex-formation usually involves vicinal hydroxyl groups, there are rare instances where this is not so; they are discussed herewith. Of all polyols, cis-inositol(1)forms the strongestcomplexeswith cations. It has, in each of its equivalent chair forms, three syn-axialhydroxyl groups that are well disposed for complex formation; however, it also contains three a,e,a sequences of hydroxyl groups, and it is not clear how much the syn-axial hydroxyl groups contribute to complex formation. A compound has become available that has three axial hydroxyl groups but no equatorial ones, namely, scyZZo-inositol monoorth~formate~~ (12). In t.l.c., it has5' RF 0.13, compared to 0.07 for cis-inositol and 0.42 for epi-inositol; the triaxial arrangement, therefore, forms a better complexing site than the a,e,a arrangement. A similar conclusion can be drawn from the europium-inducedshifts. In the spectrum of scyZZo-inositol monoorthoformate, only the signal of the equatorial hydrogen atoms5' shifts upfield on addition of Eu3+;in that of cis-inositol,both signalsshift, but that of the equatorial hydrogen atom shifts much more than that ofthe axial one.zzThe induced shift ofthe axial signal is due to complexing at the a,e,a sites; hence, it follows that there is more complexing at the syn-axial than at the a,e,a site. HO

OH

12

Even small cations complex well with cis-inosito12z;for example, on a t.1.c. plate in the Cu2+ the RF value of cis-inositol is 0.21, whereas other carbohydrates have RF > 0.80. Carbohydrates having three syn-axial hydroxyl groups are, however, rare; an example, somewhat on the borderline of carbohydrate chemistry, is muellitol [ 1,3,5-tri(3-methylbut-2-enyl)-scylloinositol] (2) which complexes strongly with cations, mainly at the secondary hydroxyl groups. Another example is 2-deoxy-2-methyl-epi-inositol (6), which forms surprisingly strong complexesz2(Mi 0.34). In its more-stable conformation,there are no good complexingsites, but inversion to the other

'*

(50) H. W. Lee and Y.Kishi, J. Org. Chem., 50 (1984) 4402-4404. ( 5 1) S. J. Angyal, unpublished results.

STEPHEN J. ANGYAL

16

chair form provides a triaxial and an a,e,a site, besides relieving the O / / C interaction. Complexing occurs at both sites, as shown by the 1.i.s. Methyl a-D-lyxofuranoside (13)forms a weak complex with cations: R, 0.65. L.i.s. showedI5that complexing occurs on 0-2, 0-3, and 0 - 5 , as first suggested by Vesala and coworkers.52In this complex, the three cationoxygen bonds would probably not be of equal length. La3+has a greater tendency to form such a complex than has Ca2+or other divalentcation~."O.~~

13

1,6-Anhydro-P-~-glucopyranose (14) complexes with cations: RF0.60. In this compound, there are two syn-axial hydroxyl groups, but there is no equatorial hydroxyl group between them; the syn-axial hydroxyl groups, in themselves, do not provide a good complexing site, as shown by the weak complexing of rnuco-inositol. N.m.r.-spectral evidence indicated that the ring-oxygen atom is also involved in complex-formation:0-2,O-4, and 0-5 form an almost equilateral triangle.l' A similar involvement of the ringoxygen atom of a five-membered ring, in addition to the two vicinal hydroxyl groups (0-3 and 0-4), has been postulated in order to explain54the 0.18) of 2,5-anhydrogalactitol(15). The evidence for this complexing (Mi postulate is not convincing: complexingmay just as well be due to coordination at 0-3,O-4, and 0-1, and 0-3,0-4, and 0-6, as in methyl a-D-lyxofuranoside (see earlier). In the crystals of the adduct of 1,4-anhydroerythritol with sodium perchlorate and sodium iodide,55such tridentate complexing does not occur. Methyl P-D-psicofuranoside(16)does not formI5a complex with Ca2+;coordination at 0-1,O-3, and 0-4 (of the a-D-lyxofuranosetype) 0

15

16

(52) A. Vesala, R. G p p i , and P. Lehikoinen, Finn.Chem. Lett., 10 (1983) 45-50. (53) H. Lonnberg, A. Vesala, and R. Kappi, Carbohydr. Res., 86 (1980) 137- 142. (54) S. J. Angyal and Y. Kondo, Aust. J. Chew., 33 (1980) 1013- 1019. (55) R. E. Ballard, A. H. Haines, E. K. Noms, and H. G. Wells, Acta Crystallogr., Sect. B, 30 (1974) 1590-1593; 32 (1976) 1577-1578.

COMPLEXES OF METAL CATIONS WITH CARBOHYDRATES

17

could occur, as well as at 0-3, 0-4, and 0-5 (ring-oxygen atom), but both would require unfavorable conformations for the furanose ring. An interesting site, because it occurs in ketopyranoses,consists of geminal hydroxyl and hydroxymethyl groups and a vicinal, equatorial, hydroxyl group; these oxygen atoms can also assume a relationship similar to the a,e,a arrangement. Weak complexing does, indeed, occur at such sites, as shown by C-(hydroxymethy1)-scyllo-inositol (17) (Mi 0.12) and 2-C-(hydroxymethyl)-myo-inositol(18) (Mi 0.08); complexing is stronger when the hydroxymethyl group is axial.” CH2OH

OH

I

I

OH

HO @OH

Ho&CH~OH

HO

OH

HO

17

18

All ofthese complexes are tridentate: in a cyclic monosaccharide, no more than three oxygen atoms can coordinate to one cation; but when there are two monosaccharide moieties in a molecule, complexing can occur at more than three oxygen atoms. A rather curious example is that of di-B-D-fmctopyranose 1,2’: 1’,2-dianh~dride~~ (19). This compound complexes readily with cations, and gives crystalline adducts with CaC12, SrCl,, BaC12, and LaCl,, whereas five isomeric D-fructose dianhydrides do not form complexes. Because the anomeric effect in the chair form would be unfavorable, the central 1,Cdioxane ring is in a skew form5’ which makes complexformation possible at 0-1,O-3,O-1’, 0-3’. In the aJ3anomer, the 1,Cdioxane ring is in a chair form, and 0-3 and 0-3’ are not close to the ring-oxygen atoms; in the furanose anomers, these oxygen atoms point outwards, no matter whether the 1,4dioxane ring is in a chair or a skew form. The L-sorbose analog, di-a-L-sorbopyranose1,2’: 1’,2-dianhydride,also forms a complex with cations. These particular anomers are, therefore, readily separated from the others, formed simultaneously, on a cation-exchange column.

19

( 5 6 ) S . J. Angyal, D. C. Craig, J. Defaye, and A. Gadelle, unpublished results. (57) C. Pedersen, personal communication.

18

STEPHEN J. ANGYAL

a-D-Allopyranosyl a-D-allopyranose (20) is a unique disaccharide: each moiety has an a,e,a sequence of oxygen atoms. As expected, it forms a quinquedentate complex with calcium ions.58This (unnatural) disaccharide is, however, the only one so far known to form a complex involving oxygen atoms of both moieties. (The proposeds9sexadentate K+ complex of lactose is sterically impossible.) One wonders whether there are such tailor-made complexingdi- or oligo-saccharidesin Nature; none has been reported so far. The cyclodextrins, of course, are such tailor-made compounds, but their combination with cations (and other molecules) depends on inclusion, rather than on specific coordination.

H

20

Disregarding the aforelisted, unusual examples, the effectiveness of complexing sites is listed in descendingorder: 1,3,5-triaxialtriol > a,e,atriol on a six-membered ring > cis-cis triol on a five-membered ring > acyclic threothreo triol > acyclic threo pair adjacent to a primary hydroxyl group > acyclic erythro-threo triol > acyclic erythro pair adjacent to a primary hydroxyl group > acyclic erythro-erythrotriol > cis diol on a five-membered ring > cis diol on a six-memberedring > trans diol on a six-memberedring. Oligo- and poly-saccharides also form complexes with cations, but very little is known about their nature. As the carbohydratechain is lengthened, the extent of complexing increases, presumably not only owing to the increasing multiplicity of complexing sites but also because of possible interchain cross-linking. Thus, although D-xylose forms only very weak complexes, D-xylan reacts with a variety of cations; the manifestation of this complex-formation is, initially, a decrease and, at higher cation :sugar ratios, an increase, in solubility.60The decreaseis probably caused by cross-linking; the increase, by extensive complexation at many sites. That the change in solubilityis not due merely to the presence of cationsis shown by the absence (58) J. Ollis, V. J. James, S. J. Angyal, and P. M. Pojer, Curbohydr. Res., 60 (1978) 219-228. (59) R. M. Munavu, B. Nasseri-Noon, and H. Szmant, Carbohydr.Rex, 125 (1984) 253-263. (60) J. D. Blake and G . N. Richards, Curbohydr. Rex, 17 (1971) 253-268.

COMPLEXES OF METAL CATIONS WITH CARBOHYDRATES

19

of such an effect when (non-complexing) ammonium ions are added. Sodium and potassium ions have very little effect. The greatest change in solubility is caused by Cu2+,Pb2+,La3+,and A13+ions. Precipitation by the addition of salts is a standard method for fractionating, and purifying, polysaccharides.61Cupric acetate is the most suitable salt for this purpose. Dextran forms crystalline, spherulitic complexes with Ca2+,Ba2+,and La3+,but not with Mg2+and Na+. The optical rotation of dextran in aqueous solution is considerably increased6Iaby the addition of La3+ and Ba2+,but only slightly by Ca2+,Mg2+,and Na+ acetates. Complex-formation appears to alter the conformation of the polymer chain. A 20% solution of dextran will dissolve 5% of Fe3+;this solution is used for the treatment of anemia.62 Amino sugars behave differently. The nitrogen atom has a great tendency to complex with transition-metal ions, and, if there is a free hydroxyl group in the vicinity, bidentate complexes will be formed that are much stronger than those of nitrogen-free sugars.63~Glucosamineforms complexes with cations in the order: Cu2+> Pb2+> ZnZ+> Ni2+> Cd2+;Ca2+and MgZ+do not form complexes of this type. Nagabushan and coworkers64put this reaction to practical use: complex-formation with cupric acetate, in a mixture of water and N,N-dimethylformamide, can prevent acetylation of amino groups that are close to a hydroxyl group. Those that are not, are acetylated; thus, selective N-acetylation can be achieved. Methyl 2-acetamido-2-deoxy-~-gluco-and -galacto-pyranoside appear to complex with lanthanide ions at the acetamido (see also, N-acetyl-a-neuraminic acid, Section IIIS).

-

2. Cations

Metal cations can be arranged45according to their increasing tendency to form complexes with carbohydrates: Li+, K+, Rb+, Na+, Mg2+,Cd2+,Sn”, Ag+, Yb3+,Ba2+,S P , Ca2+,Pb2+,and La3+.The order varies accordingto the substrate. The weak complexing with a cis pair of hydroxyl groups on a six-membered ring was found to increase in the order K+ Mg2+ CdZ+<

-

-

(61) J. K. N. Jones and R. J. Stoodley, Methods Carbohydr. Chem.. 5 (1965) 36-38. (6 la) C. Guizard, unpublished results. (62) J. S. G. Fox, R. E. King, and G. F. Reynolds, Nature, 207 (1965) 1202- 1203. (63) M. Miyazaki, S. Nishimura, A. Yoshida, and N. Okubo, Chem. Pharm. Bull., 27 (1979) 532-535, and references cited therein. (64) T. L. Nagabushan,A. B.Cooper, W. N. Turner, H. Tsai,S. McCombie,A. K.Mallams, D. Rane, J. J. Wright, P. Reichert, D. L. Boxler, and J. Weinstein, J. Am. Chem. Soc., 100 (1978) 5253 - 5254; T. L. Nagabushan,W. N. Turner, H. Tsai, R. S. Jaret, D. Schumacher, J. K. Jenkins, and J. S. Chu, Carbohydr. Rex, 130 (1984) 243-249. (65) K. Izumi, Carbohydr. Res.. 170 (1987) 19-25.

STEPHEN J. ANGYAL

20

-

-

Zn2+ Na+ < La3+ S P < Ca2+;these results were obtained by observing, with the aid of Raman spectra, the effect of the cations on the a :P ratio of D-glucose.66A certain type of complexing site is sometimes more favorable to one cation than to another; for example, with Ca2+,D-mannitol complexes more strongly than ~-arabinitoP’,but the opposite is true for Baz+. Generally, univalent cations all complex weakly; divalent ones complex better, and trivalent ones form the strongest complexes. The ionic radius of the cation is crucial: the best radius for complex-formation is 100- 110 pm (Na+,Ca2+,La3+).Larger cations complex somewhat less strongly. Smaller cations complex weakly; thus, there is considerable destabilization of the complex on going from Na+to Li+(68pm), from Ca2+to Mg2+(66pm), from Pbz+to Sn2+(93 pm), from Cd2+to Znz+(74 pm), and from La3+to Y3+(92 pm), as the radius becomes markedlf5 less than 100 pm. Cu2+(72 pm) also forms very weak complexes. La3+ions show preferences different from those of the other cations: they form stronger complexes with threo-threo hydroxyl groups in alditols, but weaker ones with OH,OH,OMe sequences on a ring, than other cations15; Yb3+and Eu3+behave ~ i m i l a r l y . ~The ’ , ~ tripositive ions appear to have a greater propensity to change the conformation of the substrate, but they are more hindered by methoxyl groups than are the other cations. The differing conduct of La3+is illustrated in Table 111, which shows the behavior of four methyl pentofuranosides in column chromatographywith several cations.42 Methyl a-D-ribofuranoside complexes at 0-1, 0-2, and 0-3; methyl a-Dlyxofuranoside,at 0-2,O-3, and 0-5; methyl P-D-ribofuranoside,at the 0-2, 0-3 pair; and methyl P-D-arabinofuranoside does not form a complex. Stannous chloride forms “complexes,” involving two hydroxyl groups, with carbohydrates in organic solvents; these tin compounds have acquired TABLEI11 Retention Volumes (cm-’)of Some Methyl Aldofuranosides on Cationexchange Columns42 Glycoside

Ca2+

Sr2+

Ba2+

Pb2+

La”

Methyl p-D-arabinofuranoside Methyl FD-ribofuranoside Methyl a-D-lyxofuranoside Methyl a-D-ribofuranoside

27.20 33.60 38.80 52.80

28.90 32.15 39.80 59.10

25.00 26.70 31.85 56.60

27.05 30.85 38.15 66.75

27.15 28.35 41.50 38.40

(66) F. Franks, J. R. Hall, D. E. Irish, and K. Noms, Curbohydr. Res., 157 (1986) 53-64. (67) L. PetruS, V. Bilik, L. Kuniak, and L. StankoviE, Chem. Zvesti, 34 (1980) 530-536.

COMPLEXES OF METAL CATIONS WITH CARBOHYDRATES

21

increasing importance as promoters of selective alkylations,68but they are not really complexes, as the tin atom is bound by covalent bonds, and the formation of this bond is slow: separate signals are seen in the n.m.r. spectrum for the complexed and the uncomplexed molecule^.^^ A ferric - D-fructose complex of definitecomposition but unknown structure has been described70;it is soluble in water and may be suitable for the treatment of anemia. The behavior of copper ions deserves some discussion. Complexing with cuprammonium ions was the historically important method for determining the conformations of sugars in their pyranose forms. For the first time, Reeves7’established a correlation between complexingability and structure. Interestingly, among the compounds investigated by Reeves, 1,6-anhydroQ-D-mannopyranose formed the strongest complex,72and it is the only one of them that has an a,e,a sequence of three oxygen atoms. Reeves found the stabilityconstant to be 250 M-’, two orders of magnitudegreater than those of complexes with other cations. Cuprammonium solution is strongly alkaline; clearly, the complexes are formed, not with neutral sugars, but with their anions. In neutral solution, cuprammonium ions complex73about as strongly as Caz+. Because it is small, Cu” (ionicradius, 72 pm) forms very weak complexes. The enthalpy of its interaction with glycosides was found to be small, indicating negligible c~mplex-formation~~; the electrophoretic mobilities of sugars in cupric sulfate solution are in~ignificant~~; and on a t.1.c. plate immersed in a cupric sulfate solution, only cis-inositol is retained to any extent.73However, electrophoretic mobilities in a solution of cupnc acetate have been reported,I2sugars have been separated on a cation-exchangeresin treated with cupric acetate,12and, on t.1.c. plates immersed in cupnc acetate solution, most carbohydrates are strongly retained.14Clearly, the cation that complexes in these cases is not Cu2+;although CuOAc+is the preponderant ionic speciesin solutionsof copper a~etate,’~ it appears73that the complexing cation is [Cu2(OH),I2+.Complexingis particularly strong with alditols, and there is some indication that quadndentate complexes are formed with suitable carbohydrate^.^^

(68) (69) (70) (71) (72) (73) (74) (75)

S. David and S. Hanessian, Tetrahedron, 41 (1985) 643-663. J. Alfdldi, R. Toman, and C. Peciar, Curbohydr. Res., 105 (1982) 258-265. S. A. Barker, P. J. Somers, and J. Stevenson, Curbohydr. Res., 36 (1974) 331 -337. R. E. Reeves, Adv. Curbohydr. Chem., 6 (1951) 107-134. R. E. Reeves, J. Am. Chem. SOC.,73 (1951) 957-959. S. J. Angyal, Absfr. Pup. Am. Chem. SOC.Meet., 196 (1988) c ~ ~ e - 4 6 . J. A. Mills, unpublished results. M. Lloyd, V. Wycherley, andC. B. Monk, J. Chem. SOC., (1951) 1786-1789.

22

STEPHEN J. ANGYAL

3. The Structures of Crystalline Complexes There are numerous examples known of crystalline adducts between sugars and various inorganic salts in stoichiometric pr0portions.4,~~ The crystal structures of many such adducts have been determined by X-ray crystallography. In these crystals, the cations are coordinated to oxygen atoms of the sugar molecules and, in most cases, also to water molecules, and the structure is reinforced by hydrogen bonds. These crystal structures provide valuable insight into the nature of cation - hydroxyl coordination; but this is an exaggerated picture, because, in the crystal, where there are only a few, and sometimes no, water molecules, the cation is forced to accept electrons from oxygen atoms which, in solution, may be only very weakly coordinated. When the crystal is dissolved in water, water molecules will usually displace the hydroxyl groups from their coordination with the cations, and only the stronger coordination will persist. Existence of such crystalline adducts, therefore, does not constitute evidence that complexformation occurs to any significant extent in solution. An example is the historically important sucrose NaBr 2 H,O, one of the first sugar structures determined by X-ray ~rystallography.’~ In solution, there is no detectable complex-formation between sucrose and sodium ions. However, when complex-formation is known to occur in solution, and a crystalline adduct can be obtained, it is very probable that the cation will be coordinated to the same site in the crystal as in solution. This has proved to be true in all cases investigated so far. Additional points of coordination are usually found in the crystal structure, but the site where the cation is coordinated to three (or more) oxygen atoms of a sugar is also the site of coordination in solution. An X-ray crystallographic investigation of the structure provides geometrical details of the coordination not available by other methods. Seven structures have been determined so far (most of them in Sydney) in which tri- (or higher) dentate coordination occurs. (Tridentate complexing has also been found in the crystal structures of some alduronic salts; see Section 111,5.) ~~ The first of these is /I-D-mannofuranose * CaC1, * 4 H,O. This a d d ~ cist of interest because aqueous solutions of D-mannose contain only 0.3%of the p-furanose form (21) in eq~ilibrium,’~and yet the adduct of this form will crystallize from a very concentratedsolution containingan excessof calcium

-

-

(76) For a review of adducts with calcium salts, see W. J. Cook and C. E. B u g , in B. Pullman and N. Goldblum (Eds.),Metal-Ligandlnteractions in Organic Chemistry and Biochemistry, Part 2, Reidel Publishing Company, Dordrecht (Holland), 1977, pp. 231 -256. (77) C. A. Beevers and W. Cochran, Proc. R. SOC.London, Ser. A, 190 (1947) 257-272. (78) D. C. Craig, N. C. Stephenson, and J. D. Stevens, Curbohydr.Res., 22 (1972) 494-495. (79) D. J. Wilbur, C. Williams, and A. Allerhand, J. Am. Chem. Soc., 99 (1977) 5450-5452.

COMPLEXES OF METAL CATIONS WITH CARBOHYDRATES

23

chloride.s0 (Daleso also described a complex of P-D-mannopyranose with calcium chloride, but repeated attempts to obtain this compound have since failed.s1)In this adduct, the calcium ion is coordinated to 0-1,O-2, and 0-3 of one D-mannose molecule, to 0-5 and 0-6 of another one, and to three water molecules; the resulting, eight-fold coordination is the most common one for calcium ions in carbohydrate adduckS2The cation is slightly closer to the water molecules than to the hydroxyl groups (this is generally true for sugar -calcium chloride adductss2)and slightlycloser to 0-2 than to 0-1 and 0-3; but none of these distances differs much from the average (244 pm). The furanose ring is in the E2 conformation. Methyl P-D-mannofuranoside CaC12 3 H 2 0is isostructuralwith the preceding adduct. The crystal structures3is closely similar, the methyl group replacing one water molecule (the one not coordinated to the cation). The geometrical arrangement of 0-1,O-2, and 0-3 is not significantly different in the two crystal structures, and the Ca-0-1 distance is only slightly increased (by 5 pm) when a methyl group replaces the anomeric hydrogen atom.

-

-

21

An example containing a pyranose ring is methyl D-glycero-a-D-@loheptopyranoside* CaC12 H 2 0 (22). The cation is here coordinated to the a,e,a sequence of 0-1, 0-2, and 0-3 of one molecule and also to 0-4,O-6, and 0-7 of another onew;it is, therefore, an example ofthe so-called “double complexing” (see Section IV,2) in a crystal structure. This is the only instance. in a crystal structure, of tridentate complexing with three hydroxyl groups that are not on consecutive carbon atoms. The orientation around the C-5 - C-6 bond is such that 0-4 and 0-6 have the same relationship as syn-axial groups on a six-membered ring, and the position of 0 - 7 relative to 0-4 and 0-6 is very similar to that of 0-2 relative to 0-1 and 0-3. All of the hydroxyl groups, as well as 0-1, are involved in complex-formation.

-

(80) (81) (82) (83) (84)

J. K. Dale, J. Am. Chem. SOC.,51 (1929) 2788-2795. J. D. Stevens, unpublished results. M. L. Dheu-Andries and S. Perez, Carbohydr.Rex, 124 (1983) 324-332. D. C. Craig and J. D. Stevens, unpublished results. D. C. Craig and J. D. Stevens, Crysf.Sfucf.Commun., 8 (1979) 161 - 166.

24

STEPHEN J. ANGYAL

22

The geometry of the a,e,a complexing site can be particularly well exploreds5by the study of epi-inositol SrCl, 5 H,O, because the crystal structure of epi-inositol(6)has also been determined.86The cation is coordinated to 0-2,O-3, and 0-4 of one epi-inositolmolecule (which constitute the a,e,a sequence)and to 0-1 and 0-6 of another molecule, and to four molecules of water; it is an example of the rare, nine-fold coordination. The a,e,a oxygen atoms, 0-2, 0-3, and 0-4, form an almost equilateral triangle. The S r - 0 distances are 263 k 8 pm; here, again, the cation is closer to the water molecules than to the hydroxyl groups. The conformation of the inositol in the complex is the same as that of free epi-inositol. There are, however, some differencesin the intramolecular0 -0 distances;in particular, the distances between 0-2,O-3, and 0-4 are shorter by 13- 16 pm. The distance between the two syn-axial oxygen atoms has been shortened from 296 to 282 pm. (In the complex of methyl D-glycero-a-D-gulo-heptopyranoside, the corresponding 0-0distance is 283 pm.) The distance between 0 - 1 and 0-6, which are coordinatedto another Sr2+ ion, is also shorterthan in epi-inositol. These results show that coordination with a cation lessens the non-bonded interactions between (syn-axialor gauche) participating oxygen atoms, and have shown this to be thereby stabilizes the complex; Bugg and Cook76*87 generally true for cation complexes of polyols. The fifth example is the sodium salt of a-L-guluronic acids8(23). In the crystal, two cations are coordinated to each sugar molecule: one to 0-1,O-2, and 0-3 and another to 0-4,O-5, and one of the carboxylateoxygen atoms. However, the Na-0-2 and the Na-0-5 distances (285 and 272 pm) are larger than the others, and it is not certain that they represent coordination.

-

-

(85) R. A. Wood, V. J . James, and S. J. Angyal, Acta Crystullogr., Sect. B., 33 (1977) 22482251. (86) G . A. Jeffrey and H. S. Kim, Ada Crystallogr., Sect. B, 27 (1971) 1812- 1817. (87) C. E. B u g and W. J. Cook, Chem. Commun., (1972) 727-729. (88) F. Mo, T. J. Brobak, and I. R. Siddiqui, Curbohydr.Res., 145 (1985) 13-24.

COMPLEXES OF METAL CATIONS WITH CARBOHYDRATES

25

23

A monosaccharide cannot provide more than three oxygen atoms for coordination to a cation, but a molecule containing two monosaccharide units (for example, a disaccharide) can offer a site for four- or five-fold coordination. The sixth example is a disaccharide especially synthesized in order to provide a site for five-fold coordination. In the a-D-allopyranosyl C U - D - ~ ~ ~ O pyranoside molecule, both moieties have an a,e,a sequence, and it was predicted that a cation would coordinateto five oxygen atoms. The disaccharide readily forms a crystalline adduct with calcium chloride and five molecules of water, and its X-ray crystallographic showed that the cation is coordinated to 0-1,0-2, 0-3, 0-2’, and 0-3’ (20). Such coordination requires practically no change in the conformation of the disaccharide, except for a decrease in the distances between the participating oxygen atoms. The distances between 0-1,O-2, and 0 - 3 are closely similar to those of the other three calcium chloride complexes discussed previously. In addition, the cation is coordinated to four water molecules; the resulting nine-fold coordination is rarely encountered. In this complex, in contrast to the five already discussed, each cation is coordinated to only one sugar molecule. In the addition compound of an epimer, a-D-glucopyranosyla-D-glucopyranoside -CaBr2 H20, which does not provide a,e,a sequences, the cation is coordinated to the hydroxyl groups offuur sugar molecules.89 In a-D-allopyranosyl a-D-allopyranoside CaC1, 5 HzO, the C- 1- 0-5 and the C-1’-0-5’ bond-lengths are unusually short (139.5 and 139.7 pm, respe~tively).~~ Jeffrey and coworkersw observed that the anomeric C - 0 bonds in aldoses are shorter than the other C - 0 bonds, owing to their partial double-bond character caused by a drift ofelectronstowards the ring-oxygen atom. In this complex, the electron attraction of the cation gives rise to the opposite effect: the C- 1 - 0-1 bonds are of normal length, but the C- 1- 0-5 bonds are shortened. The seventh crystal structure was studied in order to reveal details of an

-

-

-

(89) W. J. Cook and C. E. Bug, Curbohydr. Rex, 31 (1973) 265-275. (90) H. M. Berman, S. S. C. Chu, andG. A. Jeffrey, Science, 157 (1967) 1576-1577.

STEPHEN J. ANGYAL

26

unexpected case of c~mplex-formation.~~ Di-/?-D-fructopyranose1,2' : 1',2dianhydride (19),alone among the dianhydro-D-fructoses, forms a complex with cations in solution (see Section II1,l). The crystal-structure analysis of the SrC1, complex showed that complex-formation occurs at 0-1 and 0 - 3 of each D-fructose moiety; the complexing site consists of two overlapping sequences of three oxygen atoms on consecutive carbon atoms. The central dioxane ring is in a skew form; the resulting quadridentate coordination is quite strong (RF0.5 1). It would be of interest to study the crystal structures of alditol-cation complexes, because, in them, the conformationwould differ from that of the free alditol. However, no suitable example is known. Attempts to prepare crystalline complexes from D-glucitol and various salts have not been successful. There are other calcium chloride adducts which undoubtedly have similar tridentate complexing structures,but which have not yet been investigated; for example, a-D-gulopyranose CaC12 H 2 0 (Ref. lo), a-D-allopyranose CaC12 ? H 2 0(Ref. 9 I), and methyl a-D-gulopyranoside CaC12 2 H 2 0 (Ref. 92). It is significantthat, in the first two adducts, the sugars occur in the a form, whereas, in solution, the /Iforms are preponderant. Even if tridentate complexation does not occur, the structures of the crystalline adducts of sugars and salts provide useful information on the nature of complex-formation. In all known cases, there is coordination to at least two oxygen atoms of a sugar; frequently, there are, in the crystal structure, sugars linked to the cation by only one oxygen atom, but there is always one sugar molecule linked by at least two. In almost every instance, the two oxygen atoms form a cis pair on consecutive carbon atoms; a trans pair may be coordinated,but only if a coordinated cis pair is also present in the crystal structure. Apparently, coordination by a single oxygen atom or a trans pair offers insufficient inducement for crystals of an adduct to be formed; but a cis pair can readily serve as the core of such a structure. Sucrose NaBr 2 H 2 0constitutesa rare exception: the cation is coordinatedto 0 - 4 and 0-6 of the D-glucosyl moiety.76

-

-

-

-

-

-

m

4. Stability Constants

The measure of the strength of complex-formationbetween a cation (Xu+) and a carbohydrate(Carb)is the stability (or formation) constant, defined as K = [Carb X"+]/[Carb][X"+].The determination of the stability constants of cation - sugar complexes has not been wholly satisfactory. The constants being comparatively small, their accurate determination is difficult. The

-

(91) L. D. Hall, C. M. Preston, and J. D. Stevens, Curbohydr. Rex, 41 (1975) 41 -52. (92) H. S. Isbell, Bur. Stand. J. Rex, 8 (1932) 1-8.

COMPLEXES OF METAL CATIONS WITH CARBOHYDRATES

27

activitiesof the ions cannot be neglected at the high concentrationsrequired for complex-formation, but the activities of the complex cations are not known and, therefore, concentrations,rather than activities, have been used in calculating the “stability constants.” There is also the question of the stoichiometry of the complexes. In concentrated solutions, and with an excess of the sugar, undoubtedly some 1 :2 complexeswill be p r e ~ e n t , ~in, ~ ’ addition to the 1 : 1 complexes; this has been shown to apply particularly to cis-inositol, which has the highest stability constant.23Most authors have . ~ with ~.~~ Pb2+,three methyl glycofuobserved only 1 : 1 c o m p l e ~ e s , ~but, ranosides were found to form 1 :2 complexe~.~~ In most cases, only 1 : 1 complexes were taken into account in the calculations, without much detriment to the results.94 Despite these difficulties, approximate “stability constants,” calculated for 1 : 1 complexes, are useful, because they allow comparison of the complexing ability of various cations and various polyols, and they enable the calculation of the approximate extent of complex-formation at different concentrations. Two methods have been mainly used for the determination of the stability constants of cation - sugar complexes. The equilibriumcomposition of D-allose (and other aldoses in which 0-2 and 0-3 are cis) changes on addition of complexing cations, because only the Q anomer forms complexeswith them. From these changes, determined by study of n.m.r. spectra, the Kvalues can be ~alculated.~’ The other method involves potentiometricdetermination of the concentration ofthe cations at different concentrationsand cation :sugar In principle, the stability constants can also be determined from the 1.i.s. in the n.m.r. spectra; the “limiting shifts,” that is, the shift for complete conversion into the complex, have to be determined. However, as the sugars complex only weakly, complete conversion is not achieved, and extrapolation to the limiting shifts is uncertain. Hence, this method has rarely been used for the determination of cation - sugar complexes.46In one instance, the change of the spin - lattice relaxation-time of 23Na+,determined by pulsed n.m.r. spectroscopy, has been used to study the binding of Na+ to D-ribose and to cis-inositol.’ Stability constants for Ca2+have also been obtained by determining the solubility of calcium sulfate in the presence of various sugars?* (93) W. J. Evans and V. L. Frampton, Curbohydr. Res., 59 (1977) 571 -574. (94) M. TomaSkoviC, Z. Cimerman, Z. Stefanac, E. Pretsch, and J. Bendl, Microchem. J., 27 (1982) 372-379. (95) H. Lonnberg and A..Vesala, Curbohydr. Res., 78 (1980) 53-59. (96) L.-G. Ekstrom and A. Olin, Actu Chem. Scund., Ser. A , 31 (1977) 838-844. (97) J.-P. Morel and C. Lhermet, Can. J. Chem., 63 (1985) 2639-2643. (98) A. P. G. Kieboom, H. M. A. Buurmans, L. K. van Leeuwen, and H. J. van Benschop, Red. Truv. Chim. Pays-Bus, 98 (1979) 393-394.

STEPHEN J. ANGYAL

28

TABLEIV Stability Constants ( M - I ) of Cation-Sugar Complexes in Aqueous Solution Carbohydrate

Na+ 0.12

a-D-Allopyranose a-DAllofuranose a-D-Ribopyranose &D-Ribopyranose a-DRibofuranose a-DLyxofuranose Methyl a-D-ribofuranoside Methyl j?-D-ribofuranoside Methyl a-D-1 yxofuranoside Methyl a-Dxylofuranoside epi-Inositol a

96 9b 5b
Ca"

5.1-6.5 1.6- 5.4 4.6, 5.5 3,4, 1.4' 1.2, 1.4 0.8,0.7" 3.1, 4.4 5.3, 2.2" 0.8 1.2, 1206 0.3, 5b 0.9, 506 0.1,
Sr" 5.5 1.3

Ba2+

Pb2+

2.9 1.2

Law

References

10.4 8.7

45 45 45,95 46 45,95 46 45,95 46 46 53,95 52,53,95 53,95 52,53 3,23

6.7 1.5 4.3

3.6" 1.5 0.9"

3.4 5.7" 1.o

2406 36 370b 90b 2.1

1.8

At 52". In methanol.

-

-

Values for K of 5 M-I indicate strong complex-formation, 1 M-I represents weak complex-formation, and, when the value is 0.1, complexformation is negligible. To give some appreciation of the magnitude of the stability constant: if K = 5 M-I, the extent of complex-formation is 64% if the concentration of both the cation and the sugar is M, 27%, if it is 0.1 M, and 5% if it is 0.0 1 M. It is clear that complex-formationis significant only in concentrated solutions. As a very approximate guide, K = 5 corresponds to M i0.5 and RF0.35; K = 1 to M i0.2 and RF0.6; and K = 0.1 to M i0.05 and R, 0.8. The stability constant of D-ribose has been determined more frequently than that of any other sugar, presumably because it is the only naturally occumng aldose that complexes readily.w Most of the known stability constants are shown in Table IV. Some further stability constants are listed with herewith: rnyo-ino~itol~~ with Caz+, 0.2; 3-0-methyl-epi-ino~itol~~ CaZ+,1.6; 3-O-methyl-a-~-allo-pyranoseand - f ~ r a n o s ewith ~ ~ Caz+, 0.6 and 3.0, respectively; D-g~yCe~~a-D-gu10-heptOpy~nOSe45 with La3+, 1 1.5; methyl a-D-ribofuranoside4 with Eu3+,0.77; methyl a-D-lyxofuranoside4 with Ed+, 1.22; glycerol, D-mannitol, galactitol, and xylitollolwith

-

-

(99) An intended ab initio calculation, by the SCF method, of complex-formation between D-ribose and Na+ was,erroneously, performed on 1,4-anhydroribit01.'~The most stable arrangement was found to be the one in which the cation is simultaneously coordinated to 0-2 and 0-3. (100) H. Berthod and A. hllmann, Theor. Chim. Acta, 47 (1978) 59-66. (101) D. J. Goulding, J. Chromatogr., 103 (1975) 229-239.

COMPLEXES OF METAL CATIONS WITH CARBOHYDRATES

29

Ba2+,0.4,0.6, 1.O, and 1.1, respectively; glycerol, D-mannitol, L-arabinitol, xylitol, and ~-glucitol~* with Ca2+,0.0, 0.9, 1.2, 1.2, and I .5, respectively; D-glucito116with Ca2+,0.6 at 36" and 1.2 M-' at 4". Complex-formation generally decreases with increasing temperature. HaasIo2reported stability constants for many alditols and some sugars with Ca2+:these values are lower than those reported by others but are in the same relative order. He also reported stability constants for Cu2+that are, surprisingly,much higher than those for Ca2+.The measurements were conducted at pH 6.1; at this pH, solutions of copper salts are known103to contain a considerable proportion of the binuclear [Cu2(OH),I2+ion, which may be responsible for the strong complexing. Stability constants have been determined'Ol for reducing sugars in solution; the sugars being mixtures of anomeric forms, the constant is the weighted average of their stability constants. If the composition of the sugar is not known under the conditions of the determination, these values can merely serve as a general guide to the behavior of the sugar in solutions containing the cations. Such values are: ~-talosewith Ag+, 0.55; with Ca2+, 2.4; with La3+, 3.7; D-ribose with Ca2+, 1.6, and with Ba2+, 1.7; L-sorbose with Ba2+,0.3; and D-fructose with Ba2+,0.6 M-I. determined for By using a calorimetric method, Morel and D-ribose stability constantsthat are not in accord with other recorded values: 2.0 M-' for Ca2+,4.3 for S 3 + , 3.7 for Ba2+,2.8 for La3+,and 4.1 M-' for Gd3+.They concluded that the values of the stability constant are not related in a simple way either to the size, or to the charge, of the complexing cation. Reubenmfound some 1 :2 complex in the equilibriumbetween xylitol and lanthanide cations. His stability constants are: for Pr3+,Kl = 2.0, K2 = 0.12; for Nd3+,K I = 4.0, K2 = 0.5;and for Eu3+,Kl = 3.8, K2 = 0.77 M-I. These stability constants were determined from the 1.i.s. in the 'H-n.m.r. spectra at 39 '; calculation with these values showsthat, at the sugar concentration used for recording the spectra (0.45 M), the proportion of the 1 :2 complex is very small. For cis-inositol, constant K values were not indicating that a significant proportion of the 1 :2 complex was present. Its behavior with Ca2+can be accounted for by the values of Kl = 9 rt 3 and K2 = 50 rt 15 M-I; for S F , by Kl = 6 k 2 and K2 = 42 k 9 M-I (K, being defined as [CaIn$+]/ [CaIn2+][In]).For Na+, the K value, calculated on the assumption of only 1 : 1 complexes being p r e ~ e n tis , ~1.2, and for Mg2+,0.6 M-' (Ref. 23). (102) J. W. Haas, Jr., Marine Chem., 19 (1986) 299-304. (103) R. K. Steinhaus and C. L. Barsuhn, Inorg. Chem., 13 (1974) 2922-2929; H. Ohtaki and T. Kawai, Bull. Chem. SOC.Jpn., 45 (1972) 1735-1738. (103a) A. M. Alvarez, N. Morel-Desrosiers, and J.-P. Morel, Can. J. Chem., 65 (1987) 26562660.

STEPHEN J. ANGYAL

30

For complex-formation with lactose, K values were reported'"'' that are not consonant with other results: 0.53 with NaC1, 0.38 with LiC1, and 0.205 M-' with Ca(N03)2.These values were obtained from the change in optical rotation of lactose on addition of these salts; the change may well be caused by factors other than complex-formation. Table IV also includes some values determined in methanol as the solvent; these are very much higher (and, hence, also more accurate) than those in water, because the polyol competes with methanol, rather than with water, for outer-sphere positions on the cation. These figures explain why carbohydrates are soluble in methanol or ethanol containing high concentrations of calcium chloride, or even potassium acetate,8and in such systems as lithium chloride in 2-methoxyethanol.lo5 Sugar derivatives that are soluble in nonhydroxylic solvents form complexes with cations in those solvents even more readily; for example, methyl 2,3-O-isopropylidene-4-O-methyl-/3-~rhamnopyranoside (24) (but not its a anomer) will form a complexlMwith sodium iodide in acetone, the Na+ ion coordinatingto 0-1,O-2, and 0-3.In aqueous solution, the concentration of this complex would be neghgible. Me0

OMe

2L

Formation of complexes between sugars and sodium perchlorate in pyridine has also been reported,lo7but the stability constants do not seem to be related to their structure: L-sorbose, 7.2; methyl &D-ribopyranoside, 5.5; D-ribose, 6.8;D-glucitol, 1 1.2; and lactose 8.0 M-'. Formation ofthese complexes appears to be non-specific. The thermodynamic parameters of complex-formation between D-ribose and Ca2+have been determined by Morel and c o w ~ r k e r and s ~ by ~ ~Len~~~ kinski and Reuben.& The stability constant was found to be the result of a favorable enthalpy term (- -25 kJ mol- ') and a very unfavorable entropy term (- - 80 J K- * mol- '). Thermodynamic data for Na+, Ca2+,and La3+ complexes in methanol solution have been determined by V e d a and coworker~.~~

-

-

(104) M. L. Swartz, R. A. Berhard, and T. A. Nickerson, J. Food. Sci..43 (1978) 93-97. (105) S. J. Clarke, Curbohydr. Rex, 115 (1983) 231-233. (106) A. H. Haines, K. C. Symes, and A. G . Wells, Curbohydr.Res., 41 (1975) 85-92. (107) J. Grandjean and P. Laszlo, Helv. Chim. Actu, 60 (1977) 259-261. (108) J.-P. Morel, C. Lhermet, and N. Morel-Desrosiers,Can. J. Chem., 64 (1986) 996- 1001.

COMPLEXES OF METAL CATIONS WITH CARBOHYDRATES

31

5. Alduronic Acids

Although this article deals with complexesof neutral carbohydrates, brief mention should be made of the aldonate and alduronate anions (the undissociated acids do not form complexes), which show complexing behavior similar to those ofthe neutral polyols. Interest in these acids, and particularly in the poly(hexuronates), is due to the industrially important gel-formation of alginic acid and pectin with calcium ions, caused by complex-formation. There is an extensive literature on this subject,'0gwhich is not fully discussed here. Alduronate ions, like all a-hydroxy acid anions, form much stronger complexes with cations than do neutral sugars; for example, Kis 70 M - l for complex-formationbetween D-galacturonate ion and Ca2+ion.110Complexformation is also stronger than that with carboxylate ions lacking an oxygen atom in the a position.'l' Formation of tridentate complexes, therefore, does not necessarily contribute to the stability of such complexes; and it was found that tridentate complexing occurs in some, but not in all, cases where both a- and B-hydroxyl groups are present. There is much information available on the nature of complex-formation of hydroxy acids in the crystalline state, but much less on what happens in solution. A tridentate complex, at 0-1,O-2, and 0-3, is found in the crystal structure of calcium lactobionate112and o-arabin~nate,"~ but not in those of potassium D-gluconate114or calcium DL-glycerate.1'5It has been postulated that, in solution, DL-glycerate,D-gluconate,D-glucarate, and lactobionateall form tridentate complexes with lanthanide ions,'16 but the evidence is not conclusive.'17Unusual tridentate complexation was found at 0-1,O-3, and 0-4 in the crystals of calcium D-glucarate,'18and at 0-1,O-2,and 0 - 4 in one form of potassium D-gluconate.1'4 (109) D. A. Rees, E. R. Moms, D. Thom, and J. K. Madden, in G. 0. Aspinall (Ed.),The Polysaccharides, Vol. 1, Academic Press, New York, 1982, pp. I95 -290. ( 1 10) R. 0. Gould and A. F. Rankin, Chem. Commun., (1970) 489-490. (1 1 I ) C. A. M. Vijberg, J. A. Peters, A. P. G. Kieboom, and H. van Bekkum, Tetrahedron, 42 (1986) 167-174. (1 12) W. J. Cook and C. E. Bug, Acta Crystallogr., Sect. B, 29 (1973) 215-222. (1 13) S. Furberg and S. Helland, Acta Chem. Scand., 16 (1962) 2373-2383. ( 1 14) N. C. Panagiotopoulos, G. A. Jeffrey, S. J. La Placa, and W. C. Hamilton, Acfa Crysfallogr.. Sect. B, 30 (1974) 1421 - 1430. ( I 15) E. J. Meehan, Jr., H. Einspahr, and C. E. Bugg, Acfa Crystallogr.. Sect. B, 35 (1979) 828-832. ( 1 16) T. Taga, Y. Kuroda, and M. Ohashi, Bull. Chem. SOC.Jpn., 51 (1978) 2278-2282; T. Taga, Y. Kuroda, and K. Osaki, ibid., 50 (1977) 3079-3083. ( I 17) S. J. Angyal, D. Greeves, and L. Littlemore, Carbohydr. Res., 174 ( 1988) I2 1 - 13 1. (1 18) T. Taga and K. Osaki, Bull. Chem. Soc. Jpn., 49 ( 1976) 15 17 - 1520.

STEPHEN J. ANGYAL

32

Alginate is a polymer consisting of a-L-gulopyranosyluronateand p-Dmannopyranosyluronate subunits; pectate consists of a-D-galactopyranosyluronate units, all (1 +4)-linked. The a-L-guluronate moiety has an a,e,a arrangement of oxygen atoms, rare in natural products. To explain the dramatic changes in the properties of alginic acid on addition of Caz+ions, Smidsrsd and coworkers119and, independently, Angyal,'*O suggested the formation of a complex (25) in which 0-4, 0-5, and 0-6 of a guluronate residue and 0-2 and 0-3 of a contiguous guluronate residue in the polymer chain are coordinated to Ca2+(0-4 of the first residue is 0-1 of the second). The existence of such quinquedentate structure has never been proved, but it is in accord with modem ideas about the mechanism of gelling of alginates. However, in pectic acid, there is no a,e,a arrangement, and yet it also gels with Ca2+ions. '

/

0

25

An X-ray structure determinations8showed that, in the sodium salt of a-L-guluronic acid (23), one cation is close to 0-1, 0-2, and 0-3, while another one is close to 0-4,O-5, and 0-6 (see Section 111,3). Complexing of L-guluronic acid in aqueous solution has apparently not yet been investigated that of D-glucuronic and D-galacturonic acid has, however, been extensively studied.l l ' m The a-pyranose forms complex at the ring-oxygen atom and at one of the carboxylate oxygen atoms; in the p anomers, the ring-oxygen atom is not involved in complex-formation. It appears that, in the a,but not in the / I anomer, , electron transfer from 0 - 1 assists in the coordination of 0-5 to the cation. Whether 0-4 in the a-D-galacturonate ( 1 19) 0. Smidsrpld, A. Haug, and S. G. Whittington, Acfu Chem. Scund.. 26 (1972) 25632566. (120) S. J. Angyal, PureAppl. Chem., 35 (1973) 131-146. (121) K. Izumi, Agric. Biol. Chem.. 44(1980) 1623-1631.

COMPLEXES OF METAL CATIONS WITH CARBOHYDRATES

33

(26) is also involved in complex-formation is still controversial; available e v i d e n ~ e ”appears ~ . ~ ~ ~to indicate that it is not. In the crystal structure of the calcium-sodium and the strontium-sodium salts of D-galacturonic acid, there is no coordination between 0-4 and the cations.”’ D-Galacturonate complexes more strongly with cations than does D-glucuronate,llobut that can be explained without involving complex-formation at 0-4 in the former.117

26

All this discussion is not strictly relevant to the behavior of alginic and pectic acids. The axial-axial a-(1 4 4 ) linkage between subunits causes the polymer chain to fold, thereby bringing hydroxyl groups of one unit close to the carboxyl group of another one (asin 25). It is possible, therefore, that 0-4 is not involved in the complex-formation of Caz+with D-galacturonic acid, and yet it is involved in that with the polymer chain, together with one or two hydroxyl groups of the adjacent pyranose unit. The strength of complexformation is affected by the length of the polyguluronate chain123; complexformation is, therefore, enhanced by the steric arrangement of the polymer chain. In the widely accepted structure, denoted as the “egg-box” model,124 of calcium alginate and pectate, two buckled polymer chains are close to each other, forming cavities in which the cation is situated, holding the two chains together by chelation. Calcium alginate is often depictedIw with the cation coordinated to the five oxygen atoms (shown in 25) of each chain; this gives the cation a ten-fold coordination, which is rather rare. It is possible that not all of the five potential coordinatingpoints are actually utilized.124 In pectate, only four oxygen atoms of each D-galacturonate residue are available for complexing. Poly(gu1uronate)forms stronger complexes with SrZ+ and with BaZ+than the corresponding complexes of poly(gala~turonate).~’~ ( 122) S. E. B. Gould, R. 0.Gould, D. A. Rees, and W. E. Scott, J. Chem. SOC.,Perkin Trans.2,

(1975) 237 -240; S. Thanomkul, J. A. Hjortas, and H. Serum, Acta Crystullogr.,Sect. B, 32 (1976) 920-924. (123) R. Kohn, PureAppl. Chem., 42 (1975) 371-397. (124) G. T. Grant, E. R. Moms, D. A. Rees, P. J. C. Smith, and D. Thom, FEBS Lett., 32 (1973) 195-198. (125) A. Haug and 0. Smidsrd, Acta Chem. Scand., 24 (1970) 843-854.

34

STEPHEN J. ANGYAL

The reason for this differencein binding strength may lie in the ability of S? and Ba2+to coordinate with a greater number of oxygen atoms in poly(guluronate) than in poly(galacturonate). Important information could be obtained by study of the complexing behavior of 4-O-a!-~-gulopyranosyluronate-a-L-gulopyranuronate, which is the dimeric subunit of alginate; this study has, apparently, not yet been carried out. The segment of alginate consisting of p-D-mannopyranosyluronateunits plays little part in gel formation. On the one hand, the polymer structure being linear, there is much less opportunity for inter-unit complexing; and, on the other, it beingap-glycoside,the ring-oxygen atom would not take part in complexing. that, in oriented fibers of An X-ray crystal structure analysis has the potassium salt of gellan, a microbial polysaccharide, the potassium ion is coordinated to both carboxylate oxygen atoms of p-D-glucuronate and to 0 - 2 of an adjacent p-D-glucose in one polymer strand, to 0 - 2 of a P-D-~~Ucuronate and 0-6 of an adjacent p-~-glucosein another strand, and to one molecule of water. N-Acetyl-fi-neuraminic acid (27)is another important hydroxy acid in which several hydroxyl groups are available for complex-formation. HO

27

It forms a strong complexlZ6with Ca2+(K = 121 & 5 M-'). From analysis of electric-field shifts, it was postulated that there is a quadridentate complex involving the carboxylate group, the ring-oxygen atom, and 0 - 7 and 0-8 of the side chain; the last three form a contiguous trio1.126From similar data, Czarniecki and Thornton12' drew the conclusion that the carboxylategroup is not involved in complex-formation;this appearsunreasonable, as it would not account for the high stability-constant. Removal of the side chain, by periodate oxidation, considerably lowers the stability of the complex; re( 1 25a) R. Chandrasekaran,L. C. Puigjaner, K. L. Joyce, and S. Amott, Curbohydr.Res., 18 1

(1988) 23-40. (126) L. W. Jacques, E. B. Brown, J. M. Barrett, W. S . Brey, Jr., and W. Weltner, Jr., J. Biol. Chem., 252 (1977) 4533-4538. (127) M. F. Czarniecki and E. R. Thomton, Biochem. Biophys. Res. Commun., 74 (1977) 553-558.

COMPLEXES OF METAL CATIONS WITH CARBOHYDRATES

35

placement of the N-acetyl by an N-glycolyl group increases it.128The suggested quinquedentate complex involvingthe glycolyl group, the side chain, and the carboxylategroup with the same calcium ion appears unlikely. So is the structure suggested by Martin129for the N-acetyl-P-neuraminic acid complex, which postulates quinquedentate coordination of 0-6, 0-7, 0-8, and 0-9, besides the carboxylate group. Actually, there are probably two complexing sites, one at each end of the molecule. In Nature, N-acetylneuramink acid occurs only as the a anomer, which forms a much weaker comp l e ~ ,the ' ~ ring-oxygen ~ atom not taking part in complex-formation. It was shown that the methyl glycoside of the a anomer complexes with Gd3+at 0-8,O-9, and the carboxylategroup, whereas (the smaller) Mn2+appears to complex at two sites, the carboxylate group and the side chain.Im Similar results were ~btained'~' with 3-deoxy-~-manno-2-octulosonicacid (KDO). In heparin, a polysaccharide that contains sulfate groups in addition to the carboxylate ion, associationwith Ca2+is a delocalized process, as opposed to one involving specific binding-sites.132 This short discussion illustrates that there is still much to be learnt about the complexing behavior of acidic polysaccharides. IV. APPLICATIONS 1. Electrophoresis and Thin-layer Chromatography Carbohydrates can be identified, and their mixtures separated, by paper electrophoresis.l I Tables of Mi values in buffered calcium acetate as s u p porting electrolyte have been published for some 150 carbohydrates(cyclitols, sugars, and methyl glycosides,l 3 alditols,48 and anhydrohe~osesl~). Some typical values are shown in Table V. Better separationsare achieved by this method thanI5by t.l.c., because the movement of the complexescan be monitored over a much longer pathway; the reproducibility of the results is also better than with t.1.c. Electrophoresis is particularly suitable for the study of weakly complexing compounds; their movement is slight, and hence they give sharp spots; by contrast, in t.1.c. they move rapidly and produce more diffuse spots. For the separation of alditols, lanthanum perchlorate may be more suitable as the supportingelectr~lytel~; the separations are not only better but also more rapid.'33 (128) L. W. Jacques, B. F. Riesco and W. Weltner, Jr., Curbohydr. Res., 83 (1980) 21 -32. (129) R. B. Martin, Metal Ions Biol. Syst., 17 (1984) 1-49. (130) M. E. Daman and K. Dill, Curbohydr. Rex, 102 (1982)47-57. (131) K. Luthman, A. Claesson, L. Kenne, and I. Cdregh, Carbohydr. Rex. 170 (1987) 167-179. (132) P. Dais, Q.-J. Pen& and A. S. Perlin, Curbohydr. Res., 168 (1987) 163- 179. (133) M. E. Tate, personal communication.

36

STEPHEN J. ANGYAL TABLEV Electrophoreticand Thin-layer Mobilities of Some P o l y o l ~ ~ ~

cis-Inositol epi-Inositol Allitol D-Mannitol Galactitol D-Glucitol D-Iditol D-LyXOSe D-Ribose D-GIuco~~ D-Talose Methyl a-D-ribopyranoside Methyl &D-ribopyranoside Methyl a-D-lyxofuranoside Methyl P-D-lyxofuranoside

1

.oo

0.44 0.09 0.14 0.17 0.20 0.24 0.08 0.25 0.02 0.27 0.33 0.24 0.07 0.3 1

0.07 0.42 0.85 0.78 0.64 0.66 0.57 0.82 0.57 0.88 0.59 0.39 0.42 0.65 0.50

1

.oo

0.43 0.1 1 0.17 0.20 0.37 0.42 0.24 0.03 0.2 1

0.09 0.39 0.78 0.75 0.55 0.52 0.40 0.85 0.5 1 0.92 0.61 0.42 0.34 0.63 0.63

a Cationic movement relative to cis-inositol; paper electrophoresis in a 0.2 M solution ofcalcium acetate in 0.2 Maqueousacetic acid." Cationic movement relative to cis-inositol; paper electrophoresis in a 0.1 M solution of lanthanum perchlorate.

Although paper electrophoresishas been r e ~ o m m e n d e dfor ' ~ ~pre-testing separations before performing them on a preparative column, t.1.c. is more suitable for this purpose. Electrophoreticmobility is not strictly proportional to the extent of complex-formation; it depends also on the bulk and shape of the molecule, which influence ionic mobility. In particular, bulky substituents retard the passage of the complex ion through the electrolyte. These effects have been discussed in detail.I3T.1.c. is also affected by the bulk of substituentsbut to a much lesser extent; hence, it is more likely to predict the separations obtainable on a column.15 A somewhat similar method has been proposed by Bilisics and P e t r ~ S , ' ~ ~ namely, chromatographicseparation on 0-(carboxymethy1)cellulosepaper in the La3+,Ca2+,and Ba2+forms. This is a combination of partition and ligand-exchange chromatography, in which the usual paper chromatography is modified by the presence of cations, and it results in efficient separations of alditols and most aldoses. 2. Preparative Separations on Columns of Cation-exchange Resins Sugars and polyols had been separated on columns of cation-exchange resins in their calcium or barium form before it was realized that this separation is based on complex-formation. The method was discovered fortui(134) S. J. Angyal, G. S. Bethell, and R. J. Beveridge, Carbohydr. Rex, 73 (1979) 9- 18. (135) L. Bilisics and L. PetruS, Carbohydr. Rex, 146 (1986) 141- 146.

COMPLEXES OF METAL CATIONS WITH CARBOHYDRATES

37

tously by Felicetta and coworkers136in 1959. When separating the sugars of spent sulfite liquors from the lignin sulfonates by means of ion-exchange resins, they found that L-arabinose was separated from D-xylose. Jones and Wall'37then studied the retention of several sugars and alditols on columns of Dowex 50W-X8 (Baz+)resin, and separated several mixtures into their components. Despite these successful separations, the method was rarely used until it was realized that it is based on complex-formation between the cation on the resin and the polyol, thereby making the separations predictable. Fedoroiiko, Bilik, and Angyal, and their coworkers, then used the method extensively. In many cases, it is an excellent method of separation, because water alone is used as the eluant, the recovery is almost quantitative, the capacity of the column is, in suitable cases, very large, and the columns can be used many times without the need to regenerate them. The method has also been used successfully for the preparative separation of sugars under h.p.1.c. condition~.'~~ A detailed discussion of the method was given in Ref. 134. Caz+is the cation most frequently used; the group in Bratislava mostly employed Ba2+ columns. In some instances, Pb2+columns give better separation; SI2+ and Zn2+would probably also be suitable. PetruS and coworkers67found that La3+gives the best separation of alditols (see Section 1142). A column containing Cu2+ions (actually, probably, [Cu, (OH),]'+) was recommended by Bourne and coworkers,12but this column was eluted with cupric acetate solutions,a procedure which makes isolation of the sugarscumbersome.The separations are improved by lowering the temperature and by addition of methanol to the water serving as the eluant. These columns are, of course, inefficient in the separation of carbohydratesthat form very weak complexes(K < 0.1 A4-l). However, compounds having three suitably placed hydroxyl groups can be separated very efficiently;tens of grams can be handled by quite a small column. In the separation of hexitols, all of which can be separated from each other, optimum separation is achieved when the weight of the sample applied to the column is 5 - 15% of the weight of the column packing.67D-G~UCOX and D-fructose have been separated on a kilogram scale by the use of a Caz+column,139and this method is used on an industrial scale.140J41 The method is particularly ( I 36) V. F. Felicetta, M. Lung, and J. L. McCarthy, Tuppi, 42 (1959) 496-502. (137) J. K. N. Jones and R. A. Wall, Can. J. Chem., 38 (1960) 2290-2294. (138) K. B. Hicks, Adv. Curbohydr. Chem. Biochem., 46 (1988) 17-72. ( 139) P. E. Barker and S. Thawait, Chem. Ind. (London), ( 1983) 8 17 - 82 1. (140) T. Sakiyama, K. Nakamura, and T. Yano, Agric. Biol. Chem., 49 (1985) 2619-2625. (141) Yu. E. Kuptsevich, I. D. Stah'naya, 0. N. Sosedov, L. A. Nakhapetyan, A. Ya. Pronin, and G . 0.Larionov, Prikl. Biokhim. Mikrobiol., 21 (1985) 129- 134; Chem. Absfr., 102 (1985) 132,374.

38

STEPHEN J. ANGYAL

suitable for the isolation of a desired compound when it is not the major product of a reaction; for example, ~-talosefrom ~-galactose'~* or D-psicose from ~ - f r u c t o s e ,or '~~ for the synthesis of methyl ~D-mannofuranoSide.'44 The method is eminently suitable for the separation of cyclitols and alditols from each other. Reports of such separations achieved on ion-exchange columns are too numerous to be listed here, but some typical examples are as follows. (i) Sugars: D-galactose and ~ - t a l o ~D-altrose, e ~ ~ ~ ; ~-allose,and ~ - r i b o s e ' ~ ~ ; D - ~ ~ U C Oand S ~ ~-galactose'~~; D-xylose, D-idose, and D-gulose'46;D-sorbose and ~-tagatose'~~; ~-threo-2-pentulose,~-threo-3-pentulose,and D-erythroL-glycero-L-galacto-heptose and ~-glycero-~-talo-heptose'~~; 2-pent~lose'~~; L-fucose, and 6-deoxy-~-talose~~O; 3deoxy-3-fluoro-~-glucoseand 3-deoxy3-fluoro-~-mannose(also, the 4-anal0gs)'~';sucrose and "allo-sucrose" 152; and sucrose, inulobiose, and higher homo log^.'^^ (ii) Glycosides: the four methyl D-psicosidesand the four methyl tagat~sides'~~; and the four methyl D-nbosides.155 (iii) Alditols: D-glycero-D-gluco-heptitol and D-glycero-D1,4-anhymanno-heptitol.'56 (iv) Others: 1,6-anhydro-a-~-talofuranose, and sugars dro-P-D-talopyranose, and 1,6-anhydro-P-~-talopyranose'~~; from their 1,6-anhydride~.'~~ The sugars emerge from the column in the order listed here. Three compounds were found to emerge from the chromatographiccolumn much later than expected from their configurationsand from comparison with related compounds. It has been postulated that these compounds V. Bilik, W. Woelter, and E. Bayer, Justus Liebigs Ann. Chem., (1974) 1162- 1166. R. J. Beveridge, M.Davis, J. L. Moms, and N. Hoogenraad, Curbohydr.Res., 101 (1982) 348-349. S. J. Angyal, M.E. Evans, and R. J. Beveridge, Methods Curbohydr. Chem., 8 (1980) 233-235. V. Bilik, Chem. Zvesti, 29 (1975) 114- 118. J. R. Snyder and A. S . Serianni, J. Org. Chem.. 51 (1986) 2694-2712. S. J. Angyal and G. S. Bethell, Aust. J. Chem., 29 (1976) 1249- 1265. L. StankoviE, K. Linek, and M.Fedorofiko, Carbohydr. Res., 10 (1969) 579-583. V. Bilik, D. Anderle, and J. Alfoldi, Chem. Zvesti, 28 (1974) 668-672. J. Defaye, A. Gadelle, and S. J. Angyal, Carbohydr. Res., 126 (1984) 165- 169. J. R. Rasmussen, S. R. Tafuri, and S . T. Smale, Carbohydr. Res., 1 16 (1983) 2 1-29. L. Hough and E. OBrien, Carbohydr. Res., 84 (1980) 95- 102. M. K. Das, D. G. Streefkerk, and C. P. J. Glaudemans, Mol. Immunol., 16 (1979) 97- 100. S. J. Angyal, C. L. Bodkin, J. A. Mills, and P. M. Pojer, Aust. J. Chem., 30 (1977) 1259-1268. S. J. Angyal, C. L. Bodkin, and F. W. Parrish, Aust. J. Chem., 28 ( 1975) 1541 - 1549. S. J. Angyal and R. Le Fur, Curbohydr. Res., 126 (1984) 15-26. S. J. Angyal and R. J. Beveridge, Carbohydr. Res., 65 (1978) 229-234. S. J. Angyal and R. J. Beveridge, Aust. J. Chem., 31 (1978) 115 I - 1 155.

COMPLEXES OF METAL CATIONS WITH CARBOHYDRATES

39

complex simultaneouslywith two cations on the surface of the ion-exchange resin; this phenomenon has been called134“double complexing.” Thus, Dglycero-L-tulo-heptose (28) and ~-erythro-~-tulo-octose’~~ have been assumed to complex at 0-2,O-3, and 0-4, and also at 0-5,O-6, and 0-7; and methyl a-L-gulofuranoside(29) at 0-1,O-2, and 0-3, and also at 0-4,O-5, and 0-6. The side chain in the former compounds has the threo configuration. In paper electrophoresis (and, surprisingly, in t.l.c.), the compounds behave normally. It was postulated that, when the sugarsare absorbed on the resin, a neighboring cation on the surface will form an additional complex, even with a weak site. In each case, the two complexing sites are on the same side of the molecule; this is probably a condition for the existenceof “double complexing.”

28

29

Special consideration needs to be given to allose, gulose, and their homologs.’” For these sugars, only the a anomers form complexes; mutarotation is slow and, therefore, there is enrichment in the /?form at the front, and in the a form at the rear, of the sugar descendingthe column. Hence, the sugars emerge as wide bands. Addition oftrimethylamine to the eluant, or using the column at an elevated temperature, acceleratesmutarotation, and the sugars then emerge in narrow bands. On the other hand, by lowering the temperabut they cannot be ture to O”, the a and p anomers can be ~eparated,~’ recovered from their dilute solutions without some mutarotation. The ion-exchange resin is merely a support for the cations, and does not affect the separation. Ion-exchange resins are not the only possible supports for cations; similar separations have been achieved on sodium-calcium zeoS ~ separated from each other on an lites.’a.161D-Fructose and D - ~ ~ U C O are industrial scale by the use of barium zeolite columns.’62These separations (159) (160) (16I) (162)

V. Bilik, L. PetruS, and J. Alfdldi, Chem. Zvesti, 30 (1976) 698-702. T. M. Wortel and H. van Bekkum, Red. Truv. Chim. Pays-Bus. 97 (1978) 156- 158. J. D. Sherman and C. C. Chao, Proc. Znt. Zeolite Conf, 7th, Tokyo, (1986) 1025 - 1032. 0. Hiroyuki and 0. Masaji, Ger. Offen. 2,626,194 (1976); Chem. Absfr., 86 (1977) 74,846; H. Odawara, Y. Noguchi, and M. Ono, Jpn. Kokai 76, I 10,048 (1976); Chem. Abstr., 86 (1977) 73,066; R. W. Neuzil and J. W. Priegnitz, U.S. Pat. 4,024,331 (1977); Chem. Abstr., 87 (1977) 184,896.

40

STEPHEN J. ANGYAL

are due to complex-formation, as shown by the approximateproportionality between electrophoretic mobilities and the retention volumes on the column.161There are, however, between such predictions and the actual separations, some discrepancieswhich indicate that the retention on the column is also affected by geometric constraints imposed by pore geometries and by the position of the cation in the zeolite.

3. High-performance Liquid Chromatography Among the many columns used in the h.p.1.c. of carbohydrates, those containing a cation-exchange resin are probably the most useful. H.p.1.c. columns filled with the calcium form of an ion-exchangeresin are marketed specially for carbohydrate analyses. These columns have the advantages of giving convenient, accurate, rapid, reproducible, and sensitive analyses, of using only water as the mobile phase, and of being durable and easily regenerated. Their use was discussed in detail in this Series,13*and will therefore not be dealt with in this Chapter. These columns are also used for the separation of carbohydrates that do not form complexes with cations; in this case, the column functions by adsorption and partition. Besides Ca2+,columns in the Pb2+and the Ag+ form are also used, but these cations may react with reducing sugars. Thus, it has been reported that, at 60",ketoses are isomerized on a Pb2+c01umn.l~~ Columns in the La3+form have not been extensively used, although they are the choice for separating alditols from each other and from other carbohydrates.164 Understanding of the complexingability of various sites in carbohydrates (see Section II1,l) leads to prediction of the separations possible on h.p.1.c. columns. All eight inositol diastereoisomers have been separated on a Ca2+ column, in accordance with expectation^.^^^ 4. Nuclear Magnetic Resonance Spectroscopy

Intractable n.m.r. spectra can sometimes be resolved by the addition of paramagnetic ions if the compound forms a complex with the ions.166Eu3+ and PS+ are particularly useful; they shift the signals in opposite directions. The method can be used in aqueous solution, where shift reagents are not applicable. With the wide availability of high-field n.m.r. instruments, this method has become less useful than it was some ten years ago; but, if two (163) J. D. Baker, M. Y. Tucker, S. M. Lastick, and M. E. Himmel, J. Chromatogr.,437 (1988) 387-397. (164) H. F. Walton, J. Chromatogr., 332 (1985) 203-209. (165) K. Sasaki, K. B. Hicks, and G. Nagahashi, Carbohydr. Rex. 183 (1988) 1-9. (166) S. J. Angyal, Carbohydr. Rex, 26 (1973) 271-273.

COMPLEXES OF METAL CATIONS WITH CARBOHYDRATES

41

hydrogen atoms have the same chemical shift, even high-field instruments cannot resolve their signals. In the spectrum of xylitol, all signals are very close to each other, but gradual addition of europium chloride separates them c ~ m p l e t e l y .(A ~ *first-order ~~ 'H-n.m.r. spectrum of xylitol has been obtained167at 620 MHz.) The spectra of D-glucuronic acid,'17 methyl P-Dtalofuranoside, and D-glucaro-1,4-la~tone'~~ have been resolved by the addition of praseodymium chloride. Complex-formation with diamagnetic cations has also been used to assist in the analysis of otherwise intractable spectra. In that of cyclohexane1,2,3,4,5/0-~entol(4),the signals of the two methylene protons overlap, but they are separated by the addition of calcium i ~ n s . ~ Signals ' J ~ ~ not visible in the spectrum of ~ - a l l o appear ~e after the addition of calcium chloride.45 Complexing has been used in a different way to unravel the I3C-n.m.r. spectrum of D-tagatose. The a-pyranose is preponderant, and its signals are readily recognized. O n addition of calcium chloride, the proportion of the P-pyranose form increases until its signals are higher than all the others and are therefore readily identified.'47 The same method was used for the ~ - p s i c o s eand , ~ ~several ~ 1,2-di13C-n.m.r.spectra of 1-deoxy-~-psicose,'~~ deoxy-3-heptuloses.17b 5. Synthetic Applications

In a reversible reaction that leads to an equilibrium of several compounds, the outcome of the reaction can be changed by the addition of cations to the mixture if the reaction products have differing complexing ability.171A prime example is the Fischer glycosidation of sugars, with methanol and a strong acid, which leads to a mixture of methyl furanosidesand pyranosides. Addition of calcium chloride to the reaction mixture increases the proportion of those pyranosidesthat have an a,e,asequence, and of the furanosides that have three consecutive cis oxygen atoms. These are usually minor components of the equilibrium mixture, but they can be made the major products by the use of a large proportion of calcium chloride; fortunately, calcium chloride is readily soluble in methanol. The products are then separated on a calcium column. (167) F. Franks,R. L. Kay, and J. Dadok, J. Chem. SOC.,Furuduy Trans. 1 , 8 4 (1988) 25952602. (168) S. J. Angyal, Curbohydr. Res., 77 (1979) 37-50. (169) D. Horton and Z. Walaszek, Curbohydr. Res., 105 (1982) 95- 109. (170) S. J. Angyal, G . S. Bethell, J. E. Cowley, and V. A. Pickles, Aust. J. Chem., 29 (1976) 1239- 1247. (170a) 1. I. Cubero and M. T. P. Lbpez-Espinosa, Curbohydr. Rex, 173 (1988) 41 -52. (171) M. E. Evansand S. J. Angyal, Curbohydr. Rex, 25 (1972)43-48.

42

STEPHEN J. ANGYAL

Several methyl furanosides have been prepared154*155 in this way in yields considerably higher than those previously recorded in the literature; for example, methyl a-D-ribofuranoside, 69% (lit., 4%); methyl P-D-lyxofuranoside, 3 1% (lit. 3-5%); and methyl PD-talofuranoside, 54% (previously unknown). This increase in yield, spectacular in some cases, is due not only to the shift in the equilibrium but also to the separation, without substantial losses, of the desired component by the calcium column. In the synthesisof methyl P-D-mannofuranoside,complexingis used for a third time when the compound is isolated as its readily crystallizing complex with calcium chloride. Reaction conditions have been established under which any of the four methyl D-allosides can be prepared in good yield, in one step, from ~-allose by varying the temperature and the concentration of acid, in the presence or absence of strontium ch10ride.l~~ When none of the methyl glycosides form a complex, the dimethyl acetal of the acyclic form can sometimes be obtained in the presence of calcium ch10ride.I~~ Normally, this acetal is formed in only a very small proportion. Other reactions may also be affected by the presence of complexing cations. Methyl furanosides that complex readily are hydrolyzed by acids at a lower rate in the presence of calcium chloride than those which do not.95The first step ofthe hydrolysis, protonation by the acid, is impeded by the positive charge of the glycoside-cation complex. This retardation may be useful for the selective hydrolysis of polysaccharides. V. BIOLOGICAL IMPLICATIONS Many publications on cation - sugar complexes refer, in their introduction, to the great biological importance ofthis subject. In fact, however, there is very little evidence for the involvement of complexes of neutral carbohydrates in biological systems, in contrast to the complexes of acidic carbohydrates. Complex-formation between cations and neutral carbohydratesis weak; concentrationshigh enough to produce significant proportions of complexes are rare in biological systems. The sugars common in Nature, with the exception of D-ribose,do not complex readily; and, ofthe cationswhich have good complexing ability, only Ca2+occurs in biological systems in substantial amounts. Trace elements will not form complexes to any significant extent. Also, carbohydrates seldom occur in high concentrationsin biological systems, with the exception of D - ~ ~ U C Owhich S~ is a very poor complex(172) F. W. Pamsh, S. J. Angyal, M. E. Evans, and J. A. Mills, Curbohydr. Rex, 45 (1975)

73-83.

COMPLEXES OF METAL CATIONS WITH CARBOHYDRATES

43

former. Hence, cation complexes of monosaccharides are unlikely to have any biological significance. Complexing is more likely to occur on the surface of polysaccharides, especially on cell walls. It is widely believed that cation- carbohydratecomplexes play a role in the transport of cations through membranes, although details of this process are not known. Neutral polysaccharides may form complexes with cations to a significant extent: change in its optical rotation on addition of cations indicates that dextran does It has been pointed out that, even if a sugar does not form a complex, a polymeric chain of the same sugar may do so, due to cross-linking; this has been well establishedfor acidic p o l y s a ~ c h a r i d e s , ~but ~ ~may * ' ~ ~also be true for neutral ones. In the crystalline structure of oriented fibers of an amylose-KBr complex, the cations are situated at equal distances from two D-glucosyl units in two different chains, being coordinated to 0-3, and possibly to 0-2 and 0-4, of Apparently, there have been no systematic studies correlating chain-length and configuration with complexing ability in neutral polysaccharides. The nature of complex-formation between cations and monosaccharides is now well understood; this knowledge should form a firm basis for the study of cation - polysaccharide interactions.

( 1 73) D. A. Rees, in H. L. Kornblum and D. C. Phillips (Eds.), MTP Int. Rev. Sci. Biochem. Ser. One, Butterworth, London, 1975, pp. 1-42.

(174) J. J. Jackobs, R. R. Bumb, and B. Zaslow, Biopolymers, 6 (1968) 1659- 1670.

This Page Intentionally Left Blank

.

ADVANCES IN CARBOHYDRATE CHEMISTRY AND BIOCHEMISTRY VOL . 41

ANOMERIC AND EXO-ANOMERIC EFFECTS IN CARBOHYDRATE CHEMISTRY BY IGOR

TVAROSKA* AND TOMASBLEHAt

* Institute of Chemistry and TPoIymer Institute. Centre of Chemical Research. Slovak Academy of Sciences. *842 38 and f842 36 Bratislava. Czechoslovakia

I . Introduction ........................................................ I1 . Definition of the Anomeric and Exo-anomenc Effects ...................... 1 . AnomericEquilibria ............................................... 2 . The Energy of the Anomeric Effect ................................... 111. Experimental Data on These Effects ..................................... I . Relative Abundance of Isomers ...................................... 2 . Valence Geometry Parameters ....................................... IV. Molecular Orbital Calculations of the Anomeric Effect...................... I . Conformational Energies ........................................... 2 . The Solvent Effect ................................................. 3. The Anomeric Effect Energy from MO Calculations ..................... 4 . Coupling of Bond Lengths and Bond Angles to Torsional Angles........... 5 . Electron Distribution and Lone Pairs ................................. V. The Anomeric and Exo-anomeric Effects in Potential-Function Calculations ... VI . Nature of the Anomeric Effect.......................................... 1 . Electrostatic Interactions ............................................ 2 . Delocalization Interactions .......................................... 3. Scaling between Electrostatic and Delocalization Interactions.............. VII. Role of the Anomeric Effect in the Reactivity of Carbohydrates .............. 1 . Energy of Reaction Paths ........................................... 2 . Lone-Pair Orbital Interactions in Reactivity ............................

45 47 48 53 59 59 71 75 77 88 93 98 101 103 106 107 109 114 116 116 119

I . INTRODUCTION

The concepts of conformational analysis are fundamental to a proper understanding of the relationship between the structure and properties of carbohydrates. The general application of conformationalanalysis has been stimulated by the relative universality of its fundamental postulates. especially on the qualitative level. The leading notion is the quantification of 45

Copyright Q 1989 by Academic Ress, Inc. All rights of reproduction in any form reserved.

46

IGOR

TVAROSKAAND TOMAS BLEHA

steric interactions (bulkiness) of substituents. There is a vast number of examples where the stability of the conformations, their reactivity, or the stereochemistry of products can be explained solely by steric factors. However, as soon as application of the methods that have been used so successfullyin conformationalanalysisof acyclicand cyclic hydrocarbons to carbohydrates is attempted, it becomes apparent that some additional and quite new factors are at work. They are usually treated as separate conformational effects, and termed by special The most important ofthese factors is the anomeric effect. A basic rule of conformational analysis, deduced from study of cyclohexane, states that the equatorial position is the energetically favored orientation of a large substituent on a six-membered ring. This principle, however, cannot be applied to polar substituents at an anomeric center in aldopyranose derivatives. Electronegative substituents on the anomeric carbon atom assume a higher abundance of axial positions than could be expected from the analogy with cyclohexane derivatives. This apparently anomalous situation was first discussed by Edward," and it has been termed the anomeric effect by L e m i e ~ x The . ~ terms exo-anomeric effect6 and reverse anomeric effect' were later introduced for the orientational preference of the aglycon around the glycosidic C - 0bond, and for the enhanced trend of the quarternary nitrogen atom to adopt an equatorial orientation, respectively. The anomeric effect, first identified in carbohydrate chemistry, is now recognized as being of a more-general importance for all molecules having two (or more) heteroatoms linked to the tetrahedral center. The unusual conformational behavior of this class of compounds containing the C - X C-Y moiety, where X = N, 0, or S, and Y = Br, C1, F, N, 0, or S was denoted as the generalized anomeric effect.*The focal segments of carbohydrate compounds having oxygen atoms as the heteroatoms X and Y, namely, acetals and hemiacetals, also belong to this class. Instead of the methylene group, some other tetrahedral groups may serve as the anomeric center. This subject has received very extensive experimentaland theoretical (1) E. L. Eliel, N. L. Allinger, S. J. Angyal, and G . A. Momson, Conformational Analysis,

Wiley, New York, 1965. J. F. Stoddart, Stereochemistry of Carbohydrates, Wiley-Interscience, New York, 197 1. N. S. Zefirov, Tetrahedron, 33 (1977) 3193-3202. J. T. Edward, Chem. Ind. (London), (1955) 1102- 1104. R.U. Lemieux, in P. de Mayo (Ed.),Molecular Rearrangements, Vol. 2, Interscience,New York, 1964, pp. 709-769. (6) R. U. Lemieux, A. A. Pavia, J. C. Martin, and K. A. Watanabe., Can. J. Chem., 47 (1969) 4427 -4439. (7) R. U. Lemieux and A. R. Morgan, Can. J. Chem., 43 (1965) 2205-2213. (8) A. J. de Hoog, H. R. Buys, C. Altona, and E. Havinga, Tetrahedron, 25 (1969) 3365-3375.

(2) (3) (4) (5)

ANOMERIC AND EXO-ANOMERIC EFFECTS

41

examination. From these studies, it has become increasinglyapparent that the anomeric effect is a complex phenomenon characterized,apart from the conformational preferences, by unique variations of valence geometry, reactivity, and other properties. This behavior reflects an enhanced stereochemical sensitivity of electron distribution of the structural segments in the vicinity of the anomeric center. In this vein, it is proper to speak about the different (that is, energetic, structural, and kinetic) manifestations of the anomeric effect. The present state of knowledge of the anomeric and exo-anomeric effects in carbohydratesis discussed in this article from the standpoint oftheoretical chemistry. Detailed theoretical argumentation supported by expenmental data is used to rationalize coherently the various manifestations of the anomeric effect, and to explain its origin. The chapter is divided into seven Sections. Several possibilities of specification of the energy associated with the anomeric effect are demonstrated in Section 11. Section I11 summarizes experimentaldata concerningthe configurationaland conformational equilibria of the anomeric center, and related variations of valence geometry. Section IV presents a survey of molecular orbital calculations of the anomeric effect based mostly on model compounds. The merits and limitations of semiempirical atom-potential calculations of the anomeric effect are described in Section V. The nature of the anomenc effect and the prominent role played therein by electron lone-pairs on oxygen atoms are elucidated in Section VI. Finally, in Section VII, some correlations between the stereoelectronic structure and the reactivity of anomers are discussed. 1-339-18

11. DEFINITION OF THE ANOMERIC AND EXO-ANOMERIC EFFECTS The free-energy aspect of the anomeric effect as displayed by equilibria of isomers is the area wherein the very concept was incepted and developed, (9) W. A. Szarek and D. Horton (Eds.), TheAnomeric Effect,Origin and Consequences,ACS Symposium Series, Vol. 87, Washington, 1979. ( 10) A. J. Kirby, TheAnomeric Effect andRelatedStereoelectronicEfects at Oxygen, SpringerVerlag, Berlin, 1983. ( I I ) 1. TvaroSka and T. Bleha, Chem. Papers, 39 (1985) 805-847. ( 1 2) I. TvaroSka, in G . Naray-Szabo (Ed.), Theoretical Chemistry of Biological Systems, Elsevier, Amsterdam, 1986, pp. 283-348. (1 3) C. Romers, H. R. Buys, and E. Havinga, in N. L. Allinger and E. L. Eliel (Eds.), Topics in Stereochemistry, Vol. 4, Interscience, New York, 1969, pp. 39-97. (14) R. U. Lxmieux, PureAppl. Chem., 25 (1971) 527-548. (15) J. C. Martin, Ann. Chim., 6 (1971) 205-218. (16) R. U. Lemieux, S. Koto, and D. Voisin, in Ref. 9, pp. 17-29. (17) R. U. Lemieux, Ann. N.Y. Acad. Sci., 222 (1973) 915-934. (18) R. U. Lxmieux and S. Koto, Tetrahedron, 30 (1974) 1933- 1944.

48

IGOR TVAROSKA AND TOMAS BLEHA

and which still dominates its investigation. Obviously, this is because the energy of the anomeric effect is quite amenable to quantification. In the following treatment, the role of the anomeric and related effects on the equilibrium at the anomeric center, and the various choices for definition of the associated Gibbs energy, are briefly described. 1. Anomeric Equilibria

Based on information derived from study of cyclohexane, steric analysis predicts that the most-stable conformation of pyranoses is a chair structure with bulky substituents in equatorial (e) positions. The orientation of a substituent on the anomeric center is, however, an exception, because, at that position, preference for an axial (a) position of polar substituents has been found in several pyranose derivatives. This anomeric effect, the enhancement in population (relative to cyclohexane) of a position having an electronegativesubstituent,is illustrated in Fig. 1 by the anomeric equilibria of some a-and P-D-glucopyranosederivatives.Alkoxy, acetoxy,and halogen substituentsprefer the a over the e orientation. The only substituent having a preference for the e position is the hydroxyl group: showing 36% of the a isomer. For comparison, there is only 1 1% of the a isomer at equilibrium in the reference cyclohexane derivative,cyclohexanol;that is, D-glucopyranose exhibits about three times more of the axial isomer than could be expected based on stericgrounds. This preferencefor the axial position increasesas the electron-withdrawingcharacter of the C- 1 substituent increases, and it also depends on the other ring-substituents and on the solvent (see Table I). The preference for the axial orientation over the equatorial, displayed by an electronegativesubstituent at the anomeric carbon atom of pyranoses is equivalent to the preference of the synclinal (sc or gauche) over the antiperiplanar (upor trans) orientation about the ring 0-5 - C- 1 bond in the C-5 -05-C- I-X segment, as shown in Fig. 2. The torsional angle about this and similar bonds in acyclic compounds will be denoted as 8, and the corresponding torsional potential as V(8). a. The Exo-anomeric Effect.-The term exo-anomeric effect was introduced to describe an orientational effect on the aglycon part of a glucopyranoside,6arising from the special properties of the acetal moiety. There is no difference in the nature of the anomeric and the exo-anomeric effects; each of them just applies to a different portion of the acetal segment C-5 - 0-5 -C1 - 0-1 - C-i. The anomeric effect is related to the preference of the axial orientation of the aglycon group in glycosides, that is, to the preference of the sc arrangement about the 0 - 5 -C-1 bond, whereas the exo-anomeric effect relates to the preference of the aglycon carbon atom C-i for the sc position at

ANOMERIC AND EXO-ANOMERIC EFFECTS

49

AcO AcO

OAc AcO

OAc

12%

88% (b)

-

HO OH

H o H -

OH

64%

36%

-

HO OH

OMe

67%

33%

(4 CH~OAC

A

c

:

AcO q CI

94% FIG.1.-The

A

e

c

O

S

c

,

AcO AcO

6%

Equilibrium Compositionsof 0- and /?-D-GluCOpymIOst? Derivatives.

the C- 1 - 0-1 bond rotational potential specified by the angle a.Obviously, the most important outcome of the exo-anomericeffect concernsthe relative stability of mutual orientationsof the neighboring saccharide units in oligoand poly-saccharides. In these cases, the conformational importance of the exo-anomenc effect surpasses that of the anomeric effect.

IGOR

50

TVAROSKA AND TOMAS BLEHA

TABLEI Effects of Ring Substituents on the Anomeric Equilibrium of Aldopyranoses and Derivatives Compound

Substituent on C-1

Axial anomer (To)

HO HO HO HO HO HO HO HO HO HO HO Me0 Me0 MeO Me0 AcO AcO

32-37 67-69 20 36 32-35 71 26 41.5

AcO

84'

5

AcO AcO

86" 86"

5

AcO

91"

5

D-Glucopyranose D-Mannopyranose D-Allopyranose D-Galactopyranose D-Xylopyranose D-Lyxopyranose D-Ribopyranose 2-Deoxy-~-urubino-hexopyranose 2-O-Methyl-~-mannopyranose 2,3-Di-O-methyl-~-mannopyranose 2,3,4-Tn-O-methyl-~-rnannopyranose Methyl D-glucopyranoside Methyl D-mannopyranoside Methyl D-galactopyranoside Methyl D-xylopyranoside 1,2,3,4-Tetra-O-acetyI-~-xyIopyranose

1,2,3,4-Tetra-O-acetyl-~-fucopyranose 1,2,3,4-Tetra-O-acetyI-6deoxy-6-iodo-~-glucopyranose.

15

80 86 67 94 71 69 83" 83"

References

19-25 19,23,24 23,24 23,24 19,23,24 23,24 23,24 23,24 24 24 24 26,27 26,27 26,27 26,27 5 5

1,2,3,4-Tetra-O-acetyl-6-chloro-6-deoxy-~-glucopyranose

1,2,3,4,6-Penta-O-acetyl-~-glucopyranose 1,2,3,4-Tetra-O-acetyI-6-O-t0syl-~glucopyranose

5

Calculated from the Gibbs energy differences given in Ref. 5.

The exo-anomeric effect is illustrated in Fig. 3, which shows three staggered orientations for rotation about the glycosidic bond in both the a and p anomer of methyl D-glycopyranoside. These are referred to as (+sc, sc), (+sc, up), and (+sc, -sc) and (up, sc), (up, -sc), and (up, ap), respectively, using two torsional angles (6and a)for specification of orientations of the C-5-0-5-C-1-0-1 -C-i moiety.

+

+

(19) H. S. Isbell and W. W. Pigman, J. Res. Nutl. Bur. Stand., 18 (1937)141- 194. (20)J. H.Brewster, J. Am. Chem. Soc., 81 (1959)5475-5483. (2I ) A. S. Hill and R. S. Shallenberg, Curbohydr. Res., 11 (1969)541-545. (22)M. Mathlouthi and D. V. Luu, Curbohydr. Rex, 81 (1980)203-212. (23) S. J. Angyal, Aust. J. Chem., 21 (1968)2737-2746. (24) S. J. Angyal, Angew. Chem., Int. Ed. Engl., 8 (1 969) 157- 167. (25)F. Franks, in J. M. V. Blanshard (Ed.), Polysuccharides in Foods, Butteworth, London, 1979,pp. 33-49. (26)C. T. Bishop and F. P. Cooper, Cun. J. Chem., 41 (1963)2743-2758. (27) V. Smirnyagin and C. T. Bishop, Can. J. Chem., 46 (1968)3085- 3090.

ANOMERIC AND EXO-ANOMERIC EFFECTS

H

X

H-1

sc

51

X

aP

FIG.2.-The Orientation Around the 0 - 5 -C-1 Bond in the Axial and Equatorial Forms of Aldopyranoses.

The qualitativeconformationalanalysis of the two anomers is straightforward. In the axial series, the (+ sc, -sc) conformer is, on steric grounds, very unstable, with the Me group lying below the ring in close proximity to the two axial hydrogen atoms on the C-5 and C-3 atoms. In the equatorial anomer, the (up, sc) conformer suffers from a repulsion between the Me group and the axial hydrogen atom on C-2. The exo-anomeric effect causes

+

I

0 \Me

( +sc, + s c ) + sc. ap) ( + s c , -sc) FIG. 3.-Staggered Orientations of the Aglycon in Methyl a-and PD-Glucopyranoside Characterized by Torsional Angles 0 and @.

52

IGOR

TVAROSKAAND TOMAS BLEHA

preference for the sc conformation; therefore, in both cases, the conformers having the sc position of angle a,that is, (+sc, +sc) and (up, -sc), are expected to be preferred over those having the up position, (+sc, up) and (up, up) for alkyl glycosides. In fact, analysis of the solid-state structures28-30 revealed that most carbohydrate derivatives adopt the (+sc, -I-sc), or (up, -sc), conformation. A particularly clear illustration of the working of the exo-anomeric effect is exemplified in the structure of a,a-trehalose. In the solid state, this compound exhibits’l an approximateC2symmetry, with two glycosidic-linkage torsion angles corresponding to the orientation (+sc, sc). The preferred conformation of a,a-trehalose in solution is very similar to that in the solid state, and there is no indication of significant proportions of any another conformer^.^^ Also, the simplest acetal, (CH30)2CH2(dimethoxymethane), which played a key role in the initial development of knowledge of the anomeric effect, has been shown to exist in the (+sc, sc) c~nformation.’~ The polymeric analog of dimethoxymethane, namely, poly(oxymethylene),has the analogous synclinal, helical c~nformation.’~

+

+

b. The Reverse Anomeric Effect.- Studies of the protonation of N-glycosyl-imidazolesand -pyrimidines s h o ~ e d ’ Jthat ~ , ~the ~ presence of the positive charge on the nitrogen atom linked to the anomeric center provides a strong driving-forcefor the aglycon to adopt the equatorial orientation (see Fig. 4). This preference for the e position in excess above the value that ensued from steric analysis of cyclohexane has been termed the reverse anomeric effect.’ The carbamoyl group, having only a partial positive charge on the carbon atom, has a reverse anomeric effect large enough to shift the equilibrium in hexopyranose peracetates toward the e form36(see Fig. 4). c. The Generalized Anomeric Effect.- Although the anomeric effect, by its original definition, applies to the properties of pyranoses, it turned out later that this effect is observed in a number of polar, acyclic and cyclic compounds, especially those involving a segment of the general formula - R - X - T- Y - . The group T represents a tetrahedral (anomeric) center of S. Perez and R. H. Marchessault. Curbohydr. Res.. 65 (1978) 114- 120. I. TvaroSka and T. KoEir, Chem. Zvesri, 35 (1981) 425-440. B. Fuchs, L. Schleifer, and E. Tartakovsky, Nouv. J. Chim., 8 (1984) 275-278. G. A. Jeffrey and R. Nanni, Curbohydr. Res., 137 (1985) 21 -30. K. Bock, J. Defaye, H. Driguez, and E. Bar-Guilloux, Eur. J. Chem., 13 1 (1983) 595-600. E. E. Astrup, Acra Chem. Scund., 25 (1971) 1494-1495. P. de Santis, E. Giglio, A. M. Liquori, and A. Ripamonti, J. Polym. Sci., Purr A . 1 (1963) 1383-1404. (35) H. Paulsen, Z. Gyorgydeak, and M. Friedman, Chem. Ber., 107 (1974) 1590-1613. (36) M. Chmielewski, J. N. &Miller, and D. P. Cerretti, J. Am. Chem. Soc.. 46 (1981) 39033908.

(28) (29) (30) (31) (32) (33) (34)

ANOMERIC AND EXO-ANOMERIC EFFECTS

CONH2

AcO

53

bAc

56 ”/.

44%

AcO AcO

FIG.4.-Examples

O Ac

of Equilibria Involving the Reverse Anomeric Effect.

the types-CHR-, -CH2, -POI--, -Si(CH,)2-, and-SO2-, and X and Y are such heteratoms as N, 0, S, and also halogens in the case of the terminal substituent Y. Acetals, thioacetals, substituted sulfides, phosphates, siloxanes, and other molecules having heteroatoms in geminal, 1,3 position in the backbone, belong to this group. The general preference for the sc orientation about the T - Y bond in the system R - X - T -Y - has been termed the generalized anomeric effect*and has been reviewed.lOJ* 2. The Energy of the Anomeric Effect The energy (or, better, enthalpy) and Gibbs energy of the anomeric effect can be deduced from knowledge of the isomer equilibria. It is, however, regrettable that several energy parameters related to sc - up and similar equilibria are tacitly used as an energy measure of the anomeric effect. Consequently, the magnitude of the anomeric effect of a given substituent dependson the procedure applied at its derivation,a fact that makes comparison difficult and that can be a source of confusion. The most frequently used measure of the anomeric effect is based on the comparison of the stability of 2-substituted oxane (tetrahydropyran; THP) and cyclohexane. In general, conformationalproperties of the oxane ring are similar to those of cyclohexane, with dominance of a chair conformation. It is further presumed that steric interactions in oxane are the same as in cyclohexane, with preference for equatorial positions of bulky substituents. The Gibbs energy of the anomeric effect, AG(AE l), can be expressed as the

IGOR TVAROSKA AND TOMAS BLEHA

54

X

act"

(b)

ax

X

FIG.5.- The StandardEquilibrium in (a) Substituted Oxanes and (b) Substituted Cyclohexanes Used in the Definition of the Anomeric Effect by Eq. I .

difference of the standard, conformational Gibbs energies for substituted oxane (AG:), shown in Fig. 5a, and for the same substituent on cyclohexane, AG:; see Fig. 5b. AG(AE1) = AGO, - AG!

(1) The term “A parameter” is also used in the literature for the Gibbs energy -AG:. According to the definition in Eq. I, the anomeric effect AG(AE1) depends on the A value of the substituent, and on the temperature and solvent used in measurement of equilibria in Eq. I. This definition can be extended to multisubstituted pyran derivatives, and simple, additivity scheme^^.^^,^^ of steric interactionsof substituentsare used for the estimation of AG(AE1). These semiquantitative schemes were based on the assumptions that the pyranoid ring has the same geometry as cyclohexane and that the relative free-energies of each chair form may be obtained by summation of the interaction energies of substituentsthat are independentof one another, and by taking into account the value of the anomeric effect. The values for interaction energies were obtained experimentally from the equilibria of various cyclitols and p y r a n ~ s e s .The ~ ~ ,anomeric ~~ equilibrium between aand P-D-glucopyranose(see Fig. lb) provides a simple illustration of such calculations using a simple additive s ~ h e m e . The ~ ~ Janomers ~ of D - ~ ~ U C O pyranose differ only in the configuration at the anomeric center; consequently, only interactions involving the anomeric hydroxyl group are relevant to the conformational equilibrium. From comparison of the steric interactions in each anomer, it is seen that the a anomer has two additional 1,3 diaxial interactions (OH :H). The same two interactions are responsible for the e preference of the hydroxyl group of cyclohexanol. The addition of steric parameters predicts that the a anomer should be 3.8 kJ.mol-’ less

ANOMERIC AND EXO-ANOMERIC EFFECTS

55

TABLE I1 Axial Preferences and the Gibbs Energy of the Anomenc Effect AG(AE,) (in kJ.mol-I) in the 2Substitnted Oxanesa Group

96"

AG:

A*

Br C1 F,CCH,O CI,CCH,O Cl,CHCH,O CICH,CH20 Me0 EtO C,H,O GH9O Me,CHO Me,CO PhO AcO MeS EtS C,H,S C4H9S Me,CHS Me,CS HO (CH,),N MeHN Me0,C H,NOC

96d 96d 92' 95e 88' 77' 82 80 82 82 75 67 81 73 69 69 69 70 70 70 47 28 18 6 4d

7.5 7.5 6.3 7.5 5.0 3.2 3.8 3.4 3.8 3.8 2.7 1.7 3.8 2.5 2.0 2.0 2.0 2.1 2.1 2.2 -0.3 -2.3 -3.9 -5.8 -7.8

2.1 2.1

AC(AE,)

AC(AE,)i

References

9.6 9.6

10.8 10.8

4.21 4.11

8.0 7.5

10.3 9.8

4.11 2.9 3.3 4.28

5.8 6.7 5.8 6.2

8.1 8.3 7.6 8.5

4.2 6.4* 5.4 5.3 5.3'

3.9 4. I 1.5 -0.5 - 2.5

6.2 7.6 4.4 2.4 0.4

37 38 39 39 39 39 8 8 8 8 8 8 39 41 42 42 42 42 42 42 43 44 44 45 46

In CCI,, unless specified otherwise. The A-values are from Ref. 47, unless otherwise specified. Using A(oxane)values from Q.2. Neat. In 1,4dioxane.fRef. I. 8 Ref. 48. For the N(CH,), group, the A value ofthe N(CH,), group was used.-' In pyridine.' For CONH,, the Same value as for the MeO,C group.

(37) G. E. Booth and R. J. Ouellette, J. Org. Chem., 31 (1966) 544-546. (38) C. B. Anderson and D. T. Sepp, J. Org. Chem., 32 (1967) 607-61 1. (39) G . 0. Pierson and 0. A. Runquist, J. Org. Chem., 33 (1968) 2572-2574. (40) H. Booth, T.B. Grindley, and K. A. Khedhair, J. Chem. Soc., Chem. Commun. (1982) 1047- 1048. (41) C. B. Anderson and D. T.Sepp, Chem. Ind. (London)(1964) 2054-2056. (42) A. J. de Hoog and E. Havinga, Red. Truv. Chim. Pays-Bus, 89 (1970) 972-979. (43) A. El-Kafrawy and R. Perrand, C. R. Acud. Sci. Ser. C, 280 ( I 975) I2 19- 1221. (44) D. Barbry, D. Couturier, and G. Ricard, J. Chem. SOC., Perkin Trans. 2, (1982) 249-254. (45) E. L. Eliel, K. D. Hargrave, K. M. Pietrusiewicz, and M. Manoharan, J. Am. Chem. Soc., 104 (1982) 3635-3643. (46) I. TvaroSka, M. Hricovini, M. Chmielewski,J. Jarosz, and B. Hintze, unpublished results. (47) H.-J. Schneider and V. Hoppen, J. Org. Chem., 43 (1978) 3866-3873. (48) F. R. Jensen, C. H. Bushweller, and B. H. Beck, J. Am. Chem. Soc., 91 (1969) 344-351. (49) H. Booth and M. L. Josefowicz, J. Chem. SOC.Perkin Trans. 2 (1976) 895-901.

IGOR TVAROSKA AND TOMAS BLEHA

56

TABLE111 Axial Preferences and the Gibbs Energy of the Anomeric Effect AG(AE,) (in kJ.moV) in the 4- and 6-Methyl Derivatives of 2-Substituted Oxanesu ~~

Derivative 4-Methyl

6-Methyl

~

Substituent

%a

AGZ

HO Me0 EtO Me,CHCH,O AcO Me0,C c1 Br I Me0 EtO Me,CHO Me,CHCH, Me,O CF,CH,O HC=CCMe,O AcO MeS Me,CS Me0,C

58 82 78 79 72

0.8 3.8 3.1 3.3' 2Sd -5.2' 9.0' 11.3 11.3 3.0 2.9 2.7 3.1' 2.1 3.4 2.2 2.7d

11

96 97 97 77 16 75 78 70 80 71 75 64 66 I1

1.4

1.6 -5.2e

AqAE,)

AG(AE,Y

References

8.0 7.2

7.3 10.3 9.5

5.8 0.1 11.1 13.4 13.0 7.2 7.0

7.6 3.0 12.3 14.6 14.0 9.5 9.2

41 50 50 51 51 52 38 38 38 50 50 50

5.0

51

50 50 50 6.0 5.6 5.8 0.1

7.8 7.9 8.1 3.0

51

50 50 52

In CCl,, unless specified otherwise. Using A(oxane) values from Eq. 2. In 1,4dioxane. In acetic acid. In methanol. / N e a t .

stable than the ?/ anomer. The experimentally observed Gibbs energy difference in waterz3is 1.5 kJ.mol-L.The difference of 2.3 kJ.mol-' between the two values represents the magnitude of the anomeric effect AG(AE1). Examples of variation of the anomeric effect AG(AE 1) with the substituents are shown in Tables I1 and I11 for oxane derivatives.The application of Eq. 1 to the results of the measurements of anomeric and conformational equilibria have establishedZthat the anomeric effect decreases in approximately the following order: halogen > PhCOz > AcO > AcS > RO > RS > HO > NH2 > Me0,CO > imidazolium > pyridinium. In reality, however, steric interactions in oxane and in a cyclohexane derivative are not the same. Because C - 0 bonds are shorter than C-C bonds, repulsive interactions of an axial group on a pyranoid ring are likely to be larger than those of the same group on the cyclohexane ring, and the (50) E. L. Eliel and C. A. Giza,J. Org. Chem.. 33 (1968) 3754-3758. (5 1) C. B. Anderson and D. T. Sepp, Tetrahedron,24 ( 1968) 1707- 1716. (52) C. B. Anderson and D. T. Sepp, J. Org. Chem., 33 (1968) 3272-3276.

ANOMERIC AND EXO-ANOMERIC EFFECTS

57

anomeric effect based on Eq. I is underestimated. F r a n ~ estimated, k~~ in a new way, the steric part of the Gibbs energy difference (AGZ), for an equilibrium shown in Fig. 5a. This quantity represents the apparent size of the substituent on oxane, or the parameter A(oxane), and correlates with the AGg value.

+

(AG:), = A(oxane) = 1.53 X AGE 0.08 (2) Eq. 2 shows that the A(oxane) parameters appropriate for oxane should be 50% larger than the values currently used. Evidently, the use of the (AGZ), term in Eq. I, instead of AGE, brings about a large amplification of the anomeric effect AG(AE1) as documented in Tables I1 and 111 for substituted oxane. Two values of AG(AE1) in these Tables exemplify an essential drawback of the definition of the anomeric effect by Eq. I: its change in magnitude with the value assigned for the A parameter. The A factors may also vary with the method of their determination. For example,the A value for an OH group in CCl, was reported' to be in the range of 1.2-6.5 k.l.mol-', and this uncertainty is transferred to AG(AE1). Moreover, because the A parameters are solvent-dependent, so are the AG(AE1) values. Using Eq. I, even a qualitative decision about the presence of the anomeric effect can sometimes be ambiguous. For example, from study of 2,3,4-tri-O-acetylpentopyranosylamines, it was concluded35that the amine group does not exhibit the anomeric effect. However, a correction of the A value for this group,54 5.9 kJ.mo1-I according to Eq. 2, results in -3 kJ.mol-' larger preference of the a form than could be expected on steric grounds, and therefore, the NH2 group should exhibit the anomeric effect. The other definition implicitly utilized for an estimation of the anomeric effect is based on comparison of the Gibbs energy difference AGZ with the energy AEpFobtained from semiempirical calculations using the atom-potential functions. These methods, stemming from classical physics, vary in their complexity from a simple evaluation of steric energy by atom - atom potential^'^.^^ to detailed description of the force field in a molecule by molecular mechanics method^.^' In this concept, the energy of the anomeric effect, AE(AE2), is determined as that part of the potential energy (or Gibbs energy) that is not accounted for by the calculation procedure and is "missing" in AEpF: AGZ (53) (54) (55) (56) (57)

AEZ = AEpF

+ AE(AE2)

R. W. Franck, Tetrahedron, 39 (1983) 3251 -3252. G. W. Buchanan and V. L. Webb, Tetrahedron Lett., (1983) 4519-4520. K. S. Vijayalaksami and V. S. R. Rao, Carbohydr. Res., 22 (1972) 413-424. A. Abe, J. Am. Chem. Soc., 98 (1977) 6477-6480. N. L. Allinger, J. Am. Chem. SOC., 99 (1977) 8127-8134.

(3)

58

IGOR

TVAROSKAAND TOMAS BLEHA

Furthermore, a questionable approximation of the same entropy and volume of a and e isomers is usually assumed. Obviously, the energy of the anomeric effect, AE(AE2), depends on the quality of the method used for determination of AEpF.The extra function AE(AE2) may differ, depending on whether AGg or AEE has to be matched. Furthermore, if Eq. 3 applies for an equilibrium in a solvent, the extra term AE(AE2) also includes a contribution of the solvent effect due to its neglect or an incompleterepresentation in energy AEpF. The measures of the anomeric effect, based on Eqs. 1 -3 are of relative character, because they are expressed in reference to a standard compound or a computational method. Some absolute measure is needed for theoretical considerations, and it could be simply the positive difference of the energy of the a and e isomers or of the sc and ap conformation in model compounds.58 AE(AE3) = E, - E,

(4)

This definition does not take into account the usual preference of bulky substituents for the equatorial position in cyclic compounds, and, with the assumption of the same entropy and volume for the a and the e isomer, corresponds to the AG: value in Eq. 1. In this case, a molecule exhibits the anomeric effect if the axial position (sc orientation) is more stable than the equatorial position (up orientation). All three definitions of the anomeric effect are interrelated,but as the data required for direct recalculation of one definition into another are frequently lacking, a substituent can be characterized by the diverse data about the energy of the anomeric effect. Their comparison for various groups needs caution, and inspection as to how they were originally calculated. Definitionsbased on Eqs. I , 3, and 4 should, in principle, also apply for the exo-anomeric and reverse anomeric effects. There are, however, some problems with the practical application of Eq.I in the case of the exo-anomeric effect, because the AGg values are largely not available. For the exo-anomeric effect, the conformational equilibrium is specified by two dihedral angles, 8 and a,and the value of AGZ is needed for all six individual conformers shown in Fig. 3. Because rotation around the exocyclic bond by angle 0 is much less restricted in comparison with rotation by angle 8, a mixture of conformers was experimentally observed, with a difficult resolution of AGX into individualcomponents. Ifthe exo-anomenc effect is treated by Eq. 3, the extra term AE(AE2) should be redefined for the whole range of values of the torsional angle a.Due to the lack of experimentaldata on AGg or AEE, for each conformer in Fig. 3, the energy values calculated “correctly”, for example by some molecular orbital method, are used, instead of (58) S. Wolfe, M. H. Whangbo, and D. J. Mitchel, Curbohydr.Rex, 69 (1979) 1-26.

ANOMERIC AND EXO-ANOMERIC EFFECTS

59

those in Eq. 3. Because the rotational potential V(Q) depends on the orientation around the endocyclic, 0 - C bond, the AE(AE2) term should be considered separately for each anomer. Several functions have been p r ~ p o s e d ~ ~ - ~ ' for AE(AE2), and these are reviewed in Section VI, which is devoted to calculations of potential function. A negative value of AG(AE1) represents the reverse anomeric effect. It could, perhaps, also be defined in the context of E q . 4. Here, the reference saturated hydrocarbons exhibit negative energy AE(AE3), and the reverse anomeric effect could be defined as an excess energy (in absolute value) over that for the reference molecules. 111. EXPERIMENTAL DATAON THESE EFFECTS A considerable amount of data has been accumulated during the past two decades on the anomeric effect in terms of structure, energy, reactivity, and other properties. In this Section are discussed some pertinent data, and the ensuringgeneralizationsconcern the anomeric equilibrium and geometrical structure of isomers, concentrating on pyranoses and their models. In this respect, the multidimensional character of the anomeric effect has to be emphasized. For its full structural description, the torsional angle 8 (and @ for the exo-anomeric effect) has to be supplemented by the data on bond lengths and valence angles in the vicinity of the anomeric center.

1. Relative Abundance of Isomers

Determination of the relative representation of the a and e forms of pyranoses can be a very difficult and demanding task. In general, the abundance of isomers at equilibrium depends mainly on the type of substituent (aglycon) on the anomeric center, on the other substituentson the ring, and on the solvent. Studies of acyclic and cyclic models are of invaluable assistance in this field (see Fig. 6). Substituted dimethyl ethers bearing an electronegative group X, as in CH30CH2X,are the simplest acyclic compounds wherein the anomeric effect is operative. For C1 and F substituents, the sc orientation, with the 0 -C torsional angle - 69 - 7 1 was, from the microwave spectra,62 found to be the most stable. N.m.r. measurement^^^ ofthe C1derivativegave an energy difference of 6.3 - 8.4 kJ.mol-' between the up and sc forms. For O,

(59) R. U. Lemieux, K. Bock, L. T. J. Delbaere, S. Koto, and V. S. R. Rao, Can. J. Chem., 58 (1980) 631 -653. (60) H. Thragersen, R. U. Lemieux, K. Bock,and B. Meyer, Can. J. Chem.. 60 (1982)44-57. (61) I. TvaroSka, Carbohydr. Res., 125 (1984) 155-160. (62) M. Hyashi and H. Kato, Bull. Chem. Soc. Jpn.. 53 (1980) 2701 -2710. (63) F. A. L. Anet and I. Yavari, J. Am. Chem. SOC.,99 (1977) 6752-6753.

60

IGOR

TVAROSKAAND TOMAS BLEHA

x

FIG.6.-Axial-Equatorial Equilibrium in (a) Complex Aldohexopyranoses, (b) a Simple Derivative of Oxane, and (c and d) CorrespondingConformational Equilibria in Acyclic Model Compounds.

+

such OCH, derivativesas dimethoxymethane,the (+sc, sc) conformation with the methyl groups on opposite sides ofthe OCO plane is the most stable. The rotation around each C - 0 bond into the ap position is disfavoredH by -7.1 kJ.mol-'. Qualitatively similar trends are also observed in longer ethers, but, in most cases, the high flexibility of the chains complicates the resolution of individual isomers. Consequently, cyclic models are much more convenient, particularly substituted tetrahydropyranshaving an elec(64) T. Uchida, Y. Kurita, and M. Kubo, J. Polym. Sci., 29 (1956) 365-373.

61

ANOMERIC AND EXO-ANOMERIC EFFECTS TABLE IV ConformationalEquilibria of Tri-Oacetyl and Tri-Obeozoyl PD-Xylopyranose Derivatives as a Function of the Anomeric Groupa (see Fig. 7) 'C, Conformer at (oh) Equilibrium ~~

~

Anomeric group

R = Ac

References

R = Bz

References

H Me0 AcO BzO AcS F CI Br NH* NHAc NHCOCF,

13 19 28 39 28 80 -90 79

72 73 74 75 76 77 78

19 26 47 50

72 73 75 74

90- 100 98 90- 100

77 78 79

5'5 56 56 20b 65b 95b 956

35 35 35 35 35 35 35

N3

ImidazolyF Pyridiniumc Imidazoliumc

a In acetone unless specified otherwise. In CDCI,. a-D-Xylopyranosederivatives.

tronegative group on C-2 (see Fig. 6b). Numerous AGg values for the anomeric equilibrium in these compounds are available.6,7,37-4a.50-5z,65-7' F. Sweet and R. K. Brown, Can. J. Chem., 46 (1968) 1543- 1548. N. S. Zefirov, V. S. Blagoveschensky, I. V. Kazimirchik, and N. S . Surova, Zh. Org. Khim., 5(1969) 1150-1151. N. S . Zefirov, V. S. Blagoveschensky, I. V. Kazimirchik, and N. S. Surova, Tetrahedron,27 (1971) 31 11 -31 18. C. B. Anderson and M. P. Geis, Tetrahedron, 31 (1975) 1149- 1154. A. J. de Hoog, Org. Magn. Reson.. 6 (1974) 233-235. N. Pothier, D. D. Rowan, P. Deslongchamps,and J. K. Saunders, Can. J. Chem., 59 ( I 98 1) 1132- 1139. R. U. Lemieux and J. Hayami, Can. J. Chem., 43 (1965) 2162-2173. P. Luger, G. Kothe, K. Vangehr, H. Paulsen, and F. R. Heiker, Carbohydr.Rex, 68 (1979) 207-223. P. L. Durette and D. Horton, Curbohydr. Res., 18 (1971) 403-418. P. L. Durette and D. Horton, J. Org. Chem., 36 (1971) 2658-2669. P. L. Durette and D. Horton, Carbohydr. Rex, 18 (1971) 389-401. P. L. Durette and D. Horton, Carbohydr. Rex. 18 (1971) 419-425. H. Paulsen, P. Luger, and F. R. Heiker, in Ref. 9, pp. 63-79. P. L. Durette and D. Horton, Carbohydr. Rex, 18 (1971) 57-80. P. L. Durette and D. Horton, Carbohydr. Res., 18 (1971) 289-301.

62

IGOR

TVAROSKAAND TOMAS BLEHA

a. The Character of the Ag1ycon.-The equilibrium composition of several derivatives of oxane and pyranoses are summarized in Tables I-V. Several observations can be made on inspection of these data. The preference for the axial position increaseswith the electron-withdrawingcharacter of substituent X, and is most conspicuous for the halogen and alkoxy derivatives (see Table 11). An increase in size of the alkoxy group diminishes the preponderance of the a form, such that changingfrom a methoxyl to tert-butoxy group in 2-substituted oxane decreases the abundance of the a form by 15%. This phenomenon was found to be caused by the entropy, rather than the enthalpy, term.* The size of a substituent does not seem to influence the equilibrium of alkylthio derivatives.An enhancement of the electronegativity of X by change from the ethoxy to the trichloroethoxy derivative increases the population of the axial form from 80 to 95%. Similarly, in halogen derivativesof oxane, the a forms are the sole detectable species(seeTable 11). A tendency toward stabilization of the a form is usually quite pronounced

in the acetyl and benzoyl derivatives of fi-D-xylopyranosyl halides (see Fig. 7a) which, in solution, exist mainly or completely in the 'C.,conformation, with all substituents in axial position^.^^-*^ This was first pointed outrn for TABLE V ConformationalEquilibria of Tri-O-acetyl- and Tri-0-benzoyl-p D-ribopyranose Derivatives as a Function of the Anomeric Group" 'C, Conformer at Equilibrium (96)

Anomeric group

H Me0 EtO Me,CHO Me,CO AcO BzO AcS C1

Br NHAc N, SPh, @

R = Ac

References

R = Bz

References

24 61 61 62 54 57 56 34 94 95

46 80 81 78 74 78 77

77 79 79 79 79 75 74

98

78

38b

77 79 79 79 79 74 75 76 78 78 35 35

56b

35

56

In acetone, unless specified otherwise. In CDCI,.

(80) P. L. Durette and D. Horton, Adv. Carbohydr. Chem. Biochem., 26 (1971) 49- 126. S.Jewell, J. Org. Chem., 32 (1967) 1818- 1820.

(80a) C. V. Holland, D. Horton, and J.

ANOMERIC AND EXQANOMERIC EFFECTS

63

X

R = Ac, B z ; X =CI, Br, I

1970

81%

FIG. 7.-The Isomer Equilibrium in (a) DXylopyranose Derivatives and (b) a Related Compound Having the “Anomeric Center” Unsubstituted.

tri-0-acetyl-j3-D-xylopyranosyl chloride. In this case, the anomeric effect prevails over the unfavorable 1,3-diaxial interactions of bulky substituents. At the same time, this result indicates that the 1,3-diaxial interactions of benzoyloxy and acyloxy groups are much weaker than might be expected. This conjecture is supported by the observation72of a 19%population of the a form in a related compound having an unsubstituted anomeric center, (see Fig. 7b), and of 13% in namely, 1,5-anhydro-2,3,4-tri-0-benzoylxylitol the analogous tri-0-acetyl derivative. The preference for the axial position diminisheswith lowered electronegativity of atoms linked to the anomeric center; that is, F > 0 > N > C for the first row of the Periodic Table. For the latter two elements, N and C, the anomeric equilibrium depends on the overall polarity of the substituent. Thus, derivatives of substituted D-arabinopyranose (see Fig. 8) contain 94% of the isomer having a nitro group in the axial position at equilibrium.81N.m.r. measurement^^^ and other s t ~ d i e s ~ of J ~substituted , ~ , ~ ~ Npentopyranosyl derivativesshowed that the preference for the axial position decreases in the order NO2 > N-PPh, > N, > NHCOCF, > NH2 >

-

(8 1) B. Aebischer, R. Hollenstein, and A. Vasella, Helv. Chim.Acta, 66 (1 983) 1748 - 1754. (82) P. Finch and A. G . Nagpurkar, Curbohydr.Res., 49 (1976) 275-287.

64

IGOR

TVAROSKAAND TOMAS BLEHA

94 Yo 6% FIG.8.- The Equilibrium Composition of a Substituted2-Deoxy-~-arubinc-hexopyranosyl Nitrate.

95 %

35 7 0

65 %

p&&

i

AcO

cF3c02-

AcO

(62

cF3c02-

Ac

OAC

N

FIG.9.- Equilibriafor D-Xylopyranose DerivativesHaving N-Substituentsat the Anomeric Center.

ANOMERIC AND EXO-ANOMERIC EFFECTS

65

NHAc > NHPPh, > imidazole > imidazolium > pyridinium. The groups in the middle of the series, such as NHAc, NHCH, ,and aziridine, display a slight preference for the e form. The reverse anomeric effect is typical for imidazole and all groups having positively charged nitrogen. The variation of the anomeric equilibrium with the character of the N-substituent is illustrated in Fig. 9, and is expressedquantitativelyin Tables IV and V. Whereas, in amino-substituted&D-xylopyranosederivatives,the a form is preponderant, the e form prevails with the imidazole substituent, and n.m.r. spectroscopy showsonly the e form ofthe 'C,conformer in the case ofthe protonated imidazole ring. Carbon atom substituents on the anomeric center generally favor the equatorial position, but the minimal preference (1.4 kJ.mo1-l) of the e relative to the a position was observed for the ethynyl in oxane. A methoxycarbonylgroup45,52 gives a slightly larger population of the e isomer when linked to the oxane ring (relative to cyclohexane). This reversed anomeric effect is very pronounced for a carbamoyl group. The corresponding derivativeof oxane existsmainly (90%)in the equatorial form.&Similarly,in 2,6-anhydroheptonamide~,~~ the carbamoyl group has a strong preference for the equatorial position (see Fig. 10). For example, 3,4,5,7-tetra-O-acetyl2,6-anhydro-~-glycero-~-gluco-heptonamide in the EDconfiguration displays a considerable proportion of the 2C5conformation (56% in CDcl,), despite extensive 1,3diaxial interactions of four bulky groups. When the more-polar solvent Me2S0 was used, this compound was present almost solely in the *C,conformation.

-

CH2OAc I

L

AcO

CONH2

5

C2

FIG. 10.- The Anomeric Equilibrium of 3,4,5,7-Tetra-O-acetyI-2,6-anhydro-~-glycero-~gluco-heptonamideand Its Population in Various Solvents, Illustratingthe Reverse Anomeric Effect of the Carbamoyl Group.

66

IGOR

ACo=oAc AcO

TVAROSKAAND TOMAS BLEHA

-

A

c AcO

O

m AcO

OAc

X = H , CH3, CH21,CH2CI,CH20Ac, or CH20Ts FIG. 1 1.-Equilibria of D-Xylopyranose Tetraacetate and Its Relatives Listed in Table I.

b. The Other Ring Substituents.- It is well known that the presence and configuration of a hydroxyl group on C-2 of the pyranose ring markedly affectsthe anomeric Thus, in the case of D-mannopyranose, the axial hydroxyl group on C-2 increases the presence of the a anomer (69%)relative to that for 2-deoxy-~-arabino-hexopyranose (47.5%), which has no hydroxyl group on C-2. Conversely, when the hydroxyl group on C-2 is in the equatorial position, as in D-glucopyranose, the proportion of the a anomer decreases to 36%. These results, summarized in Table I, also show that the same trend, once termed the A2 effect,83is operative in pentopyranoses and methyl glycosides. The data for various methylated D-mannoses, given in Table I, indicate that the equilibrium composition changes in favor of the cy anomer as the degree of methylation is increased.24 The electronegativity of the substituent on C-4 also influences the anomeric equilibrium. Consequently, 2,4-dimethoxyoxane exists in methanol as an equilibrium mixture containing 80% of the isomer having45an axial methoxyl group on C-2, compared with 67 - 69% for 2-methoxy-4-methyloxane. 5 O 3 Finally, some examples of the role of the substituent at C-5 in the anomeric equilibrium may be mentioned. Anomeric equilibria for a series of substituted pentose and hexose derivatives (see Fig. 1 l), given in Table I, show that an increase in the electronegativityof the equatorial substituent at C-5 increases the axial preference of the acetyl Study of the stereoisomeric aldopyranose derivative^^^,'^-^ revealed a significant population of both chair conformers, although some limiting cases were observed in which one conformation is very strongly favored. Representative examples are given for the &D-xylopyranoseseries in Table IV, and for the fi-D-ribopyranose series in Table V, respectively. It may be seen that the anomeric effect dominates the conformational preferences, but, in general, the presence of several bulky substituentson the pyranose ring makes the anomeric equilib(83) R. E. Reeves, J. Am. Chem. Soc., 72 (1950) 1499- 1506.

ANOMERIC AND EXO-ANOMERIC EFFECTS

67

TABLE VI Influence of the Solvent on the Axial Preferences of the Hydroxyl, Methoxyl, Methylthio, Aziridinyl, and Carbamoyl Groups in 2-Substituted Oxane Derivativesa Dielectric constant Solvent Neat I ,4-Dioxane Carbon tetrachloride Benzene Carbon disulfide Chloroform Pyridine Acetone Methanol Acetonitrile Dimethyl sulfoxide Water

(E)

Oh

H@

47 32

MeSd

82,83 82 80 ?1,78

(CHJ,Ne

45 17

72 69 65,68 74 52

CONH/

31 69 66 59

28 31 29 32

55

37.5 46.7 78.3

MeO' 72 77

2.2 2.2 2.3 2.6 4.8 12.4 20.7 32.7

of the Axial Conformer

4 4 1

I

54 50 23 24

5

8

Data from Ref. 43. From Refs. 6 , s . 39,50,and 5 I . From Ref. 42. dr From Ref. 44. /From Ref. 46.

ria very intricate, and even the all-axial 'C, form can be ~ b s e r v e d ,as ~~ in. ~ ~ /3-D-xylopyranose tetraacetate (28%),or in the corresponding tri-0-benzylj?-D-xylopyranosyl acetate (47%). As already noted, this form may even preponderate in similar halogen derivatives (see Fig. 7a).

c. The Effect of the Solvent.-The variation ofthe axial preferencebased on the electronegativity of the ring substituents,and of the aglycon group, as already discussed, suggeststhat this phenomenon may be sensitiveto solvation. Table VI shows the abundance of the u form for 2-substituted oxane derivatives(see Fig. 12)in a wide range of solvents. For HO, MeO, and MeS

X

X = OH, OMe, SMe, N(CH2)2, or CONH2 FIG. 12.-Equilibrium for 2-Substituted Oxane Derivatives. (84) J. B. Lambert and S . M. Wharpy, Curbohydr. Rex, 115 (1983) 33-40,

IGOR

68

TVAROSKAAND TOMAS BLEHA

groups, the axial preference is seen to be higher in nonpolar solvents, and lessened in more-polar solvents, although the major difference is between the values in organic solvents and those in water. In dimethyl sulfoxide, the axial preference appears to be higher, as expected on the basis of correlation with the dielectric constant. As may be seen from Table VI, the conformational equilibria of oxane substituted with N(CH2)2and CONHl groups at C-2 are less sensitiveto the solvent than are those having HO, MeO, and MeS aglycon groups. The described trends in solvent effect are also apparent in the data on ~-glucopyranose.~~-~~~*~ The proportion of the a anomer of D-glucopyranose in pyridine is 45%, and in dimethyl sulfoxide it is 44%, as compared with only 32-37% in water. Conversely, in the case of the TABLEVII Illustrative Examples of the Influence of Solvents and Ring Substituents on the Gibbs Energy AWAE,) Magnitude (W.mol-') of the Anomeric Effect for Methoxyl, Hydroxyl, and Aziridinyl Groups" Group Me0

Compound Oxane 2-methoxy-

Solvent CCl, C6H6

CDCI, MeCN

HZO HO

(CH,),N

2-methoxy-4-methyl2-methoxy-6-methyl2-hydroxy-

CCl, CCl, CCI, CSZ Me,SO H*O HZO H*O HZO

D-Glucopyranose D-Mannopyranose 2-0-methyl2,3-di-O-methylH20 2,3,4,6-tetra-O-methylHZO 5-Thio-~-xylose H2O 2-Deoxy-~-arabino-hexopyranose HZO 2-(Aziridin-1-yl)oxane CCI, CSZ C6H6

CDCI, Me,SO MeCN

AWAE,)

References

10.3 10.3 9.7 8.0 6.7 10.0 9.6 6.2 4.5

8 8 8 8 8 50 50 43 43 43 43 24 24 24 24 24 84 24 44 44 44 44 44

6.0 2.6 5.0 8.4 9.2 10.1 11.0 10.8 6.3 7.5 7.6 7.8 8.0 6.9 6.9

44

a Based on the constant A(oxane) values of 6.5, 6.5, and 9.9 kJ.mol-I for the MeO, HO, and (CH,XN groups, respectively.

ANOMERIC AND EXO-ANOMERIC EFFECTS

69

N(CH2), group, the polarity of the solvent has little effect on the anomeric ratio, and the proportion of the a anomer is the same in dimethyl sulfoxide and water solution.44 Table VII presents data that illustrate the influence of the solvent and ring substituents on the anomeric effect of methoxyl, hydroxyl, and aziridinyl groups. For the purpose of this Table, the anomeric effect is defined by Eq.I. As noted, this definition takes into account the steric preference of an aglycon for the equatorial position by the A(oxane) value from Eq. 2. The comparison is, however, handicapped by the lack of accurate data on the dependence of the A values on the solvent, even if the increasein the “apparent size” of the aglycon due to solvent might be small. For example, the A value ofthe OMegroup increases from 2.5 kJ.mol-I to 4.2 kJ.mol-’ ongoing from CC14to water.’ It may be clearly seen from Table VII that the anomeric effect of HO and Me0 is higher in less-polar solvents, with dimethyl sulfoxide being the only exception. These data also document how problematic it could be to characterize the anomeric effect, or the reverse anomeric effect, of a given group by a single universal value which would serve in all cases, regardless of the other ring substituents and the solvent. In summary, experimental data on the isomeric abundances at anomeric equilibrium reveal that the preference for the axial position depends on several, interconnected factors which were clarified in surveys on carbohydrate stereochemistry,2*80 and these provided a background for ensuing theoretical studies. The elucidation of this relationship in complex carbohydrates is greatly facilitated by measurements on the simple derivatives of oxane, and qualitative trends have already been established. Table I1 illustrates several possibilities of the quantification of the energetic aspect of the anomeric effect. The procedure most frequently used, based on Eq. I, suffers from the ambiguity of the A values for the oxane ring and by their presumed variation with solvent.

d. The Exo-anomeric Equilibria.- Because of the lessened barrier of internal rotation around the exo-anomeric C -0 bond, characterizationof the conformational equilibrium of the type shown in Fig. 3 is very difficult, and complete data have not been reported so far. In contrast to the anomeric torsional angle 8,restricted to a narrow range of values, the exo-anomeric angle (D displays a much broader distribution in all six conformers in Fig. 3. Available information indicates that the (up, sc) and (+sc, sc) orientations are the most favored ones for the e and a isomers, respectively. Measured and calculated dipole moments of alkoxy and alkylthio derivatives of oxane have been ~ o m p a r e d ~by , * using , ~ ~ the coupling constants of the anomeric proton for estimation of the abundance of the a and e forms. It was concluded*that, of the six possible conformations shown in Fig. 3, only the

+

+

IGOR TVAROSKA AND TOMAS BLEHA

70

+

+

sc) conformation of the u isomer and either the (up, -sc) or (up, sc) conformation of the equatorial isomer is present. Because the dipole moment calculated for both of the latter conformations had the same value, their relative abundance could not be determined. Lemieux and cow o r k e r ~ ~did ~ . 'not ~ detect any appreciable amount of the (up, +sc) conformer in a number ofstructuresexamined, and concludedthat the (up, - sc) and (up, sc) rotamers are separated by over 8 W.mol- l. On the other hand, the changes in optical rotation for methyl 2,3-dideoxy-c~-~-glycero-pentopyranoside and its 4,6-ethylidene acetal indicated the presence of the (+sc, up) conformer for the axial isomer.6 Orientationsabout the exocyclic, C- 0 bond can be assessed by measuring the vicinal 13C-H coupling constant between the anomeric proton and the a-carbon atom of the aglycon (R) group, provided that the angular dependence is known. The vicinal, 13C-H coupling constants for the C - 0- C - H moiety show a dependence on torsional angle analogous to that for the familiar Karplus equation, and several, fairly complete, Karplus-type curves are available.16~85-87Observed values, however, reflect a thermodynamically averaged conformation that does not usually correspond to a physically real one, and separation of the contributions from different conformers cannot be achieved without making questionable assumptions. Therefore, the interpretations based on these measurements are very qualitative, and may require revision. Nevertheless, they support the conclusion that (+sc, sc) and (up, -sc) are the favored conformations of the axial and equatorial forms, respectively.*6 Additional evidence on the selection of conformations by the exo-anomeric effect is derived from the solid-state structures of It was earlier observed that the actual orientation of the anomeric alkoxyl group in pyranosides in the solid state corresponds to the (+sc, +sc) or (up, -sc) conformer, and thus proved that these conformersrespectively represent the most stable axial and equatorial forms. As already noted, a particularly clear illustration of the operation of the exo-anomeric effect comes from the nonreducing disaccharide a,a-trehalose, in which the most stable orientation about both exocyclic, C - 0 bonds corresponds to the (+sc, +sc) conformer. Analyses of carbohydrate ~ t r u c t u r e srevealed ~ ~ . ~ ~ regularities in the distribution of the torsional angle 0 that are consistent with a restriction of rotation about the exocyclic C - 0 bond. The torsional angle for equatorial isomers varies from -50 to - 1 10 with a mean value of -79.4 '. For the axial isomers, the range is 30 - 130°, with a mean value of 84.5 (see Ref. 29). (+sc,

+

+

O

O,

O

(85) G. K. Hamer, F. Balm, N. Cyr, and A. S. Perlin, Can. J. Chem., 56 (1978) 3109-31 16. (86) H. Thagersen, Ph.D. Thesis, The Technical University of Denmark, Lyngby, 1977. (87) I. TvaroSka, M. Hricovini, and E. Petrikovi, Carbohydr. Res.. 189 (1989) 359-362.

ANOMERIC AND EXQANOMERIC EFFECTS

71

A subsequent, elaborate survey of 1 1 1 carbohydratederivativesmconfirmed that the axial glycosides occur only in the conformation corresponding to (+sc, sc), but the equatorial glycosides show a 3 : 1 distribution in the ratio of (up, -sc) to (up, up) conformers. The (up, up) conformer had eluded scrutiny in previous analyses, and the presence of the (+sc, up), (+sc, -sc), and (up, +sc) conformations has not been observed. Currently, alkyloxy and alkylthiosubstituentsare mainly used as flexible, polar aglycons on the oxane ring. The observation,88however, that the azido group in the crystal structure of tri-0-acetyl-a-D-arabinopyranosyl a i d e is oriented towards the ring-oxygen atom, with a torsion angle of 76 ’,indicates the general characterof the exo-anomeric effect. Further experimentaleffort is needed in order more fully to determine the influence of solvent and pyranose-ring substituents on the exo-anomeric equilibrium.

+

2. Valence Geometry Parameters

The structural aspects of the anomeric effect manifested in the conformational variation of the valence geometry parametersin hemiacetal and acetal moieties in pyranoses and pyranosidesare receivingincreased attention. The shortening of the anomeric C -0 bond relative to its “standard” value was observed, and confirmed to be experimentally significant, some years This shortening is characteristic of any CX, (X = electronegative atom) grouping,w and is apparent, for example, in the structure of fluorometha n e ~ ,where ~ ’ the carbon- fluorine bond-length decreases from 138.5 pm in CH,F through 135.8 pm in CH,F,, and 132.6 pm in CHF,, to 13 I .7 pm in CF, . In pyranose compounds, the anomeric carbon atom parallels the central atom, and the two adjacent electronegative atoms are the ring-oxygen atom and the first atom of the aglycon group. Detailed examination of the available molecular geometry data of carbohydrates suggested that there are characteristic patterns of bond lengths and bond angles associated with particular conformation^.^^-^^^^-^^ These (88) (89) (90) (91) (92) (93)

(94) (95) (96) (97)

P. Luger and H. Paulsen, Chem. Ber.. 107 (1974) 1579- 1589. H. M. Berman, S. S. C. Chu, and G. A. Jeffrey, Science, 157 (1967) 1576- 1577. H. A. Bent, Chem. Rev.,68 (1968) 587-648. D. R. Lide, J. Am. Chem. SOC.,74 (1952) 3548-3552. M. Sundaralingam, Biopolymers, 6 ( 1968) 189 -2 13. G. A. Jeffrey, in L. E. Sutton and M. R. Truter (Us.), Molecular Structure by Diffraction Methods, Chemical Society, Special Periodical Report, Vol. 6, London, 1978, pp. 183221. G. A. Jeffrey, J. A. Pople, J. S. Binkley, and S. Vishveshwara, J. Am. Chem. Soc., 100 (1978) 373-379. G. A. Jeffrey and J. H. Yates. J. Am. Chem. SOC.,101 (1979) 820-825. G. A. Jeffrey and J. H. Yates, Carbohydr. Rex, 96 (1981) 205-213. F. Longchambon, Ph.D. Thesis, University of Paris-Nord, Bobigny, France, 1984.

IGOR

72

TVAROSKAAND TOMAS BLEHA

TABLE VIII Mean, Hemiacetal and Acetal Geometriesoin a-and /h-Pyraaoses, Methyl a-and BD-Pyranosides, and Oliosaccharidesw

Pyranos

Methyl pyranosides

Oligosaccharides

Total

Parameter

a

B

a

B

a

B

a

B

Number of structures r(C-5 -0-5)

22 144.0 142.5 139.8

12 143.4 142.7 139.2 112.1 107.1

13 143.3 142.9 138.1 143.0 112.4 107.6 113.7 -73.6

5

114.0 111.4

10 143.3 141.4 139.9 142.4 113.3 112.4 113.5 65.0

144.0 142.0 140.6 143.2 113.9 11 1.9 114.0 73.7

22 143.7 142.0 139.4 143.9 112.4 107.6 115.7 -73.5

37 143.8 142.2 140.0 142.6 113.8 111.8 113.6 69.1

47 143.5 142.4 139.0 143.6 112.3 107.5 115.0 -72.2

r(0-5 -C- 1) r(c-1-0-1) r(0-1-c-1) (C-5 - 0 - 5 4 - 1 ) (0-5 -C- 1 -0-1) (C-1-0-1 -c-1) (0-5-C-1-0-1 -C-1)

69.5

-69.3

Bond lengths in pm; bond angles and torsional angles in degrees.

patterns constitute a convincing manifestation of the anomeric effect, and can be readily discerned from the results of statistical treatment9' of pyranoses, methyl pyranosides, and oligosaccharide structures summarized in Table VIII. The most obvious feature of the experimental data on the a-anda-linkage in carbohydrates is a marked difference in the molecular geometry between the two configurations both in bond lengths and bond angles (see Table 143.8

H-1

c -2

0-5

113.6O 143 5

I

142.4

0-5

' 112 30 10750

FIG.13.-The

115.0°

H-1

Mean Valence Geometry Parameters for Aldopyranosides (from Ref. 97).

ANOMERIC AND EXO-ANOMERIC EFFECTS 145.0

142.8

109.90

73

109.80 112.60

FIG. 14.-The Valence Geometry Parameters in Two Forms of D-Xylopyranosyl Fluoride Derivatives (from Ref. 98).

VIII). The ring 0 - 5 -C- 1 and anomeric C-1-0-1 bond lengths inp-D glycosides differ appreciably. The length of the 0-5 -C- l bond approaches the standard value of 142.5 pm, and the C-1-0-1 bond is much shorter ( 139.0 pm). The two external bond-lengths are longer than normal, and are almost equal. The bond angle at the anomeric carbon atom is 107.5 that is, less than tetrahedral, and of the two angles on the oxygen atom, the glycosidic C- l - 0-l -C angle is much the larger, l l 5 .O In the a-D-glycosides, the anomeric bond is still the shortest C - 0 bond, but the difference is 1 pm less. There is no difference between the two C - 0 - C bond angles, and the bond angle at the anomeric carbon atom is 4" greater than in p-~-glycosides (see Fig. 13). The data on methyl pyranosides show similar features. In pyranoses, in contrast, the only significant differences observed between the a-and p-D anomers are in the bond angles at the ring-oxygen atom and the anomeric carbon atom, amounting to 1 14.0 and 1 1 1.4",respectively, in the a anomers, whereas they are smaller in the panomers, namely, 112.1and 107.1'. The data observed for oligosaccharides are less precise, because of the greater complexity of their structures,but they show a pattern similar to that of the methyl glycosides. Several pyranosyl halides have been studied as acetylated or benzoylated derivatives. In both the fluorides and chlorides, the equatorial carbonhalogen bonds are shorter than the axial bonds, with the data for fluorides9* illustrated in Fig. 14. The C- F bond-lengths in both derivativesofj?-D-xylopyranosyl fluoride are significantly shorter than the 143.2 pm found for the non-anomeric C - F bond in 1,3,4-tri-~-acetyl-2-deoxy-2-fluoro-~-xylop y r a n ~ s eThere . ~ ~ is also a variation of 4.4 pm in the C- 1 - 0-5 bond-lengths. Some 1 11 carbohydrate derivatives have been statistically treated,30and coupling of the C - 0 bond-lengths and C - 0- C bond angles to the orientation about the exocyclic C-1-0-1 bond (exo-anomeric effect) was demonstrated. These results are given in Table IX. The differences in bond angles and bond lengths show a small but significant variation with the torsion O ,

O .

-

-

(98) G. Kothe, P. Luger, and H. Paulsen, Actu Crystallogr., Sect. B, 35 (1979) 2079-2087. (99) G. Kothe, P. Luger, and H. Paulsen, Actu Crystallogr., Sect. B, 32 (1976) 2710-2714.

TABLE IX Mean C - 0 - C Bond Angles (Degrees) and C - 0 Bond Lengths (pm) in Carbohydrate Structures as a Function of the Torsion Angle (@) About the C - 1 - 0 4 (Anomeric) Bond" Isomer

Equatorial Axial

Number of items

@

12 13 8 17 16 7

-65 to -75 -75t0-85 65-95 55-65 65-75 75-85

Rotamer -Sc)

(UP, +SC)

(+sc, +SC)

r(C-1-0-1)

r(C-1-0-5)

(C-1-0-1 -C-i)

138.38 137.84 139.14 140.42 140.31 140.31

141.93 142.76 142.68 141.54 141.34 141.84

1 14.4 114.3 113.7 113.5 113.7 112.4

(C-1-0-5-C-5)

From Ref. 30, based on statistical treatment of I 1 I carbohydrate derivatives from the Cambridge Structural Database.

111.6 1 1 1.9 111.8 113.7 113.5 113.2

ANOMERIC AND EXO-ANOMERIC EFFECTS

75

+

angle @. In both equatorial groups, the -sc and sc torsional minima of Q, are located at angles larger than -60 and 60 respectively.

+

O ,

IV. MOLECULAR ORBITAL CALCULATIONS OF THE ANOMERIC EFFECT The description and understanding of the nature of stereoelectronic effects is an appropriate field for the application of organic quantum chemistry. Molecular orbital (MO) methods* can describe the electron distribution in molecules, and the changes in internal rotation. In principle, they give the total potential energy of individual conformers completely, without the necessity to correct for various “effects.” Quantum chemical calculations offer a deeper insight into the orbital interactions in the molecule, and reveal the factors responsible for the stabilization of any conformation. The best description of the stereochemical behavior of an isolated molecule is achieved by nonempirical ab initio calculationswith the sufficiently extended basis of the atomic orbitals, for example, 6-31G. However, an investigation of the conformationalproperties of a molecule having only two torsional angles, such as 0 and Q,, resulting in the energy map E(0, CP), representsthe multiple (even one hundredfold)repetition of a routine calculation of the energy. Therefore, in current practice, study is mostly confined to the less time-consuming methods either of ab initio methods with a minimal basis of orbitals (STO-3G), or to the semiempirical MO methods (PCILO, CND0/2, and MNDO). In both cases, a cautious approach is necessary, entailing careful comparison of computed properties for a given group of molecules with experimental data. It should be pointed out that there exists an inclination to consider ab initio results, even with the restricted or minimal basis set, as automatically superior to those of any semiempirical MO method. Calculations of a conformational energy for molecules exhibiting the anomeric effect (see later) give several examples of the deceptiveness of this claim. An optimal choice of the quantum chemical method does not solve all of the problem, however. The isolated molecule calculations can be performed by the complete or partial optimization of the molecular geometry, or by assuming fixed bond-lengthsand valence angles, with torsional angles as the only degrees of freedom. Although the optimization considerably extends the computing time, it is often unavoidable, owing to the relatively large

* MO = molecular orbital; STO-3G = the nonempirical (ub initio) method with the minimal basis using three Gaussian functions for one Slater atomic orbital: 4-3 IG, 6-3 IG, and so forth are the nonempirical methods with the extended basis, using Gaussian functions split into two groups; the semiempirical MO methods, CNDO/2 and MNDO (Complete Neglect, and Modified Neglect, of Differential Overlap, respectively); and PCILO, Perturbative Configurational Interaction using Localized Orbitals.

76

IGOR

TVAROSKAAND TOMAS BLEHA

changes of structural parameters in glycosides that are attributable to internal Finally, it is important to recognizethat energy calculations refer to isolated molecules, whereas conformational equilibria are generally measured in the liquid phase, where intermolecular interactions and solvent effects can be substantial. Therefore, the calculations should be supplemented by a procedure accounting for the influence of environment, before comparingwith experimentaldata, especiallythose for aqueous solutions.

180

60 60

-60

180

0

(degrees)

FIG. 15.-The Conformational Energy Map of Dimethoxymethane,looWith Energy Contour in kJ.mol.-l. Two Conformationsof DimethoxymethaneCorrespondingto a and e forms of 2-Methoxyoxane Are Also Shown.

ANOMERIC AND EXO-ANOMERIC EFFECTS

77

1. Conformational Energies Owing to the complexity of the internal motion of carbohydrate molecules, the elucidation of their conformational properties by MO calculations requires a lessening of dimensionality to manageable proportions. Several small acyclic molecules have therefore been used as models for ab initio or semiempirical MO studies on the structural segments of carbohydrates. On the whole, calculations reproduce all of the main structural trends and conformational preferences observed experimentally in the crystal structures of carbohydrates and in solution.

a. Acyclic Model Compounds.- The anomeric effect has been studied in several simple acyclic molecules having the general formula YCH2X,where Y is OH, OCH, ,SH, or SCH, ,and X is an electronegative group, such as C1, F, 3 H , OCH, , SCH, , NH2, and NHf ,by both ab initio and semiempirical methods. Special attention has been devoted to the - 0 - C - 0 - segment, and methanediol, methoxymethanol, and dimethoxymethane were used for the modelling of the acetal and hemiacetal moieties of carbohydrate molecules.11,58,94,100-108

A complete description of the energy of these molecules as a function of one, or two, torsion angles 0 and Q, characterizing rotations about the C - 0 or C - S bonds is given by a one- or two-dimensional torsional potential. As an example, the CND0/2 calculated potential energy surface E(0, @) for dimethoxymethanelWis shown in Fig. 15. The relevant conformations of dimethoxymethane with torsional angles fixed at 60 and 180”,corresponding to the axial and equatorial isomers of an oxane acetal, are also illustrated. The possibility of internal rotation about two C - 0 bonds is responsible for the “double” presence of the anomeric effect, resulting in the stabilization of two conformations with the methyl groups on opposite sides of the -0-C0- plane, (+sc, sc) and (- sc, - sc). The forms (- sc, sc) and (+sc, - sc), havingadjacent methyl groups on the same side ofthe - 0 - C - 0 - plane are energetically unfavorable, owing to the 1,3-diaxialtype of steric interactions.

+

+

(100) I. TvaroSka and T. Bleha, J. Mol. Struct., 24 (1975) 249-259. (101) G. A. Jeffrey, J. A. Pople, and L. Radom, Carbohydr. Rex, 25 (1972) 117- 131. (102) G. A. Jeffrey, J. A. Pople, and L. Radom, Carbohydr. Rex, 38 (1974) 81-95. (103) S. Vishveshwara, Chem. Phys. Lett., 59 (1978) 30-32. (104) L. Radom, W. J. Hehre, and J. A. Pople, J. Am. Chem. Soc., 93 (1971) 289-300. (105) D. G. Gorenstein and D. Kar, J. Am. Chem. Soc., 99 (1977) 627-677. (106) C. van Alsenoy, L. Schafer, J. N. Scarsdale, and J. 0. Williams, J. Mol. Struct. Thee chem., 86(1981) 111-117. (107) P. Bonnet, D. Rinaldi. and J. P. Marchal, J. Chem. Phys., 70 (1974) 298-302. (108) I. TvaroSka and T. Bleha, Collect. Czech. Chem. Commun., 45 (1980) 1883- 1895.

IGOR TVAROSKA AND TOMAS BLEHA

78

TABLEX Calculated Relative Energies (kJ.moI-') of Stable Conformers of R'OCHzORz with Respect to the (sc, sc) Rotamer Compared with Experimental Values for Dimethoxymetbane ~

R1 H

Rz H

____~

(sc, up)

(up, up)

References

STO-3G

11.7 9.4 10.6" 19.7 18.4" 19.8 18.8 15.7 12.6 - 1.0 6.6 19.P 10.8 10.00 13.3" 4.6 4.9 3.8 5.0 5.2 7.1 5.0 8.1' 6.2d 6.3c

28.3 27.6 26.9 46.9 46.1 49.2 44.8 43.4 43.4 9.7 13.9 22.2 32.2 31.1

58 103 103 104 101 94 103 94 102 105 105 106 106 106 I06

4-31G

H CH,

CH, CH,

~~

Method

6-31G 4-31G STO-3G 4-21G

CNDO/2

PCILO MNDO Exp.

15.8 14.9 10.9 10.9 17.2 14.2 18.8 12.4 13.0

LOO 105 107 108 I1 64 109 107 107 107

a Optimization of bond lengthsor bond angles. Complete optimization of geometry. 'Gas phase. Liquid phase. CSolution in 1 : 1 dimethoxymethane- heptane.

Furthermore, there are four minima on the conformational surface representing the conformations of the type (up, sc), and one minimum corresponding to the (up, up)conformation.The map calculated for methanediol is similar, but, in contrast to that for dimethoxymethane, the (up, up) conformer of methanediol is found to be a local maximum.1oLA potential surface has also been reportedlo2for methoxymethanol, a structural intermediate between dimethoxymethane and methanediol, with lessened symmetry due to the presence of two different rotors. For example, conformations (up, sc) and (sc, up)can be distinguished in this case, and both are only doubly degenerate. A comparison of relative energies of dimethoxymethane, methanediol,

ANOMERIC AND EXO-ANOMERIC EFFECTS

79

TABLEXI Calculated Relative Energies (W.mol-l) of Stable Conformers of ROCH,X with Respect to (sc) or (sc,sc) Orientations

X

R

Method

AEl

STO-3G

12.7 13.1 26.0 23.4 26.8" 14.9b 25.5 18.0 18.8' 2 I .4" 21.8" 10.0 18.6 20. I 22.6 21.3 5.9 15.5" 12.1" 7.7 6.3-8.4

AEz

References ~

H

F

4-3 1G

CH,

F

4-21G 6-31G 4-31G

H

C1

CNDO/2 STO-3G

CH,

CH, H

CH, CH,

-

C1

CH, NH,

NH, NH;

4-31G 6-31G 4-31G CNDO/2 exp. CNDO/2 STO-3G 4-21G 4-31G CNDO/2 CNDO/2

-1.1

-5.7b -2.6 -5.5"

58 95 95 111 111

112 95 95 95 95 95 11 58

-2.1 13.8 -1.0 -8.8 -9.1 -3.1 - 3.8

95 95 95 95 95 95 I13 63 I1 114 112 104

114 I1 11

* Optimization of bond lengths or bond angles. Complete optimization of geometry.

and methoxymethanol conformationscalculated by different methods with experimental values is shown in Table X.It is seen that conformationsof the (sc, sc) type of dimethoxymethane are predicted as preferred by both ab initio and semiempirical MO calculations, and the calculated energies are generally consistent with experimental evidence about the stability of conformers. On the basis of electron diffraction studies,33dipole moment, and Kerr constant measurements,64J09J lo it was established that (sc, sc) with both (109) M. Sakakibara, Y. Yonemura, H. Matsuura, and H. Murata, J. MoI. Sfrucf.,66 (1980) 333-337. (1 10) R. J. W. Le Fevre, A. Sundaram, and R. K. Pierens, J. Chem. Soc., (1963) 479-488.

80

IGOR TVAROSKAAND TOMAS BLEHA

angles identical, 66.3",is the most stable conformer. The data in Table X suggest that the anomeric stabilization by the two consecutive rotations about the C - 0 bonds is coupled, and not additive. The first rotation, from (up, up) to (up, sc) in dimethoxymethane, apparently leads to greater stabilization than the second rotation, from (up, sc) to (sc, sc). The results of the energy calculations for the remaining ROCH,-X m0~ecU~eSll,58,95,103,104,111-114 are presented in Table XI. The relative energies are given by the difference between the (sc) and (up) rotamers for X = F, C1, and by the differences between the (sc, sc) and (sc, up) conformers (AEl),and between the (sc, sc) and (up, up) conformers (AE,) for X = CH,, NH,, and NHf .With the exception of the NH2group, the preference for the sc orientation is confirmed by the calculations, and it decreases in the order F > OH > C1> OCH, for the 4-3 1G basis set. The energy differencescalculated by the semiempirical methods for substituted-dimethyl ethers CH,OCH, X are lower than those estimated from the ub initio calculations on HOCH2- X molecules. In methyl and amino derivatives, the preference for the up orientation increases in the order NH; > CH, > NH2. All of these results are consistent with the experimentally observed anomeric preferences in substituted oxane and in pyranoses (see Tables I - VI). A few MO calculations have been reported for the acyclic molecules RSCH2X (see Table XII), where R is H or CH,, and X is F, C1, OH, SH, OCH,, and SCH,, used as the models of the thioacetal moiety in thio sugar^.^^^' 15-' l9 The calculated potential surfaces for HSCH2SH, CH,SCH,SCH, ,HSCH,OH, and CH3SCH20CH,'149' 16r1 l9 are, in their gross features, similar to those previously obtained for methanediol and dimethoxymethane. Table XI1 shows differences in energies of RSCH,X conformers. In the case of the thioacetal segment 0- C- s, the lowest energy is found for the (sc, sc) conformer. Energies of(sc, up) and (up, up) conformers, relative to (sc, sc), are lower than corresponding values in oxygen analogs. The modelling of the dithioacetal moiety is more complicated, and the results obtained are contradictory. It can be deduced from the data, however, that the preference for the sc orientation is lessened in the rotation about the ( 1 1 1 ) L. Radom, W. J. Hehre and J. A. Pople, J. Am. Chem. Soc., 94 (1972) 2371-2381. ( I 12) L. Schafer, C. van Alsenoy, J. 0. Williams, and J. N. Scarsdale, J. Mol. Struct. Theochem., 76 (1981) 349-361. (1 13) I. TvaroSka and T. Bleha, Tetrahedron Lett., (1975) 249-252. (1 14) P. Kaliannan, S. Vishveshwara,and V. S. R. Rao, J. Mol. Struct. Theochem., 105 (1983) 359-374. (1 15) S. Vishveshwara and V. S. R. Rao, Curbohydr. Rex, 104 (1982) 2 1- 32. ( I 16) M. Ohsaku and H. Murata, J. Mol. Struct. Theochem., 85 (1981) 125- 131. ( I 17) I. TvaroSka, Chem. Zvesti, 38 (1984) 189- 197. ( 1 18) L. Nerskov-Lauritsen, F. S. Jerrgensen, and J. W. Jaroszewski, Curbohydr. Rex, 123 (1983) 1-11. ( I 19) I. TvaroSka, Collect. Czech. Chem. Commun., 49 (1984) 345-354.

ANOMERIC AND EXO-ANOMERIC EFFECTS

81

TABLEXI1 Calculated Relative Energies (kJ.mol-') with Respect to (sc)or (sc, sc) Orientations for the Stable Conformers of RSCH,X Molecules

H

X

Method

(sc,up)

(up,sc)

References

(up,up) ~~

H H H

F C1 SH

CH, CH,

SH SCH,

H

OH

CH,

OCH,

STO3G STO-3G STO-3G

4-31G CNDO/2 CNDO/2 STO-3G PClLO MNDO STO-3G

PClLO MNDO

11.0 11.1

0.8 4.3' 0.5" 9.0 - 1.3 - 3.7 4.2 - 3.4 5.0" 2.8 4.9" 2.06 1.7 9.6

5.5

4.1 9.2 3.5 19.0 5.0 11.3 7.4 - 3.0 11.1

14.1 16.6 13.9 5.5 17.5

58 58 115 115 115

115 116 I16 115 117 117,118 115 I15 115

119 119

a Complete optimization of geometry.. Optimization of bond lengths or bond angles.

C - S bond in comparison with that for the C- 0 bond. Experimental studies for these molecules120-123 gave the (sc, sc) conformation as the most stable in the solid state. The latter conformation also dominates in the liquid state, where additional conformations, (sc, up), (up, sc), and (up, up), are also present. A lessened stabilization of the sc with respect to the up orientation on rotation about the C-S bond in comparison with the C - 0 bond is supported by the dipole moment and measurements of the Kerr constant of dithioacetals.124~125These results suggest, however, that the preferred conformation of some dithioacetal derivatives in CC14 is the (sc, up). In summary, it appears that MO methods successfully predict the energy of the conformers in substituted ethers and acetals. At the same time, it is interestingthat the agreement of calculated results with the available experimental data (see Table X) seems to be better for the selected semiempirical methods than for some ub initio calculations,apparently owing to fortituous compensations. (120) (121) (122) (123) (124) ( 1 25)

M. Ohsaku, Y. Shiro, and H. Murata, Bull. Chem. SOC.Jpn., 45 (1972) 113- 121. M. Ohsaku, Bull. Chem. Soc. Jpn., 47 (1974) 965-975. H. Matsuura, K. Kimura, and H. Murata, J. Mol. Struct., 64 (1980) 281 -284. H. Matsuura, H. Murata, and M. Sakakibara, J. Mol. Sfruct.. 96 (1983) 267-275. 0.Exner, V. JehliEka, and J. Firil, Collect. Czech. Chem. Commun., 37 (1972) 466-477. A. N. Vereshchagin and 0. Exner, Collect. Czech. Chem. Commun.. 38 (1973) 690-696.

IGOR TVAROSKA AND TOMAS BLEHA

82

b. Fourier Component Analysis of Torsional Potentials.- In order to facilitate the interpretation of conformational equilibria for rotation about the C - 0 bond in compounds having the general formula CH30CH2X,it is useful to expand the potential function as a truncated Fourier expansion.111

v(e)= 0.5

+

vy(i - cos 8) 0.5 v:(i - cos 28) +0.5 vg(i - cos 381,

(5)

or v(e) = v,(e)

+ v,(e) + v3(e).

(4)

For such an asymmetric structure as dimethoxymethane with one torsion angle fixed at 60 additional sine terms are necessary, in order to account for the lack of symmetry about 8 = 180" (see Ref. 94). The individual components V,(8), V,(8), and V3(8) of the total potential function V(8) can be identified with specific physical effects of similar periodicity. For example, the onefold term, V,(8) = 0.5 Vy( 1 - cos O), moves from a maximum value to a minimum value as 8 changes by 180".The same variation with torsion showsdipolar or steric interactions. The twofold term, V2(8),changes from a maximum to a minimum as 8 changes by 90". This periodicity frequently corresponds to the change of delocalization interactions. Finally, the threefold term, V3(8),moves from a maximum value to a minimum value as it changes by 60".This is generally attributed to the intrinsic torsion potential. The onefold, twofold, and threefold components all contribute to the location of the resultant maxima and minima at the torsional potential, as illustrated in Fig. 16 for dimethoxymethaneand chloromethoxymethane.113 Values of Vg for both molecules are negative, suggesting a preference for the staggered XCOC up and sc conformations.The positive V(:indicates a preference for the synperiplanar (sp, 0 = 0") conformation over the up conformation. This is consistent with a simple dipole-dipole argument which favors the sp conformer, with the opposed dipoles of segmentsC - 0- C and C- X, over the up conformer having parallel dipoles (see Fig. 17). In chloromethoxymethane, the Vy term is lowered, owing to steric interactions between the methyl group and the chlorine atom in the sp position. The V: term, associated with delocalization interactions, is negative for chloromethoxymethane, favoring orthogonal conformation (8 = 90') over up and sp conformations, and is small and positive for dimethoxymethane. The Fourier decomposition of the potential energy reveals that the dominance of the V, term in dimethoxymethaneand that of the V, term in chloromethoxymethane is responsible for the preference for the sc conformation. The torsional potential constants of Eq. 5 for substituted dimethyl ethers CH,0CH2X, determined from the potential energy calculated by the semiempirica111J'7and ub initio MO method^,^^.^'^ are summarized in Table O ,

ANOMERIC AND EXO-ANOMERIC EFFECTS

83

4 V p = 14.7

= v; v;=-,.s

1.5

0

-4

-8

Vt= - 6 . 6

-12 I I

V,"-

I

120"

60"

(4

I

I

3"

3

Torsion angle

120"

60"

(0)

- 6.4

(b)

FIG.16.-Torsional Potential (V) in (a) Dimethoxymethane and (b) Chloromethoxymethane, and Their Decomposition by Fourier Expansion, Eq. 5 (from Ref. 113).

XIII. A qualitative rationalization of the conformational preference in CH,OCH,X, where X = F, SCH,, CH, , NH,, and NHf ,can be advanced along the same lines as for dimethoxymethane and chloromethoxymethane, with the contributions of the dipolar, delocalization,and intrinsic torsional terms. c. Calculations on Cyclic Model Compounds.-In most cases, the MO calculations on simple acyclic molecules correctly describe the preference for the sc over the up orientation. Nevertheless, the acyclic models have, besides a lack of experimental data, several evident shortcomings in order fully to represent the conformational behavior of the cyclic carbohydrate structures. To clarify this behavior further, calculations have been carried

I

180"

IGOR TVAROSKAAND TOMAS BLEHA

84

f H3C---

HI'

. . %

H

H

CH3

FIG. 17.- Dipole- Dipole Interactions in Dimethoxymethane and in Chloromethoxymethane.

out on 2-substituted oxane12,46J26-131 derivatives. In Figs. 18 and 19, the PCILO potential of rotation about the exocyclic C- 0 bond for two forms of 2 - m e t h o ~ y o x a n e ~ (MTHP) ~ ~ - ~ * ~ is compared with corresponding energies for dimethoxymethane calculated by the ub initkP4 and CND0/2 methods.loo Figs. 18 and 19 display the existence of five minima for chair conformers of 2-methoxyoxane. For the u form, the lowest minimum appeared at 63", and the next one, having -6.2 kJ.mo1-I higher energy, at 152.2". These minima correspond to the (+sc, sc) and (+sc, ap) conformers shown in Fig. 3. In the region ofthe third staggered position (+sc, -sc), there is a broad maximum in Fig. 18. For the e form, Fig. 19 shows three minima; the lowest energy is found for the (up, -sc) conformer which lies 3.1 kJ.mol-' higher than the (+sc, sc) conformer of the u form of 2-methoxyoxane. Conformations (up, +sc) and (up, ap) are disfavored by 5.7 and 6.8 kJ.mol-', respectively, relative to (ub, -sc). Based on the energies of the individual conformers, the equilibrium ratios 70.8 :6.0 : 19.9 :2.0 : 1.3 have been calculatedL2'for the distribution of the (+sc, sc), (+sc, up), (up,-sc), (up, sc), and (up, up) conformers, respectively. The calculated u to e ratio of

+

+

+

(126) (127) (128) (129) (130) (131)

+

T. Koiarand I. TvaroSka, Theor. Chim. Actu, 53 (1979) 9-19. I. TvaroSka and T. Koiir, J. A m . Chem. Soc., 102 (1980) 6929-6936. I. TvaroSka and T. Koiar, Curbohydr. Rex, 90 (1981) 173- 185. I. TvaroSka and T. Koiar, Znt. J. Quantum Chem., 23 (1983) 765-778. I. TvaroSka and T. KoiBr, J. Mol. Srruct. Theochem., 123 (1985) 141 - 154. L. Guibe, J. Augk, S. David, and 0.Eisenstein, J. Chem. Phys., 58 (1973) 5579-5583.

ANOMERIC AND EXO-ANOMERIC EFFECTS

85

TABLE XI11 Calculated Rotational Potential Constants (kJ.mol-') Describing Internal Rotation About the C - 0 Bond' in Substituted Dimethyl Ethers, XCH,OCH, X

Method

q

q

OCH,

CNDO/2 4-31G CNDO/2 4-31G CNDO/2 4-3 IG XILO CND0/2 CNDO/2 CNDO/2

14.7 20.0 13.3 6.7 3.6 38.8 4.0 -9.1 -6.2 10.0

1.5 -20.0 0.0 -17.4 -6.6 -46.7 -9.3 6.3 0.1 5.0

F C1 SCH, NH, CH, NH:

V!

References

-3.8

113 103

- 10.0 -4.4 -16.3 -6.4 -28.7 -7.1 -2.7 -3.8 -6.2

11

103 11

103 117 11 11

I1

a In OCH, and SCH, derivatives, the methyl group is in the ap position relative to the C - 0 bond.

76.8 :23.2% is in agreement with 77-83% ofthe a form of 2-methoxyoxane derivatives measured in nonpolar solvents (see Tables I1 and 111). The preponderance of conformations having the methyl group in the sc position relative to the ring-oxygen atom is a clear demonstration of the working of the exo-anomeric effect. This preference is not limited to the chair forms, but has also been found from computation in all pyranose ring forms of 2-methoxyo~ane.~~~,~~~ The MO calculations of the conformational equilibria of 2-chloro-oxane and 2-fluoro-oxane have been camed The ab initio STO-3G calculations for 2 - c h l o r o - o ~ a n econfirmed ~~~ the stabilization of the a position of the C1 atom relative to the e one. An energy difference of 5.0 kJ.mol-' is found if the C - C1 bond-length 177 pm is assumed for both conformers, and 15.5 kJ.mol-' ifthe axial C-Cl bond is lengthened to 182 pm. These results are in qualitative agreement with the energy difference of 9 kJ.mol-l observed in the pure The axial preference of both the 2-fluoro- and 2-chloro-oxane is also corroborated by the results ofthe PCILO and MNDO methods.I2For the fluoro derivative, both methods predict an energy difference of 6.7 kJ.mol-'. For the chloro derivative, the PCILO and MNDO methods give energies of 10.1 kJ.mo1-' and 13.2 kJ.mol-', respectively. A PCILO calculationMfor 2-carbamoyloxane showed that the conformation having the CONH, group in the e position is preferred over the a position. The calculated energy difference of 7 kJ.mol-' indicates a strong, reverse anomeric effect.

-

IGOR TVAROSKAAND TOMAS BLEHA

86

I

E

7

I I I I I I I

Y

u

W 9

30.C

20.1

I I

I I

\

I I

6 'CH3

I I I I

I

I I

I I I I

I I

Ii

a

1O.L

@ (degrees) FIG.18.-Torsional Potential of the Exocyclic C - 0 Bond in the Axial Form of 2-Methoxyoxane Calculatedby the K I L O Method (Full Line) and CorrespondingPotential in Dimethoxymethane (Dashed Line) Calculated by the CNDO/2 Method (Curve a)lw and by the ab initio Method (Curve b).%

ANOMERIC AND EXOANOMERIC EFFECTS

I

I

60

120

I

180

I

240

81

I

300

@

(degrees) FIG.I9.-The Same as in Fig. 18, but for the Equatorial Form of 2-Methoxyoxane.

88

IGOR

TVAROSKAAND TOMAS BLEHA

As already noted, a,&-trehaloseassumes, in the solid state2’and in solut i ~ nthe , ~sc~conformations of the two glycosidic C - 0 bonds. The conformational properties of the three trehaloseswere studied132 by using 2-(oxan2-y1oxy)oxane as a model in three isomeric forms (a, a), (e, e),and (a, e). In these compounds, the anomeric and exo-anomeric effects influence the properties of four C - 0 bonds in the C-5 - 0 - 5 -C- 1 - 0 - 1 -C-1’-0-5’-C-5’ moiety. The energy of the rotation about the C-1-0-1 or 0 - 1 -C-1’ bond depends on the orientation of the adjacent C - 0 (torsional angles and a’) bonds (8 and 0’ angles), and resembles the potential profiles calculated for the a and e forms of 2-methoxyoxane. The preferred form is the (a, a), the next is the (a, e) form, and the (e, e) form has the highest energy. The most stableconformerofall three forms is the conformerofthe a, a form, where all four C - 0 bonds are in the sc position.

+

d. Saccharides.- In a few cases, the anomeric equilibrium of monosaccharides has been treated by MO computation. The majority of calculation^^^^-^^^ on isolated D-glucopyranose incorrectly predict that the p anomer is the preferred form, and the energy differencebetween the anomers lies in the interval of 2 - 38 kJ.mol-’, according to the applied MO method. The PCILO ~alculated’~’ a :panomeric ratio for isolated D-glucopyranose of 76 :24 differs considerably from these values, and is close to the ratio of the a :e forms of 77 :23 for isolated 2-methoxyoxane (see earlier). Similar PCILO calculations on methyl a-and p-D-glucopyranosideestabl i ~ h e dthat l ~ ~the a anomer is the more stable, and the calculated energy difference is -4.2 kl.mol-’. A preference for the sc orientation about the C- 1 - 0 - 1 bond was suggested by calculations for both anomers. The ap arrangementsabout the anomeric C- 1 -0-1 bond in the a and Panomers are 4.0 and 4.6 kJ.mo1- respectively, higher in energy than the corresponding sc position.

’,

2. The Solvent Effect It is important to recognize that the energy calculations discussed refer to isolated molecules in the gas phase, whereas experimental values are mea(132) I. TvaroSka and i.Vaclavik, Carbohydr. Rex, 170 (1986) 137- 149. (1 33) Yu. A. Zhdanov, V. I. Minkin, Yu. A. Ostroumov, and G . N. Dorofeenko, Curbohydr. Rex, 7 (1968) 156- 160. (134) W. B. Neely,J. Med. Chem., 12(1969) 16-17. (135) N. Cyr, A. S. Perlin, and M. A. Whitehead, Can. J. Chem., 50 (1972) 814-820. ( 1 36) S. Melberg, K. Rasmussen, R. Scordamaglia, and C. Tosi, Curbohydr. Rex, 76 ( 1 979) 23-37. (137) I. TvaroSka and T. KoiBr, Theor. Chim. Actu, 70 (1986) 99- 114. (138) I. TvaroSka and T. KoiBr, Chem. Papers, 41 (1987) 501 -510.

ANOMERIC AND EXO-ANOMERIC EFFECTS

89

sured in solution where the effect of the medium can be substantial. In fact, significant differences in conformer population have been observed for oxane derivatives by n.m.r. in various solvents (see Section 111). As already mentioned, >77% of 2-methoxyoxane exists in the a form in nonpolar solvents, but only 52% in water (see Table VI). There exist several approaches for theoretical prediction of the effect of the solvent on c o n f ~ r m a t i o n . ' ~One ~ J ~possible procedure is based on inclusion of the solute and several solvent molecules in "supermolecule" quantum chemical calculations. Such an approach might be useful in providing information about optimal solvation sites, but the evaluation of the overall energy of solvation can hardly be obtained in this way, because of the necessity to include many solvent molecules and to perform a complete energy minimization. An alternative approach is to treat the solvent as a dielectric continuum. Computer simulations of a solution by the Monte Carlo or molecular dynamicmethods, explicitly including the carbohydratemolecule and many solvent molecules, could present the ultimate treatment of solvent effect. However, at present, such an approach to accurate evaluation of solvation energies is still far too expensive, owing to the complexity and high flexibility of carbohydrates. A quite reliable estimate of the solvent effect results from theories that combine micro- and macro-scopic parameters of solute and solvent. For example, in the solvophobic theory,14' the energy of a solute molecule in , is given as the sum of isolated molecule energy, E,, and the solvent, EWln solvation term, E, . The latter term encompasses the energy of cavity formation in the solvent to accommodate the solute, E,,, and the energy of subsequent solvent - solute interactions, Ei,. The interaction part is composed of the energy of dispersion, Edisp,and electrostatic, EeM,interactions. The final expression for E, can be written as E-1,

= Ei,

+ E, +

EeM

+ E w.

(7)

Calculations of the effect of the solvent upon the conformational properties of dimethoxymethaneLo8 based on Eq. 7 indicate that, in highly polar solvents, the ap orientation about the C - 0 bond might even be preferred. For example, the (sc, sc) conformation still prevails when parametersof CCl, are assumed in Eq. 7, with only the energy differencebetween (sc,sc)and (ap, sc) lowered by 2 kJ.mol-' in comparison to the isolated molecule. In water, ( 1 39) B. Pullman (Ed.), Environmental Efects on Molecular Structure and Properties.Jerusa-

lem Symp. Quantum Chem. Biochem., Vol. 8, Reidel, Dordrecht, 1976. (140) M. Berndt and J. S. Kwiatkowski, in G . Naray-Szabo (Ed.), Theoretical Chemistry of Biological Systems. Elsevier, Amsterdam, 1986, pp. 349 -422. (141) 0.Sinanoglu in B. Pullman (Ed.), Molecular Association in Biology, Academic Press, New York, 1976, pp. 427-445.

90

IGOR

TVAROSKA AND TOMAS BLEHA

however, the (up, up) conformation has been found as the minimum-energy structure from calculations. The analysis of individual terms of solvation energy reveal that electrostatic interactions constitute the dominant term in the solvation energy of dimethoxymethane. The lessening of the preference for the sc position with rising polarity of the medium is supported by the results of dielectric measurements of dimethoxymethane in the gaseous and liquid phases, combined, by CNDO/2 calculations.107It was found that the Gibbs energy difference of the (up, up) and (up, sc) conformations with respect to (sc, sc) gradually decreases in the succession: gaseous phase, 1: 1 (v/v) dimethoxymethane- heptane, and neat liquid (see Table X).Similarly, as in dimethoxymethane, the effect of polar solvents brings about stabilization of the up position in thio analogs of dimeth0xymethane.ll 7 * l I 9 The extent to which this kind of calculation is able to predict the effect of the solvent on conformational properties of carbohydrates has been thoroughly tested on 2-substituted oxane derivatives,12~127~-glucopyranose,~~~ and methyl a- and P-D-glucopyranoside.138 In the model applied,'27the cavity term in Eq. 7 is based on an expression taken from the Scaled Particle Theory,'42and the electrostatic term is calculated according to the reaction field theory.143The dispersion term takes into account both attractive and repulsive nonbonding interactions by using a combination of London dispersion energy and Born-type rep~1sion.l~~ The effect of the solvent on the abundance of the conformers of 2-methoxyoxane is demonstrated in Table XIV, where molar fractions of the axial form are compared with available experimental data already shown in Table V1.For a number of solvents, the agreement is remarkably good. Although the results indicate a decreased abundance of the a form of 2-methoxyoxane with increase in the dielectricconstant of the solvent, the dependence is not a simple one. The calculations also reproduce such subtle factors as the pronounced effect of chloroform when compared with other solvents of similar polarity and, conversely, a relatively weak effect of dimethyl sulfoxide in comparison to less-polar solvents. The analysis of the role of individual solvation energy terms in the total energy suggests that the conformationally most important term is the contribution of electrostatic interactions that stabilize the up conformations. Conversely,the dispersion term shows only a slight conformational dependence. The calculated abundance of the five conformers of 2-methoxyoxane quantitatively describes, for the first time, the exo-anomeric equilibrium about the C - 0 bond for the a and e forms in various solvents. At present, (142) R. A. Pierrotti, Chem. Rev., 76 (1976) 717-726. (143) R.J. Abraham and E. Bretschneider,in W. J. Onville-Thomas(Ed.),ZnternalRotation in Molecules, Academic Press, London, 1974, pp. 48 1 - 584.

TABLE XIV Calculated Molar Compositions (in %) of Conformers of 2-Methoxyoxane (at 298.2 K), and Comparison of the Calculated and Experimental Molar Compositions, x,, of the Axial Form in the Isolated State and in Solution ~~

Solvent

E"

Neat 2-methoxyoxane 1,4-Dioxane Carbon tetrachloride Benzene Carbon disulfide Chloroform Fluorobenzene Oxolane Octanol Pyridine Acetone EthanoI Methanol Acetonitrile Dimethyl sulfoxide Water

2.21 2.24 2.28 2.64 4.43 5.42 7.58 10.34 12.40 20.70 24.55 32.70 37.50 46.68 78.30

(+SC,

+sc)

70.8 67.4 69.7 69.6 68.8 63.4 66.5 64.4 67.7 61.5 62.6 60.7 55.9 58.7 61.2 35.8

(+sc, UP)

(+sc, -sc)

6.0 6.6 6.2 6.3 6.5 7.3 6.8 1.3 6.7 7.7 7.6 8.0 8.9 8.4 8.0 12.0

19.9 21.8 20.3 20.2 20.7 24.2 22.3 23.3 21.4 25.1 24.2 25.1 27.4 26.0 24.7 34.0

(UP,

-sc)

2.0 2.4 2.2 2.3 2.3 2.7 2.4 2.7 1.8 2.9 2.9 3.1 3.6 3.3 3.1 5.7

(UP, UP)

x,(calc.)

1.3 1.8 1.6 1.6 1.7 2.4 2.0 2.3 2.4 2.8 2.7 3.1 4.2 3.6 3.0 12.5

76.8 74.0 75.9 75.9 75.3 70.7 73.3 71.7 74.4 69.2 70.2 68.7 64.8 67. I 69.2 47.8

~~

a

6

= dielectric constant.

* From Refs. 6 and 8.

x.(exp.Y

82,83 82 80 71,78

72 69 65,68 74 52

IGOR

92

TVAROSKA AND TOMAS BLEHA

such detailed data on the conformer population are not available from experiments. The results given in Table XIV show that, for the u form, the abundance of the (+sc, ap) conformer in nonpolar solvents is <8%. It increases with increase in the solvent polarity, and reaches a maximum of 12%in water, which corresponds to 26% of the population of the u form. Similarly, the abundance of the (up, up) conformer in the e form does not exceed 3%in nonpolar solvents, and amounts to 12.5% in water. The latter figure means that, in water, the (up, up) conformation comprises 24% of the e form. 2-fluoro-o~ane,'~ and The preferences for the a form in 2-chloro-o~ane'~ D-glucopyranose,'37 calculated by the method mentioned127in various solvents, are presented in Table XV. In all of these compounds, the axial preference decreases with increase in the polarity of the solvent as well. In water, the e form is favored for 2-fluoro-oxane and D-glucopyranose. The values calculated for 2-chloro-oxane satisfactorily reproduce the observed axial preference of 93-96% in acetonitrile and acetone.38The anomeric ratio calculated for D-glucopyranose (seeTable XV) illustrates how comparison of the theoretical values referring to the isolated molecules with experimental data in aqueous solution might be misleading. Although, for the isolated molecule, the PCILO calculations predict 76%of the axial form, the correction for solvent brings the data in Table XV to very satisfactoryagreement with the composition in pyridine (45%of the a form), in dimethyl sulfoxide (44%), and in water (32 - 37%). Similar results of solvent polarity have been reported for methyl a- and &D-gluc~pyranoside.'~~ Whereas, for an isolated molecule, the energy difference between the anomers is 4.2 kJ.mol-', for a methanol solution, the calculation predicts a

-

-

-

TABLE XV The Calculated Preference of the (I Form of D-Glucopyranose, 2-Chlorooxane, and 2-Fluoraaxane in Various solvent^^^^^^ Solvent

Isolated molecule 1,4-Dioxane Carbon tetrachloride Chloroform Pyridine Acetone Methanol Acetonitrile Dimethyl sulfoxide Water

D-

Glucopyranose

2-Chlorooxane

2-Fluorooxane

76.0 68.5

98.3 97.8

94.0 90.6

71.4 62.5 48.9 49.9 37.3 41.9 45.7 32.2

98.0 96.9 96.3 96.5 94.1 95.1 95.9 78.4

91.7 87.5 85.6 86.1 77.6 81.0 84.1 37.0

ANOMERIC AND EXO-ANOMERIC EFFECTS

93

1.3 kJ,mol-’ energy differen~e.’~~ This correspondsto a 6390preference for of 67%. the a anomer, in agreement with the experimental value26*27 Calculation of the effect of the solvent on the abundance of the 2-(oxan-2y1oxy)oxane c o n f ~ r m e r srevealed ’~~ that the solution behavior of the (a, a) form is markedly different from that of other saccharideswhere large differences in the solvent effect of aqueous and nonaqueous solution on the equilibrium around the glycosidic linkage have been f o ~ n d . ~ ~ ~ The calculatedpopulation of the most stable conformer is 96.5%,in both the isolated molecule and in solution. The equilibrium composition of conformers in the other two forms, (a, e) and (e, e), depends on the solvent, and the solvent- solute electrostatic interactions (Eq. 7) are mainly responsible for the shift of equilibrium in s01ution.l~~ The influence of the ring and anomeric oxygen hydration on the axial preference of an aglycon has been studied for the five most stable conformers of 2-methoxyo~ane.~~~ Both the intramolecularstructure and the energy of a supermolecule formed with 2-methoxyoxaneand water molecules were calculated by the PCILO method. It was concluded that the oxygen atoms in the acetal segment act as a weak monobase in water, and that hydrogen bonding does not influence the axial preference of the methoxyl group.

3. The Anomeric Effect Energy from MO Calculations The energy difference provided by MO calculations, with or without the inclusion of solvent, serves as a direct measure of the anomeric energy, AE(AE3), when its “absolute” definition by Eq. 4 is used. However, the more-frequent use of a “relative” definition by Gibbs energy difference in Eq. I warrant an attempt to recalculate the AE(AE3) data to the values AG(AE 1). Such a procedure is, of necessity, an approximation, because the assumption that AG: = AE(AE3) neglects the entropy and volume changes of conformers owing to absence of suitable information, and the cyclohexane-based and solvent-independent A values must be used. The Gibbs energies, AG(AE1), estimated by the foregoingprocedure from PCILO calculations taking s01vent’~J~~J~’ into account for three oxane derivatives and D-glucopyranose, are listed in Table XVI. The constant values of A(oxane),namely, 3.2, 1.6,6.5, and 6.5 kJ.mo1-l corrected for the oxane ring from the A-values by Eq. 2 were used for the C1, F, CH3,and OH groups, respectively.The dependence of the anomeric effect upon the solvent follows the trends in equilibrium composition discussed in the previous Sections. The anomeric effect is maximized in the isolated molecules; the Gibbs energy, AG(AEl), gradually decreases in more-polar solvents. The major difference appears between the effect in organic solvents and in water, and, in (144) I. TvaroSka, Biopolymers, 21 (1982) 1887- 1897. (145) I. TvaroSka, Biopolymers, 23 (1984) 1951 - 1960.

94

IGOR

TVAROSKAAND TOMAS BLEHA

TABLE XVI Solvent Dependence of the Anomeric Effect, AG(AE,), Recalculated from AE(AE,), of the Chlorine, Fluorine, Methoxyl, and HydroxylGroupsa Solvent

Clb

Fb

OCH,b

OH'

Isolatedmolecule 1,4-Dioxane Carbon tetrachloride Carbondisulfide Chloroform F'yridine Acetone Methanol Acetonitrile Dimethyl sulfoxide Water

13.3 12.6

8.3 7.2

9.5 9.1

9.4 8.4

12.9 12.6 11.7 11.3 11.4 10.1 10.6 11.0 6.4

7.6 7.2 6.4 6.0 6.1 4.7 5.2 5.7 0.3

9.4 9.3 8.7 8.5 8.6 8.0 8.3 8.4 6.3

8.8 8.5 7.8 6.4 6.5 5.2 5.7 6.1 4.6

*Based on the constant A(oxane) values of 3.2, 1.6, 6.5, and 6.5 kJ.mol- for the c1,F,OCH, ,and OH groups,respectively. In 2-substituted oxane. In ~-glucopyranose.

the isolated molecule, the Gibbs energy of the anomeric effect decreases in the order of C1> OCH, = OH > F. In the exo-anomeric effect, identification of its value from MO calculations is slightly more complicated when using the definition of Eq. 4. Because there exist two energetically nonequivalent sc positions for the torsional angle @, in principle both could be used in Eq.4. In practice, however, one of the sc positions is disfavored by steric interactions with the rest of molecule; for example, sc in the e form or -sc in the u form of 2-methoxyoxane (see Fig. 3). The energy differences between conformations calculated for dimethoxymethane and related acyclic compounds," for 2-methoxyoxane conformers of different pyranoid-ring shapes,127~1u) and for the model compounds of the three t r e h a l o s e ~make ~ ~ ~ it possible to estimate quantitatively the magnitude of the exo-anomeric effect in various solvents. These estimates seem to be the only data on the energy of the exo-anomeric effect defined by Eq. 4 thus far available in the literature. Its magnitude, AE(EAE3), calculated as the difference in energies of the conformers having the aglycon group in the sc and the up position (for the angle @) varies in the isolated molecules between 5 and 10 kJ.mol-'. The exo-anomeric energy, AE(EAE3), may exceed the anomeric-effect energy, AE(AE3), when rotation by the angle @ brings about a larger stabilization of the sc position than rotation by the angle 8. For the transition from AE(EAE3) values to Gibbs energy AG(EAEl), a reference molecule has to be selected that should serve as a benchmark to

+

ANOMERIC AND EXO-ANOMERIC EFFECTS

95

assess steric interactions. 2-Ethyloxane seems to be the natural choice, but the detailed conformational energetics of this molecule is not available. A simple acyclic model, ethyl methyl ether, therefore has to be used in order to estimate the "A" values appropriate for the exo-anomeric effect. The ap orientation is favored in ethyl methyl ether'&; from its 80%population in the gas phase, as determined by electron diffraction, this implies that the Gibbs energy difference between the sc and ap rotamers is 5.1 kJ.mol-'. This value represents the steric preference ofthe methyl group for the ap orientation about the C - 0 bond, and servesas a rough estimate of an A-value for the rotation about the exocyclic, anomeric C -0 bond. Such an assumption neglects the differences in the orientational equilibrium about the C - 0 bond in cyclic and acyclic ethers, but may be used as a first approximation. The Gibbs energy ofthe exo-anomericeffect, AG(EAE1), is, therefore, defined as the sum of the energy difference for the sc and ap rotamers about the exocyclicC - 0 bond, AE(EAE3),and the A-value for the corresponding equilibrium in ethyl methyl ether. For example, the energy data for 2-methoxyoxane and the trehalose models analyzed earlier in this Section, combined with the A-value of 5.1 kJ.mol-', give an estimate of the exo-anomeric effect in isolated molecules of between 10 and 15 kJ.mol-'. For solutions, the data for the various ring-forms of the axial 2-methoxyoxane given in Table XVII show that AG(EAE1) for the exo-anomeric effect is minimal in water, and higher in less-polar media. The results for the model of a,a-trehalose, namely, the (a, a) form of 2-(oxan-2-yloxy)oxane, indicate a different behavior. 132 The magnitude of the exo-anomeric effect calculated for this compound is 15 kJ.mol-', and is not sensitiveto solva-

-

TABLE XVII The Effect of Solvent upon the Exo-anomeric Effect" (W.mol-L), AWEAE,), for Selected Ring-Shapes for the Axial Form of 2-Metho~yoxane'~ Solvent

'CI

'C,

'So

"S,

UB

B,

Isolated molecule Carbon tetrachloride Chloroform Acetone Methanol Acetonitrile Dimethyl sulfoxide Water

11.2 11.1 10.4 10.3 9.7 9.9 10.2 7.8

12.1 11.7 11.1 10.8 10.1 10.3 10.6 8.0

14.0 13.5 12.8 12.6 11.8 12.1 12.4 9.7

11.1 10.9 9.7 9.8 9.3 9.6 9.9 7.5

14.2 13.9 13.1 12.8 11.9 12.2 12.5 9.4

11.8 11.4 10.7 10.3 9.4 9.7 10.0 7.1

Based on an estimate of the A value of 5. I kJ.mol-' from ethyl methyl ether.

(146) K. Oyanagi and K. Kuchitsu, Bull. Chem. SOC. Jpn., 5 1 (1978) 2237-2242.

IGOR TVAROSKA AND TOMAS BLEHA

96

TABLEXVIII Observed and Calculated Bond Lengths (pm) and Bond Angles (degrees) for Substituted Dimethyl Ethers (CH,OCH,-X) X

0

F

c1

0

Method

sc

Exp. 4-31G 4-21G CNDO/2 MNDO 4-31G CNDO/2 MNDO Exp. 4-31G 4-21G" MNDO 4-31G MNDO CND0/2 MNDO CNDO/2 MNDO CNDO/2 MNDO CNDO/2 MNDO CNDO/2 MNDO CNDO/2 MNDO MNDO MNDO MNDO

sc

ap

NH,

NH;

CH,

SCH,

up

sc

ap

ap

up

sc

UP

UP

up

sc

ap

ap

ap ap

sc UP sc

sc a

r(C-0)

r(0-C)

r(C-X)

142.4 144.7 145.2 137.0 140.6 143.7 137.0 140.3 142.1 144.9 145.0 140.9 143.9 140.6 137.9 140.1 138.0 140.1 136.9 142.1 136.4 141.9 138.2 140.3 138.I 140.3 i 39.2 138.9 138.5

136.2 139.6 139.8 137.3 139.8 140.5 137.4 140.4 136.2 140.5 139.9 137.2 141.0 138.5 137.1 140.6 137.0 140.7 137.6 136.4 138.0 137.6 137.0 140.2 137.1 140.1 140.5 140.6 140.7

138.5 139.5 140.7 134.9 135.3 137.6 135.2 134.7 182.2 188.6 184.4 183.2 183.7 181.1 141.2 146.5 141.0 146.3 144.1 156.6 144.6 154.7 146.6 154.8 146.7 154.4 175.0 176.1 176.2

(C-0-C)

(0-C-X)

113.5

111.3

115.2 104.6 122.2

110.8 I 10.6 108.6

100.5 120.0 1 14.0

106.0 105.0 112.9

116.3 122.1

112.5 112.0

120.3 101.7 121.4 107.7 120.4 110.9 123.4 100.1 122.1 106.0 120.7 107.7 119.6 120.1 1 19.6 121.4

107.3 113.4 110.4 117.1 109.3 112.1 106.0 105.7 103.5 116.7 1 1 1.9 110.4 109.7 110.4 105.9 114.5

References 62 95 147 11 11

95 11 11

62 95 147 11 95 11 11 11

11 11 11 11 11 11 11 11 11 11

119

119 119

Chlonne atom basis set contafns d functions.

tion. The considerable values of the exo-anomeric effect mentioned document its important role in conformational equilibria of glycosidically linked compounds containing pyranoid rings. These results also explain why the oligo- and poly-saccharidesin the solid state28show a distinct preference for the sc position.

(147) J. E. B o w , M. Altman, F. R. Cordell, and Y. Dai, J. Mol. Struct. THEOCHEM., 94 (1983) 373-390. (148) I. TvaroSka and T. Bleha, Can.J. Chem., 57 (1979) 424-435.

TABLE XIX Comparison of the Calculated Bond Lengths (pm) and Bond Angles (degrees) with Experimental Values for Dimethoxymethane Conformer (sc, sc)

(an sc)

(UP, UP)

a

Method

r(C-5-0-5)

r(0-54-1)

r(C-1-0-1)

r(0-1 -C-i)

a(6)

a(])

/3

References

Exp. Exp." 4-3 lG 4-21G CNDO/2 PCILO MNDO Expa 4-31C 4-21G CNDO/2 KILO MNDO 4-3 1C 4-21G CNDO/2

140.3 143.5 144.4 144.9 137.9 138.5 140.5 143.3 143.4 144.2

1 14.2

114.2 113.1 115.9 114.5

138.5 140.6 140.6 144.4

105.5 120.0 143.5 114.4

112.6 112.3 113.9 112.4 114.9 113.2 113.6 107.9 110.9 109.5 110.0 110.9 108.2

138.4 140.3

140.3 140.4 142.3 142.2 137.6 138.4 139.9 138.3 140.0 140.9 137.7 138.4 139.3 140.6 142.0 137.8 138.4 140.0

140.3 143.1 144.4 144.9 137.9 138.5 140.5 142.7 144.3 144.8

KILO MNDO

140.3 141.6 142.3 142.2 137.6 138.4 139.9 142.8 142.5 143.2 137.9 138.5 140.5 140.6 142.0 137.8 138.4 140.0

138.4 140.3

105.8 119.3

33 94 94 I12 I48 I1 II 94 94 112 148 I1 II 94 112 148 I!

138.4 140.2 143.5 144.4

Mean values from X-ray structures of methyl aldopyranosides.

113.4 I 15.9 114.5 106.3 123.5 111.4 115.8 114.3

105.9 104.6 105.8 103.1

106.3 123.5 113.4 116.1 114.9 107.5 123.4 114.0 105.8 119.3

11

IGOR TVAROSKA AND TOMAS BLEHA

98

4. Coupling of Bond Lengths and Bond Angles to Torsional Angles

The analysis of a large body of structural information concerning the detailed geometry of the acetal and related groups in carbohydrates (see Section 111) established that there are small but significant differencesin the molecular geometry, depending on the orientation of these groups. The characteristic patterns of the variations in bond lengths and valence angles are correctly reproduced by both ab initio and semiempirical MO calculations. Most of the available results for substituted ethers CH,OCH,X are summarized in Tables XVIII and XIX, and those for 2-substituted oxane derivatives are given in Table XX. The calculated bond lengths and angles shown in Tables XVIII and XIX agree rather well with experiment, although some differences are apparent. Upon closer inspection, it is suggested that neither method can be preferred in the prediction of absolute values of bond lengths and bond angles. For example, ab initio calculations at the 4-31G or 4-21G level predict C - 0 bond lengths that are -2-4 pm longer, and the semiempirical PCILO method, 2 pm shorter,than those found experimentally.At the sametime, the PCILO calculated variations of bond lengths are less striking than the variation of experimental values. However, the trends, if not the absolute values, are reproduced correctly, and agreement is good when considering that the calculations refer to an isolated, model molecule, and experimental data are usually collected for more-complex molecules as crystalline solids. The most noticeable feature of the geometry relaxation of CH,OCH,X compounds that occurs during rotation around the C - 0 bond is the reversal of the C - 0 and C - X bond lengths. The C - 0 bond in the sc orientation is shortened, compared with its length in the ap orientation, whereas the C- X bond in the ap orientation becomes longer than in the sc orientation. In bond lengths, the most pronounced changes occur in the 0-C-X angle. The 0 - C - 0 bond angle in dimethoxymethane calculated by an ab initio method with 4-21G basis set falls from 112.4" for (sc, sc) to 109.5' for (sc, ap), and to 105.9" for the (ap, ap) conformation.122 The PCILO calculations result in similar values of 113.2, 110.9, and 105.8", respectively. The bond angle 0- C - X increases in other compounds by 1 to 5 on going from the ap to the sc orientation. Because dimethoxymethane has a 2-fold axis of symmetry in the (sc, sc) and (ap, ap) conformations,the two outer, and similarly,the two inner, C- 0 bonds are indistinguishable. In 2-methoxyoxane, the calculations suggest that all four C - 0 bond-lengths at the acetal center are also different in the

-

O

(149) (150) ( 1 5 1) (152)

P. Luger, G. Kothe, and H. Paulsen, Chem. Ber., 109 (1976) 1850- 1855. G. Kothe, P. Luger, and H. Paulsen, Carbohydr. Rex, 37 (1974) 283-292. I. TvaroSka, unpublished results. S. Perez and C. Vergelati, Acta Crystallogr.,Sect. B, 40 (1984) 294-299.

TABLE XX PCILO Calculated Bond Lengths (pm) and Bond Angles (degrees) for 2-Substituted Oxane Derivatives, Compared with Observed Data on Carbohydrates Group

B

F

sc

9

ap

CI

sc up

OCH,

sc sc

up

SCH,

sc

up -sc

up

sc

ap

ap

sc

sc

-sc sc

-sc

up

-sc

up

up

sc

UP

ap

* Experimental value.

r(C-5-0-5)

40-5-C-1)

r(C-1 -X)

139.7 145.0 139.6 142.8 139.8 145.1 139.8 142.7 139.6 143.3 139.6 139.5 143.3 139.5

138.9 136.2 139.2 140.6 137.5 138.3 138.4 141.5 139.9 141.4 139.9 140.0 142.9 140.0 142.7 139.8 138.5 144.6 138.4 138.4 138.9 143.5 139.0 138.9

136.6 138.9 136.3 136.7 183.5 185.9 182.3 175.4 139.4 139.9 139.7 139.2 138.1 139.4 139.1 139.4 191.5 179.9 191.9 191.9 190.6 180.4 190.8 190.9

139.5 139.7 144.6 139.6 139.7 139.7 144.1 139.7 139.7

a(C-5-0-5-C-1) 110.9 112.6 110.1 109.9 1 12.9 113.0 108.2 111.1 111.9 113.3 111.4 112.4 112.4 111.9 113.7 112.7 113.3 114.9 112.9 110.8 110.7 110.8

a(0-5-C-1 -X)

References

106.8 110.7 103.2 105.7 107.8 107.9 103.4 107.1 110.5 1 12.4 106.4 106.6 107.6 107.2

12 98" 12 98" 12 149" 12 150" 151 97" 151 151 97" 151 30" 151 151 152" 151 151 151 152" 151 151

102.3 110.5 111.9 110.0 108.7 105.2 108.8 105.0 104.1

IGOR TVAROSKA AND TOMAS BLEHA

\

I / /

\

\

I

60

\

I

180

I

- 60

@ (degrees)

FIG.20.-The Variation ofBond AnglesC-5-0-5-C-1 (CurvesDenoted bya)and0-5-C1 -0-1 (Denoted by b) with Torsional Angle in the Axial Form (Dashed Line) and Equatorial Form (Full Line) of 2-Methoxyoxane.

(sc, sc) and (up, up) conformations. In agreement with experiment, the endocyclic C-5 -0-5 -C-1 bond is longer than the corresponding exocyclic C- I - 0-1 - C bond (see Table XX). A striking interdependence of the bond angles and the orientation of the anomeric bond is made clear in Fig. 20 by presenting a plot of the C-5 - 0-5 C- 1 and 0 - 5 - C- 1 - 0-1 bond angles versus the angle @ for the a and e forms of 2-methoxyoxane. The former bond-angle displays the maximum variation,changingintherangesof105.5 to 115.5" and 101.5 to 108.5" forthea and e forms, respectively. As may be seen, this angle is -4" larger in the a than in the e form. In both cases, the minimum value ofthe angle pertains to

ANOMERIC AND EXOANOMERIC EFFECTS

101

the ap orientation. All of the calculated trends are in accord with the experimental data reviewed in Section I11 on the influence of conformation on the acetal g e ~ m e t r y , ~and ~ * 'confirm ~* a significant function of the anomeric and exo-anomeric effects in specification of the geometry parameters of saccharides. As a corollary to the foregoing discussion, the computations also provide an answer to the controversial question of the influence of assumed geometry on the calculated, conformational-energydifferences.The results present evidence of the necessity of at least a partial optimization, including the main bond lengths and bond angles, in the theoretical calculations of molecules exhibiting the anomeric and exo-anomeric e f f e ~ t ~ . ~ ~ J ~ ~ J ~ ~ J ~ ~ 5. Electron Distribution and Lone Pairs

Besides the energy and equilibrium geometry of conformers, the MO methods provide a great deal of useful information on the distribution of electronic charge in molecules, and on the degree of localization of individual MO's on the atomic centers. A relativeexcess or deficit ofelectrons on the atom is represented by net charges Q1. Whereas most MO's are widely delocalized by their nature, some prominent MOs, such as those involving lone-pair electrons, are centered on one atom. Symmetry of two lone-pairs on oxygen atoms in the acetal and related moieties is a crucial point in the rationalization of the nature of the anomeric effect (see Section VI). The traditional, most familiar picture of the lone-pair orbitals on the oxygen in water, alcohols, and ethers assumestwo localized orbitals of tetrahedral symmetry and of equal energy. There is, however, experimental evidence concerning the energeticnon-equivalenceof the two lone-pair orbitals in saturated, oxygen-containingmolecules155J56 that supports an alternative description of lone pairs. In this representation,one lone-pair orbital having x symmetry is perpendicular to the R -0 - R plane, and the second lone-pair orbital, having c7 symmetry in the R-0-R plane, is 1-2 eV lower in energy. Both representations are shown in Fig. 2 1. Similarly, two different pictures can be used for three lone-pair orbitals of halogen^.'^' The energetic and directional non-equivalence of lone-pair orbitals is an important factor for any situation where the symmetry of the interaction, or the energy of the lone-pair electrons, is critical, as in examination of the conformational preferences. An experimental approach to the elucidation of the character of lone-pair

-

(153) (154) (155) (156) ( 1 57)

R. U. Lemieux and K. Bock, Arch. Biochem. Biophys., 221 (1983) 125- 134. I. TvaroSka and S. Perez, Curbohydr. Rex, 149 (1986) 389-410. D. W. Sweigart and D. W. Turner, J. Am. Chem. Soc.. 94 (1972) 5599-5603. S. Cradock and R. A. Whiteford, J. Chem. SOC. Furuduy Trans. 2,68 (1972) 28 1-288. 0.Eisenstein, N. T. Anh, Y .Jean, A. Devaquet, J. Cantacuzene, and L. Salem, Tetrahedron, 30(1974) 1717-1723.

IGOR

102

(a)

TVAROSKAAND TOMAS BLEHA

(b)

FIG.2 1 .- Lone-Pair Electrons ofoxygen Shown in (a) u-z Representation,and (b) sp3-Hybridization Representation.

electrons consisted in a measurement of electron densities by X-ray and neutron diffraction techniques.I5*These difficult, but highly accurate, measurements have been carried out to date for two sugar molecules.'s9J60The electron density maps computed for the ring and anomeric oxygen atoms of j?-DL-arabinopyranose'60suggest that the spatial distribution of lone-pair orbitals on each oxygen atom corresponds to the sp3type of orbital, rather than to a and a lone-pair orbitals. The four lone-pair orbitals are not, however, identical with the lone-pair orbital of the anomeric oxygen atom, which is ap to the C- 1 - 0-5 bond, thus differing from the other three orbitals. In carbohydrates, the anomeric carbon atom is bonded to two heteroatoms having higher electronegativity than carbon. The electronegativity differences and the presence of lone-pair electrons are reflected in the electron distribution in the molecule. The ring-oxygen atom on one side, and the aglycon (X) on the other, lessen the electron density at the anomeric carbon atom, and the C 0 and C - X a-bonds are polarized in the direction of the more electronegativeatoms. On the other hand, a delocalization of electrons from lone-pair orbitals of the oxygen aglycon towards the anomeric carbon atom could take place (see Section VI). Because the spatial relationships of the lone-pair orbitals and the bonds at the anomeric center change with the conformation, there are also differencesin the electron distribution between the conformers. It was established from calculations of substituted dimethyl etherI2that the largest differencein net charge occurs at the central carbon atom and on its substituent; that is, on the atoms where a delocalization interaction occurs. A pronounced enhancement of negative charge on the oxygen atom

-

( 1 58) P. Coppens, in J. M. Robertson (Ed.), MTP Int. Rev.Sci. P h p . Chem., Ser. 2, Vol. 1 1, Butterworth, London, 1975, pp. 21-54. (159) J . C. Hanson, L. C. Sieker, and L. H. Jensen, Acfa Crystallogr., Sect. B, 29 (1973) 797 - 808. (160) F. Longchambon, H. Gillier-Pandraud, R. Weist, B. Rees, A. Mitschler, R. Feld, M. Lehmann, and P. Becker, Acta Crystallogr., Sect. B, 41 (1985) 47-56.

103

ANOMERIC AND EXO-ANOMERIC EFFECTS TABLE XXI PCILO Calculated Net Atomic Charges0and Dipole Momentsbof the Most Stable Conformers of 2-Metho~yoxane'~~ (+sc,

+sc)

291.8 -21.0 -45.0 0.4 160.0 100.6 - 170.5 - 194.2 - 78.0 0.28

(UP,+sc)

294.0 - 30.8 44.7 0.6 160.0 97.8 - 171.3 -181.4 - 75.4 2.17

299.9 -25.3 40.9 0.1 156.5 102.6 - 160.1 - 187.5 -96.5 2.13

"Q(i), in lo-C. b,u,in D, ID= 13.33563 X

293.8

- 34.5

42.6 -0.6 156.5 104.6 - 169.1 - 186.9 - 76.9 2.36

(UP,UP)

299.0 -35.2 43.4 -0.6 154.7 95.3 - 154.2 - 168.3 -94.9 3.18

mAs.

and the heteroatoms (X) was found in ethers upon conversion of the up into the sc orientation. Similarly, dependence on conformation was inferred from net charges of selected atoms and dipole moments of 2-methoxyoxaneIz8(see Table XXI). With the exception of the atoms of the a c e d segment, the changes in electron distribution by rotation about the C- 1 - 0-1 bond are small. The minimum dipole moment (0.28D) is exhibited by the (+sc, sc) conformer, where dipole moments of lone-pair orbitals on the ring-oxygen atom are oriented antiparallel to those at the anomeric oxygen atom. The largest dipole moment (3.18D) is associated with the (ap, up) conformers, with lone-pair orbitals on both oxygen atoms being oriented parallel.

+

V. THEANOMERICAND EXO-ANOMERIC EFFECTS IN POTENTIALFUNCTION CALCULATIONS

Although several MO methods successfully describe the stereochemical behavior of pyranose models, their application to the more-complex carbohydrates and oligosaccharidesis limited at present, due to reasons of economy. An alternative, more empirical approach has to be pursued, based on the summation of suitable potential functions (PF) accounting for intramolecular interactions. The intramolecular energy terms usually represent contributions from nonbonded interactions, electrostatic interactions, hydrogen bonding, torsional terms, bond stretching, and valence-angle bending terms. The firstattempts to calculatethe stability of monosaccharideisomers by the additive have shown the necessity of introduction of an extra anomeric-effectparameter, in addition to the parameters of interaction

104

IGOR TVAROSKA AND TOMAS BLEHA

between substituents. The extra parameter amounted to 1.8 -6.3 kJ.mol-’, depending on the chemical structure of the monosaccharides and on the selected set of parameters. An identical problem ariseswhen PF calculations are carried out on saccharidesand similar molecules. In conformity with Eq. 3, an extra energy contribution, AE(AE2), is assumed in order to “convert” the results of PF calculations into agreement with experiment. Obviously, the anomeric-effect correction, AE(AEZ), in the PF method depends on the set of potential functions used, and thus it is not universal. For example, extra energy, 1.7 W.mol-l, was needed in order to account correctly for equilibrium in hexo- and pento-pyran~ses,~~ and PF calculationss6of rotation around the C - 0 bond in 2-methoxyoxane required the addition of 4.6 kJ.mol- of stabilization energy in order to reproduce the experimental data. These and numerous other studies make clear that direct application of the PF method, developed primarily for nonpolar molecules, to saccharides and their models requires its modification, especially in a way to account properly for the anomeric effect. In the first stage, the modifications concerned mainly the energy of the anomeric effect, and, in later development, the valence geometry aspect. MO studies of the model compounds were instrumental in achieving the required improvement of the PF method. It was evident from MO calculations of torsional energetics of acetals.113~161 that the PF procedure must be modified by the proper incorporation of the lone-pair electrons into calculation of intramolecularenergy. The lone-pairs bring about the anisotropy of electron distribution in a molecule, such that the majority of the charge density is localized in their direction. The lone pairs may be regarded as pseudoatomshaving a size and dipole moment, and they may be incorporated into the PF scheme. A simple PF method was successfully modified161in this way, by considerationof the lone-pair dipoles in calculation of the electrostatic term. The magnitude and the localization of the lone-pair dipoles can be transferred from MO calculations of simple compounds or, alternatively, by using the universal procedure16*where the modulus and direction ofthe lone-pair dipoles is expressedin dependence on hybridization (bond angle) on the heteroatom. By using the procedure, the PF method correctly reproduced conformational energies of dimethoxymethane; that is, the extra energy term, AE(AE2), was zero in this case.’61 The developmentofthe PF method culminated in molecular mechanics, a reliable method of prediction of conformational energy and equilibrium structure of nonpolar molecules. For molecules having several heteroatoms, a modified parameterization was developed, with the lone-pair parameters as a part of the force field, and with their inclusion into all types of intramo(161) I. TvaroSka and T. Bleha, Collect. Czech. Chem. Commun., 43 (1978) 922-931. (162) I. TvaroSka and T. Bleha, Biopolymers, 18 (1979) 2537-2547.

ANOMERIC AND EXO-ANOMERIC EFFECTS

105

lecular interaction^.'^^ Using the MM2 method parameterized in this way, reasonable conformational energies were computed for pyranose models, provided that interactions of the bond dipoles were assumed in the electrostatic term.'64Computationswith net charges on atoms only, that is, lacking lone-pair dipoles, gave conformational energies in disagreement with the experimental data. An alternative way of adaptation of the PF computational scheme to the molecules exhibitingthe anomeric effect consists in the addition of a proper, preferably simple, potential term. In other words, an attempt is made in this approach to establish the extra energy term, AE(AE2),beforehand, and then to incorporate it into the PF method. In one of the first trials, an additional, twofold torsional term for C - 0 rotation in the C - 0 - C - 0 segment was assumed,'65with a bamer of - 7.6 kJ.mol-l at -90". In the next attempt to calculate the torsional potential around the glycosidicbond by the simple PF method, nonbonded interactions were combined,59with the ab initio 4-3 1G torsional potential of dimethoxymethane serving as an extra energy term. Because, in such a procedure, some interactions were evidently counted twice, the corrected approach known as HSEA (Hard Sphere Exo-Anomeric) was suggested,60where the extra energy, AE( EAE2), is expressed as a simple function of the torsional angle Q, for each anomer separately. A similar approximation of the latter extra energy was developed6' that was in analogy with Eq. I, from MO calculationsfor cyclic models, as the difference between the torsional potential of 2-methoxyoxane and that of 2-ethyloxane. The resulting different expressionsfor a and p anomers, as a function of the angle a,were subsequently used as additional energy contributions in the PF studies of oligosaccharide conformations.153~154 All of the amendmentsof the PF method so far discussed were centered on the conformational energy. The development of empirical computational methods able to account for the valence geometry changes on the anomeric center is much less straightforward. Sophisticated force-fields are required, with the numerous parameters exactly reproducing the interdependence of the bond lengths, bond angles, and torsional angles. The first force-field developed for saccharides166 correctly described the overall saccharide structure, but failed to imitate the specificvariations of geometry in the vicinity of the anomeric carbon atom with different configurations of substituents. In contrast to the conformational energy, the incorporation of lone airs^','^^ into the molecular mechanics schemes MM 1 and MM2 did not bring about improvement in the description of the anomeric center geometry. Modifica(163) (164) ( 165) (166)

N. L. Allinger and D. Y. Chung, J. Am. Chem. Soc., 98 (1976) 6798-6803. T. Koiir, A. Sarko, and I. TvaroSka, unpublished results. U. Burkert, Tetrahedron, 35 ( 1979) I945 - 195 1. S. Melberg and K. Rasmussen, J. Mol. Srrucr., 57 (1979) 215-239.

106

IGOR TVAROSKA AND TOMAS BLEHA

tion of the MM 1 by separate parameters for the ring-oxygen atom, the glycosidic oxygen atom, and the anomeric carbon atom, which, moreover, differ for the a and p anomers in the calculation of the energy terms, constituted the first successful attempt. The MM2 method has been amended in a similar way, and used to calculate the structure of various conformers of trehalose models,132c e l l o b i ~ s eand , ~ ~maltose.154 ~ The computations demonstrate the ability of the method to render the variations of disaccharide geometry with the orientation around the glycosidicbonds. In spite of the success, the modifications described were handicapped by the necessity to introduce additional parametersto an already large set of parameters. To avoid this problem, the standard C - 0 bond length in the MM2 method has been expressed168as a function of the orientation of both of the central C - 0 bonds in the acetal segment. Further testing in the future should assess the merits and shortcomings of this and other modifications of force fields. A full reproduction of the anomeric and exo-anomenc effect in conformationalenergies and valence geometry of saccharidesis an ultimate goal of this effort.

VI. NATURE OF THE ANOMERIC EFFECT In previous Sections, we have presented the experimental evidence as to the various manifestations of the anomeric effect, and have discussed the results of quantum-mechanical calculations made on model compounds and carbohydrates. In this Section, the nature of this effect will be discussed. It is obvious that any explanation of the anomeric and the exo-anomeric effect should clarify both the conformational preference, and the variations of the bond lengths and bond angles. Various rationalizations of the anomeric effect have been offered over the years; they may be roughly divided into two main groups, namely, the electrostatic and the delocalization rationalization of the anomeric effect. Besides these two dominant concepts, the anomeric effect was regarded as a result of a fine equilibrium between the electron - electron repulsion and the nucleus- electron a t t r a ~ t i o n , of ' ~ ~a barrier of internal rotation about the C-X bond170 or of the Jahn-Teller effect.171However, neither of these explanations has gained wide recognition, and, therefore, they will not be discussed here. The two prevailing rationalizations, electrostatic and delocalization, involve interactions of the (167) (168) ( 169) (170) (171)

G. A. Jeffrey and R. Taylor, J. Comput. Chem., 1 (1980) 99- 109. L. Nerskov-Launtsen and N. L. Allinger, J. Comput. Chem., 5 (1984) 326-335. S. Woife, A. Rauk, L. M. Tel, and 1. G. Csizmadia, J. Chem. Soc., B, ( I 97 1) 136 - 145. Yu. A. Zhdanov, R. M. Minayaev, and V. I. Minkin, J. Mol. Struct., 16 (1973) 357-364. R. Ponec and V. Chvalovsky, Collect. Czech. Chem. Commun.,39 (1974) 2613-2615.

107

ANOMERIC AND EXO-ANOMERIC EFFECTS

H

t

X

(6 1

(a)

FIG.22.- Dipolar Interactions in 2-Substituted Oxane Derivatives.

lone-pair electrons of oxygen (or other hetero atom). The character of those lone pairs (hybridization)is thus a very important element in the discussion. 1. Electrostatic Interactions

In the first interpretation, the anomeric effect was discussed4in terms of electrostaticrepulsion between the carbon-substituentdipole and the resultant dipole of the lone-pair orbitals on the ring-oxygen atom (seeFig. 22). The repulsive interactions are maximized in the equatorial conformer when the dipoles are parallel, and account for the preference for the axial conformer. The difference in dipole-dipole interaction energies between the axial and equatorial conformer of 2-chloro- and 2-bromo-4-methyloxane has been estimated38at 11.3 and 10.0 kJ.mol-I by use of classical electrostatics. In dimethoxymethane, the differences in electrostatic interactions calculated between the (up, up) and (up, sc) and between the (up, sc) and (sc, sc) conformers are 7 and 2 kJ.mol-', respectively.161If electrostatic interactions between the dipoles along the axes of the sp3-typelone-pair orbitals on ~ ~approach ~ ~ ~ was ~ picture~ ~ ~ ~ ~ ~ * the oxygen atoms are a s ~ u m e d , this squely termed the rabbit-ear effect.173Thus, the most stable conformer of acyclic and cyclic acetals is that having on oxygen atoms the minimum number of syn-axial, lone-pair interactionsthat can cause dipolar repulsion, with an energy of - 4 kJ.mo1-I assigned to each of such interactions. The syn-axial, lone-pair orbital interactions in the acetal segment of 2-methoxyoxane are illustrated in Fig. 23 (compare, also, Fig. 3). 2-Methoxyoxane having an equatorial methoxyl group has two conformers, (up,-sc) and (up, sc), that have a single, syn-axial, lone-pair orbital interaction, and a third, (up, up), which has two such interactions. On the other hand, in the axial form, two conformers, (+sc, -ap)and (+sc, -sc), have one interaction, and

-

+

(172) R. 0. Hutchins, L. D. Kopp, and E. L. Eliel, J. Am. Chem. Sm.,90 (1968) 7174-7175. (173) E. L. Eliel, Angew. Chem., 17 (1972) 779-791.

108

IGOR TVAROSKA AND TOMAS BLEHA

(ap, a p )

(+sc, -sc) FIG.23.-The Segment.

Rabbit-ear Effect Interactions in the Staggered Conformations of an A c e d

+

the third, (+sc, sc), has no lone-pair orbital interactions. It may be predicted, therefore, that the axial will be more stable than the equatorial conformer and that the (+sc, sc) conformer should be the most stable of the axial conformers, and the (up, -sc) conformer the most stable of the equatorial conformers, owing to unfavorable interactions between the methyl group and the axial hydrogen atom on C-2 in the (up, +sc) conformer. The electrostaticconcept predicts the stabilityof conformers in qualitative agreement with experimental observation, but quantitative accord cannot

+

ANOMERIC AND EXO-ANOMERIC EFFECTS

109

be expected from such a crude approach. The largest support for this rationalization comes from an observed lessening of the anomeric effect in morepolar solvents (see Tables VI, VII, and XIV). Also, the analysisof the individual terms in the intramolecular energy of dimethoxymethane s h o ~ ~ ~ ~ that the preference for the sc orientation about the C - 0 bond originates in the electrostaticrepulsion of the electron densities located on the two oxygen and associated bonds. However, the electrostaticconcept is unable to rationalize variations of the bond lengthsand bond anglesthat are characteristic for the anomeric effect.

2. Delocalization Interactions A second, perhaps more popular, rationalization arose from investigations of a-halo ether s t r u c t ~ r e s . ~In~these J ~ ~ compounds, *~~~ preference for the sc orientation is associated with a characteristic lengthening of the carbon - halogen bond, whereas the adjacent C -0 bond is shorter.I3A stereoelectronicexplanation is illustrated for chloromethoxymethanein Fig. 24. A stabilization of the sc orientation is attributed to delocalization or backdonation of the a-type, lone-pair orbital on oxygen to the antibonding orbital of the adjacent carbon - halogen bond. The n-type lone-pair orbital is oriented perpendicular to the C- 0- C plane, and therefore the best orientation of the carbon-chlorine bond for this delocalization is in the orthogonal (og) orientation where both orbitals are in the same plane. This interaction produces the following consequences;the lengthening of the C- C1 bond by populating of its a* antibonding orbital, the contraction ofthe C- 0 bond by increasingits double-bond character, and an opening of the 0- C-C1 bondangle compared to its tetrahedral value, because of the increased doublebond character at C. Some insight into these interactions may be obtained from the distribution of oxygen lone-pair electrons calculated by the ab initio method with 4-3 1G basis for methanol, fluoromethanol, and methanedi01.I’~In methanol itself, the a-type lone-pair electrons are subjected to back-donation into antibonding-type orbitals of suitable symmetry on the methyl group. This leads to a decrease of the n-orbital population from the value of 2.00, appropriate to water, 1.97. A comparable electron displacement occurs in fluoromethanol when the C-F bond is in the nodal plane of the oxygen lone-pair orbital. If, however, the C-F bond and lone-pair orbital are in the same plane (in the og orientation), the back-donation lowers the population to ( 1 74) (175) (176) (177)

N. Gresh, P. Clavene, and A. Pullman, Theor. Chim. Acta, 66 (1984) 1-20. E. A. C. Lucken, J. Chem. Soc., (1959) 2954-2960. G. Baddeley, Tetrahedron Lett., (1973) 1645- 1648. J. A. Pople, Tetrahedron, 30 (1974) 1605-1615.

110

IGOR

TVAROSKAAND TOMAS BLEHA

R"

I I

I I

I

&-

;I

CI

+? & j\ 01

CH3

H-

C #QC'

i'

CH3 / O \ H c C

I

H H FIG.24.-Back-donation of Lone-pair Orbital on Oxygen into the Adjacent C - 0 Bond in Chloromethoxymethane.

1.94,and leads to additionalstabilization.Similarly,the back-donation from the oxygen lone-pair orbital to the C - 0 antibonding orbital in the (sc, sc) conformation of methanediol is 0.05, which is significantly larger than the 0.03 in methanol. In resonance terms, this correspondsto double-bond- nobond structures0 - - C = O+. This simple, delocalizationapproach has been used to rationalize differences in n.q.r. frequency in a-chloro e t h e r ~ . ' ~ ' J ~ ~ The delocalization concept has been further advanced by stressing the necessity for analysis of the orbital interactionsof both lone-pair orbitals on an oxygen atom, one with n-symmetry and one with o - ~ y m m e t r y . ~ * ~ ~ ~ ~ J Even though the lone-pair orbital of the o-symmetry is more strongly bonded, its interactions with appropriate antibonding orbitals cannot be neglected. It has been shown157that, if the energetic and directional non( 1 78) S. David, 0.Eisenstein, W. J. Hehre, L. Salem, and R. Hoffmann, J. Am. Chem.SOC., 95

(1973) 3806-3807.

ANOMERIC AND EXO-ANOMERIC EFFECTS

H

X

(4

111

(b)

FIG.25.-Orientation of Two Non-equivalentLone-pairs of the u and A Type in the a and e Forms of 2-Substituted Oxane Derivatives.

equivalence of the two oxygen lone-pair orbitals is accounted for, their interactions with the antibonding C- X orbital leads to the axial preference of the electronegative substituent X. The orientations of the pertinent orbitals in the two competing conformations of 2-substituted derivatives of oxane are shown in Fig. 25. A qualitative estimate for X = C1 that the axial has a larger stabilization than the equatorial conformer. The incorporation of two nonequivalent, lone-pair orbitals into the analysis of the nature of the anomeric effect was a step of considerable importance, but qualitative reasoning still prevailed at that time. Detailed perturbation calculation of all delocalization interactions of lone-pairs has been performed for dirneth~xymethane'~~ ;to date, this is the only one conducted. The two lone-pair orbitals on oxygen atoms are engaged in mutual interactions by a through-space mechanism and in interactions with appropriately oriented antibonding orbitals of the CH2 group having n-symmetry and with o*-antibondingorbitals of two central C - 0 bonds by a through-bond mechanism.179 The relevant molecular orbitals of dimethoxymethane are illustrated in Fig. 26, but, for the sake of simplicity,only one (179) R. Hoffman, Acc. Chern. Res., 4(1971) 1-9.

IGOR TVAROSKA AND TOMAS BLEHA

112

CH) CH3

Cii,

(aPJap)

(“p ‘ 9 ) D

FIG.26.--Interactions of Lone-pair Orbitals on Oxygen Atoms with (a) Antibonding Orbitals of CH, Group and (b) Antibonding Orbital of Adjacent C - 0 Bond, for (up,up)and (up, og) Conformations of Dimethoxymethane. [For the sake of simplicity,only one oftwo perpendicular antibonding orbitals of the CH, group and the antibonding orbital of one C - 0 bond are shown.]

113

ANOMERIC AND EXO-ANOMERIC EFFECTS

of two perpendicular, n-antibonding orbitals of the CH, group and one a*-antibonding orbital of two central C - 0 bonds are included. These orbitals are separately shown in Figs., 26a and 26b. The magnitude of the individual contributions to delocalization interactions in dimethoxymethane depends on the mutual orientation of pertinent orbitals, and, therefore, on the orientation about the two inner C - 0 bonds. In the (up,ap) conformation, the lone-pair orbitals on oxygen atoms having the same symmetry are colinear and, therefore, interactionsbetween the two lone-pairs on oxygen atoms by the through-space mechanism are optimal, as well as interactions with the Ir*-antibondingorbitals of the CH, group by a through-bond mechanism. The back-donation into a a*-antibonding orbital of central C - 0 bonds is possible only from the a-type of lone-pair, because the n-type of lone-pair is perpendicular to the 0-5 -C- 1 0 - 1 plane. In the (up,og) conformation, owing to rotation about one central C - 0 bond, the spatial orientations of two lone-pair orbitals on the corresponding oxygen atom are interchanged. Consequently, the a-type of lone-pair orbital interacts by a through-space mechanism with the n-type of lone-pair orbital of the second oxygen atom, and vice versu. Similarly, interactions of these lone-pair orbitals with antibonding orbitals of the CH, group are reversed. The back-donation of lone-pairs on the oxygen atom into the a*-antibonding orbital of the C - 0 bond in the (up,og) conformation is reversed in comparison to the (up,ap) conformation. The n-type of lone-pair is in the same plane as the relevant C - 0 bond, and the back-donation into the antibonding orbital is maximal, whereas the back-donation of the a-type of lone-pair disappears. The energy of the aforementioned interactions as a function of rotation about the C - 0 bond has been calculated by perturbation MO theory.148The results show that the overall stabilization energy is rather large, -37 kJ.mol-I in the (up,up)conformation, but its conformationaldependence is not pronounced. The stabilization-energycurve resembles the V, term of Fourier expansion (see Eqs. 5 and 6), with the ogposition being 1 kJ.mo1-1 less stable than the up position.148This situation arises as a consequence of reverse conformational dependence on the n- and a-type lone-pair interactions. The foregoinganalysis does not confirm the delocalization concept as being the origin of the energetic manifestation of the anomeric or exo-anomenc effectsin acetals. On the other hand, the electron transfer associated with lone-pair orbital interactions is in complete accord with the trends observed, and it explains all of the changes in the bond lengths and valence angles, including the shortening of C - 0 bonds in (up,up) conformations, which is not obvious from the double bond-no bond interpretation of the

-

114

IGOR TVAROSKA AND TOMAS BLEHA

anomeric effect. A study of trans- 1,8-dioxadecalin by photoelectron spectroscopylsOconfirmed the results of theoretical analysis. An acetal segment fixed in the (ap, up) conformation shows two well resolved, low ionizations, at 9.08 and 9.93 eV. The difference of ionization potentials of 0.85 eV is the largest observed for acetals,155 and indicates a strong, through-space interaction of lone-pairs in the (up, up) conformation.

3. Scaling between Electrostatic and Delocalization Interactions The foregoing discussion shows clearly that neither the electrostatic nor the delocalization concept is able to describe all of the peculiarities of the anomericeffect in saccharides. The electrostaticconcept, sometimesnaively represented by the rabbit-ears effect, properly describesthe preference for the sc orientation and the decrease of this preference with increasing solvent polarity, but fails to reproduce changes in valence geometry. On the other hand, the delocalization concept, assuming non-equivalent, oxygen lonepair orbitals, completely interpretsthe changes observed in bond lengths and bond angles, but does not account for the energy preference of the sc orientation. It is apparent, therefore, that only the combination of both concepts can give a coherent explanation of all of the phenomena mentioned. A decomposition of the calculated torsional potential into components of the Fourier expansions (Eqs. 5 and 6) provides a simple technique to illustrate the balance of both concepts. As has been noted, although the decomposition is rather formal, the simple physical meaning can be ascribed to the expansion coefficients. The terms V,, V,, and V, are identified with the inherent, torsional barrier for single-bond rotation, with the delocalization interactions of lone-pair orbitals and with the interaction of dipoles in the rotating segments, respectively.111 Obviously, the shape of the torsional potential and the positions of the energy minima depend on the magnitude of the expansion coefficients. In Section IV, 1, the torsional potential of the C - 0 bond in CH,OCH,X moleculescalculatedby MO methods has already been analyzed (see Table XIII). The energy difference between the up and sc orientations can be expressedgL3 by using Eq. 5. AE(AE3) = 0.75(V?- Vs)

(8)

Hence, the anomeric effect [positive value of AE(AE3)l can appear at a suitable combination of the V? and Vq coefficients. Because the V, term is associated with the dipole-dipole interactions of polar groups, and the V, term with delocalizationinteractionsof lone-pair orbitals on heteroatoms, it may be concluded that the latter equation also effectively describesa balance ( 180) F. S. Jerrgensen and L. Nerrskov-huntsen, TetrahedronLeft.,(1982) 522 1 - 5224.

ANOMERIC AND EXO-ANOMERIC EFFECTS

I15

of the two origins of the anomeric effect, that is, of electrostaticand delocalization interactions. Depending on the character of the heteroatoms, one or the other factor may prevail. A definite appraisal can be obtained only by calculation of the V, and V, terms for a given molecule. For example, comparison of the Vy and Vq values for dimethoxymethane and chloromethyl ether (see Table XIII) showed that this balance in dimethoxymethane is considerably shifted toward electrostatic interactions, whereas, in chloromethyl ether, the preference for the sc rotamer is determined mainly by delocalization interactions. A dominance of delocalization interactions reflects a strong tendency of the chlorine substituent for back-donation,181 and explains the more-pronounced differences of bond lengths and bond angles observed between a-and P-D-pyranosyl chlorides than in other pyranose derivatives (see Table XX). The balance of electrostaticand delocalization interactionsin an isolated molecule may be perturbed by the influence of the solvent. In calculations based on Eq. 7, the analysis of solvation-energy terms suggested'*' that the electrostaticcontribution stabilizing the up orientation of the acetal segment is the conformationally dominant term. For example, in 2-methoxyoxane, the difference in energy of the (ap, ap) and (up, sc) conformers in water, compared to that in the isolated molecule, caused by solute-solvent electrostatic interactions alone, amounts to 4 kJ.mol-*. Accordingly, the interand intra-molecular, electrostatic interactions operate in reverse directions in acetals. Whereas the intramolecular, electrostaticinteractionsare responsible, together with delocalization interactions, for the appearance of the anomeric, reverse anomeric, and exo-anomeric effects, the solute- solvent electrostaticinteractionslessen their magnitude, and may even cancel them. Of course, the solvent may also influence the electron distribution and energy of MO's in a molecule. In this way, the orbital interactions of lonepairs and delocalization contributions to the anomeric effect may be scaled by the solvent, but this mechanism of the environmental effect is, in most cases, of only minor importance. Finally, a less conventional explanation of the origin of the anomeric effect may be mentioned. In a communication,4oit was stated that, contrary ~ - ~theoretical * ~ ~ - ~ ~ calculato the results from m e a ~ u r e m e n t s ~ , ~and the equilibrium constant for axial - equatorial equilibrium in 2-methoxyoxane derivatives is independent of variation in the temperature. Both anomers are therefore isoenthalpic, and the Gibbs energy difference is largely determined by the entropic term, the axial isomer having 11.3 J.K- mol- greater entropy.

-

(181) R. C. Bingham, J. Am. Chem. Soc., 97 (1975) 9743-9746.

116

IGOR TVAROSKAAND TOMAS BLEHA

VII. ROLEOF THE ANOMERIC EFFECTIN THE REACTIVITY OF CARBOHYDRATES

In the previous Sections, we have discussed the consequences of the anomeric effect for the ground state of aldopyranosesas reflected in the stability of conformers, valence structure, and electron distribution. It may be expected that intramolecular interactions involved in the anomeric effect, apart from the ground state, also influence the other points on the generalized reaction-hypersurface. As a matter of fact, it is common belief in carbohydrate chemistry that the anomeric and related stereoelectronic effects modify the reactivity of saccharides. Their actual influence on the course of reaction is, however, far from being understood. Comprehension is complicated by numerous factors, such as the flexibility of the pyranose ring, the variability of the type and localization of substituents and reacting groups on the ring, the concentration of reactants, and the character of the solvent. 182-186 Intensive exploration of this problem has resulted in formulation of qualitative rules'87allowing interpretation of the course of the whole range of reactions by using the concept of the anomeric effect. A complete description of progress in understanding the reactivity on the anomeric center lies outside the scope ofthe present article, and the reader is referred to excellent reviews. Instead, we shall focus our attention on some consequences ensuing, for reactivity, from quantum-chemical studies of saccharides and their models. 1. Energy of Reaction Paths

The observation of the significant stereoelectronic effect on reactivity (that is, a higher reaction rate of one conformer than another for stereoelectronic reasons) presumes the condition that the barrier of transition, AEL, from the less reactive (A) to the more reactive (B) conformer exceeds the activation energy of reaction, AE&,of the B conformer. When this condition is not satisfied, the two conformers rapidly interchange, because of the low (182) (183) (184) ( 185)

J. N. BeMiller, Adv. Carbohydr. Chem.. 22 (1967) 25- 108. B. Capon, Chem. Rev., 69 (1969) 407-498. E. H. Cordes and H. G . Bull, Chem. Rev., 74 (1974) 581 -603. J. Szejtli, Saurehydrolyse der glykosidischer Bindungen, Akademiai Kiado, Budapest,

1976. (186) C. Schuerch, in Ref. 9, pp. 80-94. (187) P. Deslongschamps, Tetrahedron, 31 (1976) 2463-2490. ( 188) P. Deslongschamps, Stereoelectronic Eflectts in Organic Chemistry, Pergamon, Oxford, 1983. (189) V. G. S. Box, J. Heterocycl. Chem., 19 (1982) 1939- 1966. (190) V. G. S. Box, J. Heterocycl. Chem., 20 (1983) 1641 - 1653.

ANOMERIC AND EXO-ANOMERIC EFFECTS

117

barrier, AEL, and they can react through the same transition state, that ofthe reactive conformer B; and also, their relative reactivity is determined solely by the difference of ground-state energies. However, if the barrier, AEL, is larger than the activation energy of the reactive conformer, AEg, but is smaller than the analogous activation energy of the less reactive conformer A, both conformers can again react through the transition state of conformer A, but the conformational interconversion through barrier AEL is the ratedetermining step10J91-193 for conformer A. The overall reaction of A will thus be slower than that of B, as a result of the higher activation energy, AEL, for direct reaction, although the observed rate has no direct relationship to the size ofthis barrier. Traditionally, stereoelectroniceffectsare looked for in the ground-state energiesonly,but they also operate in the transition statesof reactions and conformational interconversions. What are the energy relations in the six-membered ring of pyranoid saccharides? The energy difference of the axial and equatorial forms on the anomeric center is low, in the interval of 0- 1 1 lcJ.mol-l, dependingon the medium (see Tables 11,111, and VII). The interconformational(pseudorotational) barrier, AEL, in six-membered rings is -42-46 kJ.mo1-I. Some authorslS7 assumed an additional increase of the pseudorotational bamer due to the anomeric effect, but, in contrast, data for the ring inversion in 2,2-dimethoxyoxane194 indicate a significant decrease of the barrier by 6 9 kJ.mol-l relative to the rings, with the absence of the anomeric effect. The activation bamers in reactions are usually higher than the foregoing figures for AEL, and that explains why it is very difficult to observe the stereoelectronic effect in the reactivity of various conformers of saccharides. In order to assess the magnitude of this effect, systems conformationally much more rigid than those of the majority of saccharides have to be used. Various complex, model compounds have been d e ~ i g n e d ' ~wherein ~ - l ~ ~ the rigidity of a proper structural segment is secured by chemical fixation. It is generally accepted that the transition-state theory provides the most convenient framework for calculation of the rate constants. Its rigorous application, however, needs a knowledge of the energy hypersurface of a

-

-

N. S. Zefirov, Tetrahedron, 33 (1977) 2719-2722. J. I. Seeman, Chem. Rev.,83 (1983) 83-134. D. F. DeTar, J. Org. Chem., 5 1 (1986) 3749-375 1. C. L. Perrin and 0. Nunez, J. Chem. SOC.,Chem., Commun., (1984) 333-334. S. Chandrasekhar, A. J. Kirby, and R. J. Martin, J. Chem. Soc.,Perkin Trans. 2, (1983) 1619- 1626. (196) A. J. Kirby and R. J. Martin, J. Chem. Soc., Perkin Trans. 2, (1983) 1627-1632. (197) A. J. Kirbyand R. J. Martin, J. Chem. Soc., Perkin Trans. 2, (1983) 1633-1636. (198) A. J. Briw, C. M. Evans, R. Glenn, and A. J. Kirby, J. Chem. SOC.,Perkin Trans. 2, (1983) 1637-1640. (191) (192) (193) (194) (195)

118

IGOR

TVAROSKA AND TOMAS BLEHA

given reaction, mostly not available (with the exception of some simple, few-atom systems).From the point of view of the stereoelectroniceffect, it is sufficient to compare its involvement in the ground states versus transition states of various reaction-channels. For instance, in spite of a large absolute value of the anomeric effect, no difference in reactivity might be observed when its contributions to the ground and transition states compensate each other. Estimatesof the anomericeffect in the ground state and transition state, or in intermediates, have been used to explain the different reaction-rates of two anomers in some reactionsby employing the established reaction-mechanism. Especially suitable in this respect are reactions wherein the identical transition state is assumed for the reaction of both anomers. In this case, the axial and equatorial conformersof a given compound react at different rates, owing to a different stabilization of the ground states by the anomeric effect. Acid hydrolysis of 2-oxanyl acetals and of methyl glucosides are probably examples. The stabilization of the axial isomer in the ground state retards its reaction, by enhancement of the reaction bamer relative to the equatorial isomer, which is then hydrolyzed the faster.lS3-lS5 Actually, this reasoning lay behind the first observation of the anomeric effect? Similar stabilization of the reaction intermediates by the anomeric effect explains, qualitatively, the kinetics of conversion of derivatives of 8-D-galacturonic acid into the corresponding a-L-altruronic acid derivat i v e ~ .Hydrolysis '~~ of the a is a few times faster than that of the p anomer of aryl glycoside~,'~~ and this can probably be explained similarly. In this case, the hydrolysis is affected by the initial, proton-transfer equilibrium. Because of the reverse anomeric effect, the relative stability of the resulting conjugate acids is reversed with respect to the neutral molecule, and thepanomer is the more This means that the reactivity of the axial anomer of the conjugate acid increases (and vice versa for the equatorial anomer). The paucity of information on the mechanism of reactions, and on the structure of the transition state, and the role of the anomeric effect in its stabilization, constitutes the main reason why qualitative interpretation of reactivity as shown in the aforementioned examples is still very rare. An alternative, more-popular estimation of the relative reaction-rates of conformers is based on the lone-pair orbital interactions, and their symmetry and energy in the ground state, and could be loosely associated with the perturbation theory of chemical reactivity.200 ( 1 99) P. KovaE, J. Hirsch, I. TvaroSka, R. PalovEik, V. KovaEik, and T. Sticzay, Collect. Czech. Chem. Commun., 41 (1976) 3804-3811. (200) G . Klopman (Ed.), Chemical Reactivity and Reaction Paths, Wiley-Interscience, New York, 1974.

ANOMERIC AND EXO-ANOMERIC EFFECTS

119

2. Lone-Pair Orbital Interactions in Reactivity The lone pairs of oxygen atoms are the least strongly bound atomic orbitals, and thus are localized prevailingly in the highest-occupied MOs. Hence, it is natural that lone pairs play a central role in various interpretations of stereochemical dependence of reactivity in carbohydrates. In this respect, the most successful approach is the so-called antiperiplanar, lonepair hypothesis, derived from study of the hydrolysis of esters and a m i d e ~ . ' ~This ~J~ theory ~ is formally similar to the simple delocalization concept of the anomeric effect involving the back-donation of the oxygen lone-pair into the up antibonding C - X orbital. The fundamental presumption of the theory is that any reactive conformer must have on each oxygen atom a lone pair oriented up to the leaving group. The concept, based on sp3 hybridization on oxygen, was originally developed in order to account for the experimentally observed ratio of products in the oxidation of acetals. At present, this scheme is widely accepted, and it has been applied to the whole range of reactions on the anomeric enter.'^.^^^- 198~201 However, experimental evidence of a marked stereoelectronic effect on the reactivity of saccharidesis very difficult to obtain, because of the flexibility ofthe pyranoid ring. For example, the equatorial isomer can probably not only react through its chair conformations,but also by way of conformations close to the boat forms in which the sp3lone-pair on the ring-oxygen atom is antiperiplanar to the exocyclic C - 0 bond.I0 The largest stereoelectronic control of reactivity observed to date is found in the hydrolysis of the axial and equatorial pair of 1-(2,4-dinitrophenoxy)-9-oxabicyclo(3.3.l)nonane,198where the conformation of the equatorial form is fixed by the ring junction. As a result, the lone pair on oxygen cannot facilitatecleavageofthe C-OR bond. This compound is hydrolyzed lOI3 times slower than the corresponding, axial 2,4-dinitrophenoxy acetal. The antiperiplanar, lone-pair hypothesisis based on the presumption that the lifetime of a tetrahedral intermediate in hydrolysis is shorter than the average time of rotation about the C - 0 bond188;that is, the barrier of conformational transition exceeds the reaction barrier. However, a simple intermediate (hydrogen orthoester) has been detected,202and on this basis, the antiperiplanar, lone-pair theory was questioned.203Similarly, studies on the hydrolysis of saccharide derivative^^@'*^^^ indicated that the departure of

-

(201) V. G. S. Box, J. Heterocycl. Chem., 22 (1984) 891 -905. (202) B. Capon, J. H. Gall, and D. M. A. Grieve, J. Chem. SOC.,Chem. Commun., (1976) 1034- 1035. (203) C. L. Pemn and G. M. L. Arrhenius, J. Am. Chem. Soc., 104 (1982) 2839-2842. (204) L. Hosie, P. J. Marshall, and M. L. Sinnott, J. Chem. Soc., Perkin Trans. 2, (1984) 1121-1131. (205) A. J. Bennet and M. L. Sinnott, J. Am. Chem. SOC.,108 (1986) 7287-7294.

120

IGOR TVAROSKAAND TOMAS BLEHA

the aglycon from the anomeric oxygen atom does not require the conformations wherein a leaving group is in the up position relative to the lone pair on the ring-oxygen atom. Apparently, the interpretation of the reactivity differences based on the selection of only one type of lone-pair interaction is not satisfactory for all cases. Although interactionswith the antibonding orbital are implicitly invoked in the antiperiplanar, lone-pair hypothesis, additional, coexisting, lone-pair interactions are the basis of another qualitative approach to rationalizing many reactions of the monosaccharides.189Jw~201 This approach utilizes the fact that the energy of the most reactive MOs increases with their interactions. The magnitude of orbital interactions can only indirectly be evaluated from experiment, through geometrical changes, or photoelectronic spectra; however, they can be calculated by quantum chemistry. Perturbation MO calculations for dimetho~ymethane,'~~ elaborated in Section VI, showed that the dominant components in the frontierorbitalsof dimethoxymethane are the oxygen lone-pairs combined with the antibonding orbitals of the CH,X group. This is illustrated in Fig. 27 for three highest-occupied (HO) molecular orbitals. The theoretical analysis of orbital interactionsin dimethoxymethane148 is ls0 which predict the supported by photoelectron spectra of lY8-dioxadecalin, decreasing order of lone-pair interactions in the conformational sequence as being (up,up) > (up,sc) > (sc, sc). In all of the conformers, the energy of HOMO exceeds the energy of parent orbitals, and decreases in the same order as the energy of interaction. The enhancement of nucleophilicity of oxygen atoms relative to the noninteracting ones, and its variation with conformation, are the consequences of orbital interactions on the reactivity of molecules having an acetal segment. For example, the increase of electronegativity of the glycosidic oxygen atom by some aglycons decreases the energy of oxygen lone-pairs and their interactions. As a result, changes in the valence geometry can be observed, as, for example, in axial and equatorial aryl-2-oxanylacetals, where a linear correlation was found between the C - 0 bond length and the pK, value of the conjugated acid of the aglycon.206 Calculations148 predicted that lone-pair interactions in the (up,up) conformation corresponding to /3-glycosidesare larger than in the (sc, sc) conformation representing a-glycosides, and thus, the increase of nucleophilicity in the former isomers should be higher. The enhancement of oxygen nucleophilicitycan be demonstrated in several, simple reactions. A suitable measure of nucleophilicity of the 0 - 1 atom is its rate of protonation in (206) A. J. Briggs, R. Glenn, P. G. Jones, A. J. Kirby, and P. Ramaswamy, J. Am. Chem. SOC., 106 (1984) 6200-6206.

('p

8

'9)

FIG.27.- Symmetryof the Three Highest-OccupiedMO Orbitals in Dimethoxymethane.'"

122

IGOR

TVAROSKA AND TOMAS BLEHA

dimethyl sulfoxide, which is systematically higher in &aldopyranoses than in the corresponding a a n ~ m e r sSimilarly, . ~ ~ ~ a higher rate of scission of the C- 1 - H- 1 bond was observed in the e form of 2-methoxyoxanethan in the a form.z0sThis rate is even lower in oxane, where lone-pair interactions are absent. Lone-pair interactionsby the through-space mechanism, and the increase of the nucleophilicity of the participating atoms, have been implicated in mechanistic proposals for the oxidation of acetals, and for some reductions of glyc~loses.'~~ On the other hand, the relative reactivitiesof the anomers in S N processes ~ has been discussed in terms of through-bond interactions of the lone pairs.lWSumming up the available evidence, it is clear that both types of lone-pair interaction must be recognized, in order to rationalize the stereoreactivity of saccharides. According to the perturbation theory of chemical reactivity,200the symmetry and energy of the frontier orbitals determine the stereospecifity and the rate of reaction. We have shown that the mutual interactionsof lone pairs of oxygen atoms, and their interactionswith the properly oriented antibonding orbitals, actually affect several molecular properties of acetals, including the energy of the frontier MO and of orbitals in their vicinity. A more-rigorous account of the stereoreactivity differences should start from the complete picture of electron distribution and energies of molecular orbitals that are connected to relative rate-constants by the relations derived in the perturbation theory of chemical reactivity.z00Quantum chemical calculations made by following this procedure should help to clarify the role of the anomeric effect in the reactivity of saccharides. In closing this short and incomplete, theoretical excursion into the field of reaction stereoselectivity in carbohydrates, we have also to mention that some other aspects in this article have been treated only partially; for instance, the anomeric effect in rings other than six-membered,or in carbohydrate radicals. It is hoped that the present article has shown that many questions on the nature and various manifestations of the anomeric and related effects have already been answered. The problems still remaining to be solved will attract growing attention.zw-z'2We believe that this survey, (207) B. Gillet, D. J. Nicole, and J.-J. Delpuech, J. Chem. Soc., Perkin Trans. 2, (1981) 1329-1335. (208) V. Malatesta and K. U. Ingold, J. Am. Chem. Soc., 103 (1981) 609-614. (209) J. P.Praly and R. U. Lemieux, Can. J. Chem., 65 (1987) 213-223. (210) A. Cosse-Barbi and J. E. Dubois, J. Am. Chem. Soc., 109 (1987) 1503-151 1. (2 11 ) P. Aped, Y. Apeloig, A. Ellencweig, B. Fuchs, I. Goldberg, M. Karni, and E. Tartakovsky, J. Am. Chem. Soc., 109 (1987) 1486- 1495. (212) D. G. Gorenstein, Chem. Rev.,87 (1987) 1047- 1077.

ANOMERIC AND EXO-ANOMERIC EFFECTS

I23

based on theoretical conformational analysis, will contribute to the final solution of the puzzle named the anomeric effect. ACKNOWLEDGMENT In retrospect, the authors gratefully acknowledge the friendly and inspiring atmosphere in the laboratory of Prof. Robert H. Marchessault at the UniversitC de MontrCal, where, during their stay in the late seventies, the first thoughts of writing this article crystallized.

This Page Intentionally Left Blank

ADVANCES IN CARBOHYDRATE CHEMISTRY AND BIOCHEMISTRY. VOL. 47

13C-NUCLEAR MAGNETIC RESONANCE-SPECTRAL STUDIES OF THE INTERACTIONS OF METAL IONS WITH CARBOHYDRATES: USE OF RELAXATION PROBES BY KILIAN DILLAND R . DOUGLAS CARTER* Department of Chemistry. Clemson University. Clemson. South Carolina 29634

1. Introduction ......................................................... I1 . General Considerations Concerning Carbohydrate Structure.................. 1. Aspects of Carbohydrate FunctionalGroups ............................ 2 . Size Requirements ................................................. 3. Charge Requirements ............................................... I11. General Considerations Concerning Metal Ions ............................ 1 . Coordination Spheres ............................................... 2 . Oxidationstates ................................................... 3. Diamagnetic Species................................................ 4 . Paramagnetic Species: Shift Reagents and Relaxation Probes ............... IV . Uses of Mn2+and Gd” ................................................ 1 . Medical Uses of Gd” and Mn2+...................................... 2. Exchange of Gd” and Mn2+for Ca2+and Mg2+in Biological Systems ....... 3. Electron-Nuclear Relaxation Mechanisms for Carbon Atoms: Use of Data in Order to Calculate Metal Ion-Carbon Atom Distance Information ......... V . Gd3+and Mn” Interactions with Carbohydrates ........................... 1. Inositols .......................................................... 2 . djluconamides ................................................... 3. Simple Monosaccharidesand Glycosylated Amino Acids .................. 4 . Complex Glycopeptides ............................................. 5 . N-Acetyl-cr-D-neuraminic Acid ........................................ VI . Conclusions .........................................................

125 127 127 127 128 128 128 129 129 130 135 135 136 136 137 137 144 150 155 158 165

I . INTRODUCTION

The interaction of metal ions with carbohydrates has been a topic of interest for a number of years. In any biological system. calcium ions and

* Present address: Westvaco Corporation. Laurel. MD 20707 . 125

Copyright 0 1989 by Academic Ress. Inc. All rights of reproduction in any form m d .

126

KILIAN DILL AND R. DOUGLAS CARTER

carbohydrate molecules found in the same vicinity may be involved in calcium transport, structural support in membrane systems, cell - cell adhesion, agglutination of polysaccharides, binding of glycoproteins to the surfaces of cells, the transmission of nerve impulses, and regulation of the Magnesium ions have also been found in intracellular fluid, and are responsible for the stimulation of enzymic a~tivity.~ Furthermore, certain studies have focused on metal ion-carbohydrate interactions in the complex oligosaccharides of brain gangliosides and of glycophorin, the major sialoglycoproteinof the red-cell membrane.4-6Thus, the interaction of Ca2+ and MgZ+ ions with carbohydrate residues is of immense biological In order to investigatethese interactions, I3C-n.m.r.spectroscopyhas been used to study the binding of the metal ions to various carbohydrate residues and glycopeptides,and to extract information on the specificbinding-sites of the carbohydrate residues, in order to define further how and why metal ions interact with certain residues in biological systems. (1) W. J. Cook and C. Bugg, in B. Pelham and W. Goldblum (Eds.),Metal-Ligand Interaction in Organic Chemistry andBiochemistry, Part 2, Reidel, Holland, 1977,pp. 23 1 -256. (2) H. Sigel, Metal Ions in Biological Systems: Calcium and Its Role in Bioloe, Vol. 17,

Marcel Dekker, New York, 1984. (3) R. C. Bohinski, Modem Concepts in Biochemistry, Allyn and Bacon, Boston, 1973. (4) L. 0.Sillerud,J. H. Prestegard, R. K. Yu,D. E. Schafer, and W. H. Konigsberg, Biochemistry, 17 (1978) 2619-2628. (5) M. E. Daman and K. Dill, Carbohydr. Res., 1I I (1983) 205-214. (6) R. Prohaska, T. A. W. Koerner, Jr., 1. M. Armitage, andH. Furthmayr,J. Biol. Chem., 256 (198 I ) 578 1 -579 1. (7) H. A. Tajmir-Riahi, Carbohydr. Rex, 122 (1983) 241-248. (8) M. L. Dheu-Andries and S . Wrez, Carbohydr. Rex, 124 (1983) 324-332. (9) R. H. Kretsinger, in Ref. 1, pp. 257-263. (10) B. Birdsall, J. Feeney, and A. M. Guilian, in Ref. 1, pp. 319-324. ( I 1 ) S. J. Angyal, C. L. Bodkin, J. A. Mills, and P. M. Pojer, Ausf, J. Chem., 30 (1977) 1259- 1268. (12) J. K. Barrie and M. T. Kelso, Aust. J. Chem., 34 (1981) 2563-2568. (13) S. J. Angyal, Aust. J. Chem., 25 (1972) 1957- 1966. (14) L. W. Jaques, E. B. Brown, J. M. Barren, W. S. Brey, Jr., and W. Weltner, Jr., J. Biol. Chem., 252 (1977) 4533-4538. (15) W. J. Evans and V. L. Frampton, Carbohydr. Res., 59 (1977) 571 -574. (16) L. F. Jacques, J. B. Macaskill, and W. Weltner, Jr., J. Phys. Chem., 83 (1979) 1412- 1421. (17) M. F. Czarnierki and E. R. Thornton, Biochem. Biophys. Res. Commun.. 74 (1977) 553-588. (18) J. P. Behr and J. M. Lehn, FEBSLett., 22 (1972) 178-180. (19) J. P. Behrand J. M. Lehn, FEBSLert., 31 (1973) 297-300. (20) L. F. Jaques, B. F. Riesco, and W. Weltner, Jr., Curbohydr. Res., 83 (1980) 21 -32. (21) C. Long and B. Mouat, Biochem. J., 123 (1971) 829-836.

INTERACTIONS OF METAL IONS WITH CARBOHYDRATES

127

11. GENERAL CONSIDERATIONS CONCERNING CARBOHYDRATE STRUCTURE 1. Aspects of Carbohydrate Functional Groups It had been shown by Angyal and Davies2’ that as few as three neutral oxygen atoms suffice to define a binding site for metal ions if these oxygen atoms are in the correct steric arrangement. In carbohydrates, this may involve neutral hydroxyl groups throughout the molecule, or carboxyl groups attached to the molecule, such as are found in N-acetylneuraminic acid and similar molecules, or both. Because not all polyols complex the same metal ions to the same extent, other factors must also play a Indeed, it has been found that there are certain steric requirements for complex-formation.~4~z5 Moreover, the size and charge of the cation also appear to play a vital role in complex-formation.z6.27 General aspects of metal ion-carbohydrate binding are discussed in detail by Angyal in this volume.2E

2. Size Requirements There are two steric arrangements that have been found that will afford a binding site for a metal ionJ4One is a group of three syn-axial oxygen atoms on a six-membered ring, as in cis-inositol (see later), although this arrangement is relatively rare, The second consists of three oxygen atoms on three consecutive carbon atoms; these oxygen atoms should be so oriented that the first and second are gauche clockwise, and the second and third are gauche anticlockwise. This is most commonly found in the sequence of an axial (a), an equatorial (e),and an axial hydroxyl group on a six-membered ring in a chair conformation. This type of arrangement, according to the CahnIngold- Prelog terminology is referred to as an m,p arrangement. Although the triaxial site and the m,p arrangementsdisplay similar geometries with three oxygen atoms at a distance of 295 pm, they differ in the arrangementof the orbitals containingfree electron pairs and in the ability to bring the oxygen atoms closer These differences have led to different complexation of some metal ions.

-

(22) (23) (24) (25) (26) (27) (28)

S. J. Angyal and K. P. Davies, Chem. Cornmun.,(1971) 500-501. J. A. Mills, Biochem. Biophys. Res. Cornmun.,6 (1961) 418-421. S. J. Angyal, D. Greeves, and V. A. Pickles,-Curbohydr. Rex, 35 (1974) 165-173. S. J. Angyal, D. Greeves, and J. A. Mills, Aust. J. Chem.. 27 (1974) 1447- 1456. S. J. Angyal, Tetrahedron, 30 (1974) 1695- 1702. D. A. Hanna,C.Yeh,J. Shaw,andG. W.Everett, Jr.,Biochemtstry,22(1983)5619-5626. S. J. Angyal, Adv. Carbohydr. Chem. Biochem., 47 (1989) 1-43

128

KILIAN DILL AND R. DOUGLAS CARTER

3. Charge Requirements Indeed, the size of the metal ion has been found to be a factor in binding. The a - e- a site on a six-membered ring complexesonly cations having ionic >-80 pm. The triaxial site, found on cis-inositol, prefers catradii26,29 ions26*30 that have ionic radii B 60 pm, but < 100 pm. In the binding of the metal ion to the carbohydrate moiety,27charge also appears to play a role, and this is related to a differencein solvation enthalpy in polar solvents. Oxygen donor atoms displace with more difficulty the solvation shell for a trivalent ion in a polar solvent than for a divalent ion. This difference is removed when the study is performed in a nonpolar solvent. Consequently, it would appear that, for studies of carbohydrates in aqueous solution, the charge on the molecule must be considered.

111. GENERAL CONSIDERATIONS CONCERNING METALIONS

1. Coordination Spheres

Because the metal ions typically used to study binding to carbohydrates are from the first-row transition-metalsand from the lanthanides, it might be interesting to compare their tendencies in bonding. KarracheI3,’ summarized these differences, and obtained the following information. Lanthanide ions typically have bonding electrons in the 4f orbitals, and their ionic radii are typically 106-85 pm, whereas their counterparts in the first-row transition-metals have their electronsin the third subshell,with ionic radii of 75 to 60 pm. Lanthanide ions typically have higher coordination numbers (6- 9),and those of transition-metal ions are typically 4 or 6, although coordination numbers from 6 - 9 have been documented. Bonding in transition metals is typically ionic in nature, with no metal-ligand orbital interaction, and the preference for ligands is determined by the electronegativity of the ligand; F- > OH- > H20. On the other hand, the first-row transitionmetal ions are often covalent, with strong metal-ligand orbital interaction. Bond strengths may then be determined by the orbital interacti~n.’’,’~

-

-

(29) M. E. Daman and K. Dill, J. Magn. Reson., 60 (1984) 118- 121 (30) R. D. Carter and K. Dill, Inorg. Chim. Acta, 125 (1986) L ~ - L I1, Inorg. Chim. Acta, 108 (1985) 83-86. (31) D. G. Karracher, J. Chem. Educ., 47 (1970) 424-432. (32) R. E. Lenkinski, in L. J. Berliner and J. Reuben (Eds.), Biological Magnetic Resonance, Chapter 3, Plenum, New York, 1984, p. 6. ( 3 3 ) J. E. Huheey, Inorganic Chemistry: Principles of Structure and Reactivity, 2nd edn., Chapter 16, Harper & Row, New York, 1978.

INTERACTIONS OF METAL IONS WITH CARBOHYDRATES

129

2. Oxidation States

+

By far the most common oxidation state among the lanthanides is 3, although some 2 ions have also been f o ~ n d .For ~ ~all, the ~ ~ 3 cations of the lanthanides, the electron configuration is identical, namely, 4P5d06s0. First-row transition-metal ions, on the other hand, have variable oxidation states, dependent on their propensity to attain stable, orbital configurations, ranging from 1 to 7 oxidation states. These comparisons should be taken into account when substituting lanthanide ions for first-row transition-metal ions in order to induce chemical shifts or change the relaxation behavior of a n.m.r. spectrum to elicit ion sequestering and structural information (see later).

+

+

+

+

3. Diamagnetic Species Such diamagnetic metal ions as Ca2+ are quite prevalent in biological systems. Nuclear magnetic resonance spectroscopy has been used to study the interaction ofthese cations with various ligands in order to determine the mode and strength of binding.14*22,26,35*36 Angya126showed that downfield shifts occur in the spectra of epi-inositol upon addition of calcium chloride. These downfield shifts, known as diamagnetic shifts, are due to the effect that the charge on the ion has on the electric field surrounding the nucleus under investigation. The larger the charge on the ion, the larger the effect on the shift of the resonance. Indeed, based on theory developed by B~ckingham,’~ Angyal calculated26the shifts that result from the effects of lanthanum ions on epi-inositol. The effect of added Ca2+on the chemical shifts of sialic acid (see later) has also been used to acquire structural inf~rmation.’~ Buckingham’s theory37can be used to obtain quantitative information about diamagnetic metal-ion binding to ligands. The effect of proton shielding by the chelation of Ca2+is obtained from Eq. I.

Ad = k 4.8(cos 8)/r2

(1) where r is the length (in Angstrom units) of the vector from the charge to the middle of the C-H bond, 8 is the angle between that vector and the C-H bond, and k is a proportionality constant, obtained empirically. In the equation, either r or 8 must be known a priori; because the value is often not (34) P. N . Yocom, Adv. Chem. Ser., 71 (1967) 51. (35) E. B. Browon, W. S. Brey, Jr., and W. Weltner, Jr., Biochim. Biophys. Actu, 399 (1975) 124-130. (36) S. J. Angyal, D. Greeves, and L. Littlemore, Aust. J. Chem., 38 (1985) 1561 - 1566. (37) A. D. Buckingham, Can. J. Chem., 38 (1960) 300-307.

130

KILIAN DILL AND R. DOUGLAS CARTER

known, a reasonable assumption must be made about the geometry of metal binding to the ligand. Moreover, as may be seen from Eq. I, small errors in the measurementof the chemical shift or in approximation of the angle 8 can lead to large errors in distance measurement. Because shifts are often small (60.2 p.p.m.), this presents a serious problem; in an effort to circumvent this, paramagnetic metal ions known to mimic their diamagnetic counterparts in their chemical behavior have been used to induce larger shifts in the spectra of compounds of interest.

4. Paramagnetic Species: Shift Reagents and Relaxation Probes Paramagnetic species, primarily metal ions, can cause large changes in n.m.r.-spectral patterns and in the relaxation phenomenon, due to the interaction of the electron with the nucleus. Based on their effect on n.m.r.-spectral parameters, these paramagnetic species have been classified into two categories: shift reagents, which cause variations in the chemical shifts of compounds, and relaxation reagents, which affect the relaxation times of molecules containing these paramagnetic metal-ions. The use of these paramagnetic metal-ions has become quite common in probing the molecular structure of a variety of organic, inorganic, and biological compounds. Such n.m.r. shift reagents as NiZ+,Coz+, Fez+,low-spin Fe3+,and lanthanide(II1) ions (other than Gd3+, a relaxation reagent) are characterized by having very fast electron-spin relaxation times (T,, 3 lO-'Os), which prevents them from affecting n.m.r. relaxation phenomenon (see later), and an anisotropic distribution of charge about the The unpaired electrons of the paramagnetic species interact with the nucleus under investigation by causing n.m.r. hyperfine shifts. These hyperfine shifts are a result of dipolar interactions between the electron and nucleus, pseudocontact shifts, and by electron spin delocalizationthrough chemicalbonds, known as contact shifts. Pseudo-contact shifts can arise only in systemswhere the magnetic susceptibility of the paramagnetic center is anisotropic. This interaction appears to be dominant in certain transition-metal ions (Coz+), as well as such rareearth elements as4' Dy3+,Er3+,and Yb3+. The chemical shift from the dipo-

(38) K. Wuthrich, NMR in Biological Research: Peptides and Proteins, American Elsevier, New York, 1976. (39) R. A. Dwek, R. J. P. Williams, and A. V. Xavier, in Ref. 2, Vol. 4, 1974, pp. 61 -210. (40) 0. Jardetzky and G. C. K. Roberts, NMR in Molecular Biolofl, Academic Press,New York, 1981. (41) B. Bleany, C. M. Dobson, B. A. Levine, R. B. Martin, R. J. P. Williams, and A. V. XaGer, Chem. Commun., (1972) 791-793.

INTERACTIONS OF METAL IONS WITH CARBOHYDRATES

131

lar term has been expressed by Bleaney and coworkers4*as in Eq. 2. (2) where D, and D, are temperature-dependentconstants that vary according to the metal ion used, r is the distance between the electron spin and the nuclear spin, and 6' and 4 are the angles in the spherical polar coordinate system more fully describing the proximity of the metal ion to the nucleus under consideration. Axial symmetry, an assumption often involved, although its validity is often debated,32,42.43 reduces this equation, such that D2 = 0. Thus, pseudo-contact shiftscan be calculated only if the geometryof binding is known. Likewise, distance information can be difficult to extract from observed shifts without a priori knowledge of the geometry of binding. Fermi contact shifts may also contribute to the hyperfine shift observed in n.m.r. spectra.44These shifts arise from the delocalization of electron spin density from the extended orbital of the metal ion to the orbitals of the ligand. This shift is primarily dependent on the contact interaction constant for a given nucleus.45Contact shifts yield no distance information or structural information. Information on structural parametersthen can, or must be, obtained from shift parameters using paramagnetic probes that have little or no contribution to shift parameters from the contact mechanism. Furthermore, by making some general assumptions about the molecular geometry of binding before evaluating the data, or by obtaining this information through electron-spin resonance studies, structural information may be extracted from these induced shift studies. This has been done quite extensively for carbohydrate moieties by Angyal and coworker^.^^^"-^^^^ Paramagnetic relaxation reagents (pa.r.r.), such as Gd3+,Mn2+,Cu2+,and Cr3+,and other paramagnetic metal ions with isotropic charge-distribution, affect the nuclear relaxation rates, T, and T2,of species under investigation by n.m.r. spectroscopy through electron-nuclear spin - spin coupling. There are two classes of pa.r.r.s which are characterized by the predominant mechanism of relaxation enhancement. Nonspecific pa.r.r.s affect relaxation through an outer-sphere mechanism. This type of reagent is used primarily to shorten relaxation times and thus increase the speed of data acquisition, or to eliminate nuclear Overhauser effects (n.0.e.s) in order to allow quantitative analysis of data. Cr(acac), is (42) I. D. Campbell, C. M. Dobson, and R. J. P. Williams, Proc. R.Soc. London, Ser. A , 345 ( 1 975) 41 -59. (43) D. G. Agresti, R. E. Lenkinski, and J. D. Glickson, Biochem. Biophys. Res. Commun., 76 (1977) 71 1-719. (44) E. Fermi, Z . Phys., 60 (1930) 320-333. (45) J. P. Jesson, J. Chem. Phys., 47 (1967) 579-581.

132

KILIAN DILL AND R. DOUGLAS CARTER

probably the best known example of these types of reagent.46-56 To be useful these pa.r.r.s must be nonlabile. A second type of pa.r.r. is a specific, paramagnetic relaxation reagent. Here, a complex is formed between the paramagnetic species and the molecule of interest. In these complexes, spin-lattice relaxation is enhanced through intramolecularelectron-nuclear dipolar interaction, or scalar coupling interactions, or both. Assuming that the dipole contribution is dominant, sensitive, specificdistanceinformation can be obtained, as the dipolar electron-nuclear interaction has a dependence on the sixth power between the separation of the paramagnetic species from a given n u c l e u ~ (see ~~.~~ later). These pa.r.r.s have found increasing uses in biological systems because of the ability of certain Paramagnetic ions, such as Gd3+and Mn2+, to mimic Ca2+and Mg2+(see earlier), and they allow information as to the binding of metal ions to biologically important substances to be determined.5,29,30,59-67a (46) G. N. LaMar, Chem. Phys. Lett., 10 (1971) 230-232. (47) R. Freeman, K. G. R. Pachler, and G. N.LaMar, J. Chem. Phys., 55 (1971) 4586-4593. (48) 0. A. Gansow, A. R. Burke, and G. N. LaMar, J. Chem. Soc., Chem. Cornmun..(1972) 456-457. (49) 0.A. Gansow, A. R. Burke, andG. N. Vernon, J.Am. Chem. SOC.,94 (1972)2550-2552. (50) L. F. Farnell, E. W. Randall, and A. I. White, J. Chem. Soc., Chem. Commun., (1972) 1 159- 1 160. (51) A. J. DiGioiaand R. L. Lichter, J. Mugn. Reson., 27 (1977)431-438. ( 5 2 ) G.C. Levy,J. J. Dechter,andJ. Kowalewski,J. Am. Chem. Soc., 100(1978)2308-2314. (53) G. C. Levy,U. Edlund, and J. G. Hexen, J. Mugn. Reson., 19 (1985) 259-262. (54) G. C. Levy,U. Edlund, and C. E. Holloway, J. Mugn. Reson.. 24 (1976) 375-387. (55) G. C. Levy and J. D. Gargioli, J. Mugn. Reson., 10 (1973) 231 -234. (56) G. C. Levy and R. A. Komoroski, J. Am. Chem. Soc., 96 (1974) 678-681. (57) I. Solomon, Phys. Rev., 99 (1955) 559-565. (58) N. Bloembergen, J. Chem. Phys., 27 (1957) 572-573. (59) J. Gariepy, K. Lewis, I. D. Kuntz, B. D. Sykes, and R. S. Hodges, Biochemisrv, 24 ( 1 985) 544- 550. (60) W. L. Bigbee and F. W. Dahlquist, Biochemistry, 13 (1974) 3542-3549. (61) H. Sterle, M. Braun, 0. Schmitt, and H. Feichfinger, Curbohydr. Res., 145 (1985) 1 - 11. (62) H. K. Lannom, K. Dill, M. Denarit, J. M. Lacombe, and A. A. Pavia, Int. J. Pep?.Protein Res., 27 (1986) 67-78. (63) R. D. Carter, K. Dill, J. M. Lacombe, and A. A. Pavia, J. Protein Chem., 4 (1985) 363-373. (64a) M. E. Daman and K. Dill, Curbohydr. Rex, 102 (1982) 47-57. (64b) M. E. Daman and K. Dill, Curbohydr. Rex, 132 (1984) 335-338. (65) K. Dill, M. E. Daman, R. L. Batstone-Cunningham, J. M. Lacombe, and A. A. Pavia, Carbohydr. Rex, 123 (1983) 123-135. (66) K. Dill, M. E. Daman, R. L. Batstone-Cunningham, B. Ferrari, and A. A. Pavia, Curbohydr. Rex, 123 (1983) 137-144. (67a) K. Dill, M. E. Daman, R. L. Batstone-Cunningham, M. Denarit, and A. A. Pavia, Curbohydr. Rex, 124 (1983) 1 1 -22.

INTERACTIONS OF METAL IONS WITH CARBOHYDRATES

133

The interaction of the unpaired electron on the paramagnetic metal ion with that of the nuclear spin is analogous to the nuclear-nuclear spin complex commonly observed in nuclear magnetic resonance spectroscopy. Thus, the electron-spin relaxation (T, ) affects the nuclear-spin relaxationrates ( Tl and T2).Eqs. 3 and 4, describing the contribution to the nuclear relaxation rate from the component spin on the paramagnetic species were derived by Solomon5’ and Bl~embergen:~~

where it was assumed that ox< as, and outer-sphere relaxation was neglected. In Eqs. 3 and 4, T; and T5 are the longitudinal and transverse relaxation-timesdue to electron-nuclear spin interaction. yx is the gyromagnetic ratio of spin X being observed by n.m.r. spectroscopy, S is the total electron spin, p is the Bohr magneton, r is the electron-nuclear distance, TC1 and tc2are the effective correlation times for dipolar interaction, A is the scalar coupling constant (rad s-l), oxand w, are the Larmor frequenciesfor the nucleus and electron, and z1and z2are the correlation times for the scalar coupling,which, neglectingchemical exchange, are given by the longitudinal and transverse electron-spin relaxation-time (zl = Tl,; ‘52 = T2,). The correlation times for dipolar interaction, zCland zC2,can be determined by the combination of the rotational correlation for 7, and the electron-spin relaxati~n-time.~~ 1/zci =

l/zc

+ l/Tie

(5)

l/zc2 =

1/zc

+ 1/T2e

(6)

For paramagnetic relaxation-reagents,the electron-spin relaxation-rates are long (Tie,2e > s), so that, for small molecules, zcl = z, and zC2= 7,. Because this is the same correlation-timeas in all other terms of the nuclearspin relaxation-time T I ,where l/Tlcob) = l/TFD

+ l/TFA+ 1/TF + l/TSR+ l/TY

(7)

134

KILIAN DILL A N D R. DOUGLAS CARTER

and l/T2(ob)= l / T f D

+ l / T y A+ l / T y + l/TgR+ 1/T;,

(8)

the relaxation for nuclei near the paramagnetic center will be dominated by the electron-nuclear spin interaction (Ti) due to the large gyromagneticratio of the electron (5 12 times that of a proton). If the electron-spin relaxation is short, the dipolar contribution to nuclear relaxation will be negligible, even for nuclei near the paramagnetic center. Finally, in most applicationsof the equation, o,z, > 1 and w,z, << 1, so that the ratio T;/T; = 716 when only a dipolar interaction is considered. For small molecules, and in the absence of paramagnetic species, the values of l/T,(oh) and l/Taob) will usually be dominated by the l/TyDand l/TFDterms, respectively. Relaxation values obtained by Dais and P e r l i r ~ from ~ ’ ~ the dipolar interactions (DD) were used in ascertaining the solution structures of carbohydrates. Eqs. 3 and 4 are made up of two parts, the first involving the dipolar contribution to the relaxation rate, and the second dealing with the scalar contribution to the relaxation rate. So far, only the dipolar interaction has been considered; however, scalar coupling can also have a significant effect. On coupling Eqs. 3 and 4, it may be seen that scalar coupling will contribute to the transverse electron-nuclearrelaxation (T;) before it can be observed in the longitudinal electron-nuclear relaxation (Ti). This scalar contribution can be observed spectroscopically by extensive broadening of nuclei which are scalar-coupled to a paramagnetic metal-ion. The selective-broadeningtechnique, in which metal-ion binding to various ligands is studied through the use of paramagnetic metal ions, has been used extensively for studying the binding of metal ions to pep tide^^^,^^ and carbohydrate molecule^.^^^*-^^^^-^^ This technique allows distance information to be calculated by separating the electron-nuclear spin-relaxation contribution from the other relaxation mechanisms contained in the measured T, and T, (see Eqs. 7 and 8). This can be done by subtractingthe relaxation in the absence of the paramagnetic species from the relaxation rate as seen in Eqs. 9 and 10. l/T;

=

1/T, (with metal ion) - l/T, (without metal ion)

(9)

l/T;

=

1/T, (with metal ion) - l/T2 (without metal ion)

(10)

The resulting 1/ T; and 1/ T; values are then evaluated in terms of Eqs. 3 and

(67b) P. Dais and A. S. Perlin, Adv. Curbohydr. Chem. Bzochem., 45 (1987) 125- 168. (68) K. Dill, M. E. Daman, E. Decoster, J. M. Lacombe, and A. A. Pavia, Inorg. Chirn. Actu, 106 (1985) 203-208. (69) K. Dill, H. K. Lannom, J. M. Lacombe, M. Denarit, and A. A. Pavia, Curbohydr. Rex, 142 (1985) I 1 -20.

INTERACTIONS OF METAL IONS WITH CARBOHYDRATES

135

4, which are first so simplified that

Severalassumptionsare implicit in the selective-broadeningexperimentand in the use of Eq. 12to interpret these results. These are that the metal ions or general paramagnetic species are in fast exchange on the n.m.r. time-scale, and that only the dipolar term dominates the transverse-relaxation time. Esperson and M a r t i ~ ~ ~ showed O . ~ ~ that it is necessary to confirm that these assumptions are indeed valid before undertaking evaluation of the data based on these assumptions. Frequency- and temperature-dependentstudies may be employed to confirm that the system is in the fast-exchange limit on the n.m.r. times ~ a l e . ~The ~ * second ~ ~ * ~assumption ’ may be validated by ~ornparing’~.~~ the ratio T ; / T ; . As shown earlier, if the dipolar term is the only significant factor, the ratio Tq / T ; will be 7/6 = 1.17: 1 .OO. If values are obtained that are greater than this value, there is some contribution from electron-nuclear scalar coupling.A third implicit assumptionis that there is only one (unique) binding site on the molecule for the metal ion. IV. USESOF Mn2+AND Gd3+ 1. Medical Uses of Gdw and MnZ+

Because of their large numbers of unpaired electrons, Gd3+andMn2+have been consideredas relaxation agents to be used in magnetic resonance imaging (m.r.i.).72-77 Gd3+is used to the greatest extent, because of the inherent (70) W. G. Esperson and R. B. Martin, J. Phys. Chem., 80 (1976) 164- 184. (71) W. G. Esperson and R. B. Martin, J. Am. Chem. Soc., 98 (1976) 40-44. (72) V. M. Runge, J. A. Clanton, C. M. Lukehart, C. L. Partain, and A. E. James, Jr., Am. J. Radiol., 141 (1983) 1209-1215. (73) V. M. Runge, R. G. Stewart, J. A. Clanton, M. M. Jones, C. M. Lukehart, C. C. Partain, and A. E. James, Jr., RadioIoD, 147 (1983) 789-791. (74) V. M. Runge, J. A. Clanton, W. A. Herzer, A. C. Price, H. J. Weinman, and A. E. James, Jr., NoninvasiveMed. Imag., 1 (1984) 137- 147. (75) V. M. Runge, M. A. Foster, J. A. Clanton, M. M. Jones, C. M. Lukehart, J. M. S. Hutchison, J. R. Mollard, F. W. Smith, C. L. Partain, and A. E. James, Jr., Radiology, I52 (1984) 123-126. (76) V. M. Runge, J. A. Clanton, M. A. Foster, F. W. Smith, C. M. Lukehart, M. M. Jones, C. L. Partain, and A. E. James, Jr., Invest. Rad.. 19 (1984) 408-415. (77) V. M. Partain, J. A. Clanton, A. C. Price, W. A. Herzer, C. J. Wehr, C. M. Lukehart, C. C. Partain, and A. E. James, Jr., Physiol. Chem. Phys. Med. NMR, 16 (1984) 113- 122.

136

KILIAN DILL AND R. DOUGLAS CARTER

stabilityofthe complexes,a requirement when injected into human subjects. The complex most commonly used is the Gd3+complex with (diethylenetrinitri1o)pentaaceticacid (DTPA).77 2. Exchange of Gd3+ and Mn2+for Ca2+and Mg2+in Biological Systems As mentioned earlier, the ionic radii of Gd3+and Ca2+,and of Mn2+and Mg2+, are rather similar; furthermore, so are their coordination numbers. Therefore, it would seem plausible that they may readily be exchanged in biological systems. Mn2+can easily be substituted for Mg2+in Mg2+dependent enzymes, and the resulting enzymes are totally functional. Hence, this substitution has allowed researchers to map out enzyme sites by e.s.r. methods or by the use of Mn2+asa relaxation probe.39,78.79 The only problem is that large doses of “free” Mn2+are toxic in humans. Like many lanthanides,Gd3+can be, and has been, used2as a replacement probe for Caz+.This substitution has not been as prevalent, or as useful, as that of MnZ+for biological studies. 3. Electron-Nuclear Relaxation Methods for Carbon Atoms: Use of Data in Order to Calculate Metal Ion -Carbon Atom Distance Information

As mentioned earlier, there are two electron-nuclear mechanisms that result in relaxation of the carbon nucleus: dipolar and scalar. A dipolar mechanism can be thought of as a through-space effect, whereas a scalar mechanism is a through-bond effect. Therefore, it might be envisaged that the scalar mechanism is limited, because the transmission of electron density is rather restricted through sigma bonds. However, it does occur, and quantitative distance information is difficult to obtain from this mechanism. Eqs. 3 and 4 show the electron-nuclear relaxation-rates for the spinlattice (longitudinal) and spin - spin (transverse)relaxation-rates. In both of these equations, the first terms reflect the dipolar term, and the second part reflects the scalar interaction. The dipolar term contains a distinct r6distance term between the electron of the metal atom and the respective carbon atom; in contrast, the scalar term has none. It may be noted that the scalar term for ( T ;)- differs considerably from that for ( T ; ) - l . In fact, because of the difference, and the fact that z2/( 1 &f) is exceedingly small, the scalar contribution to T ; can be neglected, and T ; data may be used directly to calculate metal ion-carbon atom distances. However, the scalar contribution to T ; cannot be neglected, and the value varies from system to system. Therefore, before any attempts are

+

(78) A. S. Mildvan and R. K. Gupta, Methods Enzymol., 496 (1978) 327-359. (79) S. J. Opeila and P. Lu, NMR in Biochemistry, Marcel Defier, New York, 1977.

INTERACTIONS OF METAL IONS WITH CARBOHYDRATES

I37

made to use any line broadening of data in order to calculate the distance information, it must be established that the dipolar term dominates (see later). Therefore, in theory, these equations can be used directly to calculate carbon atom - metal ion distances, provided that such information as the complex correlation time is known exactly. The concept that the complex geometry is not much different for the uncomplexed system may be in error. These matters require a little guess-work, or they may be determined from relaxation parameters of a complex containing a diamagnetic species (La3+ for Gd3+). Other parameters to be concerned with are whether (a) the system is in fast exchange, (b) outer-sphere relaxation plays any role, and (c) there is more than one metal-ion binding-site. Whether the system is in fast exchange can be determined by measurement of Tf.or T ; at various field-strengths, or of temperature dependen~e.~’ Outer-sphere relaxation is often neglected, or it can be corrected for (with various degrees of success52).If two “strong” binding-sites exist, problems may occur in the calculations that may result in “no” unique binding-structure. If the foregoing problems can be overcome, and a metal-ion bindingstructure based on “exact calculations” is not wanted (or cannot be determined), the relative slopes of (T;)-I or (T;)-’ vs. [metal-ion] plots may be used to determine distance and obtain geometric information. This requires that a metal-ion- atom distance is known, and this is then used as the relative standard from which to calculate others. The use of this method is quite prevalent.80Then, from the TY or Tg values, relativedistance information can be gained by using the following equations.

V. Gd3+AND Mn2+INTERACTIONSWITH CARBOHYDRATES 1. Inositols

Inositols provide a good starting-point for studying the binding of such metal ions as Gd3+and Mn2+to carbohydrates, because each is unique, they are structurally related to carbohydrates, and they do not contain other functional groups (carboxyl, amino) which may also interact with metal (80) H. Sterk, M. Braun, and 0. Schmut, Curbohydr. Res., 145 (1985) 1 - 1 1

KILIAN DILL A N D R. DOUGLAS CARTER

138

ions. These compounds may also be used to elucidate the dominant mechanism that exists in the T ; relaxation process for carbon atoms of polyols. The two inositols that we utilized in our studies were epi-inositol(1) and cis-inositol(2).The latter was a kind gift from Dr. Stephen J. Angyal. These two compounds were used because the configuration of the hydroxyl groups provides various sizes of "cavity" for metal-ion binding. As may be seen, these related compounds differ only in the position of one hydroxyl group (on C-6). Published work by Angyal and coworkers indicated that there is a metal-ion size-requirement for binding to these cavities. Metal ions having

HO

OH

Ho OH

I

I

-

2

radii of 100 pm will readily fit into a pocket provided by the hydroxyl groups on C-2, C-3, and C-4 of cis- or epi-inositol. Metal ions with radii of <80 pm are too small for this binding pocket, but will readily fit into the cavity provided by the triaxial hydroxyl groups of cis-inositol. Apparently, charge plays little part, if any, in the binding. One way in which to establish the mechanism that dominates the T ; relaxation process for a given carbon atom is by determination ofthe T ; / T ; ratio for that carbon atom. If a single metal-binding site exists, and the metal ion - ligand interaction is in fast exchange, a limiting value of 1.17 for the T p / T ; value can be obtained if the dipolar mechanism dominates the T ; r e l a ~ a t i o n . ~Values ~ , ~ ' of 1.75 and 2.35 indicate a scalar contribution of 33 and 50%, respectively, to the T ; relaxation p r o c e ~ s . ~ ~ . ~ ' epi-Inositol is a unique compound for determining whether a dipolar mechanism dominates the T ; mechanism, and hence, line-broadening experiments (of directly bonded atoms) may be used to obtain quantitative Gd3+-ligand distance information. This occurs because the hydroxyl groups on C-2, C-3, and C-4 provide a binding site for Gd3+ion (- 100 pm radius). Figs. 1 and 2 show the effects of added Gd3+and MnZ+on the T ; values of epi-inositol. Clearly, no differentialeffect is seen on addition of MnZ+,but a substantial difference is observed for C-2, C-3, and C-4 over C-1, C-5, and C-6 when Gd3+is added. This shows conclusively that epi-inositolcontain a unique binding site for Gd3+ (3) (utilizing 0-2, 0-3, and 0-4), but not for Mn2+.

INTERACTIONS OF METAL IONS WITH CARBOHYDRATES

139

160 140 120

--

-‘L-

100

‘vl

Y

80 60 40 20 2

4

6

8 1 0 1 2

lo3 ( M ) FIG. 1 .-The Effect ofAdded Gd” on the I3C T ; Values of epi-Inositol. [The concentration of epi-inositol was 1 M (in H,O) at pH - 7. The numbers in the graph refer to the specific carbon atoms of epi-inositol, as depicted in its structure. Measurements of T, were made by using the partially relaxed, Fourier-transform (p.r.F.t.) method (1 80-7-90), with eleven 7 values. Several hundred accumulations were collected per 7 value (taken from Ref. 29).] [Gd”]

x

-

3

The effectsof Mn2+on the T ; values of epi-inositol(seeFig. 2) are dramatically diminished from those observed for Gd3+- epi-inositol. This cannot result from the fact that Mn2+has a spin of 5/2 vs. 7/2 for Gd3+,as this would only account for a factor of 1.8. This result also indicates that the interaction of Mn2+is rather weak and nonspecific, requiring only two oxygen atoms.

10

20

30

[MAx

40

50

60

lo3 ( M )

FIG.2.-The Effect of Added Mn2+on the "C T fValues of epi-Inositol.[Sampleconditions the same as described in Fig. 1 (taken from Ref. 29).]

TABLEI T ; / T $ Values for Various epi-Inositol Carbon Atoms T;/TY Carbon atom ~~~

2and4b

1 and 5"

6

3

1.43

1.66

1.48

1.47

T2values were obtained from the linewidths of the resonances. Linewidths for epi-inositol samples containing no metal ions were obtained from spectra by using a 400-Hz window, and the data were collected in 16,384 addresses (taken from Ref. 29). Overlap of resonances. Only average T , values and linewidths were obtainable.

INTERACTIONS OF METAL IONS WITH CARBOHYDRATES

141

Table I shows the T ; / T ; ratios determined for the carbon atoms of a sample of epi-inositol containing 6 mM Gd3+.The results clearly establish that the predominant mechanism for the T ; relaxation process of epi-inosito1 carbon atoms is dipolar, and this shows that, in the case of polyols and carbohydrates,the selective, line-broadening technique may be used to gain quantitative information concerning ligand -Gd3+distances. a

e

A

74.1

69.3

I

I

1

I

I

00

75

70

65

60

p.p.m. from Me4Si FIG. 3.-The Effect of Mn” on the I3C Resonances of the Protondecoupled, Naturalabundance, 13C-N.m.r. Spectrum of cis-Inositol. [The concentration of cis-inositol was 390 mM in H,O (pH 7). The vertical gain of the spectra of solutions containing large proportions of paramagnetic,relaxation reagent was increased slightly,so that broadeningeffectscould be clearly observed. (A) Sample contained no Mn2+ and required 9500 accumulations. A line-broadening factor of 1.0 Hz was used during the data processing. (B) Sample contained 2. I m M MnZ+,and required I8,351 accumulations. A line-broadening factor of 8.1 Hz was used during the data processing. (C) Sample contained 10 mMMn2+, and required 3,500 accumulations. A line-broadening factor of 15.0 Hz was applied during the data processing (taken from Ref. 30).]

142

KILIAN DILL A N D R. DOUGLAS CARTER

cis-Inositol provided an opportunity to establish size requirements for binding a metal ion to this molecule. Furthermore, it could establish the mechanism that dominates the T ; relaxation process for these elucidated complexes. Mn2+and Cu2+were used in these because their ionic radii are 80 pm, and their unpaired electron spins are 5/2 and 1/2,respectively. Figs. 3 and 4 show the effects of added Mn2+and Cuz+on the naturalabundance, I3C-n.m.r.spectra of cis-inositol. Clearly, the signals of the carbon atoms bearing an axial hydroxyl group are broadened, but those of the carbon atoms having an equatorial hydroxyl group are broadened to a lesser degree. Qualitatively,these results show that these metal ions (M) bind in the trihydroxyl pocket as depicted in 4.

-

4

On comparing the spectra in Figs. 3 and 4,it is evident that considerably more Cuz+is required in order to achieve a line broadening equivalent to that with MnZ+.Indeed, this would be expected on the basis of the correlated value of S[S 11. In comparing spins 5/2 to lj2,a difference factor of 1 1.7 should be observed. Therefore, a tremendously large proportion of Cu2+was added to the solution of cis-inositol(see Fig. 4)in order to achieve broadening; this would indicate that relaxation probes need to be metal ions that contain a large number of unpaired spins, eliminating Cuz+from any further consideration as a relaxation probe. There are several interesting features concerning T ; and T ; values obtained from the data of Mn2+and Cu2+binding to cis-inositol. The ratio of the slopes of the plots of Mn” and Cuz+ vs. (T;)-’ values was 7.5 (see Table II), which is similar to the theoretical value of 11.7 expected on the basis of unpaired electron spins of the metal ions involved. No differences were noticed29in the slopes of the plots of ( vs. MnZ+and Cu” for the carbon atoms bearing axial and equatorial hydroxyl groups (see Table 11). This could occur if these carbon atoms are equidistant from the metal ions, and therefore it could be rationalized on the basis that the metal-ion binding causes a distortion of the chair conformation, driving the system to an almost planar form.

+

INTERACTIONS OF METAL IONS WITH CARBOHYDRATES

a

74.1

143

e

69.3

I\

I \

80

75

70

65

60

p.p.m. from Me4Si

FIG.4.-The Effect of Cu2+ on the ”C Resonances of the Protondecoupled, Naturalabundance, 13C-N.m.r. Spectrum of cis-Inositol. [The concentration of cis-inositol was 390 mMin H,O, pH 7. The vertical gain of the spectra of solutions containing large proportions of paramagnetic,relaxation reagent was increased slightly,so that broadening effects could be clearly observed.(A) Same as for Fig. 3A. (B) Samplecontained 17.6 mMCu2+,and required 5,000 accumulations. A line-broadening factor of 15 Hz was used during the data processing. (C) Sample contained 500 mM Cu*+,and required 168,299 accumulations. A line-broadening factor of 15 Hz was applied during the data processing (taken from Ref. 30).]

-

The observed ratios T ; / T ; for the carbon atoms of the Mn2+- and Cu2+cis-inositol complexes were found to be 1.7 and 2.4, respectively. Clearly, this indicates the dominance of the scalar contribution to the T ; relaxation process. This makes suspect the use of these metal ions as specific, linebroadening agents in order to obtain quantitative information about metal ions- ligand distances. From the broadening, it would seem that the Mn2+and Cu2+-cis-inositol interaction is in the “fast exchange” limit. In contrast to this study, addition

KlLIAN DILL AND R. DOUGLAS CARTER

144

TABLEI1 Slopes of the Plots of ‘%(Ti)-’ vs. Metal-Ion Concentration for cis-Inositol Observed slope” Metal ion

C-2,4,6

C-1,3,5

Mn2+ cu2+ Cr3+

5.55

5.55

0.76 0.206

0.76 0.206

a In lo’s-’ For residual, nonchelated cis-inositol. Because the resonance of limited, nonchelated cis-inositolis very small at high concentrationsof CrW, scattered slopes were obtained. Resonances also broadened at higher concentrations of Crw, due to outer-sphere interactions. Sample contained 300 mM Zn” (taken from Ref. 30).

of Cr3+to a sample of cis-inositolresults in the disappearance of all the 13C resonances of cis-inositol, indicating29“slow exchange.” Cr3+has an ionic radius of - 60 pm and also readily fits into the triaxial, hydroxyl-containing pocket. 2. D-Gluconamides D-Gluconamides (5) can be synthesized by the condensation of an amine with D-gluconic acid.68Syntheticdetails and crystal structure data have been given.s1aThe synthetic compoundsthat were investigated are given by structures 6- 10. R I1

C=O

21

HCOH 31

HOCH

HEC-I HO

&C-

I OH

0

PH c -,OH cC-

4H

I OH

I

H

I OH

C-NH-CH~-

7

41

HCOH

6

SI

HCOH 61

CH20H

5

HiC-

I

HO

gH

C-

I

OH

qH

,OH

C -C - C

I

I

OH

H

PH

I

0

111

-C-NH-’CH< CH3

OH

7 (8 la) N. Darbon-Mejssonier, Y . Oddon, E. Decoster, A. A. Pavia, G . Pkpe,and J. P. Reboul, Acta Crystallogr., Sect. C, 41 (1985) 1324- 1327.

INTERACTIONS OF METAL IONS WITH CARBOHYDRATES

HC ;

4'4 C I I OH OH

gH C-

-

I HO

,OH

0 C -C-NH-'CH-'CH2I OH

-CI H

ZH

145

10

111

CH3

gCH'

\ 11C02H 1

lo

CH3

9 HiCHO I

5H 4H ,OH PH CCC -COIH OH I HI OH I

0 C-N

7

8

/CH2-CH3

\ZH2-

CH3

10 These compounds provide an unusual insight into the interaction stereochemistry of Gd3+and Mn2+with linear polyols. Although the Mn2+interaction will be scalar, it will provide qualitative information about the interaction of this metal ion with the polyol. However, it should be noted that the carbonyl and carboxyl carbon atoms are expected to broaden immediately due to the efficient transmission of unpaired electron density through the n-bond network. Earlier work by Angyal and coworkers25indicated that polyols having three vicinal hydroxyl groupswith the threo-threo configuration(11) and the gauche orientation, as does D-glucitol (C-2, C-3, and C-4) provide a unique binding site for metal ions that have ionic radii of 100 pm in the fashion shown in structure 12, in which the R groups are trans to each other. These results would tend to indicate that the gluconamides should provide a unique stereochemicalarrangement for the binding of Gd3+,but not MnZ+.

-

II

12

Proton-decoupled, 13C-n.m.r.-spectral data and resonance assignments for D-gluconamides 6-10 are provided in Table 111. These compounds

KILIAN DILL AND R. DOUGLAS CARTER

146

TABLEIll 'FChemical-Shift Data for rJ-Gluconamides6- 10

Compounds

8

6

7

C-6

176.0 74.6 71.1" 73.3" 72.5 64.0

174.2 74.6 71.7" 73.4" 72.5 64.0

175.5 74.6 71.8" 73.2" 72.5 64.0

c-7

39.1

42.9

52.4

22.6

40.7

40.6

25.7b 23.4 21.9 177.5

25.6b 23.3 22.0

Atom

C-1 c-2 c-3 c-4

c-5

C-8

c-9 c-10

c-11 c-12

9

10

175.6 74.6 71.8" 73.2" 72.5 64.1

173.1 72.5 70.7" 72.1" 72.5" 64.1 43.7 42.4

52'4

I

[

:::;

54.2

LI The assignments may have to be interchanged. Specific, m n a n c e assignments for C-9 and C-I0 were not made (taken from Ref. 68).

contain various functionalgroups (amido, amino, and carboxylgroups)that may provide a focal point for Gd3+and Mn2+interactions,and the chemical shifts of some of the resonances (C-2-C-5) for the gluconamides are very similar. This would make difficult the use of any relaxation technique ( T ;/ T ; )in order to obtain quantitative structural information. Figs. 5 and 6 show the effects of added Gd3+and MnZ+on the I3C-n.m.r. spectrum of compound 6. The numbers provided over the resonancesare the respective carbon atom assignments given in Table 111. In all cases studied, Gd3+appears to broaden preferentially the resonances pertaining to carbon atoms 2,3, and 4, before affecting the other resonances. In the case of Mn*+, all of the polyol carbon atom resonances broaden equally, indicating nonspecific binding. The qualitative data for interaction of Gd3+and Mn2+for compounds 6 - 10 are given in Tables IV and V. Clearly, the carboxyl and carbonyl functionalgroups act as transmitters of unpaired electron spin density through the n system, because immediate broadening is observed in many cases, especially on use of Mn2+.However, the selective line-broadening observed in the aliphatic polyol chain should provide qualitative information about any metal-ion binding-site. The preferential broadening ofthe resonances of C-2, C-3, and C-4 of all of

INTERACTIONS OF METAL IONS WITH CARBOHYDRATES

147

I

A

I

100

175

p p. rn.

from

l

l

I

80

I

60

I

I

40

I

1

20

MeqSi

FIG.%-The Effect of Gd” on the I3C Resonances of the Protondecoupled, Naturalabundance, I3C-N.m.r. Spectrum of 6. [The concentration of compound 6 was 168 mM in H,O, pH 7. The vertical gain of the spectra of solutions containing large proportions of paramagnetic, relaxation reagent was increased slightly, so that broadening effects could be clearly observed. (A) Sample contained no GdW,and required 27,000 accumulations. A linebroadening factor of 2.3 Hz was used during the data processing. (B) Sample contained 3.6 mMGd3+, and required 28,502 accumulations. A line-broadening factor of 3.5 Hz was used during the data processing. (C) Sample contained 7.2 mM Gd”, and required 28,9 12 accumulations. A line-broadening factor of 4.5 Hz was applied during the data processing. (D) Sample contained 12 mM Gd3+,and required 2 1,649 accumulations. A line-broadening factor of 7.0 Hz was applied during the data processing (taken from Ref. 68).]

-

the D-ghconamides indicates that they contain a unique binding site for Gd3+,as depicted in structure 12. In the case of Mn2+,addition of this metal ion to the D-gluconamide samples does not result in preferential binding of Mn2+to the polyol chain; this indicates that the interaction is nonspecific, possibly requiring only two oxygen atoms. The difference observed for the

KlLIAN DILL AND R. DOUGLAS CARTER

148

I

A

p.p.m

from

M y 9

FIG.6.-The Effect of Mn2+ on the Resonances of the Protondecoupled, Naturalabundance, W-N.m.r. Spectrum of 6. [The concentration of compound 6 was 168 mM in H,O, pH -7. The vertical gain of the spectra of solutions containing large proportions of paramagnetic, relaxation reagent was increased slightly, so that broadening effects could be clearly observed. (A) Same as in Fig. 3A. (B) Sample contained 7.2 mM Mn2+,and required 32,580 accumulations. A line-broadeningfactor of 3.0 Hz was used during the data processing. (C) Sample contained 42.0 mM Mn2+,and required 38,640 accumulations. A line-broadening factor of 5.0 Hz was applied during the data processing. (D) Sample contained 72 mM Mn2+, and required 54,838 accumulations. A line-broadeningfactor of 7.0 Hz was applied during the data processing (taken from Ref. 68).]

INTERACTIONS OF METAL IONS WITH CARBOHYDRATES TABLEIV The Effects of Added Gd” on the ‘C Resonances of Compounds 6 - 10 Compounds

6

Atom C-1 C c-3 c-4 C-5 C-6 C-I c-8 c-9 c-10 c-11 c-12

s 2

s s

s s s

m

s

m w

W

s

W

n

9

10

w s

w m s s s s

s s

s s m m

s

m w

8

7

m m ~

w n n s

W

n n n m n

W

n

a The abbreviations are: m, moderate broadening; n, no broadening; s, severe broadening; w, weak broadening; see Fig. 5 . (Taken from Ref. 68.)

TABLEV The Effects of Added Mn2+on the of Compounds 6-10

Resonances

Compounds Atom

7

6

8

9

10

c-I

s

s

s

s

s

C-2 C-3 C-4 C-5 C-6 C-7 c-8 c-9 c-10 c-11 c-12

m m m m m w

m m m m m n n

m m m m m s m n n s

m m m m m n n n n m n

m m m w m n n

a The abbreviations are: m, moderate broadening; n, no broadening; s, severe broadening; w, weak broadening; see Fig. 6. (Taken from Ref. 68.)

149

KILIAN DILL AND R. DOUGLAS CARTER

150

interaction of Gd3+and Mn2+with D-gluconamides must result from the difference in the ion size. 3. Simple Monosaccharides and Glycosylated Amino Acids

Of immense biological interest is the interaction of metal ions with monosaccharides, polysaccharides, and glycoproteins. To this end, the interaction of Gd3+and Mn2+with methyl a-and P-D-galactopyranoside(13 and 14), 3-O-a-~-Gal-~-Ser ( 1 3 , and 3-O-a-~-GalpNAc-~-Ser (16) was investigated. Compound 16 represents a common carbohydrate-amino acid linkage pair found in Nature, and compounds 13-15 are structurally related.

HO

HO

0CH3

13

14

0-

$H,-C-C02Me H

15

R = OH

16

R = NHAc

I

The D-gulucto structure of 13 and 14 was not expected to provide any strong binding site for Gd3+and Mn*+, but the interaction of Mn2+with methyl a-and P-D-galactopyranosidewas investigated by using longitudinal

INTERACTIONS OF METAL IONS WITH CARBOHYDRATES

151

3

I

5

10

15

20

25

30

[Mn2+] x 103(M) FIG.7.-The Effect of Added Mn2+on the I3C 7'; Values ofCompound 13. [The concentration of 13 was 1 M (in H,O), at pH -7.0. The numbers on the graph refer to specific carbon atoms of 13 (taken from Ref. 64b).]

electron-nuclear relaxation-rates. It seemed possible that a configurational change at C-1might influence the interaction of Mn2+with the D-galactopyranoside. Figs. 7 and 8 show the plots ofthe (T;)-I values of methyla- and P-D-galactopyranoside as a function of added MnZ+.For 13,it would appear that no unique binding site for Mn2+exists on this compound. On the other hand, C-6 of compound 14 would appear to exhibit enhanced relaxation, indicating a preferential interaction of Mn2+with 0-6. The 0-glycosylated amino acids 15 and 16 provide a basis for study of the attachment of Gd3+and Mn2+to glycopeptides. Unlike simple monosaccharides, glycosylated amino acids, and, to some extent, glycopeptides and glycoproteins, contain functional groups (amino, carbonyl, and carboxyl) that provide a primary focal point for metal-ion binding. These compounds contain N- and 0-blocking groups which should limit the interaction ofGd3+ with these functional groups. In these studies, the selective line-broadening

152

KlLIAN DILL AND R. DOUGLAS CARTER

4

3

I

5

10

15

20

25

30

[Mn“] x 103(M) FIG.&-The Effect of Added Mn2+on the IF, Tf Values ofCompound 14. [Theconcentration of 14 was 1 M (in H,O),at pH 7.0. The numbers on the graph refer to specific carbon atoms of 14 (taken from Ref. 64b).]

-

technique was used for quick acquisition of qualitative information. Gd3+is a relaxation probe for which the selective line-broadening is based on the predominance of the dipolar contribution to the TC, value. Figs. 9 and 10 respectively show the effects of GdH on the I3Cresonances of the proton-decoupled, natural-abundance, 13C-n.m.r. spectra of compounds 15 and 16. For both compounds, the resonances of C- 1 and C-2 of GalNAc and Gal appear to broaden preferentially. The resonances of C-4’ and C-6’ appear also to show slightly increased preferential broadening. On the amino acid, the acetamido methyl group signal appears also to be broadened. These results suggest that Gd3+probably interacts with the glycosidic oxygen atom (0-3) and 0-2’ for 15, and with 0-3 a1.A N-2’ for 16. Weaker interaction possibly occurs with 0-4’ and 0-6’ of these molecules. On the whole, it would then appear that even the blocked amino acids must play a role in the interaction of Gd3+with the glycosylated amino acid.

INTERACTIONS OF METAL IONS WITH CARBOHYDRATES ad

153

a a4

a I'

I

A

6.60mM Gd3+

1

18.6rnM t d 3 +

I

I

C

J

l

l

l

l

l

110 100 90 80 70 60

l

50

l

40

l

l

30 20

p.p.m. from Me,Si

FIG.9.-The Effect of Added GdN on the I3C Resonances of the Aliphatic Region of the Proton-decoupled, Natural-abundance, l3C-N.m.r. Spectrum of Model Compound 15. [The concentration ofcompound 15 was 150 mM(in H,O), pH 6.8. The vertical gain ofthe spectra containing the paramagnetic, relaxation reagent was increased slightly, so that broadening effects might be clearly observed. (A) Sample contained no GdW,and required 15,250 accumulations. A line-broadening factor of 3.0 Hz was applied during the data processing. (B) Sample contained 6.6 mMGd3+, and required 21,723 accumulations. A line-broadening factor of 3.5 Hz was applied during the data processing. (C) Sample contained 18.6 mMGdN, and required 17,262 accumulations. A line-broadeningfactor of 5.0 Hz was applied during the data processing. (D) Sample contained 36.6 mM GdN, and required 35,690 accumulations. A linebroadening factor of 8.0 Hz was applied during the processing (taken from Ref. 65).]

154

WLIAN DILL AND R. DOUGLAS CARTER

A 7.20 m M Gd3+

I

/

B

14.4 m M Gd3+

C D

110 100 90 80 70 60 SO 40 30 20

p.p.m. from Me,Si FIG. 10.-The Effect of Added Gd3+on the "C Resonances of the Aliphatic Region of the Proton-decoupled, Natural-abundance, %-N.m.r. Spectrum of Model Compound 16. [The concentration of compound 16 was 126 &(in H,O), pH 6.6. The vertical gain of the spectra containing the paramagnetic, relaxation reagent was increased slightly, so that broadening effects might be clearly observed. (A) Sample contained no Gd*, and required 13,600accumulations. A line-broadening factor of 3.0 Hz was applied during the data processing. (B)Sample contained 7.2 mM Gd*, and required 43,407 accumulations. A line-broadening factor of 4.8 Hz was applied during the data processing. (C) Sample contained 14.4 mMGd*, and required 35,131 accumulations. A line-broadeningfactor of 6.0 Hz was applied during the data processing. (D) Sample contained 2 1.6 mMGdw, and required 3 1,492 accumulations. A linebroadening factor of 8.0 Hz was applied during the data processing (taken from Ref. 65).]

INTERACTIONS OF METAL IONS WITH CARBOHYDRATES

155

4. Complex Glycopeptides The interactionsof Gd3+(and Mn2+)with complex glycopeptidesprovide a better measure of metal-ion interactionswith glycoproteins. Severaldifferent types of glycopeptide were studied: those that contain (i) a single carbo-

HC-

I

CH3

17 R 18R

= a-D-Galp = H

CO,H

I

HC-OR

1

CH3

19 R = a-D-Galp 2 0 R = H AcHN-Gly-Thr-Thr-C

I

1

R

R

H0

21 R = a-D-Galp 22 R = P-D-Galp

OR

156

KILIAN DILL A N D R. DOUGLAS CARTER

ACHN-

Gly -Ser-

I

Ser-Thr

I

I

-Thr-Gly

1 7 :

I

-C

\ N-Me

0

HO

R

R

I

R

H

23 R = a -D - Galp HO

-\

HO

Ho

0-C%-C-CO2CH3 3

H

HO

Z 0

I C02CH3

0 -CCHZ-C3

I

H

I

25

24 Z = benzyloxycarbonyl

hydrate residue (18 and 20), which are compared to the nonglycosylated species, 17 and 19, (ii) vicinally glycosylatedsites (21,22, and 23), and (iii) a disaccharide unit (24 and 25). The compounds studied gave results that elucidated several points concerning metal-ion - glycoprotein interactions: (a) do vicinally attached carbohydrates provide a unique GdN (or Mn2+)binding site?; (b) does a core disaccharide possess a unique binding site?; (c) does the formation of a peptide bond lessen the interaction of this part of the molecule with metal ions?; and (d) is there differentialmetal-ion binding to a-andB-D-gdactopyranosyl residues bound to a dipeptide? Study of the interactionsof Gd3+with compounds 17- 20 answered questions concerning the role that the peptide bond plays in the metal-ion binding-site. Furthermore, this study clearly distinguished the metal-ion carbohydrate and metal-ion - peptide interactions. Studies related to the interaction ofGd3+withcompounds 17 and 18have been published.65For the nonglycosylated species, the C-terminal Thr resonances (C-3 and C-2) and N-terminal Gly resonance (C-2) were broadened

INTERACTIONS OF METAL IONS WITH CARBOHYDRATES

157

on addition of Gd3+.This was to be expected, because of the carboxylate and a-amino groups present in these amino acids. Therefore, a similar result for compound 17, containing the a-D-Galpgroup, should be seen, in addition to a possible, unique interaction of Gd3+with the Gal group. From studiesdealing with the interactionsof Gd3+with17, the resonances of C- 1’ and C-2’ of the a - ~ - G a l pkoup appear to broaden, suggesting a binding site near these carbon atoms that is not influencedby the binding of Gd3+to the carboxyl and amino groups. Similar results to those shown were observed for the interaction of Gd3+with compounds 19 and 20.6sIn this case, the signal :noise ratio of the spectra did not allow quantitative conclusions. Examination of the behavior of compounds 21-23 should indicate whether vicinal glycosylation by D - G groups ~ produces a unique binding site and whether the anomeric state of the D - G influences ~ the metal-ion binding. Addition of Gd3+and MnZ+to solutions containing compounds 21 and 22 produced broadening of the resonances of C-6’ and C-2’, and possibly C-3’ and C-1’, of Gal in the 13C-n.m.r. spectra of these compounds. Other resonances that were broadened were those of C-3 of Thr. These results suggested binding sites for MnZ+and Gd3+in the vicinity of 0-6’, 0-2’ (0-3’), and 0-3 of Thr. Similar, but not identical, results were observeda when the interaction of Gd3+and Mn2+with 23 was studied. In this case, Gd3+appeared to interact with the oxygen atoms on C-2’, C-3’ (or C-47, C- 1’,and C-6’ of mGal, and Thr 0-3; this suggested a random interaction of Gd3+with the three a - ~ - G a l groups. Addition of Mn2+ to a solution of 23 appeared to broaden the resonances of C-2’, C-6’, and (to a lesser extent) C-1’, and, possibly, C-2‘, or C-4’ of a-Gal. The Thr C-3 resonance was also broadened. This result suggested that stronger binding sites involve a-D-Gal 0-6’ and a second site involves a-D-Gal O-2' and the glycosidic oxygen atom (0-3). A weak binding site may exist around 0-3’ (or 0-4’). The results clearly show that the degree of vicinal glycosylation will alter metal-ion binding to the simple monosaccharide, and that this may vary from metal ion to metal ion. The addition of GdW to a solution containing compound 22 results in broadening of almost all of the I3C resonances observed, although slight preferential broadening is seen for jl-~-GalC- 1’ and C-6’. In the case of the addition of Mn2+,the C-6 resonance of P-D-Gal was noticeably broadened; P-D-Gal C- 1’ and C-3’ were also broadened. This suggeststwo binding sites: one involving the glycosidic oxygen atom and P-Gal 0-3’, and another involvingjl-Gal O-6'. It would appear that the metal-ion interaction for the two anomeric Gal groups of compounds 21 and 22 differs somewhat. This may result from the vicinal attachment or to glycopeptide structural differences.

158

KILIAN DILL AND R. DOUGLAS CARTER

Metal-ion interactions with glycoproteins or isolated intact glycopeptides usually involve larger oligosaccharidesand not singly attached monosaccharides. To this end, the binding of Gd3+to derivatives of the core glycopep tides O-p-~-Gal-(1 -3)-(~-~-GalNAc--,Ser,Thr (24 and 25) was investigated. Both of these compounds contain an N-benzyloxycarbonylblocking group and methyl-esterified carboxyl groups that should inhibit any strong binding to these functional groups. For both compounds, the resonance of the acetamido methyl group is broadened considerably. Furthermore, those of C-2 of GalNAc and Thr C-3 are broadened. Slight broadening was also observed for GalNAc C-1’ and C-3’. The results indicated that a strong binding site existsfor Gd3+near to or involving N-2’ of the acetamido group, 0-3’, and, possibly, the glycosidic oxygen atom. It is quite unusual that the metal ion only appears to interact strongly with the glycosidically linked a-GalNAc residue and not with the terminal P-Gal group; this may be due, in part, to the influence of the esterified acid group and the blocked amino group. It should be noted that this binding phenomenon may also change with a more complex oligosaccharide, especially those containing functionalized carbohydrates to which “strong” metal-ion binding may occur (see later).

5. N-Acetyl-clr-D-neuraminic Acid As mentioned earlier, a-D-NeuAc (26) is a prominent carbohydrate in Nature and is found extensively on certain cell surfaces, such as those of erythrocytes, and in brain gangliosides. The structure of a-D-NeuAc indi-

H

26 R i = R 2 = H

27 Ri =H,R2=CH3 28 Ri = R 2 = C H 3

p.pm fmm Me&i FIG. 1 I .-The Effect of Gdw on the 13C Resonances of the Proton-decoupled, Naturalabundance, I3C-N.m.r. Spectrum of a-NeuAc Methyl Glycoside. [The concentration of compound 27 was 188 W i n H20, pH 6.8. The vertical gain ofthe spectra of solutions containing large proportions of paramagnetic, relaxation reagent was increased slightly, so that broadening effectscould be clearly observed. (A) Sample contained no Gd”, and required 5,000 accumulations. A line-broadening factor of 1.2 Hz was applied during processing. (B) Sample contained 12 mMGd3+,and required 5,000 accumulations. A line-broadening factor of 1.4 Hz was applied during processing. (C) Sample contained 24 mM Gd3+,and required 15,000 accumulations. A line-broadening factor of 3.0 Hz was applied duringprocessing. (D) Sample contained 49 mM Gd”, and required 15,000 accumulations. A line-broadening factor of 5.0 Hz was applied during processing (taken from Ref. 64a).]

160

KILIAN DILL AND R. DOUGLAS CARTER

-

cates that the carboxylate group present (pK, 2.6) should be a focal point for metal-ion binding. Research dealing with the interaction of Ca2+and Eu3+with P-N~uAc,’~,’~ ”,~~ composed of a-NeuAc? and oligosacderivativesof N ~ U A C ,gangliosides charides and glycopeptides of glycophorin A containing c~-NeuAc~*~ has been published. Unfortunately,one of the studies dealt with theP anomer of NeuAc (not normally found in Nature), and three dealt with the diamagnetic Ca2+,which, upon binding, only results in a few tenths of the I3Cchemicalshift change for certain resonances. In studies64adealing with derivatives of a-NeuAc (27 and 28), selective line-broadeningdata acquired upon addition ofGd3+and MnZ+were used to gain information about solution binding of Gd3+and Mn2+to these species. Compound 28 was used to permit comparisonwith the results obtained with compound 27, in order to elucidate the role that the carboxylate plays in the binding of Gd3+and Mn2+.No quantitative structural information is provided for the case of Mn2+,because of the dominance of the scalar mechanism to the T : relaxation of certain carbon atoms. The effects of graded addition of Gd3+on the I3C-n.m.r. spectra of methyl a-D-NeuAc (27) are given in Fig. 1 1. Only selected spectra are shown, although spectra containing other additions of Gd3+were also obtained. The numbers on the uppermost trace refer to the numerals assignedto the carbon atoms shown in structures 26 - 28. On a qualitativebasis, it may be seen that the resonances that are immediately broadened are those of C- 1, C-2, C-8, TABLEVI Initial and Final Coordinates of the Computer Modeling of Binding of Gd* to a-D-NeuAc Methyl Glycoside Carbohydrate-Gd” Distances (in Ay Final coordiites

Initial coordinates Atoma

x

Y

Z

X

Y

z

c -1 c-2 C-6 c-7 C-8 c-9 0-8 0-6 Gd(111)

- 1.22 0.00 0.00 0.00 - 1.41 -2.12 -2.09 0.00 -4.16

0.64 1.35 0.00 0.00 0.00 -1.22 1.17 1.35 1.77

3.98 3.48 1.50 0.00 -0.50 0.50 -0.02 1.98 1.44

- 1.22 0.00 0.00 0.00 - 1.26 - 2.3 1 - 1.72

0.64 1.35 0.00 0.00 0.63 -0.42 1.59 1.35 1.98

3.98 3.48 1.50 0.00 -0.50 -0.69 0.45 1.98 2.00

0.00

-3.79

See 29 for atom-numbering. Taken from Refs. 82 and 83.

(r,-Gd”)

3.50 4.12 4.31 4.72 3.8 1 3.90 2.6 1 3.84

-

INTERACTIONS OF METAL IONS WITH CARBOHYDRATES

161

and C-9, and the peak pertaining to the overlap of the resonances of C-4 and C-7. If a single binding site exists, as was expected from the previous data, it must involve the carboxyl group at the head of the molecule and the glycerol- 1-yl side-chain at the tail of the molecule. If the Gd3+bound to 27 is now in the “fast-exchange” limits (as was to be expected on the basis of the linewidth, temperature dependence and field strength) and T ; relaxation is dominated by a dipolar mechanism, the linewidths observed may be used to calculate a Gd’+-a-~-NeuAc binding structure. Using Eq. 14, distance information can be calculated for this system. For this equation, T, = l/nW,,, and W1,*.is the linewidth of the resonance at half height. The information needed in order to extract (Y-DNeuAc-Gd3+ distance information is a fixed Gd’+-carbon atom distance that might be used as a standard to relate to all of the other linewidth data. The literaturez7gives C - 0 -Gd3+(for carboxylate complexes)to be 350 pm; carbon bonds were assumed to have an average bond length of 150 pm. Using these values as a basis, all other relevant carbon atom distance information was calculated. The C - Gd3+distances are given in Table VI. Table VI also gives the initial coordinates for certain carbon atoms of a-D-NeuAc; these were obtained from crystallographic data, and the

(Gd 3’)

cln

\;/

“9+ I

,

/OH CH2

I

I

H

H

29

Y

KILIAN DILL AND R. DOUGLAS CARTER

162

47

-//-

p.p.m. from Me,@ Effect of Mn2+ on the ‘TResonances of the Proton-decoupled, Naturalabundance, I3C-N.m.r. Spectrum of a-NeuAc Methyl Glycoside. [The concentration of compound 27 was 198 mMin H,O, pH 7.2. The vertical gain of the spectra of solutions containing large proportions of paramagnetic, relaxation reagent was increased slightly, so that broadening effects could be clearly observed. (A) Sample contained no Mn”, and required 5,OOO accumulations. A line-broadeningfactor of 1.2 Hz wasapplied during processing.(B) Samplecontained 18 mM Mn*+, and required 5,OOO accumulations. A line-broadening factor of 1.6 Hz was applied during processing. (C) Sample contained 6 1 mMMn2+,and required 10,OOOaccumulations. A line-broadening factor of 3.8 Hz was applied during processing. (D) Sample contained 98 mM MnZ+,and required 15,000 accumulations. A line-broadening factor of 4.5 Hz was applied during processing (taken from Ref. 64a).]

FIG. 12.-The

p,p.m. from Me,Si FIG. I3.-The Effect of Mn2+ on the 13C Resonances of the Protondecoupled, Naturalabundance, l3C-N.m.r. Spectrum of a-NeuAc Methyl Ester, Methyl Glycoside. [The concentration ofcompound 28 was 99 mMin H,O, pH 6.9. The vertical gain ofthe spectra ofsolutions containing large proportions of paramagnetic, relaxation reagent was increased slightly, so that broadening effects could be clearly observed. (A) Sample contained no Mn2+,and required 5,000 accumulations. A line-broadening factor of 1.2 Hz was applied during processing. (B) Sample contained 6 mMMn2+,and required 5,000 accumulations. A line-broadening factor of 3 Hz was applied during processing. (C) Sample contained 49 mh4 MnZ+,and required 10,OOO accumulations. A line-broadening factor of 2 Hz was applied during processing. (D) Sample contained 98 mMMn2+,and required 10,000accumulations. A line-broadeningfactor of 5 Hz was applied during processing (taken from Ref. 64a).]

164

KILIAN DILL AND R. DOUGLAS CARTER

p.p.m. from Me,Si Resonances of the Proton-decoupled, NaturalFIG. 14.-The Effect of Gd* on the abundance, Y-N.m.r. Spectrum of 0-NeuAc Methyl Ester, Methyl Glycoside. [The concentration of compound 28 was 82 W i n H,O, pH 6.9. The vertical gain of the spectra of solutions containing large proportions of paramagnetic, relaxation reagent was increased slightly, so that broadening effects could be clearly observed. (A) Same as in Fig. 13A. (B) Sample contained 6 mM Mn2+,and required 5,000 accumulations. A line-broadening factor of 1.2 Hz was a p plied during processing. (C) Sample contained 24 mMGd*, and required 10,000 accumulations. (D) Sample contained 49 mMGd*, and required 10,000 accumulations. A linebroadening factor of 5.0 Hz was used during processing (taken from Ref. 64a).]

INTERACTIONS OF METAL IONS WITH CARBOHYDRATES

165

assumption was made that it has a rigid, pyranoid ring-structure.81bA computer program was used so to rotate the C-7 -C-8 and C-8 -C-9 bond angles that the distance information that was tobe obtained from the n.m.r. data, with respect to Gd3+,could be maximized. The final, computer-generated coordinates for Gd3+binding to a-D-NeuAc are given in Table VI; a strucGd3+ bind to a-D-NeuAc above the ture is depicted in f o r m ~ l a29. ~~ . ~ should ~ pyranoid ring and off to one side (behind C-6). A qualitative estimate of the interaction of Mn2+with a-D-NeuAc can be obtained from the line broadening data shown in Fig. 12. It would appear that multiple Mn2+binding sites exist on a-D-NeuAc, involving both the head and the tail of the molecule (the glycerol- 1 -yl side-chain, as well as the acetamidogroup). This is somewhat different from that observed with Gd3+, because the resonances pertaining to C-4, C-5 and the acetamido group are not broadened when Gd3+is bound. Therefore, in the case of the binding of Mn2+to a-D-NeuAc, the interaction appears to be somewhat nonselective. The necessity of the carboxyl group (C-1) for metal-ion binding to a - ~ NeuAc was also investigated, using compound 28, in which the carboxyl group is esterified. These results are shown in Figs. 13 and 14, where the effects of graded addition of Gd3+ and MnZ+on the proton-decoupled, natural-abundance, "C-n.m.r. spectra of 27 were monitored. Qualitatively, it may be seen that the C-1 and C-2 resonances are not appreciably broadened (except for general line-broadening) when Gd3+and Mn2+are added, indicating that the head of the molecule is no longer a focal point for metalion binding. In the case. of Gd3+binding, the results in Fig. 13 indicatethat an interaction occurs at the tail of the molecule, involving the glycerol-1-yl side-chain (C-7 -C-9) and the acetamido group (C-4, C-5, C-9, and -CH3). Somewhat similar resultswere also observed for the interaction of Mn2+with 27, as depicted in Fig. 14. The aforementioned results indicate that a-D-NeuAcappearsto be able to bind Gd3+and Mn2+(although less so) without the presence of the carboxyl group. The presence of this group results in enhanced binding, and, in the case of Gd3+,a unique binding structure.

VI. CONCLUSIONS It has been shown that Mn2+and Gd3+act as relaxation probes that can be used to investigate the interaction of these metal ions with carbohydrates and (81b) S. L. Flippen,ActaCr)?stallogr.,Sect. B, 29 (1981) 1881-1886. (82) M. E. Daman, Ph.D. Dissertation, Clemson University, 1984. (83) K. Dill, R. D. Carter, and M. E. Daman, Int. Syrnp. Solute-Solute-Solvent Interactions. 8th, Regensburg, W. Germany, Aug. 9- 14, 1987.

166

KlLIAN DILL AND R. DOUGLAS CARTER

glycopeptides. Only Gd3+can be used as a selective, line-broadening agent, in order to obtain distance information from line-broadening experiments, especially for atoms that are directly bonded to the metal ion. The results indicate that most of the carbohydrate residues, such as Gal and GalNAc, weakly interact with Gd3+and Mn2+.The cyclitols (inositols), on the other hand, form with these metal ions unique complexes that are based on the geometry (stereochemistry)of the multiple hydroxyl groups. Due to stereochemical requirements, a unique, metal-ion binding-site (for Gd3+)on gluconamides is also observed, although this interaction is weak. Strong binding of Gd3+and Mn2+to a-D-NeuAcis observed. In the case of Gd3+,selective line-broadening was used to determine the structure of the Gd3+-a-D-NeuAc complex. ACKNOWLEDGMENTS The authors thank Susan C. Eller for typing the manuscript.K. D. acknowledgesthe support of the National Institutes of Health (Grant GM36252-01).

ADVANCES IN CARBOHYDRATE CHEMISTRY AND BIOCHEMISTRY. VOL. 41

APPLICATION OF ANHYDROUS HYDROGEN FLUORIDE FOR THE STRUCI'URAL ANALYSIS OF POLYSACCHARIDES B Y YURIY A . m I R E L A N D

EVGENYv. VINOGRADOV

N. D. Zelinsky Institute of Organic Chemistry, Academy of Sciences of the U.S.S.R., MOSCOW, B-334, U.S.S.R.

AND

ANDREWJ. MORT

Department of Biochemistry, Oklahoma State University, Stillwater, Oklahoma 74078 I. 11. 111. IV. V. VI. VII.

Introduction ........................................................ Mechanism of Reaction between Hydrogen Fluoride and Carbohydrates. . . . . . . Techniques for Performing Solvolysis. ................................... Preparation of Monosaccharides. ....................................... Preparation of Oligosaccharides ........................................ Other Applications of Hydrogen Fluoride ................................ Conclusion .........................................................

167 168 173 174 180 200 202

I. INTRODUCTION Cleavage of glycosidic linkages with a view to obtaining monosaccharides or oligosaccharides is one of the principal chemical approaches used to establish the composition and structure of polysaccharides. Until recently, the most important methods for cleavage were hydrolysis with mineral or organic acids, and acid-catalyzed methanolysis.'** However, full depolymerization with these reagents does not always proceed satisfactorily. Under the conditions usually adopted, some monosaccharides, such as 3-deoxyaldulosonic acids, undergo decomposition. The glycosidic linkages of some other G . 0.Aspinall, in G . 0.Aspinall (Ed.), The Polysaccharides,Vol. 1, Academic Press, New York, 1982, p. 350. (2) B. Lindberg, J. Lonngren, and S. Svensson, Adv. Carbohydr. Chem. Biochem.. 31 (1975) 185 -240. ( 1)

167

Copyright 0 1989 by Academic Press, Inc. All rights of reproduction in any form ~ s e ~ e d .

168

YURIY A. KNIREL et al.

monosaccharides(such as uronic acids, and amino sugars having free amino groups or readily lost N-acyl substituents)have great stability, and often are not cleaved fully, even under relatively drastic conditions. On the other hand, in performing selective cleavage, the differences in the stability of glycosidic linkages of different monosaccharides present are often insufficient to afford oligosaccharide fragments in satisfactoryyields. A number of specific methods have been developed for selective cleavage'^^; however, their applicability may be restricted by the requirements of some specific peculiarities in the composition or structure of the carbohydrates studied. Anhydrous hydrogen fluoride has been used for the cleavage of glycosidic linkages. Application of this versatile reagent has some advantages over hydrolysis with aqueous acid and methanolysis. Hydrogen fluoride cleaves the glycosidic linkages of the majority of monosaccharides with practically no decomposition of the sugar; it does not affect the N-acyl substituents of acylamido sugars, and, under certain conditions, any 0-acyl substituents present can also be retained. Cleavage of each glycosyl linkage in a polysaccharide is highly dependent on the nature of the sugar involved, its anomeric configuration, the neighboring residues, the temperature of the reaction, and, to some extent, the duration of the reaction. Thus, experiments can often be designed to give complete depolymerization of a polysaccharide at one temperature, or, at a substantially lower temperature, to yield oligosaccharides, and, at a still lower temperature, to yield larger oligosaccharides. The present article summarizesthe data on application of hydrogen fluoride for structural analysis of polysaccharides, primarily those of bacterial origin. Some regularities in the stability of glycosidic linkages of different monosaccharides,and the mechanism of interaction between hydrogen fluoride and carbohydrates, are briefly discussed. The experimental technique for carrying out solvolysis is also described. 11. MECHANISM OF REACTION BETWEEN HYDROGEN FLUORIDE AND

CARBOHYDRATES The reactions of amylose, cellulose, inulin, starch, and xylans, and their constituent monosaccharides have been rather well ~ t u d i e d . ~ These - ~ polysaccharidesdissolve in hydrogen fluoride at temperatures above - 20 ,and undergo depolymerization to give the corresponding glycosyl fluorides; furO

(3) A. F. Sviridov and 0. S . Chizhov, Bioorg. Khim., 2 (1976) 3 15 - 350. (4) A. J. Mort and S. Parker, SERI Rep. SERI/CP/232- 1520, Proc. Int. Conf:Biotechnol. Prod. Chemicals Fuels Biomass, (1982) 51 - 64. ( 5 ) J. Defaye, A. Gadelle, and C. Pedersen, Carbohydr. Rex, 110 (1982) 217-227. (6) J. Defaye, A. Gadelle, and C. Pedersen, Curbohydr. Rex, 136 (1985) 53-65.

APPLICATION OF ANHYDROUS HYDROGEN FLUORIDE

169

oligosaccharides products of Friedel - Crafts alkylation of aromatic compounds SCHEME1.-The reactions which glycosidic linkages of aldohexoses can undergo in anhydrous liquid hydrogen fluoride.

ther conversions of these fluorides depend essentially on the reaction conditions. (see Scheme 1). The ability of hydrogen fluoride to dissolve polysaccharides is due to its formation of hydrogen bonds, which results in breaking of the hydrogen bonds between the polysaccharide molecule^.^ Subsequent depolymerization proceeds by way of protonation of the glycosidic oxygen atom and cleavage of the aglycon to give the glycosyl cation in accordance with the known mechanism of acid hydrolysis of glycosides.’ Besidesthe formation of hydrogen bonds between hydrogen fluoride and hydroxyl groups of sugars, solvolyticcleavage is also promoted through the stabilization of the resulting glycosyl cation by hydrogen fl~oride.~ Addition of the fluoride anion to the glycosyl cation affords the a-glycosyl fluoride, which is thermodynamically more stable (due to the anomeric effect). Thus, after equilibration of the reaction with cellulose (for 15 min, at amylose, and (1 44)-P-~-xylan, OO),

(7) W. G. Overend, in W. Pigman and D. Horton (Eds.), The Carbohydrates, Chemistry and Biochemistry, Vol. IA, Academic Press,New York, 1972, p. 317.

YURIY A. KNIREL et ul.

170

the solution contains only the a-D-glycosyl fluoride, together with products of secondary 0ligomerization.4,~ Secondary oligomerization occurs by glycosylation of hydroxyl groups of sugars by the glycosyl fluorides. At concentrations of sugars of < 196, this process is usually insignificant, but, for higher concentrations, arising, in particular, on evaporation of the reaction mixture, it begins to play the predominant role. Thus, evaporation of a 1% solution of cellulose in hydrogen fluoride (which is 85% D-glucosyl fluoride) yields water-soluble, branched oligosaccharidesinvolving from 2 to 14 monosaccharide residues linked mainly by a-(1+6), (1 +2), and ( 1 -3) linkages (24, 12, and 1 1%, respecti~ely.)~*~*~ Similar results were ~ b t a i n e don ~ . ~solvolysis of D-xylans with hydrogen fluoride, whereas solvolysis of inulin (see Scheme 2) proceeded differently, to give dianhydride 1, as well as other &anhydrides6 After 3 days in hydrogen fluoride at room temperature, a 6% solution of anhydrous D - ~ ~ U C Ogave S ~ rise to a yield of 22% of a nondialyzable ~ - g l u can which contained from 10 to 20% of 3,6-anhydro-~-glucose.~

-

-

D-Fructose

HF

HzOH

~

or inulin

HO

HO

H?

H?

Hb

Hb

SCHEME 2.- The reactions of ketohexose in anhydrous liquid hydrogen fluoride. (8) H. Hardt and D. T. A. Lamport, Phytochemistry, 21 (1982) 2301 -2303. (9) U. Kraska and F. Micheel, Curbohydr. Rex, 49 (1976) 195- 199.

APPLICATION OF ANHYDROUS HYDROGEN FLUORIDE

171

CH2OH

Chitin

HF

I

(OH

RO Q F

F-

NHAC

3

RO QH

~HAC

5

L

-Jn

In contrast, no secondary oligomerization was observed on cleavage of chitin1O-l2and of some other natural polysaccharides containing amino sugars, even though the hydrogen fluoride was removed from these reaction mixtures by evaporation (see later). Solvolysisof chitin12proceeds by formation of the oligomeric pyranosyl oxazolinium ion (2) existing in equilibrium with the a-glycosyl fluoride (3). The former is transformed into the (10) A. J. Mort, Ph.D. Thesis, Michigan State University (1978). (1 1) A. J . Mort and D. T. A. Lamport, Anal. Biochern., 82 (1977) 289-309. (1 2) C. Bosso, J. Defaye, A. Domard, A. Gadelle, and C. Pedersen, Curbohydr.Res., 156 ( I 986) 57-68.

I72

YURIY A. KNIREL et a/.

monomeric furanosyl oxazolinium ion (4). Hydrolysis of the ions 2 and 4 results in the oligomeric and monomeric products (6 and 5, respectively). Glycosyl fluorides are also capable of alkylating aromatic compounds by way of the Friedel- Crafts mechanism, hydrogen fluoride being an efficient catalyst of this reaction.I3Such alkylation becomes significant for the solvolysis of glycoproteins. Alkylation can be prevented by adding an excess of anisole to the reaction mixture." Alkylation may also be inhibited by the addition of methanol (which reacts with the glycosyl fluorides to form methyl glycosides).10~14 Experimentsshowed that furan derivatives, which are frequently found to occur in the products of hydrolysis of pentose-containing polysaccharides with mineral acids, are not formed on solvolysis with hydrogen fl~oride.~ The stability of glycosyl fluorides varies greatly. CK-D-G~UCOS~~ fluoride is quite stable in water at neutral pH values and can be readily obtained from the dissolution of cellulose or glucose in hydrogen fluoride, followed by quenching of the r e a ~ t i o n .Some ~ . ~ glycosyl fluorides have even been used as substrates for glycosyl transferases.Is Other glycosyl fluorides are less stable. The oligosaccharideisolated on solvolysisof the polysaccharidefrom Rhizobium japonicum bacteria (having D-mannosyl fluoride on the originally reducing end) was fully hydrolyzed16in 2 h at 20 No aminoglycosyl fluorides have been isolated in aqueous systems. However, the easily hydrolyzable 2-acetamido-2-deoxy a-D-glucopyranosyl fluoride has been detected1* by 13C-n.m.r.spectroscopy. The presence in the reaction mixture of water up to a certain limit does not affect the general composition of the reaction products. For instance, treating D-xylan or D-xylose with 90% hydrofluoric acid yields results similar to those obtained by treatment with hydrogen fl~oride.~ The presence of 6.4% of water in hydrogen fluoride decreases the rate of solvolysis of cellulose insignificantly" at 20". At the same time, hydrofluoric acid of concentration <70% cleaves the glycosidic linkages very slowly at 20 This process constitutes an acid-catalyzed hydrolysis, and, consequently, the data on such application of hydrofluoric acid for the cleavage of glycosidic linkages will not be considered in the present article. Interestingly, 40- 60% aqueous solutions of hydrofluoric acid at 0" have been used to cleave phosphoric ester linkages specifically without cleaving glycosidic linkages (see Ref. 11 for a O .

O.

-

(13) A. Wagner, in G. A. Olah (Ed.), Friedel-Crajis and Related Reactions, Vol. 4, Interscience, New York, 1965, pp. 235-254. (14) M.P. Sanger and D. T. A. Lamport, Anal. Biochem., 128 (1983) 66-70. (15) W. R. Figures and J. R. Edwards, Carbohydr. Res., 48 (1976) 245-252. (16) A. J. Mort,J.-P. Utille,G. Tom, and A. S. Perlin, Carbohydr.Res., 121 (1983)221-232. ( I 7) M.C. Hawley, K. W. Downey, S. M.Selke, and D. T. A. Lamport, Adv. Ce//uloseChem. Techno/.,Wiley-Interscience, New York, 1985.

APPLICATION OF ANHYDROUS HYDROGEN FLUORIDE

173

review) and for de-esterifyingpolysaccharides.'* After fairly prolonged treatment, a few (random?)glycosidic linkages were cleaved, resulting in a much lessened viscosity of the polymers, and thus allowed higher-resolutionn.m.r. spectra to be rec~rded.'~.'~ 111. TECHNIQUES FOR PERFORMING SOLVOLYSIS Hydrogen fluoride is a hygroscopic liquid having a melting point of - 83 and a boiling point of 19.5'. Toxic properties and its ability to cause heavy bums on contact with the skin have created a rather Vicious reputation for this reagent and have led to some limitations in its application. However, operating with hydrogen fluoride can easily be made safe by using a completely closed system. The apparatus used for handling hydrogen fluoride usually consists of a plastic vacuum-line consisting of several reaction and storage vessels connected together with tubing and stopcocks. The construction material for the apparatus is normally a fluorocarbon plastic (Kel-F or Teflon). A trap for hydrogen fluoride, or some apparatus for its neutralization, and a vacuum pump are also needed. Ready-made devices for operation with hydrogen fluoride are offered by Peninsula Laboratories and Biosearch Companies. A special device for handling micro amounts of substance has been described.I4 Detailed accounts of the experimental procedures that have been used have been published.20-22Before conducting the reaction, the hydrogen fluoride is often dried by distilling it from cobalt(II1)trifluoride or filtering it through a layer of this salt. Several procedures are employed in order to remove hydrogen fluoride from the reaction mixture: evaporation in vucuo (favored when absorbing hydrogen fluoride with a solid alkali), neutralization with a vigorously stirred, cooled suspension of calcium carbonate in dichloromethane,or precipitationof the reaction products with cold ether. A modification of the device for solvolysis that allows the quenching of hydrogen fluoride and precipitation of the reaction products in a convenient and safe way, and control of the reaction time more precisely was reported.23It should be borne in mind that some mono- and di-saccharidesare soluble in

+

(18) A. van der Kaaden, G . J. Gerwig, J. P. Kamerling, J. F. G. Vliegenthart, and R. H. Tiesjema, Eur. J. Biochem., 152 (1985) 663-668. (19) K. Leontein, B. Lindberg, J. Liinngren, and D. J. Carlo, Carbohydr. Res., 114 (1983) 257-266. (20) J. Lenard, Chem. Rev., 69 (1969) 625-638. (21) S. Sakakibara, in B. Weinstein (Ed.), The Chemistry and Biochemistry ofAmino Acids, Peptides and Proteins, Vol. I , Marcel Decker, New York, 197 1, p. 5 1. (22) J. M. Stewart and J. D. Young, Solid Phase Peptide Synthesis, Freeman, San Francisco, 1969. (23) A. J. Mort, Carbohydr. Res., 122 (1983) 315-321.

174

YURIY A. KNIREL ef al.

the hydrogen fluoride- organic solvent mixture produced during the quenching step.I6 A mixture of hydrogen fluoride with methanol can be employed for the preparation of methyl glycosides of sugars. When there is a relatively high ratio of methanol to hydrogen fluoride, the removal of the mixture from the products by evaporation is difficult, and so the hydrogen fluoride is usually removed by neutralization.

IV. PREPARATION OF MONOSACCHARIDES The use of hydrogen fluoride in establishingthe monosaccharide composition of carbohydrates was studied" for the first time in 1977. This reagent was shown to be applicable for the quantitative determination of monosaccharides in neutral and acidic plant polysaccharides (arabinogalactan from larch, and whole cell-walls from tomato). From the observation that cellulose and chitin could be depolymerized, it was proposed that solvolysiswith hydrogen fluoride could aid in determination of the sugar compositions of polymers normally difficult to hydrolyze. A comparison of recoveries of sugars from tomato suspension-culture cell-walls after hydrolysis with 2 A4 trifluoroacetic acid, with and without previous solvolysis with hydrogen fluorideshowed a great increase in the recovery of glucose (from cellulose) after the solvolysis,but some decrease in the recovery of pentoses. To offset the possibility that these losses were due to alkylation of phenolic materials or protein, 1 mL of methanol per 10 mL of hydrogen fluoride was added, to compete for the reactive glycosyl fluorides. This returned the yield of sugars to at least that expected following ordinary hydrolysis, but still gave the S ~ the depolymerization of the cellulose.'O A subseincrease in D - ~ ~ U C Ofrom quent study also showed improved recovery of sugars using solvolysiswith hydrogen fluoride in methanol, followed by hydrolysis with trifluoroacetic acid of the methyl glycosideswhich were formed. Higher yields of D-glucose and D-glucosamine, in particular, were obtained. Analysis was made of the monosaccharide composition of the cell envelope of the fungus Erysiphe graminis (mainly D-glucose, D-mannose, and D-glucosamine), and of the cell walls of its host, namely, wheat, Triticum aestivum (mainly D-glucose, D-xylose, and ~-arabinose).'~ Hydrogen fluoride has become very important for determination of the monosaccharide composition of bacterial hexosaminoglycans.Under the conditionsfor the cleavage ofglycosidiclinkages of amino sugars, the amidic linkages of their N-acyl substituentsare stable; this was shown for the first time with chitin, which, upon depolymerization with hydrogen fluoride, afforded N-acetylglucosamine.ll This feature allows avoidance of the formation of free amino groups, which inhibit the cleavage of adjacent glycosidic linkages," and it allows isolation of amino sugars in (24) R.C. G. Moggridge and A. Neuberger, J. Chem. SOC.,(1938) 745-750.

APPLICATION OF ANHYDROUS HYDROGEN FLUORIDE

175

the form of N-acyl derivatives favorable for identification. There are now a considerable number of examples of the use of hydrogen fluoride to detect, quantitate, and, in some cases, isolate amino sugars. In examinationsof the 0-specific polysaccharidesof various serotypes of Pseudomonas aeruginosa, solvolysis with hydrogen fluoride was used to show the presence of 2-acetamido-2-deoxy-~-galactose,2-acetamido-2,6-dideoxy-~-glucose~~ 2-acetamido-2,6-dideoxy-~-galactose~~ and 2-acetamido-2-deoxy-~-galacturonic acid.272-Acetamido-2-deoxy-~-mannurono-6,3-lactone was obtained28on solvolysis of the group K capsular polysaccharide of Neisseria meningitidis, which is built up of 2-acetamido-2-deoxy-~-mannuronic acid residues. A series of rare amino sugars having unusual N-acyl substituents, residues of hydroxy acids and amino acids, were also isolated by the same method. Examples of these types of compounds are 3,6-dideoxy-3-(~-glycof the cell envelope eroy1amino)-D-galactose(7)from the p~lysaccharide~~ of Eubacterium saburreum L13 and components of several 0-specific poly(8) saccharides: 4,6-dideoxy-4-[(R)-3-hydroxybutanamido]-~-galactose from’O Escherichia coli 0 : 10,2-acetamido-2,4,6-trideoxy-4-[(S)-3-hydroxybutanamido]-~-ghcose(9) from31P. aeruginosa Habs 0 :3,4,6-dideoxy-4[(S)-2,4-dihydroxybutanamido]-~-mannose(10) from32 Vibrio cholerae 0 : 1, 4-(N-acetylglycyl)amido-4,6-dideoxy-~-glucose (11) from33 Shigella dysenteriae type 7 , and 3-(N-acetyl-~-seryl)amido-3,6-dideoxy-~-glucose (12) from34E. coli 0 : 1 14.

(25) B. A. Dmitriev, N. A. Kocharova, Yu. A. Knirel, A. S. Shashkov, N. K. Kochetkov, E. S. Stanislavsky, and G. M. Mashilova, Eur. J. Biochem., 125 (1982) 229-237. (26) Yu. A. Knirel, N. A. Kocharova, A. S. Shashkov, B. A. Dmitriev, N. K. Kochetkov, E. S. Stanislavsky, and G. M. Mashilova, Eur. J. Biochem., 163 (1987) 639-652. (27) Yu. A. Knirel, E. V. Vinogradov, A. S. Shashkov, B. A. Dmitriev, N. K. Kochetkov, E. S. Stanislavsky, and G. M. Mashilova, Eur. J. Biochem., 125 (1982) 221 -227. (28) F. Mishon, J. R. Brisson, R. Roy, H. J. Jennings, and F. E. Ashton, Cun.J. Chem., 63 (1985) 2781-2786. (29) L. Kenne, B. Lindberg, C. Lugowski, and S. Svensson, Curbohydr. Rex. 151 (1986) 349-358. (30) P.-E. Jansson, B. Lindberg, M. Spellman, T. Hofsted, and N. Skaug, Curbohydr. Res., 137 (1985) 197-203. (31) Yu. A. Knirel, E. V. Vinogradov, A. S. Shashkov, S. G. Wilkinson, Y. Tahara, B. A. Dmitriev, N. K. Kochetkov, E. S. Stanislavsky, and G. M. Mashilova, Eur. J. Biochem., 155 (1986) 659-669. (32) L. Kenne, B. Lindberg, P. Unger, B. Gustafsson, and T. Holme, Curbohydr. Res., 100 ( 1982) 34 1 - 349. (33) Yu. A. Knirel, V. M. Dashunin, A. S. Shashkov, N. K. Kochetkov, B. A. Dmitriev, and I. L. Hofman, Curbohydr. Res., 179 (1988) 51-60. (34) B. A. Dmitriev, V. L. Lvov, N. V. Tochtamysheva, A. S. Shashkov, N. K. Kochetkov, B. Jann, and K. Jann, Eur. J. Biochem., 134 ( 1983) 5 17 - 52 1.

YURIY A. KNIREL et ul.

176

OCNH @OH

OH HCH

HOCH

I

OH

CH2OH

I

I I

OH

CH3

a

7

0

9

(aH

H

I

I OH AcHNCH

HOCH

I

AcHNCH2 OCHN @OH

HCH

NHAc

HOCH

HCOH CH3

OCNH b

I

HCH

Ho

I

OH

CH20H

10

I CH2OH

11

12

Some glycosidiclinkages have proved stable to hydrogen fluoride solvolysis. Attempts35to isolate monosaccharides by solvolysis of the 0-specific polysaccharide(13)of Shigellu sonnei phase 1 with hydrogen fluoride and of the capsular p~lysaccharide~~ (14) of Streptococcus pneumoniue type 1 failed. Prior acetylation of the free amino groups of the diamino sugar did not render the polymer susceptible to the solvolysis.

+3 ) + S U P ( 1+ 4)-*L-AltpNAcA-(

1-

13

(35) L. Kenne, B. Lindberg, K. Petersson, E. Katzenellenbogen,and E. Romanovska, Curbohydr. Rex, 78 (1980) 119- 126. (36) B. Lindberg, B. Lindquist, J. Lonngren, and D. A. Powell, Curbohydr. Rex, 78 (1980) 1 1 1 - 1 17.

APPLICATION OF ANHYDROUS HYDROGEN FLUORIDE -13)-eSug~-( 1

177

4)-a-MalpA-(1+3)-a-MdpA-(l-,

14

Oligosaccharide fragments containing amino sugars were isolated upon solvolysis of the 0-specific polysaccharides of P.aeruginosa G n y i 0 :3a, 3b, and 0 :3a, 3d (30)( Ref. 37), 0 :4a, 4c (32)(Ref. 38), and Habs 0 :3 (Ref. 3 1). On solvolysis of P. aeruginosa Unyi 0 :2 polysaccharide(15), only half of its amino sugars were obtained in the free state, while the other half were retained2’ as the disaccharide fragment 16. +3)-a-L-Rhap-(

1-+4)-a-~-GalpNAcA-(1-+3)-a-D-QuipNAc-( 1 -+ 15

monosaccharides + a-~-GalpNAcA-(l-+3)-~-QuipNAc

16

All of these polysaccharides contain alduronic acids. It is reasonable to suggest that the presence of the acid group confers enhanced stability on the adjacent glycosidic linkage. This suggestion was supported by experiments on the solvolysis of polysaccharides [from P. aeruginosa Gnyi 0 :4 (Ref. 38), and Habs 0 :3 (Ref 3 1)] in which the uronic acids had been carboxyl-reduced. After the reduction, hydrogen fluoride cleaved all of the glycosidic linkages, providing monosaccharides including 2-acetamido-2-deoxy-Dgalacturonamide (17)(Ref. 38), and a derivative of the diamino sugar 9 (Ref. 3 1). (37) Yu. A. Knirel, E. V. Vinogradov, A. S. Shashkov, B. A. Dmitriev, N. K. Kochetkov, E. S. Stanislavsky,and G . M. Mashilova, Eur. J. Biochem., 128 (1982) 81 -90. (38) Yu. A. Knirel, E. V. Vinogradov, A. S. Shashkov, B. A. Dmitriev, N. K. Kochetkov, E. S. Stanislavsky, and G . M. Mashilova, Eur. J. Biochem., 150 (1985) 541 -550.

178

YURIY A. KNIREL et al. CONH;, I

NHAC

17

The glycosidic linkages of derivatives of 5,7-diamino-3,5,7,9-tetradeoxyL-glycero-L-munno-nonulosonic (pseudaminic) acid proved stable towards hydrogen fluoride at room t e m p e r a t ~ r eIt. ~has ~ been possible to isolate free 5-N-acetyl-7-N-formylpseudam~nic acid (18) from P.ueruginosa immunotype 6 0-specific polysaccharide on solvolysis of the polysaccharide with hydrogen fluoridein methanol, followedby hydrolysiswith dilute acetic acid of the methyl glycosides of oligosaccharides having the derivative of pseudaminic acid on the nonreducing end.26 Peripheral sialic acids, which are derivatives of ~-amino-3,~-dideoxy-~-g~yct?ro-D-gu~ucto-nonulosonic (neuraminic) acid, could be removed".'" from glycoproteins on treatment thereof with hydrogen fluoride. However, no attempt was reported to show cleavages of the actual glycosidic linkage of the sialic acid. Cleavage of various other linkagesinternal to the sialic acid would lead to the same result. The lability of the glycosidiclinkages of the more-common neuraminic acid needs closer examination.

OH OHCHN~H

I I

HOCH CH3

18

Solvolysis with hydrogen fluoride was successfully employed to obtain partially methylated sugars in the methylation analysis of carbohydrates. (39) Yu. A. Knirel, E. V. Vinogradov, V. L. Lvov, N. A. Kocharova, A. S. Shaskov, B. A. Dmitriev, and N. K. Kochetkov, Curbohydr.Res., 133 (1984) c5-c8. (40) A. J. Mort, CellSuflace Carbohydratesand BiologicalRecognition,Liss, New York, 1978, pp. 553-561.

APPLICATION OF ANHYDROUS HYDROGEN FLUORIDE

179

Methylated 0-specific polysaccharides of Vibrio choleraeZ9and Yersinia enterocoliticu”’were depolymerized in this way. Both are homopolymers of N-acyl derivatives of 4-amino-4,6-dideoxy-~-mannose, and methylated derivatives (19 and 20) of this monosaccharide having the corresponding Nacyl substituentscould be isolated. Acid hydrolysis,usually employed for the cleavage of methylated polysaccharides would, in this case, be hampered owing to easy N-deacylation and, hence, stabilization of the glycosidic linkages by formation of the protonated methylamino

O

O

H

MeOCH MeN co I

b MeN

I

CHZ

O

H

CHO

I

CHSMe

19

20

The ability to retain N-acyl substituentson amino sugars makes hydrogen fluoride practically indispensable for methylation analysis of carbohydrates in which neighboring monosaccharideresidues are linked through hydroxyacyl groups. Thus, on analysis of one such polysaccharide (21) from Eubacterium saburreum L13, the partially methylated derivative (22) was obtained the free hydroxyl group at the residue of ~-glycericacid in the latter proved to be the linkage point of the neighboring monosaccharide to the N-acyl substituent.m (1) Methylation

oc -0CH

I

I

CH20Me

CH20H

21

OMe

22

Isolation of the alditol (23) on solvolysis with hydrogen fluoride of the methylated oligosaccharide fragment of P. aeruginosa E n y i 0 : 1Oa O-spe(4 1 ) M. Caroff, D. R. Bundle, and M. B. Perry, Eur. J. Biuchem., 139 (1 984) 195 -200.

180

YURIY A. KNIREL et al.

cific polysaccharide (25) showed that the linkage between the derivative of pseudaminic acid and the neighboring monosaccharide is accomplishedby way of the N-acyl derivative having the (R)-3-hydroxybutanoyl Similar analysis of the oligosaccharide fragment from P.aeruginosa Unyi 0 :5a, 5b polysaccharide resulted in another alditol (24) whose structure showed that the derivative of pseudaminic acid is glycosylated at 0-4, and that the hydroxyl group of the (R)-3-hydroxybutanoylresidue is, in this case, unsubstituted.26 CH20Me I

HCO Me I

HCH

I HYO Me HCN MeAc MeOCH I

HCO Me I HCH

I

HCOH

I

HCNMe MedH' CO I I AcMeNCH HCH

I

I

MeOCH HCOMe

I

CH3

23

t

CH3

24

V. PREPARATION OF OLIGOSACCHARIDES In most cases, structural elucidation of polysaccharides requires obtaining oligosaccharide fragments. It is usually much easier to determine the structure of oligosaccharides than of polysaccharides, especially since, among other advantages, they usually give high-resolution n.m.r. spectra and are amenable to mass spectrometry. As was pointed out earlier, the lability of the glycosidiclinkages of various sugars to hydrogen fluoride solvolysis varies greatly. In many cases, this allows complete cleavage of one or more linkage-types in a polysaccharide, with no cleavage of others. By going to lower and lower temperatures, more and more glycosidic linkagesbecome resistant. When working with bacterial polysaccharides, or other polysaccharides of repeating structure, this makes possible the generation of unique oligosaccharides in high yield. As was mentioned in Section IV, the glycosidic linkages of derivatives of 5,7-diamino-3,5,7,9-tetradeoxynonulosonic acids are stable towards hydrogen fluoride at 20". This allows easy generation of oligosaccharideshaving

+

APPLICATION OF ANHYDROUS HYDROGEN FLUORIDE

181

this sugar at the nonreducing terminus. Solvoly~is~~ under these conditions of the Pseudomonasaeruginosa Unyi 0 : 1Oapolysaccharide(25) proceeded highly selectively, to yield the disaccharide 26.

25

~HAC

H ~ H &H3

H

h LH3

26

Solvolysiswith hydrogen fluorideat room temperature of polysaccharides containing uronic acids and their derivatives also usually leads to useful with the 0-specific polysacchaoligosaccharides.This may be e~ernplified~~ ride (27)of Proteurn mirabilis 027. Only the glycosidiclinkages of the N-acetylglucosamine residues are cleaved at 20°,while the glycosidic linkages of

+

(42) Yu. A. Knirel, E. V. Vinogradov, A. S . Shashkov, B. A. Dimitriev, N. K. Kochetkov, E. S . Stanislavsky, and G. M. Mashilova, Eur. J. hiochem., 157 (1986) 129- 138. (43) E. V. Vinogradov, D. Pietrasik, A. S. Shashkov, Yu. A. Knirel, and N. K. Kochetkov, Bioorg. Khim., 14 (1988) 1282- 1286.

182

YURIY A. KNIREL et al.

the hexuronamides are stable. Solvolysis with hydrogen fluoride causes dephosphorylation of the polysaccharide, whereas the amino acids linked by amidic bonds, as well as N-acetyl groups, are not removed. As a result, the oligosaccharides 28 and 29 are obtained.

FW

YN7"

NHAc

HF,'0

29

Other examples are some Pseudomonas aeruginosa 0-specific polysaccharides containing amino and diamino derivatives of uronic acids. Thus, the polysaccharide (30)o f u n y i serotype0 :3a, 3d under the action of HF for 3 h at 20"underwent strictly selective~ l e a v a g e ~at' ,the ~ ~glycosidic ~ linkages of N-acetylfucosamine,to give the trisaccharide 31 in practically quantitative yield. (43a) Yu. A. Knirel, N. A. Paramonov, E. V. Vinogradov, A. S. Shashkov, B. A. Dmitriev, N. K. Kochetkov, E. V. Kholodkova, and E. S. Stanislavsky, Eur. J. Biochem., 167 (1987) 549-561.

APPLICATION OF ANHYDROUS HYDROGEN FLUORIDE

183

H3C

31

Solvoly~is~~ of the polysaccharide (32) of P. aeruginosa Unyi 0 :4a, 4c with hydrogen fluoride at room temperature led to the trisaccharide fragment 33 as the only oligosaccharide product. The glycosidic linkage of 2-acetamido-2-deoxymannuronicacid is less stable than the linkages of the other uronic acid derivativesjust considered, and even less so than those of N-acetylhexosamines. Because of this, it has been possible4 to isolate the disaccharide 35, having 2-acetamido-2-deoxyD-mannuronic acid on the reducing end, from the polysaccharide (34) of Haemophilus injluenzae type e by reaction with hydrogen fluoride at 20". The analogous disaccharides were obtained on solvolysis, under the same (44)K. Leontein, B. Lindberg, and J. Lonngren, Can J. Chem., 59 (1981) 2081-2085.

W R I Y A. KNIREL et al.

184 CONH2

I

CONH2 I

I

HF

33

conditions, of the capsular polysaccharides of Streptococcus pneumoniae types 12F (Ref. 44) and 12A (Ref. 19). The glycosidic linkage of 2-acetamido-2-deoxymannuronicacid may be retained by performing solvolysisat decreased temperature. Thus, treatment with HF for 15 min at -20" of the enterobacterialcommon antigen representing a cyclic polysaccharide involving 4 - 6 repeating trisaccharide units of the structure 36 afforded trisaccharide 37 having the uronic acid derivative on the nonreducing end.45At 20", the glycosidic linkages of this sugar (45) C. Lugowsky, E. Romanowska, L. Kenne, and B. Lindberg, Carbohydr.Rex, 1 18 (1983) 173-181.

APPLICATION OF ANHYDROUS HYDROGEN FLUORIDE

+~)-/?-DGIc~NAc-(~-A)-&D-ManpNAcA-(

185

1+

34

,3-lH3lcpNAc-( 1-A)-D-MarpNAcA 35

proved to be unstable, and, in 15 min, disaccharide 38 was formed. It is noteworthy that, at - 20", more than half of the 0-acetyl groups present in the polysaccharide were retained. Experiments on the 0-deacetylated capsular polysaccharide (39) of Neisseria meningitidis group I showed that the glycosidic linkage of 2-acetamido-2-deoxyguluronicacid is less stable than that of 2-acetamido-2-deoxymannuronic acid. Owing to favored cleavage of the glycosidiclinkage of the former sugar, the disaccharide 40,having the latter sugar on the nonreducing end, was obtained& after reaction with HF for 2 h at 25 O.

-A)+D-MqNAcA-(

1 +~)-c+D-G~c~NAc-(1+ 3 t ~ + S u e 1( 3 6 I OAc0.7

36

I

HF,-20'

I

HF,20'

38

(46) F. Mishon,J. R. Brisson, R. Roy, F. E. Ashton, and H. J. Jennings, Biochemistry, 24( 1985) 5592-5598.

186

YURIY A. KNIREL et ul.

+3>-,%-D-ManpNAcA-(l -A)-a-LGulpNAcA-(l+

39

IHF

,%-WManpNAcA-41-A)-L-GulpNAcA 40

The different stability of the glycosidiclinkage of 2-acetamido-2-deoxy-~mannuronic acid towards hydrogen fluoride at room temperature in H . injluenzaeand S.pneumoniae polysaccharideson the one hand and in the N. meningitidis polysaccharide on the other, may be accounted for by the different neighboring sugar residues. The glycosidic linkages of the commonly occurring amino sugars N-acetyl-D-glucosamineand N-acetyl-D-galactosamineare stable in hydrogen fluoride at 0" and below." Thus, partial solvolysis with hydrogen fluoride at - 30" of the Escherichia coli 078 0-antigen polysaccharide (41) cleaved the glycosidic linkages of the D-mannosyl residues, but not of the residues of N-acetylglucosamine, to give the trisaccharide fragment 42 in high yield.47

@oGlcpNAc-( 1-A)-@DGlcpNAc-(l

-A)-WManp

42

The stabilization of glycosidiclinkages in polysaccharides having an acylamido group at C-2 is higher than that of those having it in any other position. This was demonstrated by solvolysis with hydrogen fluoride at 0" of the 0-specificpolysaccharide(43) ofPseudomonasJluorescens36 1, which on led to the disaccharide 44 having 2-acetamido-2,6-dideoxy-~-galactose (47) P.-E. Jansson, B. Lindberg, G. Widmalm, and K. Leontein, Curbohydr.Res., 165 (1987) 87-92.

APPLICATION OF ANHYDROUS HYDROGEN FLUORIDE

187

the nonreducing end. No disaccharide having 3-acylamido-3,6-dideoxy-~glucose on the nonreducing terminus was f ~ r r n e d . ~ ~ , ~ ~

A3

I

t

HF

OH

44 (48) V. A. Khomenko, G . A. Naberezhnych, V. V. Isakov, T. F. Soloveva,Yu. S.Ovodov, Yu. A. Knirel, and E. V. Vinogradov, Bioorg. Khim., 13 (1987) 1641 - 1648. (49) G . A. Naberezhnych, V. A. Khomenko, V. V. Isakov, Yu. N. Elkin, T. F. Soloveva, and Yu. S . Ovodov, Bioorg. Khim., 13 (1987) 1428- 1429.

YURIY A. KNIREL et al.

188

Study ofthe 0-specific polysaccharide(45) of Pseudomonasaurantiaca 3 1 showed that the presence of a second acylamido group causes an additional stabilization of the glycosidic linkages of amino sugars. The glycosidic linkresidues were split comages of the 2-acetamido-2,6-dideoxy-~-galactosyl pletely under the action of hydrogen fluoride at room temperature, and disaccharide 46, having 2,4-diacetamido-2,4,6-trideoxy-~-glucose at the nonreducing end, was obtained as the only oligosaccharide product.50 +~)+SU~P.(

1 +~)-u-L-Fuc~NAc-( 1 -~~)-u-L-Fuc~NAc-( 1+ 45 CH3

I

b S u g ~ -1+3)-L-FucNAc (

NHAc

46

The glycosidiclinkagesof N-acetylhexosaminesare more stablethan those of their 6-deoxy derivatives (such as N-acetylquinovosamine and N-acetylfucosamine).Thus, owing to the facile cleavage of the glycosidic linkages of -A)-a-UalpNAc-(l-A)-pSugp-(

1+3)-a-~FucpNAc-(1+3)-a-D-QuipNAc-(

1+

41

I a-ffialpNAc-( l-A)-~Sugp-(l-+3)-WFucpNAc

HF,20'

p S u g p ( 1+3)-&FucpNAc 49

48

kHAc

(50) Yu. A. Knirel, G . M. Zdorovenko, S. N. Veremeichenko, G. M. Lipkind, A. S. Shashkov, and I. Ya. Zakharova, Bioorg. Khim., 14 (1988) 352-358.

APPLICATION OF ANHYDROUS HYDROGEN FLUORIDE

189

2-acetamido-2,6-dideoxyhexoses, the trisaccharide fragment 48 having Nacetyl-D-galactosamine on the nonreducing end was obtained under the action of hydrogen fluoride during 1 h at 0" on the P. aeruginosu Unyi 0 :6 polysaccharide (47). The disaccharide 49 was also generated, and it became the main product on performing solvolysis under more drastic conditions, namely, 3 h at 20". The same difference in stability of glycosidic linkages is characteristic for hexoses and 6-deoxyhexoses, as well. Thus, the lability of the glycosidic linkage of rhamnose was used to cleave selectively,by the action of hydrogen fluoride at -40°, the extracellular polysaccharide (50)from P. ueruginosu immunotype 4. The glycosidic linkages of all the hexose residues were found to be stable, and solvolysisgave, in good yield, the pentasaccharide 51, which constitutes the repeating unit of the polysa~charide.~' -+3)-p-DGlcp-(

1 +3)-&D-Manp-( 2

1 +3)-&~-Manp-(

1 +3 ja-L-Rhap-( 1 --f

T 1

a-D-Manp

so

p-~-Glcp(l-t3)-&~-Manp-(l+3 jp-D-Manp( 1-+3)-~-Rha 2

t 1 *&May

51

Experiments with the Rhizobium juponicum 311b 83 extracellular polysaccharide(52)showed that, at -40", the glycosidiclinkages of 6-deoxyhexoses may be retained, at least partially. Indeed, under these conditions, solvolysis gave a combination of the repeating unit 53 and disaccharides of rhamnose and of (4-0-methylglucosyluronica~id)rhamnose.~~ At - 23', this polysaccharide was degraded to rhamnose and the last-named disaccharide. ( 5 1) N. A. Kocharova, Yu. A. Knirel, A. S. Shashkov, N. K. Kochetkov, and G. B. Pier, J. Biol. Chem., 263 (1988) 11,291 - 11,295.

190

YURIY A. KNIREL et al.

4j ~ L - R h a g - ( l - 1 3 > - ~ L - R h a p ( l - t 4 j ~ ~ R1-1h ~ 3

T 1

44-Me-pbGlcpA 52

!

HF,-40'

/%L-Rh&

l-A)-pL-Rh+(

14)-L-Rhq

3

t 1 4-O-Me-pMlcpA

53

At - 23", there are distinct differences in the susceptibility of a and p linkages of h e x o s e ~This . ~ ~allowed an almost quantitative yield of the trisaccharide 55 to be obtained5' from another extracellularpolysaccharide(54) of R. juponicum 311b 138. A striking feature was the stoichiometric amount of acetate retained, which was easily located by using methylation with methyl trifluoromethanesulfonate coupled with reduction of the galacturonic acid methyl ester. At the even lower temperature of -4O", the predominant cleavage was at a unique site in the repeating pentasaccharide that gave rise to a high yield of the repeating unit 56. By a minor degree of loss of the D-galactosyl side-chain, the repeating tetrasaccharide unit of the backbone was obtained. Experiments with the extracellular polysaccharides from Rhizobium leguminosarum and Rhizobium trifolii again showed the marked difference in stability of the glycosidic linkages of a-and /?-linked hexoses. The polysaccharides of both of these species have the same eight-sugar repeating-unit structure (57) and, in addition, contain 0-acyl substituents. After treatment in hydrogen fluoride for 15 min at -40",they are predominantly cleaved at one site per repeating unit, namely, the sole a-linked D-glucosyl residue, to Despite the identity of the give the octasaccharides58 and 59, re~pectively.~~ fundamental sugar structure of the polysaccharides, the polymer from R. (52) A. J. Mort, Abstr. Pap. Am. Chem. Soc. Meet., 181 (1981) C A R B ~ ~ . (53) A. J. Mort and W. D. Bauer, J. Biol. Chem., 257 (1982) 1870- 1875. (54) M.-S. Kuo and A. J. Mort, Carbohydr. Rex, 145 (1986) 247-265.

APPLICATION OF ANHYDROUS HYDROGEN FLUORIDE

191

OMe

1

4

a-D-Galp 1

OAc

L

1

6 4 +3ja-D-Gl cp(l-t3)-~Mlc p-( 1+3)-cr-D-GalpA-(

1+3)-u-D-Manp-(

1+

54

I

HF, -40'

HF, -23'

OAc

I 4

e N l c p - ( 1-3 jc-D-GalpA-( 1+3 j r r M a n p 55

OMe

I 4

a--alp OAc

1

I

I

4

6

pD-Glcp-( 1 +3)-cx-&GalpA-( 1+3ja-D-Manp-( 1+3jD-Glcp 56

leguminosarum was not cleaved to the repeating unit to the same extent as that of R . trifolii. This behavior could be explained by the attachment to the a-D-glucosyl residue in the former polysaccharide of an 0-acetyl group affecting the accessibility to, or the reactivity with, hydrogen fluoride. This group is absent from the latter polysaccharide. 'H-N.m.r. spectroscopy of 58 and 59 allowed the location of all of the ester groups, and proved that these octasaccharidesend in a fluoride, not in a reducing D-glucose residue. Subjection of these polysaccharides to hydrogen fluoride solvolysis for 15 min at - 23 gave a mixture of oligosaccharidesfrom which could readily be purified a tetrasacchariderepresentingthe backbone, one representingthe sidechain, and a trisaccharide representing the side-chain minus the terminal D-galactosyl residue. The behavior of the pyruvic acetals on the O

YURIY A. KNIREL et al.

192

57

I

HF, -40'

OAc

R1 R2

1

1 1

3 2 3 1+F pD-GlcpA-( 1+4tpWGlcpA-( 1+4)-&D-Glck( 1+ 4 t * M l c p - ( 6 -r

3-h ydroxybutanoyl

1 3 1 p&Galp-( l+3)-/&WGlcp-( 1+4)-pD-Glcp-(l-+4)-~D--Glcp 6, /4 C Me' 'C02H 58

R~=R*=ACO

59

R1=R2=OH

side-chainswas not uniform. The 4,6-pyruvicacetalated galactosylgroup retained the pyruvic acetal, while the 4,6-pyruvic acetalated glucose did not. Gellan gum (60)was found to be specifically cleaved by hydrogen fluoride at -40" at the a-L-rhamnosyl linkage, to generate the tetrasaccharide 61 which was partially acetylated but also acylated by L-glyceric acid.55Treatment at -23" afforded the trisaccharide fluoride 62, which was also acylated. Some loss of glycerate, and possibly of acetate, occurred during the (55) M A . Kuo, A. Dell, and A. J. Mort, Carbohydr. Rex, 156 (1986) 173- 187.

APPLICATION OF ANHYDROUS HYDROGEN FLUORIDE

193

generation of these oligosaccharides,but there did not appear to be any acyl migration. Quantitation of acyl substituents was performed on the intact polymer, but localization was performed on the (more tractable) oligosaccharides.

A disturbing side-reaction took place during the treatment of gellan gum at both -40" and -23 The majority of the polymer did not remain as oligosaccharides. Instead, it appeared that, in most cases, after a rhamnosyl linkage had been solvolyzed,it then reacted with a nearby hydroxyl group to re-form a polymer. The resulting material contained the repeating tetrasaccharide units of the original polymer linked together in many different ways. Another instance in which unexpected oligosaccharides containing Lrhamnose were produced has been observed.56Treatment of cell walls of plant suspension-cultureswith hydrogen fluoride at -23" allowed easy isolation of an acidic disaccharide containing D-galacturonic acid and L-rhamnose. The disaccharidewas 30 percent acetylated at 0-3 of the uronic acid residue. By the unusual chemical shift of H-3 of the galacturonic acid residue, and methylation analysis, the disaccharide was shown to be the dianhydride 63. The fully acetylated &anhydride of the same sugars had been isolated previou~ly,~' after methanolysis and acetylation of a pectic polysaccharide or of the disaccharide a-D-galacturonic acid-(1+2)-~-rhamnose. O.

-

(56) P. Komalavilas and A. J. Mort, Curbohydr. Res., 187 (1989) 261 -272. (57) T. Fujiwara and K. Arai, Curbohydr. Res., 69 (1979) 97- 105.

YURIY A. KNIREL et al.

194

63

At just about the freezing point of hydrogen fluoride, in the instances tested, only pentofuranose linkages are broken. Table I shows the extent of the predictions that can be made as to which glycosidic linkages will be stable at any particular temperature in anhydrous hydrogen fluoride. Some of the data used to generate the Table are from unpublished results, and, as has been pointed out earlier, such factors as the TABLEI

Expected Labilitiee of the Glycosidic Linkages of Various Sugar Residues' Temperature of hydrogen fluoride ("C) Sugar residue

Pentofuranose Pentopyranose 6-Deoxyhexose a-Hexose &Hexox 2-Amino-2,6-dideoxyhexose 3-Amino-3,6-dideoxyhexose 4-Amino-4,6-dideoxyhexose 2,4-Diamino-2,4,6-tndeoxyhexose 2-Amino-2deoxyhexose Uronic acid Galactosaminuronic acid Mannosaminuronic acid 2,3-Diamino-2,3-dideoxyalduronic acids 5,7-Diamino-3,5,7,9-tetradeoxynonulosonic acid

<-70

-40

-23t0-20

0

20-25

+ + + + + + +

+ + + + + + + + +

+

- e f f +

-

*

-

+-

-

-

a The labilities indicated are generalizations from the limited number of cases that have been investigated. Branching, anomeric configuration, and some unidentified factors affect the labilities. Key: (+), cleavage; (-), stable; (+), depends on structural peculiarities. All of the amino sugars were tested in the N-acylated form.

APPLICATION OF ANHYDROUS HYDROGEN FLUORIDE

195

previous sugar in the chain, and branching, may have an effect on the stability of a glycosidic linkage. A particularly striking example of a sugar linkage that does not fit the labilitieslisted in Table I is the case of an N-acetyl-P-D-galactosamineresidue in the 0-specific polysaccharide (64) of E. coli 086. After reaction with hydrogen fluoride for 20 min at -5O", and reduction, the di-, tri-, and tetra-saccharide (65, 66, and 67) were isolated.58All of these contained GalNAc-ol at the reducing end, indicating favored cleavage of the P-linked amino sugar. +4)-a-L-Fucp-(

1+2&?-Mslp-(

1+3)-a-MalpNAc-( 1+3)-&ffialpNAc-(

1+

3

t 1 a-Malp 64

I

HF, -50" Na BH,

a-WGalpNAc-( 1+ 3 t M a l N A c + l + p&Gdp-( 1+3)-ewGalpNAc-( 66

65

a-&Galp(l+3)-&Malp-(

1+3)-MalNAc+l

1-13)-a-wGalpNAc-(l+3~MalpNAc+l

67

Solvolysis with hydrogen fluoride in methanol can also be employed for the preparation of oligosaccharides.In this case, essentially the same regularities in the stability of glycosidic linkages of different monosaccharides are observed as with pure hydrogen fluoride. For example, the glycosidic linkages of 6-deoxy sugars are easily cleaved, whereas those of derivatives of 2,3-diamino-2,3-dideoxyuronic and 5,7-diamino-3,5,7,9-tetradeoxynonulosonic acids are retained at room temperature. Thus, solvolysisofthe modified P.ueruginusu immunotype 7 0-specific polysaccharide (68)proceeded strictly selectively at the glycosidiclinkages of N-acetylfucosamine to give59a (58) M. Anderson, N. Carlin, K. Leontein, U. Lindquist, and K. Slettengren,Carbohydr.Res., 185 (1989) 21 1-223. (59) E. V. Vinogradov, N. A. Paramonov, and Yu.A. Knirel, unpublished results.

I96

W R I Y A. KNIREL et al.

mixture of methyl glycosides(aand Bpyranosides,and p-furanoside)of the trisaccharide fragments 69 - 71.

v 7-

-0 O "& r

I

NHAc

68

HF , CH,OH

70 R'= OMe. F? = H

71

Treatment with 1 : 1 hydrogen fluoride-methanol for 1.5 h at 20" of the polysaccharide (72) from P.aeruginosa immunotype 6 also resulted in full cleavage of glycosidiclinkages of N-acetylfucosamine and in partial cleavage of those of D-XylOSe. The linkages of 5-N-acetyl-7-N-formylpseudaminic acid were not affected.26As a result, a mixture of methyl glycosides 73 - 75 of trisaccharides and of 76 and 77 of disaccharides was formed. The formation of several methyl glycosides for each oligosaccharide fragment in these cases may be exolained by the reaction being terminated prior to equilibration.The data on solvolysiswith hydrogen fluoride in methanol of P.aeruginosa Linyi 0 :13 0-specific polysaccharide(78) gave evidencein

APPLICATION OF ANHYDROUS HYDROGEN FLUORIDE

- A j a - S u g ~ - ( 2 + 4 j ~ D - X y l p - ( 1 +3)-&D-FucpNAc-(l+ 72

1

HF,CH30H

a-S ugp-(2+4kp-tbXylp-( 1+3)-p-~-FucpNAc-( 1 +OMe 73

a-Sugp-(2+4 jp-D-Xylp( 1+3)-a-D-FucpNAc-(

1 &Me

74

a-Sugp-(2+4tp-~-Xylp-( 1+3)-p-D-FucfNAc-( 75

a-Sugp-(2+4)+D-Xylp-(

1+OMe

76

a-Sugp-(2+4ja-tbXylp-( 1 +OMe 77

sug

P

AcHN

OHCHNCH

I

1 +OMe

197

YURIY A. KNIREL ef ul.

198

favor of this suggestion.60Under relatively mild conditions, namely, 3 h at 20",a mixture of methyl a-and /?-glycosides(79 and 80) of the disaccharide fragment built up of 5,7-diacetamido-3,5,7,9-tetradeoxy-~-glycero-~-galacto-nonulosonicacid and the acetamidoyl derivative of L-fucosamine was obtained, as well as the methyl a-glycoside 81 of the trisaccharide fragment containing, in addition, N-acetylquinovosamine. The reaction at 40 within the same time period affordedonly the methyl a-glycoside (79) of the disaccharide. O

79

R'-H. $=Me0

80

R'=MeO,d=H

81

The methyl a-glycoside (82) of the disaccharide fragment, which appears to be thermodynamically the more stable, was formed6' preponderantly on (60) Yu. A. Knirel, E. V. Vinogradov, A. S. Shashkov, B. A. Dmitnev, N. K. Kochetkov, E. S. Stanislavsky, and G. M. Mashilova, Eur. J. Biochem., 163 (1987) 627-637. (61) Yu. A. Knirel, E. V. Vinogradov, A. S. Shashkov, N. K. Kochetkov, V. L. Lvov, and B. A. Dmitnev, Curbohydr. Rex. 141 (1985) cl -C3.

APPLICATION OF ANHYDROUS HYDROGEN FLUORIDE

I99

solvolysis of the polysaccharide (25) of P.ueruginosu 0: 10a with hydrogen fluoride in methanol for 3 h at 20".

0 nu

kHAc

HCOH

cr,

82

Methyl glycosides of oligosaccharides can also form at much lower temperatures in hydrogen fluoride- methanol. Gellan gum (60) gave rises5to only a single tetrasacgharide methyl glycoside (83) on treatment for 15 min at -40". OAc1n

1 6 p D - C l c p ( 1+ 4 & p M l c p A - ( 1+ I & p D - C l c p ( 1-i4)-cl-~-Rhap-( 1+OMe 2

t L-Glyceroyl

83

A branched tetrasaccharide methyl glycoside (85) was isolated from the Alcaligenesfaecalis 0-specific polysaccharide(84) by the action6*of hydrogen fluoride-methanol for 30 min at -78". Under the action of hydrogen fluoridein methanol, the carboxyl groups of acidic sugars may or may not become esterified (oligosaccharides 69-71, and 82 become methyl esters; oligosaccharides 73 -77, 79 -81, and 83 do not). Esterification is evidently dependent on structural peculiarities of the acidic sugar.

(62) Yu. A. Knirel, G. M. Zdorovenko, A. S . Shashkov, I. Ya. Zakharova, and N. K. Kochetkov, Bioorg. Khim.,12 (1986) 1530-1539.

YURIY A. KNIREL ef al,

200

+3)-eWRhw(

1+2)-*&Rhw(

l+3)-~~WRhw-( l+ 2

t

84

I

HF, CH,OH

4

a-WRhap-(l+2)-a-&Rhap(

1+3)-a-&Rhap-(

1+OMe

85

VI. OTHERAPPLICATIONS OF HYDROGEN FLUORIDE Because of the specificity of hydrogen fluoride for glycosidic linkages over amide bonds, it can be used to deglycosylate glycoproteins." The degree of deglycosylation depends on the temperature of the solvolysis. Almost all types of glycopeptide linkage have been tested. Results with glycopeptides from the plant cell-wall protein showed that the glycosidic linkages of galactose to 0 - 3 of serine and of arabinose to the oxygen atom of hydroxyproline are readily cleaved within 1 h at 0 .However, after 1 h at 0 the oligosaccharide side-chains of fetuin were only partially removed. The 0-linked sidechains (86) were removed, except for the N-acetyl-D-galactosamine linked directly to serine or threonine residues. The N-linked side-chains (87) were O

a-NeuSAc-(2+3)-p-Malp-(

O ,

1+4~a--D--Cia!~NAc( l&)-[Ser 6

T 2 a-NeuSAc

86

or Thr]

APPLICATION OF ANHYDROUS HYDROGEN FLUORIDE

20 1

87

cleaved to chitobiose units linked to asparagine residues. After 3 h at room temperature, the N-acetyl-D-galactosaminewas also cleaved from the serine and threonine residues. The chitobiose unit on the asparagine residues was split, but the N-glycosylic linkage of the last N-acetyl-D-glucosamine unit remained intact. Unfortunately, the solvolysis causes denaturation of most proteins, and so this is not a good method for determining whether the glycosylation of a protein is necessary for its biological function. Hydrogen fluoride can also be used to remove sugars from cell walls of plants.",63By immersing blocks of wood in a series of increasing concentrations of hydrogen fluoride in water (up to 80%), it was possible to remove all of the cellulose and hemicellulose and to retain a lignin skeleton which could be observed by electron micro~copy.~~ In presumably nonlignified plant-tissue, the primary cell-walls of dicots retained a cell-wall shape after hydrogen fluoridesolvolysis,'* providing evidencethat the cell-wall protein, which also remained insoluble,is crosslinkedby non-carbohydrate linkages, in addition to any possible carbohydrate cross-links. The retention of visibly intact, root-hair cell-walls after hydrogen fluoride solvolysisand trypsin treatment has been used as evidence of the presence of a substantial amount of ligninlike material in root-hair cell-walkM Hydrogen fluoride has occasionally been used to aid in the determination of structures of nucleic acids. In proving the structure of 3',5'-cyclic adenosine monophosphate, anhydrous hydrogen fluoride was used65to obtain 2-O-methyl-~-ribosefrom the methylated, cyclic AMP.65The phosphoric ester linkages and the N-glycosylic linkage to the adenine were rapidly cleaved at room temperature. At lower temperatures for brief reaction times, it appears that DNA is stable. After a 5-minute treatment of intact, dried yeast-cells in hydrogen fluoride at O", DNA could be easily extracted, and even usedMto transform Escherichia coli. (63) (64) (65) (66)

I. B. Sacks, I. T. Clark, and J. C. Pew, J. Polym. Sci., Part C, 26 (1963) 203-212. B. K. Hamilton and A. J. Mort, manuscript in preparation. D. Lipkin, W. H. Cook, and R. Markham, J. Am. Chem. Soc., 81 (1959) 6198. S. Oshiao, N. Katsura, K. Uitada, and N. Gunge, FEBS Lett., 220 (1987) 383-386.

202

YURIY A. KNIREL et al.

Hydrogen fluoride can be used to polymerize sugars. As was mentioned earlier, in the case of neutral sugars, if there is a high concentration ( 10Yo or more) of sugar in the hydrogen fluoride, or, in more dilute solutions, if the hydrogen fluoride is evaporated, thus increasing the concentration, monosaccharides polymerize to form water-soluble oligomers and polymers of mixed linkages. In this way, polymers containing almost all of the possible linkages between the sugars can be made. After methylation and hydrolysis, standards can be obtained for each individual sugar in almost all of the partially methylated forms normally encountered?' In the cases in which the anomeric linkage type has been determined, it was always a. For the amino sugars that have been investigated68(2-acetamido-2-deoxyD-glucose and 2-acetamido-2-deoxy-~-galactose), slow evaporation of their solution in hydrogen fluoride at 20" leads to some production of strictly p-( 1+6)-linked oligomers. Reaction of the amino sugars with methanol in hydrogen fluoride produced a high yield of the methyl /3-glycosides. The formation ofp-, rather than a-, glycosides was explainedby the involvement of the oxazolinium ion (2) in the condensationreaction. The strict specificity for addition to 0-6 was not explained. VII. CONCLUSION Although the use of anhydrous hydrogen fluoride in structural carbohydrate chemistry has been limited to a small number of research groups and has only been frequent during the past 10 years, a great deal of information has been derived from it. We trust that this article will bring the potential of hydrogen fluoride to the attention of many more structural carbohydrate chemists and biochemists. As we hope to have conveyed by the article, the great advantages of hydrogen fluoride over hydrolysis with aqueous acid, and many other degradative methods, are the high specificitythat can be obtained and the change which can be made in the specificity by varying the temperature of the reaction. By using hydrogen fluoride, polysaccharides can be degraded (in high yields) to smaller fragments which can be used to determine the identity and location of both N- and 0-acyl substituents on sugars, and to aid in the determination of the fundamental structure of the polymer. ACKNOWLEDGMENTS We thank Dr. Bengt Lindberg for helpful suggestions on the manuscript. A. J. Mort thanks the USDA and DOE for financial support. (67) A. J. Mort and M . 3 . Kuo, unpublished results. (68) J. Defaye, A. Gadelle, and C. Pedersen, Curbohydr. Rex, (1989), in press.

ADVANCES IN CARBOHYDRATE CHEMISTRY A N D BIOCHEMISTRY. VOL. 47

THE THERMAL DECOMPOSITION OF CARBOHYDRATES. PART I. THE DECOMPOSITION OF MONO-, DI-, AND OLIGO-SACCHARIDES

BY PIOTRTOMASIK Department of Chemistry and Physics, The Hugon Koltgtaj Academy of Agriculture, Cracow, Poland

MIECZYSEAW PAEASINSKI Department of Carbohydrate Technology, The Hugon Koltgtaj Academy of Agriculture, Cracow, Poland

AND

STANISEAW WIEJAK

College of Engineering, Opole, Poland I. Introduction

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

11. Caramel and Caramelization. ...........................................

1. Historical Background .............................................. 2. Caramel: Definition, Characteristics, Types, and Standardization ........... 3. Physical Properties of Caramel and Its Components ...................... 4. Sources for the Manufacture of Caramel. ............................... 5 . Preparation, Manufacture, and Storage of Caramel. ...................... 6. UsesofCaramel ................................................... 7. Detection of Caramel, and Some Aspects of the Analysis of Caramel ........ 8. The Chemical Nature of Caramel ..................................... 9. Biological Screening of Caramels. ..................................... Ill. The Pyrolysis of Sugars . .

203 204 204 205 214 218 226 230 234 237 267 270

I. INTRODUCTION Thermal decomposition of monosaccharides, disaccharides, and oligosaccharides, among them those originating from starch (starch hydrolyzates), 203

Copyright 0 1989 by Academic Press, Inc. All rights of reproduction in any form reserved.

204

PIOTR TOMASIK et al.

leads to a variety of products, depending on the conditions applied. Chefs and home cooks still ofien burn commercial sugar in order to obtain a product commonly called caramel. It is prepared as a flavoring, and not as a coloring ingredient, of gravies, sauces, and cakes. Such caramels differ in methods of preparation, as well as in properties, from caramels manufactured for commercial and industrial use. The latter have to be prepared at temperatures not higher than 250", under rather strictly controlled conditions, usually in the presence of one or more additives. However, the thermal decomposition of carbohydrates between 250 and 1000" has also been studied, and this aspect is of theoretical and practical interest. In this Chapter, the decomposition of saccharides by heat is discussed in two Sections, devoted respectively to caramelizationof saccharides and their pyrolysis above 250". 11. CARAMEL AND CARAMELIZATION 1. Historical Background

Caramelization belongs to the group of so-called browning reactions. These commonly occur on (a) baking, cooking, heating, roasting, and toasting of food, as well as (b) prolonged storage of fruits, vegetables, and other foods. The first kind of reaction is desirable, and has been applied as long as food has been prepared. In the second case, the reaction of browning is undesirable. The browning may be both due to enzymic and such non-enzymic processes as, for instance, the oxidation of L-ascorbicacid (see an article by Greenshields and Macgillivray and references cited therein). In any event, browning reactions had been observed and utilized for millenia, and priority for their discovery cannot be established. Schiweck2located possibly the first written note about caramel, in a period preceding the birth of Christ, by the Roman philosopher Seneca in 65 B.C. Probably the first scientificpublication on caramel was written by Peligot3in 1838, although it was preceded in 1785 by information dealing with another aspect of browning reactions, namely, humic acids resulting from treatment of sugars with mineral acids4As a matter of fact, this is an alternative route to caramel. In 1858, in his fundamental work, G e W 6 described some components (caramelan, caramelen, and caramelin)of caramel from saccharides,and the

'

(1) R. H. Greenshields and A. W. Macgillivray, Process Eiochem., 7 (1972) 1 1 - 16.

(2) (3) (4) (5) (6)

H. Schiweck, Drsch. Lebensm. Rundsch., 76 (1980) 274-280. E. Peligot, Ann. Chim.Phys., 67 (1838) 113- 176. F. K. Achard, Crell's Chem. Ann., 2 (1786) 391 -412. A. Gelis, Ann. Chim.Phys., [3] 52 (1858) 352-404. A. Gelis, Ann. Chim.Phys., [3] 65 (1862) 496-498.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I

205

polymeric (colloidal) character of caramel was then soon realized by Graham.7.sSeveral papers were published at that time to support not only the constitutionalformulasof the components characterized by Gelis but also to elucidate their structures. Studiesof these problems have continued until the present, because neither the structure of caramelan, caramelen,and caramelin nor the mechanism ofthe caramelization process is yet known in detail. A bibliography of the subject until 1961 was collected in a review by Truhaut and coworker^.^ The first studies on the thermolysis of carbohydrates, among them sugars, at temperatures above 250" had already been conducted1°in 1912, and they have been continued until the present. Thermolysis in the presence of catalysts in an atmosphere of hydrogen produces polyols,11-14 but this subject is beyond the scope of this Chapter. 2. Caramel: Definition, Characteristics, Types, and Standardization Caramel has a brown color, and it originates from various sugars when they are heated, either dry,lSYI6or in concentrated solutions. Sugars can be either heated alone or with certain additives. Products having various properties are available, depending on the temperature and duration of the heating. Those resulting from prolonged heating are readily water-soluble; a bitter taste is also characteristic. Such descriptions of caramel may be found el~ewhere.l~-*~ Caramel is, among others, designed as a food additive or ingredient; therefore, its origin and properties are laid down by the Food Laws of various (7) T. Graham, J. Chem. Soc., 15 (1862) 258-268. (8) T. Graham, Ann. Chim. Phys., [3] 65 (1862) 127-209. (9) R. Truhaut, G. Vitte, and V. Lassalle-Saint-Jean, Bull. SOC.Pharm. Bordeaux. 101 (1962) 97- 120. (10) V. 1. Ipatieff, Ber. Bunsenges. Phys. Chem., 5 (1912) 3218-3226. ( I 1) M. Abdel Akher, J. Ghali, M. S. Raouf, and M. Roushdi, Staerke, 26 (1974) 307-312. (12) G. Natta, R. Rigamonti, and E. Beati, Ber., 76 (1943) 641 -656. (13) W. H. Zartman and H. Adkins, J. A m . Chem. Soc., 55 (1933) 4559-4563. (14) C. Boelhouwer, D. Korf, and H. J. Waterman, J. Appl. Chem., 10 (1960) 292-296. (15) W. C. Moore and J. B. Davies, Methods Chem. Eng., 18 (1918) 301-302. (16) L. Poncini, Sucr. Belge, lOO(1981) 221-229. (17) W. Zipfel, Lebensrnittelrecht,Vol. 11, C. H. Beck'sche Verlagsbuchhandlung,Miinchen, c 120,g 11. (18) J. Noonan, in T. E. Furia (Ed.), CRC Handbook of Food Additives, The Chemical Rubber Co., Cleveland, Ohio, 1968, p. 38. (1 9) G. Grife, in J. Schormiiller(Ed.), Handbuch der Lebensrnittelchemie,Vol. V/1, Springer Verlag, Berlin, 1938, p. 684. (20) J. Grossfeld, in A. Juckenack, E. Barnes, E. Blayer, and J. Grossfeld (Eds.),Handbuch der Lebensrnittelchemie,Vol. V, Springer Verlag, Berlin, 1938, p. 45 I.

206

PIOTR TOMASIK et al.

countries or economic unions. For example, one of the Food Laws21lists permissible additives for the manufacture of caramel. They are acetic acid, phosphoric acid, sulfuric acid, sulfurous acid, sulfur dioxide, and ammonium, sodium, and potassium hydroxides, as well as phosphates, sulfates, and sulfites of ammonium, sodium, and potassium. The WHO/FAO Joint Expert Committee on Food additive^^^,^^ distinguishes three general kinds of caramel: (2)caramel color plain, (2) caramel color, ammonia process, and (3) caramel color, ammonium sulfite process. Both the European Technical Caramel Association (EUTECA) and the International Technical Caramel Association (ITCAY4 have standardized the properties of four classes and ten types of caramels; they are given in Table I. The content of heavy metals cannot exceed values reported in the footnote to that Table. In a report, Hellwig and coworkers26attempted to standardize particular types of caramel by means of the following tests: (a) with citric acid, (b) solubilityin 65% (v/v) ethanol, and (c) the Lassaigne test (see Table 11). They also presented gel-permeation (g. p.) chromatograms (size exclusion chromatography) of various types of caramel. Values of RBDfor zones on the chromatogramsevidently characteristic for the caramels studied are attributable to color components of caramel. These studies were directed to caramels standardized by ITCA. They allow division of these caramels into ten types and into further subtypes (see Table 111). The results published are not necessarily specific for a given type of caramel. They may be barely credible for plain caramels. For other caramels, combined analytical methods (chromatographyof dilute solutions of caramel followed by thin-layer electrophoresis,and size-exclusionchromatography) should be applied.*’ Any proof of identity of caramels (or products of the Maillard reaction in a food) is impossible without application of such complex procedures. For instance, thin-layer chromatography alone fails to distinguish between particular types of ~ a r a m e l . ~ ~ , ’ ~ (2 1) Richtlinie des Rates fur farbende Stofle, die in Lebensmitteln venvandet werden dufen vom 25 Oktober 1965, 65/469/EWG. (22) FA0 Nutr. Meet. Rep. Ser., No. 57. (23) WHO Tech. Rep. Ser.. No. 617, Geneva, 1977. (24) ITCA/EUTECA Specificationsfor Caramel Colour,Meet. EUTECA Sub-Comm. Specifications, 3rd, 4th May 1979, Paris. (25) Caramel Color, Pol. Stand.. BN-73/8083-01. (26) E. Hellwig, E. Gombocz, S. Frischenschlager,and F. Petuely, Dtsch. Lebensm. Rundsch., 77(1981) 165-174. (27) K. Tiefenbacher,R. van Eckert, W. Pfannhauser, and H. Woidich, Ernahrung, 7 (1983) 407 -4 10. (28) C. Maglitto, C. Gianotti, C. Gandini, and G. Secchi, Boll. Lab. Chim. Prov., 18 (1967) 35-43. (29) V. Pantano and R. Rossi, Boll. Lab. Chim. Prov., 20 (1969) 193- 198.

TABLE I Classes of Caramel Color' According to ITCA/EUTECAU

I

II

111

IV

Caramel color, olain Caramel color, caustic sulfite Caramel color, ammonia Caramel color, ammonia sulfite

CP-I CP-2

5-35 40-80

2-12} 15-25

55-75

<25

<0.1

<0.1

<0.01

<0.005 (
CCSl

40-80

15-25

62-82

<25

<0.1

0.15-2.5

<0.15

AC-I AC-2 AC-3 SAC-I SAC-2 SAC-3 SAC-4

60-90 100-140 150-200 35-70 75-100 I05 - I50 210-270

16-24 27-37.5 40-54 8-16) 17.5-23) 22.5 - 37 40-52

(<0.7)

<0.5

<0.015 (<0.02)

<200

1 55-75 47-57J

2oo

0.1-1.3 0.5-2.8 0.8-2.8 2.0-4.0

I

0.5-3.0 0.5-5.0 <0.3 1.5-6.5 0.3-2.01 0.8-3.2 }(<0.7) 1.0-4.0 1.O- 5.0)

0.3 I }(<0.5) 1.5)

0.08 I 0.15 }(
* The values in parentheses are FAO/WHO proposalsdating from 1980. Permissiblecontent (in m8.kgp) of heavy metalsin caramelaccordingto FAO/WHO Standard&21 are:total, 25; Cu, 20 (5); Pb, 5 (2);As, 3 (I); Hg,0. I ; Zn (20). and Sn (30).Figum in parentheses are according to the Polish Standards."

PIOTR TOMASIK et al.

208

TABLEI1 Tests for Particular Types of CaramelU Type of caramel'

Spirit caramel IA IB IC Non-ammonia caramel I1 Beer caramel 111 A 111 B I11 c Softdrink caramel IV A IV B

Citric acidb

Test alcoholu

Lassaigne

-+-

+ +

+ +

+ +

+ + + +

+

a Consult also, Tables I and I11 for classification ofthe types. Formation of a precipitate. Solubility in citric acid- monosodium phosphate at pH 2.5. Solubility in 13 :7 (v/v) ethanol-water.

The standardization of ammonia caramels presents a special task. In the course of the manufacture of caramels in the presence of ammonia and ammonium compounds, 4(5)-methylimidazole is formed, and it is n e u r o t o ~ i c . Its ~ - ~ ~o n t e n t in ~ caramel ~ - ~ ~ is limited by the Food Laws of several countries (for instance, in Austria3' and the United Kingdom38)and by the WHO regulation^^^ (see also articles by and Wooda). (The specification and analysis of ammonia caramels manufactured in the United (30) I. F. Gaunt, A. G. Lloyd, P. Grasso, S. D. Gagnoli, and K. R. Butterworth, Food Cosmet. Toxicol., 15 (1977) 509-521. (3 I ) I. G. Evans, K. R. Butterworth, I. F. Gaunt, and P. Grasso, Food Cosmet. Toxicol.. 15 (1977) 523-531. (32) G . K. Buckee and T. P. Barley, J. Inst. Brew., 84 (1958) 158- 164. (33) N. Cerny and A. Blumenthal, Z. Lebensm. Unters. Forsch., 168 (1979) 87-90. (34) J. Camevale, Food Technol.Aust., 31 (1979) 165-169. (35) R. A. Wilks, A. J. Shingler, and L. S. Thurman, J. Chromatogr., 87 (1973) 41 1-418. (36) G. Fuchs and S. Sundall, J. Agric. Food Chem., 23 (1975) 120- 122. (37) Oesterr. Lebensmittelfarbstojiverordnung.BGBI, 279J 1979. (38) M A E , FACC Report 29, Interim Report on the Review ofthe Colouring Matter in Food Regulations, 1973. (39) E. Thier, Zuckerindustrie, 105 (1980) 80-82. (40) F. Wood, BNFNutr. Bull., 7 (1983) 27-30.

TABLE 111 Analytical Characteristicsof All Classes, Sorts, Types, and Subgroups of Caramel Colors‘ Properties

Type, Carbohydrate, and Catalyst

cP2

PH Neutralization Nitric acid test Alcohol test Color intensity E& Dry substance (%) Ash (%) Sodium (mdkg) Sulfur dioxide (mdkg) Sulfates in ash (%) Total nitrogen (%) Basic nitrogen (mdkg) Formic acid (mg/kg) Glucose (%) Fructose (%) sucrose (%) 4-Methylimidazole (mgJkg) “Glucoreductone” (96) Gel-permeation analysis26 Column between R,, Maxima at R , , Minima at R,,

3.7

+

-

65 65.2 0.55 1825 n.f.* n.f. 0.08 n.f. 455 3.7 3.1 0.6 n.f. 0.15

I A. Sucrose Sodium carbonate CP 1

3.1

+32 69.6 1.16 2398 n.f. n.f. n.f. n.f. 577 14.5 11.2 n.f. n.f. 0.23

CP 2

3.1

+71 64.1 0.47 1856 n.f. n.f. n.f. n.f. 664 9.3 7.9 n.f. n.f. 0.25

0.5-1.75 0.98 1.32

0.39- 1.66 0.9 I 1.16

0.36- 1.65 0.91 1.16

1.18

I .02

1.02

CPl

I B. Sucrose Plus acid CP 1

3.9

3.6 -

29 74.4 0.09 126 n.f. n.f. n.f. n.f. 249 24.8 15.6 n.f. n.f. 0.05 0.06- 1. I3 0.35 0.8 1 0.96 0.74,0.87

-

38 70.3 0.38 1422 n.f. n.f. n.f. n.f. 384 7.0 2.0 n.f. n.f. 0.05 -0.02- 10.7 0.32 0.6 1 0.86 0.56,0.77

cP1

3.5 -

29 72.8 0.06 151 n.f. n.f. n.f. n.f. 343 16.6 3.8 0.8 n.f. 0.06

-0.020.32 0.84 0.73

1.02

TABLEIll (Continued) ~~

~~

Properties

Type, Carbohydrate, and Catalyst

CP 2

PH

c!

0

Neutralization Citric acid test Alcohol test Color intensity, E& Dry substance (96) Ash (%) Sodium (mg/kg) Sulfur dioxide (mg/kg) Sulfates in ash (%) Total nitrogen (%) Basic nitrogen (mg/kg) Formic acid (mg/kg) Glucose (%) Fructose (%) Sucrose (%) 4-Methylimidazole (mg/kg) “Glucoreductone” (96) Gel-permeation analy~is’~ Column between Rap, Maxima at hP,

Minima at Rap

6.3

+ +-

63 65.1 2.75 10146 n.f. n.f. 0.6 n.f. 614 2.8 1.8 n.f. n.f. 0.18 -0.05- 1.27 0 0.32 0.9 1 0.05 0.88

I C. Sucrose Catalyst unknown CP 2 6.1

+ +-

67 65.5 2.1s 13931 n.f. n.f. n.f. n.f. 595 2.8 2.8 0.9 n.f. 0.19 -0.05- 1.30 0 0.40 0.89 1.02 0.1 1 0.93

CP 2

3.8

11. Sucrose Sodium sulfite CCS

3.0

+-

+-

57 65.9 0.06

100 95.0 0.07 57 66 n.f. n.f.

-

n.f. n.f. n.f. n.f. 47 1 11.2 4.8

2.9 n.f. 0.1 1

-0.05-1.32 0 0.30 I .02 0.02 0.86

n.f. 200 2.3 0.4 n.f. n.f. 0.18 -0.05- 1.39 0 0.55 1.09 0.07 0.9 I

Properties

Type, Carbohydrate, and Catalyst

AC 2

PH Neutralization Citric acid test Alcohol test Color intensity, Q~ Dry substance (%) Ash (%) Sodium (mg/kg) Sulfur dioxide (mg/kg) Sulfates in ash (%) Total nitrogen (%) Basic nitrogen (mg/kg) Formic acid (mg/kg) Glucose (%) Fructose (%) Sucrose (%) 4-Methylimidazole (mg/kg) “Glucoreductone” (%) Gel-permeation analysis35 Column between R,, Maxima at R,, Minima at R,,

4.7

I11 A. Sucrose Ammonium carbonate AC 3 5.6

-

++

108 63.4 0.30 1198 n.f. n.f. 4.73 240 673 23.0 10.9 I .7 I28 0.54

194 93.6 0.58 1688 n.f. n.f. 6.7 1690 49 1 3.5 n.f. n.f. I19 0.53

-0.99- 1.66 0 0.93 1.23 0.07 1.05

-0.07- 1.75 0 0.89 1.18 0.25 1.05

I11 B. Sucrose Gaseous ammonia

AC 2

AC 1

AC 1

5.7

5.3

5.1

4.5

86 62.6 0.45 1355 n.f. n.f. 4.95 180 415 23.7 8.9 5.4 151 0.62

99 64.1 0.38 953 n.f. n.f. 4.45 220 67 1 16.1 5. I 3.1 47 0.54

++

113 70.5 0.27 n.f. n.f. 2.5 n.f. 296 10.0 8.0

n.f. 51

0.44 -0.09- 1.69 0 0.93 1.23 0.09 1.05

+-

79 60.9 0.36 I265 n.f. n.f. 4.70 250 542 25.8 10.7 n.f. 118 0.54

+-

AC 2

0.41 - 1.80 0.84 1.30

0.39- 1.73 0.9 1 1.34

0.54- 1.61 0.97 1.27

1.02

1.04

1.11

(continued)

TABLE 111 (Continued) Type, Carbohydrate, and Catalyst

Properties

AC 3

PH

N

Neutralization Citric acid test Alcohol test Color intensity, emM Dry substance (%) Ash (%) Sodium (mdkg) Sulfur dioxide (mdkg) Sulfatesin ash (%) Total nitrogen (%) Basic nitrogen (mdkg) Formic acid (mg/kg) Glucose (96) Fructose (%) Sucrose (%) CMethylimidazole(mdkg) “Glucoreductone” (%) Gel-permeation analysis35 Column between R , , Maxima at R , , Minima at RgP

6.3

++

146 71.5 3.15 10169 n.f. n.f. 5.46 20 748 9.5 4.6 n.f. 344 0.66 0.30- 1.92 0.75 0.96 1.25 0.40 0.91 1.15 1.35

I11 C. Starch Ammonia AC 2

4.6

AC 2

SAC 2

3.8

4.1

IV A. Sucrose Ammonium sulfite SAC 3

5.8

-

-

-

+

++

119 71.3 0.62 1024 n.f. n.f. 5.20 70 660 13.3 5.0 n.f. 308 0.58

I37 70.7 0.46 1214 0.0 1 n.f. 3.10 14 438 13.1 6.5 n.f. 123 0.40

78 65.0 I .80 5230 369 49.6 I .35 260 277 28.8 3.8 12.0 1 I7 0.07

123 68.3 2.18 5307 337 67.1 1.17 150 438 31.4 n.f. n.f. 40 0.09

+

0.59-2.07 1.05 I .39 1.18

+

0.30- 1.59 0.72 0.79 1.30 1.25

-0.07- 1.39 0 0.30 0.98 0.09 0.93

-0.04- 1.34 0 0.27 0.93 0.07 0.9 1

SAC 2 3.9 -

+ 98 68.4 1.96 5540 234 54.5 1.17 280 393 28.5 n.f. n.f. 37 0.09

-0.55- 1.39 0 0.27 0.93 0.9 1 0.9 1

Properties

Type, Carbohydrate, and Catalyst

SAC 2

PH

N

W

Neutralization Citric acid test Alcohol test Color intensity, emM Dry substance (%) Ash (%) Sodium (mdkg) Sulfur dioxide (mukg) Sulfates in ash (%) Total nitrogen (%) Basic nitrogen (mdkg) Formic acid (mdkg) Glucose (96) Fructose (96) Sucrose (%) 4-Methylimidazole(mdkg) “Glucoreductone” (%) Gel-permeation analysis35 Column between &p. Maxima at R , ,

IV B. Starch Ammonium suffite SAC 2

SAC 2

3.2

3.3

-

-

-

92 65.6 4.94 16183 427 58.1

I14 64. I 7.0 21731 95 1 64.2 1.26

114 76.2 2.15 609 I 538 63.5 2.08 I140 236 19.6 n.f. n.f. 94 0.08

+

0.74

350 480 29.7 n.f. n.f. 66 0.10

-0.09 -0.93 0.2 1

+

1060

577 27.6 n.f. n.f. 154 0.10 -0.09-0.95 0.27

3.2

+

-0.02-1.16 0.30

a The classification is according to ITCA/EUTECA. Data are arranged according to varying color intensity. Subgroupsare classified according to the results of gel-permeation (g.p.) analysis. The results, except those from gel-permeation analysis, are calculated on dry matter. Not found. R,* is the ratio ofthedistance which passedthe

spot ofa substance from the top ofthe column in agiven period oftime to agiven debasement ofthe mobile phase in

a gel-penneation (g.p.) column.

214

PIOTR TOMASIK et al.

Kingdom4’ and in ItalqP2 revealed that their properties satisfy regulations proposed by ITCA.) Proposals have also been made to reserve the name “caramel” for the products manufactured from saccharides in the absence of nitrogen-containing compounds. They would be considered to be flavoring ingredients. The products available from saccharides and ammonia would be called “sugar colors,” and they serve as coloring additives. Until the present time, coloring ability is the major factor characterizing all caramels, whereas, in ammonia caramels, the content of 4( 5)-methylimidazole becomes a most important factor (that is absent from nonammonia products). The content of the latter should serve for differentiation between the types of products that, until now, have all been called caramels.43

3. Physical Properties of Caramel and Its Components Ever since the investigationsof Graham,7.8it has been known that caramel is polymeric in its character. Von Elbe“ described caramel as a colloidal dispersion of a lyophobic, humic substance, in a mixture of two other substances protecting this colloid. Depending on the isoelectric point of these colloids, caramel may be roughly divided into three groups: positive, negative, and spirit caramels. Their isoelectric points lie between 5.0 and 7.0 for positive caramels, 4.0 and 6.0 for negative caramels, and below 3.0 for spirit caramels. The isoelectric points of caramels are best determinable by using an electrophoretic te~hnique,~~-~O the flocculation test with tannin,51ionic surface-active agents used in industry as a rough g ~ i d e ,the ~ ~gelatin -~~ and other method^.^^.^^ Electropositive caramels are manufactured under (41) A. L. Patey, G. Shearer, M. E. Knowles, and W. H. B. Denner, FoodAddit. Contam., 2 (1985) 107-112. (42) C. Sapetti, G. Mazza, and S. Valvasson, Riv. SOC.Ital. Sci. Aliment., 12 (1983) 287 -290. (43) G. Lehmann, Dtsch. Lebensm. Rundsch., 81 (1985) 388-389. (44) G. von Elbe, J. Am. Chem. Soc., 58 (1936) 600-601. (45) M. S. Badollet and H. S. Paine, Int. Sugar J.. 28 (1916) 23-28,97- 103, 137- 140. (46) S. M. Hauge and J. J. Willaman, Ind.Eng. Chem., 19 (1927) 943-955. (47) H. Roderer, Staerke, 7 (1955) 205 -208. (48) R. N. Greenshields, P. C. Hunt, R. Fassey, and A. W. Macgillivray, J. Inst. Brew., 75 (1969) 542-550. (49) R. N. Greenshields, Glue Gelatin Res. Assoc., RPP 77 (1970). (50) R. Truhaut, R. Costagnon, S. Lareeban, and V. Lassalle-Saint-Jean, Bull. SOC.Pharm. Bordeaux, 100(1961)261-268. (5 I ) R. Truhaut, R. Costagnon, V. Lassalle-Saint-Jean, and M. Biccsy, Bull. SOC.Pharm. Bordeaux, 100(1961)251-260. (52) J. White, Yeast Technology, Chapman and Hall, London, 1954, p. 189. (53) M. S. Badollet and H. S. Paine, Ind. Eng. Chem., 19 (1927) 1245- 1246. (54) R. E. Lothrop and H. S. Paine, Ind. Eng. Chem., 23 (1931) 328-332. (55) R. N. Greenshields, Process Biochem., 8 (1973) 17-26. (56) J. White and D. J. Munns, J. Inst. Brew., 53 (1947) 305-312.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I

215

acid conditions in the presence of inorganic catalysts, whereas electronegative products are prepared under alkaline conditions, using ammonia or, randomly, amino acids and proteins. Hence, the average molecular weight (5,000to 10,000)of electronegativecaramel is higher than that of electropositive caramel. Three main components of caramel (called caramelan, caramelen, and caramelin)were isolated by G e l i ~who , ~ used dialysis for this purpose. Other investigators refined the dialysis p r o ~ e d u r e , and ~ ~ -a~ group ~ of Russian workers60 separated these components by electrofiltration through an ionexchange membrane at an electric field strength of 20 V/cm. Other authors proposed a rough separation by employing different solubilities in various ' - ~ ~cold and hot water, but also solvents, mainly 84% aq. e t h a n ~ l , ' , ~ and 1-propanol,'"' as well as solutions of various salts.@Such techniques as fractionation by gel f i l t r a t i ~ n , and ~ ~ -adsorption ~~ on i~nexchangers~~-'* are (57) K. Smolenski and M. Werkenthin, Guz. Cukr., 62 (1928) 529-538. (58) S. K. D. Aganval and K. V. Gupta, Annu. Conv. Sugar Technol. Assoc. India, 42 (1978) G 139-GI 52. (59) A. Joszt and S. Molinski, Kolloid Beih., 42 (1935) 367-383. (60) L. D. Bobrovnik, V. N. Rudenko, V. V. Mank, and M. P. Kupchik, Izv. Vyssh. Uchebn. Zuved., Sukh. Prom., (1985) 29-31. (61) A. R. Sapronov, I n . Vyssh. Uchebn. Zuved., Pishch. Tekhnol., (1963) 33-37. (62) B. Ripp, 2. Ver. Dtsch. Zuckerind., 76 (1926) 627-635. (63) J. E. Cleland and J. B. Longenecker, U. S. Pat. 2,637,655 (1953); Chem. Abstr.. 47 (1953) 9522. (64) M. Peronnet, Ann. Chim. Anal., 26 (1944) 46-48. (65) F. Orsi, Nuhrung, 13 (1969) 53-57. (66) J. P. J. Casier, A. A. Zenouz, A. Ahmadi, and G. M. J. De Page, Flavor Foods Beverages Chem. Technol., Proc. Conf. (1978) 169-179; Chem. Abstr.. 91 (1979) 106,847~;G. Charalambous and G. E. Inglett (Eds.) Flavor of Foods and Beverages, Academic Press, New York, 1978, p. 169. (67) J. Ujszarzi, Szeszipur, 28 (1980) 45 1-456. (68) R. V. Wenn, J. Inst. Brew., 78 (1972) 404-406. (69) H. Rother, 2. Dtsch. Lebensm. Rundsch., 62 (1966) 108- 113. (70) J. Kumanotani, R. Oshima, A. Yamauchi, N. Takai, andY. Korosu, J. Chromutogr., 176 (1967) 462-464. (71) C. Tarantola, Mni Ifal.,10 (1968) 39-45. (72) J. Pagenkopf, Weinwissenschaften, 35 (1980) 68-76. (73) I. F. Bugaenko, 1. F. Bulgakova, and I. I. Pavlov, Izv. Vyssh. Uchebn. Zuved., Sukh. Prom., 42 (1968) 41 -45. (74) V. F. Selemenev,G. A. Chikin,andI. P. Shamritskaya,Izv. Vyssh. Uchebn.Zuved., Sukh. Prom., 51 (1977) 23-27. (75) A. R. Sapronov and R. A. Koltseva, Izv. Vyssh. Uchebn. Zuved., Pishch. Prom., (1975) 105- 110, 154- 157. (76) V. P. Meleshko, S. Z. Ivanov, and G. A. Chikin, CINTI Pishcheprom., (1969) 6-31. (77) V. F. Selemenev, G. A. Chikin, and N. V. Korchkova, Izv. Vyssh. Uchebn. Zuved., Sukh. Prom., 53 (1979) 22-25. (78) A. R. Sapronov, S. A. Chikin, V. P. Meleshko, andT. A. Klochkova, Izv. Vyssh. Uchebn. Zuved., Sukh. Prom., 36 (1962) 15- 17.

216

PIOTR TOMASIK et al.

useful mainly for elimination, and separation, of coloring matter from caramel. Moreover, use of Sephadex gels may indicate the molecular weight of components isolated. The use of anionic resins is compatible with the chemical stability of components being se~arated.’~ Some resins allow adsorption of caramelan, whereas caramelen and caramelin are not adsorbed. Charcoal may also be applied for this purpose.79 Caramelan was described as a brown, brittle, deliquescent solid having a bitter taste. It preponderates among the products of caramelization if this process is conducted until a 12%loss of weight is a~hieved.~ At this stage, water is the sole, volatile product.s0Caramelan melts at 138 (Ref. 8 l), 136” (Ref. 82), 144”(Ref. 83), and is readily water-soluble. Caramelen is a brown substance, much darker than caramelan, and not deliquescent. It meltss*at 1 53.5- 154”. Caramelin exists in three modifications, namely, soluble in cold water, soluble only in boiling water, and insoluble in all ordinary solvents. They are infusible materials that are much darker than caramelan and ~aramelen.~ Three high-molecular-weight components separated by Joszt and MolihskP9from caramel prepared from sucrose in the Pictet - Adrianoff vacuum processs3have the properties given in Table IV. They, and melanoidin (found in the ammonia-processedcaramels”), seem to be the sole polymeric compounds in caramel. Using a microdiffusion method, Sapronov6’ estimated the sizes of particles of caramelan (0.46 pm), caramelen (0.95pm), and caramelin (4.33 pm). The pH values of caramels are given in Table I; they constitute an important property of caramels. A high pH may indicatean incomplete “burn,” or alkali present. In either case, the tinctorial strength [defined as the absorbance at 560 nm of a 0.190 (wt/vol) solution of caramel in a quartz cell ( 10 X 10 mm)] of the product increases as the product ages. Above pH 6.0, caramel is susceptibleto attack by molds, and, below pH 2.5, it quite readily resinifies. For caramel solutions, the acid stability is commonly determined; in practice, this is a test for cloudiness of solutionsunder the influence of an acidic medium.s5 O

(79) A. R. Sapronov, S. E. Khann, and T. A. Parshina, Izv. Vyssh. Uchebn. Zaved., Sakh. Prom., 43 (1969) 24-28. (80) A. Joszt and S. Molidski, Biochem. Z., 282 (1936) 270-276. (81) E. P. Miroshnikova, G . A. Chikin, and V. M. Tobolina, Izv. Vyssh. Uchebn. Zaved., Sakh. Prom., 44 (1970) 10- 15. (82) M. Cunningham and C. Doree, Z . Ver. Dtsch. Zuckerind.. 68 (19 18) 1 - 2 1. (83) A. Pictet and N. Andriano6 Helv. Chim. Acta. 7 (1924) 703-707. (84) S. Mass, in W. Foerst (Ed.), Ullmans Encyklopedie der Technischen Chemie, Vol. 16, Urban and Schwarzenberg, Miinchen, 1965, p. 349. (85) F. W. Peck, Food Eng., 27 (1955) 99, 154-155.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I

217

TABLEIV Properties of Three High-Molecular-Weight Components of Caramel from the Pictet- Andrianoff Vacuum Process*

Water content (%) Ash (%) Organic matter (96)

c (%) H (%) Reducing sugars (%) directly after 10 min after 3 h Surface tension of I % aq. solutions = 1) Relative Viscosity of 1% aq. solution (qHfHI0 Color (Stammer units)

PH 1 pH 1 1

7.86 0.43 91.71 49.8 - 50.4 5.5-5.2 25.1 27.4 46.9 99.2

8.70 0.14 91.16 53.4-54.3 4.8-5.0 20.4 22.2 41.0 98.5

1.03

1.11

16.900 39.600

29.800 49.400

8.31 0.38 91.31 50.2-50.9 5.2-5.5 17.2 11.2 39.1 99.1 1.13 21.400 49.400

Solubility in water is also a necessary property of caramels. There seems to be a relationship between the solubility and the viscosity. Usually, the less viscous caramels dissolve the more readily. Simultaneously, they have greater color stability, shelf life, and retention of complete solubility. Such caramels are handled with a minimum of waste and effort. Lessening of the viscosity of caramels is the task of producers. Thus, some caramels have the same content of solids, and specific gravity (usually 1.315 to 1.345), as others, but flow more freely and are more stable. X-Ray diffraction investigations of aqueous solutions of caramelized sugar pointed to a marked decrease in the maximum intensity of diffraction in the range of 4 to 20 mPa-s(0.04 to 0.2 poise), which means that a change takes place in the structure of the solutions.86Solubility is evidently dependent on the ratio of caramelan to caramelen in the product. Caramel from sucrose thermolyzed for 8 h at 170- 190"is readily soluble in 84% aq. ethanol; it contains mainly caramelan. After 1 1 h of thermolysis, the solubility in such an ethanol solution decreases, accompanied by higher solubility in water at room temperature. The content of caramelen in this product is higher than that of caramelan.61Caramelin dissolves only in hot water. (86) A. Gorrelli and E. Taraglione, Nuovo Cimento, 9 (1952) 440-441.

218

PIOTR TOMASIK et a/.

The degree of caramelization influencesthe hygroscopicityof the product. If it contains some unchanged D - ~ ~ U Cmoisture O S ~ , uptake is noted. Maltose, and particularly “dextrins,” are less hygros~opic.~’-~~ Tinctorial ~ t r e n g t h ~is~one . ~of - ~the ~ most important properties of caramels. It is complemented by the hue index, which is the logarithm ofthe ratio of absorbancesmeasured at 5 10 and 6 10 nm, respectively, multiplied by 10. The resulting value indicates the “redness” of a caramel. The flavor and aroma of caramel are also important factors. They consist of two components, namely, a taste arising from the acidity, and a taste attributable to the nature of the caramel. The taste due to acidity depends on the type of caramel, and may be modified, but that due to the nature is unalterable. Caramels processed in the open are lighter, in both taste and aroma, than those from pressure-kettle processes.85It must be stated, however, that the estimation of these properties generally still undergoes subjective evaluation, like many other organoleptic properties. 4. Sources for the Manufacture of Caramel The quality of a caramel depends on its source as well as on parameters of the process of its manufacture (see, for instance, an article by Dufresnoyg5). On the other hand, studies camed out by Ramaiah and the A g a r w a l ~ ~ ~ suggested that the qualitative composition of caramel is independent of the sugar used, but is influencedby the method of preparation. This concept was supported by Cerniani.9’ D-G~UCOS~, sucrose, and lactose lose water (see Fig. 1) and their reducing properties (see Fig. 2) in somewhat different ways, but the results of these changes are finally about the same. Simultaneously, von Elbe@ had realized that the composition of caramel is independent of whether thermolysis is conducted in the air or in vucuo.These observations

(87) (88) (89) (90) (91) (92) (93) (94) (95) (96)

R. lilies, Staerke, (1953) 119- 124. Kh. U. Usmanov and V. A. Kargin, Zh. Fiz. Khim., 28 (1954) 224-228. V. N. Nikiforova and L. G. Kuznetsova, Khlebopek. Konditer. Prom., (1973) 15- 16. P. C. T. Clark, Int. SugarJ., 53 (1951) 162-164; J. Inst. Brew., 47 (1950) 254-255. W. P. Fetzer, Ind. Eng. Chem., Anal. Ed., 10 (1938) 349-353. R. P. Ward, Food Proc. Ind., (1976) 33-34. T. Koydl, O e . 4 . Z. Zuckerind. Lebensm., 47 (1918) 16-39. A. Joszt and S. Molidski, 2. Untersuch.Lebensm., 71 (1936) 19-32. X. Dufresnoy, Bios (Nancy), 9 (1978) 19-25. N. A. Ramaiah, S. K. D. Agarwal, and J. K. P. Agarwal, Curr. Sci. (India). 26 (1957)

81 -82. (97) A. Cerniani, Ann. Chim. (Rome), 41 (1951)455-464.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I

219

35

IS

301

0 100

150

200

Temperature ( O C )

Water Loss from D-GIuco~~ (Points), Lactose (Dashes), and Sucrose (Solid Line) Heated to 200" in the Air.97 FIG.].-The

may be suspect in view of results obtained by the present author^,^^.^ who found that caramels from thermolysisunder oxygen-freenitrogen and under carbon dioxide differ from one another, and from those prepared in the air. Moreover, caramels from sucrose, D-fructose, and some other saccharides differ from caramels prepared from maltose. Other evidence for differences between particular caramels was educed by Agarwal and coworkers,'m.lol who prepared caramels by heating pure sucrose, and separately, by heating D-glucose with sodium hydroxide. They separated compounds chromatographically, and recorded the infrared absorption spectra for fractions obtained. The differences between spectra are small, but significant. Sucrose and its hydrolysis products (D-glucoseand D-fructose) have been prime sources for the manufacture of c a ~ a m e 1 ~ ~despite . ~ ~ the . ~fact ~ ,that ~~.~~~ reducing sugars are caramelized more readily than nonreducing sugars. The chromatographicallycontrolled caramelization of sucrose has revealed that,

(98) M. Palasinski, P. Tomasik, and S. Wiejak, Pol. Pat. P-244 150 (1983). (99) M. Palasidski, P. Tomasik, and S. Wiejak, Staerke, 37 (1985) 308-313. (100) S. K. D. Agarwal, P. C. Johary, and D. S. Misra, Z . Zuckerind., 24 (1974) 532-535. (101) S. K. D. Agarwal and V. Gupta, Proc. Annu. Conv. Sugar Technol.Assoc. India. 47th. 1983, GI-GI6. (102) A. Gelis, C. R. Acad. Sci., 45 (1857) 590-594.

PIOTR TOMASIK ef al.

b

100

150 Tempemture

200

("c )

FIG.2.-The Variationof Reducibility of DGlucose (Points), Lactose(Dashes), and Sucrose (Solid Line) when Heated to 200" in the Air.97

in the range of 160- 170°, the high-molecular-weight fraction appears within 35 - 45 minutes and acquires its most intensive color'03after 1.5 h. Kinetic studies of the thermal decomposition of sucrose in weakly basic medium have shown that the process is comprised of two independent reactions, namely, the decomposition of glucose and D-fructose. Such decomposition is described by a first-order equation. The decomposition of g glucose under the conditionsapplied is slower than that of D-fructose and, therefore, the first process is the rate-limiting step. The rate of the thermal decomposition is temperature- (T) and pH-dependent, according to the (103) E. Ya. Matynenko, P. Ya. Mishev, and I. A. Egorov, Vinodel. Vinograd SSSR, (1980) 10- 12.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I

22 I

equations1@' log K,h-

=

17.3517 - 0.469 pH -

8435.4 - 255.1 pH T

and log K d U - = 15.5630 - 0.09 pH -

8673.6 - 204.1 pH T

Caramels from burned sucrose, D - ~ ~ U Cand OS~ D-fructose, , with some distinct but variable content of noncaramelized sugars, have the best organoleptic proper tie^.^^^,'^ On the other hand, the sugars present (mainly D-frucare responsible for the relatively lower stability of such caramels. Other mono- and di-saccharideshave also been considered as sources of ~ a r a m e l . ~ ~It Jhas ~ ~been - ~ ~found * that molasses may be used, as it already containscompounds analogousto caramel. 84,112-118 Although molasses may seem to be a reasonable source for the purpose discussed, the quality markedly influences the tinctorial properties and viscosity of the resulting carame1.lI9A disadvantagein the use of molasses is its high content of potassium compounds. Caramelization may occur during the manufacture of table sugar, perhaps through overheating.The mode of the manufacture of sugar has some influence on its behavior. Sugar from carbonation more-readily becomes brown (104) V. A. Kolesnikov and G . I. Gorokhov, Izv. Vyssh. Uchebn. Zaved., Pishch. Tekhnol., (1982) 65-70. (105) G . Linden and D. Lorient, Stuerke, 35 (1983) 400-402. (106) L. Petrov, J. Mihailovic,R. Nicolov, and L. Miskovic,Znd. Secera, 37 ( I - 2 Suppl) (1983) 39-43. 3(1914)495-497. (107) G.D.BealandH.F.Zoller,J.Am.Pharm.Assoc., (108) H. Drake-Law, J. Soc. Chem. Znd., 46 (1927) 4 2 8 ~ . (109) A. Daniel, U. S. Pat. 1,316,019 (1919); Chem. Abstr., 13 (1919) 3035. ( I 10) N. V. "LYEMPF" Leeuwarder 1js-e.n Melkproductenfabriken, Dutch Pat. 57,374 (1946); Chem. Abstr., 41 (1947) 4592. ( 1 1 1) S. Y. Kim and K. S. Chang, Nongop Kisul YongenPogo (ChunguanTaekakkyo), 3 ( I 976) 105- 119; Chem. Abstr., 91 (1979) 89,739~. ( 1 12) F. W. Zerban, Sugar Res. Found. Technol. Rep. Ser., (1947) 3. ( 1 13) M. Garino, A. Rege, and F. Rubino, Gazz. Chim. Ind. Appl., 11 (1929) 6 1-63. ( 1 14) H. S. Paine, M. S. Badollet, and J. C. Keane, Znd. Eng. Chem., 16 (1924) 1252- 1258. ( 1 15) A. Brodowski, Kolloid Beih., 29 (1929) 26 I -353. (1 16) N. S. Gulyuk, Izv. Vyssh. Uchebn. Zaved., Sakh. Prom., (1968) 18-21. ( 1 17) M. L. Wolfrom, W. W. Binkley, and L. F. Martin, Sugar, 47 (1952) 33-34. ( 1 18) G. N. Kowkabany, W. W. Binkley, and M. L. Wolfrom, Agric. Food. Chem., 1 (1953) 84-87. ( 1 19) D. J. Akkennan, Java Meded., 26 (1930) 33; Chem. Abstr., 25 (1931) 4433.

222

PIOTR TOMASIK et al.

than sugar from sulfitation. Carbonates catalyze the caramelization better than sulfites.120 Because of various economic, political, and other reasons, sucrose, invert S ~ not always available for caramel production, and sugar, and D - ~ ~ U C Oare methods have been developed for making it from oligo- and poly-saccharides. Acid,64J21-'36 base,137J38 and enzymic hydroly~is,'~~,'" as well as separation of D - ~ ~ U C Ostream S~ (mother-liquor syrup) provide very stable caramels. Maize, tapioca, sago, and potato starch all seem to be useful as sources for this purpose. Moreover, starch waste may be employed. Several other nonconventional sources for the manufacture of caramel, such as malt and soybean carbohydrates, have been proposed by Jacobs. 14' Processing of such sources leads to the preparation of a syrup containing 70 - 85% of reducing sugars. The variabilityof the sourcesfor caramelization causesa great deal of empiricism in this technology. This fact prompted several scientistsS5to study relationships between some properties of the sources and the firmness

H. T. Cheng, W. F. Lin, and C. R. Wang, ACSSymp. Ser., 215 (1983) 91 - 102. A. Di Baja, Requind., 1 (1932) 52-53,74-77; Chem. Abstr., 26 (1932) 4978. P. S. King, Ind. Center (China), 3 (1934) 319-323. G. W. Galkina, Staerke, 16 (1964) 100. R. Marinelli, Ann. Chim. Appl., 36 (1946) 46-48. J. B. Longenecker, U. S. Pat. 2,582,261 (1952); Chem. Abstr., 46 (1952) 4144. T. Ohira, E. Hara, and Y. Takagi, Jpn. Pat. 77 154,564 (1977); Chem. Abstr.. 88 (1978) 135,033~. T. Ohira, E. Hara, and Y. Takagi, Br. Pat. 1,538,016 (1979); Chem. Abstr.. 91 (1980) 73,361g. T. Ohira, E. Haraand Y. Takagi, U. S. Pat. 4,138,271 (1979); Chem. Abstr., 90(1979) 185,200q. Ajinomoto Co., Inc., Fr. Pat. 2,409,703 (1979); Chem. Abstr., 91 (1980) 20,967n. A. Maklashin, Tekst. Prom., 10 (1950) 29-31. M. Ullmann, Makromol. Chem., 10 (1953) 221-234. J. N. BeMiller,in R. L. Whistler and E. F. Paschal1(Eds.), Starch Chemistryand Technology, Vol. 1, Academic Press, New York, 1965, p. 495. G. Stadnikoff and P. Korschew, Kolloid Z., 47 ( 1929) 136- 141. H. Bergstrom, Jernkontorets Ann., (195 1) 135- 176. H. Thieleand H. Kettner, Kolloid Z., 130 (1953) 131-160. A. Sroczyfiskiand M. Boruch, Staerke, 16 (1964) 112- 118. T. Kurokawa, Jpn. Pat. 179,676 (1949); Chem. Abstr., 45 (1951) 9906a. J. N. BeMiller, in Ref. 132, p. 52 1. J. A. Starostina, V. S. Gryunov, and A. S. Sipyagin, Sb. Nauchn. Rob. Moskov. Inst. Narod. Khoz., 16 (1959) 219-224. A. L. Sokolovskii, V. N. Nikiforova, and R. Ya. Greiser, Tr. Vses. Nauchn-Issled. Inst. Konditersk. Prom., 14 (1959) 32. M. B. Jacobs, Am. Perfum.. 49 (1947) 501, 503,505.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I

223

of caramels prepared from them. Thus, Sokolovskii and coworkers'" observed that a decrease in the content of glucose in the stock parallels a decrease in the hygroscopicity, but the content of maltose has only a small effect on the firmness of caramel. The same observation has been made in regard to ready caramekE7Caramels obtained from sources prepared enzymically have a greater tendency to crystallize than those made from sources prepared by acid hydrolysis.'" Starch syrups from enzymic hydrolysis contain dextrins of lower molecular weight than "dextrins" in a syrup from acid hydrolysis. Hence, such syrups are less viscous and more hygros~opic.'~~ There has also been described the processingof saccharidesin the presence of certain additives, which play a role either as reagents or catalysts. They may be acids (acetic, citric, phosphoric, sulfurous, sulfuric, and carbonic), bases (ammonia, and sodium, potassium, and calcium hydroxides), or salts (carbonates, hydrogencarbonates,sulfates, sulfites, or phosphates of ammonium, sodium, potassium, and calcium). Some organic compounds may be used, mainly biogenic amino acids, and such sulfonic acids as taurine (2-aminoethanesulfonic acid), sulfanilic acid, and other^.^,^,^^,^^,^,^^^,^^ It has been revealed that reactions of sugars with amino acids are responsible for the development of such specific aromas as those of roasted cocoa, coffee beans, baked bread, processed tobacco, baked potatoes, and protein hydrolyzates. Such reactions, and related reactions of saccharides with amines, have been intensively studied as a way of obtaining aromas, unless amino acids and amines, respectively, are applied in catalytic amounts. The optimal conditions and kinetics of the formation of nondialyzable melanoidins responsible for the brown color of caramel is now under study in various model systems. For the D-glucose- glycine model system at 1 10 the rate constants are 4.55 X 10-5/s and 9.40 X 1O-'/s for the molar sugar: amino acid ratios 2.5 : 1 and 1 :2.5, r e ~ p e c t i v e l y . ' ~Rates ~ J ~ for other model systems decrease in the order D-glucose-arginine, glucose glycine, D-fructose-glycine, lactose-glycine, and sucrose- glycine. 147 Reactions deliver, parallelly, brown melanoidins and pyrazine derivatives, but

(142) (143) (144) (145)

T. Piemgalski, Acta Aliment. Polon., 1 1 (1985) 63-70. H. Yamamura, Jpn. Pat. 31 668 (1917); Chem. Abstr.. 12 (1918) 1259. A. S. Sultanov and Kh. U. Usmanov, Zh. Prikl. Khim., 29 (1956) 1726- 1730. M. Kuncheva and Ts.Obretenov, Nauchn. Tr. Vissh. Inst. Khranit. Vkus. Prom., Plovdiv, 32 (1985) 77-87. (146) Ts. Obretenov, M. Kuncheva, and I. Panchov, J. Food Process., Preserv., 10 (1986) 25 1-268. (147) Y. N. Kim, C. M. Kim, K. W. Han, and S . K. Oh, Hanguk YongyangSiklyongHakhoechi, I 1 (1982) 51 -56; Chem. Abstr., 98 (1983) 141,981J

224

PIOTR TOMASIK et ul.

the activation energy for melanoidin formation is lower than that for the formation of pyrazines.148 The system D-fructose- L-glutamic acid -sodium citrate has also been studied.149 It must be noted that melanoidin does not provide a stable system. Particularly in acidic media, it increases in molecular weight due to polymerization.l5O In the initial steps of browning (in the presence or absence of amino compounds),low-molecular-weightcompounds of antioxidativeactivity are formed. Their application possesses some promising features in food preservation and proces~ing.'~'-'~~ In order to protect food against a change in color (browning) on storage, Japanese scientists154have proposed the use of inhibitors. Magnesium sulfate or chloride, potassium metabisulfite, or sodium sulfite, sulfate, or polysulfate can play such a role. For instance, a mixture composed of D-fructose, L-alanine, sodium citrate, and magnesium sulfate did not change its color when kept in the dark for 30 days at room temperature. A particularly strong browning-effect is achieved when ammonia is applied, because the reaction time is shortened and the reaction temperature d e ~ r e a s e d . ~ JAliphatic ~ ~ - ' ~ ~ primary and secondary amines and diamines may replace ammonia in these reactions,158because of their catalytic effect, but they are not on the list of permitted additives. Thus, Gow Chih Yen159 studied such reactions from the point of view of producing the most inten(148) Y. H. Chung, C. K. Kim, and W. J. Kim, Hunguk Sikpum Kwuhukhoechi, 18 (1986) 55-60 Chem. Abstr., 105 (1987) 41,370~. (149) V. F. Selemenev, I. P. Shamritskaya, K. Leps, and G. Yu. Oros, Izv. Vyssh. Uchebn. Zaved., Pishch. Tekhnol., (1979) 40-44. (150) N. Okada, T. Ohta, and H. Ebine, Nippon Nogei Kuguku Kuishi, 56 (1982) 93- 100; Chem. Abstr., 96 (1982) 197,998~ ( 1 5 1) A. Huyghebaret,L. Vandewalle, and G. van Landxhoot, Recent Dev. FoodAnul., Proc. Eur. ConJ Food Chem., 1st. 1981 (Publ. 1982) 409-415; Chem. Abstr., 98 (1983) 33,2 1 1 w. (152) J. Y. Byeong, H. L. Kang, and H. L. Jong, Hunguk Susan Hukhoechi, 19 (1986) 212218; Chem. Abstr., 105 (1987) 132,293n. (153) H. Lingnert and G . Hall, Dev. FoodSci., 13 (1986) 273-279. ( 1 54) A. Korakai, M. Mihara, and K. Nakayasu, Jpn. Pat. 61 265,067 (1985); Chem. Abstr., 106 (1987) 83,330~. ( 1 55) N. V. "Lyempf" Leeuwarder Ijcen Melkproduktenfabriken,Dutch Pat. 58,900 (1947); Chem. Abstr., 41 (1947) 4872. ( 1 56) E. Hamaguchi, M. Kawamoto, and H. Ishigashi, Sci.Repts. Hyogo Univ. Agric., Ser. Agric. Chem., l(1954) 63-65; Chem. Abstr., 49 (1955) 12,024. ( 1 57) M. Komoto and H. Ishigashi, Prog. Res. Soc. Jpn. Sugar Ref: Technol.. 4 (1955) 1-8; Chem. Abstr., 50 (1956) 13,488. (158) R. P. Modi and J. S. Pai, J. Food Sci. Technol., 21 (1984) 102- 104. (159) C. Y. Gow, Chung-kuo Nung Yeh Huu Hsueh Hui Chih, 24 (1986) 229-236; Chem. Abstr.. 106 (1987) 100,941h.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I

225

sively colored caramel. Histamine, putrescine, and tyramine heated with sugars produced the best colors in the case of D-ribose, followed by D-xylose, D-glucose, D-fructose, and sucrose. An increase both of pH (from 6 to 8) and temperature favored the browning. Light, especially ultraviolet irradiation from a low-pressure mercury lamp, activates the browning, with results similar to those obtained by thermal activation.lm It has also been found'61that ultraviolet irradiation of starch in aqueous solution oxidizes it to carbon dioxide in 98% yield, but does not produce any mono- and oligo-saccharides. The observations from the reactions of amino acids with mono- and di-saccharideshave been logically employed for caramelization of molasses. A strong color, stable to tannin, sodium chloride, and alcohol was obtained at pH 9 with glycine or L-lysine as the catalyst.16* Studies have been made of the influence of y-radiation on starch, and it was found that depolymerization takes place, and water-soluble "dextrins" are formed, but the aim of those studies was not connectedwith the preparation of syrups for ~aramelization.'~~-'~' However, nonenzymic browning can be induced by y - r a d i a t i ~ n . ' ~In~ such , ' ~ ~ cases, the formation of melanoidins from sugars and amino acids competes with reaction of both types of reagent with hydroxyl radicals produced from water, the reaction medium. 174

(160) S. A. Sheldon, G . F. Russel, and T. Shibamoto, D o . FoodSci., 13 (1986) 145- 154. (161) V. G. Soyer and A. D. Semenov, Gidrokhim. Inst. Gidrokhim. Muter.. 57 (1973) 195201. (162) B. H. Ryu and B. H. Lee, Hanguk Yongyang Siklyong Hakhoechi. 10 (1981) 93- 101; Chem. Abstr., 98 (1983) 142,046s. (163) M. A. Khenokh, Zh. Obshch. Khim., 20 (1950) 1560-1567. (164) M. A. Khenokh, M. A. Kuzicheva, and V. F. Evdokhimov, Tr. 2-go Vses. Soveshch. Radiats. Khim., Akad. Nauk SSR. Otd. Khim. Nauk Moscow, 1960, (1962) 409-4 14; Chem. Abstr., 58 (1963) 4075. (165) M. Samec, Staerke, 13 (1961) 283-292. (166) V. F. Oreshko, L. F. Gorin, and N. V. Rudenko, Zh. Fiz. Khim.. 36 (1962) 1084- 1085. (167) S. E. Traubenberg, K. N. Korotchenko, and I. N. Putilova, Zzv. Vyssh. Uchebn. Zaved., Pishch. Tekhnol.. (1965) 24-28. (168) J. P. Michel, M. Rigonard, G. Berger, and L. Saint-Lebe, Staerke. 29 (1977) 254-260. (169) G. Berger, J. F. Dauphin, J. P. Michel, G. Enrico, J. P. Angel, F. Seguin, and L. SaintLebe, Staerke, 29 (1977) 80-90. (170) G . Berger, J. P. Angel, and L. Saint-Lebe, Staerke, 29 (1977) 40-47. (171) J. P. Michel, J. Raffi, L. Saint-Lebe, M. Huchette, and G . Fleche, Staerke. 32 (1980) 340,344. (172) A. +gotaand S. Bachman, Nahrung, 28 (1984) 1037-1044. (173) A. Zegota and S. Bachman, Z. Lebensm. Unters. Forsch., 184 (1987) 3-7. (174) M. L. Wolfrom and P. McWain, J. Am. Chem. SOC.,81 (1959) 1221-1223.

226

PIOTR TOMASIK et al.

5. Preparation, Manufacture, and Storage of Caramel Studies of the kinetics of caramelization made by Ramaiah, Agarwal, and 178 revealed that, in the beginning, low-molecular-weightprodcoworkers17sucts are formed, and that these then condense. The resulting, high-molecular-weight compounds decompose into the products composing caramel. These reactions are of an order > 1.32,which means that at least two reactions are involved. The character of the compounds formed depends on the temperature, and their number depends on the concentrations of the reactants. Moreover, the degree of color formation is proportional to the time of heating. Alkali evidently catalyzescaramelization.This effect is more pronounced in the case of furanoses than in the case of pyranoses. There are four concepts of caramelization, as follows. (a) Thermal treatment of pure saccharides above their melting points. This concept includes reactions under n ~ r m a l , ~ d . ~i m ~ .i 'n~i ~ h e d ,and ~~,~~ enhanced122J26J28J29 pressure. The last approach used to be applied with syrups obtained by hydrolysis of starch in order to initiate the reaction. If any built-up pressure is released,the reaction is allowed to proceed in the open, to develop all of the color, viscosity, and desired organoleptic properties. (b) Thermal treatment in the presence of catalysts. This method allows decrease in the temperature of caramelization from 190-250"in the case of thermolysis of pure sugars to 120- 130"in the case of special caramels (such as spirit caramels). Caramelization above 200"yields products of both low tinctorial power and acrid flavor. (c) Treatment of saccharides either with mineral acids or alkali. This approach is based on acid- or base-catalyzed hydrolysis of oligosaccharides, followed by a number of reactions leading to 2-furaldehyde from pentoses, and 5-(hydroxymethyl)-2-furaldehyde from hexoses, as well as to dehydrated and to condensed, matter of ~ a r a m e l . ' , ' ~Also, - ~ ~ this reaction proceeds satisfactorily at temperatures significantlylower than those applied in the process listed under (a). In this case, high pressure may also be applied.

( 175) N. A. Ramaiah, S. K. D. Agarwal, and J. K. P. Agarwal, Proc. Indian Acad. Sci. Sect. A , 45

(1957) 97- 104. ( 1 76) N. A. Ramaiah, J. K. P. Agarwal, and S. K. D. Agarwal, Proc. Annu. Conv. Sugar Technol. Assoc. India. 24, Pt. I (1956) 69-80. ( 177) N. A. Ramaiah and S. K. D. Agarwal, Proc. Annu. Conv. Sugar Technol.Assoc. India, 28 (1960) 101-110. ( 1 78) S. K. D. Agarwd and D. S. Misra, Proc. Annu. Conv. Sugar Technol.Assoc. India, 39, Pt. I1 (1973) ~ 1 0 7 - ~ 1 1 8 . (179) F. Stolle, Z. Ver. Dtsch. Zuckerind., 49 (1899) 807-841; 53 (1903) 1149-1157. (180) F. Ehrich, Z . Ver. Dtsch. Zuckerind.. 59 (1909) 746-753.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I

227

(d) Treatment with ammonia, ammonium salts, amino acids, proteins, and polypeptides. These procedures usually lead to nitrogen-containing caramel. The Maillard or similar reactionI8' is responsible for this process (see also, the article by Ripp"). This kind of caramel is characterized by enriched flavor and aroma. Certain inorganic salts acting as catalysts may also be added. All four approaches have found practical applications in industry. The reactions proceed in entirely stainless-steelequipment, namely, kettles of the open or pressure type, lines, storage tanks, fillers, agitators, and so on. The parameters of the caramelization have to be quite precisely adjusted to a given source, as well as to the product desired. The Coca-Cola Corporation proposed a device for measuring the degree of caramelization of a free-flowing material; it is a specially designed spectrophotometer for measuring absorbance in the near-infrared region.Is2 A good caramel should contain colloidal particles, which do not precipitate under storage, and this also appliesto the products improved by addition of such a caramel. For this reason, the process must be strictly controlled, as, otherwise, particles of caramel can lose their micellar character, and this is followed by their precipitation. In order to prepare a good caramel, the isoelectric point should be adjusted at the beginning of the process. It should be kept in mind that any change of isoelectric point of caramel during the process is quite complicated, and not always successful.1s3 The control of the viscosity of caramel is a quite complicated matter. The rate of evolution of water (dehydration)is one of the most important parameters which influence the properties of caramel in general.61The time of mutual contact of reagents, and manipulation of the temperature may lead to caramels of desired viscosity. The concentration and origin of syrups being caramelized is of lesser irnportan~e.'~~ Sometimes, dry caramels are required. These can be prepared either by treating hot (120")viscous caramel with ammonium carbonate, followed by adding sucrose and orthophosphoricacid, cooling to loo", and adding citric acid and sodium hydrogen~arbonate.~~~ Others proposed addition of such cereal products as rye flour, and conditioning186of the mass at 80 - 85 at pH 3.5 - 5.5, or by thickening of the liquid caramel with a mixture of starch and (181) (182) (183) (184) (185) (186)

L. C. Maillard, C. R.Acud. Sci., 154 (1912) 66-68. Coca-Cola Co., U. S . Pat. 173,008,28 (1980); Chem. Abstr., 96 (1982) 179,760e. J. Wickens, Drinks TradeJ., 29(1975) 121-122. K. G. Weckel and J. Steinke, Manuf: Confect.. 53 (1973) 24-27, 330. L. F. Lavie, Belg. Pat. 889,126 (1980); Chem. Abstr., 96 (1982) 54,2082. K. Grott, S. Gapski, and A. Zielinski, Pol. Pat. I 19,874 (1985); Chem. Abstr., 103 (1985) I59,4 15J

228

PIOTR TOMASIK et a1

d e x t r i n ~ .Ajinomotola8 ’~~ patented the production of solid caramel by extrusion of mono- and di-saccharides at 150- 300”. The terms “forcing” and “stewing” caramel describe some undesirable results of processing of sources when the temperature- time conditions are not strictly maintained. In an effort to produce a product of deeper color, it is easy to obtain so-called “strained” caramel. When caramel is manufactured in the presence of ammonia, it is easy to raise the temperature above a desirable point. Therefore, the temperature should be carefully maintained by use of an effective cooling-system. The final period of the process, the so-called “killing heat” is very important. A variety of methods is employed in this respect. The first is the application of quench water sprayed through nozzles into caramel, to bring the temperature down to 30”. However, such a caramel is not very table.^.'^' Cleland and Longeneckerlg9proposed work-up of caramelized syrup by spraying it into a large volume of 4 : 1 mixture of 85% aqueous ethanol and 20% ethyl ether; the caramel was claimed to be stable, and to possess a high tinctorial power. Other methods were described by Green~hields.~~ The current trends in the manufacture of caramel involve the contact of the flow of sugar solutionwith heat in the heat exchanger. The sugar solutionis eventuallyenriched in some catalysts.1w-194 Caramelization of plain sugars in an oxygen-free atmosphere has been reported.9*YwBecause in ammonia caramels,the content of 4( 5)methylimidazole becomes the most important factor, development of a process that offers the lowest possible level of this neurotoxin has attracted attention. 195-197 It is well known that caramels from ammonia processes used to have a higher tinctorial strength than the other caramels. Even so,the color properties of ammonia caramels may appear to be insufficient. All attempts to produce a caramel of higher tinctorial strength may result in formation of a product having other undesirable properties, among them, a higher content of 4(5)-methylimidazole. Decades ago, studies were developed on the en-

-

(187) K. K. Ajinomoto, Jpn. Pat. 5 61 1,426 (1981);FSTA, 14 (1982) IL50. (188) K. K. Ajinomoto, Jpn. Pat. 5 546,148 (1980); FSTA, 13 (1980) 10L695. (189) J. E. Cleland and J. B. Longenecker,U. S. Pat. 2,553,221 (1950);Chem.Abstr., 45 (1951) 21 18. (190) Generale Sucriere S. A., Ger. Pat. 2 135,497 (1971). (191) Generale Sucriere S. A,, Ger. Pat. 2 245,379 (1972). (192) 0. Ackermann, U. S. Pat. 3,385,733 (1968); Chem. Absrr., 69 (1968) 34,809k. (193) T. Ohira, E. Hara, and Y. Takagi, U. S. Pat. 4,138,271 (1979); Chem. Abstr., 90 (1979) 185,200~. (194) Generale Sucriere S. A., Br. Pat. 1,358,807 (1972). (195) Y. Huang, S. Zhang, and R. Yang, Tiaowei FushipiuKeji, (1983) 1 1 - 12; Chem. Abstr., 99 (1984) 103,884g. (196) S. R. Ramaswamy, U. S. Pat. 4,614,662 (1986); Chem. Abstr., 105 (1987) 224,9433’. (197) J. Bielawny, Lebensm. Ind., 33 (1986) 261-2, 266.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I

229

hancement of the color of caramel by physical methods. Ultrafiltration has appeared to be a suitable tool in this respect. 198-204 Centrifugation combined with size-exclusionchromatographyhas been ~ a t e n t e d . Ultrafiltration ~~”~~ also helps to remove from caramel some impurities, such as 4(5)-methylimidazole,but their separation is not precise enough.204Some chance ofincrease of the tinctorial strength of caramel color results from the fact that magnesium and calcium hydroxides207and calcium phosphat2°8J09better absorb caramel than melanoidins present in such colors. “Bad” caramel means not only that it has a low tinctorial strength, and precipitates from liquids that have been colored with it, but also that it may have an unpleasant, burnt, and bitter taste, and aroma. Moreover, such overburned caramels may be excessively viscous. Undesirable properties of caramel may appear even in a product that has been properly manufactured. It is believed that caramelization progresses during storage, and it is only slowed down at lower temperatures. On prolonged storage, caramel resinifies into an amorphous, irreversible gel, useless for the purposes for which it was designed.Therefore,caramel should be stored at low temperature, possibly in plastic-lined drums or barrels. If these conditions are obeyed, caramel is stable for one to even five years.85If not plastic-lined, tanks should be made of stainless steel, to prevent caramel from contamination with metal, which, with caramel components, may give some addition (198) G. N. Bollenbeck and H. M. Molotsky, U. S. Pat. 3,249,444 (1966). (199) S. Kishihawa, J. Satoshi, and M. Komoto, Nippon Nogei Kagaku Kaishi, 53 (1979) 273-275; Chem. Abstr., 92 (1980) 56,930r. (200) S. Kishihawa, M. Komoto, and D. Nomura, Nippon Nogei Kagaku Kaishi, 53 (1979) 305-31 1; Chem. Abstr., 92 (1980) 92,839J (201) S. Kishihawa, S. Fujii, and M. Komoto, Nippon Shokuhin Kogyo Gakkaishi, 27 (1980) 479-482; Chem. Abstr.. 94 (1981) 6 3 , 9 2 9 ~ . (202) A. V. Clark, D. V. Myers, andV. I. Hatch, U. S. Pat. 4,416,700(1983); Chem. Abstr., 98 (1983) 159,374~. (203) A. V. Clark, D. V. Myers, and V. I. Hatch, Eur. Pat. 70,559 (1983); Chem. Abstr., 98 (1983) 159,374~. (204) S . Fujii, S. Kishihawa, and M. Komoto, Nippon Shokuhin Kogyo Gakkaishi, 24 (1977) 236-242; Chem. Abstr., 91 (1979) 191,501s. (205) H. H. Sharton, H. M. Molotsky, and M. Hyman, Ger. Pat. 1,517,008 (1972); Chem. Abstr., 77 (1972) 150,679g. (206) M. S . Tibbets andG. J. Templeman, U.S. Pat. 4,325,743 (1982); Chem.Abstr., 96 (1982) 216,367e. (207) I. F. Bugaenko and M. S. Gouda, Izv. Vyssh. Uchebn. Zaved., Pishch. Tekhnol., (1980) 58-60. (208) S . C. Sharma and P. C. Johary, Proc. Annu. Conv. Sugar Technol. Assoc. India, 46th. (1982) ~ 1 - ~ 1 7 . (209) S . C. Sharma and P. C. Johary, Proc. Annu. Conv. Sugar Technol. Assoc. India, 47th, (1983) ~ 7 - ~ 2 3 . (210) C. R. Volmer, Z . Ver. Dtsch. Zuckerind., 45 (1895) 451 -473.

PIOTR TOMASIK et ul.

230

Some attention has been paid to the possibility of manufacture of 5-( hydroxymethyl)-2-furaldehyde by thermolysis of hexoses, or di-, as well as poly-, saccharides. It may be prepared from D-fructose in the presence of either immonium salts211or zirconium phosphate by heating in dimethyl sulfoxide.’12 As shown by heating of sucrose in anhydrous dimethyl sulfoxide, the thermolysis is first-order in s u c r o ~ e .The ~ ~reaction ~ , ~ ~ is ~ facilitated by intermolecular hydrogen-bonds to water and alcohols. The thermolysis gives a-D-glucopyranose and j?-D-fructofuranosyl carbonium ion. It is a precursor for the formation of 2,6-anhydro-~-fructofuranose.~~~~~~~ The thermolysismay also be carried out in ketones.216 It has also been shown that acidic catalystsare suitable for the manufacture of 5-(hydroxymethyl)-2-furaldeh~de.~ l7

6 . Uses of Caramel Caramel is manufacturedin order to change,and improve, the appearance (color) of many food products; enhancement of the flavor of food is also important. Thus, it is used in the preparation of meat products,218and in brewing, including vinegar making (see, for instance, a Chinese patent2I9), mineral-water manufacture, blending of spirits, whiskey, rum, and wines, as well as soft, and other, drinks. Other uses, such as browning of gravies, yeast extracts, sweets, biscuits, pickles, sauces, dog food, crisp potatoes, and sugar coloring are quite common. Caramel is commonly used in oriental cuisines for coloring and flavoring soups, gravies, and sauces, for instance, shoyu (soy sauce)22o and pastries. Caramel for such purposes is not necessarily prepared from a plain sugar, such as sucrose. Frequently, caramel is prepared from soy bean, tapioca, or sago. The presence of proteins in caramelized saccharidecontaining material enriches the flavor of the products and leads to specific organoleptic properties. The user of caramel has to select among available grades of caramels in order to achieve the result desired. Two properties of caramel are most (21 1) (212) (2 13) (2 14) (215) (2 16)

C. Gelas and J. Fayet, Curbohydr. Res., 122 (1983) 59-68. Y . Nakamura, Jpn. Pat. 80 13,243 (1980); Chem. Abstr., 93 (1980) 26,260e. 0. K. Kononenko and K. M. Herstein, Chem. Ind. Data Ser, 1 ( 1 956) 87 - 92. L. Poncini and G. N. Richards, Curbohydr. Res., 87 ( 1980) 209 - 2 17. W. Moody and G. N. Richards, Curbohydr.Res., 97 (1981) 247-255. G .Fleche, A. Garet,J.-P. Gorrichon,E. Truchot, and P. Sicard, Fr. Pat. 2,464,266 (198 1); Chem. Abstr., 96 (1981) 6552k. (2 17) H. E. van Dam, A. P. G. Kieboom, and H. van Bekkum, Stuerke, 38 (1986) 95 - 101. (218) D. Chundury and H. H. Szmant, Ind. Eng. Chem., Prod. Res. Develop., 20 (1981) 158-163.

(219) L. Guokun, Chin. Pat. 85 104,589 (1986); Chem. Abstr., 106 (1987) 4 9 , 0 0 0 ~ . (220) H. Chiba, Nippon Jozo Kyokui Zusshi, 74 (1979) 601-603; Chem. Abstr., 92 (1980) 109.18 1 k.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I

23 1

important to bear in mind. They are isoelectric properties and tinctorial strength, although the flavor is also important. Caramel for brewing (Caramel Color Ammonia Class 111) has a weaker tinctorial strength; it must be absolutely chill-proof and not cause a haze in beer. Its pH ensures that the flavor of beer shall remain unchanged. A haze may form when there is discrepancy between the isoelectric points of the beer components and the caramel. Flocculation results from discharge of colloidal micelles. Moreover, beer (positive) caramel has to withstand fermentation. It may also be used for blending lemonades, and strongly acidic foods and alcohols. Caramel Color Sulphite Ammonia Process Class IV was designed as a finishing additive for soft drinks, but not for spirits. The isoelectricpoints of most soft drinks are below 2.5, and that of the caramel should be < 1.5. Baker's types of caramel have to possess reasonable tinctorial strength, be free-flowing, and have a low viscosity, in order to afford a uniform shade of finished goods; the pH should be -4.0-4.2 The turbidity in brandy colored with caramel depends not only on the character of the micelles in the caramel but also on the content of the micelles and calcium in it. A correlationhas been reported between the content of these components in the caramel and the level of turbidity of the colored brandy.221 It has been reported that brandy spirits blended with sugar syrup or caraApart from flavoring and colormel (0.2%v/v), or both, accelerates aging.222 ing properties, caramel exhibits some stabilizing properties in respect to aspartame concentrates. Over 80% of aspartame was protected against decomposition on storage for 6 weeks at 40°, whereas a control sample then contained less than 30% of aspartame.223The pharmaceutical industry utilizes caramel for the stabilization of the color, and resistance to heat of gelatin compositions used for coatings, capsules, dragees, and the as well as to enhance the rate of dissolution, of gelatin coatings in water and saliva.225Caramel also stabilizes the dispersing property of heat-absorbing liquids comprised of aqueous propylene glycol and poly(vinylpyrro1idinone).226 Some medicinal cough-mixturesare blended with caramel, as both a coloringand a flavoringcomponent.227 Caramel is used as a permanent color (221) P. Ya. Mishev, E. A. Martynenko, I. A. Egorov, and B. N. Efimov, Prikl. Biokhim. Mikrobiol., 16 (1980) I41 - 143. (222) Ts. L. Petrosyan,L. M. Dzhanpoladyan, and R. S. Dzanazyan, Vinodel. Vinograd.SSSR, ( 1980) 24 - 26. (223) S. Sharma, Br. Pat. 2,104,369 (1983); Chem. Abstr., 98 (1983) 196,7146. (224) D. K. K. Parke, Jpn. Pat. 80 141,242 (1980); Chem. Abstr., 94 (1981) 71,5302. (225) K. Nemoto, T. Ogasawara, and S. Bessho, Ger. Pat. 3,011,044 (1980); Chem.Abstr., 94 (1980) 7716n. (226) Deinichiseika Color and Chemicals Mfg. Co., Jpn. Pat. 58,174,486(1983); Chern.Absfr., 100 (1984) 142,2941: (227) G. W. Pace, Food Technol.Aust., 22 (1970) 522-524.

232

PIOTR TOMASIK et al.

additive for general use in cosmetics. The results of short-term eye-area and ninety-day skin study indicated that caramel is a safe agent.228Because caramel exhibits a synergistic effect with cis-9-tricosene, it is employed as a component of fly-attracting baits against the domestic fly.229 Caramel in low concentration gives a stable and intense yellow color. Therefore, it was proposed as a color compatible with Yellow No. 5 (tartrazine) for nontransparent drinks.2J0The wide range of shades of color of caramel has attracted the attention of other industries. Thus, coatings and other articles having a cork-likeappearance are manufactured from thermoplastic resins colored with ~aramel.2~' Coloring of polyethylene terephthalate with caramel was also patented. Caramel may be formed in situ from carbohydrates in the presence of antimonous oxide as the catalyst; the darkbrown polymer resulting is said to be nontoxic.232 The viscosity of caramel allows its employment as an adhesive and binder for instant ~ a n d - m o l dand s ~ ~foundry ~ core for producing difficultlyaccessible, hollow cavities in castings. Caramel binds together a sodium chloride, quartz, or metal-powder matrix to a mass that binds to aluminum or aluminum-alloy castings. On casting, such a binder produces only a small volume of combustion gases. The cores are readily removed, either mechanically or by dissolution.234 The heterocyclic character of caramels from ammonia and amino acid processes may suggest that such caramels may be quaternized by metal ions to form metal coordination compounds. Indeed, such complexeswith Cuz+, Cd2+,Ni2+, and Zn2+have been found, and studied by Agarwal and cow o r k e r ~ .Because ~~~.~ caramel ~ ~ may be a transporter of ions in solution, it has been the subject of a study of its surface activity and its usefulness as a brightener in electroplating baths.237

(228) U. S . Food and Drug Administration, Fed. Regist., 46 (27 March 1981) 1894- 1895. (229) C . Heunart, Ger. Pat. 2,928,204 (1978);Chem. Abstr., 92 (1980) 175,784k. (230) C . Andres, Food Process., 41 (1980) 102. (231) G. V. Paisley, and A. Melaspina, Eur. Pat. 97,783 (1984); Chem. Abstr., 100 (1984) 86,943~. (232) M. A. Werner,A. Venerna, and M. G. H. Pisters, Eur. Pat. 61,210(1982);Chem.Abstr., 98 (1983) 5 1 8 3 ~ . (233) Daicel Chemical Industries Ltd., Jpn. Pat. 58,176,048 (1983);Chem. Abstr., 100 (1984) 72,630m. (234) W. Wischnack and A. Dobner, Eur. Pat. 19,015 (1980); Chem. Abstr., 94 (1981) 126,0352. (235) S . K. D. Agarwal, V. K. Gupta, and S. K. Upadlyay, Proc. Annu. Conv. Sugur Techno!. Assoc. India, 43rd, (1979) ~ 1 4 3 2 0 . (236) S . K. D. Agarwal and V. K. Gupta, Proc. Annu. Conv. Sugar Technol. Assoc. India, 44th, (1980) ~ 2 4 - ~ 3 0 . (237) P. Tomasik, unpublished results.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I

233

TABLE V The Application, World Production, and Acceptable Daily Intake (A.D.I.) of Four Classes of Caramels * Class

Applications

I

I1

111

IV

Spirits Brandy Candies Medicines Cookies Pastries Aromas Spices

Special spirits

Beer Malt liquor Bread Cookies Pastries soups Sauces Canned food Meat Tobacco Spices

Cola-type beverages So!? drinks Vermouths Vinegar

-

Production" World U.S.A. E.E.U.

20 5 4 -

A.D.I. (mg/kg)

1

<

not specified

45

.

80 95 50 z

0-200

~~

In % of an overall 60,000 tons per year.

The use of caramels for other than alimentary purposes is of rather marginal importance. Four classes of caramel (according to the specificationsin Table I) are offered for alimentary purposes all over the World. Table V presents the range of application of caramels of particular classes, together with their overall World consumption238in 1986. The figures of acceptable daily intake of particular caramels, quoted according to Joint FAO/WHO are also given. Expert Committee on Food The forecast for 1986 was the production of 100,000tons per year. It may, however, be presumed that the overall consumption of caramels will decrease in the near future. For instance, the Polish producer of caramels has governmental permission for production of ammonia caramels up to the end of 1989. Beyond 1989, any manufacture and consumption of ammonia caramel will become illegal. This decision was caused by the presence of 4(5)-methylimidazolein such caramel. In Austria, in 1983, such caramels could be used solely for coloring beer.27 (238) H.-D. Smolnik, Staerke, 39 (1987) 28-32. (239) Meet. Joint FAO/WHO Expert Comm. Food Additives, Geneva, 1985, Report No. 29 (1985); ICS Food Add., Sum. 85,S.4.

PIOTR TOMASIK et al

234

7. Detection of Caramel, and Some Aspects of the Analysis of Caramel The problem of detection of caramel has been reviewed t ~ i c e . ~ ~Joszt ,~'"' and M ~ l i b s kmade i ~ ~ a comparison of the methods. Thus, Jager~chmidt~~l had used either ethereal or acetone solutions of caramel-containing syrups, which were treated with resorcinol in the presence of hydrochloric acid. A red color is developed in ether, and a violet-red in acetone. Probably, the reaction is due to 5-(hydroxymethyl)-2-furaldehydepresent in the caramel. A method suggested by A m t h 0 9 ~uses ~ paraldehyde in absolute alcohol. Caramel is present if, after 24 h, there is formed a brown precipitate that reacts with phenylhydrazine hydrochlorideto give a solid insoluble in hydrochloric acid but soluble in ammonia and in alkali. The reaction is due to caramelan present in the caramel. A yellow to brown color is developed when ammonium sulfate in 96% ethanol is added and shaken with the caramel. This is the so-called Griessmayer- Aubry meth~d,"~which has some quantitative significance. In the Lichthardt an aqueous solution of tannin acidified with sulfuric acid precipitates a brown solid within 24 h. The method of F r a d i ~ employs s ~ ~ ~ dry 1-pentanol;if caramel is present, a precipitate is formed. Crampton and S i m o n ~developed ~~~ a method based on decolorization of aqueous or ethereal solutions of caramel with such powdered earths as Floridin, Tonsil, or Fuller's earth. The resulting color is determined colorimetrically. A light-yellow color and solid appear when a diluted solution of the caramel is treated with 1% aqueous stannous chloride in the presence of potassium acetate, according to Stra~b.~ The ~ ' method of Ne~sler,2~* and C a r l e gives ~ ~ ~a ~brown to orange color with fresh egg-white. The method of IhlZSOis, as a matter of fact, a variation of the Jager~chmidt~~I method. Use of pyrogallic acid in hydrochloric acid results in formation of a dark-red precipitate. In the Magalhaes method,251a light-orange color is developed when the test solution is boiled for 10 min with potassium sulfateand cotton wool. Schenk2'"'recommended (240) (241) (242) (243) (244) (245) (246)

D. Schenk, Apoth. Zfg.,29 (1914) 202-203. A. Jagerschmidt, Z . Unters.Lebensm., 17 (1909) 113- 115,269. C. Amthor, Fresenius Z . Anal. Chem.. 24 (1885) 30-33. V. Griessmayer, Pharm. Zentralhalle.2 1 (1880) 368-374. G. H. P. Lichthardt, J. Ind. Eng. Chem., 2 (1910) 389. M. N. Fradiss, Bull. Assoc. Chem. Sugar Dist., 16 (1898) 280. C. A. Crampton and F. D. Simons, J. Am. Chem. Soc.. 21 (1899) 355-358; 22 (1900)

(247) (248) (249) (250) (251)

A. Straub, Pharm. Zentralhalle, 52 (191 1) 868. J. Nessler, Weinlaube,2 (1870) 119- 122. E. Carles, J. Pharm. Chem., 22 (1875) 127. A. Ihl, Chem. Ztg., 9 (1885) 485. A. J. Magalhaes, C. R. Acad. Sci, 123 (1896) 896-897.

810-831.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I

235

a test with phenol or 2-naphthol as being very sensitive. The color with phenol is developed immediately, whereas that with 2-naphthol appears after 30 min. These chemical reactions and tests for caramel are complemented by a group of physical methods based on size-exclusion chromatography. These methods may be applied for the detection of caramel in beverages?6 beer?sz and wine.253Caramel may'be detected in breadzs4and in various slightly colored products from the sugar industry (raw sugar, molasses, sugar syrups, and the like).zssSpectral methods are most useful for these purposes. An insight into the composition of low and even higher molecular-weight components of caramel is also possible. For the investigation of high-molecular-weight fractions, size-exclusion chromatography has been the technique most e ~ p l o i t e d , 6 ~but ~ ~strong ~ ~ anion-exchange ~ - ~ ~ ~ ~ ~ resni~s2@'~ ~ ~ ~ - ~ ~ ~ and e l e c t r o p h o r e s i ~may ~ ~ ~also - ~ ~be~used. Low-molecular-weight fractions may be analyzed by various methods; among them, chromatographic methods seem to be the most convenient. Thus, Tyagunova and coworkerszMcould identify D-glucose, D-fructose, and pyruvk acid in caramel from sucrose by using paper chromatography. This analysis was preceded by partition, involving ion-exchangers, of caramel into fractions. Unreacted residual amounts of saccharides used for the caramelization may be identified by means of classical methods suitable for the determination of sugars.z65Gas - liquid chromatography was introduced into the analysis of low-molecular-weight fractions of long ago. Patey and cow o r k e r ~completely ~~~ identified 8 (and, partially, 57) components out of a (252) U.Lessig, Lebensm. Chem. Gerichtl. Chem., 38 (1984) 64-68. (253) P. Mattyasovszky and Z. Jeszenszky, Borguzdusug, 33 (1985) 105- 110. (254) L. Magrian, J. Pagenkopf, M. Schilling,and U. Sell, Dtsch. Lebensm. Rundsch.,8 I (1985) 379 - 384. (255) H. M. Saber, I. F. Bugaenko, and A. R. Sapronov, Izv. Vyssh. Uchebn. Zaved., Sukh. Prom.. (1980) 37-38. (256) E. E. Stinson and C. 0. Willits, J. Assoc. OffAnal. Chem., 46 (1963) 329-330. (257) H.-J. Schoene, Brauwissenschuft,26 (1973) 344-35 1. (258) H. Kieninger, Bruuwek 120 (1980) 560-569. (259) J. Ujszarsi, Szeszipur, 28 (1980) 46-50, 87-92. (260) D. L. Ingles and D. Gallimore, Chem. Ind. (London), (1985) 194. (261) D. Gross, Int. SugurJ., 69 (1967) 323-328. (262) D. Gross, Int. Sugar J., 69 (1967) 360-365. (263) S. K. D. Agarwal and S. K. Gupta, Proc. Annu. Conv. Sugar Technol. Assoc. India, 39 (1973), Pt.I, ~ 6 7 - ~ 7 5 . (264) V. I. Tyagunova, M. V. Rozhkova, and G. A. Chikin, Teor. Prukt. Sorbts. Protsessov, (1980) 103-106. (265) V. V. Aleksandrov, Tr. Vses. Nuuchn.-Issled.Inst. Konditer. Prom., (1941) 161 - 197. (266) H. Sugisawa and H. Edo, Chem. Ind. (London), (1964) 892-893. (267) A. L. Patey,G. Shearer,M. E. Knowles, andH. B. Denne, FoodAddit. Contam., 2 (1985) 237 - 246.

236

PIOTR TOMASIK et a1

total of 143 compounds of ammonia, alkalis, and ammonium sulfite caramels by using a preliminary extraction of the caramel with chloroformethanol and gas- liquid chromatography coupled with mass spectrometry. Several methods have been specially refined for detection and determination of particular components of caramel. Among them, the detection and determination of 4(5)-methylimidazolein ammonia caramel presents a special task. The methods of detection are based on thin-layer chromatography of extracts of caramels. They were developed on plates covered with silica gel FZs4by using 4 : 1 : 1 ether- chloroform- methanol, and 4( 5)-methylimidazole (up to 0.2 mg) was detected with a solution of sodium nitrite combined with sulfanilic acid used in a spray.z68Italian workers269proposed either an essentially similar method or, alternatively, detection with a gas- liquid chromatograph equipped with a column packed with 10%ofcarbowax 20M with 2.5% KOH on CPLA (80- 100 mesh). The most modern methods of detection offer slightly refined conditions of extraction, and improved mobile phases for the elution of thin layers; for instance, 20 :5 :5 : 1 ether chloroform- methanol - 25% ammonium hydroxide,z70 or a different column for gas- liquid chromat~graphy.~~~ The determination of 4(5)methylimidazole is possible by means of thin-layer ~hromatography,2~’,~~~ or reversed-phase ion-pair liquid chromagas - liquid ~hrornatography,2~~ t o g r a p h ~Further . ~ ~ ~ components of caramel were gradually identified, and methods for their detection became available. Thus, methods of detectionz74 and determinationz75of 2-acetyl-4(5)-( 1,2,3,4-tetrahydroxybutyl)irnidazole were published. Also, 6-methyl-3-pyridinol was found in caramels. Binder and coworkers276 suggestedthat it may be used as an indicator compound for caramels. They presented determination of that compound by means of a gas- liquid chromatograph coupled with a mass spectrometer. Chromatographic methods of detection and determinationof 2-furaldehyde and 5-(hy(268) W. H. Lam and M. Y. Takahashi, Rev.Inst. Adorfo Lutz, 39 (1979) 155- 159; Chem. Abstr.. 92 (1980) 109,2281’. (269) L. Dagna, M. Fenocchio, and G. Gasparini, Boll. Chim. Unione Ital. Lab. Prov. Par& Sci., 6 (1980) 327-334. (270) G. Lehmann and B. Binkle, Lebensm. Chem., Gerichtl. Chem., 41 (1987) 9 - 10. (271) B. Gao, Zhongguo Tiaoweipin, (1985) 6-9; Chem. Abstr., 103 (1985) 213,479r. (272) J. DingandJ.Shen, Zhonhua YufangyixueZazhi, 18(1984)364-365;Chern.Abstr., 103 (1985) 52,772t. (273) M. Thomsen and D. Willumsen, J. Chromatogr., 21 1 (1981) 213-221. (274) U. Kroplien and J. Rosdorfer, J. Urg. Chem., 50 (1985) 1 I31 - 1 133. (275) U. Kroplien, J. Chromatogr., 362 (1986) 286-290. (276) H. Binder, C . Wentzel, H. Junck, and M. Mittelbach, Z. Lebensm. Unters. Forsch., 184 (1987) 187-188.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I

237

droxymethyl)-2-furaldehydeare also available(see, for instance, Collino and V ~ l p eas, ~well ~ ~as Alfonso and coworkers278).

8. Chemical Nature of Caramel The chemical nature of caramel depends on the conditions of the caramelization and the source, at least in the terminal period of their processing. Those caramels which are prepared from plain sugars without any catalyst present the relatively simplest chemical features. In order the better to understand the complexity of caramelization specific interconversions of sugars in solution have to be kept in mind. It is well known that pyranoid and furanoid sugars are considerably more stable than the acyclic forms.279Mutarotation of sugars is autocatalyzed by protons,280-282 and, with D-glucose, the equilibrium slightly favors /3-D-glucopyranose over the a anomer. It has

been claimed that, in neutral solution, the acyclic aldose exists in -0.003% c ~ n c e n t r a t i o nwith ,~~~ its importance increasing in acidic media;284it takes part in an e q ~ i l i b r i u mthat ~ ~ includes ~ , ~ ~ ~acid-catalyzed interconversion of pyranoses and f u r a n o s e ~ .This ~ ~ ~subject , ~ ~ ~has been surveyed in detail by Ang~al.~~~ (277) F. Collino and S. Volpe, Biol. Chim. Farm., 121 (1982) 375-386. (278) F. C. Alfonso, G. E. Martin, and R. H. Dyer, J. Assoc. OffAnal. Chem., 63 (1980) 1310-1313. (279) A. G. Salamon and E. L. Goldie, J. SOC.Chem. Ind., 19 (1900) 301 -307. (280) H. Euler, Ber., 39 (1906) 344-350. (281) L. Michaelis and P. Rona, Biochem. Z., 49 (1913) 232-248. (282) N. A. Ramaiah and S. S. Katiyar, Proc. Annu. Conv. Sugar Technol.Assoc. India, 29 (1961) 77-84. (283) J. M. Los and K. Wiesner, J. Am. Chem. Soc., 75 (1953) 6346-6347. (284) E. Pacsu and L. A. Hiller, J. Am. Chem. Soc.. 70 (1948) 523-526. (285) G. Hallas, Organic Stereochemistry, McGraw-Hill, London, 1965, Ch. 2. (286) T. H. Lowry and K. Schueller-Richardson,Mechanism and Theory in Organic Chemisfry,Harper and Row, New York, 1976, p. 425. (287) C. T. Bishop and F. F. Cooper, Can. J. Chem., 40 (1962) 224-232; 41 (1963) 27432758. (288) G. G. S. Dutton and A. M. Unrau, Can. J. Chem., 40(1962) 1196-1200. (289) S. J. Angyal, Adv. Carbohydr. Chem. Biochem., 42 (1983) 15-68.

PIOTR TOMASIK et al.

238

CHZOH

CHZOH I

I

I

OH

OH

Caramel begins to form when concentrated aqueous solutions of sugars are heated290for a prolonged time at 100 . However, the first compounds were isolated from the caramel mass after heating to 120- 190" (and not above 240").The initial temperatures of decomposition of particular sugars as determined by heat-flow ~alorimetry~~' correspond to the temperaturesof their fusion.292From computerized curve-fitting of thermogravimetric data,293it was found that sugars having their structural units bonded through a-glycosidic bonds are less stable than those bonded through P-bonds. The absence of a free glycosidic hydroxyl group increases the thermal stability of the carbohydrate. Furanoses are less stable than pyranoses. Disaccharides start to decompose at a temperature 20" higher than that needed for monosaccharides. The products of caramelization are distributed between volatile and nonvolatile fractions. The composition of the volatile fraction is pretty well characterized, contrary to that of the nonvolatile fraction. Thus, neither is the structure of all compounds formed precisely known, nor are the processes which occur understood in detail (see, for instance, a review by O r ~ i ~The ~ ~composition ). of the volatile fraction from the thermolysis of sucrose is the best recognized. The profound decomposition products from the decomposition in vucuo of sucrose are water, carbon monoxide, carbon dioxide, formaldehyde, acetaldehyde, methanol, and ethanol. The detailed rates and temperature relationships suggestthat, with the possible exception of ethanol, the other products result from secondary reactions of dehydration products.295The low-molecular-weightportion of the nonvolatile fraction of the thermal degradation of sucrose contains D-fructose, ~-glucose, O

(290) (291) (292) (293) (294) (295)

M. N. Fradiss, Bull. Assoc. Chem. Sugar Dist., 16 (1898) 664. A. Raemy and T. Schweizer, Calorim. Anal. Therm., 13 (1982) 111.11.70-111.1 1.76. A. Raemy and T. Schweizer, J. Therm. Anal., 28 (1983) 95- 108. A. E. Pavlath and K. S. Gregorski, Proc. Eur. Symp. Therm.Anal., Znd, (198 1) 25 1 -254. F. Orsi, Edesipar, 36 (1985) 1-9. M. D. Scheer, Znt. J. Chem. Kinet., 15 (1983) 141- 149.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I

239

and pyruvic a ~ i d . Richardsz9' ~ ~ ~ - ~and ~ other~~~~-'O' ~ suggested that such thermal decomposition of sucrose without any catalyst added is, in fact, an acid-autocatalyzed process. Similarly, the decomposition of sucrose in aqueous solution is a proton self-catalyzed process (see the review by Ponchimz).At loo", water becomes remarkably acidicm3(pK, 12.23), and the pK, of sucrose at 90" is 30411.07. Therefore, sucrose is readily hydrolyzed into D-fructose and ~-glucose.This process is bimolecular, and the rate constant is linear in the concentration of hydrogen ion.m59306The color formation in aqueoussyrupsis a function of their concentration and tempera t ~ r e . It ~ is ~ 'well known that the pH of aqueous syrups of sucrose decreases as a function of time; they become more acidic. The pH of 0.5 M aqueous sucrose solution after 70 h at 90" reachesm83.93, due to formation of levulinic, 2-furoic, and 3-hydroxypropanoicacid and other compounds, among them ethyl lactate, 2-furaldehyde, and maltol, which could be identified by means of gas- liquid chromatography, infrared spectroscopy, and mass s p e c t r ~ m e t r yAqueous .~~ solutionsof sucrose heated above 150" exhibit the presence of further compounds, such as 5 4 hydroxymethyl)-2-furaldehyde, 2-furaldehyde, formaldehyde, acetaldehyde, butanal, and a ~ e t o n e . 2 ~ ~ ~ " ~ ~ ~ ' There are some differencesin the course and results as between caramelization in solution and in the solid state. For instance, kinetic studies in the solid are subject to various inaccuracies due to interferences from the effect of melting, and different rates on the phase boundary between solid and

(296) V. I. Tyagunova, M. V. Rozhkova, and G . A. Chikin, Teor. Prakt. Sorbts. Protsessov, 13 (1980) 103-106. (297) S . N. Richards, Int. Sugar J., 88 (1986) 145-148. (298) I. A. Egorov, V. N. Lominadze, and A. Yu. Sknpnik, Prikl. Biokhim. Mikrobiol., 10 (1974)681-687. (299) S . E. Kharin and A. R. Sapronov, Izv. Vyssh. Uchebn. Zaved., Sakh. Prom., 42 (1968) 26 - 29. (300) Z. A. Milkova, S. 2. Ivanov, and A. R. Sapronov, Izv. Vyssh. Uchebn. Zaved., Sakh. Prom., 43 (1969) 1 1 - 13. (301) V. A. Golybin and S. Z. Ivanov, Zesz. Probl. Post. Nauk Roln., 187 (1977) 89-93. (302) L. Poncini, Int. Sugar J., 82 (1980) 332-335. (303) F. H. C. Kelly and D. W. Brown, Sugar Techno/.Rev.,6 (1978) 1-47. (304) S. N. Ivanov and E. S. Lygin, Zh. Prikl. Khim., 41 (1968) 2722-2725. (305) J. G. Dawber, D. R. Brown, and R. A. Reed, J. Chem. Educ., 43 (1966) 34-35. (306) M. L. Wolfrom, A. Thompson, and C. E. Timberlake, Cereal Chem.. 40 (1963) 82-86. (307) T. Katsurai and Y. Makide, Bull. Chem. SOC. Jpn., 46 (1973) 3293-3294. (308) S. E. KhannandI. P. Palash,Izv. Vyssh. Uchebn.Zaved., Sukh. Prom., 41 (1967) 15- 17. (309) H. Ito,Agric. Biol. Chem., 41 (1977) 1307-1308. (310) R. Montgomery and L. F. Wiggins, J. SOC.Chem. Ind., (London), 66 (1947) 31 -32. (31 1) M. S. Bergdoll and E. Holmes, Food Res., 10 (1951) 50-56.

240

PIOTR TOMASIK et ul.

melt.312,313 The number of products is greater in thermal decomposition without solvent. It produces cu,P-unsaturated carbonyl compounds, 2-furaldehyde, 5 4 hydroxymethyl)-2-furaldehyde,pyruvaldehyde, 2-hydroxypropanedial (“glucoreductone”), fatty acids, succinic, fumaric, levulinic, and furancarboxylic acids; also, D - ~ ~ U Ckojibiose, O S ~ , isomaltose, nigerose, sophorose, laminarabiose, maltose, gentiobiose, cellobiose, isomaltotriose, panose, and other oligosaccharides, as well as products of reversion, and p o l y m e r i ~ a t i o n . ~LedPZ0 ~ ~ - ~ ’reported ~ the formation of 2,3-dihydro-4-hydroxy-5-methylfuran-3-oneand 2,3dihydro-3,5-dihydroxy-6-methyl-4Hpyran-4-one. The first of them readily consumes formaldehyde, to give 2,3dihydro-4-hydroxy-2-(hydroxymethyl)-5-methylfuran-3-one.A group of Japanese workers321found in caramels bis( 5,5’-formylfurfuryl) ether. As

210 230 250 270 200 310

j\ [nml FIG.3.-The

Infrared Absorption Spectran9of Sugar Stock before Caramelization(1) and of Caramel Formed from It (2). (312) G . N. Richards and F. Shafizadh,Aust. J. Chem., 31 (1978) 1825- 1832. (313) I. A. Serenkova and Yu. A. Shlapnikov, Izv. Akud. Nuuk SSSR,Khim. Ser.. (1977) 919-921. (314) M. L. Wolfrom, R. D. Schuetz, and L. F. Cavalieri, J. Am. Chem. SOC.,70 (1948) 514-517. (31 5) C. Enders and R. Marquardt, Nuturwissenschuften,29 (1941) 46-47. (316) P. E. Shaw, J. H. Tatum, and R. E. Berry, Curbohydr. Res., 5 (1967) 266-273. (317) S. Ramchander and M. S. Feather, Am. Assoc. Cereal Chem., 52 (1975) 166- 173. (318) E. F. L. J. Anet,Adv. Curbohydr. Chem., 19 (1964) 181-218. (319) K. Uobe, K. Nishida, H. Inoue, and M. Tsutsui, J. Chromutogr., 193 (1980) 83-88. (320) F. Ledl, Z. Lebensm. Unters. Forsch., 169 (1979) 176-178. (32 1) S. Fujii, M. Ishikashi, S. Kishihawa, and M. Komoto, Nippon Shokuhin Kogyo Gakkaishi, 27 (1980) 352-353; Chem. Abstr., 93 (1980) 202,808~.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I

24 1

reported by Wolfrom and coworker^,^^^-^^^ heating of D-fructose produces diheterolevulosan and di-D-fructose anhydride. Moreover, D-fructose is converted into D - ~ ~ U C O S ~ . The nonvolatile, high-molecular-weight fraction of caramel produced without any catalyst seems to contain three components. They are called caramelan, caramelen, and caramelin. Thus far, the problem of their chemical structure remains an intriguing, unsolved part of the chemistry of caramel. Several a ~ t h o r ~ ~ , have ~ ~ contributed . ~ ~ , ~to ~elucidation ~ ~ ~ of ~ the ~ , ~ ~ ~ structure of these materials, but there are still several doubts in this matter, and the authors derived only summarizing formulas. They are: 6 ClzH,,Oli - 12 HzO = 6 Cl2Hi2O9

for caramelan, and 6 CizHzzOll- 18 H 2 0 = 2 C36Hi8024

for caramelen. More light has been thrown on the problem by infrared-spectroscopic s t ~ d i e s . ~ ' ,These ~ ~ Jspectra ~ , ~ ~closely ~ resemble the corresponding spectra of sugars before caramelizationE9 (see Fig. 3). The spectra of caramelan and caramelen are also very similar to one another (see Fig. 4 and Table VI) and to the spectra of the starting materials. This means that the essential structural features of all of those compounds remain unchanged. Both caramelan and caramelen give positive tests for primary hydroxyl groups and aldehyde, and for secondary and tertiary hydroxyl groups, and the only difference found by Miroshnikova and coworkers8' lay in the test for ketose, which is positive for caramelan and negative for caramelen (see Table VII). Caramelan gives a positive Fehling test. It does not form salts, but adducts with lead oxide and with barium oxide can be isolated. Carmelen seems to absorb only slightlyin the visible region of light, contrary to caramelan (see Fig. 5). It also (322) M. L. Wolfrom and M. G. Blair, J. Am Chem. Soc., 70 (1948) 2406-2409. (323) M. L. Wolfrom, W. W. Binkley, W. L. Shilling, and H. W. Hilton, J. Am. Chem. Soc., 73 (1951) 3553-3557. (324) M. L. Wolfrom and W. L. Shilling, J . Am. Chem. Soc., 73 (1951) 3557-3558. (325) A. G. Green and A. G. Perkin, J. Chem. Soc., 89 (1 906) 8 I 1 - 8 13. (326) R. Willstatter and L. Zechmeister, Ber.. 46 ( I9 13) 240 I -24 12. (327) H. J. H. Fenton and M. Gostling, J. Chem. Soc., 79 (1901) 361-365. (328) E. Fisher, Ber., 26 (1893) 2406-2412. (329) F. Sestini, Gazz. Chim. Ital., 10 (1880) 156-245, 355-361. (330) W. B. Bottomley, Biochem. J., 9 (1915) 260-268. (331) J. B. Shumaker and J. H. Buchanan, Iowa State Coll. J. Sci., 6 (1932) 367-379. (332) A. R. Sapronov, E. F. Kozyreva, and K. N. Dunayeva, Izv. Vyssh. Uchebn.Zaved., Sakh. Prom., 42(1968)9-11.

FIG.4.-The ken Line),*l

-1 Wavenumber ( c m I Infrared Absorption Spectra of Caramelan (Solid Line) and Caramelen (Bro-

TABLE VI Infrared Absorption Spectra of Caramelan, Caramelen, and Caramelin33* Frequency (cm-I) Caramelan

Caramelen

3300

3352

3000 - 2830 2742 1707 1663 1523 1480 1403 1337 1266 1200 1018 916

2985 -2800 1704 1660 1515 1455 1395 1334 1259 1192 1036 912

Caramelin 3386 3020 - 2850

-

1707 1665 1515 1471 1395 1348

Band assignments

vOH of intra- and inter-molecular hydrogen bonds va alkyl ? vc-o ketonic vc-o enolic ? &HZ

Lf, &H,

? 1034" 920

vc-0-c

pyranose-ring

vibration 816 775

a

80 1 775

809 768

A shoulder.

242

&I-,

pyranose-ring deformation

TABLE VII Analytical Characteristics of Caramelan and Carsmeled' Result Functional group

Reaction

Primary hydroxyl and aldehydes Secondary hydroxyl

reduction of the Nessler reagent reaction with nitric acid (secondary alcohols give nitroalkanes) yellow precipitate with alcoholic KOH the Deniges reaction, a yellow or reddish precipitate with HgSO., the Legal reaction, an intense yellow color with aldehydes and ketones the Selivanov reaction, a dark-red color with resorcinol in 12% hydrochloric acid

Tertiary hydroxyl

Carbonyl

Ketose

8o

*O

Caramelan

Caramelen

+

+

+ +

+ +

+

+

+

+

+

-

c

n

I I

I

t

% 8 9) c)

V 3rl

1

8

0

hl

$ F I ?

0

Wavenumber

0

0

0

0

0 0

0 0

c n b

[ern-')

FIG.5.-The Ultraviolet Absorption Spectras1of Caramelan (Solid Line) and Caramelen (Broken Line). 243

244

PIOTR TOMASIK ef al.

possesses reducing properties, and forms adducts with lead and barium oxide^.^ Caramelen is formed when heating of a caramel source is continued at the same temperature until it loses 15% of its weight. Caramelin, the third compound isolated by G e l i ~is, ~formed when a sugar is heated until there is a 22%loss in weight. It is a polymeric substance.Caramelin formscompounds with barium oxide, reduces Fehling solution, and is precipitated from aqueous solutions by most metal salts. It seems to be formed by loss of 27 molecules of water from 6 sucrose units, as in 6 C12H22011 - 27 H 2 0 = 3 C24H26013. According to Sapronov;' the ultraviolet spectra of caramelan and caramelin differ from one another in the intensity of the absorption bands, but not in their positions (- 225 and 285 nm in the spectra of aqueous solutions). Thus far, there are insufficient reasons to accept the concept that caramelan, caramelen, and caramelin are chemical individuals. Pictet and Andrianoff83 also described a product of monodehydration of sucrose, namely, "isosaccharosan,"

-

C12H22011 - H2O = C12H20OIO

It is an optically active, extremely hygroscopic substance melting at 94.0 94.5". Heating of sugars to 200" results in the formation of polymeric materials, accompanied by the evolution of steam and carbon according to the equation 11 C12H22011= 7 CO2

+ 27 H2O + C125HtggOm.

At 230°, the degree of polymerization of caramel becomes still higher.334 Hydrocarbonswere suggested6]as being the final products of caramelization at elevated temperatures, but this suggestion has not been documented experimentally. Simultaneously,the same result could be achieved at constant temperature as the reaction time is extended. Carbon-13 cross-polarized magic-angle sample spinning nuclear magnetic resonance (c.p./m.a.s. n.m.r.) of plain caramels revealed that they contain 8 to 9% of carbonyl and aldehyde carbon atoms, 7 to 7.7%of carbon atoms of the ester group, 33 to 31% of heterocyclic and heteroaromatic carbon atoms, and 52.4 to 5 1.2% ofcarbon atoms of the alkyl groups bonded to other carbon atoms (including carbonyl carbon atoms) and oxygen.335 Careful analysis of the spectra, and comparison with existing data on this type of spectra of furan derivativesin free and polymerized strongly (333) A. Sabaneev and I. Antuschewich, Zh. Russ. Fiz. Khim. Obshch., 1 (1893) 23-31. (334) H. Schiff, Ber., 4 (1871) 908-909. (335) R. Ikan, Y. Rubinsztain, P. Ioselis, Z. Aizenshtat, R. Pugmire, L. L. Anderson, and W. R. Woolfenden, Org. Geochem., 9 (1986) 199-212. (336) G. E. Maciel, I. S. Chuang, and G. E. Myers, Macromolecules, 15 (1982) 121 8 - 1220.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I

245

suggest that the aromatic character of the high-molecular-weight fraction of plain caramels is due to the presence of the furan rings. It may be realized that the caramelization of sugars entails many competitive and consecutive reactions. The following consecutive steps in the formation of the products may be distinguished. (a) Degradation reactions, resulting in theformation of colorless or yellow compounds. In the final period of this step, carbon dioxide may be evolved. These reactions are sugar dehydration and transformation into furan derivatives. Degradation may be either acid- or base-catalyzed. At this stage, oxygen is not necessary in order that reaction may proceed. Furanoses yield 2-furaldehyde, whereas pyranoses give 5-(hydroxymethyl)-2-furaldehyde.337-341 The course of reactions is a function of both the temperature and the time. These reactions are accompanied by others and, particularly when the time of thermolysis is extended, by reactions in which atmospheric oxygen plays some role. The importance of those reactions increases in the second ~ t e p . ~Reversion ’.~~ of sugars may here play an essential role. The products of reversion may also undergo degradation in various directions. (b) Polymerizations and condensations leading to highly colored compounds. Condensationsare mainly due to aldol reactions. T ~ i l l aassumed t~~~ that many compounds in caramel are formed by polymerization of formaldehyde, but this point of view was abandoned by E h r l i ~ h . ~ ~ ~ Subtle methods of study have allowed definition of some of the earliest steps of the process. Thus, in the thermal degradation of sucrose in anhydrous dimethyl sulfoxide, considered as an aprotic solvent, scission of the glycosidic linkage is the initial step of the reaction.m2The hydrogen bond between the 1- and 3-hydroxyl groups of the D-fructosyl moiety and the glycosidic oxygen atom is suggested to facilitate this scission. It gives (Y-Dglucopyranose and a D-fructose carbocation.The latter internally cyclizes to form 2,6-anhydro-P-~-fructofuranose. Probably, this carbocation, as well as D-glucose, is the source of 5 4 hydroxymethyl)-2-furaldehyde. Thus, the problem immediately arises as to whether the dehydration in monosaccharides proceeds through dehydration to form carbocation, or takes place in the concerted step shown. Loss of a molecule of water by route “a” is sterically facile. Further elimination of water from cyclic forms is more difficult because cis-elimination would be required. The change of specific (337) (338) (339) (340) (341) (342) (343)

J. Marcusson, Angew. Chem., 32 (1919) 113- 115. V. A. Beckley, J. Agric. Sci., 1 1 (1921) 66-77. 0. Burian, Brennst. Chem., 6 (1925) 52-54. W. Eller, H. Meyer, and H. Sanger,Ann., 431 (1923) 162-177. E. Schmidt and M. Atterer, Ber., 60 (1927) 1671 - 1679. A. Trillat, 2. Ver. Dtsch. Zuckerind., 56 (1906) 95- 103. F. Ehrlich, 2. Ver. Dtsch. Zuckerind., 57 (1907) 15-21.

PIOTR TOMASIK et ul.

246

HO OH

Hocu -HzO

>QHO:

OH

OH

rotation that accompaniesthe heating of sugar^^^,^^ has been interpreted in terms of the Lobry de Bruyn -Alberda van Ekenstein tran~formation,~~~ through acyclic glucose by way of the 1,Zen04 into D-(+)-mannose and D-fructose (although this may be a naive simplification) and to take place rather in an alkaline solution. Subsequent molecules of water may be lost more readily from acyclic sugar molecules, as shown. The possible role of the catalyst may involve transformation of cyclic into acyclic sugars. The problem of dehydration of sugars was reviewed by Feather and Harris.346 H C

H

/ /H

0 -H

I

lCHOH13

( C HOH 1

C H o~H

C HZOH

I

--D [+) - glucose

1

1,2 -enol

-=---=.I

? N O HOCH ( CHOH13

I

C ?OH

I

or

c=0 I 1 C H OHI3

I

C+OH

C HZOH

D (+) -mannose -1 route a )

? route b )

0- f wctose

(344) J. E. Duschsky, Z . Ver. Dtsch. Zuckerind.. 61 (191 1) 581-608. (345) C. A. Lobry de Bruyn and W. Alberda van Ekenstein, R e d . Trav. Chim. Pays-Bas, 14 (1895) 150-165,203-216. (346) M. S. Feather and J. F. Hams, Adv. Curbohydr. Chem. Biochem., 28 (1973) 161-224.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I

241

Thermal gravimetnc and differential thermal analytic studies neither clarified these problems nor answered further questionsabout the mechanism of the following The rather complex thermogram of th glucose was interpreted by a Russian as entailing removal of hygroscopic water at 90 - I00 ', water of crystallization at 125 - 1 30°, primary dehydration water at 210-230", and pyrogenic water at 320-330".

C/H

C/ H

C - 0-H

c=o

p o

LCOH

I

3iH

-

p o

I

CH

-

___c

H COH

H COH

I HCOH I

I H COH I

CH20H

H -C

CyOH

3 - deoxyosulose

3 - deoxyoldos -2-ene

I/

'I"'

-OH

I H COH I

C H2OH

H

C /H

CH IHCO-H I

CH$H 3,4 -dideoxyosulos -3-ene

O

-H20

'2"\ H

5-1 hydroxymethyl) - 2-furaldehyde

OH 1, 6 - Anhydro -p-Dglucof uranosc -

(347) (348) (349) (350)

2-furaldehyde

E. Furukawa and F. Yoshimatsu, Kaseigaki Zasshi, 31 (1980) 246-251. R. K. Jain, K. Lal, and H. L. Bhatnagar, Indian J. Chem., 23A (1984) 828-833. P. Tomasik, S. Wiejak, and M. B*czkowicz, Staerke, 39 (1987) 94-97. G. Dunsbergs,G. A. Rossinskaya, G. Dobele, A. Mikelsone, R. Lukss, E. Heinsoo, and V. Efremov, in F. R. Vezhbitskii (Ed.), TermicheskiiAnaliz i Fazovye Ravnovesiya., Gos. Univ. Perm, USSR, 1983, pp. 1 1 - 16.

PIOTR TOMASIK et a1

248

Further rather well recognized steps of the caramelization are due to formation of oxaheterocycles. The 1,Zenediol of a hexose yields 5-(hydroxymethyl)-2-furaldehydein subsequent transformations, and, in a similar manner, pentoses give 2-furaldehyde. Either intramolecular cis-elimination or intermolecular condensations produce further molecules of water. Intramolecular cis-elimination could involve a radical mechanism, and intervention of atmospheric oxygen would be significant. Although 0 u r ~ * 9 ~ observationsshowed that this may be the case, the close spectral (ultraviolet and infrared) resemblance of caramelan, which is the major reaction product, suggests that the pyranose form is retained in caramelan. Thus, caramelan results rather from intermolecular polymerizations. This point of view

b

HO

OH

OH OH

OH

THERMAL DECOMPOSITION OF CARBOHYDRATES. I

249

received additional support from studies by Mora and c o w ~ r k e r s ,as ~~~,~~~ well as Tipson and coworkers,353who studied the acid reversion of a - ~ g l u cose in the presence of acid catalysts. They found that ( 1 +4’), (1 +6’), and ( 1 +3’) interglycosidic bands are formed when further water molecules are eliminated definitive work by Wolfrom and coworker^^^^-^^^ consolidated the essential chemistry involved. The schemepresented suggeststhe possibility of pyranose-ringopening, as well as mutarotation, which further create suitable conditions for water to be eliminated in a p-elimination. Mutarotation is practically complete before these reactions take place to any great extent.3572-Furaldehydeand, particularly, 5-( hydroxymethyl)-2-furaldehyde are commonly c o n ~ i d e r e d ~to~ * , ~ ~ ~ be precursors of the coloring matter of caramel. Telegdy-Kovats and O d 6 0 and JaniEek and coworkers361showed that the rate of formation of 5 4 hydroxymethyl)-2-furaldehydeis lower than that of the browning process, and corresponds to the rate of decomposition of a-~-glucose.Those authors assumed that 5-( hydroxymethyl)-2-furaldehydeis a secondary product of caramelization. Brown compounds of caramel are formed up to 240°, and they readily decompose at higher temperatures. Kunio and coworkers showed362that 5 4 hydroxymethyl)-2-furaldehydegives chiefly 5-methyl-2furaldehyde when heated at 400 All sugars, and products of their dehydration and condensation,also form re duct one^,^^^ which are precursors of adiketones, according to the general

-

O .

(351) P. T. Mora and J. W. Wood, J. Am. Chem. SOC.,80 (1958) 685-692. (352) P. T. Mora, J. W. Wood, P. Maury, and B. G. Young, J. Am. Chem. SOC.,80 (1958) 693-699. (353) H. W. Durand, M. F. Dull, and R. S. Tipson, J. Am. Chem. Soc., 80 (1958) 3691 -3697. (354) A. Thompson, M. L. Wolfrom, and E.J . Quinn, J. Am. Chem. Soc.. 75 (1953) 30033004. (355) A. Thompson, K. Anno, M. L. Wolfrom, and M. Inatome,J. Am. Chem. Soc.. 79 (1957) 1309-1311. (356) M. L. Wolfrom, A. Thompson, and R. H. Moore, Cereal Chem., 28 (1963) 1821 186. (357) A. Broido, Y. Houminer, and S. Patai, J. Chem. SOC., B, (1966) 41 1-414. (358) F. Fischer, Angew. Chem., 32(1919) 113-115. (359) J. Marcusson,Angew. Chem., 34 (1921) 437-438; 38 (1925) 339-341; 39 (1926) 898900. (360) L. Telegdy-Kovats and F. Orsi, Period. Polytech. Chem. Eng., 17 (1973) 373-385. (361) G. JaniEek, J. Pokorny, and J. Davidek, Chemiu iywnoSci, (Polish translation from Czech), WNT, Warszawa, 1978, p. 198. (362) K. Kunio, T. Doihara, H. Sakai, andN. Takahashi,Nippon Senbui Kosha Ghuo Kenkyusho Kenkyu Hokoku, 108 (1966) 361-364; Chem. Abstr., 66 (1967) 18,817~. (363) C . Enders, Biochem. 2.. 312 (1942) 349-353,

250 R

‘c=c

- 2 H+

/R

?’ I

I

H

H

-

PIOTR TOMASIK et al. R -0

’

R

>c-c

0

H

I

Ho

R-C-C

I

+

base

H‘

R

\c=c

-

-2Q

___c

‘c=c

‘ 0

0 ’.

\0-

/R

--

/R \O H

I R -C --

CH H‘

0

+

base H‘

H

I 0 )

R - C H-&>-CvR-C

I

R - C H -IC H - C H 2 - R

base H

0 Ii

R

scheme. Carbonyl groups are always capable of entering into various additions or condensations. One such is the well known aldol reaction. Either an acidic medium or the presence of acid catalysts accelerates the caramelization. As determined by Rossinskaya and coworker^,^^ the addition of 5 - 36% H,PO, to D-glucose decreases its melting point to 60” and shifts the temperature of dehydration, as well as of degradation, by 100150“. The following cases should be distinguished caramelization in the presence of (a) an inorganic acid, (b)an organic acid as the catalyst,and (c) either a-, y-, or Bhydroxycarboxylicacid as the catalyst. In the last case, such acids may form lactides or lactones, respectively,under the caramelization conditions. Our preliminary results suggested that such acids participate in the formation of secondary aromas similarly to a-amino acids. This problem will be discussed separately in the section devoted to certain food aromas in which carbohydrates are involved. In either case when organic acids are (364) G. A. Rossinskaya,G. Domburgs, and A. T. Cherbikova, in V. P. Karlivan (Ed.),Term. Anal., Tezisy Dokl. Vses. Soveshch., 7th, Vol. 2, Zinatne, Riga, 1979, pp. 85-86.

25 I

THERMAL DECOMPOSITION OF CARBOHYDRATES. I

involved as the catalyst, anions of these acids take place in various transformations. Generally, organic acids lead to caramels containing a higher content of dark components than inorganic acids do. Perhaps, organic acids accelerate the formation of ketoses and do not so effectivelycatalyze various features of reversion of sugars. Thus, contrary to other cases, the acid-catalyzed caramelization favors the route from a- or b-pyranoses and their acyclic forms to 1,2-enediolsby way of an acyclicketose, but not by any path from the acyclic form of an aldose to a 1,2-enediol.Similarly to the afore-described caramelizations, which apparently proceed without any catalyst, hexoses form 5-( hydroxymethyl)-2-furaldehyde.It gives both levulinic and formic acid, followed by reactions leading to dark-colored compounds. However, in acidic media, especially in the presence of inorganic acids, the acid-catalyzed rehydration of 5 4 hydroxymethyl)-2-furanolmay take place,

H-D-H

0

OH

+ H3C-C

II HC

0 C

II

+ H-C

H20\

/H

7

\O

- CH

H-O-H H-0 H3C-

I I

C

0

II

C-OH

G!- I

CH2

- H3C-i 0

II

H2C-

0

II

C -OH

I

CH2

252

PIOTR TOMASIK et ul.

which slows down the reaction toward high-molecular-weight,colored compounds. However, the composition of caramels obtained in the presence of acid catalysts is more complex. In the volatile fraction of caramel prepared from sucrose at 205' in the presence of citric acid, over 80 components are present, 27 of which have been isolated but only 19 fully characterized (acetone, methyl acetate, ethyl acetate, ethanol, 1-butanol, 1-pentanol, 1-hexanol, acetic, propanoic, isobutanoic, butanoic, 1-pentanoic, 1-hexanoic, and 1-heptanoic acids, 2-furaldehyde, 5-methyl-2-furaldehyde, and 2-furanol).It is highly possible that some of these components, such as ethyl acetate, acetic, and butanoic acid, may originate from decomposition of the citric Acid-catalyzed reversion presents a complicated process. D-G~UCOS gives rise to oligosaccharides,whereas D-fructose only d i m e r i z e ~ . ' ~ -Both ' ~ ~ reactions are reversible, but seem to contribute to the complexity of caramelizaS~ tion in an acidic medium. Reversion products, as well as D - ~ ~ U C O and D-fructose themselves, react with 5-(hydroxymethyl)-2-furaldehyde,and these reactions seem to be the source of the high-molecular-weight fraction of The process is sensitive to the nature of the acid by virtue of its effect upon the competing dehydration of sugars,namely, the rehydration of 54 hydroxymethyl)-2-furaldehyde.In the case of the catalyzing acid, its acidity is more important than the concentration,and the acidity more strongly affects dehydration than r e h y d r a t i ~ n ? ' ~ The .~~ composition ~ of the medium is also important because of its influence on the same competing processes. Hydroxylic solvents may evoke further reactions of 5 4 hydroxymethy1)-2furaldehyde with them, to afford acetals and ethers.217*37**372 In strongly acidic media, the formation of furaldehydes dominate^.^^.'^'.^^^ It could be confirmed by the characteristic chemical shifts in the ''C-cp/mas nmr spectra of such caramels. Signals in the region of 110 p.p.m. correspond to P-carbon atoms of the furan ring and those at 150 p.p.m. to the furan ring a-carbon atoms additionally bearing some exocyclic substituents. Metal salts [among them, chlorides of aluminum, chromium(III), and lanthanum] accelerate both the dehydration of sugarsand the rehydration of 54hydroxymethyl)-2-furaldehyde.However, these salts may hydrolyze, and

-

G. Goretti, A. Liberti, and C. Di Paolo, Ann. Chim. (Rome), 70 (1980) 277-284. H. C. Silberman, J. Org. Chem., 26 (1961) 1967- 1969. L. M. J. Verstraeten,Adv. Carbohydr. Chem., 22 (1967) 229-305. B. Krd, Acta Aliment. Polon.. 4 (1978) 287-296, 373-380. R. W. Binkley, W. W. Binkley, and B. Wickberg, Curbohydr.Res., 36 (1974) 196-200. B. F. M. Kuster and H. S. van der Baan, Curbohydr. Res.. 54 (1977) 165- 176. B. F. M. Kuster, Curbohydr. Rex, 54(1977) 177-183. B. F. M. Kuster and J. Laurens, Stuerke, 29 (1 977) 172 - 176. M. L. Wolfrom, R. D. Schues and L. F. Cavalieri, J. Am. Chem. Soc., 71 (1949) 35193523. (374) F. Led and T. Severin, 2. Lebensm. Linters. Forsch., 175 (1982) 262-265.

(365) (366) (367) (368) (369) (370) (371) (372) (373)

THERMAL DECOMPOSITION OF CARBOHYDRATES. I

253

give rise to higher acidity of the reaction m e d i ~ m . ~Undoubtedly, ' ~ , ~ ~ ~ under the reaction conditions, furaldehydes readily p ~ l y m e r i z e . ~ ~ ~ . ~ ~ ' In caramelization under alkaline conditions also, three cases should be distinguished, namely, caramelization in the presence of (a) either alkali or carbonates of alkali metals (mainly sodium carbonate), (b) ammonia or ammonium salts, and (c) amino acids, peptides, and proteins. The low-molecular-weight fraction of caustic caramel contains roughly 50 components, but the majority of them are present in only minute proportions. Comparativegas - liquid chromatographic studiesconducted by Patey and coworkers378on caustic and ammonia caramels showed that the lowmolecular-weightfraction of caustic caramel is very poor in components of higher retention time. Transformations of sugars in an alkaline medium used to be interpreted as being based on the 1,Zenediol, which decomposes into glyceraldehydeand a triose enediol. (On the behavior of D-fructose in alkali, see papers by Wolfrom and S ~ h u m a c h e rand ~~~ Shaw and c o w ~ r k e r s . ~ ) A~ 1,2-enediol is formed from sugars in the Lobry de Bruyn- Alberda van Ekenstein reaction. The alkaline decomposition of D-xylose and glucose affords aromatic compounds.381The 1,2-enediolis changed into pyruvaldehyde hydrate, and this, into DL-lactic acid. According to Flaig and S ~ h u l t z epyruvaldehyde ,~~~ can condense to 1,Cbenzoquinone.Reductonesgenerated by degradation of sugars polymerize to brown-colored caramel compounds, but they can also originate from condensation of pyruvaldehyde with either 5-( hydroxymethyl)-2-furaldehyde or 2-f~raldehyde.~~~ In caramels from alkaline processes, many compounds are, of course, identical with those from acidic and neutral processes. Among the products, formic, acetic, glycolic, lactic, 2-methylglyceric, 2,4-dihydroxybutanoic, isosaccharinic, and metasaccharinic acids were identified.383-386 M. A. Paul, J. Am. Chem. SOC.,76 (1954) 3236-3239. J. S. Chuang, G . Maciel, and G . Myers, Macromolecules, 17 (1984) 1087- 1090. A. H. Fawcet and W. Dodomba, Makromol. Chem., 183 (1982) 2799-2809. A. L. Patey, G. Shearer, M. K. Knowles, and W. H. Denner, FoodAdd. Contam.,2 (1985) 237-246. (379) M. L. Wolfrom and J. N. Schumacher, J. Am. Chem. SOC.,77 (1955) 3318-3323. (380) P. E. Shaw, J. H. Tatum, and R. E. Berry, J. Agric. Food Chem., 16 (1968) 979-982. (381) I. Forsskahl, T. Popoff, and 0.Theander, Carbohydr. Rex. 48 (1976) 13-21. (382) W. Flaig and H. Z. Schultze, Pflanz Dueng.. (1952) 58-59. (383) J. M. de Bruijn, A. P. G . Kieboom, H. van Bekkum, and P. W. van der Poel, Sugar Technol.Rev., 13 (1986) 21-52. (384) J. M. de Bruijn, A. P. G .Kieboom, and H. van Bekkum, Red. Trav. Chim. Pays-Bas, 105 (1986) 176-183. (385) J. M. de Bruijn, A. P. G .Kieboom, H. van Bekkum, andP. W. vander Poel, Int. SugarJ., 86(1984) 195-199. (386) J. M. de Bruijn, F. Touwslager, A. P. G . Kieboom, and H. van Bekkum, Staerke, 39 (1987) 49-52. (375) (376) (377) (378)

PIOTR TOMASIK et a[.

254

II C-OH

-

I

HOCH

I

II I

HCOH

I I

+

COH

HCOH

CH20H

I H C 01-1

CH2OH

I

C H20H HO\ C / H

II

COH

I C H20H

H-C

-

/OH

O N c /OH

I O‘H

I

C=O

___c

I

CHOH

I H3

CH3

DL-

Pyruvaldehyde hydrate

Lactic acid

0

I 0

II 0

0

OH

II

0 p- Benzoquinone

0

Kotelnikova and B ~ b r o v n i k gave ~ ~ ’ an interpretation of some of the processes that occur on caramelization in alkaline media. They based it on changes in the infrared spectra of some of the coloring substances. The products are (u) cyclic polyenes having conjugated carbon- carbon double bonds and (b) carbonyl compounds. The latter disappear during storage and, simultaneously, melanoidins are formed. Caramel is formed by a profound dehydration of sugars, which gives cyclic terpenoids having double bonds, (387) L. P. Kotelnikova and L. D. Bobrovnik, Cent. Azucur, 5 (1978) 1-6.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I

255

conjugated with primary and secondary alcoholicgroups. Based on the facts only some collected in two excellent reviews on the chemistry of sugars,388,389 elements of the overall process can be deduced as more or less speculative concepts. Based on experiments carried out on partly methylated sucrose,3w the alkaline degradation of that sugar could be described in more detail; it proceeds through two initial, competing, rate-determining steps involving the nucleophilicsubstitutionat C- 1 by 1'-and 3'-oxyanions which leads to 1and 3-O~-~-glucopyranosyl-~-fructose, respectively. The analogous substitution by the 6'-oxyanion is a side reaction. The products rapidly decompose, to afford lactic and D-glucometasaccharinic acids. The caramelization of sugars in the presence of ammonia, amino acids, peptides, and proteins gives caramels which differ from one another to a certain extent. The application of ammonia leads to contamination of the product with 4( 5)-methylimidazole,but this is effectively absent in caramels obtained with the participation of amino acids. Because such reactions are precursors of secondary food aromas, and the flavor of caramels is dependent on the amino acid used, the chemical structure of such caramels is also variable. Furans, imidazoles, and such other azaheterocycliccompounds as pyrrole, pyridine, and pyrazine derivativespresent in such caramel^^^*^^'-^^' are responsible for both flavor and aroma. They belong to the low-molecular-weight volatile fraction of caramel, which consists of a high number of individuals, several of which remain uncharacterized. Those characterized that do not contain nitrogen are essentially the same carbonyl compounds as found in the caramelizations already discussed. Their yield depends to a certain extent on whether the amino acid added is neutral (glycine, isoleucine, or valine), acidic (glutamine), or basic (lysine, or arginine). Because caramels are mainly prepared as coloring matter, a dark-brown melanoidin is a most essential and desirable result of the process. Reactions which lead to melanoidin are called the Maillard reaction.lgLThere is some controversy about the formal meaning of this reaction. considered (388) (389) (390) (391) (392) (393) (394) (395) (396) (397) (398)

I. J. Goldstein and T. L. Hullar, Adv. Curbohydr. Chem., 21 (1966) 431 -512. B. Capon, Chem. Rev., 69 (1969) 407-498. M. Manley-Hams and G. N. Richards, Curbohydr. Rex, 90 (1981) 27-40. H. Keysers, Gordiun, 73 (1973) 50. I. W. Dawes and R. A. Edwards, Chem. Ind. (London), (1966) 2203. H. Tsuchida, K. Morinaka, S. Fuji, M. Komoto, and S. Mizumo, Dev. FoodSci., 13 (1986) 85-94. A. L. Patey, J. R. Startin, P. M. Rowbottom, and G . Shearer, Food Add. Contum., 4 (1987) 9 - 15. V. Stanek, Z. Zuckerind. Boehm., 41 (1917) 298-306. H. Friedrich, Z. Zuckerind. Boehm., 41 (1917) 614-617. H. Stolzenberg,Ber., 49 (1916) 2021 -2023. G. P. Ellis, Adv. Curbohydr. Chem., 14 (1959) 63- 134.

256

PIOTR TOMASIK et al.

that the Maillard reaction encompasses all reactions of amino acids, peptides, and proteins with sugars, whereas McWeeny and coworkers3wlimited this name exclusively to reactions of these amino compoundswith reducing sugars. Such nonreducing sugars as sucrose are, on thermolysis, hydrolyzed into reducing glucose and D-fructose; thus, in fact, in the case of the reactions discussed herein, this distinction does not need to be introduced. On the other hand, the mechanisms of the formation of melanoidins from ammonia, ammonium salts, or amino acids resemble one another, and therefore, they will be discussed together. A great deal of work has been devoted to recognition of the structure of melanoidins. A wide variety of methods has been employed for this purpose. First, melanoidin shows a hyperfine structure in the e.s.r. spectrum and that means that stable free-radicals are present in caramel.- Amino acids and ammonia were also detected in nondialyzable melanoidin after acid hydrolysis,401,402 indicating that amides are present in melanoidin. Studiesby ZenouzM3showed that melanoidins may be reversibly decolorized by desolvation, to afford creamy-white solids. Melanoidin undergoes decomposition, but only to a certain extent, by sulfur dioxide?” The oxidation of melanoidins with hydrogen peroxide or ozone causes their degradation. The products are butanoic, glycolic, 2-hydroxybutanoic, and other acids. Nondialyzable melanoidins seem to be composed of saturated aliphatic fragment^.^^,^ However, the oxidation of melanoidins(also of those nondialyzable) with KMnO, gave, among others, pyrazinecarboxylic, 2,6pyrazinedicarboxylic,2,5-pyrazinedicarboxylic,oxamic,and oxalic acids,” and, in addition to those already Characterized (including use of oxidation with ozone and hydrogen peroxide), aliphatic tricarboxylic, benzoic, formic, acetic, malonic, and succinic acids. This means that amino acid, as well as sugar, structural units are present in melanoidins. Nondialyzable melanoidins subjected to pyrolysis between 300 and 400” produce aziridines and alkylpyrroles. These findings, as well as the results of oxidation with (399) D. J. McWeeny, M. E. Knowles, and J. F. Hearne, J. Sci. FoodAgric., 25 (1974) 735746. (400) C. H. Wu, G. F. Russel, and W. D. Powrie, Dev. Food Sci., 13 (1986) 135- 144. (40 1) M. V. Rozhkova,G .A. Chikin,V. I. Tyagunova, V. M. Rogozina, and 0.I. Lukina, Teor. Prakt. Sorbts. Protsessov, 14 (1981) 41-46. (402) S. Saito, Nippon Shokuhin Kogyo Gaikkaishi, 25 (1978) 400-401; Chem. Abstr., 92 (1980) 176,057n. (403) A. A. Zenouz, Agricultura, 25 (1977) 354-467. (404) V. Valter, Listy Cukrov., 67 (1951) 1 1 1 - 117. (405) H. Kato, S. B. Kim, and F. Hayase, Dev. FoodSci., 13 (1986) 215-223. (406) S. B. Kim and H. H. Park, Hanguk Susan Hakhoechi, 19 (1986) 36-44. (407) M. Komoto, H. Kato, and M. Fujimaki, Agric. Biol. Chem., 44 (1980) 677-678.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I

257

KMnO,, allow the conclusion that the major repeating unit of melanoidins probably does not consist of aromatic ring^.^^,^^ The infrared and luminescence spectra of melanoidins suggest that they possess double bonds.401,402 Extensive studies made by using the '3C-c.p./m.a.s. n.m.p. spectra confirmed previous findings, as well as throwing light on further structural details of melanoidins from various amino acids and D-glucose (or D-galactose). Thus, there are aliphatic fragments(chemicalshifts in the region below 50 p.p.m.), alkyl groups bonded to heteroatoms (chemical shifts in the region of 50 to 105 p.p.m.), alkene moieties being in the resonance, carbon atoms of aromatic and heteroaromatic moieties ( 105- 160 p.p.m.), and carboxyl groups of esters and amides ( 170- 190 p.p.m.), as well as carbonyl groups of ketones and aldehydes (190-220 p.p.m.). Carbon types in melanoidin samples could be approximately estimated in percentage^.'^^ (see Table VIII). Heteroaromatic fragments are mainly furan rings.335 Analysis of the I3C- and 15N-c.p./m.a.s. n.m.r. spectra of me la no id in^^'^.^^' led to the conclusion that the peaks at 70- 120 p.p.m. reflect the presence of conjugated enamines, such as -(C=O)-C=C-N -(C-0-)=C-C=N'=. Moreover, the peaks at 120- 170 p.p.m. seem to be mainly due to the -C=N+moiety. Pyridine and pyrazine types of nitrogen atoms do not exist in melanoidins. Therefore, it seems likely that pyridine and pyrazine compounds are present merely in the low-molecular-weight fractions. It should be noted that 13C-c.p./m.a.s. n.m.p. studies of melanoidins have revealed a great resemblance of this material to some humic acids, which henceforth should be considered to be products of the Maillard reaction.335Following these earlier suggestions, some model studies have been carried out on the formation and behavior of melanoidins in a clay-mineral matrix!" By comparison of the available knowledge on the structure of humic acids, melanoidins were deduced to consist of polymers of fulvic and humic acids, as well as polymeric kerogen. Montmorillonite provides the highest rate of formation of melanoidins. This fact nicely contributes to knowledge on the role of this mineral in the early stages of formation of organic life on Earth, There are several theories of the pathway of the formation of melanoidins. Because dimedone, which is an established analytical reagent for carbonyl

-

-

(408) H. Kato and H. Tsuchida, Prog. Food Nutr. Sci.,5 (1981) 147- 156. (409) J. J. Boon,J. W. de Leeuw, Y . Rubinsztain, Z. Aizenshtats, P. Ioselis, and R. Ikan, Org. Geochem., 6 (1984) 805-811. (410) L. Benzing-Purdie,J. A. Ripmeester, and C. M. Preston,J. Agric. FoodChem., 31 (1983) 913-915. (41 1) F. Hayase, S. B. Kim, and H. Kato, Agric. Biol. Chem., 50 (1986) 1951- 1957. (412) K. Taguchi and Y . Sampei, Org. Geochem., 10 (1986) 1081- 1089.

TABLEVIII Approximate Estimates (To) of Carbon Types in Melanoidin Sample-'B Origin of melanoidin' Gal Gal Gal Gal Gal Gal Gal Gal Glc Glc

+ +

Lys (9: 1) +Lys (1:9) Ile (1 :9) +Ile (1:l) +Ile (9:l) +Arg(1:1) Gly (9: 1) +Val (9: 1) Tyr (9: 1) +Tyr (1:9)

+ +

Carbonyl and aldehyde carbon atoms 5.7 (7.1)b -

-

9.9 8.0 4.8 5.8 7.1 4.5 7.0

Ester and amide carbonyl carbon atoms

Heterocyclic and heteroaromatic carbon atoms

C-Alkyl+ 0alkyl acetyl carbon atoms

9.4 (10.2)b 12.2 17.6 11.6 6.2 13.8 7.3 9.1 8.7 9.2

31.8 (37.8)b 21.4 14.4 24.8 29.2 20.7 39.3 34.4 40.4 51.2

53.1 (44.9)b 66.3 68.0 53.7 56.6 60.9 47.6 49.4 46.3 32.0

+

Gal, D-galactose; Lys, lysine; Ileu, isoleucine;Arg arginine; Gly, glycine; Val, valine; Tyr, tyrosine; and Glc, ~-glucose.* AFter hydrolysis.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I

+ - H20

!%gar + amine

-+

-

ti-, 0

259

g Ly cosy Larnine I

rearrangement

1- amino-;-

deoxy -2-hexulosc

Strecker degradation

5-(h ydroxymethyl)-

+

ornine

(ket imine)

aldimine

n i t r o g e n - c o n t a i n i n g p o l y m e r a n d copolymer

SCHEME 1.

compounds, prevents the melanoidin reaction,31spyruvaldehyde266and 5-( hydroxymethy1)-2-f~raldehyde~*~ have been assumed to be precursors of melanoidin. Perhaps the first consistent theory was proposed by Hodge,4I3 who distinguished 7 steps of reaction (see Scheme 1). They are: (I) condensation of a sugar with an amino acid, (2) the Amadon rearrangement,414 which is acid-catalyzed, as well as the Heyns reaction (the retro-Amadori rearrangement415),(3) the dehydration of sugars, (4) the decomposition of sugars, (5) the Strecker degradation416(which leads to amino carbonyl compounds, aldehydes, and carbon dioxide), (6) the aldol reaction, and (7) consecutive reactions in which simple amino compounds are formed (mainly (413) J. E. Hodge, Agric. Food Chern., 1 (1953) 928-943. (414) M. Amadon, AttiAccad. Nuzf.Lincei. [6] 2(1925) 337-342;9(1929)68-73,226-230; 13 (1931) 72-77, 195-199. (415) K. Heyns and K. H. Meinecke, Chern. Ber., 86 (1953) 1453- 1462. (416) A. Strecker,Ann., 123 (1862) 363-365.

I

H-N H

/H

\;/OH

3

I

HCOH

+H

I

HCOH

I

HCOH

I I CU20 H

HCOH

I

CH20H

-H20

HOC H

HOCH

HioH

H

HCO H

+

HOCH

L

HioH I HioHI HoiH-+ I

I""\ I 'ioH I

HCOH

HCOH

HCOH

HCOH

I I CH2OH

CHZOH

HNR

HNR

HC

rl

H-COH

I

HOCH

I -

R

-

HCOH

I I CH20H

HCOH

i N

-

HOCH

I HCOH I HCOH

Y

HCO

I

+

-.

I

HOCH HCOH HCO

I

CHZOH

+H+

I

HOCH

I HCOH

- -H*

II

COH

,HOCH 1

I HCOH

I

-

I

I HCOH I CH20H

HCOH

I

CHZOH

260

CH20H

H NR

HNR

I HC

HCOH

I

CH2OH

C H20U

HCOH

I I CH 20H

H OC-Cl-$.NHR

I

HO C l - 1 HCOH

I

HCO

I

-

0-H R-C-E-C

I

HI HOF\

0

R-C/

/c\

HO

26 1

H~N-C'

+

H ' HO

I

PIOTR TOMASIK ef a/.

262

aldimines, and, sometimes, ketimines). They do polymerize to melanoidins.55.417 Apart from that, a more-general scheme suggested by R e y n ~ l d s ~ (see '~.~'~ Scheme 2) has been suggested. In the Reynolds theory, 5 steps are distin-

Aldose +

rearrangement

glycosylamine

amino acid

t

-

ketos-7- ylamine

___c

diketos-I-ylamine or diamino sugars

i-

- n H20

products of condensation of amines with products of degradation

-

products 'of degradation

polymerization

SCHEME2.

-

Melanoidins

guished: (1) the reversible formation of glycosylamines, (2) the rearrangement of aldosylamines into ketosamines (the Amadori rearrangement) and ketosamines into aldosylamines (the Heyns rearrangement), (3) the formation of diketosaminesor diamino sugars, (4) the degradation of amino sugars with the loss of one or more molecules of water, and (5) the condensation of amines with the products formed in step 4, followed by polymerization. If, (jHZOH

R

I

COZH

OH

CH - C02H

-

OH

(417) D. T. Hurst, BFIMRA Scientific and TechnicalSurvey, No. 75, March 1972. (418) T. M. Reynolds, Adv. FoodRes., 12 (1963) 1-53. (419) T. M. Reynolds, Nonenzymic Browning Sugar-Amine Interaction, The AVI Publ. Co., Westport, 1969, p. 21 19.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I

263

for some reason, no cyclization is possible, the charge may be eliminated that leads to secondary amines (carbylamines). They form the corresponding carboxylicacid (path “a”) or Schiff base (path “b”).Aldosylamino acids may R

OH

I

2 H (HC OH14

CHZOH

H

R

H

- I

(HCOH)4

CH2OH

OH

R

I

H C = N - C H -COzH

I I

I

HC-

N-CH-R

(HCOH)~

C HZOH

CH20H

react with another molecule of the aldose, to give dialdosylamino acids; the products of decarboxylation of aldosylamino acids may also react. Amadori

HO CHZOH

PIOTR TOMASIK et al.

264

and Heyns rearrangements are presented in simplified forms. Diketosamines formed from diaidosylamines (or diketosamino acids), similarly to monoketosamines, decompose into diketosamino acids, 3-deoxyaldosuloses, and unsaturated glycosuloses. The decomposition may proceed through either a 1,2-enolor a 2,3-enol. The first route predominatesin acidic medium, and the second route is pronounced in either a neutral or a basic medium. Thus, the pathway may also depend on the amino acid chosen. A 2,4-diulose may react with an amino acid to give acetic acid and amino reductone. However, the same diketone may react to afford a 2,4,5-triulose,

+/

+/R H-CH-14H LI H‘ c= 0 HOCH IU

HC-NH

----JI

I

I

I

HCOH

I

I

CHZOH

CHZOH

HC= 0

I HCOH I

HC OH

I

CH20H

-

~

HCTOH

I-

HCOH

I

CH2 OH

HC= 0

HC=O

COH

I c=o I

I

Ifc= 0

17

___)

HC OH

HCOH

HC-H

I

H‘ C-0-H GHU

HO-(JOkH ‘ CH

HCOH

HC=N

I CH

I HCOH I

HCOH

I

CH2OH

+/R HC= N 1 H‘

-

c= 0

I

CH

II CH

I

HCOH

I

CH20H

HC=O

I

- - ‘J CH

I

CH

CH

HCOH

HCO

I

I

CH20H

II

I

CH 2 0 H

THERMAL DECOMPOSITION OF CARBOHYDRATES. I

-F H+

HzC-N,

I

R

C G

4

HOC-H

-

H 2 C - P NH / R I \R C OH

n11

H-0-C

I

HCOH

I

I

HCOH

CH3

-

I

In

c= 0

I

-

R H\

-1

I

N-CH

LOzH

H ’

I

1

COH

I

HC- N - CH - COZH

I/,

c=o

-

I CHZOH

c= 0

COH

R

I Y

HC- 0 - H

I CH20H

- CH-C02H

I

c-0

147

HC OH

HC- N

I

CH2OH

CH3

I

I 7 7

I

H ~ O H

I

c= 0

I

C-0-H

CHZOH

CH3

c-0

- CjRI

HCOH

CH2OH

HC-OH IF--+

I

c=o

I

HC OH

C H20H

c= 0

H -I‘,C O H

I

HC OH

265

c= 0

-- - +

I Y I

7

HC-N - C H

c=

0

I

c=o

I

I

CH20H

CH3

-C02H

H R I I HC- N - CH

-

I

COH

I

c=o

I

CH3

- C02H * CH3C02H

PIOTR TOMASIK er al.

266

CH3

CH3

c -0

c=o

c=o

HCOH

HCOH

I

I

I

In 0

HCOH

- -- - - - - *

I

m

c=

I

II

PI

I c= 0

COH

I

I

CHZOH

I

CHZOH

RI

HZN-CH-COIH

/‘\C-

+ HO-C

CH3

1

CH3

OH

C-OH

I

-

c= 0

C OH

H-COH

+

I

I

__-_

I

C-0

I

+CH$02H

HC=O

H N - CH - C02H

I

R

which decomposes into acetic acid and pyruvaldehyde. The decomposition ofthe 2,4,5-triketone is catalyzed by an amino acid (or amine). Reductone is formed either by condensation between C-2 and (2-6, or C- 1 and C-5. Finally, the mode of formation of heterocyclic low-molecular-weight products will be briefly discussed. Pyrroles are formed from some monosaccharides plus amines; however, pyrolysis of glycine, glycylproline, proline, serine, and threonine also affords pyrrole derivatives. The formation of imidazoles is favored when reducing sugars, and also a-dicarbonyl and a-hydroxycarbonylcompounds, react with ammonia. Moreover, a-, p-, and y-diketones condense with amines and aldehydes, to give five- or six-membered nitrogen h e t e r o c y ~ l e s .The ~ ~ ~formation - ~ ~ ~ of pyridine derivatives is, perhaps, due to reaction of ammonia with 2-furaldehyde derivatives, although pyrolysis of a-amino acids also yields such compounds. Pyrazines result from cyclization of D-glucosylamine, D-fructosylamine, and the like. However, reactions of substitution of the ring-oxygen atom in furan by the (420) L. Knorr, Ann., 236 (1886) 69- 115,290-332. (421) L. Knorr, Ber., 17 (1884) 1635-1642. (422) L. Knorr, Ber., 18 (1885) 299-31 1, 1568- 1569. (423) C. Paal, Ber., 18 (1885) 367-371,2251-2254. (424) M. H. Sprung, Chem. Rev.,26 (1940) 297-338. (425) H. Gutknecht, Ber., 12 (1879) 2290-2298; 13 (1880) 1 1 16- 1 1 19. (426) R. C. Jones, J. Am. Chem. Soc., 71 (1949) 78-81.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I

267

NH as well as in pyrones and pyrans by the ring-nitrogen atom428,429 to form pyridine derivatives, are also known. It is difficult to say whether, in the course of caramelization,alicyclic nitrogen compounds are formed prior to their aromatization, by subsequentdehydration, or whether azaheterocycles are formed by replacement of oxygen atoms by nitrogen atoms in the Yurev, and Baeyer, reactions, respectively.

9. Biological Screening of Caramels Ever since it was suspected that caramels are biologically active, the daily uptake levels for them have been calculated, and announced (for instance, in The unfavorable biological activity of caramels may be caused by some by-products formed in the course of caramelization. 4(5)-Methylimidazole is a well known, neurotoxic by-product present in ammonia and ammonia sulfite caramels. For this reason, these commonly used caramels are banned by the Food Laws of some countries, and, in others, serious limits have been put on accepted levels of that compound in caramels, as well as on the daily uptake of food finished with such caramels. Although 4(5)-methylimidazole can eventually be removed from caramel, or its content diminished by adsorption, extraction, centrifugation, or ultrafiltration, still other components of caramel may appear to be active in unfavorable ways. For instance, 2-acetyl-4(5)-( 1,2,3,4-tetrahydroxybutyl)imidazoleat a level of 2 mg/day induces lymph~penia.~~' The fact that caramels contain unpaired electrons (they contain free radicalsm) may make all caramels dangerous to use. This is now being intensively checked in our laboratory. The metabolism and mutagenic effects of caramels in general are now under intensive investigation, but the results of this research are not unequivocal; indeed, they are controversial, even when the mutagenicity is tested by the same tests. Thus, Aeschbal~eF~~ reported that there is little risk in the use of plain caramels, and even ammonia and caustic caramels are either neutral or only very slightly mutagenic. On the other hand, a group of Chinese reported that the Ames test run on Sulmonellu typhimurium TA 100, as well as on cultured human amnion FL cells, shows that plain caramel is mutagenic in respect to Salmonella and to human cells. It inhibited the synthesisof DNA but without damage to DNA (427) (428) (429) (430) (431) (432) (433)

Yu. K. Yurev, Zh. Obshch. Khim., 11 (1941) 1128- 1134. A. Baeyer, Ber., 43 (1910) 2337-2443. L. F. Cavalieri, Chem. Rev.,41 (1947) 525-584. S. Gergely, Ernaerhrung (Vienna), 4 (1980) 7-9. C. T. Convey and A. J. Paine, Biochem. SOC.Trans., 14 (1986) 1041 - 1042. H. U. Aeschbaker, Prog. Clin. Biol. Rex, 206 (1986) 133- 144. Y. Yu, X. Chen, C. Ding, Z. Cai, andQ. Li, Mutat. Rex, 139 (1984) 161-165.

268

PIOTR TOMASIK et al.

present. The test with plain caramel, carried out on the ovary cells of Chinese hamsters, showed induction of a relatively high frequency of chromosome breaks and exchanges in the treated cells. This clastogenic effect was diminished by liver microsomal preparations as well as by ferrous and femc comp o u n d ~ . "Mitotic ~ ~ gene conversion in the D7 strain of Sacharomyces cerevisiae under the influence of caramel and products of the Maillard reaction of lysine with D - ~ ~ U C Owas S ~ suppressed by yeast, or Fe3+or Cu2+i0ns.4~~ The ammonia caramels have been shown to be responsible for diminution of circulating lymphocyte counts in rats fed a vitamin-B,deficient diet.436,437 Such caramels release vitamin B, from rabbit brain, and inhibit ATP-dependent ki11ases.4~~ Tests on F 344 rats for carcinogenicity of ammonia caramels show no reason to attribute such an effect to them.439A similar conclusion was presented by another group of workers, who carried out their tests on B6C3F, mice fed with a 5% aqueous solution of ammonia caramel.440Jensen and coworkers,"1 who tested ammonia caramel for its mutagenicity against Salmonella typhimurium TA 100, TA 1535, and TA 98, found that only the TA 100 strain is sensitive towards caramel. The mutagenic activity of caramel is constant, independent of the stage in the industrial process; however, it increases on storage. In other studies on caramels from maltose, conducted by the micronuclei test on bone-marrow cells of mice, no mutagenicityand chromosomal aberrationscould be detected, but rather accidental and minimal mortality of rats fed with ammonia caramel suggestedthat ammonia-freecaramelsare preferable.442These results are the opposite of those obtained with Japanese caramels, which exhibit positive mutagenicity in the Ames chromosome test.443Different results may come from nonstandardized methods employed, as well as from the origin of the (434) H. F. Stich, W. Stich, M. P. Rosin, and W. D. Powrie, Mutat. Rex, 9 1 ( 198 I ) 129 - 136. (435) M. P. Rosin,H. F. Stich, W. D. Powrie,andC. H. Wu, Mutat. Res., 101 (1982) 189- 197. (436) I. F. Grant, A. G. Lloyd, P. Grasso, S. D. Gangolli, and K. R. Butterworth, Food Cosmet. Toxicol., 15 (1 977) 508 - 52 1. (437) J. G. Evans, K. R. Butterworth, I. F. Gaunt, and P. Grasso, Food Comet. Toxicol., 15 (1977) 523 - 53 1. (438) R. Spector and S. Huntoon, Toxicol. Appl. Pharmacol., 62 (1982) 172- 178. (439) A. Mackawa, T. Ogin, C. Matsuoka, H. Onodera, K. Funita, H. Tanigawa, Y. Hayashi, and S. Odashina, Food Chem. Toxicol., 21 (1983) 237-244. (440) A. Hagiwara, M. Shibata, Y. Kurata, K. Seki, S . Fukushima, and N. Ito, Food Chem. Toxicol., 21 (1983) 701-706. (441) N. J. Jensen, D. Willumsen, and J. Knudsen, Food Chem. Toxicol., 21 (1983) 527-530. (442) S. Dong, J. Xiong, and J. Huang, Tiaowei Fushipin Keji, (1984) 17- 19; Chem. Abstr., 101 (1984) 228,688r. (443) M. Ishidate, Jr., T. Sofuni, K. Yoshikawa, M. Hayashi, T. Noumi, M. Sawada, and A. Matsuoka, Food. Chem. Toxicol., 22 (1984) 623-636.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I

269

caramel. German workersw tested caustic and ammonia caramels against Salmonella typhimurium TA 1535,TA 98,TA 1537,and TA 100,and found a mutagenic effect to be caused by none. The unequivocal effect of ammonia sulfite caramel on beta amylase, namely, inhibition and activation dependent on pH, was interpreted in terms ofaggregation and dispersion of equally or oppositely charged m i c e l l e ~Studies . ~ ~ of the distribution of caramel after transfusion showed that the urine contained lower levels of D - ~ ~ U C O S ~ , D-fructose, and maltose than originally present in caramel, whereas the levels of hexitols were the same. The urinary output increased in exact proportion to the amount of caramel excreted in the urine.446Caramels have even been recommended as osmotic d i u r e t k u 7 The mutagenicity of melanoidins from the Maillard reaction has also been studied, with contradictory results; thus, Yoshida and Okamotou8 demonstrated, in the Ames test with Salmonella typhimurium TA 98 and TA 100, that the mutagenicity of creatinine in meat boiled at 3200'' increases after addition of D-glucose,meaning that mutagens can be formed by the reaction of creatinine and sugars on cooking meat. Australian workersu9 reported that caramels approved by the Australian and U.S. Food Laws are not mutagenic in respect to Salmonella strains TA 98 and TA 1538.According to Perkins and coworkers,"5othe mutagenicity of melanoidins depends on the amino acid used for their formation. Thus, melanoidins from D - ~ ~ U C O S ~ and either leucine or isoleucine were found not mutagenic, but those from lysine, methionine, alanine, histidine, and cy~teine$'~ as well as from glycine and thre~nine,"~' were slightly mutagenic against the Salmonella microsome system. Food caramels obtained by heating a sugar and a plant-protein hydrolyzate were found to be not mutagenic in the common Ames test, but melanoidin from D-glucose and arginine was slightly mutagenic. Other

(444) M. Scheutwinkel-Reichand W. van der Hude, Z. Lebensm. Unters.Forsch., 18 1 (1985) 455-457. (445) S. Fuji, H. Tsuchida, S. Kishihara, and M. Komoto, Dev. FoodSci., 13 (1986) 559-566. (446) Y. Wang, T. Matsuda, H. Onishi, Y. Shimizu, K. Inoue, and S. Tamura, Obihiro Chikusan Daigaku Gakyutsu Kenkyu Hokoku Dai-I-Bu, 14 (1985) 131 - 142; Chem. Abstr., 104 (1986) 108,355~. (447) T. Matsuda, Y. Wan&E. Satoh, Y. Shimizu, K. Inoue, and S. Tamura, Obihiro Chikusan Daigaku Gakyutsu Kenkyu Hokoku Dai-I-Bu. 14 (1986) 31 1-327. (448) D. Yoshida and H. Okamoto, Biochem. Biophys. Res. Commun., 96 (1980) 844-847. (449) A. M. Bonin and R. S. V. Baker, Food Technol.Aust., 32 (1980) 608-61 I . (450) E. G. Perkins, M. G. Becher, F. J. Geuthuer, and S. E. Martin, Food Carbohydr., 1981. Westport, Conn., 1982, pp. 458-481. (45 1) M. Jaegerstad, A. L. Reuterswaered, R. Olson, S. Grives, T. Nyhammar, K. Olsson, and A. Dahlquist, Food Chem., 12 (1983) 255-264.

270

PIOTR TOMASIK et a(.

claimed that 2,5-dimethylpyrazineand 54hydroxymethy1)-2furaldehyde are not mutagenic. These findings have also been contradicted. Caramel-containing food and beverages, as well as dried prunes and raisins, induce chromosome aberrations in tests with the ovary cells of Chinese ham~ters.4~~ Moreover, 54hydroxymethyl)-2-furaldehydeand some pyrrole derivatives have been reported to be and even strong$56mutagens against the same bacterial strains as previously. Some a ~ t h o r s ~ ~ have ' , ~pointed ~* to the fact that, even ifthe products ofthe Maillard reaction are not mutagenic, they can develop mutagenic activity in contact with the nitrates and nitrites frequently used as food preservatives, as they may produce N-nitrosoamines. It has been found that melanoidinsare antagonistic to mutagens and such carcinogens as aflatoxin BI and N-methyl-N'-nitro-N-nitro~oguanidine~~~ and various heterocyclic mutagens.460.461 High-molecular-weight fractions of melanoidin are the most effective in this respect.46o 111. THEPYROLYSIS OF SUGARS

Ever since the work of Cerniani9' in 1951, the pyrolysis of sugars has been studied at temperatures ranging from 200 to 1OOO". Above 200°, or 250°, deep changes occur in sugars, and a great variety of products is formed,461 such as carbon mon-, and di-oxide, hydrocarbons (among them alkanes, alkenes, and aromatic hydrocarbons), alcohols, aldehydes, ketones, and several furan derivatives. The number, and character,ofthe products depend, at least, on both the temperature and the time of reaction. Controversy exists in respect to the influence of the atmosphere in which the pyrolysis takes place. (452) N. Yamashita, H. Doihara, N. Jahan, and H. Omura, Nippon Eiyo Shokuryo Gukkaishi, 36 (1983) 202-204; Chem. Abstr., 100 (1984) 204,7981'. (453) H. U. Aeschbacker,C. Chapguis, M. Manganel, and R. Aeschbach,Prog. FoodNutr. Sci., 5 (1981) 279-293. (454) H. F. Stich, M. P. Rosin, R. H. C. San, C. H. Wu, and W. D. Powrie, BunburgRep., 7 (1981) 247-266. (455) H. Omura, N. Jahan, K. Shinohara, and H. Murakami, ACS Symp. Ser., 215 (1983) 537-563. (456) K. Shinohara, E. H. Kim, and H. Omura, Dew. FoodSci., 13 (1986) 353-362. (457) H. Kinae, M. Yamashita, M. Takahashi, H. Ooishi, I. Tomita, and H. Kanamori, Dew. FoodSci., 13 (1986) 343-352. (458) S. S. Jadhar and P. R. Kulkami, Arogyu (Munipul, India) 1 I (1985) 174- 176; Chem. Abstr., 104 (1986) 18,773e. (459) R. M.Chan,H. F.Stich,M. P.Rosin,andW.D. Powrie,CancerLett., 15(1982)27-33. (460) H. Kato, S. B. Kim, F. Hayase,andC. N. Wan,Agric.Biol. Chem.. 49( 1985) 3093-3095. (461) S. B. Kim, F. Hayase, and H. Kato, Dew. FoodSci., 13 (1986) 383-392.

TABLE IX Percentage Composition of Gases from the Pyrolysis of Some SugarsW Sucrose

Temperature ("C)

CO,

CO

200-250 250-300 300-350 350-400 400-450 450-500

16.29 65.75 51.99 21.52 11.85 6.08

23.39 31.01 44.24 50.12 38.06 36.15

C,H, 0.40 1.15 1.66 2.63 1.76 0.73

pGlucose

CH, 1.63 2.64 13.52 47.94 47.40

H,

CO,

CO

9.64

17.97 69.55 56.18 32.94 16.00 9.27

21.05 28.55 40.66 48.08 38.15 35.22

Cfi, 0.83 1.52 2.06 3.08 1.97 0.13

Lactose

CH,

H,

CO,

CO

11.91

71.47 63.61 50.04 21.26 11.56 6.66

21.46 32.86 44.83 47.86 33.84 35.08

0.71 1.55

15.91 43.95 43.55

C,H, 1.03 1.90 2.51 2.76 1.51

0.85

CH,

H,

1.30 2.71 22.63 49.75 49.30

7.56

272

PIOTR TOMASIK ef ul.

Bryce and GreenwoodM2studied the decomposition of maltose, isomaltose, and D - ~ ~ U C O at S ~ 300" in vucuo, and obtained acetaldehyde, furan, acetone, and 2-methylfuran as major products, together with acrolein, butanal, butanone, and 2,5-dimethylfuran. The authors concluded that the results are quite similar to those from pyrolysis conducted under nitrogen, oxygen, and in the air. On the other hand, after pyrolysis of 1 gram of D-glucose for 1 h at 350" in the air, Kunio and coworkers362isolated 2.7 grams of pyrolyzate. Moreover, they isolated 5 4 hydroxymethyl)-2-furaldehyde and 5-methyl-2-furaldehyde,which were not mentioned by Bryce and Greenwood.M2On treatment at 330" in the air, both of those compounds have been isolated from ~-glucose,D-fructose, sucrose, D-arabinose, and D-xylose. These carbohydrates exhibit different patterns in chromatographic analysis, and 5 4 hydroxymethyl)-2-furaldehydeis produced solely from aldopyranoses,whereas 5-methyl-2-furaldehyde is formed only from ketofura n ~ s e sPyrolysis .~~ of glucose caramel yields similar compounds.464Cerr ~ i a n ireported ~~ yields of carbon dioxide, carbon monoxide, methane, alkenes, and hydrogen evolved from D - ~ ~ U Clactose, O S ~ , and sucrose at temperatures varying from 200 to 500" (see Table IX). Fig. 6 showsthat there are two maxima of total gas evolution, located at -225 and 350", but the composition of the gas evolved varies as shown in Fig. 7. Under helium at temperatures of 250 to 800", decomposition is roughly similar;&' products of decomposition are the same, but the route of decomposition, and, of course, the intermediary compounds, may be different. This point of view was nicely illustrated by Prey and G r ~ b e r , ~ who ~ * *showed ~ some differences in the rate of pyrolysis, depending on whether it was conducted under nitrogen or hydrogen. Evidently, the total composition of the pyrolyzate is different. Pyrolyzate obtained under nitrogen at 400", as well as at lOOO",is entirely free from alcohols. Again, butanedione is formed only under nitrogen at 400", and not under hydrogen (see Table X). (Pyrolysisof D-glucose,and of sucrose, at 270- 300" appeared to be uninfluenced by the oxides of aluminum, chromium, and zinca8). P i ~ o nwho , ~ ~pyrolyzed a-~-glucoseand maltose, showed that, as the temperature of the process is raised from 300 to lOOO",the volume of gases (462) D. J. Bryce and C. T. Greenwood, Staerke, 15 (1964) 359-363. (463) Y. Kaburaki, U. Kobashi, T. Doihara, and S. Sugawara, Nippon Senbai Kosha Kenkyusho Kenkyu Hokoku, 108 (1966) 355-359; Chem. Abstr., 66 (1967) 18,818~. (464) H. Sugisawa,J. FoodSci., 31 (1966) 381-385. (465) E. Stahl and T. Herting, Chromatographiu, 7 (1974) 637-643. (466) V. Prey and H. Gruber, Staerke, 29 (1977) 96-98. (467) V. Prey and H. Gruber, Staerke. 29 (1977) 135- 138. (468) C. Sandomini, Ann. Chim. Appl., 20 (1930) 262-270. (469) M.Picon, Bull. SOC.Chim. Fr., (1953) 681-686.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I 2500

273

r

-I

E

2000

x 0

5 w

0

0 uW

1500

5

4

0

>

1000

150

XU

250

300

350

400

450

500

Temperature ( " C 1 FIG.6.-The Course of Total Gas Evolution9' from M l u c o s e (Points),Lactose (Dashes), and Sucrose (Solid Line) Heated from 200 to 500".

evolved does not change monotonically, and that decomposition of maltose proceeds the more readily (see Table XI).The pyrolysis of sugars has been studied because carbohydrates could be considered to be a source of technically important products. Moreover, these studies were intended to contribute to our knowledge of the chemistry of caramelization. These predictions have been fulfilled in the second case. Pyrolyzatesare very complex mixtures that are difficult to separate. To date, a total of 67 compounds has been identified in pyrolyzates of C Z - D - ~ ~ U C O Sthat ~ were obtained under various conditions.466They are furan and 23 of its derivatives, 22 aliphatic and alicyclic aldehydes, ketones, and diketones, 8 aliphatic and alicyclic alcohols, benzyl alcohol, and 12 aromatic hydrocarbons. Neither formaldehyde nor phenols have been trapped. Under nitrogen at 300 and 1000 64 and 13%of residue were noted, respectively,whereas, under hydrogen at 1O0O0, 9 and 4%of residue are found in the processes conducted without and with catalyst, respectively. O ,

PIOTR TOMASIK el al.

274

80

r

m

\.

I

I

\

-

J aJ

$ 9

a \'

_--I

.

aJ

w 3 0

E

3 0

1

..

// :9'

20

10

0 Temperature

(OC

1

FIG. 7.-The Variation of the Composition9' of the Total Gas Evolved from ffilucose (Points), Lactose (Dashes), and Sucrose (Solid Line) when Heated from 250 to 500".

The field of theory has met with more success. Thus, a balance of molecules of water produced on pyrolysisindicates that water is primarily formed both from glycosidic and any other hydroxyl groups. Heyns and coworker~ discovered ~~~ 1,4 :3,6-dianhydro-~-glucopyranose among the volaS ~300". It is not especially importile products of pyrolysis of ( Y - D - ~ ~ U C Oat tant whether an intramolecular process or intermolecular condensation gives the anhydro compound because, during pyrolysis, a number of polymerizations takes place; thus, particular compounds may be intercon~erted.4~' Any formation of ethereal linkages could not be observed up to at 300'. A more-vigorous pyrolysis evolves three water molecules least388*472 from one D - ~ ~ U C Omolecule. S~ It may be achieved by the transformation of (470) K. Heyns, R. Stute, and H. Paulsen, Carbohydr. Res.. 2 (1966) 132- 149. (471) Y. Houminer and S. Patai, cited in Ref. 474. (472) J. W. Liskowitz and B. Carroll, Carbohydr. Res., 5 (1967) 245-255.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I

215

TABLEX Liquid Productsa of F'yrolysis of D-Glucose under Nitrogen and under Hydrogen for 30 Minutes at 400 and 1OOO" 1OOO" Hz

400" Compound

Nz

Hz

Nz

Furan derivatives Furan 2-methyl3-methyl2,Sdimethyl2,3,5-trimethyl2-ethyl5-ethyl-2-methyl2,Sdiethyl2-vinyl2-methyl-5-vinyl 24 I-propenyl) 2,3-benzo24 hydroxymethy1)Alcohols Methanol Ethanol 1-Propano1 2-Propanol 2-Butanol Cyclopentanol Cyclohexanol I-methylBenzyl alcohol Aldehydes Ethanal Propanal 1-methylI-Butanal 1,3-Pentadien-5-a1 Ketones Propanone 1,3,5-Tnmethylbenzene I ,2,3-Trimethylbenzene 1,2,4-Trimethylbenzene

+ + + + + + + ++

+ + + + + + --

+ + + + + + + +-

+ + + + --

+

-

-

-

+ - + + + + - + + - + +- -+ + - + - + + + + ++

+ + + + + + + + + +

+ + + + -

+ + + + -

+ -

+ + + + - +

+ + + +

+ + + +

Compound

24 I-methylhydroxymethyl) 24 1-ethylhydroxymethyl) 2-form yl3-fOmIyl2-formyl-5-methyl2-acetyl5-acetyl-2-methyl2-propano yl2-methyl-5-propanoyltetrahydrotetrahydro-2-(hydroxymethyl) -Butanone 2-Pentanone 4-methyl3-Pentanone Cyclopentanone 2-Hexanone 3-Hexanone Cyclohexanone 2-methyl3-Buten-2-one 2-Pentenone Butanedione 2,5-Hexanedione Hydrocarbons eXylene m-Xylene pXylene Isopropylbenzene Ethylbenzene 1,2-Diethylbenzene 1,2-Dimethylnaphthalene

400" Hz

Nz

1OOO" N,

H,

-

+ + + + + + +-

+ + + + +-

+ + + + + + +-

+ + + + -

-

+ + + + + + + + + + -

+ + + + + + + + + + + + + + + + + -

+ + + + + + + + + + + + + + + +

-

+ + + + + + + + + + + + + + + + +-

-

-

+ + + + + + +-

+

PIOTR TOMASIK et al,

276

TABLEXI Comparison of the Course of Pyrolysis of a-D-Glurose and Maltosem Product

Temperature ("C)

a-D-Glucose

Maltose-l H,O

Gases evolved (cm') 300 400 500 600 700 800 900

loo0 At 400" CO, (%)

co (%)

Distillate (%) Residue (%) H,O (%)

5.8 13.9 6.6 6.2 6.5 8.8 7.1 4.6 33.0 26.0 30.5 66.6 7.2 20.9

14.4 21.8 14.0 12.4 13.1 16.7 11.6 6.6 32.7 25.0 31.2 43.7 13.6 37.0

D-glucose units into products containingfuran rings,4703473 which polymerize to give furfural resins. According to Prey and c0workers,4'~the following set of reactions is responsible for this stage of pyrolysis. The final step of pyrolysis indicates a total of five molecules of water eliminated from one ~ - g l u cose unit. This result further suggests simultaneous elimination of oxygencontaining compounds from furfural resins, as shown. The routes proposed are, of course, oversimplified, and in part, unlikely; on the other hand, the structures may be plausible. For instance, the authors never found formddehyde among the products of p y r o l y ~ i s . ~ Moreover, ~,~' 5-( hydroxymethyl)-2-furaldehyde at 400' gave 5-methyl-2-furalaldehyde and three other, unidentified products, volatile in a gas- liquid c h r o m a t ~ g r a p h , ~ ~ ~ which means that they are not high-molecular-weight products. Formaldehyde and acids were found among the products of y-irradiation of mono-, di-, and tri-saccharidesas the result of splitting of the ( 1 +2) bond in raffinose and sucrose, and of the ( 1 +4) bond in lactose.'@Therefore, it is more likely that there is agradual formation of furan derivativesby transformation within already condensed glucose units (see, for instance, structures al(473) K. Heyns and M. Klier, Carbohydr. Rex, 6 (1968) 436-448. (474) V. Prey, W. Lichberger, and H. Gruber, Staerke, 29 (1977) 60-65.

277

THERMAL DECOMPOSITION OF CARBOHYDRATES. I

HOCH2 0

HoH2c

H OH

OCHO

""";

I

OH

/

X -

278

PIOTR TOMASIK ef a1

ready discussed). The high reaction temperatures employed make it practically certain that radical mechanisms are involved. Indeed, stable, free radicals from ~-glucose,sucrose, and lactose have been detected by e.s.r. All of these free radicals were generated by ultraviolet irradiation, and all but those from such D - ~ ~ U C Osystems S~ as dextran and from inulin may be generated475 at 1 10 to 120".(Differencesin the thermal stability of sugars and starch have been studied by B a n t l i ~ ~Surprisingly, .~~~) free radicals from sugars appeared to be only slightly sensitiveto oxygen, and more to moisture.477 The free-radical character of pyrolysis is seen in the results presented by Prey and G r ~ b e rin, ~which ~ the compositions of pyrolyzates, and residues from the process carried out in an atmosphere of hydrogen in the absence, and presence, of catalysts, as well as under nitrogen, were compared. The decomposition of sugars in a photolytic process in the presence of ruthenium(1V) and titanium(1V) oxides and of platinum has been described. This process, conducted in an aqueous solution, eventually in the presence of alkali, yields hydrogen and carbon d i o ~ i d e . 4 ~ ~

(475) (476) (477) (478)

H. Hashiwagi and S. Enomoto, Chem. Pharm. Bull., 29 (1981) 913-917. G. Bantlin, J. Gasbeleucht., 57 (1913) 32-41, 55-61. G. V. Abagyan and A. S. Apresyan, Arm. Khim. Zh., 32 (1979) 850-859. T. Kawai and T. Sakata, Nature (London), 286 (1980) 474-476.

ADVANCES IN CARBOHYDRATE CHEMISTRY AND BIOCHEMISTRY, VOL. 47

THE THERMAL DECOMPOSITION OF CARBOHYDRATES. PART II.* THE DECOMPOSITION OF STARCH BY PIOTRTOMASIK Department of Chemistry and Physics, The Hugon KoMgtaj Academy of Agriculture, Cracow. Poland

STANISLAW WIEJAK College of Engineering, Opole, Poland

AND

MIECZYSEAW PALASI~~SKI

Department of Carbohydrate Technology, The Hugon KoHgtaj Academy of Agriculture, Cracow, Poland IV. Introduction ......................................................... V. Dextrins and Dextrinization ............................................ 1. HistoricalBackground .............................................. 2. Classification of Dextrins ............................................ 3. Physical Properties of Dextrins and British Gums. ....................... 4. Chemistry of Dextrinization, and Structure of Dextrins and British Gums. ... 5. Sources for the Manufacture of Dextrins and British Gums ................ 6. The Manufacture of Dextrins and British Gums ......................... 7. Uses of Dextrins and British Gums. ................................... 8. Biological Activity of Dextrins. ....................................... VI. The Pyrolysis of Starch ................................................

279 28 1 281 282 284 302 3 17 32 1 323 333 335

IV. INTRODUCTION Thermal treatment of dry starch usually leads exclusively to its depolymerization, unless the temperature applied exceeds 300". At elevated tem-

* Part I appears in this volume, pp. 203 -278. All numbers for sections,references, and tables in Part I1 continue the sequence established in Part I.

279

Copyright 0 1989 by Academic Ress, Inc. All rights of reproductionin any form reserved.

280

PIOTR TOMASIK ef ul.

peratures, depolymerization of starch is accompanied by an evolution of gaseous products that result from various thermal reactions. These reactions are mentioned in Part I of this Chapter.479The application of heat is only one among several ways for depolymerization of starch. Some oxidizing agents depolymerize starch,480although not Other depolymerizingagents are phosphorus p e n t a c h l ~ r i d and e ~ ~acetyl ~ bromide.483Starch is depolymerized y-radiation under acidic by ultraviolet irradiation,161,165.475.478,4~-488 conditions,’63-171,477,489-492 the effect of the radiation depending on the origin of the and by Boiling of aqueous starch suspensions (sizing) also effects some depolyrnerizatio11.4~~ Under the microscope, the formation of dextrins may be observedin grains(for instance, corn grain) subjected to hydrothermal treatment (boiling at lo0°).4% Some confusion may occur when depolymerizationunder the influence of acids and of alkalis is discussed. When either elevated temperature alone, or both an elevated temperature and only a catalytic amount of any acid or alkali is applied, depolymerization yields materials that have been loosely termed “dextrins,” whereas heating of starch in aqueous acidic media affords starch hydroly~ates.4~~ Dextrin itself is also present in this material. It may be isolated and determined by extraction with 2 :3 (v/v) ethanol-

(479) P. Tomasik, M. Palasinski, and S. Wiejak, Adv. Curbohydr. Chem. Biochem., (480) J. A. Radley, Starch and Its Derivatives, 4th edn., Chapman and Hall, London, 1968, Ch. 1 1 . (481) H. H. Volker, Dtsch. 2. Lebensm., 25 (1974) 61 -66. (482) H. N. Barham, E. S. Stickley, and M. J. Caldwell, J. Am. Chem. Soc.. 68 (1946) 10181024. (483) K. Freudenberg and K. Soff, Ber., 69 (1936) 1252- 1257. (484) L. Massol, C. R. Acud. Sci., 152 (191 1) 902-904. (485) J. Bielecki and R. Wurmser, Biochem. Z., 43 (1912) 154- 164. (486) F. Lieben, L. Lowe, and B. Bauminger, Biochem. Z., 271 (1934) 209-212. (487) P. Moeckel, Nuturwissenschufien,64 (1977) 224. (488) A. Merlin and J. P. Fouassier, Mukromol. Chem., 182 (1981) 3053-3068. (489) V. F. Oreshko, Zh. Fiz. Khim., 34 (1960) 2369. (490) L. Saint-Lebe, G. Berger, J. P. Michel, M. Huchette, and G . Fleche, Fr. Pat. 2,329,749 (1977); Chem. Abstr., 88 (1978) 52,250~ (491) J. Raffi, J. P. Michel, and L. Saint-Lebe, Staerke, 32 (1980) 262-265. (492) J. P. Michel, J. Raffi, and L. Saint-Lebe, Stuerke, 32 (1980) 295-298. (493) J. Pruzinec and 0. Hola, J. Rudiounal. Nucl. Chem., 118 (1987) 427-431. (494) S. Ono, Rev. Phys. Chem. Jpn., 14 (1940) 25-41. (495) H. Sobue, S. Moroyu, and N. Mitome, Cellul. Ind. (Tokyo), 12 (1936) 4- 12; Chem. Absfr.,30 (1936) 5420. (496) J. Fornal, Actu Aliment. Pol., 1 1 (1985) 141- 150. (497) R. V. McAllister, Adv. Curbohydr. Chem. Biochem., 36 (1979) 15-56.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I1

28 1

by high-performance liquid ~ h r o m a t o g r a p h y , 4or ~ . by ~ ~ use ~ of dynamic membrane^.^^ Potato starch has been reported5O’to give sugars when heated for 15 to 20 min at 200-250”. In this Section, only those processes are discussed which result in the formation of pyrodextrins (“dextrins”), namely, those reactions which take place under the influence of heat, often with the assistance of catalytic amounts of ions. A part of this Chapter is devoted to the extensive degradation of starch at temperatures up to 1000”. Reviews and chapters dealing with some aspects ofthese topics have been published by Horton,502Evans and W ~ r z b u r gGreenwood,504 ,~~~ Caesar,5o5 Act01-1,~~ and L ~ d v i gBeyond . ~ ~ ~ the scope of this article are the cyclodext r i n ~ ”(cyclomalto-oligosaccharides) ~~ available from starch by use of enzymic p r o c e ~ s e s . ~ ~ ~ , ~ ~ V. DEXTRINS AND DEXTRINIZATION 1. Historical Background

In the early 1800’s,Kir~hhoff,~lO Va~quelin,~’ and Bouill~n-Lagrange~~~ published almost simultaneouslythe results of their studies on starch modification. The first described a gummy product and sugar syrups from starch heated with acid, whereas the second and third obtained a soluble, gummy (498) T. F. Schweizer and S. Reimann, Z. Lebensm.-Unters. Forsch.. 174 (1982) 23-28. (499) J. Jakovljevic, 2.Boskov, and 2.Nikolov, Ind. Secera, 35 (Suppl. 1-2) (1981) 45-49. (499a) K. B. Hicks, Adv. Carbohydr. Chem. Biochem., 46 (1988) 17-72. (500) H. Hashimoto, K. Hara, N. Kuwabara, Y. Tanaka, and K. Ito, Jpn. Pat. 60,258,201 (1985); Chem. Abstr., 105 (1986) 45,153e. (501) H. Kiihl, Z. Spiritusind., 64 (1941) 15; Chem. Zentralbl., I1 (1941) 2188. (502) D. Horton, in Ref. 132, p. 42 1. (503) R. B. Evans and 0. B. Wurzburg, in R. L. Whistler and E. F. Paschal1 (Eds.), Starch Chemistry and Technology. Vol. 2, Academic Press, New York, 1967, p. 254. (504) C. T. Greenwood, Adv. Carbohydr. Chem., 22 (1967) 483-515. (505) G. V. Caesar, in Ref. 480, Ch. 9. (506) W. Acton, in J. A. Radley (Ed.), Starch Production Technology, Applied Science Pub lishers, Ltd., London, 1976. (507) L. Ludvig, Elelmez, Zpar., 29 (1975) 103- 109; Chem. Abstr.. 83 (1975) 162,079~. (507a) R. J. Clarke, J. H. Coates, and S. F. Lincoln, Adv. Carbohydr. Chem. Biochem., 46 (1988) 205-249. (508) Ref. 480. Ch. 10. (509) J. Szejtli, Cyclodextrins and Their Inclusion Complexes, Akademiai Kiado, Budapest, 1982. ( 5 10) G. S. C. Kirchhoff, Mem. Acad. Imp. Sci., Petersburg, 4 (1813) 27. (5 1 1) L. N. Vauquelin, Bull. Pharm. (Paris), 3 ( I 8 1 1) 49 - 58. (512) E. J. B. Bouillon-Lagrange, Bull. Pharm. (Paris), 3 (181 1) 395-398.

282

PIOTR TOMASIK et al.

product from the roasting of starch. However, the first note on a gum from the torrefaction (dry heating) of starch was datedSI31804. The common name “dextrin” was givenSI4in 1833 to the gum that resulted from treatment of starch with sulfuric acid. This name is attributable to the direction of the optical rotation of the p r o d u ~ tThe . ~ products ~ ~ ~ ~of~ ~ starch degradation by purely thermal processes are called “British gums,” as they were accidentally discovered after a fire in Dublin, Ireland, destroyed a building having potato starch stored within it.s16The truth of this story is sometimes questioned, and it may be apocryphal, as there is no account of this fire in the Dublin newspapers of that period.517 Within about half a century, industrial processes for the manufacture of dextrins were d e v e l ~ p e d . ~ ~ , ~Further years have brought several new ideas, as well as improvements in all of the technology of the manufacture, analysis, and utilization of dextrins. *69s18319

2. Classification of Dextrins

Dextnns are mixtures of various products of depolymenzation of the starting material and recombination. Their complexitycreates several problems in any classification of dextrins based on their chemical character. Therefore, their classification is best based on such physical properties as color, solubility in cold water, viscosity, opacity of a cooked dispersion on storage, and so on. An alternative classification is based on the method of preparation. Neither is unequivocal, and they have undergone some modifications. Thus, GraefeSZ0 differentiated between four groups of dextrins, as follows. (A) Pyrodextrins, prepared without any chemicals, such as (a)British gums (from maize starch) and (b) leiogome (from potato starch). (B) Pyrodextrins prepared with the aid of acids, namely, (a)white dextrins and (b) yellow dextrins. (C) Pyrodextrins prepared with the aid of alkali. ( D ) Pyrodextrins prepared with the aid of oxidizing agents. About fifteen years later, Greenwoodsw classified only three groups of dextrins: (A) white dextrins prepared from starch in the presence of an acid (513) (514) (515) (5 16)

(517) (518) (519) (520)

J. L. Roard, Ann. Chim. Phys., [ I ] 50 (1804) 220-224. J. B. Biot and J. F. Persoz, Ann. Chirn. Phys., [2] 52 (1833) 72-90. M. P. Petit,C.R.Acad. Sci., 114(1892)76-78. R. P. Walton, A Comprehensive Survey of Starch Chemistry, Vol. 1, Chemical Catalog Co., New York, 1928, p. 159. T. S. Wheeler, Chem. Ind. (London), (1959) 1014. Ref. 480, Ch. 1. C. Lintner, J. Prakt. Chem., 34 (1886) 378-394. G. Graefe, Staerke, 3 (1951) 3-9.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I1

283

catalyst for a relatively short time (3- 8 h) in the temperature range between 79 and 120°, (B) yellow, or canary, dextrins, prepared from starch in the presence of an acid catalyst in 6 to 18 h at 150to 220°, and (C) British gums prepared from starch without any catalyst for 10 to 20 h at 130 to 220”. On the other hand, A ~ t o accepted n ~ ~ the definition that British gums are products of heating of starch either alone or with a basic catalyst. Additionally, many technicians understand British gums to be the products prepared with catalytic amounts of acids. In this situation, the classification of dextrins in accord with some selected physical properties should be given priority. It parallels, to a certain extent, the foregoing classifications as, for instance, the solubility of white dextrins ranges from 1 to 95%, and their water uptake is described by a ratio of dextrin to water of 1 : 1 to 1 :5. Yellow dextrinsare water-soluble to the extent of95 to loo%, and their water uptake is 1 : 1.5 to 1 :0.75, and even less. British gums resemble white dextrins in their solubility, but simultaneously they form more-viscous solutions. The interest focused on the viscosity of dextrin solutions may lead to further differentiation between thick- and thin-boiling dextrins. Some useful classification of these products can probably be achieved by using early observationsof Komm and later extended by many workers to starch of various origins.502*523-525 They observed that iodine produces different colors with various dextrins, and that the color developed depends on the solubility of a given dextrin (see Table XII). This might be a convenient way of classification of dextrins. Simultaneously,there is no simple relationship between color and such properties as alkali-lability,content of reducing sugars, reducing value, and viscosity. However, if such a classification is accepted, statement of the origin of a dextrin appears to be essential.526 Evans and Caldwells2’introduced an index called the “Blue staining residue” (B.s.r.) which is related to the intensity of the color developed by a dextrin with iodine. B.s.r. values decrease together with the thixotropy ofgels from dextrins. It should be noted that fatty acids present in the dextrinized starch favor the formation of dextrins having a high B.s.r., that is, dextrins of good thixotropic properties. For studies on a related subject, see an article by Tsuji and M o c h i z ~ k i . ~ ~ ~ (521) (522) (523) (524) (525) (526) (527) (528)

E. Komm and U. Martin, Vorrutspjege Lebensrnittelforsch.,2 (1939) 635-649. E. Komm and U. Martin, VorrutspflegeLebensrnittelforsch.,2 (1939) 650-661. R. J. Dimler, M. A. Davis, andG. E. Hilbert,J. Am. Chem. Soc., 68 (1946) 1377- 1381. Y . Ueno, M. Izumi, and S. Kato, Staerke, 28 (1976) 77-83. K. Kainuma, T. Furukawa, and S. Suzuki, Denpun Kuguku, 20 (1973) 1-8. M. ceh, Stropnik, V. DoleEek, and S. Leskowar, Staerke, 34 (1982) 85-88. R. B. Evans and C. G. Caldwell, Staerke, 15 (1963) 448-454. S. Tsuji and Y. Mochizuki, Nippon Shohukin Kogyo Gakkaishi. 34 (1987) 513-519; Chern.Abstr., 107 (1987) 196,633d.

c.

284

PIOTR TOMASIK et al. TABLEXI1 Iodine Affinitysz4 of Dextrins over the Range 350 to 700 nm Heat treatment Heating Duration temp. ("C) (h) 120

150

170

190

2 10

230

2 4 6 8 2 4 6 8 2 3 4 5 6 8 2 3 4 5 6 8 2 4 8 1 2

It. (nm)

Color'

Absorbance

604.1 593.9 605.9 594.7 6 15. I 595.4 583.2 580.4 593.5 580.2 582.0 573.8 577.0 567.7 565.8 567.6 562.9 538.9 549.9 593.7 537.6 508.4 513.7 516.8 503.0

greenish-blue blue blue -green blue bluish-green bluish-green blue blue blue light violet blue violet violet violet violet violet purple purple purple purple purple purple-red purple purple purple-red

1.6075 1.6082 1.6335 1.6571 1.7033 1.7160 1.5964 1.6347 1.7143 1.5985 I .5933 1.5741 1.6693 1.5616 1.5798 1.5132 I .455 I 1.3159 1.3263 1.2683 1.1982 0.7459 0.6863 0.973 1 0.7882

The following relationship between the solubility (in %) of dextrin and its color with iodine was reported by Komm and Martin"': 100, yelloworange; 99.9, orange; 99.1, red-orange; 98.5, red-brown; 94.2, wine-&, 92.8, red-violet; 89.0, violet; 85.6, violet; 82.7, blue-violet; and 63. I , blue.

The lability to alkali529-531 seems to be another reasonableindex useful in the classification of dextrins.

3. Physical Properties of Dextrins and British Gums The physical properties of pyrolyzed starch obviously vary, dependingon the conditions applied in the process. The appearanceof white dextrins and (529) T. C. Taylor and G. M. Salzmann, J. Am. Chem. SOC.,55 (1933) 264-275. (530) T. C. Taylor, H. H. Flechter, and M. H. Adams, Ind. Eng. Chem., Anal. Ed., 7 (1935) 321-324. (531) T. J. Schoch and C. C. Jenzen, Ind. Eng. Chem., Anal. Ed., 12 (1940) 531 -532.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I1

285

TABLEXI11 Yield, Form, and Color of Dextrins Prepared under Various Conditions5*‘ Heat treatment Temperature (“C) ~

Time (h)

Maximal yieldo Color

Form

white white whitish white whitish-yellow whitish yellowish-white yellowish-white whitish-yellow yellowish-brown yellowish-brown dark yellowish-brown reddish-dark brown dark-brown

granule granule granule granule granule granule granule granule granule granule fragile fragile nonglazed nonglazed

(%a)

~

120 150 170

2-8 2-6 8 2-6 8

I90

210 230

3 4-8 2 4 6-8 1

2 5 8

66.75 (6) 66.70 (4) 63.31 66.40 (6) 67.02 65.30 66.19 (6) 64.85 63.05 64.78 (8) 67.75 66.35 62.80 57.41 ~

~

~

The time providing the highest yield is given in parentheses.

of slightly hydrolyzed starch is practically the same as that of the starting material, and only recognition of other physical properties can give any indication that the product differs from unprocessed starch. Deeper transformationsof starch (yellow dextrins, and British gums) result first of all in a change of color from creamy through canary yellow to “coffee with milk” and light brown. The determination of the properties of dextrins based on their color can be only roughly predicted. Ueno and coworkers524showed that changes in the appearance of British gums from corn starch depend on the time and temperature of roasting (see Table XIII). The chemical nature of these colored products seems to be closely related to the colored compounds of Commercial dextrins and British gums are more-or-less water-soluble, and form colloidal solutions, usually yellow or brown, that are more colored than the solid matter. After evaporation of these solutions, a sticky, elastic residue is obtained. Dextrins are practically odorless, except those originating from potato starch. The unpleasant odor of the latter products can be masked by addition of Bergamot (532) L. S. Solomina and A. R. Sapronov,Izv. Vyssh. Uchebn.Zaved., Pishch. Tekhnol.,(1979) 27-28. (533) 0. Lange, Bluchers Auskunftbuchfur die Chemische Industrie. 13th edn., Vol. 1, Verlag yon Walter de Gruyter Co., Berlin, 1926, p. 325.

PIOTR TOMASIK et al.

286

Dextrins used to be characterized by the following properties: solubility, alkali-lability, reducing sugars, reducing value (femcyanide number), viscosity, stability to retrogradation, iodine affinity, and beta amylolysis. Specific rotation, pH, and molecular weight, as well as the number of terminal groups, are also sometimes determined. Several authors have tried to generalize the change of some of those properties in relation to duration and temperature of dextrinization;however, their conclusionswere not alwaysin agreement with one another. Dextrinization is conducted mainly to promote aqueous flow and aqueous solubility. Solubility- time relationships, reported by many a ~ t h o r s , ~ ~ are , ~usually ~ ~ not ~ ~comparable, ~ ~ , ~ as ~ they * ~ do~ not ~ ~ ~ ~ always refer the results to a particular source. Thus, according to GrzeSkowiak,537in the case of experiments conducted at 160", the solubility of dextrins increases as a function of time, as shown in Fig. 8, depending on (a) the type of starch, (b) the size of the starch granules, and (c) the content of amylose and amylopectin. The course of dextrinization of amylopectin and amylose themselves is also slightly different, although, after 6 h, the solubility of the dextrins from them reaches 100%(see Fig. 9). It had been shown538that there is a relationship between the size of the granules, the molecular weight, and the degradation products. Several authors528*539-542 reported the results of exhaustive studies on the behavior of sorghum, wheat, maize, and potato starch in 2 A4 aqueous sodium hydroxide solution at 120"under pressure. They indicated that the size of the original starch granules is of minor importance. The amylose to amylopectin ratio may be important, as well as the structure of a coat of amylopectin on the amylose. The role of the size of the granules was s t ~ d i e dby~observation ~ ~ , ~ of ~ the ~ ~ effect ~ ~ of the first two properties upon the course of dextrinization in the presence of an acid catalyst. In this process, the size of the granules is not too important, as is also the amylose to amylopectin ratio (see Figs. 9b and c). (534) G. V. Caesar, in R. W. Kerr(Ed.), ChemistryandIndustryofStarch, 2ndedn., Academic Press, New York, 1950, p. 345. (535) G. V. Caesar and M. L. Cushing, Ind. Eng. Chem., 31 (1939) 921-924. (536) H. C. Snvastava, R.S. Parmar, and G. B. Dave, Staerke, 22 (1970) 49-54. (537) M. GrzeSkowiak, Pr. Nauk. Uniw. A. Mickiewicza, Poznari, Ser. Chemia, 28 (1978) 1 - 104. (538) W. Biltz, Z . Phys. Chem., 83 (1913) 683-707; Ber., 46 (1913) 1533. (539) M. Ceh, C?. Stropnik, V. DoleEek, and S. Leskowar, Staerke, 33 (1981) 45-49. (540) M. Ceh, C?. Stropnik, and S. Leskowar, Staerke, 30 (1978) 151, 264-268. (541) M. Ceh and Stropnik, Staerke, 27 (1975) 72,254-257. (542) E. Pertot and M. Blinc, Staerke, 24 (1972) 260-263. (543) N. B. Badenhuizen and J. R.Katz, Z . Phys. Chem., 182 (1938) 73-90. (544) A. Sroczynski and J. Skalski, Acta Aliment. Pol., 1 (1975) 39-64.

c.

287

THERMAL DECOMPOSITION OF CARBOHYDRATES. I1

a

t

100.

9080.

Wheat

70GO50-

30

t-

___c

1

2

5

3

6

b

--L-

--r90

70 8o

C

i

Maize starch

26% amylose ,,

0

1

2

3

4

starch my lose

5

6

t (h)

FIG.8.-Degree of Dextrinization at 160"of Starch, Depending on (a) Origin of Starch, (b) Size of Granules, and (c) Content of Amylose and A m y l o p e ~ t i n . ~ ~ ~

PIOTR TOMASIK et al.

288

I

60.

a

/

Amylopectin

/my

lose

50 LO 3020 10~

-

5

01

6

-1

b

100 . 90.

80. 70 . 60 .

Native starch

1

/

and dried

0 1 2 3 4 5 6 FIG.9.-Rate of Dextrinization at 160”of(a) Amylose and Amylopectin, and (b) Starch and

Pretreated

Some pretreatments of starch, such as suspending in water to form a gel, followed by drying, also influence the course, but not the result, of dextrinization (see Fig. 9b). Similarly, the course of dextrinization is different in the case of so-called “hydrogen starch,” which is monostarch hydrogenphosphate (amylophosphoric a ~ i d ) .In~the ~ case ~ , ~ of ~farina yellow-dextrins, a solubility of close to 100% is reached within the first hour of the process.5o6 The solubility of dextrins also depends on the temperature applied. Accord(545) M. Kujawski and M. Patasinski, Rocz. Technol. Chem. Zywn.,22 (1972) 79-80. (546) M. Patasinski, T. Fortuna, A. Nowotna, and M. Warchd, Acta Aliment. Pol., 7 (1981) 127- 136.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I1

289

TABLEXIV Results of Dextrinization of (1) Potato, (2) Maize, and (3) Rice Starchs52 Extinction (To)

Solubility (%)

Temperature

("C) 100 110 120 I30 140

150 160 I70 I80 I90 200 a

Reducing sugars"

1

2

3

1

2

3

1

2

0 0 0 13.3 16.6 20 23.3 40 66.6 85 95.3

0 0 0 3.3 8.3 11.6 13.3 23.3 43.3 58.3 66.3

0 0 0 0 3.3 10 13.3 25 51.7 60 90

0.22 0.26 0.19 1.52 5.78 40.85 83.68 84.47 57.80 23.87 2.92

0.35 0.39 0.63 1.14 4.94 22.47 75.75 83.27 82.50 29.34 6.80

0.23 0.23 0.49 0.98 2.03 13.07 52.75 72.50 77.31 23.19 3.01

0.80

1.90 0.90 0.80 0.95 3.30 5.35 15.00 24.50 34.00 25.90 14.50

0.45

0.30 1.15

3.10 9.00 19.10 33.90 33.50 31.20 8.90

3 0.60 0.45

0.55 1.20 1.50 4.30 15.30 26.50 39.60 21.40 7.80

In miligrams of copper precipitated from determination conducted on I g of sample.

ing to Srivastava and coworkers,536who camed out their experiments on a corn starch, soluble dextrins are already sparingly formed at 125". Other starches, as well as amylose and amylopectin, behave similarly, as shown by several a ~ t h o r ~ and , ~ ~~ o~n f, i ~r m~ e~d ~-by~~~ chromatographic ~, ~~~ * elution analysis and the Richter m e t h ~ d . ~On ~ the ~ , other ~ ~ ' hand, formation of the first soluble dextrins from potato starch was observed537after heating for 1 h at 100". The solubility of starch increases monotonically with the heating time for temperaturesof 150 to 175 At temperatures between 200 and 2 13",maximum solubility is reached, followed by a decrease interpreted as the result of secondary reactions of retrogradation of oligosaccharides formed in the first step of d e ~ t r i n i z a t i o n . ~Excellent ~ ~ . ~ ~ ~ evidence for the solubilitytemperature relationships for three different starches was given by Cerniani552(see Table XIV). Simultaneously, it may be observed that an essential role is played by the moisture present in dextrinized starch. A low moisture content allows full solubilityto be reached much faster than in the case of moister starches (see O.

(547) M. Ulmann, Kolloid Z., 130 (1953) 31 -39. (548) M. Ulmann and J. Seidemann, Adhesion. 5 (1961) 519-523. (549) E. Dworschala, Elelmiszenizsgalati Kozl.. 13 ( 1 967) 17-25. (550) M. Richter and H.-H. Stroh, Staerke, 14 (1962) 415; 18 (1966) 115-122, 176-180. (55 1) M. Richter, S. Augustat, and F. Schierbaum,Ausgewahlte Methoden der Stdrkechemie. VEB Fachbuch Verlag, Leipzig, 1967, Ch. 3.2.2. (552) A. Cerniani, Ann. Chim. (Rome), 41 (1951) 293-308.

PIOTR TOMASIK et al.

290

1500 c - A - - A I~~OC-X-X2aJoc -0-o2 1 3 O C -0- O 187”C(Z%Moisture) 2 1 3 O C (2%Moisture)-

A-A0-0

-

40

20

0

2

r,

6

Heating times in hours

FIG.10.-Effect of Temperatureand Time of Roasting on the Solubility of Starch.S36[Key: triangles, 150”;crosses, 175”;circles, 200”;and squares, 2 1 3 O . I

Fig. 10). The role of moisture in degradation was well d o c ~ m e n t e din~ ~ ~ . ~ ~ ~ the case of “hydrogen starch,” where a high content of moisture favors depolymerization by hydronium-ion catalysis. In the case of starch, it was considered536that excessive moisture causes less-selective hydrolysis of the glycosidic bonds, to give lower oligosaccharides. The latter polymerize to insoluble matter on prolonged heating. Again, the role of water may differ, depending on the origin of the starch, due to differing construction of the starch granules and, possibly, various modes of bonding of water (see earlier). Cernianis5*showed a different picture of water loss from three different starches (see Table XIV). B r i r n h a P found that gradual increase in the temperature of dextrinization up to 200’ during 3 h, followed by rnaintaining of this temperature for an additional 1 h, was the best method for achiev(553) M. Palasinski, Zesz. Nauk. WSR, Krakbw, Ser. Rozprawy. 7 (1968) 1-93. (554) F. Schierbaum and M. Palasinski, Staerke, 21 (1969) 87-91. (555) B. Brimhall, Znd.Eng. Chem., 36 (1944) 72-75.

THERMAL DECOMPOSITION OF CARBOHYDRATES. 11

'

I 110

120

130

1LO

150

100

Temperature

170

29 1

b -

180

190 200

(OC

FIG. 1 1 .-Solubility of Dextrins, Depending on Temperature of Roasting.szz

ing fully soluble dextrins. The decomposition point of soluble starch was seti5at 214". The solubility- temperature relationship for 1 h of heating of starch, with nitric acid as the catalyst,521 is presented in Fig. 1 1. The solubilityofdextrins formed can be influenced by contact with the atmosphere. Vacuum, as well as air flow, favor the presence of soluble matter in dextrinized starch (see Fig. 12).537However, the effect observed may be attributable to continuous removal of moisture. On the other hand, the results of the present aut h o r ~showed ~ ~ ~that - the ~ ~course ~ of dextrinization under nitrogen, carbon dioxide, and in air yields different dextrins of various solubilityand stability after a given period of dextrinization. Finally, if a catalyzing acid is applied, the course of dextrinization and changes in solubility depends on the acid applied, namely, on its pK, value. Acids of higher pK, acceleratethe solubility (see Fig. 13), but they also yield (556) M. Palasihski, P. Tomasik, and S. Wiejak, Staerke, 38 (1986) 221 -224. (557) M. Palasinski, P. Tomasik, and S. Wiejak, Pol. Pat. P-246 327 (1984). (558) P. Tomasik, M. Bgczkowicz, and S. Wiejak, Staerke, 38 (1986) 410-413.

PIOTR TOMASIK ef al.

292

15

j

in

10

air stream

I

05-

}

/

.

In air

I

in air

i

i

in mcuurn

i

0

1

2

3

4

5 Time ( h ;

FIG. 12.-Dextrinization of Potato Starch, Depending on Time and Atmosphere of Roasting.537

THERMAL DECOMPOSITION OF CARBOHYDRATES. I1

293

t I

FIG. 13.-Solubility of Dextrins from Potato Starch, Depending on Roasting Time at 165" and pH 2.9. [Starchwas acidifiedwith ( 1) hydrochloricacid, (2) nitric acid, and (3) phosphoric acid.]562

-

darker d e ~ t r i n s . ~ Also, ~ ~ ,increase ~ ~ ' ~of~ the ~ ~concentration - ~ ~ ~ of the catalyzing acid parallels the increase of the solubility of the dextrins prepared.s20,537 The effect of the nature of the catalyzingacid is due to structural changes in the dextrins being formed, as documented by Sroczydski and S k a l ~ k i . ~They ~ ' , ~determined, ~~ as a function of the roasting time and pH, the average molecular weight, the number ofterminal groups, the number of glucose units, and the number of branches in the dextrins from potato starch prepared in the air at 165". Relevant results are given in Table XV. Obviously, the solubility of dextrinized material is also influenced in a different manner by other catalysts, such as halogens and alkalis. The alkali-labilit~~~ is an index that gives valuable information on the structure of pyrodextrins, mainly on the content of reducing-terminal groups. It varies in a nonuniform manner, depending on the conditions already listed. Fig. 14 shows the change of alkali-lability of yellow farina dextrin and both white and yellow dextrins from corn starch as a function of the roasting time. The alkali-lability reaches a maximum on passing from white to yellow dextrin, and decreases on going to more-extensively dextrinized material.s35Fig. 15 presents the variation of the alkali-labilitywith the temperature of the process. This index was reported537to increase with increase in the concentration of the catalyzing acid. (559) 0. Wolff, Staerke, 2 (1950) 138-202. (560) A. Sroczynski and J. Skalski, Zesz. Nauk. Politech. Lbdz., Chem. Spozyw., 24 (1974) 45-58. (561) A. Sroczynski and J. Skalski, Rocz. Technol. Chem. Zywn.,24 (1974) 261 -267. (562) J. Skalski and A. Sroczydski, Zesz. Probl. Post. Nauk Roln., 159 (1974) 21 1-218.

PIOTR TOMASIK er al.

294

TABLE XV Changes in Properties of Dextrins Acidified with (1) Hydrochloric, (2) Nitric, and (3) Phosphoric Acid, and D e ~ t r i n i z e d at ~ ~165" '

Time of dextrinization (min)

pH of dextrin

1

2

30 60 80 90

2.9 2.9 2.9

3.0 2.9

100

2.8 2.8 2.8

2.9 2.8 3.0

2.8 2.7

3.1 2.9

120 140 150 160 180 210 220 240

Medi molecuk weight (MJ

3

1

2

3.0 2.8

8620 8070 7050

7535 7010

2.9 2.9

6720 6005 7010

9490 9434

8570 9480

34,645 26,880

4 5 7

14,185

8875 I 1,775

2.8

5 6

3 8 9

16,525

1 7 9 13

2

3

8 9 12

7

10

12 II

7

II

9 8

7

11

17

13 13

23

15

19 15

23 21

12 8

1

21

5

17

3 1

14 0

4 9

3 3 2

3 3 3

3 4

2 3

23

4 4

15

3 5

4

17 7

8 6 3

2

17

15

8

1

15

9

9

9

9790

2.9

2

8115

2.7 2.7 2.3

1

23.705 7800 8870 5290

Number of segments

Number of end groups

3

Number of D-glucosyl units in one segment

13

Reducing sugars (ferricyanidenumber) of yellow farina dextrin and those formed in the acid-catalyzed dextrinizationof potato starch vary similarly to the alkali lability index, namely, it passes through a maximum,563 to decrease very slowly as the time passe^.^^'.^^^ The ferricyanide number increases almost monotonically for British gums from corn starch, and the slope (see Fig. 16) depends on the temperature,536as well as on the catalyzingacid (see Fig. 17). The reducing sugar- temperature characteristics are different for various starches (see Table XIV). The high reducing activity ofdextrin is due to low-molecular oligosaccharide constituents rather than to monosaccharide~.~~~ Almost all authors discussing the topic agree with one another that the viscosity of dextrins decreases quite suddenly in the first hour of dextrinization, to achieve an almost flat curve on further roasting. This is the case for British gums, as well as for acid-catalyzed dextrinization.A slight increase of viscosity is noted in the case of materials roasted for over 3 h at temperatures above 200".This may be due to secondary polymerization of degradation products. On the other hand, the viscosity of yellow dextrins from corn (563) H. Riiggeberg, Staerke, 4 (1952) 78-83. (564) H. Isaka, M. Kushiya, R. Hasegawa, and T. Komuro, Eisei Shikensho Hokoku, 103 (1985) 177- 180; Chem. Abstr., 104 (1986) 184,978~.

8

! i i

I

GO 80 40

100

I

I

:

i / i I

1

1

:

-..

I

I

:

1

I

I

1

I

0

1

2

3

4

Time ( h ) FIG.14.- Variation of Solubility, Alkali-lability, Reducing Sugars, and Viscosity, as Function of Time, for Yellow-Farina Dextrin.m

20

-

1 7 5 T -xX2oooc -0-o21 30c -0- 0187°C(20/~Masiure)-A -A2130C(2%~oisture1-00-

PIOTR TOMASIK et a/.

296

1

150OC- A--A175oc - x-x200oc- 0 - 0-

213oc- 0-0 187OC (2% Moisture) - A -A 213°C(T/~Moisture)- 0 - 0 -

-

50 I

al

n

0

2

L

6

8

Heating time in hours FIG.16.- Effect of Temperature and Time of Roasting on the Reducing Value (Femcyanide Number) of Starch.sN(See Fig. 10 for notation.)

1

2 3 t (h) 1 FIG. 17.-Changes of Reducibility of Dextrins from Potato Starch Depending on Time of Roasting at 165"and pH 2.9. [Starch acidified with (1) hydrochloricacid, (2) nitric acid, and (3) phosphoric

-

THERMAL DECOMPOSITION OF CARBOHYDRATES. I1

291

$1

B 2m 'I!

>

I

I

' c

2

3 t (h) FIG. 18.-Change of Viscosity of Dextrins from Potato Starch Depending on Time of Roasting at 165 and pH 2.9. [Starch acidified with (1) hydrochloricacid, (2) nitric acid, and (3) phosphoric 1

-

starch increases significantly within the first hour, and subsequently drops.534The nature of the catalyzing acid also has its effect on the viscosity ofthe resultingde~trins,5~-~~~ as is shown in Fig. 18. Nitric and hydrochloric acid have about the same effect, whereas phosphoric acid exerts a milder effect. Solutions of dextrins increase in their v i s ~ o s i t y . ~ ~ ' The , ~ * ~time ~ ~viscosity increase relationship possesses various patterns, depending on the type of dextrin and the concentration of its solution. The iodine affinity is due to the formation of colored complexes with amylose. The color of this complex depends on both the concentration of the iodine in the solution and the kind of starch. Amylose binds 20% (v/v) of iodine, to develop a blue color, whereas amylopectin binds only 0.5 to 1% (v/v) of iodine to give a red-violet color. Starch which does not contain any amylose gives a red color with i ~ d i n e . ~Thus, ~ * ~evaluation ~' of the degree of dextrinization based on the observations of Komm and Martin521needs standardizationofthe method and approach,because ofthe variabilityin the origin of the starch. Table XVI shows that methods in use to date are not equivalent to one another. The color-development characteristics are again dependent on the origin of the

-

(564a) M. Samec, Kolloid Beih., 4 (1912) 132- 142. (565) M. Samec and F. HoelTt, Kolloid Beih., 5 (1913) 147-210. (566) S. Winkler, Stuerke, 14 (1962) 168-175. (567) S. Winkler, Stuerke, 15 (1963) 102- 105.

298

PIOTR TOMASIK et al. TABLEXVI Affinity of Various Modified Starches for Iodine568 Affinity for iodine (%) Method Starch

Maize, waxy Light-yellow, dextrin of 97% solubility Carboxymethylated(d.s. 0.5) White dextrin of 64% solubility Acetylated, medium-substituted Soluble Maize, oxidized with hypochlorite Potato, unmodified Maize, soluble Pea

Calcium chloride

0.1 0.2 0.6 0.9

1.2 2.7 4.2 5.0 5.4

13.8b

Standard 0 0 0 0 4.7"

1.5 3.5 4.1 5.0 15.0

a The high value is due to deacylation in alkaline medium. The low value is due to limited solubility of the sample in calcium chloride solution.

-

Beta amylase is an exoenzyme that degrades amylose chains from the nonreducing end, to liberate maltose. The official notation of this enzyme is (1 4)-a-~-glucanmaltohydrolase, EC 3.2.1.2. The changes as a function of time in the beta-amylolytic index (a percentage of the conversion into maltose) for corn starch is shown in Fig, 19. A small increase in the case of dextrinization at lower temperatureswas interpreted as the result of cleavage of hydrogen bonds, which makes some regions of hydrolyzed material more easily accessible to enzymes.536 Some changesin the physical properties of roasted starch undergo organoleptic control (such as taste549and appearance),552and others can be observed by the variation in, and loss of the powder X-ray diffraction pattern of, the original granules.568The first change in this pattern is manifested by a loss of sharpness at 140- 1SO" in the case of potato, tapioca, and wheat s t a r ~ h . ~The ~ ~ change , ~ ' ~ observed corresponds to a rapid increase in the cold-water solubility of starch. At 2 10- 220", starch becomes amorphous, and the X-ray diffraction pattern disappears. These changes are due to loss of water, as confirmed by thermogravimetric analysis (TGA) of starch. This (568) C. C. Kesler and E. T. Hjermsted, Methods Carbohydr. Chem., 4 (1964) 304-306. (569) J. R.Katz, Red. Trav. Chim. Pays-Bas, 53 (1934) 555-560. (570) J. R. Katz and A. W. Weidinger, Z . Phys. Chem., Abt. A, 184 (1939) 100- 122.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I1

150' 1750 2000 213 0 187 O 213O

0

2

299

-A--A-x-x-0- U-

00-

12°/~rnoisture)-A(2%moisture ) - 0

L

6

A-

-0 -

8

Heating time in hours FIG.19.-Effect of Temperature and Time of Roasting on the Beta-Amylolysis Value of (See Fig. 10 for notation.)

technique has been applied for studying the thermal behavior of starches of various origins. The results, collected in Table XVII, are not comparable,as, sometimes,the origin of the starch remains unknown; on the other hand, the conditions of the measurement are not analogous to one another, and this The origin of can seriously influence the results of the the starch plays, however, the most essential r 0 1 e . ~According ~ ~ * ~ ~to~ studies574on native maize starch, an endothermicprocess in the range up to 150" (571) D. Costa and C. Costa, Chem. Znd. (Milan), 33 (1951) 71-76. (572) A. T. Perkins and H. L. Mitchell, Trans. Kunsus Acud. Sci.. 60 (1957) 437-440. (573) H. Morita, Anal. Chem., 28 (1956) 64-67; 29 (1957) 1095-1097. (574) A. I. Zhukhman and V. A. Kovalenko, In. Vyssh. Uchebn. Zuved., Sukh. Prom., (1983) 47 - 50. (575) P. L. Waters, Coke Gus, 20 (1958) 341-343. (576) M. Bgczkowicz, P. Tomasik, and S. Wiejak, Staerke, 38 (1986) 339-341. (577) M.C.P.Varma,J. Appl. Chem., 8(1958) 117-121. (578) G. Chesters and S. 0. Thompson, Science. 133 (1961) 275-276. (579) S. Yamada, N. Nagashima, and A. Kawabata, Nippon Kusei Gukhishi, 38 (1987) 647650; Chem. Absfr., 107 (1987) 234,999n.

PIOTR TOMASIK ef ul.

300

TABLE XVII Results of Differential Thermal Analysis of Starches of Various Origins Starch

Endotherins"('C)

Exotherins" ("C)

References

? ?

175(s), 280(s) 210(s), 320(ms) 110(m), 260(vs), 295 (vs) 247(w), 36 l(w) 130(m), 280-31O(s) 80(vs), 220-250(s), 270(m), 325(s) 125(m), 275-305(s) 144(m), 288(m) 100(m), 258(s), 269(vs), 283(vs) 249(w), 265(w), 289(w) 1 15(s) 155(m), 260(ms) 140(s), 285(m) 315(w), 340(w), 373(w), 382(w) 247(w)

325(ms) 510(m) 365(s), 480(ms) 330(w), 380(w), 396(w) 330-370(ms), 470(ms), 525(m)

57 1 572 573 558 573

4 10- 500(m)

574 573 575

Rice Maize

Potato

Potato (benzene-dried) ? Pea Wheat Rye

600(s) 460- 525(m)

322(w)

576 558 577 573 575 558 558

(m) = medium; (ms) = medium small; (s) = small; and (vs) = very small.

reflects the loss of capillary and absorbed water. Between 200 and 220", chemically bonded water is evolved. A small peak at about 250" was interpreted as exothermic combustion of lipids complexed by the starch. Further peaks are caused by decomposition of starch, with evolution of carbon-containing, volatile products. This process begins rather suddenlyat 270", and is over at 310". Finally, intwosteps, at 310-330" and 330-500", thecombustion of starch is complete. The picture of thermolysis of acid-hydrolyzed maize starch is very similar. More-essential differences are visible in the region between 350 and 450 Amylopectin maize starch decomposes thermally in a slightly different way which can be distinguished at the region of higher temperature, that is, from 350". The acid hydrolysis of such starch makes it less stable to heat; moreover, the differential thermal analysis (DTA) curve shows the presence of a more chemically differentiated mixture. As shown by the present a ~ t h o r s , ~atmospheric ~ ~ , ~ ' ~ oxygen does not affect DTA, DTG, and TG curves to such an extent that oxidation can be accepted as an important process in the range up to 300". The dextrinization of starch does not change the points of thermal effects, that is, an endothermic effect, with its maximum at -80" (loss of humidity), a very small endothermic effect between 200 and 222", which ceases at the stage of fully soluble O.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I1

30 1

dextrin, and an exothermic peak at 222 -234” without loss ofweight. Intensive loss of weight begins from 234”. The degree of dextrinization influences the magnitude of the thermal effect. The thermal effects linearly decrease as the degree of dextrinization (the solubility of dextrins) increases.580 , of the products of dextrinization is randomly The specific rotation, [a], estimated. This index is not of any special interest. As reported for a maize [a]i5 (and its variation)is nonuniformly dependent on the temperature of dextrinization. Afier a given time of heating, it increases together with temperatureofdextrinization(fromabout+ 160”at 120°Cto+ 198”at as a 230°C). Only dextrinization at 210°C causes a drastic decrease of [a], function of the time of dextrinization; in other cases, the variation of this index is small. The decrease of [a], is interpreted as the result of the formation of new p-Dbonds by D-glucosyl residues. Schulz and JacobiS8’opened up a new chapter in the study of the optical properties of oligo- and poly-saccharides by optical rotatory dispersion (0.r.d.) and circular dichroism (c.d.) measurements for amylose and amylopectin. Structural changes of starch and “hydrogen starch” granules on heating have been d i s c ~ s s e d . ~ ”These , ~ ~ changes ~ , ~ ~ ~facilitate the next stage of dextrinization, which is chemical in character. Some of the properties of dextrins may be easily and effectively modified. One of them is the solubility of dextrins. The hydrolysis of dextrins in the presence of a reversed micelle has been camed o ~ t as a~part~of model ~ studies on the depolymerization of polysaccharides, and of poly- as well as oligo-saccharidewaste, for the manufacture of ethanol and livestock feed. It was found that dextrin may be readily hydrolyzed in benzene by means of 4-dodecylbenzenesulfonic acid. The rate of hydrolysis is 285 times that in aqueous solution. The reaction rate of hydrolysis in such a system depends strongly on the degree of polymerization (d.p.) of dextrin in the d.p. range up to 15. In the range from 15 to > 130, the rate is almost constant. The solubilizingbehavior of the micelle is dependent on its hydrophile -lipophile balance (dielectric constant of the solvent). Relationships between the dielectric constants of solvents and either the maximum amount of solubilized water, the rate constant of hydrolysis (which is pseudo-first-order), or the critical micelle concentration586are presented in Fig. 20.

-

(580) V. A. Kovalenko, A. 1. Zhukhman, 1. I. Kuznetsova, and N. N. Tregubov, Izv. Vyssh. Uchebn. Zaved.. Sakh. Prom., (1985) 47-49. (581) R. C. Schulz and E. Jacobi, Rev. Roum. Chim., 25 (1980) 1059- 1068. (582) C. T. Greenwood, Adv. Curbohydr. Chem., 1 1 (1956) 335-393. (583) K. Arai and Y. %wars, Bull. Chem. Soc. Jpn.. 55 (1982) 836-841. (584) K. Arai and Y. Ogiwara, J. Appl. Polym. Sci., 28 (1983) 3309-3312. (585) K. Arai and Y. Ogiwara, J. Appl. Polym. Sci., 29 (1984) 4399-4401. (586) K. Arai and Y. Ogiwara, Bull. Chem. SOC. Jpn., 51 (1978) 182-184.

~

~

~

302

PIOTR TOMASIK ef al.

2

i.0

3 4

5

G

3 4

5

G

I

1

I

2.2

2.L

2.6

Dielectric constant

’

‘ , Critical

micelle concentration,mo\x~-I Rate constant k, x 10-3min-1 Maximum amount of solubilized water, molx L-’

FIG.20.-Critical Micelle Concentration (at Room Temperature), Rate Constant (at 60°), and Maximum Amount of SolubilizingWater (at Room Temperature) as a Function of Dielectric Constant of Medium for the Process of Hydrolysis of Dextrin by Means of Dodecylbenzenesulfonicacid.585( 1, in hexane; 2, in cyclohexane; 3, in CCI,; 4, in benzene; 5, in toluene; and 6, in exylene.)

The same a ~ t h o r s prepared ~ ~ ~ - copoly(viny1 ~ ~ ~ alcohol -styrenesulfonic acid) resins which catalyze the hydrolysis of carbohydrates, among them dextrin and sucrose. Also, cation-exchanger membranes (from radiochemical grafting of styrene followed by reaction with chlorosulfonic acid) were patented for the hydrolysis of dextrin to ~ - g l u c o s eThe . ~ ~results discussed may be useful for introducing subtle modifications into dextrins already prepared. A reviewS9lon modified (converted) starches appeared in 1987. 4. Chemistry of Dextrinization, and Structure of Dextrins and British Gums The first chemical changes in heated starch begin when the water present in starch is forced to be evolved. As shown by many authors (see articles by (587) K. Arai, Y. Ogiwara, and C. Kuwabara, J. Appl. Polym. Sci., 25 (1980) 2935-2941. (588) K. Arai, Y. Og~wara,and C. Kuwabara, J. Polym. Sci., Polym. Chem. Ed., 19 (1981) 1885-1889. (589) K. Arai, Y. Ogiwara, and C. Kuwabara, J. Appl. Pofym. Sci.,27 (1982) 1601 - 1605. (590) Agency of Industrial Sciences and Technology, Jpn. Pat. 82,119,805 (1982); Chem. Abstr., 97 (1982) 217,601j. (59 1) 0.B. Wurzburg, in 0.B. Wurzburg (Ed.), ModiJiedStarches: Properties and Uses, CRC, Boca Raton, FL, 1987, pp. 17-40.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I1

303

Schierbaum and references cited therein592-595) and by Lechert and cow o r k e r ~two , ~ ~types ~ of binding of water should be distinguished. The capillary, condensed water seems to be one ofthem. Water molecules reside in the helix, and form a bridge between the hydroxyl groups at C-2 and C-6.597 Lechert and S c h ~ i eassumed r ~ ~ ~ that this point of view is not of any general validity; moreover, the hydroxyl group at C-3 may also be involved. Such a status of water rationalizes the findings of various authors576*593-595*5w-m1 that water present in the starch matrix protects it against changes that are probably mainly due to retrogradation. Its elimination by drying causes some structural changes, and the swelling temperature of such a starch decreases. This is, however, insufficient to convert starch into a water-soluble form. The complete dehydration of starch can be achieved up to 140160 , and deep decomposition begins at - 260 . The mechanism of action of water on starch on heatingmlahas been examined by nuclear magnetic resonance spectroscopy.598By observation of the splitting of the deuteron resonance in D,O-deuterated starch, an anisotropic motion could be deduced. It indicated that, in starch containing <20% of water, the starch structure is hydrolytically attacked above loo", but that, at a water content >4 1%, irreversiblechanges occur below 90". The mobility of water in different starches has also been studied; a reviewm2was published in 1981, and this was followed by two article^.^^,^ A separateproblem is presented by starch that contains fatty acids (lipids); for instance, maize, rice, and wheat starch. The naturally occurring lipids form inclusion complexes with amylose that exhibit texture and morphology different from those of native starch granules. These differences are reflected by the behavior of starch at a relatively low temperature; for instance, gelatinization at 120". The process involved is a high-temperature retrogradation, with participation of the proton dissociated from the complexed fatty acid residue. A Clz acid complexed in the helical structure of O

O

-

(592) (593) (594) (595) (596)

F. Schierbaum, Staerke, 12 (1960) 237-242. F. Schierbaum and K. Taufel, Staerke, 14 (1962) 233-238. F. Schierbaum and K. Taufel, Staerke, 14 (1962) 274-275. F. Schierbaum, K. TLufel, and M. Ulmann, Staerke, 14 (1962) 161- 167. H. Lechert, W. Maiwald,R. Koethe, and W. D. Baler, J. FoodProcess. Preserv., 3 (1979) 275 -299. (597) J. A. Blackwell, J. A. Sarko, and R. H. Marchessault,J. Mol. Biol.. 42 (1969) 379-383. (598) H. Lechert and I. Schwier, Stuerke, 34 (1982) 6- 1 1 . (599) F. Schierbaum, Staerke, 12 (1960) 257-264. (600) F. Schierbaum and K. TLufel, Staerke, 14 (1962) 41 1-415. (601) F. Schierbaum and K. Taufel, Staerke, 15 (1963) 52-56. (601a) 0.Theander and D. A. Nelson, Adv. Carbohydr. Chem. Biochem., 46 (1988) 273-326. (602) H. T. Lechert, in L. B. Rockland and G. F. Stewart (Eds.), Water Activity, Influences on Food Quality. Academic Press,New York, 1981, pp. 223-245. (603) H. Lechert and I. Schwier, Staerke, 34 (1982) 6- 11. (604) D. S. Smith, Diss. Abstr., Znt. B, 43 (1982) 1029- 1030.

PIOTR TOMASIK ef al.

304

amylose dissociatesat 85 but the octadecanoic acid ( c t g ) complex does not dissociate below 120 . For this reason, differences in the behavior of lipidcontaining starch at elevated temperature may be attributed to the level of lipids and their c o m p o ~ i t i o n . ~ ~ * ~ Further chemical changes that occur during dextrinization should be discussed separately for the case of formation of British gum, and for variously catalyzed thermal processes. Thus, when starch (1) is heated in the presence of moisture, random hydrolytic scission at the (1 46), and more readily at the ( 1 4), linkages has been proposed to occur in the branched chains of amylopectin, to give more-linear structure^,^^^^^ leading to the observed decrease in viscosity. D - G ~ U C O oxocarbonium S~~ ions (2) were postulated as intermediates, but free radicals (3) formed by homolytic fission at the (1 44)-glycosidic bond are also feasible. O,

O

-

VH C H20H

CYOH

q u

,

OH

0

/

0

I

0

I

2

1

3

OH

Support for a free-radical mechanism for dextrinization is afforded by comparison of the processes conducted in air with those carried out under a neutral a t m ~ s p h e r eand ~ ~under ~ , ~ ~vacuum.6o8 ~ The absence of oxygen accelerates dextrinization, and thus, the overall process yielding dextrins is not merely oxidation; furthermore, free radicals are trapped by molecular oxygen, a biradical. Pyrolysis of a-D-glucose at 300"to 1,4 :3,6-dianhydro-P-~(605) (606) (607) (608)

T. Davies, D. C. Miller, and A. A. Procter, Staerke, 32 (1980) 149- 158. M. Kugimiya, J. W. Donovan, and R. Y. Wong, Sfaerke, 32 (1980) 265-270. R. W. Kerr and F. C. Cleveland, Staerke, 5 (1953) 261 -266. I. E. Puddington, Can. J. Res., Sect. B, 26 (1948)415-431.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I1

305

TABLE XVIII Some Properties of British Gums556 Atmosphere Air

fiOpertY

Solubility (%) Viscosity (Pas) of 40% aq. solution Viscosity (Pas) after 24 h Degree of thickening Reducing sugars (%) Average chain-length PH Extinction X lo--’ at 420 nm

Nitrogen

Carbon dioxide

90.0 0.173

98.8 0.126

90.7 0.134

98.7 0.108

90.6 0.243

99.0 0.138

0.459

0.136

0.351

0.120

0.153

0.148

165.0 1.87 53.5 5.11 85

7.9 1.70 58.8 4.72 140

164.0 2.26 44.2 5.25 85

11.1 2.24 44.6 5.00 120

53 1.O 1.58 63.3 5.28 97

7.4 1.65 60.6 5.00 112

glucopyranose,609pyrolysis of starch-like material in an inert and pyrolysis of starch in the air, in carbon dioxide, and in an inert gasSS6-ss8 showed that there is a considerablelikelihood that free radicals are involved. The results shown in Table XVIII indicate that dextrinization in an oxygenfree atmosphere leads to essentially different dextrins. Moreover, the probability of the free-radical character of dextrinization is not surprising,in view of chemical resemblance to caramelization, in which the existence of free radicals was proved.400,47s.477 Special properties of such free radicals may be anticipated. On the other hand, their properties should be similar to those present in caramel. Indeed, the kinetics of changes of free radicals prepared from starch by radiolysis depends on the water content, but is only slightly dependent on atmospheric oxygen. The characters of the radicals are independent of the origin of the radiolyzed starch. The distribution of free radicals between amorphous and crystalline zones of starch is The effect of oxygen is the result of formation of peroxides, which transform the starch into various products. The primary radicals resulting from the cleavage of the C - H, C- OH, and CO - H bonds decomposeby dehydration in a- and /I-eliminations and cleavage of the C-C and C - 0 bonds in the D-glucosyl residues and the D-glucosidic bonds.612 (609) F. H. Holmes and C. J. G. Shaw, J. Appl. Chem., 11 (1961) 210-216. (610) L. J. Carlson, U.S. Pat. 3,235,541 (1966); Chem. Abstr., 64 (1966) 16,1221: (61 1) J. J. Raffi and J.-P. L. Agnel, J. Phys. Chem., 87 (1983) 2369-2373. (61 la) H. Frahm, Ber., 74 (1941) 622-635. (612) M. N. Sultankhodzhaeva, K. A. Korotchenko, M. A. Yakubova, V. A. Sharpatyi, and P. K. Khabibullaev, Proc. Tihany Symp. Radiat. Chem., 1982, 5 (1983) 1079- 1084; Chem. Abstr., 99 (1983) 101,569s.

306

PIOTR TOMASIK et al.

The varied retrogradative ability of dextrins is striking. Even former opinions on the mechanism of dextrinization to afford true British gums did not neglect the first step of hydrolysis, and assumed either intermolecular, (1 6) bond-formation between two chains, or intramolecular 1,6dehydration, to form 1 , 6 - a n h y d r o - ~ - ~ - g l u c o s e .At ~ ~290 ~ ~-~326 ~ ~ ~ in ~~~~~’~-~ vucuo, the latter is formed from potato starch, amylose, and amylopectin, as well as combined starch-D-glucose material through cleavage of the (1 4)-a-~-glucosidicbonds.249,6L6,617 The thermal stability of these compounds increases in the order: amylose < starch < amylopectin. Wolfrom and coworkers618 educed evidence that pyrodextrins contain D - ~ ~ U C O Sunits ~ 1 6), and B-D-( 1 2) linkages. The bonded through Q-D-( 1 6), FD-( units which are chains are terminated by 1,6-anhydro-P-~-glucopyranose 1 44)-glucosyl linkages. Based on the exotherbound preferably by a-~-( mal effect of hydrolysis of amylose, conversion of the a-~-( 1 4) linkages ~ ~ , ~ ~and ~ into the more-stable CY-D-(1 6) linkages was ~ l a i m e d . 6Bedford Gardiner6” reported that pyrolysis of amylose gives 1,4:3,64anhydro-~glucose and 1,4 :3,6dianhydro-~-mannose. Further reactions leading to British gums are as follows. (a)The hydroxyl groups at C-2, C-3, or C-6 Of D-glucoseunits are attacked either by oxocarbonium ions (2) or by free radicals (3), to produce (1 2), (1 3), and (1 6) glycosidic bonds (transglycosylation). This leads to branching and cross-linking.Because of stereochemicaland, possibly, thermodynamicconsiderations, formation of (1 6)-glycosidic bonds is favored over that of other types of bonds. 1,6-Anhydro-/%~-glucopyranose is formed at this stage,621-626and it readily polymerizes. As shown by Wolfrom and

-

-

-

-

-

-

-

--+

-

- -

(612a) A. Pictet and J. Sarasin, Helv. Chim. Acta, 1 (1918) 78-96. (613) J. Metzger, 2.Anal. Chem., 308 (1981) 29-30. (614) A. K. Ganguly, J. Chem. SOC.,Chem. Commun..(1979) 148-149. (615) G. Zemplh and A. Gerecs, Ber., 64 (1931) 1545- 1554. (6 16) D. J. Bryce and C. T. Greenwood, Stuerke, 17 ( 1965) 275 -278. (617) I. A. Wolff, D. W. Olds, and G. E. Hilbert, Adv. Carbohydr. Chem., 7 (1952) 37-40. (618) A. Thompson and M. L. Wolfrom, J. Am. Chem. Soc., 80 (1958) 6618-6620. , (1951) (619) M. L. Wolfrom, E. N. Lassettre, and A. N. O’Neill, J. Am. Chem. Soc..73 595-599. (620) G. R. Bedford and D. Gardiner, J. Chem. SOC.,Chem. Commun., (1965) 287-288. (621) A. Pictet, Helv. Chim. Acta, 1 (1918) 226-230. (622) H. Pringsheim and K. Schmalz, Ber., 55 (1922) 3001 -3007. (623) J. C. Irvine and J. W. H. Oldham, J. Chem. Soc., 127 (1925) 2903-2922. (624) D. S. Carvalho,J. W. Prins, andC. Schuerch, J. Am. Chem. Soc.,81 (1959) 4054-4058. (625) J. D. Geerdes, B. Lewis,and P. Smith, J. Am. Chem. Soc.. 79 (1957) 4209-4215. (626) G. M. Christensen and F. Smith, J. Am. Chem. SOC.,79 (1957) 4492-4495.

307

THERMAL DECOMPOSITION OF CARBOHYDRATES. I1

c ~ w o r k e r s ,in ~ ~this ~ -process, ~ ~ ~ any formation of ( 1

0

-

6) linkages is not

1 1

OH

7 (627) M. L. Wolfrom, A. Thompson, and R. B. Ward, J. Am. Chem. Soc., 81 (1959) 46234625. (628) M. L. Wolfrom, A. Thompson, and R. B. Ward, 2nd. Eng. Chem., 53 (1961) 217-218. (629) A. Thompson, M. L. Wolfrom, and E. J. Quinn, J. Am. Chem. Soc., 75 (1953) 30033004. (630) M. L. Wolfrom, A. Thompson, R. B. Ward, D. Horton, and R. H. Moore, J. Org. Chem., 26 (1961) 4617-4620.

PIOTR TOMASIK et al.

308

favored. The compounds formed in such polymerization are mainly 4-0-aD-, 4-0-p-~-,2-0-a-~-,and 2-O-P-~-ghcopyranosyl-(1,6-anhydro-P-~were glucopyranose). No 3-0-~-glucopyranosylanhydrod~saccharides found. (b) In the next stage, either of two reactions must be taken into account; they are 7 7 +-5, according to Wolfrom and coworkers,628or 6 3 5, according to Srivastava and and recombination628reactions.

+

+

6+7

-

j

CYOH

I '

I

0

-

OH

I

I

8

Transglycosylation seems to be the most typical and predominant reaction among the group of reactions constituting the dextrinization process. The evidence for it, as well as for the mechanism just presented, was provided by the results of methylation analysis631of corn British gum (see Table XIX). Incidentally, the formation of ether-type bonds was que~tioned.~' At temperatures up to 250", small proportions of various volatile products are evolved, among them, carbon mon- and di-oxide, acetaldehyde and other lower aldehydes, ketones, and some methylfurans, as well as water. The yield of these compounds increases with the t e m p e r a t ~ r e . ~This ~.~~~,~~~ phenomenon will be discussed in Section VI, together with the mechanism of their formation. It was suggestedthat processes yieldingthese compounds are responsible for the small increase in the viscosity of dextrins and British gums after prolonged heating.634 A different chemistry must be involved in the formation of pyrodextrins, that is, the dextrinization of starch in the presence of catalysts. The methylation analysis of various p y r o d e ~ t r i n s ~indicated ~ ~ - ~ ~this ~ .well ~ ~ (see ~ Tables (631) S . K. Fischer, Diss.Abstr., 19 (1959) 2450. (632) C. T. Greenwood, J. H. Knox, and E. A. Milne, Chem. Ind. (London), (1961) 18781879. (633) D. J. Bryce and C. T. Greenwood, Staerke, 15 (1963) 166- 170. (634) D. J. Bryce and C. T. Greenwood, Staerke, 15 (1963) 285-290. (635) G. M. Christensen, Diss. Abstr., 19 (1959) 2450-2451.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I1

309

TABLEXIX Methylation Analysis of a Corn British Cum631 @Methyl-wglucose

2,3,4,6-tetra2,3,6-tri2,4,6-tri2,3di2,6-di2-mono3-mOnO6-mono-

%

6.5 83.8 0.3 4.5 4.1

]

0.5

XX and XXI); it revealed differences between the results for pyrodextrins obtained in the presence of an acid catalyst and for British gums. If the problem of roasting of starch in the presence of a catalyst is to be discussed in detail, dextrinization under the influenceof acids, alkalis, and oxidants must be considered separately. Acid catalysts certainly favor hydrolysis, at least in the first stage of dextrinization. The (1 6) bonds are disintegrated, and shorter chains are This corformed from (1 4) intermolecularlybonded D-glucosyl responds to the formation of white dextrins (8).However, it was noted637that

--

TABLEXX Methylation Analysis"of P y r o d e ~ t r i n s ~ ~ ~ . ~ ~ 6 ~ ~ ~ ~ Pyrodextrin from

0-Methyl-D-glucose (%)

2,3,4,6-tetra2,3,4-tri2,3,6-tri2,4,6-tri2,3-di2,6di3,6-di2-mono3-mono6-mono-

~~

Corn amylose

Amylopectin

Starch

12.7

16.4

72.9 5.2 5.3 3.6

65.5 2.1 8.4

16.5 2.6 57.3 1.2 6.3 10.0 3.2

0.2 0.2

1.2 0.6 0.3

5.5

1.5

0.8 0.5

Methylation conducted in the presence of an acid catalyst.

(636) R. W. Ken, Pup. Trade J., I 15 (1942) 30- 34. (637) M. A. Swanson and C. F. Con, J. Biol. Chew.. 172 (1948) 797-804.

PIOTR TOMASIK el al.

310

CHPH

-

OH

-

CH20H

OH

OH

OH

OH

-

(1 6)-a-glucosidic bonds are more stable to hydrolysis than (1 4)-aglucosidic bonds. Thus, ( 1 4) linkagesare converted into (1 ---* 6) bonds in such short-chain polymers. As shown by Wolfrom and c0workers,6~*,~~~ after acid hydrolysis of amylose and amylopectin, maltotrioses and isomaltose could be isolated. Moreover, in the degradation products of amylopectin, 6 - ~ - c Y - D - g l U C O p y ~ n O S y l - D - ~ U C O s(panose) e was found. GraefeW cited structures of yellow dextrins having (1 + 6)4- and (1 6)-a-~-glucosidic bonds, as well as (1 ---* 4) and (6 6) bonds, but at least the last suggestion has not been proved. Srivastava and Krishnamurthyal presented details of the dextrinization of tamarind-kernel polysaccharide. They compared its structure with that of the dextrin prepared from it. This comparison showed

-

-

TABLE XXI Generalization of Results of Methylation Analysis of Pyrodextrin and British Gum625*626.63i~635 From

&Methyl-D-glucose (%)

TetraTriDiMono-

F'yrodextrin

British gum

16.5 61.1 19.5 2.8

6.9 84.1 8.6 0.5

(638) A. Thompson and M. L. Wolfrom, J. Am. Chem. SOC.,73 (1951) 5849-5850. (639) M. L. Wolfrom, J. T. Tyree, T. T. Galkowski, and A. N. ONeill, J. Am. Chem. SOC.,73 (1951) 4927-4929. (640) G. Graefe, Staerke, 2 (1950) 27-31. (641) H. C. Shrivastava and T. N. Krkhnamurthy, Sfaerke, 24 (1972) 369-372,405-412.

u0

PIOTR TOMASIK et al.

312

1300

I

I

1100

93

I

,

700 650

Wavelength (cm FIG.2 1 .-Infrared Spectra, from 1300 to 650 cm-', of D e x t r i n ~[A, . ~ Native ~~ starch; B, corn starch heated for 2 h at 170";C, corn starch heated for 2 hat 190";D, corn starch heated for 2 h at 210";and E, corn starch heated for 2 h at 23OO.l

- -

dextrinization at 153" to comprise a complex of processes of partial debranching of less-hindered side-chains, exchange of the sugar units, and (1 4) to (1 6) transglycosylationin the fundamental chain, as shown in structure 2, representing one monomeric unit of such a dextrin. Structural changes in the course of dextrinization find only a minor response5" in the infrared-spectral characteristics of dextrins (see Fig. 2 1). Excellent articles have been published on the infrared absorption (including the far4.r. r e g i ~ n )and ~ ~Raman-scattering -~~ spectra of mono-, oligo-, and poly-saccharides. Detailed band-assignments for these spectra commend attention on these spectroscopic techniques as useful tools in recognizing (642) A. Gatat, Acta Biochim. Pol., 27 (1980) 135- 142. (643) T. W. Bamtt, Spectrochim.Acta. Part A , 37 (1981) 233-239. (644) S. K. Husain, J. B. Hasted, D. Rosen, E. Nicol, and J. R. Birch, ZnfraredPhys.,24 (1984) 209 - 2 1 3. (645) J. J. Cael, J. L. Koenig, and J. Blackwell, Biopolymers, 14 (1975) 1885- 1903.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I1

313

subtle changes taking place on dextrinization. Other spectral techniques have only limited application in structural studiesof starch and the processes which occur on heating, because of the limited solubility of starch and its components (amylose and amylopectin). Among spectral techniques, nuclear magnetic resonance provides the most valuable information. The 'Hand I3C-n.m.r. methods provide the same type of information, but the former is preferred, because of its higher sensitivity. Unfortunately, in both cases, spectra are recorded for solutions, and consequently, studies are limited to soluble starch. Neither 'H- nor I3C-n.m.r. spectra provide any information on the distribution of branch points. Therefore, a combination of partial enzymic degradation with n.m.r.-spectral analysis allowed inquiry into details of the fine structure of amylopectin and degraded starch.646First, the 13C-n.m.r.spectrum was recorded for soluble starch containing 10- 15% of amylopectin; it showed six signal^.^' Extensive studies carried out by other groups of investigator^^*-^^' allowed them to ascribe the chemical 3), and (1 6) a-bonded carbon shifts at 100 p.p.m. to (1 -P 4), (1 atoms in the sequence of upfield shift in that order. The group of chemical shifts from 80 to 70 p.p.m. belongs to the C-4, (2-3, C-2, and C-5 atoms. The chemical shift at 87 p.p.m. belongs to the oxygen-substituted C-3 atom. Upfield, at 60 p.p.m., the chemical shift of the C-6 atom is located. Moredetailed structural assignments may be conducted by comparison with the relevant spectra of other polysaccharides of established structures. A review652on this subject was published in 1981. One year earlier, the carbon13 cross-polarized magic angle spinning nuclear magnetic resonance (13Cc.p./m.a.s. n.m.r.) technique was applied for ascertaining the characteristics of wheat grain, flour, gluten, and wheat-protein p o ~ d e r s . 6 ~ ~ Severalquestionsarise about the role of the acid catalyst in dextrinization. Such catalysts are'known to promote a carbonium ion mechanism of sugar polymerization in which some rings can be opened. Perhaps, in the case of dextrinization,both the heterogeneity of the process and steric reasons protect the rings against attack by the acid. The nature of the acid catalyst and The mild action the mode of its interaction also play an essential role.560-562 of polyphosphoric acid in dextrinization was interpreted in terms of the possibility of its esterification with D-glucosyl units. Some volatile acids,

-

-

-

(646) (647) (648) (649) (650) (651) (652) (653)

-

-

M. J. Gidley, Curbohydr. Rex, 139 (1985) 85-93. D. E. Dorman and J. D. Roberts,J. Am. Chem. Soc., 93 (1971) 4463-4472. S. R. Erlander, M. M. Purvinas, and H. L. Griffin, Cereal Chem.. 45 (1968) 140- 153. S. R. Erlander and R. Tobin, Mukromol. Chem., 11 1 (1968) 212-225. H. J. Jennings and I. C. P. Smith, J. Am. Chem. Soc., 95 (1973) 606-608. P. Colson, H. J. Jennings, and I. C. P. Smith, J. Am. Chem. Soc., 96(1974) 8081 -8087. P. A. J. Gorin, Adv. Curbohydr. Chem. Biochem., 38 (1981) 13-104. I. C. Baiann and H. Foerster, J. Appl. Biochem., 2 (1980) 347-354.

PIOTR TOMASIK et ul.

314

TABLE XXII Composition and Characteristicsof Yellow D e x t r i O ~ ~ ~ ~ Yieldof fraction

Number of glucose unitsin one segment

Mean molecuk weight (M,.)

Number ofend groups

Nurnberof segments in molecule 3 5 7 27 3 3 5 9 17

2 2 3 4 3 3 3 5 4 5

Product

(%)

Dextrin of low viscosity obtained at 180"

7.5 4.2 16.7 31.9 36.5 8.3 3.5 10.8 34.0 38.5

980 1650 3660 8850 14,630 1760 4160 6450 16,600

2 3 4 6 14 2 2 3 5 9

<5

48,689

13

25

12

<5

51,471

15

29

11

Dextrin of medium viscosity obtained at 180"

Macromolecular dextrins obtained with hydrochloric acid at 125" Macromolecular dextrins obtained with nitric acid at 125"

1400

I1

such as hydrochloric acid, can gradually escape from the reaction mixture, especially when the temperature is being raised. The acid strength of a given ~ the ~ ~ case - ~ ~of~ acid seems to be quite important; it is well d o c ~ m e n t e d in dextrins from potato starch with either hydrochloricacid or nitric acid as the catalyst. The first catalyst gives dextrins of lower viscosity (85.0 mPa.s) and more color than dextrins produced with nitric acid. The viscosity ofthe latter was much higher ( 1 38.0 mPa.s). Both dextrins were separated into five fractions each. The yields of the fractions,their mean molecular weight, and such structural details as number of end groups, number of segments in a molecule, and number of D-glucosyl units in one segment are given in Table XXII.Independent of the acid catalyst, high-molecular-weight,water-insoluble fractions are also formed. Some information on their structures is presented in the same Table. However, not only may the acidity of the acid be a useful property in acidic dextrinization of starch; the anion of the catalyzing acid may be very important in that process. Thus, for instance, (654) J. Skalski, Acta Aliment. Pol., 1 1 (1985) 79-84. (655) K. Nowakowska, Acta Aliment. Pol., 1 1 (1985) 71 -76. (656) A. Sroczynski and K. Nowakowska, Nuhrung, 5 (1986) 475-480.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I1

315

12-tungstosilicicacid exhibits high catalytic activity, among others, because its heteropolyanions form complexes with d e ~ t r i n . ~ ~ ’ The role of base catalysts is not clear. Alkalis are presumed to deprotonate the hydroxyl groups on C-2 and C-3, mainly to form alkali-star~h,~~ but further stages of dextrinization have not been recognized. The results of alkaline hydrolysis are different for amylose and amylopectin. Amylose is depolymerized into “isosaccharans,” formate, and lactate. The branched structure of amylopectin is much more resistant to alkaline degradation. “Peeling” proceeds from the reducing terminus along a main chain, and branches remain intact.132 The kinetics of alkalinethermochemical degradation of polysaccharides was as, in this second-order reaction, organic carboxylic acids were available. Ammonia was found to increasethe rate of dextrinization. At 200°, yellow dextrins additionally produce the aroma of roasted peanuts.659This suggests formation of such aza-aromatic nitrogen heterocyclesas imidazole, pyridine, and pyrazine derivatives. The dextrinization of rice starch in the presence of urea has been attempted by Egyptian workers6@’ However, the reaction conditions applied, namely, up to 60 min at 165 did not cause fundamental changes in the starch. The authors obtained only products of high swellability, having a nitrogen content of 0.64 to 2.83%, probably bonded in the form of a starch carbamate of unrecognized structure. The preliminary step of the reaction includes the formation of a complex, shown to be useful as a ruminant feed.661 The authors of this Chapter have developed studies on dextrinization of potato starch in the presence of several a-amino acids. Conducted at 160220”,this processyields light to dark brown solids bearing secondary aromas specific for roasted, baked, or fried foods. These aromas are dependent on the amino acid used, as well as on both the time and temperature of such reactions.662Umano and ShibamotoM3showed the role of an amino acid (glycine) in the process of roasting of starch at 290”. Comparison of the composition of the reaction mixtures resulting from heating of starch alone and with glycine gave the following results. The content of aliphatic hydrocarbons (mainly cyclohexane) is lower in the process with glycine. The overall content of ketones also decreases, and, instead of 4-methyl-3-penten-3-one,2-hydroxy-2-methyl-1,4-pyrone (maltol), and 2-ethyltetrahydro-3-hydroxy-1,4-pyrone, which are major products from

-

O ,

K. Arai and Y. Ogiwara, J. Appl. Polym. Sci.. 29 (1984) 3087-3095. J. M. Krochta, J. S. Hudson, and S. J. Tillin, Prepr. Pap. Am. Cbem. SOC.,Div. Fuel Cbem., 32 (1987) 148-156; Chem. Abstr., 106 (1987) 178,285s. M. Patasihski, G. Pyre, P. Tomasik, and S. Wiejak, Pol. Pat. P-245 518 (1983). I. Abd El-Thalouth, M. A. El-Kashouti, and A. Hebeish, Staerke, 33 (198 1) 306- 3 10. (661) L. Fodor, S. Gal, and J. Matyas, Gabonaipar, 27 (1980) 105- 109. (662) P. Tomasik and W. Zawadzki, unpublished results. (663) K. Umano and T. Shibamoto, Agric. Biol. Cbem.. 48 (1984) 1387- 1393.

PIOTR TOMASIK et al.

316

starch itself, 2-methyl-3-buten-2-one and 4-penten-2-one, as well as maltol, are major products when thermolysisis carried out with glycine. The yield of maltol decreases to almost 50%of that obtained from starch. There is also a significant decrease in the yields of aldehydes and mononuclear aromatic hydrocarbons, whereas the yields of naphthalene and tetralin dramatically increase. 1,3-Dimethoxybenzene and 2,5dimethyl- 1,Cnaphthoquinone, which are absent among the products from starch itself, are now formed in considerable proportion. The yield of phenols increases, 2-ethoxy- and 2methoxy-phenolsbeing major products of the reaction with glycine, whereas they are absent among the products from pure starch. The yields of furans decrease in the reaction with glycine, and various nitrogen heterocyclic compounds are formed instead. Among them, 2,5dimethylpyrrole, pyrazine, pyridine, 2-ethyl-4,5-dimethylimidazole,and 2,4dimethylpyrrole are major products, with the yields decreasing in that sequence. The mutagenicity of the products obtained in this manner should be noted. These products show a dose-related, mutagenic activity against Sulmonellu typhimurium strains TA 98 and TA 100. It was also shown that products obtained from starch alone under the same conditions, namely, 40 min at 290" in an atmosphere of nitrogen, do not exhibit any mutagenic effectm On the other hand, another report665stated that tar materials from potato starch heated to a high temperatureare weakly mutagenic. It was even suggestedthat such polysaccharides as starch inhibit the formation of mutagenic products from cooked proteins and products of the reaction of sugars with proteins.M6 Dextrinization in the presence of such oxidants as halogens and hypochlorites predictably r e s ~ l t ~in ' -oxidation ~~ of the C- 1 atoms of reducingS~ to carboxyl groups, as shown in structure 3. terminal D - ~ ~ U C O residues

OH

OH

OH

3 (664) C.-I. Wei, K. Kitamura, andT. Shibamoto, FoodCosmet. Toxicol., 19 (1981) 749-751. (665) M. Nagao, M. Honda, Y. Seino, T. Yahagi, T. Kawachi,andT. Sugimura,CancerLett.,2 (1977) 335-337. (666) N. E. Spingarn, C. T. Garvie-Gould, and L. A. Slocenn, J. Agric. Food Chem., 3 I ( 1983) 30 I - 304. (667) M. Samec and M. Blinc, Kolloid Beih.. 38 (1933) 48-56. (668) M. Samec, Kolloid 2.. 64 (1933) 32 1 - 323. (669) S. K. Fischer and F. Piller, Staerke, 29 (1977) 232-235.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I1

317

Stronger oxidation leads to oxidation of the C-6 atoms to carboxylgroups, as Only well as of the hydroxyl groups on the C- 1, C-2, and C-3 oxidations useful in dextrinization are considered here. Oxidized starches yield darker d e ~ t r i n sMore-advanced .~~~ discussionof the problems of starch oxidation is given in the Radley monograph.676 The following general conclusions may be drawn from the literature on the structure of dextrins and British gums, at least of those prepared in the air. One nonreducing end-group is present per - 12 D-glucosyl units. In a maize starch, one such group is present per each 24-30 units. Amylose is virtually absent, and the degree of polymerization is 66. Hence, dextrins are low-molecular-weightproducts having 4 - 5 branches per molecule, with chains composed of 8 D-glucosyl units.

-

-

5. Sources for the Manufacture of Dextrins and British Gums The origin of the starch is very important for the manufacture of dextrins and British gums, as well as for the quality and properties of the final products. The sources most utilized are maize, potato, rice, and tapioca, and, sometimes, also sago, and waxy maize. The source of dextrin used in a given country depends on the economic conditions therein. In the Soviet Uni0n,6~~ for instance, an increased consumption of maize starch for this purpose is noted, whereas, in Poland, potato starch is the sole source of dextrin. In the Federal Republic of Germany, 2 1% of potato starch, 62% of maize starch, and 17% of wheat starch constitute the total amount of starch available in that country.678Dextrinization of such more-exotic starches as those of ragi (Eleusine caracana) and jowar (Sorghum vulgarae) has also been studied.679In some countries, such starches are industrially processed in considerable amounts; for instance, barley in Finland, sorghum in the Sudan, arrowroot in Brazil, and sago in Indonesia. Dextrins of the best quality used to be manufactured from tapioca. Their dispersionsare clear, stable, tasteless, and odorless, and have superior adhesiveness. Comparative studies of the quality of dextrins from tapioca, potato, (670) (671) (672) (673) (674) (675) (676) (677)

G. Felton, F. F. Farley, and R. M. Hixon, Cereal Chem., 15 (1938) 678-689. F. F. Farley and R. M. Hixon, Znd. Eng. Chem., 34 (1942) 677-681. C. C. Unruh and W. 0. Kenyon, J. Am. Chem. Soc., 64 (1942) 127-131. R. W. Kerr, J. Am. Chem. Soc.. 72 (1950) 816-820. F. Ehrlich and R. Guttmann, Biochem. Z., 159 (1933) 100- 109. J. Kucera, A. Burelova, M. Pancek, and B. Fecak, Farm. Obz., 52 (1983) 35-41.

Ref. 480, Ch. 11. Ministerstvo Pishchevoi Promyshlennosti, Pzshch. Prom., Ser. 5 , Krukhm.-Putoch. Prom., 3 (1983) 1 - 19. (678) Anonymous, Kurfoflelwirfschuft,39 (1986) 4. (679) D. B. Wankhede and S. Umadevi, Stuerke, 34 (1982) 162- 165.

318

PIOTR TOMASIK et ul.

and maize were published by Solomina and Tregubov.680Cheaper dextrins and British gums are manufactured from maize and potato starch (farina). Maize dextrins are matte and imperfect in the glossiness of their films, and their dispersions are more viscous than those of farinaceousdextrins. Better properties of this kind are exhibited by farinaceous dextrins; however, they have an unpleasant odor, particularlywhen they belong to the class of yellow dextrins. The quality of the dextrins desired, and selection of the most reasonable technologies, cannot be based merely on the type of starch used. Geographical location can seriously influence the properties of a given type of starch, such as structure of the granules, amylose to amylopectin ratio, and so on.537,68* The conditions of dextrinization should be carefully selected, adapted not only to a given type of but also to such variable properties as seasonal, climatic, and agricultural conditions. More-essential variables are the pH after acidifying the starch, and the moisture content, and a minor consideration is the granule size-distribution. Farina starch is the most difficult to employ, and requires the most work in processing. This is due not only to the morphological characteristicsof the granules, but also to their phosphoric acid content, which depends on their size. Calculation of the amount of acid necessary for dextrinization is based on the acid factor, determined potentiometrically by titration of the starch, suspended in deionized water, with 0.1 M hydrochloric or 0.05 M sulfuric acid. The acid factor is read, on the curve of pH related to the volumesmof acid added, at the point of pH 3.0. Preliminary modification ofthe sourcein order to obtain dextrins of better quality can be applied in the form of steaming and rolling of the starchy grai11s.6~~ Storage of the source materials may also be essential for the high quality of dextrin prepared from them. Optimal storage conditions can be selected based on the thermodynamic characteristics of the de~trins.~~~ The manufacture of pyrodextrins and British gum-like products requires addition of some kind of catalyst. Among them, the acidic catalysts, hydrochloric, phosphoric, and nitric acid, are the most frequently used. Comparative studies have shown that phosphoric acid (the weakest of the three)exerts the least effect on farina starch.560Treatment of starch with sulfuric acid, sulfurous acid, sulfur dioxide, hydrochloric acid, mixtures of hydrochloric

(680) L. S. Solomina and N. N. Tregubov, Izv. Vyssh. Uchebn. Zuved., Sukh. Prom., (1983) 52-54. (681) Ref. 480, Ch. 4. (682) R. L. Datta and P. K. Chatterjee, Chem. Age (London), 46 ( 1942) 189- 19 1,20 I - 202. (683) M. Amario, G . Piva, and M. A. Beghian, Zootech. Nutr. Anim., 5 (1979) 316-326. (684) S. G. Tarasov, Zzv. Vyssh. Uchebn. Zuved., Sukh. Prom., (1975) 73-77.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I1

319

and phosphoric or nitric and phosphoric as well as such carboxylic acids as lactic and oxalic acid have also been l i ~ t e d . ~Dextrinization ~~,~’~ of a 25% aqueous suspension of starch with carbon dioxide under high pressure1@may also be included in the group of acid-catalyzed processes. The second group of catalysts covers all alkalis and other basic catalysts, including sodium hydroxide and such hydrolyzable salts of weak acids as carbonates, hydrogencarbonates, perchlorates, hypochlorites, and so on. Hypochloritesalso act as oxidants. They provide dextrins of excellentclarity and luster in aqueous solution.687 ammonium compounds,688 and urea689should also be considered. Many authors have reported the use of catalytic amounts of such oxidants as halogens,670,674.6w-693 nitrogen and h y p o ~ h l o r i t e s , ~and ”,~~~ ozone diluted with airSmshould also be mentioned. Some attention has been paid to the role of various salts as catalysts. These studies were performed by Bryce and G r e e n w o ~ d ~in ” ,order ~ ~ ~to establish the effect of the addition of 2% of various salts on the decomposition of amylomaize starch at 225 to 337 into carbon mon- and di-oxide and water; the results will be discussed in Section VI. Manganese(I1) chloride in the presence of nitric acid,682or aluminum salts696should be mentioned as catalysts for starch dextrinization. Aqueous solutions of various salts depolymerize starch granules at room temperature. The anions of these salts have a more pronounced effect than the cations.697Russian developed studies on dextrinization in the presence of either alkalinized or acidified alum; they obtained either white or light-yellowdextrins of high quality (685) R. Schirner, M. Leissner, and H. Rothfuss, East Ger. Pat. 158,111 (198 1); Chem. Abstr., 98 (1983) 200,149r. (686) K. W. Stevenson and S. Cooke, W. Ger. Pat. 1,938,394 (1970); Chem. Abstr., 72 (1970) 1 13,050~. (687) R. K. Samuel, R. J. Kozlowski, and F. J. Pustek, U.S. Pat. 4,549,909 (1985); Chem. Abstr., 104 (1986) 1 1 1,730~. (688) J. E. Clegg and F. C. Hilliard, U.S. Pat. 2,127,205 (1938); Chem. Abstr.. 32 (1938) 8183. (689) G. V. Caesar, U.S. Pat. 2,131,724 (1938); Chem. Absfr.,32 (1938) 9341. (690) R. W. Kerr, U.S. Pat. 2,108,862 (1938); Chem. Abstr., 32 (1938) 3188. (691) C. Ekrquist, U.S. Pat. 1,851,749 (1932); Chem. Abstr., 26 (1932) 3135. (692) A. 0. Fuller, U.S. Pat. 1,937,752 (1933); Chem. Absfr., 28 (1934) 1214; U.S. Pat. 1,942,544 (1934); Chem. Abstr., 28 (1934) 1888. (693) A. J. Bulfer and C. C. Gapen, U.S. Pat. 2,287,599 (1942); Chem. Absfr.,37 (1943) 280. (694) J. Potze and P. Hiemstra, Stuerke, 6 (1963) 217-225. (695) D. J. Bryce and C. T. Greenwood, J. Appl. Polym. Sci., Pt. C, 2 (1966) 159- 173. (696) K. Pearl andF. Steinitzer,Ger. Pat. 456 841 (1927); Chem. Zentrulbl..I1 (1)(1928) 1035. (697) M. Samec, Kolloidchemie der Stuerke, Steinkopf Verlag, Dresden, 1927. (698) L. S. Solomina, E. A. Shtyrkova, and N. N. Tregubov,Zzv. Vyssh. Uchebn.Zuved., Sukh. Prom., (4) (1979) 40-42. (699) L. S. Solomina, E. A. Shtyrkova,and N. N. Tregubov,Zzv. Vyssh., Uchebn.Zuved., Sukh. Prom., (4) (1979) 52-54.

-x M I""IL

II 1/ //

/

/-

0

60

120

180

2LO

Time (min) FIG.22.- Relationship between Amylolytic Index, Reducing Sugars,and Solubility of Dextrin, and Time of Dextrinization of Maize Starch in the Process Involving Alum.701[Curves from I to 5 correspondto temperaturesofthe processwhich decreaseregularly from 180 to 140" in decimal ranges, respectively.]

THERMAL DECOMPOSITION OF CARBOHYDRATES. I1

32 1

with a minimum of grit and specks. The content of nonalkalinized alum, that is, of alum which liberates a strong acid on hydrolysis, varied from 0.25 to 1.5%, calculated on dry starch. Dextrinization is accelerated by a higher content of alum, and less-viscous material is obtained. Simultaneously, the color of the resulting product is inten~ified.~'"The properties of dextrins depend on the reaction time and temperature of heating70'.702 in the manner presented in Fig. 22. Dextrinization of starch treated with alum proceeds in a slightly different way on using infrared radiation.703

6. The Manufacture of Dextrins and British Gums The process of manufacture of dextrins consists of eight essential stages. The first is addition of the catalyst. It is, perhaps, the most important manip ulation, because the catalyst must be distributed very evenly on the starch granules, and simultaneously, the penetration of catalyst into the whole volume of starch has to be maintained, as otherwise, severalproblems will be encountered in the second stage of operation. Depending on the catalyst, l7 Swellvarious techniques for its administration are in use~83,689-691,693,704-7 ing of the starch prior to addition of the catalyst was patented.713 Maturation is the next stage in the preparation of dextrins. It is necessary to allow the catalyst to penetrate into the starch granule, in order to correct imprecise distribution of acid, should this have happened in the first stage. The matured source is then subjected to drying, to avoid hydrolytic scission of the starch in the first period of roasting (as well as gelatinization, or L. S. Solomina, Zzv. Vyssh. Uchebn. Zuved., Sakh. Prom., (10) (1979) 49-50. L. S . Solomina and N. N. Tregubov, Izv. Vyssh. Uchebn. Zuved., Sakh. Prom., (1) (1980) 5 1-53, L. S. Solomina, Izv. Vyssh. Uchebn. Zuved., Sukh. Prom., (3) (1985) 47-49. V. A. Kovalenok, I. I. Kuznetsova, and N. N. Tregubov, Zzv. Vyssh. Uchebn. Zaved., Sukh. Prom., ( 11) ( 1981) 47 -49. S. G. Moms, U S . Pat. 2,359,378 (1944); Chem. Abstr., 39 (1945) 832. W. Browing and J. Barlow, U.S. Pat. 773,469 (1904); H. J. Roberts, in Ref. 503, Ch. 13. A. Fielding, Br. Pat. 22,455 (1910); H. J. Roberts,in Ref. 503, Ch. 13. H. Wulken, U.S. Pat. 993,011 (191 1); W. Acton, in Ref. 506, Ch. 15. H.C.Gore,U.S.Pat. 1,335,162(1920);Chem. Abstr., 14(1920) 1617. Haake Akt. Ges., Ger. Pat. 273,420 (1912); W. Acton, in Ref. 506, Ch. 15. R. W. G. Stutzke, U.S. Pat. 1,516,512 (1924); Chem. Abstr., 19 (1925) 416. P. A. Singer, US. Pat. 1,564,979(1925); Chem. Abstr., 20 (1926) 516. R. J. Brindle, U.S. Pat. 1,505,656 (1924); Chem. Abstr., 18 (1924) 3290. H. T. Boehm Akt. Ges., Ger. Pat. 252,827 (1912); Chem. Zentralbl., I1 (2) (1912) 1795. H. E. Bode, US. Pat. 2,156,488 (1939); Chem. Abstr., 33 (1939) 6084. M. D. Rozenbroek, Br. Pat. 544,245 (1942); Chem. Abstr., 36 (1942) 6368. Y. Komai, K. Mifune, and A. Inubashi, Jpn. Pat. 70 07,946 (1970); Chem. Abstr., 73 ( I 970) 36,8242. W. A. Scholten, Br. Pat. 1,140,189(1969); Chem. Abstr., 70 (1969) 79,363.

PIOTR TOMASIK et al.

322

aggregation), as these phenomena are very undesirable. If white dextrins are required after the acidificationwith acid (for instance, nitric acid, d 1.07),the starch may be subjectedto seasoning, followed by two-stage drying in drums. In the first stage, on increasingthe temperature from 60 to 1 15 the content of moisture is lowered to 5%, and, in the second stage, at 75 the moisture content reaches 2%. The total drying time should be 25 min. Potato starch yields white dextrin of 50% If British gums are prepared, drying may constitute the first step of manufacture. Several methods have been described, and patented, for carrying out the operation of the next stage, namely, roasting. A Hagen cooker7I9for conventional roasting is constructed of a series of jacketed metal pans heated by steam coils. This method is neither convenient nor safe. More effective is the rotary-kiln system.5MSome particular processes and improvements have been ~ l a i m e d , ~ ~ among O-~~ them, ~ microwave727and d i e l e ~ t r i c heat~~~-~~~ ing, as well asjetting of the starch under pressure at heated plate^.^" Infrared More-modern technological solutions irradiation can also be applied.701,735 are an autoclave system equipped with stimng (the Blattman process736); this is a typical, one-container process in which acidification, maturation, drying, roasting, and cooling can be conducted in one vessel. Further devel-

-

-

O,

-

O

K. Baranowski, Pol. Pat. 107,385 (1980); Chem. Abstr.. 94 (1981) 67,599~. W. J. Rowe and C. Hagen, U.S. Pat. 2,332,345 (1943); Chem. Abstr., 38 (1944) 1660. N. C. Phillips, U.S. Pat. 1,894,570 (1933); Chem. Abstr., 27 (1933) 2599. A. C. Horesi, U.S. Pat. 2,274,789 (1942); Chem. Abstr., 36 (1942) 4366. A. Sroczynski and J. Skalski, Pol. Pat. 55,470 (1968). A. Sroczynski and J. Skalski, Przem. Spozyw., 19 (1965) 648-654. G. P. Krause, Ger. Pat. 549,7 1 1 ( 1934); J. A. Radley, in Ref. 506, Ch. 15. B. Marlewska, J. Stasihski,Z. Wertz, and M. Politowski, Pol. Pat. 72,357 (1974); Chem. Abstr., 84 (1976) 76,097~. (726) Okawara Mfg. Co., Ltd., Jpn. Pat. 80,156,600 (1980); Chem. Abstr., 94 (1981) 105,262J (727) P. Verberne and T. Vlot, Br. Pat. 1,425,624 (1976); Chem. Abstr., 85 (1976) 34,947j. (728) S. Neuman, U.S. Pat. 2,494,191 (1950); Chem. Absrr., 44 (1950) 2780. (729) C. Ziegler, R. Kohler, and H. Ruggeberg, U.S. Pat. 2,818,357 (1957); Chem. Abstr., 52 (1958) 12,438. (730) C. M. Lakshmanan, B. Gal-Or, and H. F. Hoerlscher,Ind. Eng. Chem., Prod. Res. Dev., 8 (1969) 261-267. (731) C. M. Lakshmanan, B. Gal-Or, and H. F. Hoerlscher, Stuerke, 22 (1970) 221 -227. (732) C. M. Lakshmanan and H. F. Hoerlscher, Ind. Eng. Chem., Prod. Res. Dev., 9 (1970) 57-59. (733) C. M. Lakshmanan and H. F. Hoerlscher, Staerke, 22 (1970) 261 -264. (734) R. Kiyama and H. Kinoshita, Rev. Phys. Chem. Jpn.. 22 (1952) 18-21. (735) A. I. Syroedov, S. G. Ilyasov, A. I. Zhushman, V. A. Kovalenok, and N. D. Lukin, USSR Pat. 576,884 (1977); Chem. Abstr., 89 (1978) 131,472r. (736) M. A. Staerkle and E. Meier, U.S. Pat. 2,698,937 (1955); Chem. Abstr., 49 (1955) 4315.

(718) (719) (720) (721) (722) (723) (724) (725)

THERMAL DECOMPOSITION OF CARBOHYDRATES. I1

323

opment brought the fluid-bed p r o ~ e s s ,which ~ ~ ~possesses - ~ ~ several advantages over those already discussed. The manufacture of either dextrin or British gum may be carried out in the same manner. To protect the dextrins against further, undesirable conversions,the roasted product is cooled in the next stage of the manufacture. Subsequent stages are remoistening, blending, and packing. An apparatus for cooling and moistening dextrin was patented by Schirner and coworkers.743Remoistening is carried out in order to remove absorbed gases. Blending is necessary to obtain standard batches with regularity. Because of the great variety of properties of the sources, and several subtleties of the dextrinization procedures, it is rather difficult, without blending, to obtain products that correspond to exact standards.

7. Uses of Dextrins and British Gums Dextrins serve first of all as adhesives. In several countries, they are used to gum stamps and labels, and as a liquid glue for various purposes. At present, they are applied in mixtures with other compounds, such as poly(viny1 acetate)or poly(viny1alcohol), or both, and fillers, for instance, clay. Gradually, dextrin is being replaced in adhesives by synthetic polymers. Reviews on dextrins and dextrin-based adhesives have been written by Auguin and A n ~ a r tas , ~well ~ ~ as by Malet~ky.’~ Another common application of dextrins is as dressings for linen, laces, thickeners in finishing of textiles, leather, and buckram, and for the warp in the fabrication of textiles. The optimal composition of a thickener for printing paste can be designed by means of a mathematical-statistical method proposed by D e n k ~ vThe . ~ ~content ~ of dextrinsin an adhesive can be determined by gas- liquid chromatographyof

(737) (738) (739) (740) (741) (742) (743) (744) (745) (746) (747)

R. E. C. Frederickson, U.S. Pat. 3,003,894 (1961); Chem. Abstr., 56 (1962) 2624. R. E. C. Frederickson, U.S. Pat. 2,845,368 (1957); Chem. Abstr., 53 (1959) 2658. CPC International Inc., Neth. Pat. 75,13,763 (1976); Chem. Abstr.. 86 (1977) 9 2 , 2 2 2 ~ . R. Schirner, G. Bernhardt, T. Bernhardt, I. Niedner, H. Fischer, G. Freyer, and H. Rothfuss, W. Ger. Pat. 2,318,035 (1974); Chem. Abstr., 82 (1975) 6 0 , 3 9 8 ~ . R. Schirner, G. Bernhardt, T. Bernhardt, I. Niedner, H. Fischer, G. Freyer, and H. Rothfuss, USSR Pat. 644,838 (1979); Chem. Abstr., 90 (1979) 153,7150. A. W. Blomberg and J. E. Jackson, Belg. Pat. 882,718 (1980); Chem. Abstr., 94 (1981) 105,427~. R. Schirner, H. Fischer, M. Leissner, W. Lange, and P. Dittmar, E. Ger. Pat. 143,929 (1980); Chem. Abstr., 94 (1981) 177,057r. R. Schirner,M. Leissner, and H. Rothfuss, E. Ger. Pat. 157,347 (1982); Chem.Abstr.,98 (1983) 109,227d. H. Anguin and M. Ansart, Muter. Tech. (Paris), 72 (1984) 281 -283. A. Maletsky, Adhesion. 24 (1980) 347-349. A. Denkov, Textilveredlung, 19 (1984) 78-82.

324

PIOTR TOMASIK et a].

the pyrolytic products.748As thickeners, dextrins are applied in the manufacture of dyes and other colors, paints, ink, and sized water-paint. They are sizing agents for wall-paper, felt, paper, matches, glass fibers, fireproof mortars, ceramic glaze, fireworks, explosives, and softeners for fabrics. Some attempts have been made to replace dextrin by syntheticpolymers. Thus, sizes for transfer paper may contain poly(viny1alcohol), 0-( hydroxypropyl)cellulose, 0-methylcellulose, or saponified acrylonitrile- itaconic acid- methyl methacrylate copolymer instead of dextrin. Paper sized with A these agents has good brightness and only a low tendency to product having good penetrability, water-solubility, viscosity, stability, and high film-elongation is available by dextrinization of corn starch with urea, and mixing of that dextrin with talc, an emulsion oil, and glycer01.~~ Dextrin may be added to some drilling fluids. The use of dextrins as the binder for many purposes is very common. Depending on circumstances, such binders play different roles. Its main function in briquettes, bricks, and ceramics is to facilitate hardening. In the case of sand molds for casting,not only is shorteningofthe time of hardening achieved, but also, application of high temperature to, and mechanical disintegration of, the layer becomes easier. Addition of dextrin to such material lessens the temperature expansion of the material. As an additive to cement, dextrin is also responsible for water setting in concrete and generally improves its workability and ~ollapsibility.~~~ -754 Several pharmaceuticals, herbicides, pesticides, insecticides, disinfectants,bleachingagents, and enzymes are camed on dextrins playing the role of a matrix or excipient. Matrices for biologically active substances have been described by S ~ h r o e d e r Such .~~~ enzymes as l y ~ o z y m e , sarcosine ~ ~ ~ , ~ ~~~x i d a s e serine , ~ ~ ~ p e ~ t a s e and ,~~~ (748) G . Stoev, Kh. Alaminov, and G. Borisov, Khim. 2nd. (SoJia),56 (1984) 267. (749) A. E. Anderson and V. A. Volkov, Bum. Prom., (1) (1984) 15-16. (750) H. Yuan and Y. Shi, Chin. Pat. 85,108,465 (1986); Chem. Abstr., 107 (1987) 42,Ol In. (751) T. Dusmuradov, Sh. M. Rakhimbaev, M. N. Golubev, B. I. Gavrikov, and A. D. Dzhuraev, Dokl. Akad. Nauk. Tadzh. SSR, 27 (1984) 110- 112. (752) Denki Kagaku Kogyo K.K., Jpn. Pat. 59,121,145 (1984); Chem. Abstr., 102 (1985) 50,144v. (753) V. S. LaFay and S. Neltner, Trans. Foundrymen’s Soc.. 94 (1986) 271 -276. (754) G . L. Datta and N. Jain, Indian Foundry J., 33 (1987) 21 -24. (755) U. Schroeder, Biomaterials (Guilford,Engl.), 5 (1984) 100- 104. (756) T. Matsumoto, T. Yokoyama, H. Aikawa, and Y. Odaka, Jpn. Pat. 79,135,215 (1979); Chem. Abstr., 92 (1980) 135,439k. (757) F. S. Seiyaku K. K., Jpn. Pat. 82,46,922 (1982); Chem. Abslr., 96 (1982) 2 2 3 , 2 9 8 ~ ~ . (758) Noda Institute for Scientific Research, Jpn. Pat. 80,34,001 (1980); Chem. Abstr., 93 (1980) 68,683~. (759) K. Ishida, K. Kamyama, H. Yamada, and S. Sato, Jpn. Pat. 60,244,288 (1985); Chem. Abstr., 104 (1986) 213,248q.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I1

325

other^,^^-^^^ as well as hormones,763may also be closed in dextrin matrices. The presence of dextrin provides the proper level of moisture in a given product. The additon of dextrin to pellets and the coatings of pellets makes these pharmaceuticals easily disintegrable under the influence of saliva or gastric acids. According to Indian nondextrinized starch is a better binder than dextrin itself. Apart from its obvious uses, gelatinized starch plus dextrin has been patented as a medium which, on spraying, controls dust and spontaneous combustion in the drying, handling, transporting, and storing of Such coatings used to be formed from polymers containing dextrin as an additive. Binders are formed either solely from dextrin or with dextrin as a component. Enhanced water-absorbance of absorbent polymers may be achieved.766Coatings for special purposes may be formed by suspending certain mineral powders in dextrin. For instance, a suspension of corundum and zircon serves for the lining of glass-melting and glass-annealing furnace~.~~~ Dextrin is extensivelyused in the manufacture of photosensitivematerials (developers, thermal recording materials, optical information recording media, and photohardenable printing and transfers). Among other uses, dextrin diminishes the absorptivity and flabbiness of diazo copying paper.768 Dextrin is also applied as the component of protective colloids,lubricating greases, corrosion inhibitors, and scale-preventing formulations for water treatment. The stabilization of colloidal AgCl increases the sensitivityof the determination769of chloride anions by titration with AgN03. (760) Y. Kasahara, H. Suzuki, and Y. Ashihara, Eur. Pat. 94,777 (1983); Chem. Abstr., 100 ( 1984) 48,086t. (761) Fujirebio Inc., Jpn. Pat. 59,104,324 (1984); Chem. Abstr., 101 (1984) 77,5086. (762) J. S. Harrison, Belg. Pat. 895,827 (1983); Chem. Abstr., 99 (1983) 154,828g. (763) S. M. Cady, R. Fishbein, U. Schroeder,and H. Eriksson, Eur. Pat. 193,9 17 (1985); Chem. Abstr., 105 (1986) 197,215q. (764) A. K. Bandyopadhyay,B. Chandhuri,and P. K. Bhattachajee, Indian J. Pharm. Sci., 42 (1980) 54-55. (765) T. Y. Yang, US. Pat. 4,642,196 (1987); Chem. Abstr., 106 (1987) 199,100t. (766) G. F. Fanta, W. M. Doane, and E. I. Stout, U.S. Pat4,483,950(1984); Chem. Abstr., 102 ( 1985) 47,667n. (767) I. G. Orlova, E. V. Degtyareva, L. V. Miroshnik, E. D. Lisovaya, and V. V. Kosheleva, Ogneupory, (9) (1980) 45-48. (768) B. Dutkiewicz, T. Gajda, and A. Czajkowski, Pol. Pat. 99,075 (1978); Chem. Abstr., 92 (1980) 224,267e. (769) 2.Yamaguchi, Gijutsu Joho-Shizuoka-kenEisei KenkyoSentu, 2 (3)( 1984) 3-4; Chem. Abstr.. 103 (1985) 47,441e.

326

PIOTR TOMASIK et al.

Several reports have been devoted to the role of dextrins as electroplating brighteners for various metal coatings. Dextrins act either as s~rfactant~~O or as camers ofthe ions to the place ofdischarge.The presence ofdextrin lowers cathodic polarization at a given current density. Comparativelylarge values of throwing power were obtained after addition of dextrin to the bath; some other compounds (for instance, L-serine or D - ~ ~ U C act O Ssimilarly. ~) In the case of simultaneous deposition of Ni - Mn - Zn alloy, addition of dextrin has a different influence on each of these metals; the content of deposited nickel and manganese increases in the presence of dextrin, whereas that of zinc decreases.771The effects of dextrin are, however, chimeric and strongly dependent on the deposited metal, as well as on the composition ofthe bath. Thus, contrary to the observations on the deposition of the Ni-Mn-Zn alloy, addition of dextrin to the bath for deposition of a Ni-Cd-Zn alloy causes an increase in the cathodic polarization. The content of nickel and zinc in the deposit increases, and that of cadmium decreases. Dextrin does not provide a smooth, bright layer of alloy; it is grayish, uneven, and cracked.772Japanese workers773reported that a plate of pure titanium utilized as the anode immersed in a solution containingphosphate and dextrin took on a beautiful, gold appearance. The typical concentrationof dextrin in the bath varies from 0.5 to 6 g/L. The use of dextrin in combination with poly(viny1 alcohol) is claimed to lead to improved, fine-grained layers, at least in bright-zinc ele~troplating.~~~ In some cases, the addition of dextrin seems to have a positive effect upon the corrosion resistance of material electroplated with cadmium775*776 and with titanium alloyed with zinc or cadmium.776Voltamperometric analysis of dextrins in a galvanic bath was described by a group of Russian workers.777 The addition of dextrin to fuel corn position^^^^^^^^ has also been described. Dextrins have been mentioned as a source for the manufacture of tartaric (770) S. I. Berezina, N. B. Berezin, and N. V. Gudin, Zashch. Met., 21 (1985) 572-576. (771) R. Kashyap, S. K. Srivastava, and S. C. Srivastava, Surf: Coat. Techno!., 28 (1986) 129-137. (772) S. W. Srivastava, R. Kashyap, and S. C. Srivastava, Acta Chim. Hung., 117 (1984) 3-9. (773) T. Togaya and M. Suzuki, Shika Giko, 12 (1984) 1477- 1480 Chem. Abstr., 102 (1985) 191,110t. (774) T. V. Venkatesha, J. Balachandra, and S. M. Mayauna, Met. Finish., 83 (8) (1985) 33 - 36. (775) V. P. Maksimchuk and M. T. Sopoleva, Zushch. Met., 23 (1987) 678-681. (776) L. I. Dylaeva, N. B. Chertovskikh,and K. B. Usenko, Zushch. Met., 20 (1984) 805-806. (777) 0.A. Aydashkina, R. Yu. Bek, and Yu. B. Kletenik, Izv. Sib. Otdel. Akad. NaukSSSR. Ser. Khim. Nauk, (2) (1985) 68-72. (778) T. Horiba, K. Iwamoto, M. Kawana, K. Fujita, and K. Tamura, Eur. Pat. 92,802 (1983); Chem. Abstr., 100 (1984) 10,072~. (779) V. Y. Han, W. Ger. Pat. 3,044,687 (1981); Chem. Abstr., 96 (1981) 71,9288.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I1

327

and as a substrate for enzymic modification.780They may also be enzymically converted into ~-glucose.~~l Another application of dextrin in synthesis results from the ability of the propagation of free radicals. It was demonstratedthat radical polymerization of either methyl methacrylate or methacrylic acid in dextrin does not give optically active polymer, but the product was 7 1% isotactic. Dextrin seems to propagate radicals in its asymmetric surrounding perhaps through an electrostatic or hydrogen-bonding interaction.782It should be mentioned that polysaccharides, among them dextrin, form interpolymer complexes with other polymers, such as poly(ethy1eneoxide) or poly( l-vinyl-2-pyr~olidinone). These well defined complexes may exist because of hydrogen bonding. Thus, both of the aforementioned polymers form 1 : 1 complexes with dextrin, but poly( 1-vinyl-2-pyrrolidone)forms 2 :3 complexes783with either dextran or inulin. Low- and high-molecular-weightcarbohydrates are capable of formation of complexeswith metals.784Such complexes also exist in food systems, with consequences as regards the taste, nutritional value, and technological quality of foods. The taste of iron salts may be masked by complexing with carbohydrates at high pH and be exposed in the same manner at low pH. Complexation of the Cu2+ion protects vitamin C from copper-catalyzed Dextrin may be considered to be a useful complexing agent for ferric hydroxide. Iron ion in such a complex is readily bioavailable, although the bioavailability is reported to be decreased in comparison with normal stores of iron and, especially, with quick-release preparation^.^^^-^^^ Exhaustive studies of such complexes were performed by Erni and coworkers.789

-

(780) A. Sroczybski, M. Boniek, and T. Piengalski, Rocz. Technol. Chem. Zywn., 16 (1969) 7-20. (781) Agency of Industrial Sciences and Technology, Japan Organo Co. Ltd., Jpn. Pat. 58,81,793 (1983); Chem. Abstr., 99 (1983) 138,189~. (782) K. Nozaki, Y . Matsubara, M. Yoshihara, and T. Maeshima, Makromol. Chem.. Rapid Commun.. 5 (1984) 723-724. (783) H. Ohno, H. Takinishi, and E. Tsuchida, Makromol. Chem., Rapid Commun., 2 (1981) 5 1 1-5 15. (784) J. A. Rendleman, Food Chem., 3 (1978) 47-49, 127- 162. (785) H. Cross, T. Pepper, M. W. Kearsley, and G. G. Birch, Staerke, 37 (1985) 132- 135. (786) H. C. Heinrich and R. Fischer, Therapiewoche, 34 (1984) 4225-4226, 4228, 4231 4232,4236-4240. (787) H. C. Heinrich, R. Fischer, E. E. Gabbe, and N. Theobald, Klin. Wochenschr.,61 (1983) 103- 110. (788) W. Forth and W. Schneider, Med. Klin.. 80 (1985) 697-699. (789) 1. Emi, N. Oswald, H. W. Rich, and W. Schneider, Arzneim.-ForschJDrug Rex, 34 (1984) 1555-1559.

328

PIOTR TOMASIK et a1

The iron preparations made from iron hydroxide and dextrin (see a patent790 for their easy preparation) were improved from the point of view of the bioavailability of iron by addition of hydroxy carboxylic a ~ i d s .Prepa~ ~ ~ , ~ ~ ~ rations of iron with dextrin may be used as magnetic particles for diagnostic purposes in parenteral a d m i n i ~ t r a t i o n .Dextrin ~ ~ ~ . ~in ~~ the complexes mentioned may play a role as complexing agent as well as stabilizer, but the mechanism of the action is not yet clear. For instance, the complexes of zinc and ~ - g l u c i t o las , ~well ~ ~as ofiron and ~-glucitol’~~ are stabilized by addition of d e ~ t r i nit; may ~ ~ ~possibly be a co-complexingagent. Complexesof aluminum as well as borax,798with dextrin have also been studied. Because starch itself rather strongly (and selectively)bonds metal ions,7wfurther studies of complexes of metals with dextrins may be anticipated. The problem of solvation of dextrins is still unsolved in general. It was documented that even such commonly used solvents as ethanol exhibit an unusually strong interaction with dextrin, as determined by lH-n.m.r. spectroscopy.800The interaction of dextrin with normal alcohols from methanol to 1-hexanol was studied by determination of their infinite-dilutionactivitycoefficients in dextrin - water mixtures.801Studiescarried out on the adsorption of dextrin on oxidized coal revealed that hydrophobic moieties in dextrin are involved in that process.802The possibility of hydrophobic bonding has made dextrin an interesting component of the media used for flotation. Dextrin acts as a flocculant (depressant). When combined with a proper additive providing action by means of hydrogen bonding, reverse flotation (790) J. Verdagner Bronsoms, M. Vich Figa, and J. Fabrega Dalman, Span. Pat. 517,998 (1984); Chem. Abstr., 105 (1986) 66,441n. (791) Nippon Zoki Pharmaceutical Co. Ltd., Jpn. Pat. 60,06,701 (1985); Chem. Abstr., 102 (1985) 1 7 , 6 5 9 ~ (792) T. Kurosaki, K. Ota, H. Matsura, and K. Sawada, Jpn. Pat. 61,28,502 (1986); Chem. Abstr., 105 (1986) 158,809s. (793) H. Gries, W. Meutzel, C. Zurth, and H. J. Weinmann, Eur. Pat. 186,616 (1986); Chem. Abstr., 105 (1986) 120,790r. (794) H. Gries, W. Meutzel, C. Zurth, and H. J. Weinmann, W. Ger. Pat. 3,443,251 (1986); Chem. Abstr., 105 (1986) 120,773n. (795) I. H. Siddiqui and S. A. H. Zaidi, Pak. J. Sci. Ind. Res., 29 (1986) 408-41 1. (796) J. Jafri, S. A. Khan, and S. A. H. Zaidi, Puk. J. Sci.Ind. Res.. 23 (1980) 237-239. (797) I. H. Siddiqui, J. Jafri, R. Jafri, and S. A. H. Zaidi, Puk. J. Sci. Ind. Res., 29 (1986) 95-101. (798) R. Marakami, Kenkyu Hokoku-Kurnumoto, Kogyo Daiguku, 7 (1982) 35-39; Chem. Abstr., 97 (1982) 57,431~. (799) N. Perisic-Janjic,V. Canic, and S. Radosavljevic,Chromutogruphiu,17 (1983)454-455. (800) A. Kaji and M. Muamo, Desalination, 54 (1985) 351-359; Chem. Abstr., 104 (1986) 20,948j. (801) A. LebertandD. Richon,J. Agric. FoodChem., 32(1984) 1156-1161. (802) J. D. Miller, J. S. Laskowski, and S. S. Chang, Colloids Sud, 3 (1983) 137- 15 1.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I1

329

may be achieved, as described in the case of separation of iron pyrites from cod8O3and of coal itself.804The proportion of dextrin added as a depressant in any flotation system used may vary from 30 to 500 mg/L. Because dextrins are carbohydrates, they are accepted in nutrition as a "generally-recognized-as-safe" ( G U S ) , direct ingredient for food and cosmetic^.^^^-^^^ The sole limitation arises from the method of their manufacture, mainly from the catalyst added, and retrogradation to a certain extent, from the degree of their gelation. In the United States, starches oxidized with hypochlorite are added to margarine.811In some countries, only those dextrins which are manufactured with hypochlorite, phosphoric acid and its sodium salts, sulfuric acid, and sodium hypochlorite are accepted in n ~ t r i t i o n . ~ ' ~ - ~ ' ~ In the cosmetic industry, dextrin is applied as the moisture-maintaining component of powders for the skin.815Makeups, and gelating agents for lipsticks and other cosmetics and ointments, involve esters of dextrin with higher fatty acids; for instance, dextrin h e x a d e ~ a n o a t e . ~ ' ~ - ~ l ~ (803) J. D. Miller, C. L. Lin, and S. S. Chang, Coal Prep. (Gordon Breach), 1 (1 984) 2 1 - 38. (804) C. J. Im and F. F. Aplan, in A. R. Swanson (Ed.),Proc. Aust. CoalPrep. ConJ, Ist, Broken Hill Propr. Co. Ltd., Wallsend, Australia, (1981) 183-204. (805) Joint FAO/WHO Expert Committee on Food Additives, Report No. 15, Geneva, June 1971. (806) Joint FAO/WHO Expert Committee on Food Additives, Report No. 17, Geneva, October 1973. (807) U.S. Food and Drug Administration, Fed. Regist., 27 March 1979, 44 (60) 18,24618,249. (808) U.S. Environmental Protection Agency, Fed. Regist., 19 Aug. 1980, 45 (162) 55,19855,200. (809) U.S. Food and Drug Administration, Fed. Regist., 20 Aug. 1982, 47 (162) 36,43736,640. (810) LJS. Food and Drug Administration, Fed. Regist., 15 Nov. 1983, 48 (221) 51,9075 1,909. (81 1) L.Chalmers,MunufChem. AerosolNews, 39(8)(1968)23-28,39(9)(1968)31-36,42. (812) G. Graefe, Staerke, 26 (1974) 145-153. (813) M. GrzeSkowiak and B. Nowicki, Pol. Pat. 39,987 (1956). (814) Pudding Flour, Pol. Stand., BN/75 8084-01. (815) Nippon Surfactants Industry Co. Ltd., Jpn. Pat. 59,193,812 (1984); Chem. Abstr., 102 (1985) 67,237~. (816) H. Tamai, T. Aya, Y. Fujita, and S. Suzuki, Jpn. Pat. 60,146,81 I (1985); Chem. Abstr., 104 (1986) 10,396~. (817) K. Yadosaki and T. Tsuchiko, Jpn. Pat. 61,56,115 (1986); Chem. Abstr., 105 (1986) 29,806q. (818) T. Ikeda, M. Omura, and M. Tanaka, Jpn. Pat., 61,236,716 (1986); Chem. Abstr., 106 (1987) 7 2 , 7 1 9 ~ . (819) S. Mon, S. Kuwata, and T. Mayuzumi, Jpn. Pat. 62,121,764 (1987); Chem. Abstr., 107 (1987) 121,124e.

330

PIOTR TOMASIK et al.

In nutrition, dextrins are more widely applied. First of all, they are used as sweeteners in compositions with sugar. An increased content of dextrin (up to 22%) increases the storage stability of the Dextrins are often recommended for diabetic diets.824Addition of dextrin preserves starchy foods825-828 ;the addition preserves its flavor and aroma, as well as its shape. Dextrin prevents aggregation of frozen food into and makes food (potato and other vegetable chips) crisp when either immersed, or cooked, in dextrin, preferably in dextrins from the alkaline depolymerization of s t a r ~ h . The ~ ~preserving - ~ ~ ~ ability of dextrin has been extended to preventing the browning of food due to the presence of sugars and amino acids in it. Dextrin is an ingredient in compositions with NaC1, D - ~ ~ U C O S ~ , sodium succinate, and L-glutamic acid, as well as 1,5’-ribon~cleotides.~~~ Dextrins as additives find application in the various fields of food manufacture. When they are added to candies, the latter readily disintegrate.8” In sausages835 and canned dextrin improves the quality (organoleptic properties). In such oriental foods as tofus3’and t e m ~ e h , dextrin 8 ~ ~ provides a nutritious component also acting as a binder. In special breads having a (820) M. Schulman and E. Pitchon, Eur. Pat. 36,738 (1958); Chem. Abstr.. 96 (1982) 67,527e. (821) M. Gliksman and B. N. Wankler, Br. Pat. 1,411,664 (1975); Chem. Abstr.. 84 (1976) 149,564d. (822) M. Gliksman and B. N. Wankler, U.S. Pat. 1,030,395 (1978); Chem. Abstr., 89 (1978) 128,091k. (823) M. Aoi and M. Okumura, Jpn. Pat. 62,100,300 (1987); Chem. Abstr., 107 (1987) 1783I I h. (824) N. N. Godbole, J. Diabet. Assoc. India, 19 (1979) 9- 11. (825) M. Kawamura, Jpn. Pat. 79,105,249 (1979); Chem. Abstr.. 92 (1980) 57,085~. (826) Toppan Printing Co., Ltd., Jpn. Pat. 60,70,053 (1985); Chem. Abstr., 103 (1985) 122,0246. (827) Nikken Chemicals Co., Ltd., Jpn. Pat. 57,166,945 (1982); Chem. Abstr., 98 (1983) 33,432n. (828) J. A. Radley, Industrial Uses of Starch and Its Derivatives, Applied Science Publishers, Ltd., London, 1976, Ch. 2. (829) Chiba Mill Co., Ltd., Jpn. Pat. 59,51,733 (1984); Chem. Abstr., 101 (1984) 53,681h. (830) Kyu-Pi Co., Ltd., Jpn. Pat. 82,58,869 (1980); Chem. Abstr., 97 (1982) 54,3010. (831) Amano Corp., Jpn. Pat. 60,37,940 (1985); Chem. Abstr., 103 (1985) 70,110e. (832) A. Kotani, Jpn. Pat. 61,96,960 (1986); Chem. Abstr., 105 (1986) 96,282d. (833) H. Ogawa and S. Imamura, Eur. Pat. 90,356 (1983); Chem. Abstr., 99 (1983) 21 1,394~. (834) T. Jado and K. K. Daido, Jpn. Pat. 8 1,53,624(1979); Chem. Abstr.. 95 (198 I ) 1 13,796~1. (835) J. Pyrcz and W. Pezacki, Fleischwirtschuft, 55 (1975) 1431- 1434, 1437- 1440. (836) Is0 Jiman K.K., Jpn. Pat. 82,141,262 (1982); Chem. Abstr., 97 (1982) 214,5840. (837) Machida Shokuhin K.K., Jpn. Pat. 59,28,448 (1984); Chem. Abstr., 100 (1984) 208,198~1. (838) U. H. Nishi and K. Inoue, Jpn. Pat. 61,152,256 (1986); Chem. Abstr., 105 (1986) 189,740s.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I1

33 1

high content of molasses, a specific dextrin speeds the reaction between gluten and water; it prevents any increase in mixing time and any lessening of dough stability,both ofwhich would result from the high content of added molasses.839In bakery products, dextrins and dextrin alcohols prevent staling.840(The identification of dextrins in bakery products was described by Menger.841) The use of dextrins as binders is quite common, especiallyin the manufacture of noodle^,^^,^^^ but also in some crackers,844 powdered vegetab l e ~and , ~fresh ~ ~ fish foods.846Dextrin in beer decreases the colloidal stability, but simultaneouslyimproves its foam stability, flavor, and other organoleptic properties.847The complexesof metal hydroxidesor metal oxideswith dextrin treated with C 0 2 at G 25 constitute a useful carbonating agent for beverages.848Dextrins are useful in the manufacture of low-caloric, puffy, and foam-like f o ~ d ~some , ~imitations ~ - ~ of ~ such ~ foodstuffs as cheese,8s3 fat- or oil-replacement~,~~~ taste-modified protein h y d r ~ l y z a t eand , ~ ~other ~ nutrients such as granular diet formulation^,^^^.^^^ and bulking agents.858 O

E. F. Glabe and M. H. Silverbrandt, Baker’s Dig., 54 (3) (1980) 8- 10, 12. Q. P. Corp., Jpn. Pat. 60,49,737 (1985); Chem. Abstr., 103 (1985) 70,0912. A. Menger, Getreide, Mehl, Brot, 34 (1980) 143- 146. Kanegafuchi Chemical Industry Co., Ltd. and Morikawa Shokuhin K.K., Jpn. Pat. 57,186,448 (1982); Chem. Abstr., 98 (1983) 106,0196. (843) Y.Takekoshi, Jpn. Pat. 81,75,073 (1981); Chem. Abstr., 95 (1981) 113,7881. (844) S. Minemura, R. Kimura, and S. Kitamura, Jpn. Pat. 79,143,552 (1979); Chem.Abstr., 92 (1980) 109,423~ (845) Hasegawa T. Co., Ltd., Jpn. Pat. 59,205,932(1984); Chem.Abstr., 102(1985) 147,793~. (846) Kabushiki Kaisha Ueno Seiyaku Oyo Kenkyusho, Jpn. Pat. 59,187,744 (1984); Chem. Abstr., 102 (1985) 1 1 1,909. (847) M. Duran-Chavarria, Ing. Cienc. Quim.. 6 (1982) 173-175. (848) R. S. C. So and A. A. Willi, Eur. Pat. 185,196(1986); Chem.Abstr.. 105(1986) 207,853~. (849) E. Farkas, U.S. Pat. 3,821,428 (1974); Chem. Abstr., 82 (1975) 15,554~. (850) Kyu-Pi Co., Ltd., Jpn. Pat. 58,89,138 (1983); Chem. Abstr., 99 (1983) 86,910~. (851) Sand B. Shokuhin Co., Ltd., Jpn. Pat. 60,06,159 (1985); Chem. Abstr.. 102 (1985) 202,909~. (852) T. Yoshida, M. Matsudaira, and T. Takahashi, Jpn. Pat. 62,9 1,502(1987); Chem.Abstr., 107 (1987) 2 3 5 , 1 9 8 ~ (853) G . A. Zwierzan, N. L. Lacourse, and J. M. Lenchin, U.S. Pat. 4,608,265 (1986); Chem. Abstr., 105 (1986) 207,854~. (854) J. M. Lenchin, P. C. Trubiano, and S. Hoffman, U.S. Pat. 4,510,166 (1985); Chem. Abstr., 103 (1985) 5,2411’. (855) B. Lieske, G. Konrad, and W. Schultze, E. Ger. Pat. 217,981 (1985); Chem.Abstr., 103 (1985) 70,095d. (856) Asahi Chemical Industry Co., Ltd., Jpn. Pat. 60,49,776 (1985); Chem.Abstr., 103(1985) 86.8 176. (857) Ajinomoto Co., Inc., Jpn. Pat. 59,55,831 (1984); Chem. Abstr., 101 (1984) 28,309k. (858) Ajinomoto Co., Inc., Jpn. Pat. 59,53,015 (1984); Chem. Abstr., 102 (1985) 202,901e. (839) (840) (841) (842)

332

PIOTR TOMASIK et al.

Dextrin is one of the components of a parented-nutrition solution in which it plays the role of sugar.859 Dextrins either compete with, or are used in conjunction with, cyclodextrins (cyclomalto-oligosaccharides)in applications as encapsulating agents. They allow encapsulation of ethanol to afford dried alcoholic beverages, even those having a high level of alcoho1,860,861 powdered and vinegar-containing condiment The clarity of drinks is maintained by added dextrins.864Encapsulation allows preservation of the flavor of freeze-dried fruits,865extracts of medicinal dried eggs,867and instant tea,868and permits the storing of flavors for foods and bever* strawa g e ~ .The ~ ~color ~ .of~processed ~ ~ collagen or gelatin hydr0lyzates,8~ and meat873may also be retained. Another application of dextrins is as a source for the manufacture of ~ a r a m e l . ' * ~ The J ~functionalproperties of dextrins in food have been reviewed by Sander874and M e d ~ a l f . ~ ~ ~ In various countries, the ratio of the amount of dextrins used for nutritional purposes to the total amount of dextrins available is, of course, different. In the Federal Republic of Germany, -40% of the dextrins manufactured per annum is consumedby the food industry,and the rest is utilized for Nippon Oils and Fats Co., Ltd., Jpn. Pat. 60,25,934 (1985);Chem. Abstr., 102 (1985) 226,05 Id. Sat0 Shokuhin Kogyo K.K., Jpn. Pat. 82,132,876 (1982); Chem. Abstr., 97 (1982) 196,835~. Sat0 Shokuhin Kogyo K.K., Jpn. Pat. 82,122,788 (1982); Chem. Abstr., 97 (1982) 16 I ,070r. Sat0 Shokuhin Kogyo K.K., Jpn. Pat. 82,91,187 (1982); Chem. Abstr., 97 (1982) 90,4296. Sato Shokuhin Kogyo K.K., Jpn. Pat. 82,129,667 (1982); Chem. Abstr., 97 (1982) 197,119~. M. Moriya, Jpn. Pat. 79,39,457 (1979);Chem. Abstr., 92 (1980) 145,294h. T . Lovric and A. Pozderovic, Prehrumbeno-Tehnol. Rev., 22 (1984) 107- 115. J . Baluch, V. Zatloukalova, and K. Martinovic, Czech. Pat. 232,144 (1986); Chem. Abstr., 106 (1987) 66,143n. S. Maekawa, R. Ishihara, and Y. Asano, Jpn. Pat. 61,21,046 (1986);Chem. Abstr., 104 (1986) 205,837r. M. Kobayashi, S. Shiraishi, K. Matsuhara, and K. Yamashita, Jpn. Pat. 61,146,150 (1986);Chem. Abstr., 105 (1986) 170,988g. 0.Prohaszka,A. Hager Veres, G. Horvath, and I. Varga Kiss, Hung. Pat. 38,805 (1986); Chem. Abstr., 107 (1987) 6070x. B. King, R. Wyler, and J. Solms, in D. G . Land and H. E. Nurnsten (Eds.), Prog. Flavour Res., (Proc. Weurman Flavour Res. Symp.), 2nd, 1978, (1978) 327-335. Nichibi Co., Ltd., Jpn. Pat. 81,68,378 (1981);Chem. Abstr., 95 (1981) 148,920g. F. Mitsui, Jpn. Pat. 60,53,577 (1985);Chem. Abstr., 104 (1986) 128,674j A. Nakajima, Jpn. Pat. 61,67,440 (1986);Chem. Abstr., 105 (1986) 5429t. E. H. Sander, Cereal Foods World, 27 (1982) 527; 28 (1983) 257. D. G. Medcalf, Prog. Biotechnol., 1 (1985) 355-362.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I1

333

chemical and technical purposes.678Within the current decennial, a steady interest in dextrins may be deduced from the number of papers published and patents claimed in the world literature. The number of new processes and technological solutions has clearly decreased, with publication of very few new patents on the manufacture of dextrins, but merely improvements of earlier technologies. Any current statistics on the world production of dextrins and their use is lacking. However, it may be realized from more-detailed examination of the decennial information that dextrins are continuously utilized as binders, thickeners, and sizing agents, as well as additives for electrochemical baths for electrolysis and electroplating. However, the role of dextrins as adhesives is decreasing, as they are being steadily replaced by synthetic polymers and polycondensates. The introduction of dextrins into the Federal Register of allowed additives to food has opened up a new field of research. This direction of study has been especially appreciated in Japan, and the majority of patents on the application ofdextrins in nutrition comes from that part of the World. Another area of application in which an increased interest is noted is the use of dextrins as encapsulating agents for retention of the flavor, aroma, and color of processed food, although there is serious competition from cyclomaltodextrins.Increased interest is observed in the manufacture and use of dextrins from the enzymic hydrolysis of starch.

8. Biological Activity of Dextrins The origin and the structure of dextrins do not induce any particular doubts as to their safe application in nutrition. They should be metabolized like other carbohydrates,although the ability of dextrins to afford complexes with certain metal ions, and, possibly, enzymes, might cause temporary perturbations in some processes in living organisms. Thus far, results on the uptake and digestion of dextrins have not suggested any need for rules regarding them as additives and ingredients for food; this is reflected by aforementioned regulation^.^^^-^^^ Indeed, an issue of the United States Environmental Protection Agency876exempted dextrins from the requirement of a tolerance, under the Federal Food, Drug, and CosmeticsAct, when used as a surfactant, suspending agent, or dispersing agent in pesticide and insecticide formulations applied in animals. Moreover, dextrin has been investigated as a potential antiulcer agent. Studiescamed out on four experimental ulcers in rats revealed that, among four polysaccharides applied (dextrin, ginseng pectin, heparin, and chondroitin 4-sulfate), the action of dextrin was the most potent, as shown by a decrease of the level of acid and of (876) U.S. EnvironmentaI Protection Agency, Fed. Regist., 10 Dec. 1986, 5 1 (237) 44,46644,467.

334

PIOTR TOMASIK ef ul.

the activity of pepsin. All four polysaccharides had no influence on the biochemical parameters of pilocarpine-induced gastric secretion.877Comparative studies on the effect of high-fat and high-carbohydrate diets on liver and muscle glycogen repletion in exhaustivelyexercised rats showed that the latter diet is more effectivefor glycogen repletion in liver and skeletal muscle; dextrin is superior to starch in this respect. Probably, the high-carbohydrate diet decreases adipose-tissue l i p o l ~ s i sThis . ~ ~ point ~ of view did not receive support from studies on the metabolism of adipose tissue in sheep fed with grass with the addition of either dextrin or ~-glucose.Both carbohydratesin S~ this effect stronger.879 the diet intensified lipogenesis, and D - ~ ~ U C Omade Studiesconducted on the effect of simple and complex carbohydrateson the arginine utilization of young rats revealed that differences in growth and weight-increase are only dependent on the carbohydrate applied in the case of an arginine-deficient diet. Guar gum, a complex carbohydrate, contributed to thegrowth, whereas sucrose, dextrin, and wheat brain were neutral in this regard.88oAccumulation of liver lipid in rats fed with an excess of L-lysine was lessened when dextrin was added to meals, instead of sucrose.881 The level of serum cholesterol in the organisms of squirrels, spiders, and rhesus monkeys fed with a diet low in saturated fat was lowered when dextrin replaced sucrose.882Dextrin was also found to control a low level of cecal biotin in growing chickens.883 Generally, dextrin is considered safe for human beings; however, some unexpected, but easily rationalized, negative effects have been found. Thus, workers exposed to dextrin dust during production suffer from an increase in dental caries.884 The effect of dextrin on lower organisms has also been studied. Thus, Penaeus monodon Fabricusjuveniles fed with such carbohydratesasdextrin, maltose, sucrose, molasses, cassava starch, corn starch, or sago-palm starch showed the highest mortality with maltose and molasses. The highest sur-

(877) X. Cheng, A. Liu, and B. Wang, YaoxueXuebao,20(1985) 571 -576; Chem. Absfr., 103 (1985) 189,544q. (878) M. Suzuki, S. Saitoh, M. Yashiro, and J. Hariu, J. Nutr. Sci. Vitaminol., 30 (1984) 453 -466. (879) J. Pearce and L. S. Piperova, Comp. Biochem. Physiol., B, 78 (1984) 565-567. (880) E. A. Ulman and H. Fisher, J. Nutr., 113 (1983) 131-137. (881) P. Hevia,E. A.Ulman,F. W.Kari,andW. J.Visek,J. Nutr., llO(1980) 1231-1239. (882) S. R. Srinivasan, B. Radhakrishnamurthy, E. R. Dalferes, Jr., and G. S. Berenson, Atherosclerosis (Berlin),(1980) 359- 364. (883) K. D. Bauer and P. Griminger, Poult. Sci., 59 (1980) 1493- 1498. (884) E. Szponar, Czus. Stomatol., 33 (1980) 857-864.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I1

335

viva1 rate was afforded by sucrose and d e ~ t r i nAmong . ~ ~ ~ carbohydratesfed, namely, D-glucose, sucrose, dextrin, and starch, dextrin provides the highest net protein-retentionin tilapia Oreochromis niloticus.886 In the case of studies on the effect of carbohydrates on the growth and sporulation of Clostridium perfringens,it was found that, depending on the strain, either dextrin or the lower sugars constitute the best energy source.887In the biosynthesis of and R h i z o p ~ s , ~ ~ ~ maltase and glucoamylase by Aspergillus avamorza88~889 maltose and dextrin are the best sources of carbon; other carbohydrates tested were much less effective for this purpose. Dextrin is also an effective source for Saccharomyces diastaticusaWand the mushroom Termitomyces c l y p e a t u ~which , ~ ~ ~produces a single endoxylanase [(1 4)-p~-xylanxylanohydrolase, EC 3.2.1.81. It was, however, shown not only that the source of carbon influences the quality of the enzyme, but that different enzymes may be obtained. Dextrin is a convenient carbon source for the production of neomycin B by Streptomyces m a r i n e n ~ i s . ~ ~ ~

-

VI. THEPYROLYSIS OF STARCH Studies on the pyrolysis of starch have attracted attention since 1913. In a historically first report, B a n t l i r ~determined ~~~ the following yields of products from rice starch decomposedat temperaturesraised within 7 h from 100 to 500": 12%of coke, 30% ofwater, 3% oftar, 5% of aceticacid, 6%of various aldehydes, 1.1 % of ketones, 13%of carbon dioxide, 8% of carbon monoxide, and some hydrogen and ethylene. SandominiM8was the first to study the influence of metal oxides (of Al, Cr, and Zn). He did not observe any appreciable effects of these additives on the decomposition at 270 to 300" of several organic compounds, among them starch and cellulose. On the other hand, essential differences have been noted when pyrolysis is (885) F. Piedad-Pascual,R. M. Coloso, and C. T. Reeicardo,Aquaculture, 31 (1983) 169- 180. (886) J. Anderson, A. J . Jackson, A. J. Matty, and B. S. Capper, Aquaculture, 37 (1984) 303-314. (887) L. E. Sacks, Appl. Environ. Microbiol., 46 (1983) 1169- 1175. (888) L. G. Serichenko, L. S. Losyakova, E. P. Moskvicheva, and A. N. Mazurenko, Fermentn. Spirt. Prom., (8) (1984) 33-35. (889) F. Pestana and F. J. Castillo, MIRCEN J. Appl. Microbiol. Biofechnol., 1 (1985) 225237. (890) L. I. C. De Figueroa and M. R. G. De van Broek, Appl. Microbiol. Biotechnol.,21 (1985) 206 - 209. (891) M. Mukhejee and S. Sengupta, J. Gen. Microbiol., 131 (1985) 1881-1885. (892) P. Ellaiah and C. N. Rao, Indian Drugs,24 (1987) 319-320 Chem. Abstr., 107 (1987) 132,500~.

PIOTR TOMASIK er al.

336

camed out at 600" without a catalyst, as well as with either zinc chloride or sodium hydroxide.893Appreciable decomposition of starch to products of molecular weight lower than that of dextrins can already be observed at 200" at atmospheric pressure.6o8Application of pressure decreases the temperature needed for such decomposition; thus, starch heated in a steel vessel at I55 to 165 yields 5-( hydroxymethyl)-2-furaldehyde,but the yield is Decomposition of starch in the range of 200 to 500" shows two maxima in the yield of gaseous products evolved; they lie at about 250 and 400". The first is due to evolution of carbon dioxide and carbon monoxide (2 : 1), and the second, to formation of methane and unsaturated hydrocarbons. The susceptibility of decomposition depends on the origin of the starch. Among three starches at -250", their stability falls in the order: potato > rice > maize, whereas, at -4OO", the order is potato > maize > rice. The yield of gases evolved at 500" is highest for rice starch, followedby potato, and maize starch. The composition of the gases vanes as shown in Fig. 23. For comparison, the course of decomposition of D-glucoseand sucrose is also presented in this F i g ~ r e . ~The ~,~ P-D-glucosidic '~ bonds are always more stable to heat (by 70- 100°C)than a-D-glucosidic bonds.895 Further studies on the decomposition of starch at various temperatures were conducted by P i ~ o n He . ~ characterized ~ the composition of gases evolved from pea, wheat, rice, and potato starch, as well as potato amylopectin and soluble starch from potatoes, when heated at temperatures from 300 to 1000" (see Table XXIII). It may be seen that there are quite remarkable differences in the pyrolysis of various starches. However, these differences cannot be ascribed solely to the kind, but also to the variety, of plant that produced the starch. The thermal properties of the starch from two different varieties of a given plant differ more than those of starches from two kinds of starch (see Table XXIV).469 Semiquantitativestudies on the yield of volatile products from the pyrolysis of various carbohydrates in the range of 200 to 800" have been cond ~ ~ t e d The . ~ most ~ * information ~ ~ ~ * ~about ~ ~ the pyrolysis of starch was published by Bryce and Greenwood.462.6'6,633,634.695.898 They identified the O

(893) F. Shafizadeh and Y. Z. h i , J. Org. Chem., 37 (1972) 278-284. (894) K. Am, Nippon Nogei Kagaku Kaishi, 10 (1934) 1201 - 1203. (895) L. Dugviliene, L. S. Salbraikh, G. Domburg, and T. N. Skripchenko, Khim. Drev., (1) (1979) 48-55. (896) H. Taj, R. M. Powers, and T. F. Protzman, Anal. Chem., 36 (1964) 108- 1 10. (897) A. Berton, Chim. Anal., 47 (1965) 502-51 I . (898) D. J. Bryce and C. T. Greenwood, J. Appl. Polym. Sci., Pt. C, 2 (1966) 149- 158.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I1

331

1M3 90

->

-> -$

80

70

'D

4

60

0

> W m

c

50

u

3 -0

g

40

IJl

0

s

30 20

10

0

150

203

250

3b

350

Temperature

400

4%

5k

(" C )

FIG.23.-course of Decomposition of Saccharides Presented as the Relationship between Temperatureand Volume of Evolved Gases (CO,, CO, CH,, and C,,H2,). [Potato starch (solid line); maize starch (dashes); and rice starch (points).552]

following 17 species, roughly in the order of their retention times in a gasliquid chromatograph: carbon dioxide, acetaldehyde, methyl formate, furan, propanal, ethyl formate, acetone, acrolein, 2-methylfuraq butanal, butanone, 3-pentanone, 2,5-dimethylfuran, methanol, pentanal, 2-hexanone, and ethanol. Table XXV presents the quantitativeanalysis ofsome of these products from the pyrolysis at 300"of starch (and other carbohydrates for comparison). The yields of these products significantly i n ~ r e a s e ~as~ @ ~ the pyrolysis temperature is raised, even to 800". 1,6-Anhydr0-3,4dideoxy-

PIOTR TOMASIK et al.

338

TABLEXXIII Products of 4rl.olysisq9 of Starch of Various Origins at 300-1OOO" Starch Products

Gas evolved (%) at 300" 400" 500" 600" 700" 800" 900" 1OOo"

co, (%I co (%)

Heavy distillate Solid residue Steam and light vapor

Pea

Wheat

Rice

Potato

8.8 9.6 3.1 3.3 4.3 4.1 4 2.6 25.3 31.2 84.7 4 8.9

8.3 14.6 5.5

5.9 10.4 6.1 6.3 6.9 9 6 4.1 22.5 32.5 14 9 14.3

8.8 13 7.8 7.1 8.2 10.4 1.2 5 22 34

5.5

6.9 1.6 1.3 5.1 25 31 13.5

1.5 14

I0 9.8 15.5

Potato mylopectin 10.9 31.1 16.4

Potato starch (soluble) 11.5 52.1 21.2 22.1 25 31.2 18.1 10.2

15

20.3 18.5 15.4 8.2 22 33 41 11.6 28.6

20.2 22.8 31.5

~-~-g~yceru-hex-3-enopyranos-2-u~ose ("levoglucosenone") was also isolated from the p y r o l y ~ a t e . ~ ~ ~ . ~

The fragmentation of soluble starch in an atmosphere of isobutane in a mass spectrometer revealed the formation of a quasimolecular ion of levoglucosan (m/z 163) which develops ions of m/z 145 and 127 as a consequence of the loss of two water molecules. Fragmentation in an atmosphere of ammonia gives a quasimolecularion of m/z 342,which probably belongs to an anhydromaltose. Fragmentation in an atmosphere of Freon (CF,Cl,) givesgo'a series of negative ions, m/z 359 of anhydromaltose Cl-, m/z 2 15

+

(899) F. Shafizadeh and P. P. S. Chin, ACS Symp. Ser., 39 (1916) 119. (900) D. Gardiner,J. Chem. Soc., C, (1966) 1473- 1476. (901) J. Metzger, Z. Anal. Chem., 308 (1981) 29-30.

THERMAL DECOMPOSITION OF CARBOHYDRATES. I1 TABLEXXIV Gaseous and Liquid Fractions Resulting from Pyrolysisw9 at 1OOO" of Starch from Various Sources Heavy distillate

Volume of gases

Starch pyrolyzate

(%I

(%I

Pea Bean Chestnut (India) (Italy) Maize Banana, green Oat Ricea Wheat" Rye Barley Arrowroot Potatob Rice amylopectin Farina amylopectinc Soluble starch

84.1 82.25 81.3 16.65 11.15 76.5 76.6 17.5-12.4 14.5-73.0 73.85 11.3 69 16.1-48.6 56.5 44.3-4 1 .O 20.2

34.4 42.2 42.3 53.1 48.4 41.9 49.8 50.9-57.6 55.5-58.0

58.2 58.2 63 53.4-121.8 84.8 93.9- 136.4 192.6

a Depending on the variety. Range from four estimations. Determined for 14 varieties. < Determined for 2 varieties.

TABLEXXV Amounts of Volatile ProductsM2from Starch and Related Materials after Pyrolysis for 18 h at 300" Volatile product"

Formaldehyde Acetaldehyde Butanal Acrolein 2-Furaldehyde Acetone Butanone Formic acid Acetic acid Furan 2,5-Dimethylfuran a

Starch

Amylopectin

Amylose

1300 400 10 15 3000 230 85 2600 300 395 35

1400 460

1100

15

15

5

15 2500 335 60

20 3500 90 15

20

210 20

225 10

In 10' mol.g-' of saccharide pyrolyzed.

240

D-Glucose

250

130 15 230 15

339

PlOTR TOMASIK et al.

340

+

+

of glucose C1-, m/z 197 of levoglucosan C1-, m/z 179 of dianhydroglucose C1-, m/z 161 of 5 4 hydroxymethyl)-2-furaldehyde,and m/z 143 of a dianhydroglucose (m/z 179 - HCl). Amylose, as well as amylopectin, pyrolyzed at a pressure of 2.00 to 2.66 kPa at 420" yields four nonvolatile compounds, the yield of which, given in % (in parentheses), does not depend much on the source pyrolyzed. These compounds are 1,6-anhydro-h-~-glucopyranose (28.8 and 24.7), 1,6-anhydro-h-D-glucofuranose(2.0 and 2. l), 1,4 :3,6-dianhydro-a-~-glucopyranose (1.2 and 1.6), and 5-(hydroxymethyl)-2-furaldehyde(0.4 and 0.4) from both pyrolyzed sources.m The mechanisms proposedm for the formation of 1,6-anhydro-P-~-glucopyranose and 1,6-anhydro-~-~-glucofuranose are as follows, and for carbony1 compounds, the mechanism of their formation seems to bew2 as shown. Bryce and Greenwood898studied the kinetics of formation of the major volatile fraction from potato starch, and its components. They limited their interest to the temperature range from 156 to 337" and to the formation of water, as well as of carbon mon- and di-oxide. The results revealed the following facts. Stability toward pyrolysis within the first 20 minutes of the process falls in the order: amylose < starch < amylopectin < cellulose. Autocatalysis is absent, as shown by Puddington.608Both carbon mon- and di-oxide are evolved as a consequence of each of two first-order reactions. The initial one is fast, and the second is slow. The reasons are not well understood, but they probably involve some secondary physical effects. The amount of both carbon oxides is a direct function of the quantity of water produced from any polysaccharide, which, furthermore, is independent of the temperature. The activation energy for the production of carbon monand di-oxide reaches 161.6 kJ/mol, and is practically independent of the polysaccharide formed. At the limiting rates, the approximate ratios of water :carbon dioxide :carbon monoxide were found to be 16 :4 : 1 for amylopectin, 13:3 : 1 for starch, 10:3 : 1 for amylose, and 16 :5 : 1 for cellulose. The same authors695developed studies on the effect of some inorganic salts on the pyrolysis of amylomaize starch at 220 to 340" in a high vacuum. They tested the effect of 2% of the additives mono- and di-sodium phosphate, borax, sodium chloride, sodium hydrogencarbonate, and lithium, potassium, copper(11),calcium, and magnesium chlorides. This effect seems to be due to degradation of individual D-glucosyl units, instead of scission of glucosidic bonds with subsequent formation of anhydro sugars. Each of the salts lowers the threshold temperature of pyrolysis. At the lowest temperature, anionic changes in the additive produced the greatest differencesin the

+

(902) G. A. Byme, D. Gardiner, and F. H. Holmes, J. Appl. Chem.. 16 (1966) 81 -88.

341

THERMAL DECOMPOSITION OF CARBOHYDRATES. I1 C HZOH

Ho

SO HO

OR

1

5

11 CH20H I

'2

HO

2

G

J,

HO

3

OH

1 3 4

OH

8

342

PIOTR TOMASIK et al.

1

5

11

11

H2C-OH /

I

Vj.\.. HO

9

11

il

11 OH

H2C-

/

OH

OH

I

12

10 mode of production of gases. Monosodium phosphate, copper(11)chloride, and magnesium chloride act as catalysts, whereas the others act chemically. Carbohydrate materials present in tobacco leaves produce, on smoking, furan derivativesand cyclic all of which are highly responsible for the flavor and aroma of the smoke. New, more-sophisticated dextrins have become attractive for both theoretical and practical reasons. They are availableeither from chemicallymod-

(903) W. s. Schotzbauer, R. F. Arrendall, and 0. T. Chortyk, Beitr. Tabakforsch. Int., 13 (1985) 74-80.

343

THERMAL DECOMPOSITION OF CARBOHYDRATES. I1

H,O R ”

OH

OH

RO

0I i

Glycolaldehytk

Furan

Glyoxal

Acrolein

Hydroxy pyr uval dehyd e

ified starches by dextrinization, or from dextrins by treatment in various chemical reactions. In some cases, dextrinization and chemical modifications are carried out simultaneously. Dextrin has been applied for grafting of some polymers. Thus 0-(carboxymethy1)cellulose gave, on grafting with dextrin, a polysaccharide gum having unusual rheology.!” Ethylenically unsaturated dextrin prepared by reaction with derivatives of acrylamideW5ymis a durable, hydrophilic photopolymer. The use in cosmeticsof d e ~ t r i n esterified ~ ~ ~ - with ~ ~ fatty ~ acids has already been mentioned. (904) D. J. Sikkema, J. Appl. Polym. Sci., 30 (1985) 3523-3529. (905) A. D. Rousseau and L. W. Reilly, Jr., U.S. Pat. 4,451,613 (1984); Chem. Abstr., 101 (1984) 112,726b. (906) E. A. Fohrenkamm and A. D. Rousseau, US.Pat. 4,5 1 1,646 (1985); Chem.Abstr., 103 (1985) 14,571~.

This Page Intentionally Left Blank

ADVANCES IN CARBOHYDRATE CHEMISTRY AND BIOCHEMISTRY. VOL. 47

THE MACROSTRUCI’URE OF MUCUS GLYCOPROTEINS IN SOLUTION BY STEPHEN E. HARDING Department of Applied Biochemistry and Food Science, University oflvottingham, Sutton Bonington LEI2 5RD, England I. Introduction ........................................................ 11. Composition.. ...................................................... 111. Primary and Secondary Structure: The Basic Unit of the Mucus Glycoprotein . . IV. Tertiary Structure: Assembly of Basic Units. .............................. 1. Thiol Reduction: “Subunits”. ........................................ 2. Branched Models for Much Structure ................................. 3. Mucin Molecular Weights ........................................... 4. Linear Models for Much Structure. ................................... V. The Gross Conformation of Mucus Glycoproteins in Solution ................ Mucins as Polyelectrolytes. ............................................. VI. Mucin Heterogeneity.. ................................................ 1. Polydispersity and Self-AssociationBehavior. ........................... 2. Interactions with Other Macromolecules ............................... VII. Summary and Prospects. ..............................................

345 347 349 352 352 353 356 367

370 373 374 377 380 380

I. INTRODUCTION The importance of mucus glycoproteins in health and disease is undisputed: their biological function is closely related to their conformation, and yet they provide a macromolecular system that is very difficult to analyze. This article describes, in terms of both structural integrity (susceptibilityto degradation phenomena) and physical behavior (very high thermodynamic non-ideality, discrete and quasi-continuous polydispersity, and possible self-association phenomena), how these difficulties have been addressed, focussing particularly on the suitability and difficulties associated with certain physical techniques. The various models for mucus glycoprotein “macrostructure” (namely, assembly, gross conformation, and heterogeneity) in solution -which have been based on various interpretations of the data provided by these techniques- are discussed, and the most likely model is assessed. 345

Copyright 0 1989 by Academic Press, Inc. All rights ofreproductionin any form reserved

346

STEPHEN E. HARDING

Over the past few years, considerable interest has arisen, across a wide spectrum of scientific disciplines, concerning the structure and behavior of mucus glycoproteins or “mucins.” The importance of their role in health and disease is indi~putable,’-~ and yet they provide a heterogeneousmacromolecular system very difficult to a n a l y ~ eFor . ~ convenience,the term “heterogeneity” is used here in its widest sense to describe any system in which the solute species do not have a single value of molecular weight, no matter what the origin of this variation may be. As a result, our knowledge of these molecules is some two decades behind that of other more “fashionable” macromolecules,such as proteins or nucleic acids, wherein the subtleties of individual structure are now quite well understood. The principal difficulty in attempting any form of structural or physicochemical analysis of mucins arises from three fundamental properties (see, for example, Refs. 6 - 10): I, they are highly non-idealin the thermodynamic sense; 2, they are polydisperse(that is, they consist of components of different molecular weight and partial specific volume that are not in chemical equilibrium with each other); and 3, they may be self-associatingin solution (although there is growing evidence to suggest that they are not). Largely because of these difficulties, but also because of sample variability through enzymic and possible mechanical degradation phenomena, there has been some level of disagreement in the literature as to the size, subunit composition, mode of assembly, and gross conformation of the native macromolecule. This was highlighted in a penetrating series of papers which appeared’08 in 1984. The purpose of the present article is to assess critically the literature, and formulate what may be the most likely model for a mucus glycoprotein in solution. Knowledge of the structure and behavior of mucus glycoproteinsin solution is crucial for a proper understandingof their behavior in situ, where they often exist in a more concentratedform. It is these molecules which generally dictate the physical properties of mucus (namely, high viscosity, viscoelastic (1) F. Avery-Jones, Br. Med. Bull. 34 (1978) 1 - 16. (2) A. Allen, Trends Biochem. Sci., 8 (1983) 169-173. (3) A. Allen, in L. R. Johnson (Ed.), Physiology ofthe Gastrointestinal Tract, Raven Press, New York, 1981, pp. 617-639. (4) H. R. P. Miller, J. F. Huntley, and G . R. Wallace, Immunology, 44 (1981) 419-429. (5) 1. Carlstedt,J. K. Sheehan, A. P. Corfield, and J. T. Gallagher,Essays Biochem., 20 (1985) 40-76. (6) S. E. Harding and J. M. Creeth, ZRCSMed. Sci., 10 (1982) 474-475. (7) J. M. Creeth and C. G . Knight, Biochem. J., 105 (1967) 1135- 1145. (8) J. M. Creeth and C. G. Knight, Chem. Soc. Spec. Publ., 23 (1968) 303-313. (9) S. E. Harding,Biochem. J., 219 (1984) 1061 - 1064. (10) S. E. Harding, Biophys. J., 47 (1985) 247-250. (10a) Biochem. Soc. Trans., 12 (1984) 612-621.

MACROSTRUCTURE OF MUCUS GLYCOPROTEINS

347

and gel characteristics1-3,11). Mucins from a wide variety of sources seem to have the same basic physical properties and hence, it is reasonable to assume that they have the same basic structure in solutioni2*13 but vary in their molecular It is not the purpose ofthis articleto discuss in detail the gelation properties of these macromolecules, as consideration of this can be found elsewhere (see, for example, Ref. 12). The glycoproteins found in submandibular and submaxillary secretions will also not be considered. Although these substances are also referred to as mucins, they have somewhat different characteristics (namely, a considerably lower proportion of carbohydrate, shorter carbohydrate chains of different structure and composition) and properties (such as a lower viscosity).16A study has howeveri7shown that they have a greater overall homology with other mucins in terms of assembly and conformation than was perhaps originally presumed. The suitability of certain physical techniques (such as low-speed sedimentation equilibrium in the analytical ultracentrifuge, and electron microscopy) will be assessed and comment will be made regarding the inherent difficultiesof others, such as light-scattering,calibrated gel chromatography, and free-boundary diffision in the ultracentrifuge, interpretations from which have led to some of the disagreements in the literature. Particular emphasis will be laid on the utility of low-speed, sedimentation-equilibrium procedures in the analytical ultracentrifuge for characterizing the molecular size and heterogeneity of mucus glycoproteins, and some developments that make the technique particularly suited for characterizing these molecules will be described. However, before the macrostructure is considered, it may be helpful to review briefly some well established facts about mucin composition and primary structure.18 11. COMPOSITION

-

Native mucus secretions normally contain 1% of saltsand other dialyzable components,0.5 - 1% of proteins, a similar proportion of carbohydrate(11) J. M.Creeth,Br. Med. Bull., 34(1978) 17-24. (1 2) A. Silberbergand F. A. Meyer, in E. N. Chantler,J. B. Elder, and M. Elstein (Eds.),Mucus in Health & Diseuse,Vol. 11, Plenum, New York, 1982, pp. 115- 133. (13) I. Carlstedt and J. K. Sheehan, Biochem. Soc. Trans., 12 (1984) 615-617. (14) F. A. Meyer, Biochim. Biophys. Acta, 493 (1977) 272-282. (15) J. K. Sheehan and I. Carlstedt, Ciba Found. Symp., 109 (1984) 157- 172. I Biol. . Chem., 252 (1977) 3791 -3798. (16) H. D. Hill, J. A. Reynolds, and R. L. Hill, . (17) R. Shogren,A. M. Jamieson, J. Blackwell, and N. Jentoft, Biopolymers. 25 (1986) 150515 17;see also, R. L. Shogren,A. M. Jamieson, J. Blackwell, P. W. Cheng, D. G. Dearhn, and T. F. Boat, ibid., 12 (1983) 1657- 1675. ( 1 8) The present use of the terms “primary,”“secondary,” and “tertiary” structure for mucins does not necessarily correspond to their usage as applied to proteins.

348

STEPHEN E. HARDING

rich glycoprotein, and - 95% of water.I1Although comprising less than 1% of the total mucus secretion, it is this glycoprotein or mucin component which gives the mucus its characteristic high viscosity and viscoelastic characteristics." The mucin component is normally extracted by using a two- (or more)stage density-gradient ultracentrifugation in cesium salts,19the glycoprotein component having a buoyant density (in CsCl) of 1.5 g/mL as compared with - 1.3 g/mL for proteins, and 1.7 g/mL for nucleic acids. Detailed extraction protocols have been g i ~ e n ' ~ in , ~which . ~ ' the importance of meticulous inclusion of protease inhibitors has been clearly demonstrated. A mucus glycoprotein is composed, typically, of - 80% of carbohydrate which, for humans, is restricted to 5 monosaccharides: L-fucose (L-FUC, l), N-acetylgalactosamine(GalNAc, 2), N-acetylglucosamine(GlcNAc, 3),galactose (Gal, 4) and N-acetylneuraminic acid (NeuAc, 5).2,11*22

-

-

a

HO

AcNH

HO

OH

2 1

HO

OH

AcNH

4

3

OH

5 (19) J. M. Creeth, K. R. Bhaskar, J. R. Horton, I. Das, M. T. Lopez-Vidriero, and L. Reid, Biochem. J., 167 (1977) 557-569. (20) I. Carlstedt, H. Lindgren,J. K. Sheehan, U. Ulmsten, and L. Wingerup, Biochem. J., 2 I 1 (1983) 13-22. (21) I. Carlstedt, J. K. Sheehan, U. Ulmsten, and L. Wingerup, in Ref. 12, pp. 273-274. Gfycoconjugutes. ), Vol 1, Academic (22) W. Pigman, in M. I. Horowitz and W. Pigman (as. Press, New York, 1977, p. 132.

MACROSTRUCTURE OF MUCUS GLYCOPROTEINS

349

TABLEI Amino Acid CompositionU(Mo1/100 Mol) of 2 Mucus Glycoproteins from an Ovarian Cyst (“603” and “485”) ~

Aminoacid

603

485

ASP Thr Ser Glu

4.9 19.0 16.3 4.7 15.1 8.5 12.7

2.8 27.1 18.1 4.1 16.0 5.8 9.6

nda

nda

4.3 0.3 1.5 2.1 0.7 1.0 3.2 1.9 3.2

3.7 0.4 1.8 1.9 0.5 1.0 2.8 1.3 2.6

7.6

12.0

Pro

GIY Ala CYS Val Met Ile

Leu TYr Phe His LYS Arg Total peptide content (%) n.d., not determined.

The protein moiety is considered to consist of a single polypeptide chain about which the carbohydrate is built. Approximately one in every three residues is either L-serine or L-threonine, and the 0-3 atoms of these provide the sites for glycosidic linkage. Understandably, therefore, amino acid composition data (see Table I) reveal a large proportion of L-serine and L-threonine. The significance of the large percentage of L-proline will be discussed in the following Section. 111. PRIMARY AND SECONDARY STRUCTURE: THEBASICUNITOF THE

Mucus GLYCOPROTEIN The molecular weights of mucus glycoproteins range from -0.5 X lo6to 16.0 X lo6,and it is now widely accepted that the mucins of higher molecular weight are made up12,22.24,25 of multiplesofa basic unit having a molecular (23) J. M. Creeth, B. Cooper, A. S . R. Donald and J. R. Clamp, Biochem. J., 211 (1983) 323-332. (24) S. E. Harding, J. M. Creeth, and A. J. Rowe, Proc. Znt. ConJ Glycoconjugutes, 7th, Olson-Reklambyra, Sweden, 1983, pp. 558-559. (25) A. Silberberg, Biorheobgy, 24( 1987) 605-614.

350

STEPHEN E. HARDING

FIG.1.-Schematic Mucin Basic Unit.2234[The continuous line represents the polypeptide, and the attached chains, the oligosaccharides.Although two regions devoid of carbohydrate are shown, there may be only one.]

weight of (0.4-0.6) X lo6:protease digestion (and thiol reduction, but see later) of a wide range of mucins generallyproduces materials having molecular weights of this order.12,25-27 The basic nit^^,^^*^^ ofa mucin is a single polypeptide chain which consists of two (or three) distinct regions, namely, one heavily glycosylated central core and one (or two) end peptide segments which are rich in cysteine and acidic groups but virtually devoid of carbohydrate (see Fig. 1). The central glycosylated region is rich in serine, threonine, and proline, with many multi-branched oligosaccharides ranging in length from 5 to 30 residues. Insofar as “secondary structure” is concerned, there seems to be no evidence of regularly folded structures normally associated with globular proteins (ahelices, j3 sheets, and the like). The polypeptide is, however, presumed, for three reasons, to adopt a loosely coiled structure. Firstly, it accounts for the high levels of proline in these macromolecules (see Fig. 2). A prolyl residue is most commonly found in the trans configuration in polypeptides: from minimum-energy considerations,28there are two allowed conformations, a compact form (w = - 55”) and an extended form (w = 145”). The relative ease with which an isolated prolyl residue can adopt the compact form makes it ideal for producing bends and turns in the polypeptide backbone.28An alternative explanation for the large proportions of proline could be its use in the formation of helical and fibrous structures. However, there is no evidence for helical structures (in mucins) corresponding to polymers or copolymers of proline as found in, for example, collagen.

+

(26) M. Scawen and A. Allen, Biochem. Soc. Trans., 3 (1975) 1107- 1109. (27) J. M. Creeth, unpublished results. (28) P. R. Schimmel and P. J. Flory, J. Mol. Biol., 34 (1968) 105- 120.

MACROSTRUCTURE OF MUCUS GLYCOPROTEINS

35 1

FIG. 2.-Schematic Diagram of a Portion of a Polypeptide Chain Containing an Isolated Proline Residue (from Refs. 28 and 29). [An isolated proline residue can produce a “kink” or “turn” in a polypeptide chain, if the C-c“ bond angle (v)is --55”.]

A second reason supporting a loosely coiled structure for the basic unit is that such a structure would be more efficient at engulfing and immobilizing local solvent than an extended form.24.29 The third line of support for a coiled domain is from hydrodynamic data on T - d o m a i n ~These . ~ ~ are formed by protease (namely, trypsin)digestion of the native mucin and are essentially equivalent to a basic unit but are lacking in naked peptide. Values for the ratio of the sedimentation concentration regression coefficient,&, to the intrinsic viscosity, [q],are close to the value

GalNAc GlcNAc Gal Fuc NeuAc FIG. 3.-Variability12 of Carbohydrate Side-Chain Composition for Eleven Mucins (Relative to GalNAc).

(29) C. R. Cantor and P. R. Schimmel,Biophysical Chemistry. Part I, Freeman, San Francisco, 1980, p. 270. (30) S. E. Harding, A. J. Rowe, and J. M. Creeth, Eiochem. J., 209 (1983) 893-896.

352

STEPHEN E. HARDING

expected for a spheroidal or randomly coiled molecule (see, for example, Refs. 31 and 32). Although amino acid composition data are relatively invariant from mucin to mucin, much greater variability seems to exist as far as the carbohydrate moiety is concernedI2: Fig. 3 compares composition data for a number of different mucins normalized to the relative amounts of GalNAc present in a linkage. The first sugar in the chain is always GalNAc, and the chains are often terminated by L-fucose or NeuAc, with at least one branch per chain on average. It is now well established, from, for example,the results of gas- liquid chromatography (g.1.c.) and n.m.r. spectroscopy, that considerable microheterogeneity,both in length and complexity, exists among the side chain^.^^-^^ However, as Silberberg and MeyerI2 pointed out, mucus function may not necessarily depend upon a specific sequence in the oligosaccharide side-chains, as the protein moiety contains the more-specific featuresneeded to build up a supramolecularstructure,and thus produce the requisite physical properties. IV. TERTIARY STRUCTURE: ASSEMBLY OF BASICUNITS

How are the basic units arranged to form the macrostructure of the larger mucins? Evidently, because of the nature of the degradation products of pronase digestion, the linkage is by way of the naked peptide regions. 1. Thiol Reduction: “Subunits”

It was originally suggested26that the links between the basic units are through intermolecular disulfide bridging between cysteine residues in the naked end regions. Reduction of mucins by thiols produced the M, -500,000 forms (see, for example, Refs. 26 and 36). However, Creeth27 observed a variety of forms in the 0.5-2.0 X lo6 region, and subsequent observations on cervical and other mucins by Carlstedt and Sheehan’3J5,31yielded, in the presence of guanidine hydrochloride (Gum HCl), species having forms M, -2 X lo6 which they referred to as (31) J. K. Sheehan and I. Carlstedt, Biochem. J., 217 (1984) 93- 101. (32) J. M. Creeth and C. G. Knight, Biochim. Biophys. Acta, 102 (1965) 549-558. (33) H. Van Halbeek, L. Dorland, J. F. G .Vliegenthart,W. E. Hull, G. Larnblin, M. Lhermitte, A. Boersma, and P. Roussel, Eur. J. Biochem., 127 (1982) 7-20. (34) H. Van Halbeek, L. Dorland, J. F. G. Vliegenthart, J. Montreuil, B. Fournet, and K. Schmid, J. Biol. Chem., 256 (1981) 5588-5590. (35) G . Larnblin, M. Lhermitte, A. Klein, P. Roussel, H. Van Halbeek, and J. F. G. Vliegenthart, Biochem. SOC.Trans., 12 (1984) 599-600. (36) A. S. Mall, D. A. Hutton, R. M. Coan, L. A. Sellers,and A. Allen, Biochem. Soc. Trans., 16 (1988) 585-586.

MACROSTRUCTURE OF MUCUS GLYCOPROTEINS

353

TABLEI1 Terminology for Mucus Glycoprotein Components Component

Description

Basic unit

basic mucin building block, M, 5 X lo5;glycosylated centralcore region plus 1 or 2 regions of naked peptide product remaining after treatmentof much with proteases; equivalentto a basic unit minus naked peptide product remaining after reduction of mucin by thiols, whatever form this may take

T-domain Subunit

-

“subunits.” Table I1 distinguishes between “basic unit” and “subunit”; henceforth in this article, the term subunit refers to the macromolecular entity produced by reduction of the native mucin by thiols, whatever the value of MIfor this entity may take. Subsequent action of trypsin produced forms having MI -0.3-0.4 X lo6 that are referred to as T - d ~ m a i n s . ’ ~ * ’ ~ * ~ ’ These workers argued that the earlier, low M, values for the subunit could have been a result of proteolytic or mechanical degradation phenomena occumng during and after the extractionprocess, thus strengtheningthe case for the meticulous inclusion of protease inhibitors. It would appear, therefore, that the linkage between basic units (equivalent to T-domains plus naked peptide) is not exclusively through disulfide bridge attachment, but rather by peptide, or some other, linkage. Mall and c~workers’~ demonstrated, however, that pig gastric mucin, extracted in the presence of inhibitors but analyzed in the absence of Gu HC1, gives a value of 500,000 for . ~ ’the presence the molecularweight of the subunit. It has been ~ l a i m e d ~that of Gu HCl produces irreversible aggregationphenomena leadingto anomalously high values (-2 X lo6) for the subunit, although this observation conflicts with data for whole bronchial r n ~ c i n (see s ~ ~Section IV,2). Silberberg25demonstrated that the linkages between basic units or subunits could be lectin-like: intramolecular disulfide bridging could stabilize the naked peptide into a particular conformation which forms a binding site for a specific, but unusual, sugar sequence on the side chain of the adjacent unit. Whether the units are linked into branched or linear arrays has been the subject of much debate which will now be considered.

-

-

2. Branched Models for Mucin Structure One ofthe first proposalsas to the form ofthis tertiary structure, at least for one particular mucin, was a branched structure, referred to as a “star” or (37)A. Allen, A. Bell, M. Mantle, and J. P. Pearson, in Ref. 12, pp. I 15- 133. (38) S.E. Harding and J. M. Creeth, Biochim. Biophys. Acta, 746 (1983)114- 119.

STEPHEN E. HARDING

354

M r

- 2 x106

Cleavage by pepsin to degraded subunits

..

/b

subunit

[Mr-.5

‘YO

protein core

”

Mi-70000 protein joined by one or more disulfide bridges to each subunit

1061

-5

8.’

Reduction by 0.2 M 2-rnercaptoethanol t o subunit and Mr- 70000 protein

v

.

FIG.4.-Branched Model for Pig Gastric

“windmill” formz*3*37 (see Fig. 4). This model was based on several years of exhaustive work on pig-gastric mucin (p.g.m.) by Allen, Pain, and in which a spheroidal model for the gross conformation was successfully predicted,40 in agreement with others.14*15,24*30,31 Their model for the assembly was supported by (i) the susceptibilityofthe mucin to attack by proteases or thiols4’; (ii) estimates of molecular weights for the native and reduced pig-gastric mucinz*26; (iii) end-group amino acid analysis of the components3’; and (iv) the discovery of a 70,000 molecular weight, link protein?’ A similar protein (M, 1 10,000) was discovered by the Forstners and coworkers42for mucus glycoprotein from small intestines: these workers also considered where this protein could be located in both branched and linear models. Using the value of -2 X lo6 obtained for the molecular weight of the native molecule, and 0.5 X lo6 for the reduced form, it was concluded that

-

(39) (40) (41) (42)

M. Mantle, D. Mantle, and A. Allen, Biochem. J., 195 (1981) 267-275. A. Allen, R. H. Pain, and T. Robson, Nature. 264 (1976) 88-89. J. P. Pearson, A. Allen, and C. W. Venables, Gastroenterology,78 (1980) 709-715. M. Mantle, M. Potier, G. G. Forstner, and J. F. Forstner, Biochim. Biophys. Acta, 88 1 (1986) 248-257.

MACROSTRUCTURE OF MUCUS GLYCOPROTEINS

355

there must be four basic units. Treatment with 2-mercaptoethanolyielded the forms having a molecular weight of -0.5 X lo6;also released was a protein of molecular weight 70,000, as judged from sodium dodecyl sulfate-gel electrophoresis. A branched or "star" model was proposed as a likely form of assembly,26and this was later modified to accommodate the M, = 70,000 protein in the center.'"' This model has been used by other w0rkers.4~ From results on bovine-cervical, human-ear, and bronchial mucus, Meyer and Silberberp suggested very tentatively that this four-unit form may be a rather general building-block for other mucins. Although, as a first estimate, this model provided a good fit to the data then available, subsequent analyses suggested that the branched form is unlikely, for and other mucins. Crucial to the model for p.g.m. is the assump tion of a four basic-unit form and a molecular weight of 2 X lo6.Molecular weight values for p.g.m. vary widely, depending on the technique used and the preparation procedure. For example, for p.g.m. extracted by using 6 M Gu HCl and proteinase inhibitors, Creeth and Co~per,"~ using lowspeed sedimentationequilibrium,obtained a molecular weight of 9 X 106, not 2 X lo6. Carlstedt and Sheehan13obtained values for the M, of p.g.m. as high as 45 X lo6(in the presence of 6 MGu HCI) by using light-scattering,and they indicated that the earlier, lower value of 2 X lo6 was a result of inadequate precautions against protease or mechanical degradation. On the other hand (and parallel to the discussions over the size of subunits; see Section IV, l), Allen and c ~ w o r k e r sargued ~ ~ * ~that the higher values of the M, for native p.g.m. and other mucins represented non-covalently bound aggregates, a result of some irreversible, anomalous association caused by the presence (or subsequent removar7) of 6 M Gu HCl. However, this effect on p.g.m. has not been reproduced by others,13and no such aggregation (or dissociation) phenomenon has been observed for bronchial and ovarian-cyst m ~ c i n s . ~ ~ The effect of possible mechanical disruption of covalent linkages during high-shear solubilization of p.g.m. mucins from the gel state (and especially without the presence of 6 M Gu HCl) was also demonstrated by Carlstedt and Sheehan13for p.g.m. It has been pointed out,"*however, that this is not a problem for other large glycoconjugates, such as proteoglycans.

-

-

-

-

-

-

(43) (44) (45) (46)

V. A, Bloomfield, Biopolymers, 22 (1983) 2141-2154. F. A. Meyer and A. Silberberg, Ciba Found. Symp., 54 (1978) 203-218. J. M. Creeth and B. Cooper, Biochem. Soc. Trans., 12 (1984) 618-621. A. Allen, D. A. Hutton, D. Mantle, and R. H. Pain, Biochem. Soc. Trans., 12 (1984)

6 12-6 15. (47) D. Snary, A. Allen, and R. H.Pain, Biochem. J., 141 (1974) 641 -646. (48) J. E. Fitzgerald, G . G . R. Green, F. W. Stafford, J. P. Birchall, and J. P. Pearson, Clin. Chim. Acta, 169 (1987) 281 -298.

356

STEPHEN E. HARDING

Although difficulties in sample integrity would appear to be the most likely explanation for discrepancies in the measured size of mucins, problems associated with the methodology of the physical techniques employed may also have contributed to them. For example, the low values of molecucould possibly be explained by some lar weight obtained difficulties in the particular method employed, difficulties manifested by correlating distributions of sedimentation coefficienP9with distributions of molecular weight? for flexible, linear polymers, M, is not a linear function of the sedimentationcoefficient,s, but rather,50M, s2, so the mean value of s would not necessarily correspond to the mean M,. Absolute values for molecular weights were usually obtained from the Svedberg equation51*52 by combining measurements of s with the (translational) diffusion coefficient, D, measured by free-boundary spreading in the analytical ultracentrifuge (see, for example, Ref. 52). This procedure has inherent difficulties when applied to these substances. Diffusion measurements on polydisperse materials are difficult to interpret, and the broad molecular-weight distribution of the slowly diffusing mucins makes it a possibility that high-molecular-weightmaterial is sedimented out of solution during the long time-periods used, although no significant losses have thus far been observed5%in the majority of cases.

3. Much Molecular Weights What, then, is a suitable method for determiningthe molecular weight of a mucin? Because molecular weight is such an important parameter in the evaluation of mucin macrostructure,some of the other procedures that have been used, and the difficulties and possible pitfalls encountered, will be considered here. a. Light-Scattering.-In a series of papers on c e r v i ~ a l ' and ~~~ other ~-~~ mucins (see, for example, Ref. 56), Carlstedt and Sheehan also used the Svedberg equation. However, they measured the (Z-average)diffusion coefficient, D,, in a different way, by quasi-elastic light-scattering(q.1.s.). They also used total-intensity light-scattering, where the intensity scattered by a (49) R. H. Pain, Symp. SOC.Exp. Biol., 34 (1980) 359-376. (50) C. F. Tanford, Physical Chemistry of Macromolecules, Wiley, New York, 1961, p. 382. (5 I ) T. Svedberg and K. 0. Pedersen, The Ulfrucenfrt$uge,Oxford University Press, 1940. (52) Ref. 50, p. 380. (52a) R. H. Pain, personal communication. (53) I. Carlstedt, H. Lindgren, and J. K. Sheehan, Biochem. J., 213 (1983) 427-435. (54) J. K. Sheehan and I. Carlstedt, Biochem. J., 217 (1984) 93-101. (55) J. K. Sheehan and I. Carlstedt, in Ref. 24, pp. 599-600. (56) I. Carlstedt and J. K. Sheehan, in Ref. 24, pp. 580-581.

MACROSTRUCTURE OF MUCUS GLYCOPROTEINS

357

glycoprotein solution is measured as a function of the concentration and the angle; using a biaxial, extrapolation procedure to zero angle and zero concentration (Zimm plot), it is possible to obtain values for the weight-average molecular weight M, and the root-mean-square radius, &. Both of these light-scattering procedures have found very wide application to a variety of macromolecular systems, but it is fair to say that they are extremely difficult to apply to such heterogeneoussystems as mucins, largely because of problems of dust and traces of large aggregates, particularly for measurements at low angles. In the case of q.1.s. to minimize these effects, an angle of 90" is often employeds7;although this does not lead to any appreciable error for rigid, non-spherical particles, extrapolation to zero angle is normally necessary for the flexible mucins, because of the possible finite contribution to the observed autocorrelation data from rotational diffusion p h e n ~ m e n aUnfortunately, .~~ at low angles, any supramolecularaggregates will seriously affect the extrapolation, and, as a result, molecular-weight values can be on the high side, depending on the extent of contamination. The Zimm-plot technique, like q.l.s., also involves a difficult extrapolation to zero angle. Another difficulty is that its application assumes that the particles are Rayleigh - Gans - Debye scattererss8(namely, that there is no change of phase or other distortionsofthe incident radiation by the particle). For the larger mucins in particular, this may not be the case. Apparent agreement between Zimm plots and the Svedberg equation (using D, values measured by q.1.s.) can be misleading, in that the same effects producing high M, values (and high & values) from the Zimm method would also contribute to lower D, values (and, hence, higher M, values from the Svedberg equation). The dramatic influence of even small proportions of aggregates on the results from q.1.s. has been clearly demonstrated by, for example, Preston and coworkerss9in related studies on proteoglycans. Another difficulty that is often not reported is the contribution to error caused by concentration measurement (of the unsolvated solute); concentrations can rarely be measured to better than 5%, and will contribute error in both the Zimm plot and the values for the refractive increment used for evaluation of the constant. If q.1.s. and the Svedbergequation are used, errors in concentration will also be manifested in the extrapolations of the diffusion and sedimentation coefficients. (57) R. E. Godfrey, P. Johnson, and C. J. Stanley, in D. B. Sattelle, W. I. Lee, and B. R. Ware (Eds.), Biomedical Applications ofLaser Light Scattering, Elsevier, Amsterdam, 1982, pp. 373-389. (58) For larger particles, see, for example, S. H. Chen, M. Holz, and P. Tartaglia,Appl. Opt., 16 (1977) 187-194. (59) G . S. Harper, W. D. Comper, and B. N. Preston, Biopolymers, 24 (1985) 2165-2173.

358

STEPHEN E. HARDING

Because of these constraints, light-scatteringtechniques should not, where possible, be the method of choice: were light-scatteringto be used, confirmation of results by using an independentprocedure would be desirable. Agreement with molecular weights from, for example, low-speed sedimentation equilibrium would also give greater confidence in other potentially useful parameters from light-scattering, such as the equivalent Stokes radius, r,, from q.l.s., and also the root-mean-square radius, &, and the second thermodynamic virial coefficient, B, from Zimm plots. It should also be pointed out that light-scattering may be the only absolute method applicable to species having molecular weights larger than 15 X 1O6 (the upper limit for accurate measurements from low-speed sedimentation equilibrium procedures, unless ultra-short solution columns are used). Despite these difficulties, and only after the employment of meticulous preparative procedures,’s reproducible results were obtained by Carlstedt and Sheehan for T-domains, subunits, and whole mucins, which appear to be in agreement with other data. For whole pig-gastric mucin, for example, they obtained a molecular weight of -45 X lo6, some 20 times the value reported by Allen and coworkers.36Lower values ( 15 - 40 X 106)have also been ~ b t a i n e d ,presumably ~~,~~ because of sample variability. Using lowspeed sedimentation equilibrium, Creeth and Cooper45obtained a lower value (-9 X lo6).

-

b. Relative Techniques.- Relative techniques for the determination of molecular weight, such as gel electrophoresis and gel-permeation chromatography (g.p.c.), also have their difficulties. They are referred to as “relative techniques” for macromolecular molecular weight analysis, because they require calibration using standards of known molecular weight. G.p.c. is useful for giving a qualitative indication of size distribution (after adequate correction for diffusion broadening) but, because of difficultiesof obtaining standards of similar size and conformation, results can only be relative. An adequate calibration procedure is, however, now available that avoids the problem of inappropriate standards by combining g.p.c. with low-speed sedimentation equilibrium measurements.60,61 An important development, not yet applied to mucins, is the availability of an instrument (Wyatt Technology, Santa Barbara, CA, U.S.A.) facilitatingg.p.c. measurements on-line to a multi-angle, laser light-scatteringdetector. This serves two purposes; it 1, provides an on-line “clarification” of macromolecular solutions prior to light-scattering, overcoming the principal difficulty already referred to in (60) A. Ball, S. E. Harding, and J. R. Mitchell, In?. J. Biol. Macromol.. 10 (1988) 259-264. (6 1 ) S. E. Harding, in G. 0.Phillips, D. Wedlock,and P. Williams(Eds.), Gums & Stabilisers in theFoodIndustry, Vol IV, IRL Press, Oxford, 1988, pp. 15-23.

MACROSTRUCTURE OF MUCUS GLYCOPROTEINS

359

FIG.5.- Rayleigh Interference Profiles from Low-speed Sedimentation Equilibrium. [(a) Native much (bronchial gp 376); (b) reduced gp 376; (c) immunoglobulin G. In (a), note the steep rise of the fringes near the cell base, but the finite slope at the meniscus.]

Section IV,3a, and 2, facilitates a direct (and absolute) visualization of molecular-weight distributions. With gel electrophoresisas a quantitative tool for size determination, the problem is more serious: unlike nucleic acids, mucins do not have a natural uniform charge :length ratio, and furthermore, unlike for unglycosylated proteins, sodium dodecyl sulfate does not bind uniformly. The technique appears to have some use, however, as a probe for possible mucin - protein intera~tion.’~ c. Low-speed Sedimentation Equilibrium.- Arguably the most powerful technique for measuring mucin molecular weights (provided that sample molecular weights are 5 15 X lo6)is the technique of low-speed sedimentation equilibrium using Rayleigh interference optics: the inherent effect of the ultracentrifugal field can be put to particular use in helping resolve the components of a heterogeneoussystem, provided that the effects of thermodynamic non-ideality can be properly taken into account. The low- or intermediate-62-64 speed procedure is normally the method of choice, where the speed is sufficiently low to ensure adequate resolution of the fringes near the cell base.62Fig. 5 gives a comparisonof Rayleigh interferencepatterns for (62) J. M. Creeth and S. E. Harding, J. Biochem. Biophys. Methods, 7 (1982) 25-34. (63) D. C. Teller, T. A. Horbett, E. G. Richards, and H. K. Schachman, Ann. N. Y.Acad. Sci., 164(1969)66-101. (64) H. K. Schachman, Ultracentrifugation in Biochemistry, Academic Press,New York, 1958.

360

STEPHEN E. HARDING 1.5

3.0

I

I

I

-

*

-C

**

-

-1.0

** -1.5

0

4

1.0

~

*

-

**

-

7

-7

-C

I

**

*

t

**

-

**

**

b 1

I

I

0

1

1.0

t

3.0

* z ax

s d 1

0.5

0.0 0

E

0

1.0

1

1

1

1

c (mg/mL)

1

3.0

FIG.6.-Low-speed Sedimentation Equilibrium of Mucus Glycoproteins. [(a) I d vs. plot for the chronic bronchitic, bronchial m u c h BM GRE, M,- 6 X 106 (Ref. 9); (b) I d vs.
<

-

+, -

-

a native mucin, a reduced mucin, and immunoglobulinG at sedimentation equilibrium under these conditions: even before full data analysis, the steep rise at the cell base, but finite slope at the meniscus for the native mucin (see Fig. 5a), suggest the presence of considerable heterogeneity. At equilibrium, the concentration at the air - solvent meniscus remains finite, but can be found without too much difficultyby mathematical manipulation of the fringe data.62A typical plot of In J vs. where J is the absolute concentration (in fringe numbers) and = (r2 - a2)/(b2- a2) [r being the radial displacement from the center ofthe rotor, and a and b the correspond-

<

<,

MACROSTRUCTURE OF MUCUS GLYCOPROTEINS

36 I

ing positions of the meniscus and cell base, respectively] for a bronchial mucin (BM Gre, M,= 6.0 X lo6)from a chronic bronchitis patient is given in Fig. 6a. The pronounced upward curvature is symptomaticof heterogeneity. In some cases, the effects of thermodynamic nonideality can obscure the effects of heterogeneity, to give a linear, pseudo-ideal plot, as shown in Fig. 6b for a bronchial mucin (CF PHI) from a cystic fibrosis patient. Weight-average molecular weights for particular, whole-solute distributions, MGcan be readily obtained by using a particularly directly determinable point average62(the “star average,” M*), without the requirement for an independent estimate for the initial concentration. M* is defined, for solute at a given radial position, r, by: r

\

2 RT 02(1-Vp)’

M*(r) =

(1)

where j is the concentration in fringe numbers (relative to that at the meniscus), J, is the meniscus fringe number, and the other parameters have their usual meaning. A plot of M* vs. is given for CF PHI in Fig. 6c. ME is obtained from the limiting value of M* at the cell base,62and several values : of mucins obtained in this way, together with other values using for the M different techniques, are included in Table III.’3,14J0~35~~~47~56,65-69 The partial specific volume, V,(for mucins, normally in the range of 0.60 to 0.65 mL/g) has to be determined separately: V is usually obtained’O from a plot of solution density versus concentration. The accuracy with which this value can be obtained (compare Section IV,3,a for refractive increment measurements) will largely be affected by errors in concentration measurement, although the final value can be checked by calculation by using the Traube rule (see, for example, Ref. 7 l), if the amino acid and carbohydrate composition of the mucin are known. Other useful molecular-weight information is readily obtainable: for example, point weight average molecular weights, M,, can be obtained by

<

(65) J. M. Creeth, B. Cooper, A. S. R. Donald, and J. R. Clamp, Biochem. J., 211 (1983) 323-332. (66) P. A. Feldhoff, V. P. Bhavanandan, and E. A. Davidson, Biochemistry, 18 (1979) 24302434. (67) R. A. Gibbons and F. A. Glover, Biochem. J., 73 (1959) 217-225. (68) T. K. S . Mukkur, D. L. Watson, K. S. Saini, and A. K. Lascelles, Biochem. J., 229 (1985) 4 19-428. (69) J. P. Pearson, R. Kaura, W. Taylor, and A. Allen Biochim. Biophys. Actu, 706 (1982) 221-228. (70) 0. Kratky, H. Leopold, and H. Stabinger Methods Enzymol., 2 7 (1973) 98- 110. (71) R. A. Gibbons, in A. Gonschalk (Ed.), Glycoproteins: Their Composition, Structure & Function, Vol 5 , Part A, Elsevier, Amsterdam, 1972, pp. 31 - 157.

TABLE I11 Molecular Weights of Mucus Glycoproteins GIycoprotein

Buffer

I‘

Ovarian cyst (human) 603/AmSsolc

phosphate-chloride

0.10

485C 603/43- 5@

sodium acetate phosphate - chloride

0.10 0.10

+M NaCl + 6 MGu.HCI

sodium acetate

phosphate-chloride

0.10

0.10

+M NaCl +6 MGu.HCI

+M NaCl Chronic bronchitic “BM GRE”

0.10

Technique

References

low-speed sedimentation equilibrium

31

0.90 0.84 0.85

1 1 1 1 2 2 2

0.41 1.14

I 2

8

2.4 1 2.25 2.18

3-4 3-4 3-4

31 31 31

0.56 0.51 0.55 0.58

+MNaCI +6 MGu.HCI 603/43-5@ (pronase-treated) 316 Bronchial (human) Cystic fibrosis “CFF’HI”

d

10-‘XM,

+0.4 M CaCI, +5 mgl fucose/mL + 5 me/ GlcNAc/mL +6 MGu*HCl

1.80

3-4

23 31 31 31

ultrashort-column sedimentation equilibrium low-speed sedimentation equilibrium

30

9

6.2

11-12

6.0

11-12

9

6.2 5.5

11-12 11-12

9 9

Chronic bronchitic “BM GRE” (fractions 13- 15)

0.10

Chronic bronchitic “BM GRE’ (fractions 13- 15, thiol-reduced) Chronic bronchitic Asthmatic (lowdensity fraction) (highdensity fraction) Cervical (human)

phosphate

phosphate- chloride

10- 11 10- 11

+6 MGuaHCl

1.1 18

2 40

5.6 4.6

10- 11 9- 10

10.6 2.1

20 3-4

+6 MGu*HCI

W o\ W

Trypsindigested lithium citrate (citric acid) phosphate-bufferedsaline

Bovine (pronasedigested) (pronase

5.3 5.0

phosphate-buffered saline

reduced

Bovine (estrus) Bovine (pregnancy) Bovine Bovine (reduced)

+6 MGu-HCI

+ reduced)

1.6 0.29 0.38 4.16 3.93 16.4 5.2 4.6

3-4 1 1 8 8 35 10 10

0.84 0.50 0.47 0.29

1 1 1

8.7

45 45

light-scattering high-speed sedimentation equilibrium

light-scattering high-speed sedimentation equilibrium

45 13

66 66 54 54

sedimentation-diffusion

54 54 54 67 67 14 14 14

1

light-scattering sedimentation-diffusion light-scattering sedimentation-diffusion

14 14 14 14

18

low-speed sedimentation equilibrium

45

light-scattering intrinsic viscosity light-scattering

Gastric (pig)

phosphate-chloride

0.10

(continued

TABLE I11 Molecular Weights of Mucus Glycoproteins (continued) Glycoprotein

Buffer

reduced papain-digested

reduced Small intestine (human) native papain-digested

native sheep Gall bladder (human) papain-digested a

I'

10+'XMr

n6

Technique

References

2.1

4

46,47

2.3 45.0 39

4 70

sedimentation-diffusion low-speed sedimentation equilibrium light-scattering sedimentation- diffusion

0.5

1

0.5 2.5

I 4-5

1.7 0.24

3-4 I

5.0

10- I 1

0.5

1

light-scattering

46,47 15,56 56 46 46 56

c.g.c.d c.g.c.d high-speed sedimentation equilibrium (absorption optics)

38 38

c.g.c.d

69

68

I = ionic strength. n = approximate number of basic units, based on molecular weight of basic unit of 0.5 X 106. Protease inhibitors not used in extraction procedure. c.g.c. = calibrated gel chromatography.

MACROSTRUCTURE OF MUCUS GLYCOPROTEINS

365

using sliding strip fits to the log fringe concentration versus radial displacement squared plots,g and an example for two different loading concentrations, co, of CF PHI is given in Fig. 6d. A potential pitfall if sedimentationequilibrium is used is a failure to allow properly for the effects of thermodynamic non-ideality, which tends to diminish measured molecular weights and to mask heterogeneity. The effects on M: are normally minimized by using the lowest possible loading concentration (as low as 0.2 mg/mL if 30-mm path-length cells are employed). Alternatively, point weight averages can be extrapolated to zero concentration (J = 0), to yield a value independent of non-ideal or associative phenomena. The ideal value obtained in such a way may, however, be biased towards the lower end of the molecular weight distribution, but this bias can be minimized by using short solution columns and low speeds, and, in extreme cases, by extrapolating the value so obtained to zero gravitational field. Another way of coping with non-ideality is to combine weight average values with number, z-, and higher-order average^^'.^^ if the precision in the data justifies this: it is possible to obtain number-average and Z-average whole cell average molecular weights (although these are less precise) and also point (M, and M,) and compound molecular weights (My,,Myz):

My2and My2point averages are free from first-order, non-ideality effects: measured values are, however, generally reliable only if a laser-light source can be employed to generate the interference fringes, or if accurate on- or off-line data-capture procedures are a ~ a i l a b l e . ~ ~ - ~ ~ High-speed or meniscus-depletion equilibrium techniques have been attempted by employing scanning absorption o p t i ~ s , and ~ ~ by - ~using ~ Rayleigh interferen~e.~~.~' Several values have been included for comparative purposes in Table 111. In this method, the meniscus region is essentially depleted of sample (namely, zero concentration), and this simplifies the D. Roark and D. A. Yphantis Ann. N. Y. Acad. Sci., 164 (1969) 245-278. D. C. Teller, Methods Enzymol., 27 (1973) 346-441. S. E. Harding and A. J. Rowe, Optics & Lasers in Engineering, 8 (1988) 83-96. S. E. Hardingand A. J. Rowe, in J. Tyrer andG. T. Reid (Eds.), Automatic FringeAnalysis, Open Tech. Press, Loughborough, U.K., 1987, pp. 187-200. (76) S. E. Harding and A. J. Rowe, Biochem. SOC.Trans., 15 (1987) 1046- 1047. (77) H. Woodward, B. Horsey, V. P. Bhavanadan, and E. A. Davidson, Biochemistry, 21

(72) (73) (74) (75)

(1982) 694-701.

366

STEPHEN E. HARDING

interpretation of the data if Rayleigh interference optics is used, because there is no need to measure meniscus concentrations. However, no real advantages are accrued by attempting meniscus depletion if absorption optics are used, because this treatment gives a direct record of meniscus absorbance. In addition, for such polydisperse materials as m u c h , it is normally not possible to deplete the meniscus properly without losing optical registration at the cell base.37,62 A common pitfall is to assume depletion conditions when this is clearly not valid.78 A further difficulty in attempting meniscusdepletion conditions with polydisperse materials is that the effective, thermodynamic, second virial coefficient, B, can be greatly e n h a n ~ e d ~ , ~ ~ ? B,@=B(l +PM2/12+- - -), (3) B being the “static” virial coefficient,8A = ( 1 - Vp)& (b2 - a2)/2RT,p the solvent density, and o the angular velocity. This “speed dependence” effect can be minimized by low speeds and short solution columns.8i80A problem peculiar to absorption optics (applicable to protein systems in general) is the difficulty in obtaining an accurate baseline due to, for example, anomalous absorption onto cell window^^'-^^: this difficulty is often ignored in the literature. Rayleigh interference optics is by far the most suitable optical system. Furthermore, if a laser-light source is used, as opposed to the mercury-arc sources commercially supplied, fringe resolution can be greatly e n h a n ~ e d .This, ~ ~ - together ~~ with the possibility of using a commerciallaser gel-~canner,~~ the Ultroscan (LKF3 Instruments, Bromma), for data capture, and a simple Fourier a l g ~ r i t h m for ~ ~data - ~ ~analysis (enabling fringe displacements to be measured to an accuracy of 1/300th of a fringe, even with fringes generated by a mercury-arc light-source) opens up new possibilities

(78) J. Hinnie and A. Serafini-Fracassini,Biopolymers, 25 (1986) 1095- 1107. (79) H. Fujita, Foundations of UltracentrifugeAnalysis, Wiley, New York, 1975, Chap. 5. (80) H. Suzuki, in B. Ranby (Ed.),Physical Chemistry of Colloids and Macromolecules, Blackwell Scientific, Oxford, 1987, pp. 101- 109. (8 1 ) S . P. Spragg, The PhysicalBehaviourofMacromoleculeswith Biological Functions.Wiley, New York, 1980, pp. 66-68. (82) A. J. Rowe, Protein EnzymeBiochem., 106 (1984) 1-37. (83) P. H. Lloyd, Optical Methods in Ultracentrifugation,Electrophoresis and Difliion, Oxford University Press, 1974, pp. 86-87. (84) R. C. Williams, Methods Enzymol., 48 (1978) 185-191. (85) T. M. h u e and D. A. Yphantis, Biophys. J., 25 (1979) 164. (86) D. A. Yphantis, Methods Enzymol., 6 1 ( 1979) 3 - 12. (87) R. C. Williams, Anal. Biochem., 48 (1972) 164- 171. (88) T. M. h u e , D. A. Yphantis, and D. G . Rhodes, Anal. Biochem., 143 (1984) 103- 112.

MACROSTRUCTURE OF MUCUS GLYCOPROTEINS

367

for detailed analysis of the nature of the heterogeneity and thermodynamic non-ideality.10*89-93 4. Linear Models for Mucin Structure A model alternative to the branched assembly proposed earlier is one in which the basic units of the mucin are assembled in linear arrays. The first direct evidence for this structure came from electron microscopy. Initial progress was made by Slayter and c o w ~ r k e r s , using ~ - ~ ~the technique of rotational platinum-shadowingafter air-drying bronchial and other mucin solutions onto mica. This procedure revealed a variety of conformations: in the main, linear structureswere seen, with no evidence of branched forms. However, these linear structuresgenerally lay in parallel arrays which could well have resulted from the high shearingand surface-tension forces on airdrying which may have uncoiled them from their true solution-conformation. On the other hand, this procedure did have the advantage of revealing the probable nature of the assembly (namely, linear). In a subsequent study on bronchial mucins, in which extraction procedures were used to minimize the effect of proteases, flexible, linear forms were again made visible, although, if negative staining with uranyl acetate was employed, the material appeared as aggregate^.^^ Poorly defined images using negative staining were also reported for ovariancyst and cystic-fibrosis bronchial mucins.99Other workers also demonstrated linear forms on using both rotary shadowinglmJo1and staining procedures.lolJoZ (89) H. Kim, R. C. Deonier, and J. W. Williams, Chem. Rev., 77 (1977) 659-690. (90) P. W. Chun, S. J. Kim, J. D. Williams, W. T.Cope., L. H. Tang, and E. T. Adams, Jr., Biopolymers, 1 1 (1972) 197-214. (9 1) M. L. Johnson, J. J. Correia, D. A. Yphantis, and H. R. Halvomn, Biophys. J.,36 (198 1) 575-588. (92) S. H. Tindall and K. C. Aune, Anal. Biochem., 120 (1982) 71 -84. (93) B. K. Milthorpe, L. W. Nichol, and P. D. Jeffrey, Biophys. Chem., 3 (1975) 169. (94) G. Lamblin, M. Lhermitte, P. P. Roussel, and H. S. Slayter,Biochimie, 6 1 (1979) 23 -43. (95) H. S. Slayter, A. G. Cooper, and M. C. Brown, Biochemistty, 13 (1974) 3365-3371. (96) H. S. Slayter, in J. R. Harris (Ed.), Electron Microscopy ofProfeins, Vol. 1, Academic Press, New York, 1981, pp. 197-254. (97) P. Roussel, G. Lamblin, N. Houdret, M. Lhermitte, and H.S. Slayter, Biochem. Soc. Trans., 12 (1984) 617-618. (98) H. S. Slayter, G. Lamblin, A. LeTreut, C. Galabert, N. Houdret, P. Degand, and P. Roussel, Eur. J. Biochem., 142 (1984) 209-218. (99) S. E. Harding, A. J. Rowe, and J. M. Creeth, Biochem. J., 209 (1983) 893-896. (100) A. Mikkelsen, B. T. Stokke, B. E. Christensen, and A. Elgsaeter, Biopolymers, 24 (1985) 1683- 1704. (101) J. K. Sheehan, K. Oates, and I. Carlstedt, Biochem. J., 239 (1986) 147- 153. ( 102) A. 0.Jenssen, 0. Harbitz, and 0. Smidsred, Eur. J. Resp. Dis., 6 1 (1 980) 7 1 - 76.

368

STEPHEN E. HARDING

B

C

D 0.2 1

0

1

2

3

4

Contour length (grn)

FIG.7.- Electron Microscopy of Mucins. [(A) Unidirectionally platinum-shadowedCF PHI macromolecule, after airdrying onto mica (from Ref. 30). a, “low profile areas; b, linear extensions. (B) Unidirectionally shadowed pig gastric mu& ( E M )macromolecule after spreading on 6 M guanidine hydrochloride and air-drying.lol(C) CF PHI, unidirectionally platinum-shadowed after fixation, and critical-point drying.IM (D) Distribution of contour lengths for PGM.I0I]

MACROSTRUCTURE OF MUCUS GLYCOPROTEINS

369

Applications of similar, metal-replica preparation procedures to a bronchial mucin from a cystic fibrosis patientw also revealed images of linear assemblies (see Fig. 7A). However, instead of using rotary shadowing techniques, unidirectional shadowing was used. Low-profile areas, with linear extensions, were seen; the low-profile areas could clearly be derived from highly swollen particles of low asymmetry which flattened onto the mica upon drying; such structures would not be as clearly visible on using rotary although similar, “puddled” structures have been observed.94 The linear extensionslinking these “low-profile” areas do not appear to be flattened, possibly because they contain less carbohydrate,and hence are less swollen through solvation prior to drying. It has been tentatively suggested that the “naked” regions linking the basic units/T-domains could be located within these “low glycosylation” e~tensions.2~ The form shown in Fig. 7A would therefore correspond to a mucin molecule consisting of three basicunit forms. It should be stressed, however, that single, nonglycosylated, polypeptide chains having no apparent secondary structure would not be visible with the resolution used here. On the other hand, there could be a small amount of residual glycosylation in this region, insufficient, at least in patches, to protect from protease attack, but sufficient to allow visibilization by unidirectional shadowing. This suggestion is, however, highly speculative, and may be disproved or confirmed by using stains specific for carbohydrate-free regions.Io3 Much ofthe earlierwork on mucins using electron rnicrosc~py~~-~’ was on relatively small macromolecules (M, 2 X 106). A comprehensive study by Sheehan and coworkerdo’on larger bronchial, cervical, and gastric mucins, using a variety of preparative procedures, appears to leave little doubt that the general mode of assembly for mucins from a variety of sources must be linear. Nodular forms in the flexible linear chains were often, but not always, made visible. Indeed, Fig. 7B’O’ shows an image of a p.g.m. molecule using a preparation procedure similar to that for Fig. 7A, but after spreading on 6 M Gu * HCl; no evidencefor any nodular or low-profileareas is apparent. These workers,Io1in agreement with other^,^*.^ also observed considerable polydispersity in all their samples; for example, Fig. 7D’O’ gives the distribution of contour lengths measured for p.g.m., which is of the same log-normal form as the predicted molecular-weight distribution reported by Creeth and Cooper45for this mucin from low-speed sedimentation equilibrium (see Section VI,1 and Figs. 10 and 12). The original postulate of a linear model for basic unit/T-domain and subunit assembly in ~olution’~-’~ was based on light-scatteringstudies on the

-

(103) J. K. Sheehan and I. Carlstedt, Proc Znt. Symp. Glycoconjugates, 9th. 1989 (in press).

370

STEPHEN E. HARDING

X’

n h

FIG. 8.-Linear Model for Cervical Mucin, Consisting of Basic Units or “T-Domains” Linked by Naked Peptide Regions, Some by Disulfide Bridging. [Arrowed regions are susceptible to thiol attack. From Ref. 53. Key: T, T-domain; X or X’, “basic unit”; and Y,“subunit.”]

high-molecular-weight cervical mucin (see Fig. 8); for example, the Burchard p parameterlWwas used, and this had been proposed as a measure of the extent of branching in a polymer; it is obtained as the ratio of & to the Stokes radius r, (from q.1.s. diffusion-coefficient measurements): the values obtainedHfor whole mucins and subunits ( 1.8 and 1.6, respectively) appear consistent only with a linear assembly. A similar, light-scattering study performed on submaxillary mucins (which have side-chain compositions and structures different from those of other mucins) gave very similar results,17 suggesting that there is a close homology of these with other mucins. V. THEGROSS CONFORMATION OF Mucus GLYCOPROTEINS IN SOLUTION

What domains do these mucin assemblies described herein occupy in solution? How do they interact with each other and with other molecules to form the mucosal barrier? In answer to the first question, there is growing evidence to suggest that they occupy enormouslyexpanded, and thus, at low concentration (51%), readily overlapping, spheroidal domains in solution. Although the electron-microscopic technique of airdrying onto mica prior to platinum shadowing has yielded valuable information about the mode of assembly or secondary structure, it can yield misleading information about the gross conformation in solution, because of, for example, the very high shearing and surface-tension forces encountered by these highly solvated structures on air-drying onto mica. For example, it could be inferred from some earlier studiesw-%that, in the native form, bronchial (104) W. Burchard, M. Schmidt, and W. H. Stockmayer, Macromolecules. 13 (1980) 1265-

1272.

MACROSTRUCTURE OF MUCUS GLYCOPROTEINS

37 1

mucins behave as semi-flexiblerods. However, parallel arrays of such rods are often seen, strongly suggestive of alignment through shearing. An attempt has been made to circumvent this problem by using glycerol prior to drying,lo5giving images of many complex conformations, both globular and linear. However, the effect on the native conformation of replacing the large proportion of solvent normally associated with mucins with concentrated glycerol remains open to question. Hallett and coworkedMused, instead, a critical-point drying procedure with prior fixation using glutaraldehyde. Images of a cystic-fibrosis, bronchial glycoprotein using such a procedure (see Fig. 7C106)yielded large, roughly spherical particles of dimensions very reasonably consistent with their prediction. Using a value of 2 X lo6 for the molecular weight, a value for the swelling due to solvation, vJV [where v, is the solvated specific volume and V is the partial (anhydrous) specific volume] of 100 times was obtained.lMOnly with less-complete fixation was any evidence of elongated structures apparent. Again, however, questions could be raised as to how "native" the conformation remains after fixation. The strongest evidence for a solvated, loosely coiled domain comes from hydrodynamics and related techniques: I . Estimation of the thermodynamic, second virial coefficientby using the technique of ultrashort-column sedimentationequilibrium under conditions where charge effects have been largely suppressed gives a value that is consistentwith a molecular expansion of - 100in solution.8*30 2. Values for both the intrinsic viscosity, [q],and the concentration-dependence regression-coefficientof sedimentation, k, (corrected for radial dilution effects and solution density; see, for example, Ref. 107)are very high.8*300J837'JM High [q]or high k, can arise from asymmetry or expansion, but the ratio k,/[q] has been shown by several workers to be a function of shape alone.'07-' The calculated value of - 1.4- 1.6 for ovarian-cyst' and cervical rnucid4appearsto be consistentwith a coiled, spherical domain. Sheehan and C a r l ~ t e d obtained t~~ similar values for subunits and T-domains. 3. Double log plots of $, D,, s, and [q]vs.molecular weight, M, ,for whole mucins, subunits,and T-domainsgave slopesconsistent with a coiled s t r u ~ t u r e(see ~ ~ Fig. . ~ ~9).

-

-

''

(105) M. C. Rose, W. A. Voter, C. F. Brown, and B. Kaufman, Biochem. J., 222 (1984) 37 1 - 317. (106) P. Hallett, A. J. Rowe, and S. E. Harding, Biochem. Soc. Trans., 12 (1984) 878-879. (107) S. E. Harding and P. Johnson, Biochem. J., 231 (1985) 543-548. (108) M. Wales and K. E. van Holde, J. Polym. Sci., 14 (1954) 81 -86. (109) P. Y. Cheng and H. Schachman, J. Polym. Sci., 16 (1955) 19-30. ( I 10) J. M. Creeth and C. G. Knight, Biochim. Biophys. Acfu, 102 (1965) 549-558. ( I 11) A. J. Rowe, Biopolyrners, 16 (1977) 2595-261 1.

372

STEPHEN E. HARDING

b

6

7

log M,

FIG.9.-“Double Log Plots” of (a) Intrinsic Viscosity, (b) the Reciprocal of the Diffusion Coefficient, and (c) Sedimentation Coefficient Data versus Molecular Weight for Human Cervical M u c i n ~[Key: . ~ ~ 0 and 0, “whole” mucins; Wand 0, “subunits”; and A,T-domains. Molecular weights determined from Zimm plots (filled symbols)or the Svedbergequation using QLS (open symbols). Values for the slopes are in all cases consistent with a randomcoil model and not with a rigid sphere or a rod.]

Besides being enormously expanded and spheroidal, the mucins also appear very flexible: evidence for this derives from the ease with which these molecules are deformed with shear, as shown in streaming-birefringence e~periments,~*J’~ from the concentration dependence of viscosity,’ and the temperature-dependence of sedimentation coefficients.’ Mikkelsen and cow o r k e r P confirmed these observations by using the dependence on field strength of electric birefringence relaxation-phenomena. Evidence for local flexibility in the carbohydrate side-chains has come from, for example, n.m.r.-spectral s t ~ d i e s . ~ ~ ,l4~ ~ , ’ Finally, the existence of mucins in solution as highly expanded, spheroidal domains is also consistent with their biological function as space-filling l 3 9 l

(1 12) R. A. Gibbons and F. A. Glover, Biochem. J.. 73 (1959) 217-225. (1 13) M. Barrett-Bee, G. Bedford, and P. Loftus, Biosci. Rep., 2 (1987) 257-263. ( 1 14) M. Barrett-Bee, G. Bedford, and P. Loftus, in Ref. 12, pp. 107- 1 1 1.

MACROSTRUCTURE OF MUCUS GLYCOPROTEINS

373

-

water immobilizer^'-^: at relatively low concentrations (below l%), the domains normally overlap,lL5corresponding to the onset of gelation.'1,115 Mucins as Polyelectrolytes One important feature of mucins is their behavior as polyelectrolytes; for example, this has been postulated as a possible control-mechanism in the periodic changes in cervical mucus.116This behavior is a result of the relatively large proportion of acidic sugar residues (NeuAc) normally present. In viscosity studies (at a concentration of 2.0 mg/mL), these effects can normally be suppressed at ionic strengths 1.0 M, as demonstrated for bronchial and ovarian-cyst glycoproteins by Harding and Creeth37using either NaCl or Gu HCl. No deleterious effectson the glycoprotein by these compounds (up to I = 6.0 M for Gus HC1) were observed, consistent with the observations of Snary and coworker^.^^,^^' The effects of NaCl and Gu * HC1were indistinguishable,with only a modest conformationalchange in the range I = 0.1 - 1.O M , a typical polyelectrolyte effect.IL8The effects (including those of 6 M Gu HCl) were found to be fully re~ersible.~~ These observations imply a degree of flexibility in the macrostructure, a conclusion in agreement with other physical properties indicating a loose, highly solvated, greatly expanded structure. It was also observed37that a relatively low ionic strength (I = 0.1 M) is sufficient to suppressthe primary salt effect, so that correct molecular-weight information can be obtained at I = 0.1 Mif low concentrationsare employed in sedimentation-equilibrium measurements. However, those physical measurements that are especially sensitive to non-ideality should be performed at ionic strengths not less than 0.5 M, unless extrapolationsto infinite dilution can be employed.37 Tam and Verdugo116observed that changes in pH over the range of 6.5 8.0 sharply affect the swelling of cervical mucus, and suggested that changes in the polyelectrolyte nature of the gelatinous, mucin framework (for a Donnan equilibrium) are responsible. It was later pointed out, however,37 that changes in the mucin moiety are not likely to be directly responsible, as only the small number of histidine residues (and not the NeuAc) could possibly be titrated in this range. The protein content of the secretion probably cannot be ignored in this instance.

-

-

-

( 1 15) ( 1 16) ( 1 17) ( 1 18)

J. M. Creeth, Mod. Prob. Pued., 19 (1988) 34-45. P. Y. Tam and P. Verdugo, Nature, 292 (1981) 340-342. D. Snary, A. Allen, and R. H. Pain, Eur. J. Biochem., 24 ( 197 I ) 183 - 189. L. H. Kent, B. R. Record, and R. G. Wallace, Phil. Trans. R. SOC.,Ser. A, 250 (1957) 1-43.

STEPHEN E. HARDING

374

VI. MUCINHETEROGENEITY

Mucins are highly heterogeneoussubstances(see Table IV), on at least two counts. 1. They are polydisperse in a quasi-continuous sense (arising from variability in the carbohydrate c o m p o s i t i ~ n . ~ ~2.~They ~ ~ ~ are ~~~~~~~). polydisperse in a discrete sense, arising from variability in the numbers of basic units.1°J2 Furthermore, they are known to interact in solution with other components of mucus secretions, such as lysozyme,120J21 and, in addition, they were originally considered to self-associatein solution,6.122although subsequent evidence suggested otherwise?*45A further complication is that they are extremely non-ideal in the thermodynamic sense, arising from the very high solvent-association. Table V'*8,30,123-125 compares the nonideality parameter, BM,, where B is the thermodynamic second virial coefficient, for mucins with that for other macromolecules.Finally, besides being heterogeneous with respect to molecular weight,126they are heterogeneous with respect to density.12' One of the most powerful techniques for the investigation of macromolecular heterogeneity is the analytical ultracentrifuge, in particular, sedimentaTABLEIV Terminology for Mucus Glycoprotein Heterogeneity Term Heterogeneity

Polydispersity Primary polydispersity Secondary polydispersity

Definition Any macromolecular system where the solute species do not have a single value for the molecular weight, no matter what the origin of the variation may be. Can include polydispersity or self-association behavior. The presence of non-interacting components in a macromolecular system of different molecular weight or density, or both. Quasicontinuous distribution of much molecular weights arising from variability in carbohydrate sidechain composition. Discrete distribution of mucin molecular weights arising from vaxiation in numbers of basic units.

( 1 19) P. Roussel, P. Degand, G. Lamblin, A. Laine, and J. J. Lafitte, Lung, 154 (1978) 241 260. ( 120) J. M. Creeth, J. L. Bridge, and J. R. Horton, Biochem. J.. 18 1 (1 979) 7 17 - 724. ( 1 2 1) A. 0. Jenssen, Ph.D. Thesis, Univ. Trondheim, Norway, 1980. ( 122) J. M. Creeth, Biochem. Soc. Trans., 8 (I 980) 520- 52 1. (123) C. F. Tanford, Ref. 50, p. 196. (124) S. E. Harding, Int. J. Biol. Mucromol., 3 (198 1) 341 -342. (125) C. H. Emes and A. J. Rowe, Biochim. Biophys. Actu, 537 (1978) 110- 124. (126) R. A. Gibbons, Br. Med. Bull., 34 (1978) 34-38. (127) J. M. Creeth and J. R. Horton, Biochem. J., 161 (1977) 449-463.

MACROSTRUCTURE OF MUCUS GLYCOPROTEINS

375

TABLEV Comparative Thermodynamic Non-ideality of Macromolecular Solutions Particle

BM, (mL * g-')

References

Spherical unhydrated protein' Hemoglobin Myosin GP 376 CF PHI

2.9 4.8 52.6 300 300

123 124 125 7,s 30

a

Based on a partial specific volume of 0.73 mL-g-l.

tion equilibrium, especially when it is used in conjunction with such other techniques as electron microscopy. Heterogeneity can be characterized by combining whole-cell average molecular weights (M:, Mt, and MP) in the form of the Herdan relation^^^*,^^^:

where a, and a, are respectively the number and weight-average standard deviations ofthe distribution,whatever form this may take, or by using ideal, molecular-weight moments see Eq. 2 and Ref. 72) and other diagnostic combinations of point average M, , M,, and M, values.73 There have been many theoretical attempts to characterize heterogeneity by using quasi-elasticlight-scattering(see, for example, Ref. 130), and sizedistribution, software packages are available from commercial manufact u r e r ~ .The ~ ~ 'so-called 'polydispersity factor,' PF, namely, the Z-averaged, normalized variance of the distribution of diffusion coefficient^,^^^^^^^ has proved a useful guide to the homogeneity of a sample. However, because q.1.s. is so sensitive to declarification from dust and aggregates (see, for example, Refs. 57 and 59), a realistic application of the polydispersity factor and the more-sophisticated, size-distribution a n a l y s e ~ ' ~in J ~order ' to obtain quantitative information about mucin heterogeneity appears to lie in the future. Furthermore, the analysesdeveloped do not distinguish between self-associationand genuine polydispersity, and do not, in general, take into (128) J. M. Creeth and R. H. Pain, Prog. Biophys. Mol. Biol.. 17 (1967) 217-287. (129) G. Herdan, Nature, 163 (1949) 139. (130) S. W. Provencher, Mukromol. Chem., 180 (1979) 201 -209. (1 3 1) See,for example, Malvern Instruments Bulletin PBo41 2-85 (1985). (132) D. E. Koppel, J. Chem. Phys., 11 (1972) 4814-4820. ( 1 33) P. Pusey, in H. Z. Cummins and E. R. Pike (Eds.),Photon CorrelationundLight Beating Spectroscopy, Plenum, New York, 1974, p. 387-428.

STEPHEN E. HARDING

376

0

0

10

20

30

50

40

60

20,w

0

I

I

I

2

4

6

I 1

8

FIG. 10.-(a) Distribution of SedimentationCoefficients for Pig Gastric Mucin (from Ref. 49). (b) Corresponding, Estimated Distribution of Molecular Weights, Using the Relation s a M0.5.

considerationthermodynamic non-ideality;the approximation is also made that the diffusing species have similar conformations. Nevertheless, as a relative technique for demonstrating changes in a mucin sample (for example, detecting an interaction), q.1.s. would appear to be a useful tool.’” Another useful way of visibilizingheterogeneity is by using sedimentation velocity and using the approximation that all sedimenting specieshave similar conformations: for example, Pain and coworker^^*^^ used a procedure (and only after several additional approximation^)^^^^^^^ for calculating sedi(134) M. Smedley, PhD Thesis, Brighton Polytechnic, 1987. (135) V. C. Hascall and S. W. Sadjera, J. Biol. Chem., 244 (1969) 2384-2396. (136) H. Fujita, Foundations of UltracentrgugeAnalysis, Wiley, New York, 1975, pp. 166206.

MACROSTRUCTURE OF MUCUS GLYCOPROTEINS

377

mentation coefficient(s) distributions to demonstrate heterogeneity of pig gastric mucin (see Fig. 10a). To demonstrate what this could mean in terms of a size distribution, the assumption (valid for flexible polymers) is here used that M, a s2, and that an s value of 33 X s is approximately equivalent to a molecular weightMof 2.5 million, to generate the equivalent molecular-weightdistribution of Fig. 1Ob. The form of the latter distribution is similar to that of the distribution of contour lengths shown in Fig. 7D. 1. Polydispersity and Self-Association Behavior

It is normally impossible to distinguish between the effects of self-association and polydispersity (that is, non-interactingcomponents of different M t or density, or both) from a single sedimentation-equilibrium experiment (see, for example, Ref. 137). However, it is known, both from the apparent nature of the biosynthetic process” and from combining the results of more than one sedimentation-equilibrium experiment using the diagnostic technique of non-overlap of plots of point weight average molecular weight versus concentration plots (see Fig. 6D) for different loading concentrat i o n ~ ~that . ’ ~ polydispersity in mucins is significant. Although it was origithere is now nally considered that mucins might also be self-associating,6,’22 increasingevidence that this is not significant, at least in dilute solution: this

FIG. 1 I.-Rayleigh Equilibrium Interference PatternsP for BM GRE in (a) PhosphateChloride Buffer Containing 0.4 M CsCI; (b) Phosphate-Chloride Buffer + 5 mg of Fucose/ml; and (c) Phosphate-Chloride Buffer 5 mg of GlcNAc /mL.

+

(137) J. M. Creeth and S. E. Harding, Biochem. J., 205 (1982) 639-641.

318

STEPHEN E. HARDING

conclusion was reached from the results of sedimentation-equilibrium experiments wherein potential sites of association were competitively blocked with i n h i b i t ~ r s . ~ * ~ ~ There are two possible sites for self-association: on the glycosylated regions, there may be hydrophobic patches arising from possible localized clusters of L-fucose or GlcNAc residues. Such sites would be competitively blocked by having a swampingconcentrationof ~-fucoseor GlcNAc present in the solvent. However, in such an experiment using a multi-channel cell, there is little difference between the Rayleigh fringe distributions (see Fig. 1 l), or the calculated values for MC (all -6 X 106; see Table 111), for a glycoprotein, from the bronchial secretion ora chronic bronchitis patient, in the presence of L-fucose or GlcNAc, compared with non-diswciating solvent. A similar experiment performed in 6 M Gu HC1 to block any hydrophobic sites on the naked peptide regions again yielded an MC of 6 X 106: there is therefore little evidence of self-association in this mucin. These observations were supported by similar experiments on a related bronchial mucin and pig gastric m u ~ i n , 4and ~ also lower-molecular-weightmucin~.~' The observed heterogeneity of mucins in dilute solution would appear, therefore, to be due mainly to polydispersity, and not to self-association phenomena: it could be inferred from this conclusion that the gel-forming tendency of mucins is mainly achieved through molecular overlap of their huge, swollen domains, and not by virtue of association phenomena. To characterize the solute distributionsobserved at sedimentation equilibrium in terms of actual molecular-weight distributions is, however, not easy, largely because of the difficulties introduced by the presence of the high thermodynamicnon-ideality. An indirect procedure for modelling the polydispersity of p.g.m. was used by Creeth and Cooper45by assumingthe solute distributionto be the same as if it were an isodesmic, self-associationprocess (see Fig. 12). This makes use of the principle that a reacting system cannot be distinguished, in a single experiment, from a non-reacting system that contains the same distribution of molecular weights. The form of the distribution agrees well with the distribution of contour lengths from electron microscopylol(see Fig. 7D),and also the form of Fig. lob calculated from sedimentation-velocitydata, although the actual values from Fig. lob are lower (see Section IV,2). A direct procedure for handling non-ideal, polydispersedistributions has now been derived,'O and has been applied, after making certain assumptions, with some success to discrete distributions of molecular weights for bronchial mucins. Assuming that the observed polydispersity of a cystic fibrosis glycoprotein (CF PHI) was due entirely to variation in the numbers of mucin basic units, it was possible to obtain representations of the observed solute distributions by using a three-component, non-ideal fit (see Fig. 13), based

-

MACROSTRUCTURE OF MUCUS GLYCOPROTEINS

0.16

r

379

M i = 8.85

0.12

0.04

0 6

0

12

18

24

30

FIG.I2.-Molecular Weight Distribution for Pig Gastic M~c in:~Calculated by Using the Constants for a Hypothetical Isodesmic Association. [The value marked by an arrow corresponds to the weight-average molecular weight for the whole solute distribution, ML]

on the earlier observations from electron microscopy.30Because of the relatively large amount of computer time required to achieve such fits (even on a large, mainframecomputer), it is not practical at the present time to perform (arguably, more representative)log-normal distribution fits, because of scaling problems, at present insurmountable, caused by the very large range in molecular weights. The algorithms are now available, however, and it is hoped that it would be possible to fit to such distributions as soon as the hardware is improved.

0 ' 0

I

I

1.0 c

<

4

FIG. I3.-Plot of I d vs. for Bronchial Mucin CF PHI.'O [Lines fitted correspond to a three-component fit (of molecular weights, 1.2 X 106, 1.8 X 106, and 2.4 X 106) for various values of the thermodynamic, second vinal coefficient, B X 10-4 mL. mol. s-~].

380

STEPHEN E. HARDING

Point average molecular-weight distributionsalso provide a very valuable insight into mucin polydispersity, although this, like the direct method, normally requires data ofthe highest quality (for example, obtained by using a laser light-source, and accurate data-capture and fringe-analysis procedures; compare Section IV,3). Again, as noted in Section IV,3 mention should be made of two methods, not yet applied to mucins, involvingg.p.c. linked with an “absolute” molecular-weight technique to give molecularweight distributions: 1. “off-line” calibration of g.p.c. columns using lowspeed sedimentation equilibrium60s61 and 2. “on-line” calibration of g.p.c. using multi-angle laser light-scattering.137a

2. Interactions with Other Macromolecules Although the mucins (which constitute at least 30%of the non-water part of mucus) largely dictate the physical properties of whole mucus, their interactions with other components play an important role. For example, such interactions are considered to play an important cause of changes occumng in cervical mucus during the estral cycle.37,’l6 Although it has been suggested that immunoglobulin IgA may be involved in the stabilization of mucus gels,138~139perhaps the clearest evidence for mucin - protein interaction is for bronchial mucin with lysozyme. This has been demonstrated by Creeth and coworkers120by using sedimentation-equilibriumtechniques. Rheological techniques have also proved to constitute a powerful probe for such interactions: the viscometric data of Jenssen and coworkers12L,140 supported the observations of Creeth and associates,12oand also suggested the interaction to be electrostatic in nature. A rheological investigation by Smedley’” demonstrated a calcium ion-dependent interaction of alginate secreted from Pseudomonus aeruginosu with bronchial mucins. The use of various polymer systems, including alginates with intestinal mucins, for increasing the transit time of drug-carrier systems14’is currently being explored. VII. SUMMARY AND PROSPECTS Mucins from a wide range of sources show the same basic structural properties: a heavily glycosylated basic unit of M, -0.5 X lo6, linked by covalent (or other sufficiently long-lived) bonds into linear arrays that are polydisperse with respect to numbers of basic units (discrete polydispersity) (137a) P. J. Wyatt, D. L. Hicks, C. Jackson, and G . K. Wyatt, Am. Lab.. June, 1988. (138) J. R. Clamp, Biochem. SOC.Trans., (1977) 1579-1581. (139) E. Puchelle, J. M. Zahm, and R. Havey, Bull. Physio-Puthol. Respir., 9 (1973) 237. (140) 0. Harbitz, A. 0. Jenssen, and 0.Smidsred, Eur. J. Resp. Dis., 65 (1984) 512-520. (141) M. F. Anderson, S. E. Harding, and S. S. Davis, Biochem. Soc. Trans., 1989 (in press).

MACROSTRUCTURE OF MUCUS GLYCOPROTEINS

38 I

and extent of glycosylation (quasi-continuous polydispersity). In solution, they occupy highly expanded and flexible, spheroidal domains (space-filling at concentrations5 1%), and behave as classicalpolyelectrolytes. Self-association phenomena in solution does not appear to be significant, but bronchial mucins have been shown to interact with such other components of mucus as lysozyme and alginates. Some of the questions that remain open, such as whether the apparently similar molecular architecture between mucus glycoprotein from different sources and species is purely fortuitous or has a definite genetic basis, will be answered once detailed information concerning the peptide core of the “basic unit,” the gene or genes which code for it, and the post-translational processes responsible for complex glycosylation and sulfation is available. Progress is continuallybeing made in all these areas, usingdirect methods for deglycosylatingthe peptide core,142 and using mRNA probes. 143,144 For example, for the related case of submaxillary glycoproteins, evidence for tandemly repeated sequences of amino acid residues has been presented from studying apomucin cDNA.lU We can reasonably look forward to the expectation that the growing information from these genetic probes into the basic unit structure, when coupled with detailed information from the physical probes concerning the gross conformation, polyelectrolyte behavior, and heterogeneity of the glycoprotein macrostructures, will provide us with a clear understanding of the structural- functional relationships of mucins in health and disease. ACKNOWLEDGMENTS The author thanks Dr. J. M. Creeth, Dr. R. H. Pain, Professor A. Allen, Dr. J. K. Sheehan, and Dr. I. Carlstedt for their comments and suggestions, and other colleagueswho have made helpful comments. The support of the Lister Institute of Preventive Medicine is greatly appreciated.

( 1 42) N. Mian, K. Marks, J. G. Widdicombe, J. R. Davies, and P. Richardson, Biochem. SOC.

Trans., 14(1986) 114-115. (143) R. Nichols, D. M. Carlson, and J. E. Dixon, Biochem. Biophys. Res. Commun., 149 (1987) 244-248. (144) C. S. Timpte, A. E. Eckhardt, J. L. Abernathy, and R. L. Hill, J. Biol. Chem., 263 (1988) 1081-1088.

This Page Intentionally Left Blank

AUTHOR INDEX Numbers in parentheses are footnote reference numbers and indicate that a n author’s work is referred to although the name is not Cited in the text.

A

Abagyan, G. V., 278,280(477), 305(477) Abd El-Thalouth, I., 3 15 Abdel Akher, M., 205 Abe, A., 57, 104(56) Abernathy, J. L., 38 1 Abraham, R. J., 90 Ackermann, O., 228 Acton, W., 281,282(506), 283(506), 286(506), 288(506), 295(506), 318(506), 321, 322(506) Adams, E. T., Jr., 367 Adams, M. H., 284,286(530), 293(530) Adkins, H., 205 Aebischer, B., 63 Aeschbach, R., 270 Aeschbacker, H.U., 267,270 Agarwal, J. K. P., 218,226 Agaxwal, S.K. D.,215,218,219,226,232, 235,241(100), 255(58) Agnel, J.-P. L., 305 Agresti, D. G., 131 Agric, J., 245 Ahmadi, A., 215,235(66) Aikawa, H., 324 Aizenshtats, Z., 244, 257(335), 257 Ajinomoto, K. K., 228 Akkerman, D. J., 221 Alaminov, Kh., 324 Alberda van Ekenstein, W., 246 Aleksandrov, V. V., 235 Alfiildi, J., 2 I, 38, 39 Alfonso, F. C., 237 Allen, A., 346, 347(2, 3), 348(2), 350, 352, 353, 354, 355(26, 36,40), 356(26, 36, 38, 39), 359(38), 361, 363(69), 364(38, 46), 371(38), 373, 376(46), 377(46) Allerhand, A., 22 Allinger, N. L., 46, 47, 57, 105, 106, 107(13), 109(13) Altman, M., 96

Altona, C., 46, 53(8), 55(8), 67(8), 69(8), 91(8), 107(8), 115(8) Alvarez, A. M., 29 Amadori, M., 259 Amario, M., 318, 321(683) Amthor, C., 234 Anderle, D., 38 Anderson, A. E., 324 Anderson, C. B., 55,56,61,65(51, 52), 67(51), 85(38), 115(38,41, 51,52) Anderson, J., 335 Anderson, L. L., 244,257(335) Anderson, M., 195 Anderson, M. F., 380 Andrasko, J., 2, 27(5) Andres, C. 232 Andrianoff, N., 216,219(83), 226(83), 244(83) Anet, E. F. L. J., 240 Angel, J. P., 225, 28q169, 170) Anguin, H., 323 Angyal, S. J., 2, 4, 5,6, 7, 8(22, 23), 9, ll(15, 17,27, 31), 13, 14, 15, 16, 17, 18, 19(45), 20(45), 2 I, 24,25(58), 26(5 I , 56), 27(23), 28(3,23,45), 29(23), 31, 32, 33(117), 35(13, 15, 17,48), 36, 38, 39(51, 134),40,41,42,46,47(1), 50, 54(23, 24), 56(23), 65(23,24), 68(23), 92(23,24), 103(23,24), 126, 127, 128(26), 129, 131(22, 24, 25, 26, 36), 132(28), 134(28), 139(28), 140(28), 237 Anh, N. T., 101, 1 lO(157) Anno, K., 249 Ansart, M., 323 Anthonsen, T., 10, 1 l(33) Antuschewich, I., 244 Aoi, M., 330 Aplan, F. F., 329 Apresyan, A. S., 278,280(477), 305(477) Arai, K., 193,301,302,315 Armitage, I. M., 126

383

384

AUTHOR INDEX

Amott, S., 34 Arpalahti, J., 12, 20(41), 21(41) Arrendall, R. F., 343 Asano, Y., 332 Aschard, F. K., 204 Ashihara, Y., 325 Ashton, F. E., 175, 185 Am, K., 336 Aspinall, G. O., 31, 33(109), 167, 168(1) Astrup, E. E., 52, 79(33), 97(33) Atterer, M., 245 Augk, J., 84, 85(131), 1 lO(131) Augustat, S., 289 Aune, K. C., 367 Avery-Jones, F., 346, 347( I), 373( 1) Aya, T., 329,344(816) Aydashkina, 0. A,, 326

B Bachman, S., 225 Bipkowicz, M., 247, 291, 299, 300(558, 576), 303(576), 305(558) Baddeley, G., 109 Badenhuizen, N. B., 286 Badollet, M. S., 214, 221 Baeyer, A., 267 Baiann, I. C., 3 13 Baker, J. D., 40 Baker, R. S. V., 269 Balachandra, J., 326 Ball, A., 358,380(60) Ballard, R. E., 16 Baluch, J., 332 Balm, F., 70 Bames, E., 205 Bandyopadhyay, A. K., 325 Bantlin, G., 278, 335(476) Baranowski, K., 322 Barbry, D., 55, 6 I (44), 63(44), 67(44), 69(44), 115(44) Bar-Guilloux, E., 52, 88(32) Barham, H. N., 280 Barker, P. E., 37 Barker, S. A., 21 Barley, T. P., 208 Barlow, J., 321 Barrett, J. M., 34, 126, 129(14), 160(14)

Barrett, T. W., 312 Barrett-Bee, R. A., 372 Banie, J. K., 126 Barsuhn, C. L,. 29 Basler, W. D., 303 BaS~n-Noori,B., 18 Batstone-Cunningham, R. L., 132, 134(65, 66,67a), 153(65), 154(65), 155(65) Bauer, K. D., 334 Bauer, W. D., 190 Bauminger, B., 280 Bayer, E., 38 Bed, G. D., 22 1 Beati, E., 205 Beattie, J. K., 5, 29(16) Becher, M. G., 269 Beck, B. H., 55 Becker, P., 102 Beckley, V. A., 245 Bedford, G. R., 306, 372 Beevers, C. A., 22 Beghian, M. A., 3 18, 32 l(683) Behr, J. P., 126 Bek, Yu. B., 326 Belhouwer, D., 205 Bell, A., 353, 354(36), 355(36), 356(36) BeMiller, J.N., 52,65(36), 116,222,315 (132) Benbow, J. A., 9 Bendl, J., 27 Bent, H. A., 7 1 Benzing-Purdic, L., 257 Berenson, G. S., 334 Berezin, N. B., 326 Berezina, S. I., 326 Bergdoll, M. S., 239 Berger, G., 225,280 Bergstrom, H., 222 Berhard, R. A., 30 Berliner, L. J., 128, 130(32), 131(32) Beman, H. M., 25,7 1 Berndt, M., 89 Bernhardt, G., 323 Bernhardt, T., 323 Berquist, C., 319, 321(691) Berry, R. E., 240,253 Berthod, H., 28 Berton, A., 336 Bessho, S., 23 1 Bethell, G. S., 36, 39( 134), 41

AUTHOR INDEX

Beveridge, R. J., 36, 38, 39(134), 42(144) Bhaskar, K. R., 348 Bhatnagar, H. L., 247 Bhattacharjee, P. K., 325 Bhavanandan, V. P., 361,363(66), 365 Biccsy, M., 214 Bielawny, J., 228 Bielecki, J., 280 Bigbee, W. L., 132 Bilik, V., 20, 37(67), 38, 39 Bilisies, L., 36 Binder, H., 236 Bingham, R. C., 115 Binkle, B., 236 Binkley, J. S., 7 1, 76(94), 77(94), 82(94), 84(94), 97(94) Binkley, R. W., 252 Binkley, W. W., 221,241,252 Biot, J. B., 282 Birch, G. G., 327 Birch, J. R., 312 Birdsall, B., 126 Bishop, C. T., 50, 237 Blackwell, J. A., 303, 3 12, 347, 370( 17) Blagoveschensky, V. S., 6 1 Blair, M. G., 24 1 Blake, J. D., 18 Blanshard, J. M. V., 50 Blayer, E., 205 Bleany, B., 130 Bleha, T., 47, 53( 1 l), 77, 80, 83( 1 13), 84( IOO), 86( loo), 88( 108), 94( 1 I), 96, 97(11, 148), 104, 109(161), 111(148), 113(148), 114(113), 120(148),121 ( 148) Blinc, M., 286, 316 Blitz, W., 286 Bloembergen, N., 132, 133(58) Blomberg, A. W., 323 Blumenthal, A., 208 Boat, T. F., 347 Bobrovnik, L. D., 215, 254 Bock, J., 101, 105(153) Bock, K., 52, 59, 88(32), 105(59), 105(60) Bode, H. E., 321 Bodkin, C. L., 38, 42(154, 155), 126 Boehm, H. T., 321 Boersma, A., 352, 372(33) Boggs, J. E., 96 Bohinski, R. C., 126

385

Bollenbeck, G. N., 229 Boniek, M., 327 Bonin, A. M., 269 Bonnet, P., 77, 9q107) Boon, J. J., 257 Booth, G. E., 55, 61(37), 85(37), 115(37) Booth, H., 55,6 1(40), 1 15(40) Borisov, G., 324 Boskov, Z., 281 Bosson, C., 171 Bottomley, W. B., 241 Bouillon-Lagrange, E. J. B., 28 1 Bourne, E. J., 4, 21(12), 37(12) Box,V. G. S., 116, 120(189, 190, 201) Boxler, D. L., 19 Braun, M., 132, 137 Bretschneider, E., 90 Brewster, J. H., 50, 92(20) Brey, W. S., Jr., 34, 126, 129, 160(14) Bridge, J. L., 374, 38q120) Bnggs, A. J., 117, 119(198), 120 Briggs, J., 5 Brimhall, B., 290, 306(555) Brindle, R. J., 321 Brisson, J. R., 175, 185 Brobak, T. J., 24 Brodowski, A., 22 1 Broido, A., 249 Browing, W., 321 Brown, C. F., 371 Brown, D. R., 239 Brown, D. W., 239 Brown, E. B., 34, 126, 129(14), 160(14) Brown, M. C., 367, 369(95), 370(95) Brown, R. K., 61 Brownstein, S., 8 Browon, E. B., 129 Bryce, D. J., 272, 304(462), 306, 308, 319, 336, 339(462), 340(695, 898) Buchanan, G. W., 57 Buchanan, H. J., 241 Buckee, G. K., 208 Buckingham, A. D., 6, 129 Bugaenko, I. F., 2 15, 229, 235 B u g , C. E., 22,24, 25,26(76), 31, 126 Bulfer, A. J., 319, 321(693) Bulgakova, I. F., 215 Bull, H. G., 116, 118(184) Bumb, R. R., 43 Bundle, D. R., 179

386

AUTHOR INDEX

Burchard, W., 370 Burelova, A., 3I7 Burian, O., 245 Burke, A. R., 132 Burkert, U., 105 Bushweller, C. H., 55 Butterworth, K. R., 208,268 Buurmans, H. M. A., 27,29(98) Buys, H. R., 46, 53(8), 55(8), 67(8), 69(8), 91(8), 107(8, 13), 109(13), 115(8) Byeong, J. Y., 224 Byme, G. A,, 340

C

Cady, S. M., 325 Cael, J. J., 312 Caesar, G. V., 28 1,286,293(535), 297(534), 301(534), 319, 321(689) Cai, Z., 267 Caldwell, C. G., 283 Caldwell, M.J., 280 Calloud, F., 1 Campbell, I. D., 131 Canic, V., 328 Cantacuzene, J., 101, 1 10(157) Cantor, C. R., 35 1 Capon, B., 116, 118(183), 255 Capper, B. S., 335 Carla, E., 234, 306(249) Carlin, N., 195 Carlo, D. J., 173, 184(19) Carlson, D. M., 38 1 Carlson, L. J., 305 Carlstedt, I., 346, 347, 348, 352, 353(13, 15, 31), 354(15, 31), 355(13), 356, 358(54, 55), 361(13,56), 363(13,54), 364(15, 5 5 , 56), 365(53, 54, 5 9 , 367, 368( 101), 369(53, 54, 55, 56, 101), 370(53, 54), 371(53, 54), 378(101) Camevale, J., 208 Caroff, M.,179 Carroll, B., 274 Carter, R. D., 6, 11, 128, 132(30), 134(30), 140(30), 143(30), 165 Carvalho, D. S.,306, 308(624) Casier, J. P. J., 215, 235(66) Castillo, F. J., 335 Cavalieri, L. F., 240,252, 259(314), 267

Ceh, M., 283,286 Cerniani, A., 21 8,219(97), 220(97), 22 1(97), 246(97), 272(97), 274(97), 289, 290(552), 297(552), 298(552), 336(97), 337(552) Cerny, N., 208 Cemtti, D. P., 52,65(36) Chalmers, L., 329 Chan, R. M.,270 Chandhuri, B., 325 Chandrasekaran, R., 34 Chandrasekhar, S., 117, 119(195) Chang, K. S., 221 Chang, S. S.,328, 329 Chantler, E. N., 347 Chao, C. C., 39,40(161) Chapguis, C., 270 Chattejee, P. K., 318, 319(682) Chen, S. H., 357 Chen, X., 267 Cheng, H. T., 222 Cheng, P. W., 347 Cheng, P. Y., 371 Cheng, X., 334 Cherbikova, A. T., 250 Chertovskikh, N. B., 326 Chesters, G., 299 Chiba, H., 230 Chikin, G. A., 215, 216, 235, 241(81), 242(81), 243(81), 256,257(401) Chikin, S. A., 2 I5 Chin, P. P. S.,338 Chizhov, 0. S., 168 Chmielewski, M.,52, 5 5 , 61(46), 65(36), 67(46), 84(46), 85(46), 1 15(46) Chortyk, 0. T., 343 Christensen, B. E., 367, 372( 100) Christensen, G. M.,306,308 Chu, S. S. C., 25 Chuang, I. S., 244 Chuang, J. S.,253 Chun, P. W., 367 Chundury, D., 230 Chung, D. Y., 105 Chung, Y. H., 224 Chvalovsky, V., 106 Cimerman, Z., 27 Claesson, A., 35 Clamp, J. R., 349, 36 1, 362(23), 380 Clanton, J. A., 135, 136(77)

AUTHOR INDEX

Clark, A. V., 229 Clark, I. T., 201 Clark, P. C. T., 218,228(90) Clarke, R. J., 281 Clarke, S. J., 30 Clavene, P., 109 Clegg, J. E., 319 Cleland, J. E., 21 5, 228 Cleveland, F. C., 304, 308(607) Coan, R. M., 352, 353(35), 361(35) Coates, J. H., 28 I Cochran, W., 22 Collino, F., 237 Coloso, R. M., 334 Colson, P., 3 13 Comper, W. D., 357, 375(59) Convey, C. T., 267 Cook, W. H., 201 Cook, W. J., 22,24, 25,26(76), 31, 126 Cooke, S., 319 Cooper, A. B., 19 Cooper, A. G., 367, 369(45, 9 3 , 370(95), 378(45), 379(45) Cooper, B., 349, 358(45), 36 I, 362(23), 363(45), 374(45) Cooper, F. F., 237 Cooper, F. P., 50 Cope, W. T., 367 Coppens, P., 102 Cordell, F. R., 96 Cordes, E. H., I 16, 1 18(184) Corfield, A. P., 346 Con, C. F., 309 Correia, J. J., 367 Costa, C., 299, 300(571) Costa, D., 299,300(571) Costagnon, R., 214 Couturier, D., 55, 61(44), 63(44), 67(44), 69(44), 1 15(44) Cowley, J. E., 41 Cradock, S., 101 Craig, D. C., 17, 22, 23, 26(51, 56) Crampton, C. A., 234 Creeth, J. M., 346, 347, 348, 349, 350, 35 I , 352, 353, 354(30), 355(37), 358(45), 359, 360(62), 361,362(8,23, 30, 37), 363(45), 366(8, 37,62), 367, 368(30), 369(24,45), 371, 372(7, 30), 373,374, 375, 377,378(45), 379(30,45), 380(37, 120)

387

Cross, H., 327 Csizmadia, I. G., 106 Cdregh, I., 35 Cubero, I. I., 41 Cummins, H. Z., 375 Cunningham, M., 216,226(82), 241(82) Cushing, M. L., 286,293(535) Cyr, N., 70, 88 Czajkowski, A., 325 Czamiecki, M. F., 34,35(127), 126, 160(17)

D Dadok, J., 41 Dagna, L., 236 Dahlquist, A., 269 Dahlquist, F. W., 132 Dai, Y.,96 Daido, K. K., 330 Dais, P., 35, 134 Dale, J. K., 23 Dalferes, E. R., Jr., 334 Daman, M. E., 35, 126, 128, 132, 134, 140(29), 144(29,68), 146(68), 147(68), 148(68), 149(68), 151(64b), 152(64b), 153(65), 154(65), 155(65), 159(64a), 160(64a), 162(64a), 163(64a), 164(64a), 165 Daniel, A., 22 1 Darbon-Mejssonier, N., 144 Das, I., 348 Das, M. K., 38 Datta, G. L., 324 Datta, R. L., 318, 319(682) Dauphin, J. F., 225,280(169) Dave, G. B., 286,289(536), 290(536), 294(536), 295(536), 296(536), 298(536), 299(536), 308(536) David, S., 21, 84, 85(131), 110 Davidek, J., 249 Davidson, E. A., 361, 363(66), 365 Davies, J. B., 205, 38 1 Davies, K. P., 2, 7(3), 13(3), 28(3), 127, 129(22), 131(22) Davies, T., 304 Davis, M., 38 Davis, M. A., 283 Davis, S. S., 380

388

AUTHOR INDEX

Dawber, J. G., 12,239 Dawes, I. W., 255 de Bruijn, J. M., 253 De Figueroa, L. I. C., 335 de Hoog, A. J., 46, 53(8), 5 5 , 61, 67(8,42), 69(8,42), 91(8), 107(8), 115(8,42) de Leeuw, J. W., 257 de Mayo, P., 46, 50(5), 107(5) De Page, G. M. J., 215, 235(66) de Santis, P., 52 De van Broek, M. R. G., 335 Dearbon, D. G., 347 Dechter, J. J., 132 Decoster, E., 134, 144, 146(68), 147(68), 148(68), 149(68) Defaye, J., 17,26(56), 38, 52,88(32), 168, 169(5), 170(5,6), 171, 172(5), 202 Degand, P., 367, 369(98), 374 Degtyareva, E. V., 325 Delbaere, L. T. J., 59, 105(59) Dell, A., 192, 199(55) Denarit, M., 132, 134 Denkov, A., 323 Denne, H. B., 235, 236(267) Denner, W. H. B., 214,253 Deonier, R. C., 367 Deslongchamps, P., 61, 116, 117(187), 119(187, 188) Desreux, J. F., 9, 1 l(29) &Tar, D. F., 117 Devaquet, A., 101, llO(157) Dheu-Andries, M. L., 23, 126 Di Baja, A., 222 DiGioia, A. J., 132 Di Paolo, C., 252 Dill, K., 6, 11, 35, 126, 128, 132, 134, 14q29, 30), 143(30), 144(29,68), 146(68), 147(68), 148(68), 149(68), 15 1(64b), 152(64b), 153(65), 154(65), 155(65), 159(64a), 160(64a), 162(64a), 163(64a), 164(64a), 165 Dimitriev, B. A., 181 Dimler, R. J., 283 Ding, C., 267 Ding, J., 236 Dittmar, P., 323 Dixon, J. E., 38 1 Dmitriev, B. A., 175, 177, 178, 180(26, 39), 182, 183(38), 196(26), 198 Doane, W. M., 325

Dobele, G., 247 Dobner, A., 232 Dobson, C. M., 130, 131 Dodomba, W., 253 Doihara, H., 270 Doihara, T., 249,272 DoleEek, V., 283, 286 Domard, A,, 171 Domburgs, G., 250, 336 Domovs, K. B., 3 Donald, A. S. R., 349, 361, 362(23) Dong, S., 268 Donovan, J. W., 304 Doree, C., 216, 226(82), 241(82) Dorland, L., 352, 372(33, 34) Dorman, D. E., 3 13 Dorofeenko, G. N., 88 Downey, K. W., 172 Drake-Law, H., 221 Driquez, H., 52,88(32) Dufresnoy, X., 218, 219(95) Dugvilliene, L., 336 Dull, M. F., 249 Dunayeva, K. N., 241 Dunsbergs, G., 247 Duran-Chavama, M., 33 1 Durand, H. W., 249 Durette, P. L., 61, 62, 65(72, 73, 74, 75, 76, 78, 79, 80), 69(80) Duschsky, J. E., 246 Dusmuradov, T., 324 Dutkiewicz, B., 325 Dutton, G. G. S., 237 Dwek, R. A., 130, 136(39), 137(39) Dworschala, E., 289 Dyer, R. H., 237 Dylaeva, L. I., 326 Dzanazyan, R. S., 23 t Dzhanpoladyan, L. M., 231 Dzhuraev, A. D., 324

E Eckhardt, A. E., 38 1 Edlund, U., 132 Edo, H., 235, 259(266) Edward, J. T., 46, 107(4), I18(4) Edwards, J. R., 172

AUTHOR INDEX

Edwards, R. A., 255 Efimon, B. N., 23 1 Efremov, N., 247 Egorov, I. A., 220,231,239 Ehrich, F., 226 Ehrlich, F., 245, 317, 319(674) Eisenstein, O., 84, 85(131), 101, 110 Ekstr6m, L.-G., 27 Elder, J. B., 347 Elgsaeter, A., 367, 372(100) Eliel, E. L., 46, 47, 55, 56, 61(45, 50), 65(45, 50), 67(50), 107, 109(13), 1 15(45, 50) El-Karawy, A,, 55,61(43), 67(43), 115(43) El-Kashouti, M. A., 3 15 Elkin, Yu, N., 187 Ellaiah, P., 335 Eller, W., 245 Ellis, G. P., 255 Elstein, M., 347 Emes, C. H., 374 Enders, C., 240,249,259(3 15) Enomoto, S., 278, 305(475) Enrico, G., 225,280(169) Eriksson, H., 325 Erlander, S. R., 3 13 Emi, I., 327 Esperson, W. G., 135, 138(70,71) Euler, H., 237 Evans, C. M., 117, 119(198) Evans, I. G., 208 Evans, J. G., 268 Evans, M. E., 38,41, 42 Evans, R.B., 281,283 Evans, W. J., 27, 126 Evdokhimov, V. F., 225,276(164), 280(164) Everett, G. W., Jr., 127, 128(27) Evine, H., 224 Exner, O., 8 1

F Fabrega Dalman, J., 328 Fanta, G. F., 325 Farkas, E., 331 Farley, F. F., 3 17 Farnell, L. F., 132 Fawcet, A. H., 253

Fayet, J., 230 Feather, M. S., 240, 246, 252(346) Fecak, B., 3 17 Fedoroiiko, M., 38 Feeney, J., 126 Feichfinger, H., 132 Feld, R., 102 Feldhoff, P. A., 361, 363(66), 365(66) Felicetta, V. F., 37 Felton, G., 3 17, 3 19(670) Fenocchio, M., 236 Fenton, H. J. H., 241 Fermi, E., 131 Ferrari, B., 132, 134(66) Fetzer, W. P., 218 Fielding, A., 321 Figures, W. R., 172 Finch, P., 5, 63 Firil, J., 81 Fischer, E., 241 Fischer, F., 249 Fischer, H., 323 Fischer, R., 327 Fischer, S. K., 308, 309(63I), 3 16 Fishbein, R., 325 Fisher, H., 334 Flaig, W., 253 Fleche, G., 225,230,280 Flechter, H. H., 284,286(530), 293(530) Flippen, S. L., 165 Row, P. J., 350 Fodor, L., 315 Foerst, W., 216,221(84) Foerster, H., 3 13 Fohrenkamm, E. A., 344 Fornal, J., 280 For&n, S.,2, 27(5) Forsskahl, I., 253 Forstner, G. G., 354 Forstner, J. F., 354 Forth, W., 327 Fortuna, T., 288 Foster, M. A., 135 Fouassier, J. P., 280 Foumet, B., 352 Fox, J. S. G., 19 Fradiss, M. N., 234, 238 Frahm, H., 305 Frahn, J. L., 3, 4(6) Frampton, V. L., 27, 126

389

390

AUTHOR INDEX

Franck, R. W., 57 Franks, F., 20,41, 50 Frederickson, R. E. C., 323 Freeman, R., 132 Freudenberg, K., 280 Freund, E. H., 3 Freyer, G., 323 Friedman, M., 52, 57(35), 63(35), 65(35) Friedrich, H., 255 Frischenschlager, S., 206,218(26), 223(26), 235(26) Fuchs, B., 52, 71(30), 74(30), 76(30), 98(30) Fuchs, G., 208 Fujii, S.,229, 240, 255, 269 Fujimaki, M., 256 Fujita, H., 366, 376 Fujita, Y., 329, 344(816) Fujiwara, T., 193 Fuller, A. O., 3 19 Funita, K., 268 Furia, T. E., 205 Furthmayr, H., 126 Furukawa, E., 247 Funtkawa, T., 283,289(525)

G

Gabbe, E. E., 327 Gadelle, A., 17, 26(56), 38, 168, 169(5), 170(5,6), 171, 172(5),202 Gagnoli, S. D., 208 Gajda, T., 325 Gal, S., 315 Galabert, C., 367,369(98) Galat, A., 3 12 Galkina, G. W., 222 Galkowski, T. T., 3 10 Gallagher, J. T., 346 Gallimore, D., 235 Gandini, C., 206 Gangolli, S. D., 268 Ganguly, A. K., 306 Gansow, 0. A., 132 Gao, B., 236 Gapen, C. C., 319, 321(693) Gapski, S., 227 Gardiner, D., 306, 338, 340 Caret, A., 230

Gargioli, J. D., 132 Gariepy, J., 132 Garino, M., 221 Garvie-Could, C. T., 316 Gasparini, G., 236 Gaunt, I. F., 208 Gavrikov, B. I., 324 Geerdes, J. D., 306, 308(625) Geis, M. P., 6 1 Gelas, C., 230 Gelis, A., 204,215(5), 216(5), 219,226(5), 241(5, 102), 244(5) Gerecs, A., 306 Gergely, S., 267 Gerwig, G. J., 173 Geuthuer, F. J., 269 Ghali, J., 205 Gianotti, C., 206 Gibbons, R. A., 361, 363(67), 371(71), 372, 374 Gidley, M. J., 313 Giglio, E., 52 Gill, C. H., 2 Gillier-Pandraud,H., 102 Giza, C. A., 56,61(50), 65(50), 67(50), 115(50) Glabe, E. F., 331 Glaudemans, C. P. J., 38 Glenn, R., 1 17, 1 19(198), 120 Glickson, J. D., 131 Gliksman, M., 330 Glover, F. A., 361, 363(67), 372 Godbole, N. N., 330 Godfrey, R. E., 357, 375(57) Goldblum, N., 22, 24(76), 26(76) Goldblum, W., 126 Goldie, E. L., 237 Goldstein, I. J., 255, 274(388) Golubev, M. N., 324 Golygin, V. A., 239 Gombocz, E., 206,218(26),223(26),235(26) Good, B. W., 9, 1 l(29) Gore, H. C., 32 1 Gorenstein, D. G., 77 Goretti, G., 252 Gorin, L. F., 225,280( 166) Gorin, P. A. J., 9, 313 Gorokhov, G. I., 221 Gorrelli, A., 2 17 Gomnchon, J.-P., 230

AUTHOR INDEX Gostling, M., 24 1 Gottschalk, A., 361, 371(71) Gouda, M. S., 229 Gould, R. O., 31, 33 Gould, S. E. B., 33 Goulding, D. J., 28,29( 101) Gow, C. Y., 224 Graefe,G., 282,286(520),293(520), 310,329 GHfe, G., 205 Graham, T., 205,214(7, 8) Grandjean,J., 30 Grant, G. T., 33 Grant, I. F., 268 Grasdalen, H., 10, 1 l(33) Grasso, P., 208,268 Green, A. G., 241 Greenshields, R. H., 204, 214, 221(55), 222(55),223(55),224( I), 226( I), 259(55) Greenwood, C. T., 272,281,299(504), 301, 304(462), 306, 308, 319, 336, 339(462), 340(695,898) Greeves, D., 5 , 6, 7, 8(22), 9, ll(17, 31), 14, 15(18, 22), 16(17), 31, 32(117), 33(117), 35(17,48), 41(48, 117), 127, 129, 131(36) Gregorski, K. S., 238 Greiser, R. Ya., 222,223(140) Gresh, N., 109 Gries, H., 328 Griessmayer, V., 234 Griffin, H. L., 313 Griminger, P., 334 Grindley, T. B., 55, 61(40), 115(40) Grives, S., 269 Gross, D., 235 Grossfeld, J., 205 Grott, K., 227 Gruber, H., 272, 273(466), 275(466), 276 GrzeSkowiak, M., 286, 287(537), 288(537), 289(537), 291(537), 292(537), 293(537), 318(537), 329 Gudin, N. V., 326 Guibe, L., 84,85(131), llO(131) Guilian, A. M., 126 Guizard, C., 19,43(61a) Gulyuk, N. S., 221 Gunge, N., 201 Guokun, D., 230 Gupta, K. V., 2 15 Gupta, R. K., 136

39 1

Gupta, S. K., 235 Gupta, V. K., 232 Gupta, V., 2 19 Gustafsson, B., 175 Gutknecht, H., 266 Guttmann, R., 317, 3 19(674) Gyorgydeak, Z., 52,57(35), 63(35), 65(35)

H Haas, J. W., Jr., 29 Hagen, C., 322 Hager Veres, A., 332 Haines, A. H., 16, 30 Hall, G., 224 Hall, J. R., 20 Hall, L. D., 26 Hallas, G., 237 Hallett, P., 371 Halvorson, H. R., 367 Hamar, G. K., 70 Hamilton, B. K., 201 Han, K. E., 223 Han, V. Y., 326 Hanaguchi, E., 224 Hanessian, S., 2 1 Hanna, D. A., 127, 128(27) Hanson, J. C., 102 Hara, E., 222,226(126, 128), 228, 332(127) Hara, K., 28 1 Harbitz, O., 10, 11(33), 367, 380 Harding, S. E., 346, 349, 350(24), 35 1, 353, 354(25, 30), 355(37), 358, 359, 360(62), 361(30), 362(9,30, 37), 365, 366(37, 62, 64, 74, 75, 76), 367, 368(30), 369(24,45), 371, 372(30), 373(37), 374, 377, 378(9, lo), 379( 10, 30), 380 Hardt, H., 170, 172(8) Hardy, G. E., 12 Hargrave, K. D., 55, 61(45), 65(45), 115(45) Hariu, J., 334 Harper, G. S., 357, 375(59) Hams, J. F., 246, 252(346) Hams, J. R., 367, 369(96), 370(96) Hamson, J. S., 325 Harvey, R., 380 Hascall, V. C., 376 Hasegawa, R., 294

392

AUTHOR INDEX

Hashimoto, H., 28 1 Hashiwagi, H., 278, 305(475) Hasted, J. B., 312 Hatch, V. I., 229 Hatsura, H., 328 Haug, A., 32, 33 Hauge, S. M., 214 Havinga, E., 46, 53(8), 55,61(42), 67(8,42), 69(8,42), 91(8), 107(8, 13), 109(13), 1 15(8,42) Hawley, M. C., 172 Hayami, J., 61 Hayase, F., 256,257,270 Hayashi, M., 268 Hayashi, Y., 268 Hearne, J. F., 256 Hehre, W. J., 77, 80, 110, 114(111) Heiker, F. R., 61,63(72), 65(72, 77) Heinrich, H. C., 327 Heinsoo, E., 247 Hellwig, E., 206,218(26), 223(26), 235(26) Hennings, H. J., 185 Herdan, G., 375 Herstein, K. M., 230 Herting, T., 272,336(465) Herzer, A., 135, 136(77) Heunart, C., 232 Heveish, A., 315 Hevia, P., 334 Hexen, J. G., 132 Heyns, K., 259,274,276 Hickman, R. J., 8(23), 27,28(23), 29(23), 41(23) Hicks, D. L., 380 Hicks, K. B., 37, 40, 281 Hiemstra, P., 3 19 Hikin, G. A., 239 Hilbert, G. E., 283, 306 Hill, A. S., 50, 88(21), 92(21) Hill, H. D., 347 Hill, R. L., 347, 381 Hiller, L. A., 237 Hilliard, F. C., 3 19 Hilton, H. W., 241 Himmel, M. E., 40 Hinnie, J., 366 Hintze, B., 55,61(46), 67(46), 84(46), 85(46), 115(46) Hiroyuki, O., 39 Hixon, R. M., 317, 319(670)

Hjermsted, E. T., 298 Hjortas, J. A., 33 Hodge, J. E., 259 Hodges, R. S., 132 HoelR, F., 297 Hoerlscher, H. F., 322 Hoffman, S., 331 Hoffmann, R., 110, I 1 1 Hofman, I. L., 175 Hofsted, T., 175 Hola, O., 280 Holland, C. V., 62 Hollenstein, R., 63 Holloway, C. E., 132 Holme, T., 175 Holmes, E., 239 Holmes, F. H., 305, 340 Holz, M., 357 Honda, M., 3 16 Hoogenraad, N., 38 Hoppen, V., 55 Horbett, T. A., 359 Horesi, A. C., 322 Horowitz, M. I., 348 Horsey, B., 365 Horton, D., 41,47, 61, 62, 65(72, 73, 74, 75, 76, 78, 79, 80), 69(80), 281, 283(502), 307 Horton, J. R., 348, 374, 380( 120) Horvath, G., 332 Houdret, N., 367, 369(97,98) Hough, L., 38 Houminer, Y., 249,274 Hricovini, M., 55,61(46), 67(46), 70, 84(46), 85(46), 1 15(46) Huang, J., 268 Huang, Y., 228 Huchette, M., 225, 280 Hudson, J. S., 3 15 Huheey, J. E., 128 Hull, W. E., 352, 372(33) Hullar, T. L., 255, 274(388) Huntley, J. F., 346 Huntoon, S., 268 Hunt, D. T., 262 Husain, S. K., 3 12 Hutchins, R. O., 107 Hutton, D. A., 352, 353(35), 361(35), 364(46), 376(46), 377(46) Huyghebaret, A., 224

AUTHOR INDEX

Hyashi, M., 59, 96 Hyman, M., 229 I

Ihl, A., 234 Ikan, R., 244,257 Ikeda, T., 329, 344(8 18) Illies, R., 218, 223(87) Ilyasov, S. G., 322 Im, C. J., 329 Imanura, S., 330 Inagaki, F., 9 Inatome, M., 249 Ingles, D. L., 235 Inoue, H., 240 Inoue, K., 269,330 Inubashi, A., 321 Ioselis, P., 244, 257 Ipatieff, V. I., 205 Irish, D. E., 20 Irvine, J. C., 306 Isaka, H., 294 Isakov, V. V., 187 Isbell, H.S., 3, 12(10), 26, 50,92(19) Ishida, K., 324 Ishidate, M., Jr., 268 Ishigashi, H., 224, 228( 157) Ishihara, R., 332 Ishikashi, M., 240 Ito, H., 239 Ito, K., 281 Ivanov, S. N., 239 Ivanov, S. Z., 215,239 Izumi, K., 19, 32, 33(121) Izumi, M., 283, 285(524), 289(524), 301(524), 3 12(524) J

Jackobs, J. J., 43 Jackson, A. J., 335 Jackson, C., 380 Jackson, J. E., 323 Jacobi, E.,301 Jacobs, M. B., 222 Jacques, J. W., 34 Jacques, L. F., 126 Jacques, L. W., 35

393

Jadhar, S. S., 270 Jado, T., 330 Jaegerstad, M., 269 Jafri, J., 328 JQerschmidt, A., 234 Jahan, N., 270 Jain, N., 324 Jain, R. K., 247 Jakovljevic, J., 28 1 James, A. E., Jr., 135, 136(77) James, V. J., 18,24,25(58) Jamieson, A. M., 347,370(17) JaniEek, G., 249 Jann, B., 175 Jann, K., 175 Jansson, P.-E., 175, 186 Jaques, L. F., 126, 160(20) Jaques, L. W., 126, 129(14), 160(14) Jardetzky, O., 130 Jaret, R. S., 19 Jarosz, J., 55, 61(46), 67(46), 84(46), 85(46), I I5(46) Jaroszewski, J. W., 80 Jean, Y., 101, llO(157) Jeffrey, G. A., 24, 25, 52, 71, 76(93, 94,95, 96), 77, 78(101, 102), 80(95), 84(94), 86(94), 96(95), 106 Jeffrey, P. D., 367 JehliEka, V., 8 1 Jennings, H. J., 175, 313 Jensen, F. R., 55 Jensen, J. H., 102 Jensen, N. J., 268 Jenssen, A. O., 367,374,380 Jentoft, N., 347 Jenzen, C. C., 284,286(531) Jesson, J. P., 131 Jeszensky, Z., 235 Jewell, J. S., 62 Johary, P. C., 219,229,241(100) Johnson, L. R., 346, 347(3), 354(3), 373(3) Johnson, M. L., 367 Johnson, P., 357, 371, 375(57) Jones, J. K. N., 19, 37, 38( 137) Jones, M. M., 135 Jones, R. C., 266 Jong, H. L., 224 Jsrgensen, F. S.,80, 114, 12q180) Josefowicz, M. L., 55 Joszt, A., 215,216,218,234(94), 241(80)

AUTHOR INDEX

394 Joyce, K. L., 34 Juckenack, A., 205 Junck, H., 236

K Kaburaki, Y., 272 Kahn, S. A., 328 Kainuma, K., 283,289(525) Kaji, A., 328 Kaliannan, P., 80 Kamerling, J. P., 173 Kamyama, K., 324 Kanamori, H., 270 Kang, H. L., 224 U p p i , R., 12, 16, 20(41, 42, 44), 21(41), 27(40,41), 28(52, 53), 30(52) Kar, D., 77 Kargin, V. A., 218 Kari, F. W., 334 Karlivan, V. P., 250 Karracher, D. G., 128 Kasahara, Y., 325 Kashyap, R., 326 Katiyar, S. S.,237 Kato, H., 59,96,256,257,270 Kato, S., 283,285(524), 289(524), 301(524), 3 12(524) Katsura, N., 201 Katsurai, T., 239 Katz, J. R., 286,298, 306(569, 570) Katzenellenbogen,E., 176 Kaufman, B., 37 1 Kaura, R., 361, 363(69) Kawabata, A., 299 Kawachi, T., 3 16 Kawai, T., 29,278 Kawamoto, M., 224 Kawamura, M., 330 Kay, R. L., 41 Kazimirchik, I. V., 61 Keane, J. C., 221 K m l e y , M. W., 327 Keininger, H., 235 Kelly, F. H. C., 239 Kelso, M. T., 5,29( 16), 126 Kenne, L., 35, 175, 176, 179(29), 184 Kent, L. H., 373 Kenyon, W. O., 317,319(672)

Kerr, R. W., 286,297(534), 301(534), 304, 308(607), 309, 317, 319, 321(690) Kesler, C. C., 298 Kettner, H., 222 Keysers, H., 255 Khabibullaev, P. K., 305 Kharin, S. E.,216,239 Khedhair, K. A., 55,61(40), 115(40) Khenohk, M. A., 225,276(164), 280(163, 164) Kholodkova, E. V., 182 Khomenko, V. A., 187 Kieboom, A. P. G., 9, 10, 14,27,28(49), 29(98), 31,230,252(217), 253 Kim, C. K., 224 Kim, C. M., 223 Kim, E. H., 270 Kim, H., 367 Kim, H. S., 24 Kim, S. B., 256,257,270 Kim, S. J., 367 Kim, S. Y., 221 Kim, W. J., 224 Kim, Y. N., 223 Kimura, K., 8 1,98( 122) Kimura, R., 33 1 Kinae, H., 270 King, B., 332 King, P. S., 222, 226(122) King, R. E., 19 Kinoshita, G., 322 Kirby,A. J.,47, 53(10), 116(10), 117, 119(10, 195, 196, 197, 198), 120 Kirchhoff, G. S. C., 28 1 Kishi, Y., 15 Kishihara, S., 269 Kishihawa, S., 229 Kitamura, K., 316 Kitamura, S., 331 Kiyama, R., 322 Klein, A., 352 Kletenik, Yu.B., 326 Klier, M., 276 Klochkova, T. A., 2 15 Klopman, G., 118 Knight, C. G., 346,352, 362(8), 366(8), 371, 374(7,8) Knirel, Yu,A., 175, 177, 178, 18q26, 39), 181, 182, 183(38), 187, 188, 189, 195, 196(26), 198, 199 Knorr, L., 266

AUTHOR INDEX

Knowles, M.E., 214,235,236(267), 256 Knowles, M.K., 253 Knox, J. H., 308 Knudsen, J., 268 Kobashi, U., 272 Kocharova, N. A., 175, 178, 180(26), 189, 196(26) Kochetkov, N. K., 175, 177, 178, 180(26, 39), 181, 182, 183(38), 189, 196(26), 198, 199 Koenig, J. L., 3 12 Koerner, T. A. W., Jr., 126 Koethe, R.,303,322 Kohler, R.,322 Kohn, R., 33,43( 123) Kolesnikov, V. A., 22 I Koltseva, R. A., 21 5 Komai, Y., 32 1 Komalavilas, P., 193 Komm, E., 283,284(521), 286(522), 291(521, 522), 293(521), 294(521), 297(521) Komoto, M.,224,228(157), 229,240, 255, 256,269 Komura, T., 294 Kondo, Y.,16 Konigsberg, W. H., 126 Kononenko, 0. K., 230 Konrad, G., 33 1 Kopp, L. C., 107 Koppel, D. E., 375 Korakai, A., 224 Korchew, P., 222 Korchkova, N. V., 215, 216(77) Korf, D., 205 Kornblum, H. L., 43 Korosu, Y ., 2 15 Korotchenko, K. A., 305 Korotchenko, K. N., 225, 28q167) Kosheleva, V. V., 325 Kotani, A,, 330 Kotelnikova, L. P., 254 Kothe, G., 61, 63(72), 65(72), 73, 98, 99( 150), 1 18( 150) Koto, S., 47, 59,68(18), 92(18), 105(59) KovaE, P., 1 18 KovaEik, V., 1 I8 Kovalenko, V. A., 299,300(574), 301,321, 322 Kovayashi, M.,332 Kowalewski, J., 132

395

Kowkabany, G. N., 22 I Koydl, T., 2 18 KO%, T., 52, 70(29), 71(29), 76(29), 84, 85(126, 129), 88,90(127, 137, 138), 92(127, 137), 93(127, 129, 137, 138), 94(127, 130), 103(128), 105, 115(127, 128, 129) Kozlowski, R. J., 3 19 Kozyreva, E. F., 241 Kratky, O., 36 1 Krause, G. P., 322 Kretsinger, R. H., 126 Krishnamurthy, T. N., 310 Krochta, J. M.,315 K d , B., 252 Kroplien, U., 236 Kubo, M.,60, 79(64) Kucera, J., 3 17 Kuchitsu, K., 95 Kugimiya, M.,304 Kiihl, H., 28 1 Kujawski, M.,288,301(545) Kulkami, P. R., 270 Kumanotani, J., 2 15 Kuncheva, M.,223 Kuniak, L., 20, 37(67) Kunio, K., 249, 272(362) Kuntz, I. D., 132 KUO,M.-S., 190, 192, 199(55), 202 Kupchik, M.P., 215 Kuptsevich, Yu.E., 37 Kurita, Y., 60, 79(64) Kuroda, Y.,31 Kurosaki, T., 328 Kushiya, M., 294 Kuster, B. F. M.,252 Kuwabara, C., 302 Kuwabara, N., 281 Kuwata, S., 329,344(8 19) Kuzicheva, M. A., 225,276(164), 28q164) Kuznetsova, I. I., 301,321 Kuznetsova, L. G., 218, 240(89), 241(89) Kwiatkowski, J. S., 89

L Lacombe, J. M., 132, 134, 144(68), 146(68), 147(68), 148(68), 149(68), 153(65), 154(65), 155(65) Lacourse, N. L., 331

396

AUTHOR INDEX

LaFay, V. S., 324 Lalitte, J. J., 374 Lai,Y. z., 335 Laine, A., 374 Lakshmanan, C. M., 322 Lid, K., 247 Lamar, G. N., 132 Lambert, J. B., 67 Lamblin, G., 352, 367,369(94,97, 98), 370(94), 372(33), 374 Lamport,D.T.A., 170, 171, 172, 173(14), 174(11, 14), 178(11), 186(11),200(11), 201(11) Land, D. G., 332 Lange, O., 285, 319(533) Lange, W., 323, 327(533) Lannom, H. K., 132, 134 Lara,W. H., 236 Lareeban, S., 214 Larionov, G. O., 37 Larsen, B., 10, 1 l(33) Lascelles, A. K., 361, 363(68) Laskowski, J. S.,328 Lassalle-SaintJean, V., 205, 214, 223(9) Lassettre, E. N., 306 Lastick, S. M., 40 Laszlo, P., 30 h u e , T. M., 366 Lavie, L. F., 227 Le Fur, R., 38 Lechert, H., 303 Lechert, H. T., 303 MI, F., 240,252, 253(374) Lee, B. H., 225 Lee, H. W., 15 Lee, w. I., 357, 375(57) Leeuwarder Ijcen Melkproduktenfabriken, N. V., 221,224,235(155) Lehikoinen, P., 16,28(52, 53), 30(52) Lehmann, G., 214,236 Lehmann, M., 102 Lehn, J. M., 126 Leissner, M., 319,323 Lemieux, R. U., 46, 50(5), 52(7, 14), 59, 61, 63(7, 14),68(16, 17, 18), 70(16, 17), 91(6), 92(18), 101, 105(59,60, 153), 107(5,6, 14) Lenard, J., 173 Lenchin, J. M., 33 1 Lenkinski, R. E., 13,27(46), 28(46), 30(46), 128, 130(32), 131

Leontein, K., 173, 183, 184(19,44), 186, 195 Leopold, H., 361 Leps, K., 224 Leskowar, S.,283, 286 Lessig, U., 235 LeTreut, A., 367,369(98) Levert, A., 328 Levine, B. A., 130 Levy, G. C., 132 Lewis, B., 306, 308(625) Lewis, K., 132 Lhermet, C., 27, 30 Lhermitte, M., 352, 367, 369(94), 370(94), 372(33) Li, Q., 267 Liberti, A., 252 Lichberger, W., 276 Lichhardt, G. H. P., 234 Lichter, R. L., 132 Lide, D. R., 71 Lieben, F., 280 Lieske, B., 33 1 Lin, C. L., 329 Lin, W. F., 222 Lincoln, S. F., 28 1 Lindberg, B., 167, 173, 175, 176, 179(29), 183, 184, 186 Linden, G., 22 I Lindgren, H., 348, 356, 365(53), 369(53), 370(53), 371(53) Lindquist, B., 176 Lindquist, U., 195 Linek, K., 38 Lingnert, H., 224 Lintner, C., 282 Lipkin, D., 20 1 Lipkind, G. M., 188 Liquori, A. M., 52 Liskowitz, J. W., 274 Lisovaya, E. D., 325 Littlemore, L., 5, 6, 7(17), 9, 11(17), 15(18), 16(17), 31, 32(117), 33(117), 35(17), 41(117), 129, 131(36) Liu, A,, 334 Lloyd, A. G., 208,268 Lloyd, M., 21 Lloyd, P. H., 366 Lobry de Bruyn, C. A., 246 Loftus, P., 372 Lominadze. V. N.. 239 Long, C., 126

AUTHOR INDEX Longchambon, F., 71, 72(97), 76(97), 98(97), 102 Longenecker, J. B., 215,222,228 Lijnnberg, H., 12, 16,20(41), 21(41), 27, 28(95) Lijnngren, J., 167, 173, 176, 183, 184(19,44) Mpez-Espinosa, M. T. P., 41 Lopez-Vidriero, M. T., 348 Lorient, D., 22 1 Los, J. M., 237 Losyakova, L. S., 335 Lothrop, R. E., 214 Lovric, T., 332 Lowe, L., 280 Lowry, T. H., 237 Lu, P., 136 Lucken, E. A. C., 109, 1 lO(175) Ludvig, L., 28 I Luger, P., 61, 61(77), 63(72), 65(72, 77). 71. 73,98,99( I50), I 18(150) Lugowski, C., 175, 179(29), I84 Lukehart, C. M., 135, 136(77) Lukin, N. D., 322 Lukina, 0. I., 256, 257(401) Lukss, R., 247 Lung, M., 37 Luthman, K., 35 Luu, D. V., 50,92(22) Lvov, V. L., 175, 178, 180(39), 198 Lygin, E. S., 239

M McAllister, R. V., 280 McCarthy, J. L., 37 McCombie, S., 19 McConnell, H. M., 10 Macaskill, J. B., 126 Macgillivray, A. W., 204,224(1), 226(1) Maciel, G., 253 Maciel, G. E., 244 Mackawa, A., 268 McWain, P., 225 McWeeny, D. J., 256 Madden, J. K., 31,33(109) Maekawa, S., 332 Maeshima, T., 327 Magalhaes, A. J., 234 Maglitto, C., 206 Magrian, L., 235

397

Maillard, L. C., 227,255(181) Makide, Y., 239 Maklashin, A., 222 Maksimchuk, V. P., 326 Maletsky, A., 323 Mall, A. S., 352, 353(35), 361(35) Mallams, A. K., 19 Manganel, M., 270 Mank, V. V., 215 Manley-Hams, M., 255 Manoharan, M., 55,61(45), 65(45), 115(45) Mantle, D., 354, 356(38), 359(38), 363(38), 364(46), 37 1(38), 376(46), 377(46) Mantle, M., 353, 354, 355(36, 38), 356(36, 38), 359(38), 363(38), 364(38), 371(38) Marakami, R., 328 Marchal, J. P., 77, 90( 107) Marchessault, R. H., 52,70(28), 7 1(28), 76(28), 303 Marcusson, J., 245,249,276(359) Marinelli, R., 222 Markham, R., 201 Marks, K., 38 I Marlewska, B., 322 Marquardt, R., 240,259(315) Martin, G. E., 237 Martin, J. C., 46,61(6), 67(6), 69(6), 70(6), 91(6), 107(6) Martin, L. F., 221 Martin, R. B., 35, 130, 135, 138(70,71) Martin, R. J., 117, 119(195, 196, 197) Martin, S. E., 269 Martin, U., 283,284(521), 286(522), 291(521, 522), 294(521), 297(521) Martinovic, K., 332 Martynenko, E. A., 231 Masaji, O., 39 Mashilova, G. M., 175, 177, 180(26), 18I, 182(37), 183(38), 196(26), 198 Mass, S., 216, 221(84) Massol, L., 280 Mathlouthi, M., 50,92(22) Matsubara, Y., 327 Matsuda, T., 269 Matsudaira, M., 33 1 Matsuhara, K., 332 Matsumoto, T., 324 Matsuoka, A., 268 Matsuoka, C., 268 Matsuura, H., 81,98(122) Matty, A. J., 335

398

AUTHOR INDEX

Mattyasovszky, P., 235 Matulevicz, C., 5 Matyas, J., 3 15 Matynenko, E. Ya., 220 Maury, P., 249 Mayauna, S. M., 326 Mayuzumi, T., 329,344(8 19) Mazurenko, A. N., 335 Mazza, G., 214 Medcalf, D. G., 332 Meier, E., 322 Meinecke, K. H., 259 Melaspina, A., 232 Melberg, S., 88, 105 Meleshko, V. P., 215 Menger, A., 33 1 Merlin, A., 280 Metzger, J., 306, 338 Meutzel, W., 328 Meyer, B., 59, 105(60) Meyer, F. A., 347,349(12), 350(12), 352(12), 354(14), 361(14), 363(14), 377(12) Meyer, H., 245 Mian, N., 38 1 Miawald, W., 303 Michaelis, L., 237 Micheel, F., 170 Michel, J. P., 225, 280 Mifune, K., 321 Mihailovic, J., 22 1 Mihara, M., 224 Mikelsone, A., 247 Mikkelsen, A., 367, 372( 100) Mildvan, A. S., 136 Milkova, Z. A., 239 Miller, D. C., 304 Miller, H. R. P., 346 Miller, J. D., 328, 329 Mills, J. A., 3, 4, 5, 11(15), 14, 13(13, 15), 15(13), 16(15), 21, 35(13, 15,48), 36(13, 15), 38,41(48), 42, 126, 127 Milne, E. A., 308 Milthorpe, B. K., 367 Miltome, N., 280 Minayav, R. M., 106 Minemura, S., 331 Minkin, V. I., 88, 106 Miroshnik, L. V., 325 Miroshnikova, E. P., 216,241(81), 242(81), 243(81)

Mishev, P. Ya., 220,23 1 Mishon, F., 175, 185 Miskovic, L., 22 1 Misra, D. S.,219, 226, 241(100) Mitchel, D. J., 58, 77(58), 80(58), 81(58), 110(58) Mitchell, H. L., 299, 300(572) Mitchell, J. R., 358, 380(60) Mitschler, A., 102 Mitsui, F., 332 Mittelbach, M., 236 Miyazaki, M., 19,29(63) Miyazawa, T., 9 Mizumo, S., 255 Mo, F., 24 Mochizuki, Y., 283,286(528), 289(528) Modi, R. P., 224 Moeckel, P., 280 Moggridge, R. C. G., 174, 179(24) Molinski, S., 215,216,218,234(94), 241(80) Molotsky, H. M., 229 Monk, C. B., 2 1 Monoroski, R. A., 132 Montgomery, R., 239 Montreuil, J., 352, 372(34) Moore, R. H., 249,307 Moore, W. C., 205 Mootsky, H. M., 229 Mora, P. T., 249 Morel, J.-P., 27, 29, 30 Morel-Desrosiers, N., 29, 30 Morgan, A. R., 46, 52(7), 61(7), 63(7) Mori, S., 329, 344(819) Morinaka, K., 255 Morita, H., 299,300(573) Moriya, M., 332 Moroyu, S., 280 Moms, E. R., 31, 33 Moms, J. L., 38 Moms, S. G., 321 Momson, G. A., 46,47( 1) Mort, A. J., 168, 170(4), 171, 172, 173, 174(10, 1 l), 178, 186(1 I), 189(23), 190, 192, 193, 199(55),200(1 I), 201,202 Moskvicheva, E. P., 335 Mouat, B., 126 Muamo, M., 328 Mukhejee, M., 335 Mukkur, T. K. S., 361, 363(68) Munavu, R. M., 18 Munns, D. J., 214

AUTHOR INDEX

Murakami, H., 270 Murata, H., 80,81,98(122) Myers, D. V.,229 Myers, G. E., 244 Myers, G., 253

N Naberezhnych, G. A., 187 Nagabushan, T. L., 19 Nagahashi, G., 40 Nagao, M., 3 16 Nagashima, N., 299 Nagpurkar, A. G., 63 Nakajima, A., 332 Nakamura, K., 37 Nakamura, Y., 230 Nakayasu, K., 224 Nakhapetyan, L. A., 37 Nanni, R., 52 Naray-Szabo, G., 47, 84(12), 85( 12), 89, 90(12), 92(12), 98(12), 102(12) Natta, G., 205 Neely, W. B., 88 Nelson, D. A., 303 Neltner, S.,324 Nemoto, K., 23 1 Nessler, J., 234 Neuberger, A., 174, 179(24) Neuman, S., 322 Neuzil, R.W., 39 Nichol, L. W., 367 Nichols, R.,38 1 Nickerson, T. A., 30 Nicol, E., 3 12 Nicolov, R.,22 1 Niedner, I., 323 Nikiorova, V. N., 2 18, 222, 223( 140), 240(89), 24 l(89) Nikolov, Z., 28 1 Nishi, U.H., 330 Nishida, K., 240 Nishimura, S., 19,29(63) Noguchi, Y., 39 Nomura, D., 229 Noonan, J., 205 Norris, E. K., 16 Norris, K., 20 Nsrskov-Lauritsen, L., 80,8I( I 18), 106, 114, 120(180)

399

Noumi, T., 268 Nowakowska, K., 314 Nowicki, B., 329 Nowotna, A., 288 Nozaki, K., 327 Nunez, O., 1 17 Nurnsten, H. E., 332 Nyhammar, T., 269

0

OBrien, E., 38 ONeill, A. N., 306, 3 10 Oates, K., 367, 368(101), 369(101), 378(101) Obretenov, Ts., 223 Odaka, Y., 324 Odawara, H., 39 Oddon, Y., 144 Ogawa, H., 330 Ogm, T., 268 Ogiwara, Y., 301,302,315 Oh, S. K., 223 Ohasawara, T., 23 1 Ohashi, M., 3 1 Ohira, T., 222, 226(126, 128), 228,332(127) Ohno, H., 327 Ohsaku, M., 80,8 I Ohta, T., 224 Ohtaki, H., 29 Okada, N., 224 Okamoto, H., 269 Okubo, N., 19, 29(63) Okumura, M., 330 Oldham, J. W. H., 306 Olds, D. W., 306 Olin, A., 27 Ollis, J., 18, 25(58) Olson, K., 269 Olson, R.,269 Omura, H., 270 Omura, M.,329, 344(818) Ona, M., 39 Ona, S., 280 Onishi, H., 269 Onodera, H., 268 Ooishi, H., 270 Opella, S. J., 136 Oreshko, V. F., 225,280 Orlova, I. G., 325 Orsi, F., 215, 235(65), 238, 249, 332(360)

400

AUTHOR INDEX

Onville-Thomas, W. J., 90 Osaki, K., 31 Oshiao, S., 20 1 Oshima, R., 2 15 Ostroumov, Yu. A., 88 Oswald, N., 327 Ota, K., 328 Ouellette, R. J., 5 5 , 61(37), 85(37), 115(37) Overend, W. G., 169 Ovodov, Yu. S., 187 Oyanagi, K., 95

P Paal, C., 266 Pace, G. W., 231 Pachler, K. G. R., 132 Pacsu, E.,237 Pagenkopf, J., 2 15,235 Pai, J. S., 224 Pain, R. H., 354, 356, 361(47), 364(46,47), 373, 375, 376(46,49), 377(46) Paine, A. J., 267 Paine, H. S., 214,221 Paisley, G. V., 232 Palash, I. P., 239 Palasidski, M., 219,228(98,99), 245(98, 99), 247(98,99), 280,288,290,29 1, 301(545), 303(556, 557), 305(556, 557), 315, 319(659) PalovEik, R., 118 Pancek, M., 3 17 Panshov, I., 223 Pantano, V.,206 Paramonov, N. A., 182, 195 Park, H. H., 256 Parke, D. K. K., 231 Parker, S., 168, 170(4), 172(4) Parmar, R. S., 286,289(536), 290(536), 294(536), 295(536), 296(536), 298(536), 299(536), 308(536) Panish, F. W., 38,42 Parshina, T. A., 216 Partain, C. C., 135, 136(77) Partain, V. M., 135, 136(77) Paschall, E. F., 222,281, 315(132) Patai, S., 249,274 Patey, A. L., 214, 235, 236(267), 253, 255 Paul, M. A., 253

Paulsen, H., 52, 57(35), 61,63(35, 72), 65(35, 72, 77), 71, 73,98,99(150), 118(150), 274,276(470) Pavia, A. A., 46,61(6), 67(6), 69(6), 70(6), 91(6), 107(6), 132, 134, 144, 146(68), 147(68), 148(68), 149(68), 153(65), 154(65), 155(65) Pavlath, A. E., 238 Pavlov, I. I., 215 Pearce, J., 334 Pearl, D., 319 Pearson, J. P., 353, 354, 355(36,40), 356(36), 361, 363(69) Peciar, C., 21 Peck, F. W., 216,218(85), 229(85) Pedersen, C., 17, 168, 169(5), 170(5,6), 171, 172(5), 202 Pedersen, K. O., 356 Pelham, B., 126 Peligot, E., 204 Pelmore, H., 9 Peng, Q.-J., 35 P&pe,G., 144 Pepper, T., 327 PCrez, S., 23, 52, 70(28), 71(28), 76(28), 98, 99(152), 101, 105(154), 106(154), 126 Perisic-Janjic, N., 328 Perkin, A. G., 241 Perkins, A. T., 299, 300(572) Perkins, E. G., 269 Perlin, A. S., 35, 70, 88, 134, 172, 174(16) Peronnet, M., 215,222(64), 223(64) Perrand, R., 55,61(43), 67(43), 115(43) Penin, C. L., 117 Perry, M. B., 179 Persoz, J. F., 282 Pertot, E., 286 Pestana, F., 335 Peters, J. A., 9, 11, 31 Petersson, K., 176 Petit, M. P., 282 Petr&kovl, E., 70 Petrosyan, Ts. L., 23 1 Petrov, L., 36, 37(67), 39, 22 1 PetruS, L., 20 Petuely, F., 206,218(26), 223(26), 235(26) Pew, J. C., 201 Pezacki, W., 330 Pfannhauser, W., 206,233(27) Phillips, D. C., 43

AUTHOR INDEX Phillips, G. O., 358, 380(61) Phillips, N. C., 322 Pickles, V. A., 5, 7, 8(22), 11(17), 14(22), 15(22), 16(17), 35(17), 41, 127 Picon, M., 272, 336(469), 339(469) Pictet,A., 216,219(83),226(83),244(83), 306 Piedad-Pascual, F., 334 Pier, G. B., 189 Pierrotti, R. A., 90 Pierson, G. O., 55,61(39), 67(39), 1 15(39) Piengalski, T., 223,327 Pietrasik, D., 181 Pietrusiewicz, K. M., 55, 61(45), 65(45), 84(46), I 15(46) Pigman, W. W., 50,92( l9), 348, 349(22), 350(22) Pike, E. R., 375 Piller, F., 3 I6 Piperova, L. S., 334 Pisters, M. G. H., 232 Pitchon, E., 330 Piva, G., 318, 321(683) Pojer, P. M., 18, 25(58), 38, 42(154), 126 Pokorny, J., 249 Politowski, M., 322 Poncini, L., 205,230,239,245(302) Ponee, R., 106 Pople, J. A., 71, 76(94), 77, 78(101, 102), 80, 82(94), 84(94), 86(94), 97(94), 109, 1 14(I 1 1) Popoff, T., 253 Pothier, N., 61 Potier, M., 354 Potze, J., 319 Powell, D. A,, 176 Powers, R. M., 336 Powrie, W. D., 256, 268, 270, 305(400) Pozderovic, A., 332 Prestegard, J. H., 126 Preston, B. N., 357, 37359) Preston, C. M., 26, 257 Pretsch, E., 27 Prey, V., 272,273(466), 275(466), 276 Price, A. C., 135, 136(77) Priegnitz, J. W., 39 Pringsheim, H., 306 Prins, J. W., 306, 308(624) Procter, A. A., 304 Prohaska, R., 126 Prohaszka, O., 332

401

Pronin, A. Ya., 37 Protzman, T. F., 336 Provencher, S. W., 375 Pruzinec, J., 280 Puchelle, E., 380 Puddington, I. E., 304,308(608), 336(608), 340(608) Pugmire, R., 244,257(335) Puigjaner, L. C., 34 Pullman, A., 109 Pullman, B., 22,24(76), 26(76), 89 Pullmann, A., 28 Purvinas, M. M., 313 Pusey, P., 375 Pustek, F. J., 319 Putilova, I. N., 225, 280(167) Pyr6, G., 3 15, 3 19(659) Pyrcz, J., 330

Q Quinn, E. J., 249, 307

R Radhakrishnamurthy, B., 334 Radley, J. A., 280, 281, 282(506), 283(506), 286(506), 288(506), 295(506), 3 18(506), 322(506), 330 Radom, L., 77,78(101, 102), 80, 114(11I) Radosavljevic, S., 328 Raemy, A., 238 Raffi, J. J., 225, 280, 305 Rakhimbaev, Sh. M.,324 Ramaiah, N. A., 218,226,237 Ramaswamy, P., 120 Ramaswamy, S. R., 228 Ramchander, S., 240 Ranby, B., 366 Randall, E. W., 132 Rane, D., 19 Rankin, A. F., 31 Rao, C. N., 335 Rao, V. S. R., 57, 59, 80, 81(1 IS), 104(55), 105(59) Raouf, M. S., 205 Rasmussen, J. R., 38 Rasmussen, K., 88, 105

402

AUTHOR INDEX

Rauk, A., 106 Reboul, J. P., 144 Record, B. R., 373 Reed, R. A., 239 Reeicardo, C. T., 334 Rees, B., 102 Rees, D. A., 3 I , 33,43 Reeves, R. E., 2 1,66 Rege, A., 221 Reichert, P., 19 Reid, G. T., 365 Reid, L., 348 Reilly, C. N., 9, 1 l(29) Reilly, L. W., Jr., 344 Reimann, S., 28 1 Rendleman, J. A., Jr., 2, 3, 22(4), 30(8), 327 Reuben, J., 9, 11(30), 13, 27(46), 28(46), 29(30), 30(46), 41(30), 128, 130(32), 13l(32) Reuterswaered, A. L., 269 Reynolds, G. F., 19 Reynolds, J. A., 347 Reynolds, T. M., 262 Rhodes, D. G., 366 Ricard, G., 55,61(44), 63(44), 67(44), 69(44), 115(44) Rich, H. W., 327 Richards, E. G., 359 Richards, G. N., 18,230,240,255 Richards, S. N., 239 Richardson, P., 381 Richon, D., 328 Richter, M., 289 R i m , B. F., 35, 126, 160(20) Rigamonti, R., 205 Rigonard, M., 225,280(168) Rinaldi, D., 77,90(107) Ripamonti, A., 52 Ripmeester, J. A., 257 Ripp, B., 215,219(62), 227(62) Roard, J. L., 282 Roark, D., 365, 377(72) Roberts, G. C. K., 130 Roberts, H. J., 321 Roberts, J. D., 3 13 Robertson, J. M., 102 Robertson, R. E., 10 Robson, T., 354, 356(39) Rockland, L. B., 303

Riiderer, H., 214 Romanovska, E., 176, 184 Romers, C., 47, 107(13), 109(13) Rona, P., 237 Rosdorfer, J., 236 Rose, M. C., 371 Rosen, D., 3 12 Rosin, M. P., 268, 270 Rossi, R., 206 Rossinskaya, G. A., 247, 250 Rother, H., 21 5,235(69) Rothfuss, H., 3 19, 323 Roushdi, M., 205 Rousseau, A. D., 344 Roussel, P. P., 352, 367, 369(94,97,98), 370(94), 372(33), 374 Rowan, D. D., 61 Rowbottom, P. M., 255 Rowe, A. J., 349,350(24), 351,354(25,30), 361(30), 362(30), 365, 366, 367, 368(30), 369(24), 371, 372(30), 374, 379(30) Rowe, W. J., 322 Roy, R., 175, 185 Rozenbroek, M. D., 32 1 Rozhkova, M. V., 235,239,256,257(401) Rubino, F., 22 1 Rubinsztain, Y., 244,257 Rudenko, N. V., 225,280( 166) Rudenko, V. N., 215 Ruggeberg, H., 294,322 Runge, V. M., 135 Runquist, 0. A., 55, 61(39), 67(39), I15(39) Russel, G. F., 225,256, 305(400) Ryu, B. H., 225

S

Sabaneev, A., 244 Saber, H. M., 235 Sacks, I. B., 201 Sacks, L. E., 335 Sadjera, S. W., 376 Saini, K. S., 361, 363(68) Saint-Lebe, L., 225,280 Saito, S., 256,257(402) Saitoh, S., 334 Sakai, H., 249,272(362)

AUTHOR INDEX Sakakibara, M., 8 1 Sakakibara, S., 173 Sakata, T., 278 Sakiyama, T., 37 Salamon, A, G., 237 Salem, L., 101, 110 Salgraikh, L. S., 336 Salzrnann, G. M., 284 Samec, M.,225,280(165),297,316,319 Sampei, Y., 257 Samuel, R. K., 3 19 San, R. H. C., 270 Sander, E. H., 332 Sandomini, C., 272, 335(468) Sanger, H., 245 Sanger, M. P., 172, 173(14), 174(14) Sapetti, C., 214 Sapronov, A. R., 2 15, 2 16, 2 17(61), 227(61), 235,239, 241, 244(61), 285 Sarasin, J., 306 Sarko, J. A., 303 Sasaki, K., 40 Sashkov, A. S., 188, 189 Sato, S., 324 Satoh, E., 269 Satoshi, J., 229 Sattelle, D. B., 357, 375(57) Saunders, J. K., 61 Sawada, K., 328 Sawada, M., 268 Scarsdale, J. N., 77, 80, 97(112) Scawen, M., 350, 352(26), 354(26), 355(26), 356(26) Schachman, H. K., 359,371 Schafer, D. E., 126 Schafer, L., 77,80, 97( 1 12) Scheer, M. D., 238 Schenk, D., 234 Scheutwinkel-Reich, M., 269 Schierbaum, F., 289,290, 303 Schiff, H., 244 Schilling, M.,235 Schimmel, P. R., 350, 35 1 Schirner, R., 319, 323 Schiweck, H., 204,223(2) Schleifer, L., 52, 71(30), 74(30), 76(30), 98(30) Schmalz, K., 306 Schmid, K., 352 Schmidt, E., 245

Schmidt, M., 370 Schmitt, O., 132 Schmut, O., 137 Schneider, H.-J., 55 Schneider, W., 327 Schoch, T. J., 284,286(531) Schoene, H.J., 235 Scholten, W. A., 321 Schotzbauer, W. S., 343 Schroeder, U., 324,325 Schueller-Richardson,K., 237 Schuerch, C., 116, 306,308(624) Schues, R. D., 252 Schuetz, R. D., 240,259(314) Schulman, M., 330 Schultze, H. Z., 253 Schultze, W., 331 Schulz, R. C., 301 Schumacher, D., 19 Schumacher, J. N., 253 Schweizer, T. F., 28 1 Schweizer, T., 238 Schwier, I., 303 Scordamaglia,R., 88 Scott, W. E., 33 Searle, F., 4, 21(12), 37(12) Secchi, G., 206 Seeman, J. I., 117 Seguin, F., 225,280(169) Seidemann, J., 289 Seino, Y., 316 Seiyaku, F. S., 324 Selemenev, V. F., 215,216(77), 224 Selke, S. M., 172 Sell, U., 235 Sellers, L. A., 352, 353(35), 361(35) Semenov, A. D., 225 Sengupta, S., 335 Sepp, D. T., 5 5 , 56, 61(38, 41, 51, 52), 65(51, 52), 67(51), 85(38), 115(38, 51, 52) Serafini-Fracassini, A., 366 Serenkova, I. A., 240 Senanni, A. S., 38 Serichenko, L. G., 335 Sestini, F., 24 1 Sevenn, T., 252,253(374) Shafizadeh, F., 240,335,338 Shallenberg, R. S., 50, 88(21), 92(21) Shamntskaya, I. P., 215, 224

403

404

AUTHOR INDEX

Sharma, S. C., 229,23 1 Sharpatyi, V. A., 305 Sharton, H. H., 229 Shashkov, A. S., 175, 177, 178, 180(26, 39), 181, 182, 183(38), 196(26), 198, 199 Shaw, C. J. G., 305 Shaw, J., 127, 128(27) Shaw, P. E., 240,253 Shearer, G., 214,235,236(267), 253,255 Shechan, J. K., 346, 347,348, 352, 353(13, 15, 31), 354(15, 31), 355(13), 356, 358(54, 55), 361(13,56), 363(13, 54), 364(15, 55, 56), 365(53,54, 55), 367, 368(101), 369(53, 54, 55, 56, IOl), 370(53, 54), 371(53, 54), 378(101) Shelton, S. A., 225 Shen, J., 236 Sherman, J. D., 39, 40(161) Shi, Y., 324 Shibamoto, T., 225, 315, 316 Shilling, W. L., 241 Shimizu, Y., 269 Shingler, A. J., 208 Shinohara, K., 270 Shiraishi, S., 332 Shiro, Y., 81 Shlapnikov, Yu. A., 240 Shogren, R. L., 347, 370(17) Shrivastava, H. C., 310 Shtyrkova, E. A., 319 Shumaker, J. B., 241 Sicard, P., 230 Sicker, L. C., 102 Siddiqui, I. H., 328 Siddiqui, I. R., 24 Sigel, H., 126, 128(2) Sikkema, D. J., 343 Silberberg,A., 347, 349, 350(12), 352(12), 353(25), 374(12), 377(12) Silberman, H. C., 252 Silverbrandt, M. H., 33 1 Simons, F. D., 234 Sinanoglu, O., 89 Singer, P. A., 32 1 Sinnema, A., 10, 14, 28(49) Skalski, J., 286,293,296(562), 297(560, 561, 562), 313(560, 561, 562), 314, 3 18(560), 322 Skaug, N., 175 Skripchenko,T. N., 336

Slayter, H. S., 367, 369(94, 95, 96, 97, 98), 370(94, 95) Slettengren, K., 195 Slocenn, L. A., 316 Smale, S. T., 38 Smedley, M., 376,380( 134) Smidsrd, O., 10, 11(33), 32,33, 367, 380 Smirnyagin, V., 50 Smith, D. S., 303 Smith, F., 306, 308(626) Smith, F. W., 135 Smith, I. C. P., 313 Smith, P., 306, 308(625) Smith, P. J. C., 33 Smoledski, K., 2 15 Smolnik, H.-D., 233 Snary, D., 355, 361(47), 364(47), 373 Snyder, J. R., 38 So, R. S. C., 331 Sobue, H., 280 Soff, K., 280 Sofuni, T., 268 Sokolovshii, A. L., 222, 223(140) Solms, J., 332 Solomina, L. S., 285,318,319,321,322(701) Solomon, I., 132, 133(57) Soloveva, T. F., 187 Somers, P. J., 21 Sopoleva, M. T., 326 %rum, H., 33 Sosedov, 0. N., 37 Soyer, V. G., 225 Spector, R., 268 Spellman, M., 175 Spingarn, N. E., 316 Spoormaker, T., 14, 28(49) Spragg, S. P., 366 Sprung, M. H., 266 Srinivasan, S. R., 334 Srivastava, H. C., 286,289(536), 290(536), 294(536), 295(536), 296(536), 298(536), 299(536), 308(536) Srivastava, S. C., 326 Srivastava, S. W., 326 Sroczynski, A., 222,286,293,296(562), 297(560, 561, 562), 313(560, 561, 562), 314, 318(560), 322 Sroczydski, A., 327 Stabinger, H., 361 Stadnikoff, G., 222

AUTHOR INDEX Staerkle, M. A., 322 S a l , E., 272, 336(465) Stah'naya, I., 37 Stanek, V., 255 Stanislavsky, E. S., 175, 177, 180(26, 39), 181, 182, 183(38), 196(26), 198 StankoviE, L., 20, 37(67) Stanley, C. J., 357, 375(57) Startin, J. R., 255 Stasinski, J., 322 Stefanac, Z., 27 Steinhaus, R. K., 29 Steinitzer, F., 3 19 Steinke, J., 227 Stephenson, N. C., 22 Sterk, H., 137 Sterle, H., 132 Stevens, J. D., 22, 23, 26 Stevenson, J., 2 1 Stevenson, K. W., 319 Stewart, G. F., 303 Stewart, J. M., 173 Stewart, R. G., 135 Stich, H. F., 268, 270 Stich, W., 268 Stickley, E. S., 280 Sticzay, T., 1 I8 Stinson, E. E., 235 Stockmayer, W. H., 370 Stoddart, J. F., 46,47(2), 48(2), 56(2), 69(2) Stoev, G., 324 Stokke, B. T., 367, 372(100) Stolle, F., 226 Stolzenberg, H., 255 Stoodley, R. J., 19 Stout, E. I., 325 Straub, A., 234 Strecker, A., 259 Streefkerk, D. G., 38 Stroh, H.-H., 289 Stropnik, c.,283,286 Stute, R., 274, 276(470) Stutzke, R. W. G., 321 Sugawara, S., 272 Sugimura, T., 3 16 Sugisawa, H., 235,259(266), 272 Sultankhodzhaeva, M. N., 305 Sultanov, A. S., 223, 319(144) Sundall, S., 208 Sundaralingam, M., 7 1

405

Surova, N. S., 6 1 Sutton, L. E., 71,76(93) Suzuki, H., 325,366 Suzuki, M., 326,334 Suzuki, S., 283,289(525), 329, 344(816) Svedberg, T., 356 Svensson, S., 167, 175, 179(29) Sviridov, A. F., 168 Swanson, A. R., 329 Swanson, M. A., 309 Swartz, M. L., 30 Sweet, F., 61 Sweigart, D. W., 101, 114(155) Sykes, B. D., 132 Symes, K. C., 30 Symons, M. C. R., 9 Syroedov, A. I., 322 Szarek, W. A., 47 Szejtli, J., 116, 118(185), 281 Szmant, H. H., 18,230 Szponar, E., 334

T Tafuri, S. R., 38 Taga, T., 31 Taguchi, K., 257 Tahara, Y., 175, 177(31) Taj, H., 336 Tajmir-Riahi, H. A., 12, 126 Takagi, Y., 222,226(126,128),228,332(127) Takahashi, M., 270 Takahashi, M. Y., 236 Takahashi, N., 249,272(362) Takahashi, T., 33 1 Takai, N., 2 15 Takekoshi, Y.,331 Takinishi, H., 327 Tam, P. Y., 373, 380(116) Tamai, H., 329, 344(8 16, 8 18) Tamura, S., 269 Tanaka, M., 329, 344(818) Tanaka, Y., 28 1 Tanford, C. F., 356, 374 Tang L. H., 367 Tanigawa, H., 268 Taraglione, E., 2 17 Tarantola, C., 215, 235(71) Tarasov, S. G., 318

406

AUTHOR INDEX

Tartaglia, P., 357 Tartakovsky, E., 52, 71(30), 74(30), 76(30), 98(30) Tate, M. E., 35 Tatum, J. H., 240,253 Tiiufel, K., 303 Taylor, R., 106 Taylor, T. C., 284,286(530),293(530) Taylor, W., 361,363(69) Tel, L. M., 106 Telegdy-Kovats, L., 249, 332(360) Teller, D. C., 359, 365, 375(73) Templeman, G. J., 229 Tempte, C. S., 38 1 Thanomkul, S., 33 Thawait, S., 37 Theander, O., 253,303 Theobald, N., 327 Thiele, H., 222 Thier, E., 208 Thmersen, H., 59, 70, 105(60) Thom, D., 31,33 Thompson, A., 239,249,306,307, 308(628), 3 10 Thompson, S. O., 299 Thomsen, M., 236 Thornton, E. R., 34, 35(127), 126, 16q17) Thurman, L. S., 208 Tibbets, M. S., 229 Tiefenbacher, K., 206, 233(27) Tiesjema, R. H., 173 Tillin, S. J., 315 Timberlake, C. E., 239 Tindall, S. H., 367 Tipson, R.S.,249 Tobin, R., 3 13 Tobolina, V. M., 216, 241(81),242(81), 243(81) Tochtamysheva,N. V., 175 Togaya, T., 326 Toman, R., 2 1 Tomasik, P., 219,228(98,99), 232,245(98, 99), 247,280,291,299, 300(558,576), 303(556, 557, 576), 305(556, 557, 558), 315, 319(659) TomaSkoviE, M., 27 Tomita, I., 270 Tom, G., 172, 174(16) Tosi, C., 88 Touwslager, F., 253

Trahaut, R., 214 Traubenberg,S. E., 225,280(167) Tregubov,N.N., 301,318,319,321, 322(701) Trillat, A., 245 Trubiano, P. C., 331 Truchot, E., 230 Truhaut, R., 205,223(9) Truter, M. R., 71,76(93) Tsai, H., 19 Tsuchida, E., 327 Tsuchida, H., 255,257,269 Tsuchiko, T., 329,344(817) Tsuji, S., 283,286(528), 289(528) Tsutsui, M., 240 Tucker, M. Y., 40 Turner, D. W., 101, 114(155) Turner, W. N., 19 TvaroSka, I., 47, 52, 53(1 I), 55,61(46), 67(46),70, 71(29), 76(29), 77,80, 81( 117, 119), 83( 113), 84,85(12, 46, 126, 129), 86(100), 88,90(12, 117, 119, 127, 137, 138), 92(12, 127, 137), 94(11, 127, 130, 132), 95( I32), 96,97( I 1, 148), 98,99(151), 101, 102(12), 103(128), 104, 105(61, 154), 106(132, 154), 109(161), 111(148), 113(148), 114(113), 115(46, 127, 128, 129), 118, 120(148), 121(148) Tyagunova, V. I., 235,239,256,257(401) Tyree, J. T., 3 10 Tyrer, J., 365

U

Uchida, T., 60, 79(64) Ueno, Y., 283, 285(524), 289(524), 301(524). 3 12(524) Uitada, K., 201 Ujszarzi, J., 215, 235 Ullmann, M., 222 Ulman, E. A., 334 Ulmann, M., 289,303 Ulmsten, U., 348 Umadevi, S., 3 17 Umano, K., 3 15 Unger, P., 175 Unrau, A. M., 237 Unruh, C. C., 317, 319(672)

AUTHOR INDEX Uobe, K., 240 Upadlyay, S. K., 232 Usenko, K. B., 326 Usmanov, Kh. U., 218,223,319(144) Utille, J.-P., 172, 174(16)

V Viclavik, L., 88,93( 132), 94( 132), 95( 132), 106( 132) Valter, V., 256 Valvassori, S., 214 van Alsenoy, C., 77, 80,97(112) van Bekkum, H., 10, 14,28(49), 31, 39, 230,252(217), 253 van Benschop, H. J., 27,29(98) van Dam, H. E., 230,252(217), 253(217) van der Baan, H. S.,252 van der Hude, W., 269 van der Kaaden, A., 173 van der Poel, P. W., 253 van der Toorn, J. M., 10, 14,28(49) van Eckert, R., 206,233(27) Van Halbeek, H., 352, 372(33, 34) van Holde, K. E., 37 I van Landschott, G., 224 van Leeuwen, L. K., 27,29(98) Vandewalle, L., 224 Vangehr, K., 61,63(72), 65(72) Vankatesha, T. V., 326 Varga Kiss, I., 332 Varma, M. C. P., 299,300(577) Vasella, A., 63 Vauquelin, L. N., 28 1 Venables, C. W., 354, 355(40), 361(40) Venema, A., 232 Verberne, P., 322 Verdagner Bronsoms, J., 328 Verdugo, P., 373, 380(116) Veremeichenko, S. N., 188 Vereshchagin, A. N., 8 1 Vergelati, C., 98, 99( 152) Vernon, G. N., 132 Verstraeten, L. M. J., 252 V e d a , A., 12, 16, 20(41, 42, 44),21(41), 27,28(52, 53,95), 30(52) Vezhbitskii, F. R., 247 Vich Figa, M., 328 Vijayalaksami, K. S., 57, 104(55)

407

Vijberg, C. A. M., 31 Vinogradov, E. V., 175, 177, 178, 180(39), 181, 182, 183(38), 187, 195, 198 Visek, W. J., 334 Vishveshwara, S., 71, 76(94), 77,80, 81( I 1 9 , 82(94), 84(94), 86(94), 97(94) Vitte, G., 205, 223(9) Vliegenthart, J. F. G., 173, 352,372(33,34) Vlot, T., 322 Vnogradov, E. V., 181 Voisin, D., 47 Volker, H. H., 280 Volkov, V. A., 324 Volmer, C. R., 229 Volpe, S., 237 von Elbe, G., 214,215(44), 218(44) Voter, W. A., 371

W

Wagner, A., 172 Walaszek, Z., 41 Wales, M., 37 1 Wall, R. A., 37, 38( 137) Wallace, G. R., 346 Wallace, R. G., 373 Walton, H. F., 40 Walton, R. P., 282 Wan, C. N., 270 Wang, B., 334 Wang, C. R., 222 Wang, Y., 269 Wankhede, D. B., 3 I7 Wankler, B. N., 330 Warchol, M., 288 Ward, R. B., 307, 308(628) Ward, R. P., 2 18 Ware, B. R., 357, 375(57) Watanabe, K. A., 46,61(6), 67(6), 69(6), 70(6), 91(6), 107(6) Waterman, H. J., 205 Waters, P. L., 299,300(575) Watson, D. L., 361,363(68) Webb, V. L., 57 Weckel, K. G., 227 Wedlock, D., 358,380(61) Wehr, C. J., 135, 136(77) Wei, C.-I., 316 Weidinger, A. W., 298, 306(570)

408

AUTHOR INDEX

Weigel, H., 3,4, 5,21(12), 35(11), 37(12) Weinman, H. J., 135 Weinmann, H. J., 328 Weinstein, J., 19 Weist, R., 102 Wells, A. G., 30 Wells, H. G., 16 Weltner, W., Jr., 34, 35, 126, 129, 160(14) Wenn, R. V., 2 I5 Wentzel, C., 236 Werkenthin, M., 215,255(58) Werner, M. A., 232 Wertz, Z., 322 Whangbo, M. H., 58,77(58), 80(58), 81(58), 1 lO(58) 67 Wharpy,'S. 'M., Wheeler, T. S., 282, 294(517) Whistler, R. L., 222, 28 1, 3 15(132) White, A. I., 132 White, J., 2 14 Whiteford, R. A., 101 Whitehead, M. A., 88 Whittington, S. G., 32 Wickberg, B., 252 Wickens, J., 227 Widdicombe, J. G., 38 1 Wiejak, S., 2 19, 228(98, 99), 245(98, 99), 247,280,291,299,300(558, 576), 303(556, 557, 576), 305(556,557, 558), 315, 319(659) Wiemalm, G., 186 Wiesner, K., 237 Wiggins, L. F., 239 Wilbur, D. J., 22 Wilkinson, S. G., 175, 177(31) Wilks, R. A., 208 Willaman, J. J., 214 Willi, A. A., 33 1 Williams, C., 22 Williams, J. D., 367 Williams, J. O., 77, 80,97(112) Williams, J. W., 367 Williams, P., 358, 380(61) Williams, R. C., 366 Williams, R. J. P., 11, 130, 131, 136(39), 137(39) Willits, C. O., 235 WillstBtter, R., 241 Willumsen, D., 236,268

Wingerup, L., 348 Winkler, S., 297 Wischnack, W., 232 Woelter, W., 38 Woidich, H., 206,233(27) Wolfe, S., 58, 77(58), 80(58), 81(58), 106, 1 lO(58) Wolff, I. A., 306 Wolff, O., 293 Wolfrom, M. L., 22 1, 225, 239, 240,241, 249,252,253,259(3 14), 306, 307, 308(628), 3 10 Wong, R. Y., 304 Wood, F., 208 Wood, J. W., 249 Wood, R. A., 24 Woodward, H., 365 Woolfenden, W. R., 244,257(335) Wortel, T. M., 39 Wright, J. J., 19 Wu, C. H., 256,268,270,305(400) Wulken, H., 321 Wurmser, R., 280 Wurzburg, 0. B., 28 I , 302 Wiithrich, K., 130 Wyatt, G. K., 380 Wyatt, P. J., 380 Wycherley, V., 2 1 Wyler, R., 332

X Xavier, A. V., 130, 136(39), 137(39) Xiong, J., 268

Y Yadosaki, K., 329,344(8 17) Yahagi, T., 3 16 Yakubova, M. A,, 305 Yamada, H., 324 Yamada, S., 299 Yamaguchi, Z., 325 Yamamura, H., 223 Yamashita, K., 332

AUTHOR INDEX

Yamashita, M., 270 Yamashita, N., 270 Yamauchi, A., 215 Yang, R., 228 Yang, T. Y., 325 Yano, T., 37 Yashiro, M., 334 Yates, J. H., 71, 76(95,96), 80(95), 96(95) Yeh, C., 127, 128(27) Yocom, P. N., 129 Yokoyama, T., 324 Yoshida, A., 19, 29(63) Yoshida, D., 269 Yoshida, T., 331 Yoshihara, M., 327 Yoshikawa, K., 268 Yoshimatsu, F., 247 Young, B. G., 249 Young, J. D., 173 Yphantis, D. A., 365, 366, 367, 377(72) Yu, A,, 239 Yu, G., 224 Yu, R. K., 126 Yu, Y., 267 Yuan, H., 324 Yurev, Yu. K., 267

409 Z

Zahm, J. M., 380 Zaidi, S. A. H., 328 Zakharova, I. Ya., 188, 199 Zartman, W. H., 205 Zaslow, B., 43 Zatloukalova, V., 332 Zawadzki, W., 315 Zdorovenko, G. M., 199 Zechmeister, L., 241 Zefirov, N. S., 46,47(3), 61, 1 17, 1 19(3) Zegota, A., 225 Zempltn, G., 306 Zenouz, A. A,, 215,235(66), 256 Zerban, F. W., 221 Zhang, S., 228 Zhdanov, Yu. A., 88, 106 Zhukhman, A. I., 299, 300(574), 301, 322 Ziegler, C., 322 Zielinski, A,, 227 Zipfel, W., 205 Zkorovenko, G. M., 188 Zoller, H. F., 221 Zurth, C., 328 Zwierzan, G. A., 331

This Page Intentionally Left Blank

SUBJECT INDEX A

Alditols complex-formation, I3 - 14 Alcaligenesfaecalis, 0-specific polysacchaisolation, 179- 180 ride, 199 separations on ionexchange columns, 38 2-Acetamido-2-deoxy-D-galacturonamide, 177- 178 Aldohexopyranoses, complex, axial2-Acetamido-2deoxyguluronicacid, equatorial equilibrium, 59-60 glycosidic linkage, 185 Aldohexoses, reactions of glycosidic link2-Acetamido-2deoxymannuronicacid, ages, 169 glycosidic linkage, 183- 184, I86 Aldopyranoses 2-Acetamido-2,6dideoxy-D-galactose, 186anomeric equilibrium, ring substituent 187 effects, 48 - 50 2-Acetamido-2,6-dideoxy-L-galactosyl orientation, 48, 5 I residues, glycosidic linkages, 188 Aldopyranosides, valence geometry 2-Acetamido-2,6dideoxyhexoses,glycosidic parameters, 72 linkages, 188- 189 Aldosylamino acids, 263 2-Acetamido-2,4,6-trideoxv-4-Io-3Alduronate ions, complex-formation, 3 1 - -. , hydroxybutanamidol-Dglucose, Alduronic acids, 3 1 -35 175-176 binding strength, 33- 34 5-N-Acetyl-7-N-formylpseudaminic acid, egg-box model, 33 178 polysaccharides containing, 177 N-Acetylfucosamine,glycosidic linkages, quinquedentate complex, 34- 35 195-196 tridentate complexation, 3 1 N-Acetyl-Dgalactosamine, 195,200-20 1 Alginate, 32 N-Acetylglucosamineresidues, 181- I82 Alkali-lability, dextrins, 293, 295 4-(N-Acetylglycyl)amido-4,6dideoxy-D-glu- a-D-Allopyranosyl a-D-allopyranose, 18 cog, 175-176 a-D-Allopyranosyl a-D-allopyranoside, 25 N-Acetylhexosamines,glycosidic linkages, Ames test 188 caramel, 267 N-Acetyl-&newamink acid, 34, 158- 165 melanoidins, 269 Amino acids binding to Gd”, computer modeling, 160- 161 in mucus glycoproteins, 349 Gd3+effect on l3C-n.m.r., 159- 160, reactions with sugars, 223 164-165 a-Amino acids, dextrinization in presence interaction with Ca2+and Eu”, 160 of, 315 Mn2+effect on ’F-n.m.r., 162- 163, 165 Amino sugars, 19 structure, 158 N-acyl substituents, 179 3-(N-Acetyl-L-seryl)amidc-3,6dideoxy-Dreaction with methanol, 202 glucose, 175- 176 Amylolytic index, dextrins, 320- 32 I Acid factor, 3 18 Amylopectin Acyclic model compounds, conformational degradation products, 3 10 energy, 77-81 rate of dextrinization, 286,288 Adhesives, dextrins as, 323 Amylose Aglycon pyrolysis, 340 character, 62 - 65 rate of dextrinization, 286,288 stagered orientations, 50- 5 1 I ,6-Anhydro-3,4dideoxy-&D-glycer~hexAlditol-cation complexes, crystal structures, 3-enopyranos-2-ulose, 338 26 2,5-Anhydrogalactitol, I6 41 1

412

SUBJECT INDEX

1,6-Anhydro-&Dglucofuranose,formation, 340 1,6-Anhydro-&Dglucopynose, 16,340 1,6-Anhydro-fiD-glucose,306 1,6-Anhydro-&D-mannopyranose,2 I 1,5-Anhydro-2,3,4-tri-~t~nzoylxylitol, 63 Anomeric effects, 46-47, see also Isomers (up,up)conformation, 1I 3 - 1 14 (uppg) conformation, 1 13 - 1 14 carbohydrate reactivity, 1 16 antibonding orbital, 120 ground and transition states, 1 18 lone-pair orbital interactions, 1 I9 reaction-path energy, 116- 118 definitions, 58 delocalization interactions, 109- 114 scaling with electrostatic interactions, 114- 115 electrostatic interactions, 107 109 scaling with delocalition interactions, 114- 115 energy difference between up and sc orientations, 114 energy of, 53- 59 A parameter, 53, 57 measures, 58 2-substituted oxanes, 55 - 56 methoxyl, hydroxyl, and aziridinyl groups, solvent effects, 68-69 molecular-orbital calculations, 75 - 103 a6 initio methods, 75 anomeric energy effect, 93-97 bond length and angle coupling to torsional angles, 98- 101 conformational enerpies, see Conformational energy electron distribution and lone pairs, 101- 103 molecular geometry optimization, 75-76 solvent effect, 88-93 nature of, 106- 107 potential-function calculations, 103- 106 reverse, see Reverse anomeric effect solvation-energy terms, 1 15 valence geometry parameters, 7 1 -75 aldopyranosides, 72 -73 bond length and angle patterns, 7 1 -72 C-0-C bond, 73-74 D-xylopyranosyl fluorides, 73

-

Anomeric energy, from MO calculations, 93 - 96 solvent dependence, 93-94 transition from AE to AG, 94- 95 Anomeric equilibria, 48 - 53 aldopyranose orientation, 48,5 1 anomenc effects, 52 - 53 C-4 substituent electronegativity,66 C-5 substituent effects, 66 A2 effect, 66 exo-anomeric effect, 48 - 52 a-and &Dglucopyranose derivatives, 48-49 reverse anomeric effects, 52 -53 ring substituent effects, 48, 50 substituted oxanes and cyclohexanes, 54 Antiperiplanar, lone-pair hypothesis, 119-120 Antiulcer agent, dextrins as, 333-334 A parameter, 54 &DL-Arabinopyranose, oxygen atoms, 102 Aspartame, stabilizing properties of caramel, 23 1 Aziridinyl groups, anomeric effects, solvent effects, 68-69

B Beta amylase, 298 Binder, dextrins as, 324- 325 Blue staining residue, dextrins, 283 Bond angles, coupling to torsional angles, 96- 101 Bond length, coupling to torsional angles, 96- 101 Branched models, m uc h structure, 353-356 British gums, see Dextrins Bronchitis patient, mucus glycoproteins, 360- 36 I Browning, see Caramelization

C

Ca2+ chelation, proton shielding by, 129 exchange of Gdw and Mn2+for, 136 interaction with N-acetyl-a-Dneuraminic acid, 160

SUBJECT INDEX

Calcium alginate, 33 Caramel, see also Caramelan; Caramelen; Caramelin; Melanoidins acceptable daily intake, 233 as adhesive and binder, 232 ammonia carcinogenicity,268 standardization, 208 from ammonia processes, 228 -229 analysis, 235-237 aroma, 2 18 background, 204-205 bad, 229 Baker’s types, 23 I biological screening, 267 - 270 for brandy, 23 1 for brewing, 23 1 “(2-c.p./m.a.s. n.m.r., 244-245 chemical nature, 237-267 acid-catalyzed reversion, 252 aqueous sucrose solution pH, 239 carboxylic acid or Schiff base paths, 263 caustic caramel, 253 diketosamine decomposition, 264-266 formation of heterocyclic low-molecular-weight products, 266-267 intramolecular cis eliminations, 248 metal salt effects, 252-253 mutarotation, 249 oxaheterocycle formation, 247 - 248 from plain sugars, 237 reductones, 249-250 thermal stability, 238 volatile and nonvolatile fractions, 238-239 clastogenic effect, 268 colloidal particles in, 227 colors, analytical characteristics,206, 209-2 13 composition, 2 15 in presence of acid catalysts, 252 cosmetic use, 23 1 -232 definition, 205 degree of polymerization, 244 detection, 234-236 dry, 227-228 electronegative, 2 15 electropositive, 2 14 firmness, 222-223 flavor, 2 I8

413

gas-liquid chromatography, 233-236 hygroscopicity, 223 isoelectric points, 214,227 isoelectric properties, 23 1 killing heat, 228 manufacture, 228 manufacturing sources, 2 18 - 225 inhibitor use, 224 light and browning, 225 melanoidin formation, 223 -224 from molasses, 22 1 nonconventional sources, 222 in presence of ammonia, 224-225 y-radiation effect, 225 reducing properties, 2 18,220 from saccharides, 223 from sucrose and its products, 2 19 - 22 1 water loss, 218-219 mutagenicity, 267 - 269 pH, 216-217 pharmaceutical industry, 23 1 physical properties, 214-218 from Pictet - Andrianoff vacuum process, 216-217 plastic resins colored with, 232 presence of proteins, 230 quaternized by metal ions, 232 range of shades of color, 232 size-exclusion chromatography, 235 for soft drinks, 231 solubility, 2 17 standardization, 205 -206,208 storage, 229 sugar colors, 2 14 thin-layer chromatography, 236 tinctorial strength, 2 18,231 types, 206-208 applications, 233 tests for, 206, 208 ultrafiltration, 229 undesirable properties, 229 in urine, 269 uses, 230-233 viscosity control, 227 world production, 233 Caramelan analytical characteristics, 24 I, 243 infrared absorption spectra, 24 1 -242 from intermolecular polymerizations, 248-249

414

SUBJECT INDEX

physical properties, 216 ultraviolet absorption spectra, 241,243 Caramelen analytical characteristics, 241,243 chemical nature, 241,244 infrared absorption spectra, 241 - 242 physical properties, 216 ultraviolet absorption spectra, 24I, 243-244 Caramelin chemical nature, 244 infrared absorption spectra, 241 -242 physical properties, 216 Caramelization acidic medium or acid catalysts, 250-252 background, 204-205 by-products, 267 concepts, 226-227 condensation, 245 degradation reactions, 245 degree of, 218 during manufacture of table sugar, 221- 222 kinetics, 226 polymerization, 245 in presence of ammonia, amino acids, peptides, and proteins, 255 products of, 238-240 solution versus solid state, 239-240 sucrose, 219- 220 under alkaline conditions, 253-255 Carbamoyl group, reversed anomeric effect, 65 Carbohydrates, see also Metal ions anomeric carbon atom, 102 charge requirements, 128 functional groups, 127 reactivity, see Anomeric effect size requirements, 127 Carbon atom anomeric, in carbohydrates, 102 electron-nuclear relaxation methods, 136-137 Cation-exchangecolumns methyl aldofuranosides, retention volumes, 20 metal cation-carbohydrate complexes applications, 36-40 C-F bond, anomeric, 73 ' F - H coupling constants, vicinal, 70 Chitin, solvolysis, 171 - 172

Chloromethoxymethane backdonation of lone-pair orbital, 109- 110 dipole-dipole interactions, 82,84 torsional potential, 82-83 2-Chlorooxane u form preference in solvents, 92 conformational equilibria, 85 CH,OCH,X compounds, geometry relaxation, 98 Coatings, dextrins as, 325 C-0 bond anomeric, shortening, 71 exo-anomeric effect, 70 C-0-C bond, exeanomericeffect, 73-74 Condensation, caramel, 245 Conformational analysis, 45-47 Conformational energy anomeric effect, 77- 88 acyclic model compounds, 77-8 1 cyclic model compounds, 83-88 dimethoxymethane, 78-79 methanediol, 78 - 79 methoxymethanol, 78- 79 ROCHZX, 79-80 RSCHZX, 80 - 8 1 saccharides, 88 torsional potentials, 82-85 map, dimethoxymethane, 76-78 Coordination spheres, metal ions, 128 Copper ions, complex-formation, 21 Cosmetics, dextrins in, 329 C d + , effect on cis-inositol T-n.m.r., 142- 143 Cyclic model compounds, conformational energy, 83-88 Cyclitols induced shifts of proton signals, 7 lanthanide-induced shifts, 9- 10 Cyclohexane-1,2,3,4,5/0-pentol,8,41 Cyclohexanes, substituted, anomeric equilibrium, 54

D

Deglycosylation,glycoproteins, 200 Delocalization interactions, see also Lonepair orbitals anomeric effect, 109- I 14

SUBJECT INDEX scaling with electrostatic interactions, 114-115 perturbation calculation, 11 1 stabilization energy, 1 13 2-Deoxy-Durubino-hexopyranosyl nitrate, substituted, equilibrium composition, 63-64 2-Deoxy-2-methyl-epi-inositol, 8, 15 Depolymerization, starch, 280 Dextran, 19 Dextrinization acid factor, 3 I8 catalysts, 318-319 chemistry, 302 - 3 17 acid catalyst role, 3 I 3 - 3 14 acid hydrolysis, 309 3 10 action of water on starch, 303 aldehyde yield, 3 16 base catalyst role, 3 15 branching and cross-linking, 306 - 308 ITor IH-n.m.r., 313 free-radical mechanism, 304-305 hydrolytic scission, 304 hydroxyl groups, 303 phenol yield, 316 in presence of a-amino acids, 3 15 in presence of oxidants, 3 16- 3 17 starch containing fatty acids, 303 structural changes, 312-313 tamarind-kernel polysaccharide, 310-312 thermolysis with glycine, 3 15- 3 16 transglycosylation, 308 conditions of, 3 18 degree of, 286-287 potato, maize, and rice starch, 289, 291 -292 products, specific rotation, 30 1 purpose, 286 rate, 286, 288 salts as catalysts, 3 19, 32 I Dextrins alkali-lability, 293, 295 amylolytic index, 320- 32 1 as antiulcer agent, 333-334 biological activity, 333-335 on humans, 334 on lower organisms, 334-335 blue staining residue, 283 classification, 282-284 critical micelle concentration, 30 1 - 302

-

415

"generally-recognized-as-safe", 329 historical background, 281 -282 hydrolysis, 30 1 infrared spectra, 312-313 iodine affinity, 283-284,297-298 manufacture, 32 1 -323 maturation, 32 1 - 322 methylation analysis, 308 309 mutagenicity, 3 I6 organoleptic control, 298 in presence of catalyzing acid, 29 1,293 properties, 284 302 atmosphere effects, 305 catalyzing acid effects, 293 -294 modification, 301 reaction time and temperature of heating, 320- 32 1 reducibility changes, 294, 296 regulations, 333 retrogradative ability, 306 roasting, 322 solubility, 288-289 moisture content and, 289-290 temperature relationship, 290-291 in solution, 285 solvation, 328 sources for manufacture, 3 17- 32 I UXS, 323- 333 adhesives, 323 as binders, 324-325 coatings, 325 as complexing agent, 327 - 328 in cosmetics, 329 as depressant in flotation system, 328 - 329 electroplating brighteners, 326 as encapsulating agents, 332 in food, 329-332 free radical propagation, 327 in fuel compositions, 326 - 327 matrices, 324-325 in protective colloids, 325 replacement by synthetic polymers, 324 thickness, 323-324 viscosity changes, 294,297 of solutions, 283 yellow, composition and characteristics, 314 yield, form, and color, 285

-

-

416

SUBJECT INDEX

5,7-Diacetamido-3,5,7,9-tetradeoxy-D-gly- 1,2-Enediol, sugar transformations in cerd-galacto-nonulosnic acid, 198 Diamagnetic species, 129- 130

5,7-Diamino-3,5,7,9-tetradeoxynonulosonic acids, 180-181

1,4:3,6-Dianhydro-&Dglucopyranose, 306 - 308 1,4:3,6-Dianhydro-Dmannose, 306 3,6-Dideoxy-3-(L-glycroylamino)-D-galacto=, 175-176 4,6-Dideoxy-4-[(S)-2,4-dihydroxybutanamido]-D-mannose, 175- 176 4,6-Dideoxy-4-[(R)-3-hydroxybutanamido]D-galactose, 175- 176 Differential thermal analysis, starch, 299 - 300 Di-&Dfructopyranose 1,T: 1’24anhydride, 17,26 Diketosamine, decomposition, 264- 266 Dimethoxymethane bond lengths and angles, 97-98 conformational energy map, 76-78 dipole-dipole interactions, 82, 84 highest-occupied MO orbitals, 120- 12I molecular orbitals, I 1 1 relative energies, 78 - 79 torsional potentials, 82 - 83 Dimethyl ethers, substituted bond length and angles, 96,98 net charge, 102- 103 rotational potential constants, 82-83, 85 1,8-Dioxadecalin, photoelectron spectra, 120 Dipolar interaction, correlation times, 133 Dipole-dipole interaction, energies, 107 Dia-L-sorbopyranose 1,2’: 1’,2dianhydride, 17

alkaline medium, 253 erythro-erythro configuration, 14 Escherichia coli 078,O-antigen polysaccharide, 186 Ethyl methyl ether, ap orientation, 95 Ed+, interaction with N-acetyla-Dneuraminic acid, 160 Eubacterium saburreum, polysaccharides, 179 Exo-anomeric effect, 58 C-0 bond, 70 definition, 48 MO calculations, 94- 96 potential-function calculations, 103- 106 Exo-anomeric equilibria, 69 -7 I

F Farina starch, dextrinization, 318 Fermi contact shifts, 131 Femcyanide number, starch, 294,296 2-Fluorooxane a form preference in solvents, 92 conformational equilibria, 85 Food, dextrins in, 329-332 Fourier component analysis, torsional potentials, 82-85 Free radicals, propagation, dextrins, 327 D-Fmctose decomposition, 220-22 1 separation, 39-40 Fuels, dextrins in, 326-327 2-Furaldehyde, formation, 247 - 248

G

E Egg-box model, 33 Electron distribution, lone-pair orbitals and, 101-103 Electron-nuclear relaxation methods, carbon atoms, 136- 137 Electron-spin relaxation, 133- 135 epi-inositol, I38 - 140 Electrophoresis, metal cation - carbohydrate complexes applications, 35 - 36 Electroplating brighteners, dextrins as, 326 Electrostatic interactions. anomeric effect. 107- 109

p-D-Gal, 157 a-D-Galp group, 157 3-Oa-D-GalpN&-LSer, Gdw and Mn2+ effect, 151-154 3-Oa-D-Gal-L-Ser, Gdw and Mn2+effect, 151-154 y-irradiation, saccharides, 276 (+)-gauche-(-)-gauche arrangement, 14 Gd3+ N-acetyl-cY-D-neuraminicacid binding to, computer modeling, 160- 161 effect on 13C-n.m.r., 159- 160, 164- 165

SUBJECT INDEX effect on D-gluconamides IC-n.m.r., 146- 147, 149 electron-nuclear relaxation methods, 136-137 exchange for Ca2+and Mg2+in biological systems, 134 ept-inositol electron-spin relaxation effect, 138- 139 interactions with complex glycopeptides, 155- 158 with glycosylated amino acids, 151- 154 with inositol, see Inositols medical uses, 135- 136 Gel electrophoresis, mucus glycoproteins, 359 Gellan gum, 192- 193, 199 Gel-permeation chromatography, mucus glycoproteins, 358 Gibbs energy, anomeric effect, 53- 59 D-Gluconamides, 144- 150 13-C-chemical-shiftdata, 145- 146 I3C-n.m.r. Gd" effect, 146- 147, 149 Mn2+effect, 146, 148- 149 structures, 144 D-Glucopyranose a form preference in solvents, 92 anomeric equilibrium, solvent effects, 68 anomers, 54-55 derivatives, equilibrium compositions, 48-49 D-Glucosamine, 19 D-Glucose acid-catalyzed reversion, 252 composition of evolved gases, 272,274 course of total gas evolution, 272-273 decomposition, 220-221,272 pyrolysis, 272 course of, 273,276 liquid products, 272,275 volatile products, 274,276 reducing properties, 2 18, 220 separation, 39-40 water loss, 218-219 a-D-Glucose, pyrolysis, 304- 305 D-Glucosyl oxocarbonium ions, 304 Glycopeptides, complex, interactions with Gd)+ and Mn2+, 155- 158 Glycoproteins, deglycosylation, 200 a- and /?-D-Glycopyranose, anomers, 54-55

417

Glycosidic linkages cleavage, hydrogen fluoride, 168 lability, 178, 180, 189 sugar residues, 194- 195 Glycosylation by D-Gal groups, I57 hydroxyl groups, 170 Glycosyl fluorides, stability, 172 Griessmayer-Aubry method, 234 Ground state, anomeric effect, 118 a-L-Guluronic acid, 24 - 25 complexing, 32

H Hard sphere exo-anomeric, 105 D-glycereLtalo-Heptose, 39 Herdan relations, 375 Hexitols, separation, 37 Hexosaminoglycans, bacterial, monosaccharide composition, 174- 175 High-performance liquid chromatography, metal cation-carbohydrate complexes applications, 40 HSEA approach, 105 Hydrogen bonds, between hydrogen fluoride and hydroxyl groups of sugars, 169 Hydrogen fluoride glycoprotein deglycosylation, 200-201 mechanism of reaction with carbohydrates, 168 173 presence in reaction mixture of water, 172 solvolysis, 173- 174 in methanol, 195 sugar polymerization, 202 removing from cell walls, 20 1 Hydrogen starch, moisture and degradation, 290 Hydrolysis acid-catalyzed, 172 dextrins, 301 Hydroxyl groups anomeric effects, solvent effects, 68-69 in dextrins, 303 glycosylation, 170 5-(Hydroxymethyl)-2-furaldehyde, formation, 249 manufacture, 230 reaction with reversion products, 252

-

418

SUBJECT INDEX

2-C-(Hydroxymethyl)-myo-inositol, 17 C-(Hydroxymethyl)-scybinositol,17

I

L Lactose reducing properties, 2 18,220 stability constants, 30 water loss, 218-219 Lanthanide ion complex-formation, 20 coordination spheres, 128 oxidation state, 129 Light-scattering procedures, mucus glycoproteins, 356-358 Linear models, mucin structure, 367-370 Lone-pair orbitals (up,up) conformation, 120 electron distribution and, 101- 103 orbital interactions, 110 interactions in reactivity, 119- 123 oxygen distribution, 109 interactions, I 1 1 - 1 12 n - t w , 109 through-space mechanism, 122 Low-speed sedimentation equilibrium, mucus glycoproteins, 359- 367

cis-Inositol, 4, 8, 138 complex-formation, 15 T-n.m.r. Cu2+effect, 142- 143 Mn2+effect, 141- 142 plots of i3C-(Tt)-1 vs. metal-ion concentration, 142- 144 stability constants, 29 trihydroxyl pocket, 142 epi-Inositol, 7, 138 diamagnetic shifts, 129 electron-spin relaxation, 138 - 140 T;/TC, values, 140- 141 scyllo-Inositol monoorthoformate, 15 Inositols interactions with Gd* and MnZ+,137- 144 metal-ion size-requirement for binding, I38 epi-Inositol. SrCl, * 5 H,O, 24 Intramolecular ciselimination, 248 Iodine affinity, dextrins, 283-284,297-298 M Iron, dextrin as complexing agent, 327-328 Maillard reaction, 255 -256 Isomaltose, decomposition, 272 products, 269-270 Isomers Maize relative abundance, 59 -7 I dextrins from, 3 17- 3 18 aglycon character, 62-65 dextrinization and properties, 320-32 1 exo-anomeric equilibrium, 69-7 1 Maize starch, dextrinization, 289 /?-D-ribopyranose derivatives, 62 Maltose ring substituents, 66-67 decomposition, 272 solvent effect, 67-69 pyrolysis, course of, 273, 276 PDxylopyranose derivatives, 6 1 PD-Mannofuranose CaCI, -4H,O, 22 -23 PDxylopyranosyl halides, 62 -63 D-Mannopyranose, axial hydroxyl group on 2,3-O-Isopropylidene4O-methyl-~Lrham- C-2,66 nopyranose, 30 Melanoidins Isosaccharosan, 244 I3C- and i5N-cp./m.a.s. n.m.r. spectra, 257 carbon types, 257-258 formation, 223-224,257,259-262 mutagenicity, 269- 270 K nondialyzable, 256-257 Ketohexose, reactions in anhydrous oxidation, 256-257 precursors, 259 hydrogen fluoride, 170 Ketopyranoses, 17- 18 structure, 256

.

SUEUECT INDEX

Metal cation-carbohydrate complexes, 1 - 3, see also Metal - sugar complexes alduronic acids, 3 1 - 35 applications cation-exchange resin columns, 36 -40 electrophoresis, 35 - 36 high-performance liquid chromatography, 40 n.m.r. spectroscopy, 40-41 synthetic applications, 41 -42 thin-layer chromatography, 35- 36 biological implications, 42-43 cations, 19-21 complexing sites, 12- 19 a, e, a sequence, 12-13 alditols, 13- 14 syn-axial hydroxyl groups, 15 - 16 dextran, 20 effectiveness of, 19 cis-inositol, 15 ketopyranoses, 17- 18 methoxyl group replacing hydroxyl group, 14 crystalline complex structures, 22-26 a-D-allopyranosyl a-Dallopyranoside, 25 a-L-guluronic acid, 24-25 epi-inositol SKI, * 5H20, 24 BD-mannofuranose * CaCI, .4H,O, 22-23 methyl D-glycero-cu-D-guleheptopyranoside CaCI, .H,O, 23 methyl BD-mannofuranoside*CaCI2.3H,O, 23 in solution, 22 electrophoretic mobilities, in cupric sulfate solution, 2 I stability constants, 26- 30 cation-sugar complexes, 28 1:1 and 1:2 complexes, 27 in methanol, 30 methods of determination, 27 reducing sugars, 29 Metal ions, see also specific ions binding site, 127 coordination spheres, 128 diamagnetic species, 129- 130 interactions with carbohydrates, 125- 126 a, e, a sequence, 127

419

m,p arrangement, 127 selective-broadening technique, 134 oxidation states, I29 paramagnetic species, 130- 135 quaternization of caramel, 232 Metal salts, reactions of 5-(hydroxymethyl)2-furaldehyde and, 252-253 Metal - sugar complexes electrophoretic mobilities, 4- 5 n.m.r. spectroscopy, 5 - 11 syn-axial hydroxyl groups, 8 complexes with diamagnetic cations, 6-9 complexes with paramagnetic cations, 9-11 contact and pseudocontact shifts, 9- 10 polyol~,5 - 6 screening constant change, 6 shifts, 5 optical rotation, 12 paper electrophoresis, 3 - 5 stability constants, 28 thin-layer ligand-exchange chromatography, 5 Methanediol, 78 - 79 Methoxyl groups, anomeric effects, solvent effects, 68 - 69 Methoxymethanol, relative energies, 78- 79 2-Methoxyoxane acetal segment, rabbit-ear effect interactions, 107- 108 axial, exo-anomeric effect, 95 bond angle variation with torsional angle, 100

conformers, solvent effect, 90-92 net atomic charges and dipole moments, 103 torsional potential, 84,86-87 Methyl aldofuranosides, cation-exchange columns, retention volumes, 20 Methyl D-allosides, synthesis, 42 Methylation analysis, 308 - 3 10 Methyl furanosides, synthesis, 42 Methyl a-and /3-D-galactopyranoside, Gdw and Mn2+effect, 150- 152 Methyl a-and j?-Dglucopyranoside aglycon orientations, 50- 5 1 conformational energy, 88 Methyl a-glycoside, 198- 199

420

SUBJECT INDEX

Methyl a-D-gulopyranoside, 10 Methyl D-glycero-cr-D-gubheptopyranoside-CaCI,-H,O, 23 4(5>Methylimidazole, 267 Methyl a-D-lyxofuranoside, I6 Methyl j?-D-mannofuranoside.CaCI,* 3 H20,23 Methyl ED-psicofuranoside, 16 - 17 Methyl a-and E D pyranosides, mean, hemiacetal and acetal geometries, 72 Mg2+,exchange of Gd3+and Mn2+for, I36 Micelles, critical concentration, dextrins, 301 -302 Mn” effect on IF-n.m.r. N-acetyl-ol-D-neuraminicacid, 162-163, 165 D-gluconamides, 146, 148- 149 cis-inositol, 141- 142 electron-nuclear relaxation methods, 136-140 exchange for Ca2+and Mg2+in biological systems, 134 epi-inositol electron-spin relaxation effect, 138-140 interactions with complex glycopeptides, 155 - 158 with inositol, see Inositols with methyl a-and /3-D-galactopyranoside, 150- 152 medical uses, 135- 136 Molasses, caramel from, 22 1 Molecular mechanics, 104- 106 Molecular orbital calculations anomeric effect, see Anomeric effect solvent effect a form preference, 92 2-methoxyoxane conformers, 90-92 solvophobic theory, 89 supermolecule quantum chemical calculations, 89 Monosaccharides composition of bacterial hexosoaminoglycans, 174-175 glycosidic linkages, 167- I68 preparation, for analysis, 174- I80 Mucins, see Mucus glycoproteins Mucus glycoproteins, 345 -347 amino acid composition, 349 basic unit, 349- 352

coiled structure, 350-352 containing proline residue, 350-35 1 prolyl residue, 350 carbohydrate sidechain composition variability, 35 I - 352 components, terminology, 353 composition, 347 - 349 electron microscopy, 368- 369 gross conformation in solution, 370-373 heterogeneity, 374- 380 Herdan relations, 375 InJ versus (, 378-379 interactions with other macromolecules, 380 molecular weight distribution, 378-379 polydispersity and self-association behavior, 377-380 sedimentation coefficientdistribution, 376-377 terminology, 374 thermodynamic non-ideality, 374- 375 importance, 345 - 346 molecular weights, 349,356-367 distribution, 378- 379 light-scattering, 356-358 low-speed sedimentation equilibrium, 359 - 367 meniscus-depletion technique, 365 - 366 partial specific volume, 36 1 Rayleigh interference optics, 366 relative techniques, 358-359 scanning absorption optics, 365 - 366 thermodynamic non-ideality e f f i , 365 virial coefficient, 366 mucin component, 348 as polyelectrolytes, 373 properties, 346 Rayleigh equilibrium interference patterns, 377-378 secondary structure, 350-35 1 tertiary structure branched models, 353-356 linear models, 367 - 370 molecular weights, 356 - 367 star average, 361 subunits, 352-353 T-domains, 353 thiol reduction, 352-353 Muellitol, 6, I5 Mutagenicity

SUBJECT INDEX caramel, 267-269 melanoidins, 269-270 Mutarotation, caramel, 249

N Neisseria meningitidis, Odeacetylated capsular polysaccharide, 185 n.m.r. spectroscopy metal cation-carbohydrate complexes applications, 40 - 4 I metal - sugar complexes, see Metal-sugar complexes 0

Octasaccharides, 190, 192 D-erythro-Ltalo-Octose, 39 Oligomerization, secondary, polysaccharides, 170 Oligosaccharides caramel production, 222 cation complexes, 18- 19 containing Lrhamnose, 193 mean, hemiacetal and acetal geometries,72 methyl glycosides, I99 preparation of, 180- 200 Oxaheterocycles, formation, 247 - 248 Oxane alkoxy and alkylthio derivatives, exo-anomeric effect, 69 -70 derivative, axial - equatorial equilibrium, 59-60 substituted, anomeric equilibrium, 54 2-substituted anomeric equilibrium, 67-68 axial preferences and Gibbs energy, 55-56 bond lengths and angles, 98-99 dipolar interactions, 107 non-equivalent lone-pairs, 1 I 1 solvent effectson axial preference, 67 - 68 Oxane ring, conformational properties, 53 2-(Oxan-2-yloxy)oxane,conformers, 93 Oxidation states, metal ions, 129 Oxygen lone-pair electrons, I01 - 102 lone-pair orbitals, backdonation

42 1

chloromethoxymethane, 109- 110 distribution, 109 interactions, 1 1 1 - 1 12 nucleophilicity, enhancement, 120, 122

P Paper electrophoresis,metal-sugar complexes, 3 -5 Paramagnetic species, 130- 135 dipolar interaction, 133 electron-spin relaxation, 133- 135 Fenni contact shifts, 131 hyperfme shifts, 130 pseudo-contact shifts, 130- 131 relaxation agents, 130- 135 scalar coupling, 134 shift agents, 130- 131 Pectate, 32 Perturbation theory, chemical reactivity, 122 pH, caramel, 216-217 n-antibonding orbitals, CH, group, 1 13 Pictet -Andrianoff vacuum process, 216-217 Polydispersity, self-association behavior and, mucus glycoproteins, 377 - 380 Polyelectrolytes,mucus glycoproteinsas, 373 Polyhydroxy compounds, cation complexes, 11 Polymerization caramel, 245 s u p , hydrogen fluoride, 202 Polyols electrophoretic and thin-layer mobilities, 35 - 36 electrophoretic behavior, 4 threo-threo configuration, 145 tridentate complexes, 10- 1 1 Polysaccharides, 167- 168 alduronic acids in, 177 alkaline thennochemical degradation, 3 15 caramel production, 222 cation complexes, 18- 19 complexing, on surface, 43 containing uronic acids, solvolysis, 181 secondary oligomerization, 170 Potato starch dextrinization, 289,29 1 -292 dextrins from, 3 17- 3 18 solubility, 29 I, 293

SUBJECT INDEX

422

Potential-function calculations, anomeric and exo-anomeric effects, 103- 106 Prolyl residue, 350 Proteum mirabilis, 0-specific polysaccharides, 181-182 Proton shielding, by chelation of Caz+, 129 Pseudocontact shifts, 130- 131 Pseudomonas aeruginosa extracellular polysaccharide, 189 methylated oligosaccharide fragment, 179- 180 @specific polysaccharides, 175, 177, 182- 183, 196, 198 Pseudomonas aurantiaca 3 1, Qspecific polysaccharide, 188 Pseudomonasfluorescens 36 I, 0-specific polysaccharide, 186- 187 Pyranoid saccharides,energy relations, 117 Pyranoses shift-changes, 7 - 8 mean, hemiacetal and acetal geometries, 72 Pyranosides, a,e,a sequence, 41 Pyrodextrins manufacture, 3 18 methylation analysis, 308- 3 10 Pyrolysis CY-D-~~UCOS~, 304- 305 starch, 335-344 s u m , 270-278 Pyrolyzate, composition, 272-273

R Rabbit-ear effect, aced segment of 2-methoxyoxane, 107- 108 Rayleigh interference patterns, mucus glycoproteins, 359-360 Reaction paths, energy, 116- 118 Reductones, 249-250 Relaxation reagents, 131 - 135 nonspecific, 131- 132 specific, 132 Reverse anomeric effect, 59 definition, 52-53 a-L-Rhamnosyl linkage, 192- 193 Rhizobium japonicum 311b 83, extracellular polysaccharide, 189- 190 Rhizobium leguminosarum, extracellular polysaccharides, 190- 192

Rhizobium triforii,extracellular polysaccharides, 190, 192 &D-Ribopyranose, 13 derivatives, conformational equilibria, 62 D-Ribose, stability constants, 29 Rice starch, dextrinization, 289 Rotational platinum-shadowing, mucus glycoproteins, 367

S Saccharides caramel from, in presence of additives, 223 conformational energy, 88 course of decomposition, 336-337 force-field, 105 y-irradiation, 276 reactivity hydrolysis, I 19- 120 stereoelectroniceffect, 1 19 solid-state structures, 70 thermal treatment above melting points, 226 treatment with mineral acids or alkali, 226 Scalar coupling, 134 Self-associationbehavior, polydispersity and, mucus glycoproteins, 377 - 380 Shift reagents, I 30 - 13 1 relaxation times, 130 Shigella sonnei, 0-specific polysaccharide, 176 Solubility, caramel, 217 Solvation energy, 89-90, 115 Solvolysis chitin, 171 - 172 hydrogen fluoride in methanol, 195 techniques, 173- 174 Solvophobictheory, 89 Specific rotation, products of dextrinization, 30 1 Starch action of water on, 303 alkali-lability, 293, 295 beta-amylolysis value, 298 -299 containing fatty acids, 303- 304 decomposition, 279-28 1 degree of dextrinization, 286-287 depolymerization, 280 differential thermal analysis, 299- 300

SUBJECT INDEX femcyanide number, 294,296 y-radiation effect, 225 pyrolysis, 335 - 344 course of decomposition, 336-331 effect of inorganic salts, 340 fragmentation, 338-339 gaseous and liquid fractions, 336,339 kinetics, 340 origin of starch and, 336 pressure and temperature, 336 products, 336,338 volatile products, 336,339 rate of dextrinization, 286, 288 transformation, see Dextrins Streptococcus pneumoniae, capsular polysaccharide, 116- 171 Sucrose aqueous syrups, pH, 239 caramelization, 219-220 crystalline adducts, 2 decomposition, 239 reducing properties, 2 18,220 thermal degradation, 220,245- 246 water loss, 218-219 sugars colors, 2 14 polymerization, hydrogen fluoride, 202 pyrolysis, 270-218 composition of evolved gases, 212, 214 composition of gases, 21 1-212 furan derivatives, 216,218 total gas evolution, 212-273 water formation, 274 reactions with amino acids, 223 removal from cell walls, 201 residues, glycosidic linkage lability,

194-195 reversion, 245 separations on ion-exchange columns, 38 stock, before caramelization, 240- 24I transformation from cyclic to acyclic, 246 transformations in alkaline medium, 253 Svedberg equation, 351

423

Tdomains, mucus glycoproteins, 353

3,4,5,1-Tetra-O-acetyl-2,6-anhydro-~-glycero-lgluco-heptonamide, anomeric equilibrium, 65 Thermal decomposition, 203-204 Thin-layer chromatography, metal cationcarbohydrate complexes applications,

35-36 Thin-layer ligand-exchange chromatography, metal - sugar complexes, 5 Thiol reduction, mucus glycoproteins tertiary structure, 352-353 threo-threo configuration, 14, 145 Torsional angles, coupling of bond lengths and bond angles to, 96- 101 Torsional potentials, Fourier component analysis, 82-85 Transglycosylation, 308 Transition state, anomeric effect, 1 18 cqa-Trehalose conformation, 52 exo-anomeric effect, 10 sc conformation, 88

2,3,4-Tri-O-acetylpentopyranosylamines, 51 Tri-0-acetyl-&D-xylopyranosylchloride, 63 1,3,5-Tri(3-methylbut-2-enyl)-scylbinositol, 6, 15

U Uronic acids, polysaccharides containing, solvolysis, 181

V

Vibrio cholerae, 0-specific polysaccharides, 119

W Water, action on starch, 303

T Tamarind-kernel polysaccharide, dextrinization, 310-312 Tapioca, dextrins source, 3I7- 3I8

X DXylan, 18 Xylitol, stability constants, 29

424

SUBJECT INDEX

D-Xylopyranose amino-substituted,65 derivatives having N-substituents, equilibria, 64-65 Dxylopyranose tetraacetate, anomeric equilibria, 66 DXylopyranosyl fluorides, valence geometry parameters, 73 PDXylopyranosyl halides, isomer equilibrium, 62-63 DXylose, 18

Y Yellow-farina dextrin, alkali-lability, 293, 295 Yersinia enterocolitica, 0-specific polysaccharides, 179

z Zimm-plot technique, mucus glycoproteins, 351-358