COFFEE Recent Developments Edited by R.J. Clarke and O.G. Vitzthum
b
Blackwell Science
# 2001 by Blackwell Science Ltd Editorial Offices: Osney Mead, Oxford OX2 0EL 25 John Street, London WC1N 2BS 23 Ainslie Place, Edinburgh EH3 6AJ 350 Main Street, Malden MA 02148 5018, USA 54 University Street, Carlton Victoria 3053, Australia 10, rue Casimir Delavigne 75006 Paris, France Other Editorial Offices: Blackwell Wissenschafts-Verlag GmbH KurfuÈrstendamm 57 10707 Berlin, Germany Blackwell Science KK MG Kodenmacho Building 7±10 Kodenmacho Nihombashi Chuo-ku, Tokyo 104, Japan Iowa State University Press A Blackwell Science Company 2121 S. State Avenue Ames, Iowa 50014-8300, USA The right of the Author to be identified as the Author of this Work has been asserted in accordance with the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of the publisher. First published 2001 Set in 9.5/11 Ehrhardt by DP Photosetting, Aylesbury, Bucks Printed and bound in Great Britain by MPG Books Ltd, Bodmin, Cornwall The Blackwell Science logo is a trade mark of Blackwell Science Ltd, registered at the United Kingdom Trade Marks Registry
DISTRIBUTORS
Marston Book Services Ltd PO Box 269 Abingdon Oxon OX14 4YN (Orders: Tel: 01235 465500 Fax: 01235 465555) USA and Canada Iowa State University Press A Blackwell Science Company 2121 S. State Avenue Ames, Iowa 50014-8300 (Orders: Tel: 800-862-6657 Fax: 515-292-3348 Web www.isupress.com email:
[email protected] Australia Blackwell Science Pty Ltd 54 University Street Carlton, Victoria 3053 (Orders: Tel: 03 9347 0300 Fax: 03 9347 5001) A catalogue record for this title is available from the British Library ISBN 0-632-05553-7 Library of Congress Cataloging-in-Publication Data Coffee: recent developments/edited by R.J. Clarke and O.G. Vitzthum p. cm. Includes bibliographical references and index. ISBN 0-632-05553-7 1. Coffee. I. Clarke, R.J. (Ronald James) II. Vitzthum, O.G. TP645 .C65 2001 663'.93Ðdc21 00-049815 For further information on Blackwell Science, visit our website: www.blackwell-science.com
Contents Preface vii List of Contributors ix 1 Chemistry I: Non-volatile Compounds 1 1A Carbohydrates 1 A.G.W. Bradbury 1.1 Introduction 1 1.2 Green coffee 1 1.2.1 Low molecular weight carbohydrate 1 1.2.2 High molecular weight carbohydrate 3 1.3 Roast coffee 6 1.3.1 Low molecular weight carbohydrate 6 1.3.2 High molecular weight carbohydrate 7 1.4 Soluble coffee 8 1.4.1 Low molecular weight carbohydrate 8 1.4.2 High molecular weight carbohydrate 12 1.5 Reactions of carbohydrates on roasting 13 1.6 Functional properties of coffee carbohydrates 14 1.6.1 Role in soluble coffee processing 14 1.6.2 Foam 15 1.6.3 Coffee fiber 15 References 15 1B Acids in Coffee 18 H.H. Balzer 1.7 Quantitative data on organic acids in green coffee 18 1.8 Determination of organic acids in roasted coffee 19 1.9 Acid formation mechanisms 23 1.9.1 Acetic, formic, lactic, glycolic and other carbohydrate derived acids 23 1.9.2 Quinic acid 23 1.9.3 Citric and malic acid 25 1.9.4 Phosphoric acid 25 1.10 Acid increase on storage 26 1.11 Volatile acids 26 1.12 Acid content and sensory characteristics 27
1.12.1 Total acidity and sour taste 1.12.2 Acid content and acidity 1.12.3 Roast kinetics References 1C Lipids K. Speer and I. KoÈlling-Speer 1.13 Introduction 1.14 Coffee oil 1.14.1 Determination of total oil content 1.14.2 Isolation of coffee oil for detailed analysis 1.15 Fatty acids 1.15.1 Total fatty acids and fatty acids in triglycerides 1.15.2 Free fatty acids 1.16 Diterpenes in the lipid fraction of robusta and arabica coffees 1.16.1 Free diterpenes 1.16.2 Diterpene fatty acid esters 1.16.3 Diterpenes in the lipid fraction of roasted coffees 1.16.4 Diterpenes in coffee: health aspects 1.17 Sterols 1.18 Tocopherols 1.19 Other compounds 1.20 Coffee wax References 2 Chemistry II: Non-volatile Compounds, Part II S. Homma 2.1 Amino acids and Protein 2.1.1 Amino acids 2.1.2 Amino acid derivatives 2.1.3 Protein 2.2 Fate of chlorogenic acid derivatives during roasting 2.2.1 Quinic acid moiety 2.2.2 Cinnamic acid derivative moiety 2.3 Antioxidative compounds in coffee brew 2.3.1 Compounds occurring naturally in green beans
iii
27 28 29 30 33 33 33 33 34 34 34 35 36 38 38 39 41 41 42 44 45 46 50 50 50 51 51 54 54 56 57 57
iv
Contents
2.3.2
Effect of roasting on antioxidative activity 2.4 Colored macromolecular compounds 2.4.1 Characterization of colored polymers 2.4.2 Characterization of the zincchelating compounds in coffee brews References 3 Chemistry III: Volatile Compounds W. Grosch 3.1 Introduction 3.2 Methodology 3.2.1 Isolation of the volatile fraction 3.2.2 Screening for potent odorants 3.2.3 Enrichment and identification 3.2.4 Quantification 3.2.5 Aroma models and omission experiments 3.3 Raw coffee 3.3.1 First studies 3.3.2 Potent odorants 3.3.3 Content and OAVs of odorants 3.3.4 Contaminants causing off-flavour 3.4 Roasted coffee 3.4.1 Concentration of important odorants 3.4.2 Evaluation of key odorants 3.4.3 Arabica versus robusta coffee 3.4.4 Influence of degree of roast 3.4.5 Aroma changes during storage 3.5 Coffee brew 3.5.1 Extraction yield of potent odorants 3.6 Formation of odorants 3.6.1 Mono- and dicarbonyl compounds 3.6.2 Furanones 3.6.3 Alkylpyrazines 3.6.4 Phenols 3.6.5 Thiols 3.7 Conclusions References 4 Technology I: Roasting R. Eggers and A. Pietsch 4.1 Introduction 4.2 Roasting methods and their parameters 4.2.1 General
58 58 58 62 65 68 68 69 69 69 71 72 73 73 73 73 73 75 75 75 77 77 79 79 80 80 82 82 83 84 84 84 85 85 90 90 90 90
4.2.2 4.2.3 4.2.4 4.2.5
Conventional roasting Fluidized bed roasting Fast roasting Detection of optimum degree of roast 4.3 Bean behaviour during roasting 4.3.1 Bean temperature, mass and moisture 4.3.2 Swelling and structure 4.3.3 Decaffeinated coffee 4.4 Heat and mass transport 4.4.1 Complexity of the process 4.4.2 Specific heat of coffee 4.4.3 Thermal conductivity 4.4.4 Heat uptake of the bean 4.4.5 Temperature profiles in the bean 4.4.6 Heat transfer from gas to bean and overall heat transfer coefficient 4.5 Some aspects on future scientific research 4.6 Industrial roasting equipment 4.6.1 Traditional roasters 4.6.2 Fluidized beds 4.6.3 Packed bed roasting 4.6.4 Roasting with heated cooling gas 4.6.5 Technical data and capacities 4.6.6 Roaster patents 1986±99 References 5 Technology II: Decaffeination of coffee W. Heilmann 5.1 Introduction 5.2 Solvent decaffeination 5.3 Water decaffeination 5.4 Supercritical CO2 decaffeination 5.5 Liquid CO2 decaffeination 5.6 Decaffeination with fatty material 5.7 Special developments 5.7.1 In-home decaffeination 5.7.2 Coffee with adjusted caffeine content 5.8 Caffeine recovery from activated carbon 5.9 Economic aspects References 6 Technology III: Instant Coffee R.J. Clarke 6.1 Introduction 6.1.1 Instant coffee in the market place
91 91 92 92 93 93 94 95 96 96 96 97 97 98 99 100 101 101 101 104 104 104 104 107 108 108 109 110 113 118 118 119 119 119 119 122 123 125 125 125
Contents
v
6.1.2 6.1.3
New technology The legacy of Professor H.A.C. Thijssen 6.1.4 Legislation and standardization 6.2 Processing 6.2.1 General 6.2.2 Roasting/grinding 6.2.3 Extraction 6.2.4 Freeze concentration of extracts 6.2.5 Thermal concentration and volatile compound recovery 6.2.6 Volatile compound handling 6.2.7 Reverse osmosis 6.2.8 Spray drying and agglomeration 6.2.9 Freeze drying 6.2.10 Aromatisation 6.2.11 Spent grounds disposal 6.2.12 Grading, storage and blending of green coffees 6.2.13 Liquid extracts 6.3 Physical properties of volatile compounds 6.3.1 Important physical properties in relation to instant coffee processing 6.3.2 Tables of physical properties References 7 Technology IV: Beverage Preparation: Brewing Trends for the New Millennium M. Petracco 7.1 Introduction 7.2 Extraction methods 7.2.1 Decoction methods 7.2.2 Infusion methods 7.2.3 Pressure methods 7.3 Beverage characterization 7.3.1 Physical and chemical characteristics 7.3.2 Organoleptic characteristics 7.4 Modified coffee beverages 7.4.1 Coffee±milk admixtures 7.4.2 Canned coffee beverages 7.4.3 Flavoured coffee beverages References 8 Health Effects and Safety Considerations B. Schilter, C. Cavin, A. Tritscher and A. Constable 8.1 Introduction 8.2 Objectives and scope
125 125 126 126 126 127 127 128 129 130 130 130 131 132 132 132 132 133 133 137 137 140 140 141 143 144 145 151 151 157 160 160 161 161 162 165 165 165
8.3 8.4
Coffee Coffee 8.4.1 8.4.2 8.4.3 Coffee 8.5.1
consumption and cancer Human data Experimental data Conclusions 8.5 and cardiovascular disease Myocardial infarction or coronary death 8.5.2 Arrhythmias 8.5.3 Caffeine and blood pressure 8.5.4 Serum cholesterol 8.5.5 Serum homocysteine 8.5.6 Conclusions 8.6 Coffee and bone health 8.6.1 Calcium metabolism 8.6.2 Osteoporosis 8.6.3 Conclusions 8.7 Reproductive and developmental potentials of coffee and caffeine 8.7.1 Congenital malformations 8.7.2 Neurodevelopmental effects 8.7.3 Low birth weight, growth retardation and prematurity 8.7.4 Spontaneous abortion 8.7.5 Fertility 8.7.6 Conclusions 8.8 Emerging benefical health effects 8.8.1 Neuroactivity 8.8.2 Chemoprotection 8.9 Coffee consumption ± safety considerations 8.10 Conclusions References 9 Agronomy I: Coffee Breeding Practices H.A.M. Van der Vossen 9.1 Introduction 9.1.1 World production increase 9.1.2 Selection and breeding before 1985 9.1.3 New developments 9.2 Genetic resources 9.2.1 World collections 9.2.2 Species relationships 9.2.3 Conservation 9.3 Breeding 9.3.1 General objectives and strategies 9.3.2 Productivity 9.3.3 Quality 9.3.4 Resistance to coffee leaf rust 9.3.5 Resistance to coffee berry disease
166 166 166 167 168 169 169 170 170 171 172 172 173 173 173 174 174 174 175 175 176 177 177 177 177 178 178 179 179 184 184 184 184 185 186 186 186 188 189 189 189 191 192 193
vi
9.3.6 Resistance to other diseases 9.3.7 Resistance to nematodes 9.3.8 Resistance to insect pests 9.3.9 Drought tolerance 9.4 Propagation of new cultivars 9.4.1 Seeds 9.4.2 Clonal propagation Abbreviations References 10 Agronomy II: Developmental and Cell Biology M.R. Sondahl and T.W. Baumann 10.1 Overview 10.2 Organ development and the allocation of defense compounds 10.2.1 Introduction 10.2.2 The leaf 10.2.3 The fruit 10.3 Purine alkaloid formation in coffee cell cultures 10.3.1 Introduction 10.3.2 Callus culture 10.3.3 Suspension culture 10.4 New advances in cell and organ culture 10.4.1 Brief review of the literature 10.4.2 New advances 10.5 Coffee scale-up by micropropagation 10.5.1 Mass production of somatic embryos 10.5.2 Applications 10.6 Somaclonal variation and new breeding lines 10.6.1 Definitions and examples 10.6.2 Coffee somaclonal variation program 10.6.3 Commercialization of new varieties 10.7 Summary Abbreviations References 11 Agronomy III: Molecular Biology J.I. Stiles 11.1 Introduction
Contents
194 195 195 196 196 196 196 197 197
11.2 11.3 11.4 Appendix 1.1 1.2
202 202 202 202 203 205
1.3 1.4 1.5 Appendix 2.1
207 207 207 208 209 209 209 212 213 215 217 217 217
2.2 2.3 Appendix 3.1
218 219 220 220 224 224
Index
3.2
Coffee genes Transformation systems for coffee Prospects References 1 International Standards Organization (ISO) Glossary relating to coffee and its products Green coffee (guides and sampling procedures) Instant coffee (sampling procedures) Methods of test (chemical or physical) General comments 2 International Coffee Organization (ICO) The International Coffee Agreement 1994 2.1.1 Background 2.1.2 Priorities 2.1.3 Coffee development projects 2.1.4 Promotional activity 2.1.5 Involvement of the private sector 2.1.6 Statistics and information 2.1.7 Global research network on coffee 2.1.8 Economic studies and publications 2.1.9 Towards a new Agreement in 2001 Conclusions Statistical information 3 Units and Numerals Units 3.1.1 SI base units 3.1.2 Some derived SI units used in engineering 3.1.3 Some prefixes for SI units 3.1.4 Some conversions of SI and non-SI units References Numerals ± cardinal
225 230 231 233 235 235 235 235 235 236 238 238 238 238 238 238 239 239 239 239 239 239 240 242 242 242 242 243 243 245 245 246
Preface A considerable period of time has now passed since the publication of six volumes upon all the technical aspects of coffee (Chemistry ± Technology ± Physiology ± Agronomy ± Related Beverages ± Commercial and Technico-legal) by Clarke and Macrae (eds), 1985± 8; and indeed since other major comprehensive volumes on coffee by Clifford and Willson (eds), 1985; and by Sivetz and Desrosier in 1979. Sadly, Robert Macrae was to die at an early age in 1995, and his literary and scientific work is sorely missed. The present editors considered that the time is now well due for an update on the topics covered by these volumes. Since 1987, seven Colloquia, organised by the Association Scientifique Internationale du CafeÂ, on coffee, have been held, with some 500 papers published in their Proceedings. Accordingly, we have been fortunate enough to secure the contributions of more than 15 internationally respected coffee scientists and technologists around the world to provide our readers with generally new information in the various areas in which they are expert. We have arranged for three updating chapters on the non-volatile and volatile compounds present, including some new ones, but especially the important ones, now known more clearly, to determine flavour. There follows four updating chapters on coffee technology; one, specifically, on some new developments in roasting techniques; two others on instant coffee and decaffeination processing; whilst another reflects recent developments in home/catering beverage preparation, especially of the true Italian espresso type. The
physiological effects of coffee drinking continue to interest and concern both the scientific and general public, so that the numerous publications and investigations of the last decade are considered in a comprehensive chapter on coffee and health. Agronomic aspects continue also to show considerable progress in the areas of coffee plant breeding, by conventional and tissue culture techniques; and to show rapid developments in a detailed understanding of the genes present in the coffee plant, and their function, and so to the possibilities and actualities of genetic engineering for desired characteristics, for example caffeine-free. Finally, some other topics relevant to coffee, that is, recent activities of some international organisations, such as the International Coffee Organization, are reviewed in a lengthy Appendix. Any opinions expressed by our contributors are those of the contributors themselves, and do not necessarily reflect those of the editors. R.J. Clarke M.A. (Oxon), Ph.D(Hons), C.Eng., F.I.Chem.E, F.I.F.S.T. Consultant, Chichester, UK O.G. Vitzthum Honorary Professor, Technical University of Braunsweig Scientific Secretary, Association Scientifique Internationale du CafeÂ, Paris (ASIC) Formerly Head of Coffee Research, Kraft Jacobs Suchard, Bremen, Germany
vii
List of Contributors Dr Hartmut H. Balzer, Kraft Foods R and D Inc., Unterbibergerstrasse 15, D-81737, Munich, Germany (1B Acids in Coffee) Professor Thomas W. Baumann, Institute of Plant Biology, University of Zurich, Switzerland (10 Agronomy II: Developmental and Cell Biology) Dr Allan G.W. Bradbury, Kraft Foods R and D Inc., Unterbibergerstrasse 15, D-81737, Munich, Germany (1A Chemistry I: Non-volatile Compounds) Christophe Cavin, Nestle Research Center, Department of Quality and Safety Assurance, PO Box 44, Vers chez-les Blanc, CH1000 Lausanne 28, Switzerland (8 Health Effects and Safety Considerations) Dr Ronald J. Clarke, Ashby Cottage, Donnington, Chichester, West Sussex, PO20 7PW, UK (6 Technology III: Instant Coffee) Anne Constable, Nestle Research Center, Department of Quality and Safety Assurance, PO Box 44, Vers chez-les Blanc, CH1000 Lausanne 28, Switzerland (8 Health Effects and Safety Considerations) Mr C. Pablo R. Dubois, International Coffee Organization, 22 Berners Street, London, W1P 4DD, UK Professor Dr.-Ing. Rudolf Eggers, Technical University Hamburg-Harburg, Eissendorfer Strasse 38, D-21073 Hamburg, Germany (4 Technology I: Roasting) Professor Werner Grosch, Burgstrasse 3B, D-85604, Zermeding, Germany (3 Chemistry III: Volatile Compounds) Dr Wolfgang Heilmann, Hollerender Weg 50, D-28355 Bremen, Germany (5 Technology II: Decaffeination) Professor Selichi Homma, Ochanomizo University, Department of Nutrition and Food Science, Ohtsuka 2-1-1 Bunkyo-ko, Tokyo, 112-8610, Japan (2 Chemistry II: Non-volatile Compounds Part II) Dr Isbelle KoÈlling-Speer, Institut fuÈr Lebensmittelchemie, Facultat fuÈr Mathematik and Naturwissenschaften. Technische Universitat Dresden, D01062 Dresden, Germany (1C Lipids)
Marino Petracco, Illycaffe Spa, Via Flavia 110, 34147 Trieste, Italy (7 Technology IV: Beverage Preparation: Brewing Trends for the New Millennium) Dr.-Ing. Arne Pietsch, Eurotechnica IngenieurbuÈro GmbH, Bartemeide, Germany (4 Technology I: Roasting) Dr Benoit Schilter, Nestle Research Center, Department of Quality and Safety Assurance, PO Box 44, Vers-chez-les Blanc, CH1000 Lausanne 28, Switzerland (8 Health Effects and Safety Considerations) Dr Maro Sondahl, Fitolink Corporation, 6 Edinburgh Lane, Mount Laurel, 08054 NJ, USA (10 Agronomy II: Developmental and Cell Biology) Professor Karl Speer, Institut fuÈr Lebensmittelchemie, Facultat fuÈr Mathematik und Naturwissenschaften. Technische UniversitaÈt Dresden, D-01062 Dresden, Germany (1C Lipids) Dr John Stiles, Integrated Coffee, Technologies Inc., 4 Waterfront Plaza, Suite 575, 500 Ala Moana Boulevard, 96813 Honolulu, Hawaii, USA (11 Agronomy III: Molecular Biology) Angelika Tritscher, Nestle Research Center, Department of Quality and Safety Assurance, PO B ox 44, Vers-chez-les Blanc, CH1000 Lausanne 28, Switzerland (8 Health Effects and Safety Considerations) Dr Herbert Van der Vossen, Steenhuil 18, 1606 CA Venhuizen, The Netherlands (9 Agronomy I: Coffee Breeding Practices)
ix
Chapter 1
Chemistry I: Non-volatile Compounds 1A:
Carbohydrates
A.G.W. Bradbury Kraft Foods R and G Inc., Munich, Germany 1.1 INTRODUCTION
(1) Sucrose contents for the arabica samples varied from 6.25% to 8.45% and those for the robusta samples from 0.9 to 4.85% (apart from two low values of 0.9 and 1.25%, the robustas analyzed fell in the range given above). (2) Robustas contained more reducing sugars than arabicas. (3) Apart from sucrose and one robusta which contained a minimal content of maltose (0.01%), there was no evidence of other simple oligosaccharides such as the `flatulent sugars', raffinose or stachyose, in the green beans.
This review summarizes the literature in the field of coffee carbohydrate chemistry with an emphasis on work done since the chapter on the subject by Trugo (1985). The importance of the carbohydrate fraction in coffee is evident in its high content; on a dry weight basis, it constitutes about half of the green coffee bean. Low and high molecular weight carbohydrate components are present in green coffee; these both participate in the extensive chemical changes associated with coffee roasting. Carbohydrates are also present at a high level in roasted and soluble coffee products. In this chapter, the contents and properties of the carbohydrate fractions in green, roast and soluble coffees will be described.
It would be expected that the action of endogenous enzymes during ripening or processing would lead to the presence of monosaccharides derived from the polysaccharides and, indeed, small yields of arabinose were found in all coffees, and mannose was found in most coffees, although none of the samples contained free galactose. The sugar data described above was obtained mainly by use of HPLC or GC-based methods. However, the recent advent of ion exchange chromatography coupled with ampometric detection allows excellent resolution with low detection limits, and it is now the preferred technique for the analysis of sugars in coffee, particularly in commercial soluble products (see section 1.4.1). By use of this technique, Zapp (personal communication) obtained sucrose contents between 2.60% and 3.02% for three green Indonesian robustas coffees. He also obtained values of 6.6% for a dry processed arabica (Brazil) and 7.02% (New Guinea) and 6.5% (Mexico) for two wet processed arabica samples. It would be expected that sucrose content would increase with degree of ripening and this was apparent with defective coffee beans, where for both immatureblack and immature-green Brazilian beans, sucrose
1.2 GREEN COFFEE 1.2.1 Low molecular weight carbohydrate The principal low molecular weight carbohydrate or sugar in green coffee is sucrose; the monosaccharide content is relatively low. Published values show much variation among bean types, although, in general, arabican varieties tend to contain about twice as much sucrose as robustas. Most literature values for sucrose are in the range of approximately 2% to 5% for robusta beans and 5% to 8.5% for arabicas (summarized by Clifford 1985). Silwar and LuÈllman, (1988; LuÈllman & Silwar, 1989) used an HPLC-based method to determine the low molecular weight carbohydrate profile of 20 green coffee samples from 13 different producer countries) (Table 1.1). The key findings from their work can be summarized as follows: 1
2
Table 1.1 1988.) Coffee
Coffee: Recent Developments
Mono- and disaccharide content of green arabica and robusta coffees. (From Silwar & LuÈllman, Sucrose
Fructose
Glucose
Mannose
Arabinose
Rhamnose
Total
Arabica Colombia Colombia Salvador Brazil Brazil Kenya Kenya Tanzania Ethiopia Ethiopia New Guinea East India
8.20 8.30 7.30 6.65 6.30 8.45 7.05 7.55 6.30 6.25 7.70 6.50
0.15 0.07 0.02 0.15 0.15 0.02 0.03 0.20 0.40 0.25 0.07 0.04
< 0.01 0.30 < 0.01 < 0.01 < 0.01 < 0.01 < 0.01 0.45 0.40 0.45 < 0.01 < 0.01
ND ND ND 0.02 0.10 <0.01 0.06 0.08 ND ND ND ND
< 0.01 0.05 0.09 0.15 0.07 0.07 0.07 0.05 < 0.01 0.04 0.06 0.10
ND ND 0.02 ND ND 0.01 ND ND ND ND 0.01 ND
8.35 8.72 7.43 6.87 6.62 8.55 7.21 8.33 7.10 6.99 7.84 6.64
Robusta Madagascar Cameroun CoÃte d'Ivoire CoÃte d'Ivoire Indonesia Indonesia Philippines Philippines
3.90 3.20 3.40 0.90 1.25 3.00 4.00 4.85
0.25 0.30 0.35 0.55 0.25 0.20 0.40 0.35
<0.01 <0.01 0.20 0.50 <0.01 0.35 0.35 0.50
ND ND ND ND ND 0.06 0.02 ND
0.12 0.09 0.09 0.15 0.05 0.07 0.10 0.04
0.02 0.01 0.02 < 0.01 0.01 < 0.01 < 0.01 < 0.01
4.29 3.60 4.06 2.10 1.56 3.68 4.87 5.74
All values are expressed as % dry basis. ND: not detected.
levels were at a third and fifth of the level of normal beans, respectively, (Mazzafera, 1999). In another study, black bean occurrence in Indian coffee was attributed to low carbohydrate accumulation during bean development, resulting in sucrose levels about half of normal beans (Gopal & Venkataramanon, 1974). This was also observed for Vietnamese robustas, where black beans sorted from a sample contained 0.9% sucrose; normal beans from the same batch contained 4.0% sucrose (Zapp, personal communication). Hydrolysis of sucrose to give glucose and fructose was observed on steam treatment of green coffee by Luger and Steinhart (1995). Yields of the invert sugars increased with degree of steaming, as did the small yields of the polysaccharide hydrolysis products, mannose and galactose. The keto sugar psicose was also produced at high steaming levels, presumably due to a rearrangement of fructose. An HPLC method, which used an anion exchange column, was developed for the determination of inositol tri-, tetra-, penta- and hexaphosphate (phytic acid) in coffee by Franz and Maier (1993). Green beans contain the hexaphosphate as well as lower levels of
pentaphosphates. Total inositol phosphate contents in robusta beans tended to be higher than those in arabica beans where, for the four samples of each monitored, ranges were in the range of 0.34 to 0.40% and 0.28 to 0.32%, respectively (Franz & Maier, 1994). A presteaming treatment of a Salvador arabica sample led to some hydrolysis of the phosphate esters as indicated by the tri- and tetra-phosphate derivatives formed (Franz & Maier, 1994). Polyalcohols have also been found in green coffee; Noyes and Chu (1993) found low levels of mannitol in blends of Brazilian robustas and arabicas (average content 0.027%). A series of diterpene glycosides, the atractylglycosides, was identified in green and roast coffee in the 1970s by the group of Spiteller (Ludwig et al., 1975). The contents of the principal components of this compound class, termed KAI, KAII and KAIII, were determined for coffee by Maier and Wewetzer (1978). Bradbury and Balzer (1999) have shown, by using LC-MS, that each of these glycosides in green coffee actually contains two carboxyl groups at C-4, and consequently the compounds were identified as carboxyatractylglycosides. The structures of the
Chemistry I: Non-volatile Compounds: Carbohydrates
compounds are given in Fig. 1.1. The molecules readily underwent decarboxylation, and mild thermal conditions, such as steaming treatment, led to the quantitative removal of a carboxyl group from each glycoside. The lability of the carboxyl group probably explains why earlier studies only observed the atractyl forms. Measured contents for different bean types confirmed the previous finding by Maier and Wewetzer (1978) that arabica beans contained significantly more of the glycosides than robusta beans (Table 1.2). They were also able to identify three isomers of CKAIII, although levels were too low for accurate quantification. Other similar glycosides have also been found in coffee.
R1 H
R2
Compound
MW
CKA II
526
CH3 CH3
CKA III
610
CH3 CH3
CKA I
772
H
H
O k ±C±CH2±CH
b±D±Glucp(1?)
O k ±C±CH2±CH
Fig. 1.1 Carboxyatractylglycosides in green coffee. (From Bradbury & Balzer, 1999.)
1.2.2 High molecular weight carbohydrate Although various types of polysaccharide have been proposed as coffee bean components over the years, it has now been established that the coffee bean polysaccharide fraction is dominated by three polymer types: arabinogalactan, mannan (and/or galactomannan) and cellulose. Bradbury and Halliday (1990) confirmed this by applying polysaccharide structural analysis to ground, defatted, desugared green beans which had been solubilized in 4-methylmorpholine Noxide. The carbohydrate linkage pattern so obtained (Fig. 1.2) represented a combination of three polysaccharide types: a highly branched arabinogalactan, a linear mannan with a low degree of substitution and an unsubstituted glucan (cellulose). `Total carbohydrate' analysis (the yield and type of the monosaccharides produced by acid hydrolysis) of a series of different arabica and robusta bean types (Bradbury & Halliday, 1987) showed that the mannan and cellulose contents
3
of arabica and robusta beans were similar (ca 22% and 7%, respectively), whereas arabinogalactan contents of the former were typically lower (ca 14% versus 17%). Values obtained by Zapp (personal communication) were similar, although slightly lower mannan contents were obtained for robustas (Table 1.3). The described polymer contents are based on the assumption that all of the galactose is associated with the arabinogalactan polymer and do not allow for the small fraction of galactose which is bound to the mannan fraction (see Section 1.2.1(b)).
(a) Arabinogalactan structure The first attempt to characterize the structure of coffee arabinogalactan (AG) was by Wolfrom and Patin (1965). They isolated the polymer from green coffee (the yield was about 1%, on a bean weight basis) and proposed a b1±3 linked backbone of galactose units with frequent single (arabinofuranosyl) and two (galactopyranosyl linked at C-3 to arabinofuranosyl) unit side chains linked at the C-6 position to galactopyranosyl units b1±3 linked in the main chain. This structure qualifies the polymer as a `Type II' AG. In their study on the structure of the coffee bean polysaccharides, Bradbury and Halliday (1990) confirmed the Type II classification of the AG, but suggested that the polymer contained additional structural features. These included non-terminal arabinose units (linked 1±5), 1±6 linked galactose units and terminal galactose units. Partial hydrolysis of an AG-enriched preparation by means of weak acid showed that, as the arabinose content decreased due to the higher hydrolytic lability of the glycosidic linkages to furanosides (i.e. those to arabinose) as opposed to those to pyranosides (i.e. those to galactose), then the terminal galactose content as well as that of the 3±6 linked galactose increased. This confirmed the earlier claim of galactose involvement in side chains, but also indicated that galactose±galactose linkages (1±6) were also present, probably in side chains. Fischer et al. (1999) isolated a cell wall material (CWM), which contained all of the polysaccharide fraction, from green coffee beans; 20% of the CWM was water-soluble and contained soluble AG. Comparison of their linkage analysis for this preparation with that obtained for the coffee AG-enriched fraction isolated in the study described above (Table 1.4) shows similarities. The AG isolated from CWM had a higher level of branching (branched galactose:non-branched galactose 1:1.15 versus 1:3.5). This difference could result from a heterogeneous population of AG
4
Coffee: Recent Developments
Table 1.2 Content of carboxyatractylglycosides and atractylglycosides in coffee (g/kg). (From Bradbury & Balzer, 1999.) Green
Wet processed arabicas
Costa Rica I Guatemala EI Salvador Colombia Excelsa I Colombia Excelsa II Kenya Ethiopia Harar Ethiopia Djimmah Papua Neuguinea Indonesia Sumatra Jamaica Blue Mountain Burundi Zimbabwe Brazil Minas Brazil Santos Brazil Old Crop (1 year) Brazil Old Crop (5 years) Vietnam Indonesia Decaffeinated Indonesia
Dry processed arabicas
Robustas
Peaks Linkage position 1 2 3 4 5 6 7 8 9 10
5 6
t t t 5 4 3 4 6 2,4 4,6 3,6
Methyl position
Mono type
2,3,5 2,3,4,6 2,3,4,6 2,3 2,3,6 2,4,6 2,3,6 2,3,4 3,6 2,3 2,4
Arabinose Mannose Galactose Arabinose Mannose Galactose Glucose Galactose Mannose Mannose Galactose
I.S.
1 10 2 34
0
5
10
89 7
15 20 25 Time (min)
30
35
Fig. 1.2 Linkage analysis of green robusta coffee beans. GC separation of partially methylated alditol acetates. (From Bradbury & Halliday, 1990.)
Steamed
CKAII
CKAII
2.0 1.9 2.0 2.4 2.2 2.2 2.0 1.0 1.8 2.0 1.4 2.0 1.0 2.2 2.2 2.1 2.1 0.2 0.1 0.2
0.2 0.4 0.1 1.0 0.9 0.5 1.1 3.1 1.0 1.0 1.1 0.9 0.5 1.9 2.4 1.3 1.3 0.2 0.6 0.8
KAII
KAI
1.9 1.5
0.8 0.6
1.6
1.5
Roasted KAII
KAI
1.4 1.4
0.2 0.4
1.2
0.1
1.2
0.9
0.03
0.06
polymers in the coffee bean. The actual degree of branching cannot be determined because of the incorporation of an unknown amount of unbranched galactose units in side chains. There was also no indication of 1±6 linked galactose in the CWM AG. Free coffee bean AG is very soluble and does not impart much viscosity to an aqueous solution. This is typical behaviour of Type II AGs where, for example, larch wood AG, despite a molecular weight range of ca 40 000 (Prescott et al., 1995), can be readily solubilized at levels of 50% (wt) in water. This high solubility can be attributed to the open helical form of the b(1±3) galactan main chain and the frequent, irregular side chains, which restrict opportunities for inter- or intrahydrogen bonding in aqueous solution. Despite the high solubility of the polymer it is not readily extractable from ground green coffee. For example, even after heating in 20% NaOH overnight at 1008C, 45% of the arabinogalactan remained in the insoluble fraction (Bradbury & Halliday, 1990). This suggests that the polymer is covalently bound to less soluble components of the coffee cell wall matrix. A likely candidate is protein, which is known to be covalently linked to Type II AGs in certain plant systems (Prescott et al.,
Chemistry I: Non-volatile Compounds: Carbohydrates
5
Table 1.3 Polysaccharide analysis of green coffees (% dry weight basis).1 From Zapp, personal communication; Bradbury & Halliday, 1987). Arabinose
Galactose
Glucose
Mannose
Total
Reference
Arabica El Salvador Colombia I Ethiopia Colombia II Brazil El Salvador
3.6 3.4 4.0 3.4 3.5 3.4
22.5 22.2 21.3 21.6 20.7 21.5
6.7 7.0 7.8 ND ND ND
10.7 10.4 11.9 10.5 10.0 10.7
43.5 43.0 45.0 ND ND ND
16 16 16 5 5 5
Robusta India Ivory Coast Sierra Leone Cameroons Indonesian I Indonesian II
4.1 4.0 3.8 4.0 3.9 4.3
21.9 22.4 21.7 19.2 19.5 19.6
7.8 8.7 8.0 ND ND ND
14.0 12.4 12.9 13.4 11.9 13.8
48.2 48.3 46.9 ND ND ND
16 16 16 5 5 5
1
Determined by acid hydrolysis; values correspond to anhydro-monosaccharides. ND: not determined.
1995). Protease treatments did increase AG extraction yield, but much of the polymer remained insoluble (Bradbury & Halliday, 1990). The exact chemical structure of coffee AG still remains to be defined. In order to elucidate this, it will first be necessary to obtain a high purity preparation by identifying the nature of the matrix±AG linkages and by utilizing a chemical or enzymatic treatment to cleave them and allow solubilization of the polymer. Gel permeation chromatography can then be used to give pure preparations, which could be characterized by detailed structural analysis used in combination with specific hydrolytic enzymes. Table 1.4 A comparison of the results of linkage analysis of arabinogalactan isolated from green coffee by Bradbury and Halliday (1987) and Fischer et al. (1999). Linkage
Bradbury & Halliday (1987) (mole %)
Fischer et al. (1999) (mole %)
t-rhap t-araf 5-araf 2-araf t-galp 3-galp 6-galp 3,6-galp
Ð 15.4 9.8 2.1 5.0 44.5 1.9 21.5
4.7 28.6 17.2 Ð 3.9 24.5 Ð 21.1
(b) Mannan structure The first structural characterization of mannan isolated from green coffee beans was by Wolfrom and Patin (1961). Their isolated polymer was identified as a b(1± 4) linked mannan, with a molecular weight of approximately 7000 and one galactose unit side chain for every 47 mannose units, i.e. about one galactose unit per molecule. This mannan represented 5% of the green bean dry weight and thus about 20% to 25% of the total green bean mannan. Bradbury and Halliday (1990) also isolated a linear mannan from robusta green beans with even fewer galactose side chains (ca 1 in 130) at a yield of 9.5% of the bean weight. This mannan was obtained from green bean holocellulose (Wolfrom et al., 1960) by a series of extraction steps prior to extraction with 10% alkali and subsequent neutralization with HCl. Thus, about half of the total bean mannan was shown to be a linear b(1±4) linked polymer with very few galactose side chains. The same authors showed evidence of mannan side chain activity in their total bean structural analysis (1 in 18 mannose units linked at the C-2 position; 1 in 16 mannose units linked at C-6), but did not show evidence of such side chains in any isolated fractions and attributed their observations to possible undermethylation of the mannan fraction. Fischer et al. (1999) used precipitation with barium hydroxide to obtain a mannan fraction free from arabinogalactan from their soluble CWM pre-
6
paration (see previous section). Structural analysis indicated single unit galactose side chains at the C6 position of about 1 in 30 of the mannose units. End-group analysis (frequency of terminal mannose units) confirmed a low average number molecular weight of the polymer of about 3500. In another study (see Section 1.3.2) structural analysis also indicated that galactose side chain density was higher in the soluble mannan fractions found in roast coffee extracts compared to those in insoluble fractions. Evidence suggests that coffee bean mannan molecules, in terms of chemical structure, are present as a heterogeneous mixture of unsubstituted and substituted mannans where the frequency of galactose side chains is generally quite low. Galactose substitution levels are discussed in more detail in the roast coffee high molecular weight carbohydrate section (1.3.2). Both `mannan' and `galactomannan' are used in the literature for the coffee polymer and based on the structural analyses, there is justification for the use of both terms. The low level of galactose substitution in the majority of the coffee polymer molecules and the relatively low molecular weights generally tend to favor the mannan term. However, galactomannan is also an appropriate nomenclature for the lower content, higher galactose side chain density, soluble polymer found in roast and soluble coffee extracts (see Sections 1.3.2 and 1.4.2).
(c) Other polysaccharides The glucose produced by the saccharification of green beans desugared by aqueous ethanol extraction is all derived from cellulose, a b(1±4) linked glucan (Bradbury & Halliday, 1990; Trugo & Macrae, 1985). Contents of 6.7% to 7.8% were determined for three arabica coffees and 7.8% to 8.7% for three robustas (Table 1.3). Although the presence of minute quantities of monosaccharides such as xylose and rhamnose in acidic hydrolysates of green beans offers evidence for other polymers, their content is small; values were similar for both arabica and robusta types and averaged 0.2 and 0.3%, respectively (Bradbury & Halliday, 1987). Thus, polysaccharides common to the plant kingdom such as starch or pectin are present, if at all, at only low levels in mature coffee beans. It has been suggested that small quantities of residual parchment in the green coffee beans (known to contain polysaccharide yielding xylose on hydrolysis) cannot be discounted.
Coffee: Recent Developments
1.3 ROAST COFFEE 1.3.1 Low molecular weight carbohydrate On roasting, sucrose is rapidly degraded and its content is minimal in a `normal' roast coffee. Trugo and Macrae (1985) showed that for both an arabica and a robusta sample, sucrose loss for a light roast was about 97% and for a dark roast about 99%. Hughes and Thorpe (1987) found sucrose contents of 0.24% and 0.33% for two conventionally roasted coffees and 0.08% and 0.11% for two conventionally roasted decaffeinated coffees. The sucrose hydrolysis products, glucose and fructose, are formed on roasting but these reducing sugars degrade thermally even more rapidly than sucrose itself so their contents are also minimal in normal roasts. Noyes and Chu (1993) found an average total of only 0.1% sugars in 21 roasted Brazilian arabica/robusta blends. Most of this was sucrose with a content of 0.08%. Mannitol content was 0.017%, which indicated some degradation of this polyol during roasting (average green bean content 0.027%). The keto sugar psicose, presumably formed by rearrangement of fructose, was detected at low levels in roast coffee (Blanc & Parchett, 1989). As the molecule of water required to cleave the glycosidic bond by hydrolysis is not always available under high temperature roasting conditions, it would be expected that non-hydrolytic scission may occur so as to give anhydro-sugars such as those obtained by the pyrolysis of oligo- or polysaccharides such as cellulose (Maga, 1989). However, although 1,6-anhydro-glucose has been detected in roast coffee (Bradbury, unpublished results), determined contents were less than 0.01%. Franz and Maier (1994) showed that, on roasting, dephosphorylation of inositol phosphates took place. At early stages of roasting, hexaphosphate contents decreased, whereby those of penta-, tetra- and triphosphates increased. At higher degrees of roast, levels fell off slowly with contents decreasing with degree of phosphorylation. Apart from dephosphorylation, the compounds were relatively stable; 75% of the original green bean inositol phosphate content was still present in an E1 Salvador sample roasted to 16% organic roast loss. Solubility increased with decreasing degree of phosphorylation, whereas, under typical roast coffee percolation conditions, IP3 was completely soluble; only up to 80% of the IP6 derivative was extracted. The atractyloglycosides KA I, II and III were formed on roasting by quantitative decarboxylation of the corresponding carboxyatractyloglycoside forms
Chemistry I: Non-volatile Compounds: Carbohydrates
(Bradbury & Balzer, 1999). These derivatives have been shown to have good stability and typically they are only about 50% degraded under `normal' roast conditions (Maier & MaÈtzel, 1982; Bradbury & Balzer, 1999) (Table 1.2)
1.3.2 High molecular weight carbohydrate The coffee bean cell walls are thick and are responsible for the hardness of the bean. Microscopy studies have shown that much of the cell wall macrostructure survives conventional roasting conditions (Wilson et al., 1997). As the polysaccharides are the principal cell wall components, they should be relatively stable to roasting and this has been confirmed analytically. Application of the complete coffee polysaccharide structural analysis method used by Bradbury and Halliday (1990) to give GC `fingerprints' of the methylated sugars representative of polysaccharide linkages showed that all of the linkages present in green coffee were maintained after roasting. Nevertheless, as has previously been shown by the group of Thaler (1975), there is some degradation of the polysaccharide fraction on roasting. Total carbohydrate analysis of robusta beans at different roasts (Fig. 1.3) indicated that the relative lability of the polysaccharide sugar units was in the order arabinose > galactose > mannose > glucose. Arabinose, which occurs principally as a terminal unit on the AG side chains was most labile. Cellulose is one of nature's most stable polysaccharides and this is reflected in the minimal bound glucose loss on roasting. Roasting improves the extractability of the coffee polysaccharides. Leloup and Liardon (1993) have shown how the solubility of arabinose-, galactose- and mannose-containing molecules, at 958C, was increased
Fig. 1.3 Total carbohydrate analysis of green Cameroon robusta beans at three degrees of roast. (From Zapp, personal communication.)
7
by ca 1 g/100 g in green to ca 6 g/100 g in roasted Colombian arabica beans. Arabinogalactan was solubilized as arabinose, arabinose-containing oligomers or as reduced molecular weight polymer (as compared to green coffee). The ratios of Ara/Gal in extracted polymer were reduced from 1:7 (green beans) to 1:12 (roasted beans). Galactomannan (and/or mannan) extraction was negligible from green beans, but ca 2.5% (bean weight basis and about 12% of the bean polymer) was extractable from roast beans at 958C. Size exclusion chromatography indicated molecular weight ranges of 200 to 50 000 and 800 to 80 000 for the AGand galactomannan-derived fractions in the roast coffee extracts, respectively. Bradbury (unpublished results) extracted three roasted coffee samples, and added ethanol (three volumes to one aqueous extract) to precipitate a polymer fraction from which a mannan-rich preparation was isolated via precipitation with Fehling's solution. Linkage analyses indicated a degree of branching ranging from 12.9 to 26.4 for the three samples (Table 1.5). The proportion of end units indicated number average DPs (degrees of polymerization) ranged between 14.8 and 26.2, which was considerably lower than the values derived from the size exclusion chromatography profiles given above. It must be noted that the isolated complexes constituted less than 5% of the total roast bean mannan and are not representative of the whole mannan structure. It is apparent that the side chain branching density was higher in the soluble mannan fraction compared to the major residual fraction, which was not extracted under the conditions used. Navarini et al. (1999) used a detailed fractionation scheme to separate polysaccharide-rich fractions from dark roast arabica extracts (at 908C). Fractions were characterized by total carbohydrate analysis and 1 H and 13 C NMR was used to identify linkage positions in the bound monosaccharide residues. Addition of ammonium sulphate to the coffee extract produced a precipitate, which was redissolved, dialyzed and treated with isopropanol to give Fraction A. With a molar percentage of 75%, mannose was the main constituent of this fraction, which was obtained in 2.2% yield from the beans. Fraction A was treated with 0.1M NaOH (3 h, room temperature) and during the subsequent dialysis step, 12.4% of the material precipitated out to give Fraction Ans. This fraction contained 96.3% carbohydrate, of which 95% (molar basis) was bound mannose. The reason for the insolubilization was not clear, although the authors suggested it was a result of the alkali-induced removal of bound proteinaceous
8
Coffee: Recent Developments
Table 1.5 Linkage analysis of mannans isolated from roast coffee (roast colour (medium) 12 colour units) extracts. From Bradbury (unpublished data). Linkage4
Mannan1 Yield (% wt)
t-man
t-gal
4-man
4,6-man
Dp2
DB3
Vietnam Robusta
0.27
0.043
0.086
1
0.041
26.2
26.4
Uganda Robusta
0.73
0.070
0.104
1
0.090
17.9
12.9
Colombia Arabica
0.81
0.080
0.068
1
0.053
14.8
21.4
Coffee
1
Isolated from Fehling's complex. Average number degree of polymerization. 3 Average degree of branching based on [4,6-man] = No of branch points. 4 Mole, relative to 4-man 2
material. Another possibility is the formation of the insoluble form of mannan by crystallization. The authors suggested that galactose and arabinose residues were covalently associated with the mannan polymer, but because of their low content they were unable to obtain structural evidence. Barium hydroxide precipitates b1±4-linked mannan (Wolfrom et al., 1961) and treatment of Fraction As (freeze-dried supernatant after removal of Fraction Ans with this reagent produced a sediment and a soluble polymer. Molar percentages in the latter were 76.9% galactose, 13.4% arabinose and 9.7% mannose. Linkage analysis corresponded well with the structural features described earlier for coffee AG (Bradbury & Halliday, 1990). The authors claimed that the findings of their study could be extrapolated to green coffee bean polysaccharides, and it would be interesting to apply their approach to a green coffee sample. However, it must be considered that roast-induced structural changes could be induced in the polysaccharide fraction, particularly in view of the dark roasting conditions used in this study. Also, the structures of the polysaccharides isolated in this study represent only a minor part of the total coffee bean polysaccharide and are not necessarily characteristic of the whole fraction.
1.4 SOLUBLE COFFEE 1.4.1 Low molecular weight carbohydrate In recent years, there has been a large amount of activity in the area of sugar profiling of commercial
soluble coffee products. This can be attributed to first, the enormous range of coffee-based soluble products commercially available and the need to have an analytical means of classifying them, and, second, the need to monitor the presence of non-coffee bean components in a product and establish a method which allows the definition of analytically defined limits for a nonadulterated product. The advent of high resolution, sensitive methods for carbohydrate analysis has been of much benefit in this field. As described in Section 1.3.2, the sugar content of roasted coffee, and thus that of extracts at atmospheric pressures, is low. In commercial soluble coffee processes, the elevated temperatures used lead to hydrolysis of the polysaccharide fraction reflected in the increased levels of the constituent monosaccharides arabinose, galactose and mannose. This generation of reducing sugars is apparent in the material balance data for roast to soluble coffee (Brazilian arabica and robusta blends) extracted in two stages (708C and 1908C, each for 20 minutes) of Noyes and Chu (1993). Average contents (21 trials) of roast and corresponding extract were: 0.0042% and 0.666% (arabinose), 0 and 0.137% (galactose) and 0.006% and 0.36% (mannose). Blanc et al. (1989) monitored the effect of autoclave extraction time and temperature (subsequent to `atmospheric extraction' at 1008C) on the free and total carbohydrate profile in arabica and robusta coffees (Table 1.6). Arabinose, the most labile of the bound monosaccharides, was released most readily; galactose and mannose extract contents increased with extraction time and temperature until under the most extreme conditions (1908C, 240 minutes) carbohydrate degradation was significant.
Chemistry I: Non-volatile Compounds: Carbohydrates
9
Table 1.6 Free and total carbohydrate contents of autoclave extracts made from arabica (Santos) roasted coffee (% dry weight basis). First stage extraction conditions for all samples: 30 min, 1008C. (From Blanc et al., 1989.) Conditions of 2nd stage extraction
Free carbohydrate
Total carbohydrate
Min
8C
Ara
Fru
Man
Glu
Gal
Xyl
Ara
Man
Glu
Gal
30
150 160 170 180 190 160 170 180 160 170 180 190
1.35 1.74 1.89 1.44 0.52 2.17 1.19 0.64 1.44 0.62 0.03 0.00
0.11 0.10 0.11 0.17 0.47 0.11 0.24 0.77 0.24 0.57 1.08 0.03
0.05 0.08 0.17 0.43 1.42 0.32 0.82 2.54 0.84 2.02 2.82 0.43
0.09 0.10 0.09 0.10 0.19 0.11 0.14 0.36 0.16 0.27 0.70 0.01
0.10 0.19 0.38 0.79 1.61 0.82 1.55 2.36 1.79 2.31 1.44 0.01
0.14 0.16 0.21 0.17 0.21 0.17 0.20 0.13 0.18 0.14 0.06 0.08
6.29 6.00 5.34 3.55 2.53 4.26 2.65 1.03 2.70 2.01 1.72 2.05
10.66 8.95 9.93 13.09 19.50 10.23 13.40 15.50 11.30 13.70 9.36 5.91
1.86 1.40 1.32 1.29 1.73 1.29 1.45 1.87 1.36 1.86 2.00 1.66
12.76 18.40 20.60 18.00 14.63 20.93 16.40 10.08 17.00 11.48 4.61 3.19
120 240
Glucose and fructose contents also increased, some glucose may have resulted from cellulose that had been partially depolymerized by roasting but the atractylglycosides (Section 1.3.1) were a likely source. The resulting fructose was probably a thermally generated rearrangement product of glucose and mannose (via the 1,2-enediol intermediate form). Leloup et al. (1997) monitored the effect of degree of roast on the carbohydrate composition of roast coffee extract extracted at high temperature conditions (1808C, 20 minutes). Although the total extracted carbohydrate was similar for light, medium and dark roasts, the carbohydrate composition both in terms of type (Table 1.7) and size distribution (Fig. 1.4) showed significant differences. For light and dark roasts, total yields of arabinose plus galactose were 22.4% and 15.4% and for mannose 8.7% and 16.0%, respectively. Thus AG was apparently degraded more at higher roasts, supporting earlier observations that AG was more labile than mannan on roasting. The relatively stable mannan became more accessible as the tough cell wall matrix was weakened with increasing degree of roast. HPLC showed the presence of oligomannans DP 1 to 7 in all extracts, indicative of mannan hydrolysis during extraction. In a recent patent, Gerhard-Rieben et al. (1999) used the carbohydrate molecular weight profile to indicate the degree of hydrolysis occurring during their coffee extraction process, which incorporated three extraction
stages and a thermal hydrolysis stage. Most hydrolysis, apparent in the content of monosaccharides and oligosaccharides produced from the polysaccharides, occurred at the thermal hydrolysis stage. Zapp and Kuhn (1997) have used MALDI-TOF mass spectrometry to identify mannodextrins from DP 2 to 14 in a commercial sample of soluble coffee. Each oligomer gave a doublet (sodium and potassium adducts) at molecular weights oligomer +23 and 39, respectively (Fig. 1.5). By assuming a linear mass peak area response/molecular weight relationship based on a calibration using mannodextrins DPs 2, 3, 6 and 7, they were able to quantify the contents of DPs 2 to 14 in the sample. Yields decreased steadily with increasing molecular size. As separation was based on molecular weight, it was not possible to determine whether galactose was present in the oligomers. There was no evidence of arabinose-containing oligomers (the molecular weight would be 30 less than the corresponding pure oligohexaose for every bound arabinose unit), even though the total arabinose content in the sample was 3.65%. Total galactose and total mannose contents in the sample were 17.11 and 16.09%, respectively. The reducing sugars in soluble coffee were determined as their enantiomers after derivatization with chiral phenylethylamine using capillary electrophoresis (Noe et al., 1999). The carbohydrate profiling of soluble coffee offers a means of identifying the
10
Coffee: Recent Developments
Table 1.7 Extraction yields and carbohydrate composition (in g/100 g of dry extract) of coffee extracts. (From Leloup et al., 1997.) Ara
Gal
Glu
Xyl
Man
Fru
Total + other
Extract 1: Light roast, extraction yield 38.4% DP 1 1.70 0.92 DP 2 to 6 0.00 4.02 DP > 6 0.67 15.04 Total 2.38 19.99
0.21 0.49 0.21 0.91
0.02 0.07 0.01 0.10
0.31 3.45 4.92 8.68
0.18 0.00 0.00 0.18
3.64 8.03 20.92 32.59
Extract 2: Medium roast, extraction yield 39.2% DP 1 1.45 0.83 DP 2 to 6 0.00 4.41 DP > 6 0.20 12.53 Total 1.65 17.77
0.14 0.60 0.04 0.79
0.01 0.07 0.00 0.08
0.32 4.49 5.60 10.41
0.14 0.00 0.01 0.15
3.16 9.57 18.44 31.17
Extract 3: Dark roast, extraction yield 39.1% DP 1 0.90 0.60 DP 2 to 6 0.00 3.08 DP > 6 0.57 10.24 Total 1.47 13.92
0.08 0.40 0.22 0.70
0.01 0.06 0.02 0.08
0.28 3.84 11.84 15.96
0.14 0.00 0.00 0.14
2.29 7.37 23.05 32.71
Carbohydrate (g/100g R and G)
presence of non-coffee bean adulterants such as husks in commercial soluble coffee products. Blanc et al. (1989) showed that free glucose and fructose values were higher in products where husks had been added to the coffee prior to extract. They also showed that by incorporating the xylose content after acid hydrolysis, they could detect the presence of roasted hulls, in which case the contents of reducing sugars, glucose and fructose, were much lower due to degradation. Levels of the carbohydrate alcohols, inositol and mannitol (the latter identified in coffee for the first time) determined by GC analysis of their trimethylsilyl derivatives, were also enhanced by the presence of husk extracts (Davis et al., 1990). A mannitol level in excess of 0.3% in the
30 25
Coffee oligosaccharides
Dark roast Medium roast Light roast
20 15 10 5 0
10
100
1000
10000
100000
1000000
Molecular weight
Fig. 1.4 GPC profile of extracts from Colombian arabica coffee at three degrees of roast. (From Leloup et al., 1997).
product was shown to be evidence of husk adulteration. Berger et al. (1991) used an enzymatic method to determine the glucose and fructose contents of soluble coffees under three different conditions: as is, and after acid hydrolysis under weak and strong acid conditions. In this way they were able to establish the presence of adulterants such as chicory (the inulin present yielded fructose on weak acid hydrolysis), cereal products and caramel, in addition to other adulterants such as figs, maltodextrins, glucose syrups, starch, unroasted husks and parchments. The method, however, is not intended to identify the type or determine the amount of adulterants. Noyes et al. (1991) used a statistical design to monitor the contribution of husk type and content, coffee blend, degree of roast and extraction conditions on the carbohydrate profiles (determined by GC of the trimethylsilylethers) in soluble coffees. They confirmed earlier work where increased husk levels led to higher contents of total xylose, mannitol and free glucose and fructose. The other factors had less influence and considerable variation in carbohydrate levels was observed over a range of samples. They summarized that an intimate knowledge of raw materials and processing conditions was needed in order to establish product limits. The advent of high-performance anion exchange chromatography with pulsed amperometric determination greatly facilitated the determination of carbohydrates in soluble coffee (Prodolliet et al., 1992). The
1499.7Da '(hexose)9 + Na'
1337.7Da '(hexose)8 + Na'
1176.0Da '(hexose)7 + Na'
1014.1Da '(hexose)6 + Na'
852.3Da '(hexose)5 + Na'
690.3Da '(hexose)4 + Na'
528.4Da '(hexose)3 + Na'
1661.2Da '(hexose)10 + Na'
1500
1191.9Da
1800
11
868.6Da
2100
706.7Da
2400
545.0Da
2700
366.3Da '(hexose)2 + Na'
3000
382.5Da
3300
245.4Da 'N-Acetyl-D-galactosamin + Na' 261.3Da 'N-Acetyl-D-galactosamin + K'
Chemistry I: Non-volatile Compounds: Carbohydrates
1200 900 200
400
600
800
1000
1200
1400
1600
1800
m/z
Fig. 1.5 MALDI mass spectrum of soluble coffee, N-acetylgalactosamine is the internal standard. (From Zapp & Kuhn, 1997.)
technique combines the unique resolution capacity of the pellicular anion-exchange polystyrene-divinylbenzene resin separative column with the sensitivity and specificity of an amperometric detector. Analysis is performed under alkaline conditions, where a complete profile of all relevant low molecular weight carbohydrates can be made in one run. Figures 1.6 and 1.7 show free carbohydrate (before hydrolysis) and total carbohydrate (after hydrolysis) profiles. The method was used in a series of detailed studies (Prodolliet et al., 1995a, 1995b), and also in an interlaboratory study (Prodolliet et al., 1995c), to characterize the carbohydrate profile of soluble coffee and has subsequently been adapted by the International Standards Organization as a standard method for this purpose (ISO 11292), and as first action by the Association of Official Analytical Chemists (AOAC). Several workers have utilized the technique for coffee carbohydrate profiling in soluble coffee. Oestreich-Janzen (1995) showed that
both the high performance anion-exchange chromatography (HPAEC) method and an HPLC procedure, which used a cation-exchange column, could be used to give comprehensive profiles for a variety of coffee substitutes and coffee-like beverages. Oligosaccharide profiling by GC analysis of oxime/trimethylsilyl derivatives has been used to demonstrate chicory and malt derived solids in coffee containing beverages, (Kundel et al., 1998). Limits of acceptability for the relevant low molecular weight carbohydrates have been defined by AFCASOLE (the European Soluble Coffee Manufacturer's Association), described in a publication issued by the British Ministry of Agriculture, Fisheries and Food ((MAFF), 1995) (Table 1.8). In a major survey by MAFF, the free and total carbohydrate profiles of 344 commercial soluble coffees, determined using the ISO procedure, and recommended limits, indicated that 50 (nearly 15%) of the coffees exceeded the 97.5%
Coffee: Recent Developments
confidence level for one or more indicator carbohydrates, and were thus considered adulterated. The adulterated coffees fell into two main groups: those with a high total xylose and glucose, which indicated adulteration with skins or husks; and those with a high total glucose content and a low xylose content, which indicated adulteration with a starch based material.
Fructose Ribose
Mannose
Mannitol
mV
Rhamnose
180
Galactose Glucose
Arabinose
12
1.4.2 High molecular weight carbohydrate
100 0
10
20
30
40
50
Minutes
Xylose
Mannitol
mV
Glucose
800
Mannose
Rhamnose Arabinose Galactose
Fig. 1.6 HPAEC chromatogram of typical total carbohydrate profile of a pure soluble coffee. (From Prodolliet et al., 1995b.)
100 0
10
20
30
40
50
Minutes
Fig. 1.7 HPAEC chromatogram of typical free carbohydrate profile of a pure soluble coffee. (From Prodolliet et al., 1995b.)
Table 1.8 Maximum limits for indicator carbohydrates (`M') and experimental variability interval (`l') for authentic soluble coffee expressed on a % dry matter basis. (Adapted from MAFF, 1995.) Indicator carbohydrates Total glucose
Total xylose Free mannitol Free fructose
`M' limit
`I' interval (+)
`M + l' limit
1.80 0.40 0.30 0.60
0.47 0.18 0.17 0.37
2.27 0.58 0.47 0.97
Although as described in Section 1.4.1, substantial hydrolysis of the polysaccharide fraction occurs during extraction, studies have indicated the presence of polymeric carbohydrate in coffee extracts obtained under commercial conditions. These include precipitation by solvent (Leloup et al., 1997), molecular weight determinations by gel permeation chromatography (gpc) and precipitation of polysaccharide containing mannose and galactose with Fehling's solution (Ara & Thaler, 1976). The extractability of the polysaccharides as a function of extract yield/temperature and bean type was extensively studied in earlier studies by Thaler's group (1975) and more recently by Leloup and Liardon (1993). Ara and Thaler (1976, 1977) used Fehling's solution to precipitate a series of complexes from coffee extracts prepared from arabica and robusta beans at a range of roast colors. Analysis of the acid hydrolyzates gave mannose as the dominant hexose, but relatively high yields of galactose were also obtained. On the assumption that all the galactose was bound to the mannan, the degree of branching was relatively high. Mannose: galactose ratios varied for preparations from light to roast from 3.0 to 29.5 (Santos) and 9.5 to 21.3 (Colombian) for arabicas and 1.7 to 10.0 (Ivory Coast) and 1.55 to 13.2 (Angola) for robustas. The robustas showed a higher density of side chains. Although roasting led to a lower side chain frequency, that for mannans isolated over a range of extraction yields (36.4 to 53.2% for Santos arabicas, 38.8% for Angolan robustas) did not show significant variation. However, covalent linkage of galactose to the mannan polymer was not proven in these studies. It is apparent that the level of galactose substitution in the mannan polymer solubilized in the commercial soluble coffee extraction process is higher than the average level of substitution for the whole mannan fraction, i.e. the level of substitution of the residual unextracted fraction. This, as already suggested by Clifford (1985), would impart more solubility to the mannan fraction. Precipitates formed by addition of ethanol to coffee autoclave extracts yielded somewhat more galactose than mannose for samples derived from light, normal
Chemistry I: Non-volatile Compounds: Carbohydrates
and dark roasted coffees (Leloup et al., 1997). This indicated the presence of arabinogalactan-derived polymer material, even though, due to hydrolysis during extraction, the molecular weight of most of the sample was probably in the oligosaccharide size range. The residual solids (`spent grounds') remaining after the commercial extraction process contain a large polysaccharide fraction. Stahl and Turek (1991) described a process which utilized acid hydrolysis of commercial grounds (mannan content 25%, cellulose 15%) to convert the mannan to produce mannose. A subsequent reduction step (hydrogen/reduced nickel catalyst) was described for the conversion to mannitol. They showed how acid hydrolysis of the mannan polymer was significantly faster than that of cellulose and applied reaction conditions were adjusted (high temperature, short time) so as to allow mannan conversion with very little concurrent cellulose hydrolysis. A related process for the production of a mannose hydrolyzate was also developed (Fulger et al., 1985); (see also Chapter 6 in this book, Section 6.2.11).
13
either as the volatile aroma compounds, or as nonvolatile taste compounds. Maillard reaction products are responsible for the brown product colour. The reactive breakdown products of carbohydrate can also react with the side chains of the basic amino acids in coffee protein. Henle et al. (1996) have claimed that up to 30% of the total arginine reacts with methylglyoxal to give bound imidazolinone (I, Fig. 1.8). In another paper from the same laboratory (Henle et al., 1997), cross-links formed by the reaction of arginine and lysine side chains by reaction with pentose (arabinose in coffee) were identified in roasted coffee (II, Fig. 1.8). Although the concentration was low (11 to 40 mg/kg protein) this reaction could contribute to the formation of melanoidins. These reaction products were detected using an amino acid analyzer following acid hydrolysis as used for analysis of proteins (for a sample containing 40±50 mg of protein: 6N HCl (10 ml, 1108C, 23 h).
1.5 REACTIONS OF CARBOHYDRATES ON ROASTING Sucrose in coffee beans is rapidly degraded on roasting. Early reaction products are the invert sugars, fructose and glucose, as well as 1,6 anhydro-glucose, arabinose and erythritol. These primary reaction products then react in a number of ways: . Fragmentation to form low molecular weight products such as aliphatic acids. Most of the acids generated during coffee bean roasting are formed during the early stages. Ginz et al. (2000) have shown that the carbohydrate fraction, in particular sucrose, was the main precursor. The principal acids generated from carbohydrate were formic, acetic, glycolic and lactic (in order of decreasing molar yield). . Dehydration (caramelization) to form the numerous heterocyclic compounds such as hydroxymethylfurfural. Many compounds of this class are volatile and important contributors to coffee aroma. Some of them are also reactive and can polymerise to give melanoidin-type molecules (Tressl et al., 1998). . Interaction with amino acids or protein to give Maillard products which may be either polymeric (melanoidins) or of low molecular weight. The latter are important contributors to coffee flavor,
Fig. 1.8 Protein-bound structures generated under coffee roasting conditions. (from Henle et al. 1996 and Hofmann et al. 1999.)
The brown complex melanoidin fraction can constitute up to about 30% of a coffee beverage and has been the subject of various investigations. In coffee, carbohydrate and protein are the major precursors for melanoidin formation, although chlorogenic acids also contribute. (See also Chapter 2 in this book, from the amino acid point of view.) In a comprehensive ongoing study, Hofmann has been investigating the chemical mechanisms involved in melanoidin formation. He has referred to the growing number of coloured low molecular weight
14
compounds found in his and other recent studies, which are formed by Maillard reactions between amino acids and carbohydrates (Hofmann, 1998). The ability of certain chromophores to link to lysine side chains in non-colored protein, thereby leading to a colored melanoidin-type polymer, has been demonstrated. These studies have also suggested that another important step in melanoidin formation was the formation of radical containing linkages between protein units (Hofmann, 1999). In model reaction studies between proteins and glyoxal a free radical containing protein bound cross-link (CROSSPY) was identified (III, Fig. 8). The free radical signal of the cross-link resembled that detected in roast coffee which complied with the reaction of lysine side chains with glyoxal, a known carbohydrate degradation product under roasting conditions, to form CROSSPY (Hofmann et al., 1999). Electron spin resonance measurements have shown that a precipitate formed by addition of methanol contained the majority of free radicals detected in a soluble coffee sample. Spectroscopic identification of carbohydrate as the principal component of the fraction coupled with the similarity of the free radical spectrum obtained by the thermal treatment of sucrose suggested that carbohydrate was the precursor of the free radicals in the coffee product (Gonis et al., 1995). Steinhart and Packert (1993) used gel permeation to separate the melanoidins in the hot water extracts of an arabica and a robusta roasted coffee. Each sample gave four coloured fractions with the peak retentions corresponding to molecular weights in the range 1000 to 60 000 daltons. Preparative thin layer chromatography was used to separate each fraction into distinct bands, from which additional subfractions were obtained. Analysis of the acid hydrolyzates indicated the presence of bound carbohydrate, sourced principally from the polysaccharides in the melanoidins. Mannose dominated in the molecular weight gpc fraction, whereas galactose and arabinose as well as glucose and, to a minor extent, rhamnose were concentrated in the lower molecular weight fractions. There was no significant difference between gpc patterns obtained for untreated beans, and those that had been pretreated with steam before roasting (Moller & Steinhart, 1993). It would be expected that as much of the polysaccharide is degraded at the later stages of roasting, many of its degradation products would be incorporated in the high molecular weight melanoidin products. An exception would be the labile terminal arabinose units on the AG polymer of which scission begins at early stages of the roasting process. Stahl and
Coffee: Recent Developments
Parliment (1993) have suggested that this process was involved in the formation of the important aroma compound, furfural mercaptan. They proposed that released arabinose eliminated water to give furfural, which reacted with protein-generated H2 S. De Mario et al. (1995) also suggested AG as a precursor of furfural by monitoring its formation on roasting of a polysaccharide-rich fraction free from sucrose isolated from green coffee beans. A carbohydrate material balance indicated that 42% of the bound arabinose was degraded whereas the levels of the other carbohydrates were affected only slightly. In a later publication, the same group (De Mario et al., 1996) used their green bean extract fractionation technique to show that the yields of furfural and other furans were higher from the degradation of sucrose than from the degradation of polysaccharide during coffee roasting.
1.6 FUNCTIONAL PROPERTIES OF COFFEE CARBOHYDRATES 1.6.1 Role in soluble coffee processing (a) Extract viscosity Commercial soluble coffee extracts typically contain about 30% total carbohydrate, present at various stages of hydrolysis. Consequently, the physical properties of the carbohydrate are a key factor in soluble coffee processing. It is fortuitous that commercial soluble coffee solids have a molecular weight distribution that allows for ease of processing while affording good product stability. Complete hydrolysis of the polysaccharide fraction would result in higher product hygroscopicity, while higher molecular weights would give high extract viscosities. Nevertheless, some viscosity reduction at high solids could facilitate extract processing. Coffee AG has a very low viscosity in aqueous solution and the mannan fraction presumably contributes more to product viscosity. Ehlers (1980) showed how treatment with a mannanase enzyme reduced coffee extract viscosity, facilitating evaporation to a higher soluble content before spray or freeze drying and thereby lowering production costs. Jardine and Moretti (1993) have shown that treatment of coffee extract with a commercial polysaccharide hydrolyzing enzyme reduced the viscosity in the range of 1.66 to 2.56, for commercial coffee extracts with soluble solids content between 45% and 62%. They used the treatment in order to improve the freeze-drying of extract. A patent application for the use of an immobilized mannan-hydrolyzing enzyme to reduce the
Chemistry I: Non-volatile Compounds: Carbohydrates
molecular weight of the mannan fraction in coffee extract for purposes of viscosity reduction in soluble coffee processing has been granted (Nicolas et al., 1999 see also Chapter 6).
(b) Sedimentation The formation of sediments during the processing of freshly prepared extract is a well known phenomenon in the coffee industry. Bradbury and Atkins (1997) showed that such sediments contain mannan as the major component. They used X-ray crystallography to show that the polymer underwent crystallization to an insoluble form during processing. The crystallization process was driven by higher temperatures and extract concentrations. The crystallized mannan was identified as the `mannan I' type, i.e. that which is identified with `Ivory nut' mannan and usually associated with the lower molecular weight forms of the polymer. Crystallization was attributed to hydrogen bond formation between linear, non-branched regions of the mannan molecule. The low total galactose contents of the sedimented molecules were indicative of long linear sections in the molecules which would promote interhydrogen bonding and formation of crystalline regions. Other mannan molecules do not crystallize so readily and remain solubilised in the extract by virtue of their structure, i.e. lower molecular weight and/or a higher density of galactose side chains.
1.6.2 Foam Nunes et al. (1997) have shown that the stability of espresso foams was directly related to the concentration of polysaccharide present. For a series of espresso coffees prepared from Brazilian arabicas and Ugandan robustas, those corresponding to the degrees of roast of 9.7 and 7.65 (organic roast loss), respectively, gave maximum foam stability and also the highest polysaccharide contents. The polysaccharides were precipitated from the espresso extracts using ethanol (55% and 75%) and their constituent monosaccharides (mannose > galactose > arabinose in order of yield) determined by acid hydrolysis, derivatization and GC. The foam stabilizing effect was attributed to viscosity imparted to the extract by galactomannan. Further investigation by these workers (Nunes & Coimbra, 1998) led to the identification of a high molecular weight fraction (ca 2000 kDa, as estimated by GPC) in the precipitates formed in 55% ethanol. Although the content of this fraction was low (0.3% to 0.9% of the total solids) its content correlated well with maximum
15
foam stability, reaching a maximum in the Ugandan robusta 7.65 roast degree sample. A complex structure resulting from roasting induced linkages between polysaccharides, proteins and phenolic compounds was suggested for the material. In another study, Petracco et al. (1999) showed that the polysaccharide containing fraction A (ca 80% total mannose) isolated from dark roast coffee by Navarini et al. (1999), described in Section 1.3.2, showed good foam stabilizing properties (see also Chapter 7).
1.6.3 Coffee fiber The coffee polysaccharides are not hydrolyzed by mammalian enzymes and have thus the potential to act functionally as dietary fiber. Feeding studies have suggested that an AG-rich preparation (coffee fiber) isolated from roast coffee extracts can lower colon cancer risk (Rao et al., 1998).
REFERENCES Ara, V. & Thaler, H. (1976) Investigations on coffee and coffee substitutes. XVIII. Dependence of yield and structure of polymeric galactomannan on the type and degree of coffee roast. Z. Lebensm. Unters.-Forsch., 161, 143±50. Ara, V. & Thaler, H. (1977) Investigations on coffee and coffee substitutes. XIX. Dependence of yield of polymeric galactomannan on the degree of extraction of coffee extract. Z. Lebensm. Unters.-Forsch., 164, 8±10. Berger, A., Bruelhart, M. & Prodolliet, J. (1991) Determination of adulteration in pure soluble coffee by enzymatic sugar determination. Lebens. Wissens. Tech., 24, 59±62. Blanc, M.B., Davis, G.E., Parchet J.-M. & Viani, R. (1989) Chromatographic profile of carbohydrates in commercial soluble coffees. J. Agric. Food Chem., 37, 926±30. Blanc, M. & Parchet, J.M. (1989) Identification and quantification of a new carbohydrate in coffee chemistry. In: Proceedings of the 13th ASIC Colloquium (Paipa), pp. 191±5. ASIC, Paris, France. Bradbury, A.G.W. & Atkins, E.D.T. (1997) Factors affecting mannan solubility in roast coffee extracts. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 128±32. ASIC, Paris, France. Bradbury, A.G.W. & Balzer, H.H. (1999) Carboxyatractyligenin and atractyligenin glycosides in coffee. In: Proceedings of the 18th ASIC Colloquium (Helsinki), pp. 71±8. ASIC, Paris, France. Bradbury, A.G.W. & Halliday, D.J. (1987) Polysaccharides in green coffee beans. In: Proceedings of the 12th ASIC Colloquium (Montreux), pp. 265±9. ASIC, Paris, France. Bradbury, A.G.W. & Halliday, D.J. (1990) Chemical structures of green coffee bean polysaccharides. J. Agric. Food Chem., 38, 389±92.
16
Clifford, M.N. (1985) Chemical and physical aspects of green coffee and coffee products. In: Coffee, Botany, Biochemistry and Production of Beans and Beverage (eds M.N. Clifford & K.C. Willson), pp. 305±74. Croom Helm, London. Davis, G.E., Garwood, V.W., Barfuss, D.L., Hussini, S.A., Blanc M.B. & Viani, R. (1990) Chromatographic profile of carbohydrates in commercial coffees. 2. Identification of mannitol. J. Agric. Food Chem., 38, 1347±50. De Mario, C.A.B., Trugo, L.C., Aquino Neto, F.R. & Moreira, R.F.A. (1995) Arabinogalactan as a potential furfural precursor in roasted coffee. Int. Food Sci. Tech., 29, 559±62. De Mario, C.A.B., Trugo, L. C., Aquino Neto, F. R., Moreira, R. F. A. & Alviano, C. S. (1996) Composition of green coffee water-soluble fractions and identification of volatiles formed during roasting. Food Chem., 55, 203±7. Ehlers, G.M. (1980) Possible applications of enzymes in coffee processing. In: Proceedings of the 9th ASIC Colloquium (London), pp. 267±71. ASIC, Paris, France. Fischer, M., Reimann, S., Trovato, V. & Redgwell, R.J. (1999) Structural aspects of polysaccharides from Arabica coffee. In: Proceedings of the 18th ASIC Colloquium (Helsinki), pp. 91±4. ASIC, Paris, France. Franz, H. & Maier, H.G. (1993) Inositol phosphates in coffee and related beverages. I. Identification and methods of determination. Deutsche Lebensm.-Rundsch., 89, 276±82. Franz, H. & Maier, H.G. (1994) Inositol phosphates in coffee and related beverages. II. Coffee beans. Deutsche Lebensm.Rundsch., 90, 345±9. Fulger, C.V., Stahl, H.D., Turek, E.J. & Bayha, R. (1985) Production of a mannan oligomer hydrolysate. US Patent 4 508 745. Gerhard-Rieben, E., Lebet, C. R., Leloup, V. & Schlecht, K. (1999) Coffee extraction process and patent. US Patent 5 897 903. Ginz, M., Balzer, H.H., Bradbury, A.G.W. & Maier, H.G. (2000) Formation of aliphatic acids by carbohydrate degradation during roasting of coffee. Eur. Food Res. Technol., 211, 404±10. Gonis, J., Hewitt, D.G., Troup G., Hutton D.R. & Hunter C.R. (1995) The chemical origin of free radicals in coffee and other beverages. Free Rad. Res., 23, 393±9. Gopal, N.H. & Venkataramanan, D. (1974) Studies on black bean disorder in coffee. Ind. Coffee, 9/10, 259±67. Henle, T., Deppisch, R. & Ritz, E. (1996) The Maillard reaction ± from food chemistry to uraemia research. Nephrol. Dial. Transplant, 11, 1718±22. Henle, T., Schwarzenbolz, U. & Klostemeyer, H. (1997) Detection and quantification of pentosidine in foods. Z. Lebensm. Unters.-Forsch. A., 204, 95±8. Hofmann, T. (1998) Studies on melanoidin-type colorants generated from the Maillard reaction on protein-bound lysine and furan-2-carboxaldehyde ± chemical characterization of a red coloured domaine. Z. Lebensm. Unters.-Forsch. A., 206, 251±68. Hofmann, T. (1999) 4-Alkylidene-2-imino-5-[4-alkylidene-5oxo-1,3-imidazol-2-inyl]aza-methylidene-1,3-imidazolidine ± a novel colored substructure in melanoidins formed by Maillard
Coffee: Recent Developments
reactions of bound arginine with glyoxal and furan-2-carboxyaldehyde. J. Agric. Food Chem., 46, 3896±901. Hofmann, T., Bors, W. & Stettmaier, K. (1999) Radical-assisted melanoidin formation during thermal processing of foods as well as under physiological conditions. J. Agric. Food Chem., 47, 391±6. Hughes, W.J. & Thorpe, T.M. (1987) Determination of organic acids and sucrose in roasted coffee by capillary gas chromatography. J. Food Sci., 52, 1078±83. ISO 11292 (1997) Instant coffee ± determination of free and total carbohydrates ± method using HPAEC. International Standards Organization, Geneva, Switzerland. Jardine, J.G. & Moretti, R.H. (1993) Enzyme treatment of coffee extract to reduce viscosity in instant coffee manufacture. Boletim da Sociedade Brasileira de Cienca e Technologia de Alimentos, 27, 14±23. Kundel, W., Kunz, M. & Martin, D. (1998) Determining oligosaccharide spectra in food products. Lebensmittelchemie, 52, 72± 73. Leloup, V., De Michieli, J.H. & Liardon, R. (1997) Characterization of oligosaccharides in coffee extracts. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 120±27. ASIC, Paris, France. Leloup, V. & Liardon, R. (1993) Analytical characterization of coffee carbohydrates. In: Proceedings of the 15th ASIC Colloquium (Montpellier), pp. 863±5. ASIC, Paris, France. Ludwig, H., Obermann H. & Spiteller, G. (1975) On new diterpenes found in coffee. In: Proceedings of the 7th ASIC Colloquium (Hamburg), pp. 205±10. ASIC, Paris, France. Luger, A. & Steinhart, H. (1995) Carbohydrates in steam treated coffee. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 366±71. ASIC, Paris, France. LuÈllman, C. & Silwar, R. (1989) Investigation of mono- and disaccharide content of Arabica and Robusta green coffee using HPLC. Lebensm. Gericht. Chem., 43, 42±3. MAFF (1995) Authenticity of soluble coffee. Food surveillance paper No. 46. Ministry of Agriculture, Fisheries and Food, London. HMSO, London. Maga, J.A., (1989) Thermal decomposition of carbohydrates. In: Thermal Generation of Aromas (eds T.H. Parliament, R.J. McGorrin & C.-T. Ho) pp. 32±9. ACS Symposium Series. ACS, Washington. Maier, H.G. & MaÈtzel, U. (1982) Atractyligenin and its glycosides in coffee. In: Proceedings of the 10th ASIC Colloquium (Salvador), pp. 247±51. ASIC, Paris, France. Maier, H.G. & Wewetzer, H. (1978) Determination of diterpeneglycosides in coffee beans. Z. Lebensm. Unters.-Forsch., 167, 105±107. Mazzafera, P. (1999) Chemical composition of defective coffee beans. Food Chem., 64, 547±54. Moller, A. & Steinhart, H. (1993) Studies of melanoidins in coffee beverages. Lebensmittelchemie, 47, 16±17. Navarini, L., Gilli, R., Gombac, A., Abatangelo, A., Bosco, M. & Toffanin, R. (1999) Polysaccharides from hot water extracts of roasted Coffea arabica beans. Isolation and characterization. Carbohydr. Polym., 40, 71±81. Nicolas, P., Raetz, E., Reymond, S. & Sauvageat, J.L. (1999)
Chemistry I: Non-volatile Compounds: Carbohydrates
Hydrolysis of coffee with immobilized beta-mannanase. European Patent Application EP 0676145. Noe, C.R., Lachmann, B., Mollenbeck, S. & Richter, P. (1999) Determination of reducing sugars in selected beverages by capillary electrophoresis. Z. Lebensm. Unters. Forsch. A., 208, 148±52. Noyes, R.M. & Chu, C.M. (1993) Material balance on free sugars in the production of instant coffee. In: Proceedings of the 15th ASIC Colloquium (Montpellier), pp. 577±82. ASIC, Paris, France. Noyes, R.M., McCarthy, J.P. & Oram, C.P. (1991) The variation of xylose, mannitol, and free sugar levels in instant coffee. In: Proceedings of the 14th ASIC Colloquium (San Francisco), pp. 202±10. ASIC, Paris, France. Nunes, F.M. & Coimbra, M.A. (1998) Influence of polysaccharide composition in foam stability of espresso coffee. Carbohyd. Polym., 37, 283±5. Nunes, F.M., Coimbra, M.A., Duarte, A.C. & Delgadillo, I. (1997) Foamability, foam stability, and chemical composition of espresso coffee as affected by the degree of roast. J. Agric. Food Chem., 45, 3238±43. Oestreich-Janzen, S. (1995) Carbohydrate profiles in beverages like coffee: methods and objects of investigations. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 286±91. ASIC, Paris, France. Petracco, M., Navarini, L., Abatangelo, A., Gombac, V., D'agnolo E. & Zanetti, F. (1999) Isolation and characterization of a foaming fraction from hot water extracts of roast and ground coffee. In: Proceedings of the 18th ASIC Colloquium (Helsinki), pp. 95±105. ASIC, Paris, France. Prescott, J.H., Enriquez, P., Jung C., Menz E. & Groman, E.V. (1995) Larch arabinogalactan for hepatic drug delivery: isolation and characterization of a kDa arabinogalactan fragment. Carbohyd. Res., 278, 113±28. Prodolliet, J., Blanc, M.B., Bruelhart, M., Obert, L. & Parchet J.M. (1992) Determination of soluble carbohydrates in coffee by high-performance anion-exchange chromatography with pulsed amperometric detection. In: Proceedings of the 14th ASIC Colloquium (San Francisco), pp. 211±19. ASIC, Paris, France. Prodolliet, J., Bruelhart, M., Blanc, M.B., et al. (1995a) Adulteration of soluble coffee with coffee husks and parchments. J. AOAC Int., 78, 761±7. Prodolliet, J., Bruelhart, M., Lador, F., et al. (1995b) Determination of free and total carbohydrate profile in soluble coffee. J. AOAC Int., 78, 749±61. Prodolliet, J., Bugner, E. & Feinberg, M. (1995c) Determination of carbohydrates in soluble coffee by anion-exchange chromatography with pulsed amperometric detection: interlaboratory study. J. AOAC Int., 78, 768±82.
17
Rao, C.V., Chou, D., Simi B., Herching, K. & Reddy, B.S. (1998) Prevention of colonic aberrant crypt foci and modulation of large bowel microbial activity by dietary coffee fiber, inulin and pectin. Carcinogenesis, 19, 1815±19. Silwar, R. & LuÈllman, C. (1988) The determination of mono- and disaccharides in green Arabica and Robusta coffees using high performance liquid chromatography. CafeÂ, Cacao, TheÁ, 32, 319±22. Stahl, H.D. & Parliment, T.H. (1993) Generation of furfuryl mercaptan in cysteine-ribose model systems in relation to roasted coffee aroma. In: Proceedings of the 15th ASIC Colloquium (Montpellier), pp. 607±15. ASIC, Paris, France. Stahl, H. & Turek, E. (1991) Acid hydrolysis of spent grounds to produce D-mannose and D-mannitol. In: Proceedings of the 14th ASIC Colloquium (San Francisco), pp. 339±48. ASIC, Paris, France. Steinhart, H. & Packert, A. (1993) Melanoidins in coffee. Separation and characterization by different chromatographic procedures. In: Proceedings of the 15th ASIC Colloquium (Montpellier), pp. 593±600. ASIC, Paris, France. Thaler, H. (1975) Structures of macromolecules in coffee. In: Proceedings of the 7th ASIC Colloquium (Hamburg), pp. 175± 87. ASIC, Paris, France. Tressl, R., Wondrak, G.T., Garbe, L.-A., Rewicki, D. & KruÈger, R.-P. (1998) Pentoses and hexoses as sources of new melanoidin-like Maillard polymers. J. Agric. Food Chem., 46, 1765± 76. Trugo, L.C. (1985) Carbohydrates. In: Coffee, Vol. 1, Chemistry (eds R.J. Clarke & R. Macrae) pp. 83±114. Elsevier Applied Science, London. Trugo, L.C. & Macrae, R. (1985) The use of the mass detector for sugar analysis of coffee products. In: Proceedings of the 11th ASIC Colloquium (LomeÂ), pp. 245±51. ASIC, Paris, France. Wilson, A.J., Petracco, M. & Lilly, E. (1997) Some preliminary investigations of oil biosynthesis in the coffee fruit and its subsequent re-distribution within green and roasted beans. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 92±9. ASIC, Paris, France. Wolfrom, M.L., Laver, M.L. & Patin, D.L. (1961) Carbohydrates of the coffee bean. II. Isolation and characterization of a mannan. J. Org. Chem., 26, 4533±35. Wolfrom, M.L. & Patin, D.L. (1964) Isolation and characterization of cellulose in the coffee bean. J. Agric. Food Chem., 12, 376±7. Wolfrom, M.L. & Patin, D.L. (1965) Carbohydrates of the coffee bean. IV. An arabinogalactan. J. Org. Chem., 30, 4060±63. Wolfrom, M.L., Plunkett, R.A. & Laver, M.L. (1960) Carbohydrates of the coffee bean. J. Agric. Food Chem., 8, 58±65. Zapp, J. & Kuhn, R. (1997) Maldi-MS, a new analytical technique and its potential for coffee analysis. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 141±9. ASIC, Paris, France.
1B
Acids in Coffee
H. H. Balzer Kraft Foods R and D Inc., Munich, Germany 1.7 QUANTITATIVE DATA ON ORGANIC ACIDS IN GREEN COFFEE
contains data obtained by capillary isotachophoresis for malic acid and citric acid for several green coffee origins. The data were reported by Scholze and Maier (1984). Analysis of the arabica coffees gave an average of 5.6 g/kg for malic acid and 12.3 g/kg for citric acid, while values for robusta coffees averaged 3.0 g/kg for malic acid and 8.6 g/kg for citric acid. Kenyan coffee had the highest concentration of malic acid among the different arabicas. Robusta levels were less than half that of Kenyan beans. After the chlorogenic acids, citric acid represents the next highest acid concentration in green beans. Levels of citric acid in Kenyan coffees tend to be lower than in Central American arabicas, which in view of the malic acid data suggests that, as these acids are part of the plant metabolism, ripening could be less advanced in the Kenyan coffee. The contents of phosphoric acid were also determined by isotachophoresis (Table 1.9) (Scholze & Maier, 1983). In general, arabica coffees contained less phosphoric acid (average 1.3 g/kg) than robusta varieties (average 1.7 g/kg). The silylation/GC method was further optimized by
Early data on acids in green coffee came from Marbrouk and Deatherage (1956); Lentner and Deatherage (1959) and Nakabayashi (1978a) including citric, malic, oxalic, tartaric, pyruvic and acetic acids. The first quantitative data for quinic acid in green coffee were given by Kampmann (1981) and Kampmann and Maier (1982) (Table 1.9). The acids were silylated and then separated from other coffee constituents by capillary gas chromatography (GC). Arabica green coffee contained 5.5 g/kg quinic acid on average and robusta 3.5 g/kg. Their contents were presumably also affected by factors such as age, processing and fermentation. Quinic acid has also been determined in steamed beans using the same technique (silylation/ GC) (Hucke & Maier, 1985). After steaming, the content of quinic acid was 15±40% higher compared to an untreated control. Besides chlorogenic and quinic acid, the major acids in green coffee are malic and citric acids. Table 1.9
Table 1.9 Contents (g/kg) of quinic, malic, citric and phosphoric acids of different green coffees. (Data from Kampmann & Maier; 1982; Scholze & Maier, 1983, 1984.) [7,8]. Provenance Santos Arabica Burundi Arabica Kenya Arabica Colombia Arabica Mocca Arabica Burundi Robusta Angola Robusta Togo Robusta Guinea Robusta Ivory Coast Arabusta Ivory Coast Excelsa Ivory Coast Stenophylla Liberica 1 2
Quinic acid1
Malic acid2
Citric acid2
Phosphoric acid2
5.6 5.7 4.7 5.5
6.15 5.12 6.62
13.81 13.0 11.65
1.07 1.11 1.37
4.60 3.78 2.78 2.47
10.55 10.01 9.17 6.70
1.47 1.42 2.16 1.63
4.53 5.22 2.92 3.90
9.09 10.90 10.02 10.63
1.73 1.46 2.19 1.76
3.5 3.1 3.9
GC analysis after silylation. Isotachophoresis analysis.
18
Chemistry I: Non-volatile Compounds: Acids in Coffee
19
tents of citric, malic, lactic, acetic and pyruvic acids were determined by titration. Woodman et al. (1967) extracted acids with butanol/chloroform, cleaned them through column chromatography and determined them as their methylesters by GC. Roasted coffee contained 9 g/kg citric acid, 5 g/kg acetic acid and 4 g/ kg lactic acid. Gas chromatography (GC), intra-red reflectance (IR) and mass spectrometry (MS) of esters were used to identify 12 acids; eight of them were reported for the first time. Itaconic, citraconic and mesaconic acids were identified as degradation products of citric acid. Maleic acid and fumaric acid were produced by malic acid degradation. Data for two roasted arabicas (Colombia and Santos) were reported by Feldman et al. (1969). Acids were determined by GC as their methylester and trimethyl silyl (TMS) derivatives. Organic acids have also been determined by liquid chromatography as their phenyl esters (Moll & Pictet, 1980); however, no quantitative data were given. Enzymatic tests were used by Blanc (1977). Carboxylic acid data were given for citric, malic, lactic, pyruvic and acetic acids in two arabica coffees during roasting (Table 1.12). It is necessary to know the degree of roast in order to compare data from different authors. Although not reported by Blanc, it is assumed that the degree of roast is given on an `as is' basis, which incorporates the moisture in the green coffee (typically 8 to 13%). This will be referred to as roast loss (RL). It is preferable to specify the percentage of dry matter loss, which can be given as roasting loss on a dry basis (organic roast loss, ORL). A comprehensive summary of the work carried out before 1985 was given by Woodman (1985). Scholze and Maier showed that the main acids of roasted coffee, including formic, malic, chlorogenic, citric, acetic,
Engelhardt and Maier (1984, 1985a), so that additional acids such as succinic acid, glycolic acid, lactic acid and phosphoric acid could be detected. When this technique is compared with isotachophoresis, the data of the former tended to be slightly lower. Organic acids in both arabica and robusta raw coffees were also determined by van der Stegen and van Duijin (1987) using HPLC with UV detection at 210 nm. The contents are given with their range and mean values in Table 1.10. More coffee acids were identified when free flow electrophoresis was used as a clean up and concentration step. This technique was applied by BaÈhre (1997) and BaÈhre and Maier (1999). After freeze drying, the acids were separated by silylation/GC and identified by mass spectroscopy (MS). Several minor acids, which occur in green coffee in low concentrations, could be determined by this technique (Table 1.11). Some acids were identified in coffee for the first time. The total amount of these minor acids was below 0.6 g/kg.
1.8 DETERMINATION OF ORGANIC ACIDS IN ROASTED COFFEE Marbrouk and Deatherage (1956) and Lentner and Deatherage (1959) established that the acidity of coffee depends upon the degree of roast. Clements and Deatherage (1957) used paper chromatography for the separation of caffeic, quinic and chlorogenic acids and column chromatography to separate citric, malic, tartaric, oxalic, pyruvic and acetic acids. Organic acids in malt and chicory coffee substitutes and mixtures of them with coffee were analyzed by SchormuÈller et al. (1961). Acids were isolated by distillation and separated by GC. After elution from the GC column, the con-
Table 1.10 Acid contents (g/kg) dry basis in different green arabica and robusta coffee samples analyzed by HPLC/UV. (Data from van der Stegen & van Duijn, 1987.) Arabica Acid Citric Malic Quinic Succinic Formic Acetic tr = trace amounts.
Robusta
Range of 8 samples
Mean
Range of 7 samples
Mean
5.0±14.9 2.6±6.7 3.3±6.1 tr±1.5 tr±1.4 tr
7.6 4.1 4.3 0.5 0.7 tr
3.3±10.1 1.8±7.3 1.6±8.6 0.5±3.5 tr±3.9 tr±2.0
6.5 3.4 3.6 1.1 1.3 0.5
20
Coffee: Recent Developments
Table 1.11 Acid contents (g/kg) in green Kenyan and Colombian arabica samples analyzed by GC/MS after silylation and pre-concentration with free flow electrophoresis. (Data from BaÈhre & Maier, 1996, 1999.) Kenyan Acid Arabonic Succinic 2-Oxopropionic Citraconic Erythro-3-desoxypentonic Threo-3-desoxypentonic 2,4-Dihydroxybutyric 3,4-Dihydroxybutyric Erythronic Fumaric 2-Furanoic Gluconic Glutaric Glycerinic 3-Hydroxybenzoic 4-Hydroxybenzoic 2-Hydroxybutyric 3-Hydroxybutyric 4-Hydroxybutyric Hydroxyacetic 2-Hydroxyglutaric 5-Hydroxymethyl-2-furanoic 3-Hydroxypropionic Itaconic Maleic Mannonic Mesaconic ribo,arabo-Metasaccharinic Methylsuccinic Lactic Nicotinic 2-Oxobutyric 2-Oxoglutaric 4-Oxovaleric Ribonic Shikimic Threonic Tartaric
Colombian
Unsteamed
Steamed
Unsteamed
Steamed
12 38 4 tr 13 3 tr 5 4 22 ND 67 2 7 8 7 tr 3 5 5 25 ND 4 4 6 24 3 ND 3 75 ND tr ND ND 12 tr 8 2
18 47 14 tr 12 4 1 29 20 15 2 51 4 20 10 7 tr 3 3 29 15 1 9 3 7 18 2 11 3 93 ND tr 15 4 14 10 14 4
10 31 4 ND 9 2 1 3 3 22 tr 50 2 7 ND ND ND 2 4 4 17 ND 2 1 ND 16 1 ND 2 42 ND ND ND tr 9 tr 4 2
19 49 18 ND 11 5 2 18 22 18 2 48 2 39 ND ND ND 2 2 50 12 3 15 ND ND 15 1 14 2 82 ND ND 12 tr 10 tr 8 3
ND = not detected, tr = trace.
glycolic and phosphoric acids, can be determined by capillary isotachophoresis (Scholze & Maier 1982, 1983, 1984; Scholze 1983). Hucke (1984) and Hucke and Maier (1985) used silylation/GC to determine quinic acid content: values of 8.7 to 16.6 g/kg were determined in 13 commercial roasted and ground (R&G) blends and 24.6 to 46.4 g/ kg in soluble coffee (SC). In roasted coffee, quinic acid
content increased proportionally together with its corresponding lactone (quinide) up to a maximum of 15% organic roast loss. Quinide could not be detected in 14 green coffees and in four steamed green coffees. Organic acids and sucrose were determined in roasted coffee by capillary gas chromatography, as reported by Hughes and Thorpe (1987). Gas chromatography was used to analyse 27 organic acids
Chemistry I: Non-volatile Compounds: Acids in Coffee
21
Table 1.12 Change in the contents of carboxylic acids (g/kg, dry basis) during roasting for Tanzanian arabica green coffee analyzed by enzymatic tests. (Data from Blanc, 1977).
RL (%) 11.6 12.15 13.45 13.80 15.10 17.25 17.75
Citric acid
Malic acid
Lactic acid
Pyruvic acid
Acetic acid
8.7 7.8 7.4 6.8 7.2 5.1 5.5
3.9 2.8 3.4 2.7 2.1 2.4 2.4
0.8 0.3 1.5 3.2 1.2 3.2 1.0
1.7 0.9 1.3 1.4 1.3 1.3 1.4
1.1 1.0 2.4 2.7 3.1 2.5 2.2
RL = roast loss.
Table 1.13 Acid contents (g/kg) of a commercial R&G coffee sample analyzed by GC after silylation. (Data from Engelhardt & Maier, 1985a.) Acid Malic Quinic Citric 2-Furanoic Maleic Phosphoric 3,4-Dihydroxybenzoic 2,5-Dihydroxybenzoic Formic Chlorogenic Acetic Glycolic Mesaconic Pyroglutamic Succinic Citraconic Fumaric Itaconic Lactic Tartaric
Content 2.16 7.93 6.72 0.14 0.09 1.78 0.08 0.23 2.01 21.4 5.15 1.32 0.08 0.87 0.07 0.54 0.12 0.16 0.88 0.04
by trimethylsilylation. Among these organic acids were pimelic, phthalic, suberic, sebacic, ascorbic, iso-ferulic, dodecandioic and sinapinic acids, which were described in coffee for the first time. The concentration of quinic acid ranged from 8.15 to 9.74 g/kg in caffeinated and decaffeinated coffees. A sample clean-up for nonvolatile acids by gel electrophoresis was described by Engelhardt (1984). After freeze drying, the acids were transferred into the trimethylsilyl derivatives, which were separated by GC and identified by MS. This method was used for five major and 14 minor acids.
The acid contents of a commercial roasted coffee determined by this methodology are given in Table 1.13 (Engelhardt & Maier, 1985b). Analysis of the main organic acids in coffee by a simple routine method was carried out by van der Stegen and van Duijin (1987) using HPLC/UV at 210 nm. The method was successfully applied to citric, malic, quinic, succinic, glycolic, lactic, formic, acetic and fumaric acids. As a limiting factor, phosphoric acid could not be analyzed by this method. Sample clean-up was performed with anion exchange resins. Acid contents for a commerical coffee sample are given in Table 1.14; however, the degree of roast was not reported. Succinic acid was determined for the first time in coffee. Table 1.14 Acid contents (g/kg) in commercial R&G coffee samples analyzed by HPLC/UV. (Data from van der Stegen and van Duijn, 1987.) Acid Citric acid Malic acid Quinic acid Succinic acid Formic acid Acetic acid Glycolic acid Lactic acid
Range of 17 samples 4.3±7.0 1.0±3.9 8.9±15.0 1.9±8.0 1.8±2.5 3.6±5.5 1.7±4.9 0.0±1.8
A recently developed technique, capillary electrophoresis (CE), offers an additional means to determine the major aliphatic organic acids. Capillary electrophoresis has a better separation ability than HPLCbased methods and the analysis is rapid. Its effectiveness has been shown for the analysis of organic acids in
22
Coffee: Recent Developments
various types of foods and beverages. For organic acids, a UV-absorbing buffer system is set at a pH at which the acids are essentially fully ionized. A CE procedure which incorporates this approach for the analysis of coffee was described by Weers et al. (1995). Tables 1.15 and 1.16 illustrate the contents of the major aliphatic (and phosphoric) acids in Colombian arabica and Indonesian robusta (EK I) beans as a function of organic roast loss (ORL). The sample range was from green to roasted at the limit of organoleptic acceptability. The principal acids in the green beans were malic, citric and quinic. As roasting progressed, the content of the first two decreased, whereas those of quinic and the other aliphatic acids increased. As mentioned before, yield of quinic acid increases with roasting due to the cleavage of chlorogenic acid (Maier, 1987). At higher roasts, the yields of formic acid and acetic acid began to fall off. The decrease of the former, because of its higher volatility, was initiated at a lower
ORL. Yields of the other aliphatic acids tended to level off at higher roasts. At even higher roasts (not studied here) the yields of volatile acids continued to fall off, whereas those of less volatile acids levelled off. A comparison of acid yields as a function of roast loss for the Colombian arabicas and Indonesian robustas (Tables 1.15 and 1.16, respectively) revealed some interesting differences. The content of phosphoric acid was significantly higher in robusta beans, which is in line with the results in Table 1.9. Franz and Maier (1994) found that four robusta samples all had higher total phosphorous and phytic acid contents than four arabica samples. The quinic acid content in robusta coffee rose faster during roasting, and therefore the initial differences observed in green beans between the species diminished. This observation reflects the higher chlorogenic acid content in robusta beans (Clifford, 1985). Apart from phosphoric acid and quinic acid, the yields of the other aliphatic acids
Table 1.15 Acid contents (g/kg) as a function of organic roast loss (ORL, %) for a Colombian arabica analyzed by capillary electrophoresis. (Data from Weers et al., 1995.) Acid Formic Malic Citric Succinic Glycolic Acetic Lactic Phosphoric Quinic
0.0 (Green)
3.3
4.2
5.1
6.4
ND 4.02 13.11 ND ND 0.29 ND 1.45 6.87
2.29 3.29 11.1 0.28 1.28 3.76 0.73 1.88 9.96
2.53 3.11 9.53 0.25 1.64 4.18 1.00 1.69 8.81
2.47 2.49 7.66 0.36 1.81 4.86 1.35 2.30 8.94
2.28 2.00 6.34 0.33 2.02 4.98 1.30 2.18 9.12
ND = not detected.
Table 1.16 Acid contents (g/kg) as a function of organic roast loss (ORL, %) for an Indonesian robusta analyzed by capillary electrophoresis. (Data from Weers et al., 1995.) Acid Formic Malic Citric Succinic Glycolic Acetic Lactic Phosphoric Quinic ND = not detected.
0.0 (Green)
3.25
4.15
5.0
6.5
0.18 2.47 13.50 0.13 ND 0.15 ND 2.79 4.70
0.86 2.40 13.38 0.23 0.47 1.98 ND 4.05 10.01
1.31 2.20 12.26 0.37 0.86 2.73 0.51 5.47 10.82
1.48 1.93 10.79 0.43 1.04 3.0 0.71 6.35 12.32
1.61 1.45 8.38 0.45 1.30 3.25 0.77 5.95 13.19
Chemistry I: Non-volatile Compounds: Acids in Coffee
in roasted samples tended to be higher in the Colombian beans. Barlianto (1990) and Barlianto and Maier (1994) identified and determined as many as 64 acids in coffee substitutes such as roasted chicory (48 new) and 60 (47 new) in roasted barley malt. Most of these were only present in trace amounts. BaÈhre and Maier confirmed the presence of many of these acids in coffee (BaÈhre, 1997; BaÈhre & Maier, 1996, 1999). These authors also described a convenient sample clean-up based on free flow electrophoresis. Using this method, organic acids could be separated from non-acidic and high molecular compounds in roasted and soluble coffees. Acid identification was achieved by GC/MS after freeze drying and trimethylsilylation. Thirty-eight acids have been identified and quantitated in coffee (Table 1.17). Of these, 18 acids were identified for the first time and five others were quantified for the first time in coffee, including 3-hydroxypropionic, 2-oxobutyric, glyceric, 2.4-dihydroxybutyric, 5-hydroxymethylfuran-2-carboxylic and 2-hydroxyglutaric acids. Table 1.17 contains average contents of six commercial blends of roasted coffee (including one steamtreated sample), one espresso blend and one soluble coffee. The corresponding values of the espresso blend were, in many cases, higher than the R & G mean value. This indicates a more extended acid formation by prolonged heating, but the contents of acids derived from carbohydrates were lower in many cases, indicating their destruction under these conditions. The contents of soluble coffee were, on average, 3.6 times higher than the mean contents of the R & G blends. The silylation/GC method described is suitable for most coffee acids and extremely valid for the determination of minor acids, when a pre-concentration/ clean-up step like free flow field electrophoresis is used. However, this method could not be used for volatile acids, especially formic acid and acetic acid, due to the high volatility of the silylated derivatives.
1.9 ACID FORMATION MECHANISMS As already indicated in Table 1.17, there are several different sources and precursors of acids found in coffee. Some aspects of acid formation will be discussed in the following.
1.9.1 Acetic, formic, lactic, glycolic and other carbohydrate derived acids Several publications have investigated the formation of acids from various carbohydrates in model systems.
23
Olsson et al. (1978) reported on the acid formation from carbohydrate precursors under non-enzymatic browning. Nakabayashi (1978) carried out model experiments with sucrose under roasting conditions. Formic, acetic and lactic acids were identified in the roasting mixture. Formation of acids, lactones and esters through the Maillard reaction has been investigated by Beck et al. (1990) through model roasting of a glucose/glycine mixture. Suggestions for acid formation mechanisms were made. Additional investigations made by Barlianto (1990) and Barlianto and Maier (1994), using model systems of fructose, maltose and inulin, have revealed that, during roasting, various acids were derived from these carbohydrates. Twenty-one compounds formed from a carbohydrate source are listed in Table 1.17. These acids, together with formic and acetic acid, seemed to be generated during roasting from carbohydrate precursors that are available in coffee. Their contents were negligible in green coffee. In a paper by Ginz et al. (2000) sucrose was confirmed as the principal green bean precursor of these acids. This correlates with the higher content of carbohydrate derived acids in arabica coffees (sucrose: 70 to 80 g/kg) compared to robusta coffees (sucrose: 30 to 40 g/kg). As an example, Colombian arabica contained 4.86 g/kg acetic acid and 2.47 g/kg formic acid at 5.1% ORL (Table 1.15) while Indonesian robusta contained only 3 g/kg acetic acid and 1.48 g/kg formic acid (Table 1.16). A major fraction of acidity generated on coffee roasting can be attributed to the formation of the four aliphatic acids: formic, acetic, glycolic and lactic. The roast kinetics of these principal acids are given in Fig. 1.9. Acetic and formic acids are formed upon roasting up to a maximum at 2408C roast temperature and decrease on further roasting. This can be explained by the high volatility of both compounds. Lactic and glycolic acids, by contrast, continue to increase from 2408C even up to 2808C. Arabinose, erythrose and 1,6-anhydroglucose were identified as intermediate reaction products of sucrose thermal degradation and also subsequently served as precursors for acid formation. Isotopic labelling experiments indicated that known degradation pathways could be used to explain the formation of the four aliphatic acids from sucrose.
1.9.2 Quinic acid As outlined above, quinic acid is already present in green coffee and its content is slightly increased during roasting. The precursors are chlorogenic acids (CGAs)
24
Coffee: Recent Developments
Table 1.17 Acid contents (mg/kg) in 6 commercial R&G coffees, espresso and soluble coffee analyzed by GC/ MS after silylation and pre-concentration with free flow step electrophoresis. (Data from BaÈhre & Maier, 1996, 1999; Bahre, 1997.) Commercial R&G blends Acids derived from
Mean
Lowest value
Highest value
Espresso
Soluble
838 188 15 16 13 11 21 10 473 91 51 33 85 109 194 85 126 117 12 25 19
540 157 10 10 ND 4 7 tr 330 60 41 8 53 70 93 64 84 62 6 5 16
1360 211 25 30 21 18 45 22 617 104 63 67 119 143 312 105 143 177 20 37 24
767 186 6 6 ND ND 8 3 548 57 99 16 257 283 357 116 91 47 7 52 36
1892 427 15 15 28 13 15 12 1443 230 215 65 508 617 589 328 236 342 50 167 139
8 7 21
3 ND 3
11 13 32
26 9 42
57 16 105
Citric acid Citraconic Glutaric Itaconic Mesaconic Succinic
272 13 145 40 61
233 10 119 29 53
332 15 197 54 67
325 20 182 116 87
798 52 456 269 270
Malic acid Fumaric Maleic
97 60
86 55
108 70
157 79
553 286
Trigonelline Nicotinic
46
4
82
114
255
Unknown precursors 2-Hydroxybutyric 3-Hydroxybutyric 4-Hydroxybutyric 2-Hydroxyglutaric 2-Oxoglutaric Tartaric
27 6 8 9 46 5
19 3 2 7 38 3
33 7 14 9 59 8
45 8 29 16 46 ND
112 15 50 33 96 ND
Carbohydrates Glycolic Glyceric Erythronic Threonic Arabonic Ribonic Gluconic Mannonic Lactic 3-Hydroxypropionic 2,4-Dihydroxybutyric 3,4-Dihydroxybutyric erythro-3-Deoxypentonic threo-3-Deoxypentonic Metasaccharinic 2-Furanoic 5-Hydroxymethyl-2-furanoic 2-Oxopropionic (pyruvic) 2-Oxobutyric 4-Oxovaleric Methyl succinic Chlorogenic acids 3-Hydroxybenzoic 4-Hydroxybenzoic Shikimic
ND = not detected, tr = trace.
Chemistry I: Non-volatile Compounds: Acids in Coffee
25
degradation products, citric acid yields mainly citraconic, glutaric, itaconic, mesaconic and succinic acids, while malic acid is responsible for the formation of fumaric and maleic acid (BaÈhre, 1997; BaÈhre & Maier, 1996, 1999).
1.9.4 Phosphoric acid
Fig. 1.9 Acid contents (g/kg) of formic, acetic, lactic and glycolic acid as a function of temperature for Indonesia robusta (EK I) analyzed by capillary electrophoresis. Coffee beans were roasted with a fluidized bed roaster for 3 minutes. (Data from Ginz et al., 2000.)
(Blanc, 1977). The quinic acid increase observed was lower than expected from CGA degradation. The kinetics were studied in detail by Leloup et al. (1995), revealing that under roasting conditions the major products are chlorogenic acid lactones and quinic acid lactones, rather than quinic acid. A comprehensive study of all quinic acid and quinic acid lactone (quinide) isomers was carried out by Scholz-BoÈttcher and Maier (1991). All possible isomers were generated under roasting conditions of coffee. Six isomers of quinic acid and seven of quinide were quantitated, (+)-quinic acid has the highest concentration in medium roasted coffee (ORL = 5 to 8%), being in the range of 6.63 to 9.47 g/kg. Scyllo-quinic acid is in the range 0.27 to 0.67 g/kg under the same conditions, while concentrations of other isomers like neo-quinic acid, meso-quinic acid I, meso-quinic acid II and (+)-epi-quinic acid were below 0.2 g/kg. The corresponding lactones were found in a concentration range of 2.29 to 8.49 g/kg for (+)-g-quinide and 0.17 to 0.59 g/kg for scyllo-d-quinide, while other isomers were below 0.3 g/kg (see also Chapter 2).
1.9.3 Citric and malic acid As shown above, these two acids are already present in green coffee beans in a range of concentrations according mainly to the botanical variety. Roasting leads to a continuous degradation; some degradation products are listed in Table 1.17. As well as other
Franz and Maier (1994) have shown that some of the phosphoric acid (phosphate) present in coffee extracts is produced by hydrolysis of phytic acid (IP6) and other inositol phosphates (IP5±IP1) during extraction and storage. In green coffee, mainly IP6 and a small quantity of IP5 were found. On roasting, IP6 was decomposed, while the amounts of other inositol phosphates increased at the beginning and decreased at higher roast degrees (Fig. 1.10).
Fig. 1.10 Contents (g/kg) of inositol phosphates in Salvador arabica coffee as a function of organic roast loss analyzed by HPLC. (Data from Franz & Maier, 1994.)
In addition to other phosphor-containing compounds, phytic acid and its derivatives, the inositol phosphates (IP5±IP1), play a major role in coffee. Phytic acid was analyzed in four soluble coffees and five varieties of coffee beans (green and roasted) (McKenzie, 1984). It was found that roasting coffee beans reduced their phytate content. All soluble coffee samples had a phytic acid content of at least 0.6 g/ 100 kg. The content in green coffee ranged from 2.2 g/ kg in Colombian, 2.3 g/kg in Brazilian Santos and 3.4 g/kg in Tanzanian to 3.5 g/kg in Papua New Guinea beans. Roasting of the beans resulted in a reduction in phytate content of 35% (Colombian) to 53% (Tanzanian). Inositol phosphates with six (IP6), five (IP5), four
26
(IP4) and three (IP3) phosphate groups have been determined by means of HPLC in several green and roasted coffees (Maier, 1993). The inositol phosphates were extracted with hot water or cold hydrochloric acid and isolated by anion exchange followed by purification on RP-18. HPLC employed an ion exchange column, a water/sodium nitrate/magnesium nitrate gradient, post column reaction with ferric perchlorate and detection at 310 nm (Franz & Maier, 1993). By this technique, several isomers of IP4 and especially IP5 could be resolved. Alternative methods for detection of phosphate, inositol phosphates and their corresponding isomers are based on capillary gas chromatography (March et al., 1996) and capillary electrophoresis (CE) combined with mass spectrometric determination (Maier et al., 1983).
1.10 ACID INCREASE ON STORAGE A slight increase in acidity can soon be perceived if a coffee brew is stored, even at room temperatures. This effect increases with time and significantly with rising temperature, as reported in 1963 by Sivetz. Many authors subsequently reported the same effect ± an increase of acidity during storage which could be detected by a decrease in the pH and a corresponding increase in titratable acidity (Cros et al., 1980; Walkowski, 1981). However, no explanation was put forward at that time. According to van der Stegen and van Duijn (1987), the quinic acid increase is the driving factor in acidity development, accounting for 25% of the figure. The increase in concentration of all acids, after heating for 24 hours at 958C, is given in Table 1.18. For comparison, the average content of commercial R&G coffees is also given. First assumptions that oxidation of simple aldehydes was responsible for the increase in acidity were ruled out early, because purging freshly brewed coffee with air did not increase acidity. The reason for the observed acid increase was discussed by Maier et al. (1984b): the hydrolysis of esters and quinic lactone, which were formed during roasting of coffee, seemed to be responsible. One indication was the fact that dark roasted coffee showed the biggest pH decrease. Acid analysis with isotachophoresis showed that all acid concentrations increased at a temperature > 608C. It was also found that the major fraction of acid precursors was water soluble while an additional 20% of acid precursors were obtained when coffee was heated together with coffee that had been extracted with water
Coffee: Recent Developments
Table 1.18 Acid contents (g/kg) of a commercial R & G coffee beverage and its increase after heating for 24 hours at 958C analyzed by HPLC/UV. (Data from van der Stegen & van Dujin, 1987.) Acid Citric Malic Quinic Succinic Formic Acetic Glycolic Lactic
Commercial coffee, mean of 17 samples
Acid increase
5.9 2.7 10.4 4.0 2.2 4.3 2.6 1.0
0.2 0.1 2.8 0.9 0.5 1.0 0.4 0.3
several times. For esters of other acids, the formation of monoesters with carbohydrates was suggested, a reaction that was shown to occur when acids and carbohydrates are heated together (Maier & Ochs, 1973). Both quinic acid and its corresponding lactone (quinide) can be analyzed by silylation/GC (Hucke & Maier, 1985). Their contents were determined in a range of different coffees. It was shown that the lactone content decreased significantly in a coffee brew which was heated for 48 hours at 808C, while the content of quinic acid increased by nearly the same amount. Additional investigations were carried out by Dalla Rosa et al. (1990), confirming the findings of Maier and colleagues. The acidity development described was accompanied by an increase in perceived sourness and thus had an undesirable effect on coffee quality. Bradbury et al. (1998) were able to inhibit the development of acidity on storage by using alkali. If the pH of a coffee brew was raised above 8, esters and lactones were hydrolyzed. Following neutralization, storage studies showed that the previously observed drop in pH no longer occurred.
1.11 VOLATILE ACIDS The term `volatile acids' describes a class of aliphatic compounds with a carboxylic functionality that can be isolated from coffee by simultaneous distillation and extraction (SDE) or by vacuum distillation. These acids can be detected by GC without further derivatization. Several of these acids also exhibit a contribution to the aroma profile of coffee; examples are the rancid/ sweaty smelling compounds such as 2- and/or 3-
Chemistry I: Non-volatile Compounds: Acids in Coffee
methylbutyric acid (Holscher et al., 1990). The most comprehensive determination of volatile acids in green and roasted coffee was reported by WoÈhrmann et al. (1997). With the exception of formic acid and acetic acid, the contents of volatile acids have been determined in green and roasted coffee by SDE, ion exchange and GC/FID detection. As peaks of 2- and 3-methylbutanoic acid could not be discriminated by normal GC, a chiral cyclodextrin phase was used for separation. The analysis of steam-treated coffee samples by the same methodology was described by SchroÈder et al. (1997). Contents of volatile acids were lower in steam-treated samples in most cases. 3-Methylbutanoic and (S)-2methylbutanoic acid could be used as indicators for steam treatment in unknown samples. The contents of volatile acids in a roast series are shown in Table 1.19 for an arabica coffee and Table 1.20 for a typical robusta coffee (WoÈhrmann, 1991). Coffees were roasted to an organic roast loss of between 0 and 9.97%. Some acids were identified for the first time in roasted coffee. Most of the contents increased with increasing degree of roast, but some of the higher saturated fatty acids did not. It was also shown that straight chain fatty acids C5 to C10 were formed by hydrolysis of fatty acids.
27
1.12 ACID CONTENT AND SENSORY CHARACTERISTICS 1.12.1 Total acidity and sour taste The acidity or sourness of coffee brews has (together with aroma and bitterness) always been recognized as an important attribute of their sensory quality. In general, the acids present in coffee are responsible for about 11% of the green beans' weight and for 6% of roasted coffee beans' weight (Maier, 1987). Especially in high quality beans (arabica), roasted to light or medium roast degrees, a major taste is sourness (Clifford, 1989). Kenyan coffee beans are well known for their well developed acidic character which is often described as `fine acidity' (Vitzthum, 1976). Dark roasted arabicas, on the other hand, characteristically show less acidity, so that bitterness becomes the dominating taste. Wet processed (`washed') coffees are higher in acidity than unwashed, dry processed varieties (i.e. Brazilian arabicas or robusta coffees). Robusta coffees have the lowest acidity. Thus, there was an early interest in characterizing acids in roasted coffee. In arabica coffee varieties, the pH of the brew is between 4.85 and 5.15. They are more sour than robustas, which typically have a brew pH in the range
Table 1.19 Contents (mg/kg) of volatile acids in roasted arabica coffee as a function of organic roast loss (ORL, %). Coffee extracts were obtained by simultaneous distillation extraction and then analyzed by GC/MS. (Data from WoÈhrmann, 1991.) Acid Propanoic 2-Methylpropanoic Butanoic 2-/3-Methylbutanoic 2-Methyl-propenic Pentanoic trans-2-Butenic 3-Methylpentanoic 3,3-Dimethylpropenic Hexanoic trans-2-Methyl-2-butenic 2-Ethylhexanoic Heptanoic Octanoic Nonanoic Decanoic Dodecanoic Tetradodecanoic Pentadodecanoic ND = not detected.
0 (Green)
3.68
6.02
8.17
9.97
ND 0.43 0.32 38.27 ND 0.31 ND 0.32 5.68 0.83 0.38 ND 0.31 0.50 0.76 0.20 0.22 2.95 0.70
2.93 0.69 1.55 65.18 0.15 0.45 0.44 0.64 15.84 1.11 0.95 0.06 0.32 0.33 0.61 0.38 0.31 3.31 0.84
2.02 0.28 4.02 70.20 0.89 0.44 0.98 0.82 15.45 0.99 1.37 0.16 0.45 0.48 0.47 0.34 0.31 3.32 0.72
3.45 0.47 4.49 78.01 1.28 0.51 0.93 0.97 14.24 0.71 1.59 0.25 0.36 0.24 0.85 0.36 0.31 4.74 0.95
1.71 0.58 5.0 74.48 1.03 0.57 0.68 0.99 10.65 0.88 2.22 0.48 0.46 0.45 0.53 0.47 0.40 6.24 1.24
28
Coffee: Recent Developments
Table 1.20 Contents (mg/kg) of volatile acids in roasted robusta coffee as a function of organic roast loss (ORL, %). Coffee extracts were obtained by simultaneous distillation extraction and then analyzed by GC/MS. (Data from WoÈhrmann et al., 1997.) Acid Propanoic 2-Methylpropanoic Butanoic 2-/3-Methylbutanoic 2-Methylpropenic Pentanoic trans 2-Butenic 3-Methylpentanoic 3,3-Dimethylpropenic Hexanoic trans-2-Methyl-2-butenic 2-Ethylhexanoic Heptanoic Octanoic Nonanoic Decanoic Dodecanoic Tetradodecanoic Pentadodecanoic
0 (Green)
4.41
5.76
7.50
8.57
9.66
ND 0.55 0.43 31.89 ND 0.55 ND 0.62 7.77 3.17 0.37 ND 0.34 0.50 0.69 0.20 0.32 3.07 1.64
3.87 0.89 1.41 63.67 0.68 0.56 2.37 1.12 18.38 3.97 0.66 0.11 0.38 0.37 0.15 0.51 0.36 2.53 1.05
2.78 0.83 1.79 72.52 0.43 0.59 2.25 1.14 15.79 3.62 0.96 0.12 0.49 0.48 0.44 0.53 0.39 2.18 0.77
1.56 0.35 2.86 80.10 0.85 0.73 2.52 1.46 14.81 3.70 1.21 0.12 0.83 0.85 0.63 0.63 0.41 2.37 0.82
2.80 0.63 3.23 70.11 0.89 0.68 2.02 1.48 12.62 3.03 1.49 0.16 0.83 0.78 0.46 0.67 0.44 2.49 0.74
1.33 0.45 3.89 64.99 0.88 0.73 1.63 1.43 10.02 2.46 1.57 0.15 0.94 0.91 0.47 0.83 0.36 2.69 0.78
ND = not detected.
of 5.25 to 5.40. The roasting conditions and the bean type are important, but green bean processing and age also influence pH (Werner & Kohley, 1965a,c,d). The relationship between coffee acidity and sour taste has been studied intensively. Although there is no doubt that hydrogen ion concentration is associated with perceived acidity, many studies have shown that there is only moderate correlation between sour taste and pH. Total acidity of a coffee brew, which is expressed by titratable acidity (TA), has been demonstrated to show better correlation to sourness than pH (BaÈhre & Maier, 1996). However, there was some debate about the best end point for pH titration. Ciurea and VoÈsgen (1985) and Wurziger and Drews (1983a,b) suggested titration to pH 7.0. The former calculated that, at pH 7, at least 99% of the coffee acids are significant in the dissociated form. For citric, malic and other diprotic acids only the first dissociation step was considered. According to the titration curve of individual acids, a titration of phenolic protons starts at pH values greater than 8. A correlation of acid taste with pH and titratable acidity was made by Maier et al. (1983) using 26 brews of different roasted coffees and 36 solutions of different soluble coffees. A poor correlation was found with pH value, with a correlation coefficient (r2 ) of a linear regression of 0.53. A fairly good correlation existed
between acid taste and TA titrated to pH 8 (r2 = 0.85). The correlation was even better with TA titrated to pH 7 (r2 = 0.89) and best if titrated to pH 6 (r2 = 0.92). The correlation would even be better if the coffees were of equal taste. The deviations of the points from the regression line are partly due to the more or less bitter taste (Maier et al., 1984a). Maier et al. (1983) suggested that titration to an end-pH of 6 gave the best correlation, because human saliva around the tongue has a pH in the region of 6. Thus, sour taste could be related to neutralization of acids in the mouth. Much research by other groups has been directed towards defining sourness and the mechanism by which it is perceived by the receptors in the mouth (Noble et al., 1986; Ganzevles & Kroeze, 1987). Studies on the sourness of different acids indicate that free protons (as represented by pH) contribute to sourness as well as bound protons, probably by their release at the receptor site. Both free and bound protons are represented by TA (Shallenberger, 1996). These findings are in agreement with the observed correlation of sour taste and TA.
1.12.2 Acid content and acidity Acids found in coffee are primarily aliphatic carboxylic acids, including chlorogenic acids which are reported
Chemistry I: Non-volatile Compounds: Acids in Coffee
in another section (see Chapter 2). Phosphoric acid, phytic acid and the other phosphorylated inositols together with alicyclic and heterocyclic acids are also considered. The parameter that governs the contribution of a particular acid to pH and titratable acidity is the pKa value. The lower the pKa value, the stronger the acid, i.e. the lower the pH at which the molecule dissociates to give protons and anions in solution. The pKa values of the principal coffee acids are shown in Table 1.21. Typical concentrations of the different acids in a brew were published by Clifford (1989). Citric acid, phosphoric acid (and phytic acid), quinic acid, chlorogenic acid and malic acid are the most acidic. For citric, malic and phosphoric acids, more than one proton has to be considered. On the basis of the quantitative acid data in coffee, it can be calculated which of the individual acids are mostly responsible for the acid taste. Because titratable acidity (an end-pH of 6) is affected by the consumption of hydroxide between the pH of the coffee brew and pH 6, one can assume that the order of alkali consumption by the individual acids is also the order of their contribution to acid taste. The consumption of hydroxide by each individual acid can be calculated as the corresponding dissociation of acid groups between brew pH (ca 4.9) and mouth pH (ca 6). These calculations need only the concentration of the individual acid, the dissociation constants and the pH values of the range, i.e. the pH of the coffee brew to pH 6 (Engelhardt & Maier, 1985b). The 22 acids determined
29
by Engelhardt and Maier (1985a) contribute 67% (roast coffee, R & G) and 72% (soluble coffee, SC) to the total titratable acidity with an end-pH of 6. In the first place, citric acid (12.2% R & G/10.7% SC), acetic acid (11.2%/8.8%) and the high molecular weight acids (8%/9%) contribute to the total acidity. Chlorogenic acid (9%/4.8%), formic acid (5.3%/ 4.6%), quinic acid (4.7%/5.9%), malic acid (3.9%/ 3%) and phosphoric acid (2.5%/5.2%) contributions should also be mentioned. However, other acids also need to be considered in terms of acidity. Although most of the acids in coffee extracts are present as anions, which are not perceived as sour, they already contribute to acidity by providing protons. Thus, an increase in the content of any of the coffee acids leads to a lowering of pH and an increase in titratable acidity. The relationship between beverage pH, titratable acidity, buffer capacity of mineral content and organic and phosphoric acids was discussed in detail by Clifford (1989). A model was proposed which could be used to explain the variations associated with different types of green beans and the variations introduced by different degrees of roast.
1.12.3 Roast kinetics Acidity in coffee is generated during roasting, as indicated by a drop in extract pH from 5.7±6.0 to 4.9± 5.5 (Werner & Kohley, 1965b). The main acidity is
Table 1.21 Published data for pKa values of typical coffee acids and acid content of a Colombian arabica R & G coffee brew (mmol/100 ml). (Data from Clifford, 1989.) Acid Chlorogenic Citric Quinic Phosphoric Formic Acetic Malic Glycolic Lactic Pyroglutamic
Proton
1 2 3 1 2 3 1 2
pKa
Typical content (mmol/100 ml)
3.4 3.14 4.77 6.39 3.4 1.96 7.21 12.30 3.75 4.73 3.4 5.05 3.83 3.89 3.32
96±291 75±189 123±242 65±108 130±159 74±226 58±76 51±100 22 27
30
Fig. 1.11 Development of titratable acidity as a function of organic roast loss for a Colombian arabica and an Indonesian robusta (EK I). (Balzer, 1999, unpublished data.)
generated at the beginning of the roasting process (Fig. 1.11). A fast increase in TA on roasting was observed from green to 4.2% ORL, followed by a smaller decrease as roasting proceeded (Balzer, unpublished data).
REFERENCES BaÈhre, F. & Maier, H.G. (1996) Electrophoretic clean-up of organic acids from coffee for the GC/MS analysis. Fresenius J. Anal. Chem., 355, 190±93. BaÈhre, F. (1997) Neue nichtfluÈchtige SaÈuren im Kaffee. Dissertation, Technical University of Braunschweig, Germany. BaÈhre, F. & Maier, H.G. (1999) New non-volatile acids in coffee. Dtsche Lebensm.-Rundsch., 95, 399±402. Barlianto, H. (1990) SaÈuren in Zichorie und Malz. Dissertation, Technical University of Braunschweig. Barlianto, H. & Maier, H.G. (1994) Acids in chicory roots and malt. Lebensm. Unters.-Forsch., 198, 215±22. Beck, J., Ledl, F., Sengl, M. & Severin T. (1990) Formation of acids, lactones and esters through the Maillard reaction. Z. Lebensm. Unters.-Forsch., 190, 212±16. Blanc, M. (1977) Les acides carboxyliques du cafeÂ. In: Proceedings of the 8th ASIC Colloquium, pp. 73±8. ASIC, Paris, France. Bradbury, A.G.W., Balzer, H.H. & Vitzthum, O.G. (1998) Stabilization of liquid coffee by treatment with alkali. US Patent Application No. 98300217.1±2114. Buscher, B.A.P., van der Hoeven, R.A.M. Tjaden, U.R., Andersson, E. & van der Greef, J. (1995) Analysis of inositol phosphates and derivatives using capillary zone electrophoresis-mass spectrometry. J. Chromatogr. A712, 235±43. Ciurea, I.C. & VoÈsgen, W. (1985) Uber die elektrometrische SaÈuregrad-Bestimmung von Kaffee-Extrakten. In: Proceedings
Coffee: Recent Developments
of the 11th ASIC Colloquium (LomeÂ), pp. 197±203. ASIC, Paris, France. Clements, R.L. & Deatherage, F.E. (1957) A chromatographic study of some of the compounds in roasted coffee. Food Res., 22, 222±32. Clifford, M.N. (1985) Chlorogenic acids. In: Coffee, Vol. 1, Chemistry (eds R.J. Clarke & R. Macrae). pp. 153±202, Elsevier Applied Science, London. Clifford, M. (1989) What factors determine the intensity of coffee's sensory attributes. Tea Coffee Trade J., 8, 35±9. Cros, E., Fourny, G. Guyot, B., Rouly, M. & Vincent, J.C. (1980) Changes in roasted Arabica coffee stored in four model packagings. Changes in the volatile fraction. Comparison with a control. CafeÂ, Cacao, TheÂ, 24, 203±25. Dalla Rosa, M., Barbanti, D. & Lerici, C.R. (1990) Changes in coffee brews in relation to storage temperature. J. Sci. Food Agric, 50, 227±35. Engelhardt U.H. (1984) NichtfluÈchtige SaÈuren im Kaffee. Dissertation, Technical University of Braunschweig. Engelhardt, U.H., Maier, H.G. (1985a) SaÈuren des Kaffees. XI. Anteil einzelner SaÈuren an der titrierbaren GesamtsaÈure. Z. Lebensm. Unters.-Forsch., 181, 20±23. Engelhardt, U.H. & Maier, H.G. (1985b) SaÈuren des Kaffees. XII. Anteil einzelner SaÈuren am sauren titrierbaren GesamtsaÈure. Z. Lebensm. Unters.-Forsch., 181, 206±9. Feldman, J.R., Ryder, W.S. & Kung J.T. (1969) Importance of nonvolatile compounds to the flavor of coffee. J. Agric. Food Chem., 17, 733±9. Franz, H. & Maier, H.G. (1993) Inositolphosphate in Kaffee und Kaffeemitteln. I. Identifizierung und Bestimmungsmethode. Dtsche Lebensm.-Rundsch., 89, 276±82. Franz, H. & Maier, H.G. (1994) Inositolphosphate in Kaffee und Kaffeemitteln. II. Bohnenkaffee. Dtsche Lebensm.-Rundsch, 90, 345±9. Ganzevles, P.G.L. & Kroeze, J.H.A. (1987) The sour taste of acids. The hydrogen ion and the undissociated acid as sour agents. Chem. Senses, 12, 563±75. Ginz, M., Balzer, H.H., Bradbury, A.G.W. & Maier H.G. (2000) Formation of aliphatic acids by carbohydrate degradation during roasting of coffee. Eur. Food Res. Technol., 211, 404± 410. Holscher, W., Vitzthum, O.G. & Steinhart, H. (1990) Identification and sensorial evaluation of aroma-impact-compounds in roasted colombian coffee. CafeÂ, Cacao, TheÂ, 34, 205±12. Hucke, J. (1984) ChinasaÈurelacton im Kaffee. Dissertation, Technical University of Braunschweig. Hucke, J. & Maier, H.G. (1985) ChinasaÈurelacton im Kaffee. Z. Lebensm. Unters.-Forsch., 180, 479±84. Hughes, W.J. & Thorpe, T.M. (1987) Determination of organic acids and sucrose in roasted coffee by capillary gas chromatography. J. Food Sci., 52, 1078±83. È ber ChinasaÈure im Kaffee. Dissertation, Kampmann, B. (1981) U Technical University of Braunschweig. Kampmann, B. & Maier, H.G. (1982) SaÈuren des Kaffees. I. ChinasaÈure. Z. Lebensm. Unters.-Forsch., 175, 333±6. Leloup, V., Louvrier, A. & Liardon, R. (1995) Degradation mechanisms of chlorogenic acids during roasting. In: Proceed-
Chemistry I: Non-volatile Compounds: Acids in Coffee
ings of the 16th ASIC Colloquium (Kyoto), pp. 192±8. ASIC, Paris, France. Lentner, C. & Deatherage F.E. (1959) Organic acids in coffee in relation to the degree of roast. Food Res., 24, 483±92. McKenzie, J.M. (1984) Content of phytate and minerals in instant coffee, coffee beans and coffee beverages. Nutr. Repts. Int., 29, 387±95. Maier, H.G. (1987) The acids of coffee. In: Proceedings of the 12th ASIC Colloquium (Montreux), pp. 229±37. ASIC, Paris, France. Maier, H.G. (1993) Status of research in the field of non-volatile coffee components. In: Proceedings of the 15th ASIC Colloquium (Montpellier), pp. 567±76. ASIC, Paris, France. Maier, H.G., Balcke, C. & Thies, F.-C. (1983) Die SaÈuren des Kaffees. VI. AbhaÈngigkeit des sauren Geschmacks von pHWert und SuÈregrad. Lebensm. Gerichtl. Chem., 37, 81±3. Maier, H.G., Balcke, C. & Thies F.-C. (1984a) SaÈuren des Kaffees. X. Einfluû des bitteren Geschmacks auf den sauren. Dtsche Lebensm.-Rundsch., 80, 367±9. Maier, H.G., Engelhardt, U.H. & Scholze, A. (1984) SaÈuren des Kaffees. IX. Zunahme beim Warmhalten des GetraÈnks. Dtsche Lebensm.-Rundsch., 80, 265±8. Maier, H.G. & Engelhardt U.H. (1985) Determination of nonvolatile acids in coffee: comparison of capillary isotachophoresis and capillary gas chromatography. Fresenius Z. Anal. Chem., 320, 169±74. Maier, H.G. & Ochs, H. (1973) Bildung von Estern aus GenubsaÈuren und Zuckern bzw. Polyalkoholen. SuÈsswaren, 18, 925±8. Marbrouk, A.F. & Deatherage, F.E. (1956) Organic acids in brewed coffee. Food Tech, 10, 194±7. March, J.G., Forteza, R. & Grases F. (1996) Determination of inositol isomers and arabitol in human urine by gas chromatography-mass spectrometry. Chromatographia, 42, 329±31. Moll, H.R. & Pictet G.A. (1980) La chromatographie liquide haute performance appliqueÂe aÁ certains constituants speÂcifiques du cafeÂ. In: Proceedings of the 9th ASIC Colloquium (London), pp. 87±98. ASIC, Paris, France. Nakabayashi, T. (1978a) Chemical studies on the quality of coffee. VI. Changes in organic acids and pH of roasted coffee. (Nippon Shok. Kogyo Gak., 25, 142±6.) Jap. Soc. Food Sci. Technol., 25, 142±6. Nakabayashi, T. (1978b) Chemical studies on the quality of coffee. VII. Formation of organic acids from sucrose by roasting. (Nippon Shok. Kogyo Gak., 25, 257±61. Jap. Soc. Food Sci. Technol., 25, 257±61. Noble, A.C., Philbrick, K.C. & Boulton, R.B. (1986) Comparison of sourness of organic acid anions at equal pH and equal titratable acidity. J. Sens. Stud., 1, 1±8. Olsson, K., Pernemalm, P.A. & Theander, O. (1978) Formation of aromatic compounds from carbohydrates. VII. Reaction of D-Glucose in slightly acidic, aqueous solution. Acta. Chem. Scand. B, 32, 249±56. Scholz-BoÈttcher, B.M. & Maier, H.G. (1991) Isomers of quinic acid and quinides in roasted coffee: indicators for the degree of roast. In: Proceeding of the 14th ASIC Colloquium (San Francisco), pp. 220±29. ASIC, Paris, France. Scholze, A. (1983) Quantitative Bestimmung von SaÈuren in Kaffee
31
durch Kapillar-Isotachophorese. Dissertation, Technical University of Braunschweig. Scholze, A. & Maier, H. G. (1982) Quantitative Bestimmung von SaÈuren in Kaffee mittels Kapillar-Isotachophorese. Lebensm. Gerichtl. Chem., 36, 111±12. Scholze, A. & Maier, H.G. (1983) Die SaÈuren des Kaffees. VII. Ameisen, AÈpfel-, Citronen- und EssigsaÈure. Kaffee Tee Markt 33, (22) 3±6. Scholze, A. & Maier H.G. (1984) SaÈuren des Kaffees. VIII. Glykol- und PhosphorsaÈure. Z. Lebensm. Unters.-Forsch. 178, 5±8. SchormuÈller, J., Brandenburg, W. & Langner H. (1961) Organische SaÈuren in Kaffee-Ersatzstoffen sowie in TrockenExtraktpulvern aus Kaffee-Ersatzstoffen und Kaffee. Z. Lebensm. Unters.-Forsch., 115, 226±35. SchroÈder, I., Stern, G., Hojabr-Kalali, B., Schliekelmann, K. & Maier, H.G. (1997) Volatile minor acids in coffee. II. Detection of steam treatment, Dtsche Lebensm.-Rundsch. 93, 216±18. Shallenberger, R.S. (1996) The AH,B glycophore and general taste chemistry. Food Chem., 56, 209±14. Sivetz, M. (1963) Coffee Processing Technology. Vol. II. AVI, Westport. van der Stegen, G.H.D. & Duijn, J. (1987) Analysis of normal organic acids in coffee. In: Proceedings of the 12th ASIC Colloquium (Montreux) pp. 238±46. ASIC, Paris, France. Vitzthum, O.G. (1976) Chemie und Bearbeitung des Kaffees. In: Kaffee und Coffein (ed. O. Eichler.) pp. 3±64. Springer-Verlag, Berlin. Walkowski, A. (1981) Changes in factors determining coffee bean quality during storage. Lebensm. Ind., 28, 75±6. Weers, M., Balzer, H., Bradbury, A. & Vitzthum, O.G. (1995) Analysis of acids in coffee by capillary electrophoresis. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 218±23. ASIC, Paris, France. Werner, H. & Kohley, M. (1965a) Untersuchungen uÈber den SaÈuregehalt von Roh- und RoÈstkaffee verschiedener Herkunft. Kaffee Tee Markt 15(2), 6±9. Werner, H. & Kohley, M. (1965b) Untersuchungen uÈber den SaÈuregehalt von Roh- und RoÈstkaffee verschiedener Herkunft. Kaffee Tee Markt 15(3), 6±12. Werner, H. & Kohley, M. (1965c) Untersuchungen uÈber den SaÈuregehalt von Roh- und RoÈstkaffee verschiedener Herkunft. Kaffee Tee Markt 15(4), 6±10. Werner, H. & Kohley, M. (1965d) Untersuchungen uÈber den SaÈuregehalt von Roh- und RoÈstkaffee verschiedener Herkunft. Kaffee Tee Markt 15(5), 5±10. WoÈhrmann, R. (1991) Gehalte FluÈchtiger SaÈuren in Kaffee, Malz u. Zichorie. Dissertation, Technical University of Braunschweig. WoÈhrmann, R., Hojabr-Kalali, B. & Maier, H.G. (1997) Volatile minor acids in coffee. I. Contents of green and roasted coffee. Dtsche Lebensm.-Rundsch, 93, 191±4. Woodman, J.S., Giddey, A. & Egli, R.H. (1967) The carboxylic acids of brewed coffee. In: Proceedings of the 3rd ASIC Colloquicion (Trieste), pp. 137±43. ASIC, Paris, France. Woodman, J.S. (1985) Carboxylic acids. In: Coffee, Vol. 1,
32
Chemistry (eds R.J. Clarke & R. Macrae) pp. 266±89 Elsevier Applied Science, London and New York. Wurziger, J. & Drews, R. (1983a) Zur lebensmittelrechtlichen Beurteilung von GetraÈnken aus RoÈstkaffee, Teil I. Kaffee Tee Markt 33(15), 3±9.
Coffee: Recent Developments
Wurziger, J. & Drews, R. (1983b) Zur lebensmittelrechtlichen Beurteilung von GetraÈnken aus RoÈstkaffee, Teil II. Kaffee Tee Markt, 33(16), 3±8.
1C
Lipids
K. Speer and I. KoÈlling-Speer Institute of Food Chemistry, Technical University Dresden, Germany 1.13 INTRODUCTION
1.14 COFFEE OIL
Since the comprehensive review of the lipids present in green coffee, and also in roasted coffee products, by Folstar (1985), there have been a considerable number of further investigations, into each of the individual components in these lipids. The lipid content of green arabica coffee averages some 15% on a dry basis, whilst robusta contains much less, around 10%. Most of the lipids, the coffee oil, are located in the endosperm of the coffee bean (Wilson et al., 1997), while only a small amount, the coffee wax, is in the outer layer. Maier (1981a) presented data from Kaufman's work (1962±4) which showed some 75% triglycerides in 100% green arabica coffee oil, with therefore a high percentage of unsaponifiables, including about 19% total free and esterified diterpene alcohols, about 5% total free and esterified sterols, and the remainder, very small quantities of other substances, such as tocopherols; all are considered separately in the following sections. These particularly often quoted data were, however, based upon only a 7% solvent extraction. A number of new components of analytical importance have since been found, notably 16-O-methylcafestol.
1.14.1 Determination of total oil content
Table 1.22 al., 1966.)
The percentage yield of crude lipid has long been known to be a function, not only of the nature of the bean, but also of the conditions of extraction, particular particle size/range and surface area, choice of solvent and duration of the extraction (Folstar, 1985). One standard method is that given by the Association of Official Analytical Chemists (AOAC, 1965). The Soxhlet extraction is carried out over 16 hours with a 35±508C boiling range of petroleum ether. In the method of the German Society for Lipid Science (DGF), published in 1952, which is similar to the method of the International Union of Pure and Applied Chemistry, published in 1966, the material is ground, then dried at 1058C for 30±35 minutes (if the moisture content exceeds 10%), and extracted for 4 hours with petroleum ether (40±558C boiling range). Streuli et al. (1966) and Streuli (1970) treated the ground green coffee with acid prior to extraction; this method became an official Swiss method. Streuli et al. (1966) compared the three above mentioned methods, with the results as shown in Table 1.22, in which the Swiss method generally gave the highest results.
Percentage of coffee oil in dry matter determined by three different methods (data from Streuli et Swiss official method (1973)
AOAC1 method (1965)
DGF2 method (1952)
Green coffee Congo Santos Madagascar
10.3 15.9 9.2
7.8 13.6 7.6
8.4 14.1 10.5
Roasted coffee Congo Santos Madagascar
11.4 16.3 10.6
10.9 15.1 9.7
11.9 16.4 12.0
1 2
Association of Official Analytical Chemists. Deutsche Cresellschaft fuÈr Fettwissenschaft.
33
34
Coffee: Recent Developments
Because of the great differences in yield, the term `coffee oil' needs to be explicitly defined. Values for the content of oil in roasted coffee are higher compared with those in the green coffees from which they were derived, due to the overall dry matter content loss on roasting, which varies with the level of roast. The loss of actual lipid matter is, however, low (Vitzthum, 1976).
1.14.2 Isolation of coffee oil for detailed analysis In order to obtain a coffee oil to study its chemical composition in detail, direct solvent extraction without acid treatment is necessary. According to Picard et al. (1984), different authors have used variously diethyl ether, petroleum ether with different boiling point ranges, n-hexane and a mixture of diethyl ether and nhexane. The results are not comparable because variable amounts of other more polar and non-lipid substances, such as caffeine, were extracted, according to the solvent used. Picard et al. (1984) observed that, with increasing extraction time, the oil extracted from a robusta coffee with hexane/diethylether increased for 6 and 8 hours (11.4% and 11.6%, respectively) and then slightly decreased for 10 and 12 hours (11.0% and 10.9%, respectively). Furthermore, Folstar et al. (1975) demonstrated that the yield obtainable on solvent extraction depends on the particle size to which the coffee is finally ground (Table 1.23). Speer (1989) extracted ground coffee of a particle size smaller than 0.63 mm, selecting tertiary butyl methyl ether as a safer extraction solvent than diethyl ether. His method was adopted as a part of the DIN (German Standards Institute) method 10779 (DIN, 1999) and the method is as follows. Roasted coffee beans are coarsely ground in a regular coffee mill and passed through a 0.63 mm sieve, 5 g of the sieved material is then powdered together with sodium sulfate in a mortar and extracted with tertiary butyl methyl ether in a Soxhlet (4 hours), siphoning six or seven
times per hour. The solvent is evaporated and the residue is dried to constant weight (1058C). Longer extraction times (6, 8 or 10 hours) do not increase the lipid content. For green coffee beans, grinding in the mill is prepared together with dry ice.
1.15 FATTY ACIDS 1.15.1 Total fatty acids and fatty acids in triglycerides For the most part, the fatty acids are to be found in the combined state; most are esterified with glycerol in the triglycerides, some 20% are esterified with diterpenes and a small proportion is to be found in the sterol esters. The total fatty acid composition of coffee oil has been the subject of many investigations. Table 1.24 is a review of the results obtained for green coffees. According to Maier (1981a), review 1 in the table summarizes the data from Pokorny and Forman (1970), Roffi et al. (1971), Streuli (1970), Vitzthum (1976), Wurziger (1963). Lercker et al. (1996) tabulated the data from Calzolari and Cerma (1963), Carisano and Gariboldi (1964), Hartmann et al. (1968), Chassevent et al. (1974) according to Folstar (1985) in review 2. Folstar et al. (1975) and Speer et al. (1993) investigated the fatty acids in detail. They analysed the fatty acids in the triglycerides of coffee beans and in the diterpene esters. The fatty acids in sterol esters were determined by Picard et al. (1984). For separating the different lipid classes Folstar et al. used a Florisil column. Speer et al. isolated the triglycerides by means of gel permeation chromatography, transesterified them with potassium methylate and chromatographed the methylated fatty acids using a 60 m fused silica capillary column coated with RTX 2330. There are no significant differences between the total fatty acids and the fatty acids in triglycerides (Table 1.24). During roasting there were only slight changes in the fatty acid composition (Vitzthum, 1976). More recently, Casal et al. (1997) reported that in
Table 1.23 Oil content (%) of dry matter of green Colombian arabica beans (Data from Folstar et al. 1975; Folstar 1985.) Sieve size (mm)
Direct extraction with petroleum ether (40±608C) for 6 h)
Extraction according to the Swiss official method
0.15±0.42 0.42±0.60 0.60±0.85
15.51 13.10 9.36
15.54 15.66 14.06
Chemistry I: Non-volatile Compounds: Lipids
Table 1.24
35
Gas chromatographic data on the fatty acid composition (%) of oil from green coffee beans. Total fatty acids Review 11
C12:0 C14:0 C14:1 C15:0 C15:1 C16:0 C16:1 C17:0 C17:1 C18:0 C18:1 C18:2 C18:3 C19:0 C20:0 C20:1 C20:4 C21:0 C22:0 C22:1 C22:2 C23:0 C24:0 C28:0
Traces 0.1±2.3 0.1±0.5 0.1±1.7 0.1±0.9 16.8±38.6 0.2±4.0 Traces±0.6 Traces±0.3 4.5±13.1 7.6±18.9 30.5±50.4 0.3±6.0
Review 22 Traces
Fatty acids in triglycerides From dewaxed green beans3 0.2
Robusta (n = 9)4
Arabica (n = 4)4
Traces
Traces
Traces
Traces
Traces Traces 30.7±41.5
33.3
Traces Traces 6.6±10.6 7.6±11.9 36.6±45.9 1.1±2.7
27.2±32.1 Traces Traces
26.6±27.8 Traces Traces
7.3 6.6 47.7 1.7
0.7±6.7 Traces±0.4 0.7
0.3±3.3
2.5
5.8±7.2 9.7±14.2 43.9±49.3 0.9±1.4 Traces 2.7±4.3 0.2±0.3
5.6±6.3 6.7±8.2 52.2±54.3 2.2±2.6 Traces 2.6±2.8 Traces±0.3
73.0 70.4 70.3 Traces±1.0 Traces Traces
0.3±6.4
0.5
Traces 0.3±0.8
Traces 0.5±0.6
Traces±0.3
Traces
Traces 0.3±0.4
Traces 0.2±0.4
1
Maier (1981a). Folstar (1985). Folstar et al. (1975). 4 Speer et al. (1993). 2 3
arabica and robusta coffee the roasting process increased the trans fatty acid levels, specifically the contents of C18:2ct and C18:2tc . Furthermore, Folstar (1985) studied the positional distribution of the fatty acids in the triglyceride molecule. A technique was used whereby sn-1,2 (2,3)-diglycerides, sn-2-monoglycerides and fatty acids were obtained from triglycerides through partial deacylation using pancreatic lipase. It was shown that the unsaturated acids, especially linoleic acid, are preferentially esterified with the secondary hydroxyl position in glycerol.
1.15.2 Free fatty acids The presence of free fatty acids (FFA) in coffee has been described by various authors (Kaufmann & Hamsagar, 1962a; Calzolari & Cerma, 1963; Carisano & Gariboldi, 1964; Wajda & Walczyk, 1978). However, all
the data are expressed as the acid value, a common but indirect determination procedure used in the analysis of fat. In the case of coffee this titration method is only very approximate, for it includes not only the free fatty acids themselves but other acid compounds as well. Thus Speer et al. (1993) developed a method to determine the free fatty acids directly. Using the gel chromatographic system with BioBeads S-X3, mentioned above, the coffee lipids, extracted with tertiary butyl methyl ether, can be divided into three individual fractions: a fraction containing the triglycerides, a fraction containing the diterpene fatty acid esters and one containing the free fatty acids. The latter were converted with BF3 /methanol and determined by capillary gas chromatography as methyl esters. Nine different free fatty acids were detected (Speer et al., 1993); they are very uniformly distributed in both the robusta and arabica coffees. In both coffee species
36
the main fatty acids are C18:2 and C16 . It was also possible to detect large proportions of C18 , C18:1 , C20 and C22 , whereas there were no more than traces of C14 , C18:3 and C24 . Differences between arabica and robusta coffee only become visible when their stearic acid and oleic acid contents are compared on chromatograms (Fig. 1.12). While the proportion of stearic acid is noticeably smaller than that of oleic acid in the robustas, the percentages of these two acids in the arabica coffees are almost equal. The ratio stearic acid/oleic acid may give a first indication of robusta in coffee blends.
Fig. 1.12 GC chromatograms of methylated free fatty acids (IS = internal standard for heptadecanoic acid ethyl ester).
The free fatty acid contents of different green robusta and arabica coffee samples are shown in Fig. 1.13. All the arabica coffees had a somewhat lower free fatty acid content than the robustas, within the range of approximately 1 to 1.5 g per 100 g lipid.
Coffee: Recent Developments
In coffees roasted at different temperatures, there were scarcely any changes in the amount and distribution of the individual fatty acids. Only the linoleic acid content decreased slightly as the roasting temperature increased. Using a roasting series of a Madagascar coffee, the differences between the FFA contents determined directly by chromotography and those determined indirectly via the acid value could be demonstrated (Fig. 1.14). While the difference is only about 360 mg for green coffee, it becomes steadily greater as the roasting temperature increases; at the last roasting stage the difference is 1300 mg. This can be explained by the fact that more and more acid compounds are released as the temperature rises. These are primarily phenolic degradation products of the chlorogenic acids that considerably distort the results for the free fatty acids.
Fig. 1.14 Contents of free fatty acids in a green and roasted Madagascar robusta coffee roasted at different temperatures for 212 minutes in each case: determined by GC and by acid value.
In an analysis of 12 commercial coffee samples on the German market, several were found to contain free fatty acids in the range 0.8 g to 1.8 g/100 g lipid (SchluÈter, 1992). In a study of nine French coffee blends with significant Robusta portions, the levels of free fatty acids were between 1.2 and 2.5 g/100 g lipid (LuÈtjohann, 1993).
1.16 DITERPENES IN THE LIPID FRACTION OF ROBUSTA AND ARABICA COFFEES
Fig. 1.13 Contents of free fatty acids in green robusta and arabica coffees.
The main diterpenes in coffee are pentacyclic diterpene alcohols based on the kauran skeleton. The research teams of Bengis and Anderson (1932), Chakravorty et al. (1943a,b), Wettstein et al. (1945), Haworth and
Chemistry I: Non-volatile Compounds: Lipids
37
Johnstone (1957), and Finnegan and Djerassi (1960) worked for several years to identify the structure of two of the coffee diterpenes, namely kahweol and cafestol. Both are sensitive against acids, heat and light, and kahweol in particular is unstable in the purified form. In 1989, 16-O-methylcafestol was isolated from robusta coffee beans and its structure was elucidated by synthesis (Speer & Mischnick, 1989; Speer & Mischnick-LuÈbbecke, 1989). With 16-O-methylkahweol, another O-methyl diterpene has been found in robusta coffee beans (KoÈlling-Speer & Speer; Speer et al., 2000). The structural formulae of these diterpenes are illustrated in Fig. 1.15.
Fig. 1.16 Contents of diterpenes in the unsaponifiable matter of different green arabica and robusta coffees.
Fig. 1.15 Structural formulae of the diterpenes.
Another important group of diterpene derivatives found in coffee is the atractylosides, which are mainly present as glycosides (Obermann & Spiteller, 1976; Maier & Wewetzer, 1978; Maier & MaÈtzel, 1982; Aeschbach et al., 1982; Bradbury & Balzer, 1999). They are discussed in detail in this chapter in the section on carbohydrates. Arabica coffees contain cafestol and kahweol, robusta coffees cafestol, small amounts of kahweol and, additionally, 16-O-methylcafestol (16-OMC), which was found only in robusta coffee beans (Speer & Mischnick-LuÈbbecke, 1989; Speer & Montag, 1989; Speer et al., 1991a) (Fig. 1.16). The absence of 16-OMC in arabica coffee beans was confirmed later by White (1995), Frega et al. (1994), and by Trouche et al. (1997). Because of its stability even during the roasting process, 16-O-methylcafestol has become the ideal quality characteristic for reliably detecting robusta in arabica coffee blends (Speer et al., 1991a).
The presence of 16-O-methylkahweol in various robusta coffees both in green and roasted beans was clearly identified using different spectroscopic methods (KoÈlling-Speer & Speer, Speer et al., 2000). These findings are in contrast to those of De Roos et al. (1997) who described it, only tentatively introduced as a 16-Omethyl derivative of kahweol, as being present exclusively in Coffea stenophylla. Wahlberg et al. (1975) isolated and identified ent-16kauren-19-ol, another diterpene alcohol, but without a furan ring, in the beans of Coffea arabica. As mentioned above, the three diterpenes cafestol, kahweol, and 16-O-methylcafestol are mainly esterified with various fatty acids. In order to analyze the total amount, coffee oil was saponified and the diterpenes were then determined in the unsaponifiable matter by means of GC (Speer & Mischnick-LuÈbbecke, 1989; Frega et al., 1994) or faster by RP-HPLC with acetonitrile/water as eluent (Nackunstz & Maier, 1987; Speer, 1989; White, 1995; Trouche et al., 1997) (Fig. 1.17). In Germany a validated method of 16-OMC determination in roasted coffee has recently been
38
Coffee: Recent Developments
after saponification, a similar result was obtained. In arabicas, the proportions ranged from 0.7% to 2.5%, in robustas the proportions of the free diterpenes were slightly higher at 1.1% to 3.5%.
1.16.2 Diterpene fatty acid esters
Fig. 1.17 HPLC chromatograms of green arabica and robusta coffees. Conditions: column 250 6 4 mm, Nucleosil 120-3 C18, eluent: acetonitrile/water (50:50), detection: UV 220 nm (Speer, 1989).
published as the DIN method No 10779 (DIN, 1999) of the German Institute for Standardization. This DIN method is based on that by Speer (1989), which makes it possible to detect less than 2% robusta in mixtures with arabica. It should be mentioned here that although 16-Omethylcafestol was not detectable in arabica coffee beans, it has been clearly identified in other parts of the arabica coffee plant, for instance in the leaves (KoÈllingSpeer & Speer, 1997).
1.16.1 Free diterpenes In their free form, the diterpenes cafestol, kahweol, and 16-OMC occur only as minor components in coffee oil. Quantifying them requires an effective separation from the major compounds which interfere with the analysis. Using the gel permeation chromatographic system described for the free fatty acids, the free diterpenes could be analysed by subsequent RP-HPLC (Speer et al., 1991b; KoÈlling-Speer et al., 1999). In arabica coffees, both free cafestol and free kahweol were determined in amounts of about 50 to 200 mg per kg dry matter, with mostly more cafestol than kahweol. In robusta coffees, the free cafestol contents ranged from about 50 to 100 mg per kg coffee, slightly higher than the 16-OMC contents of 10 to 50 mg. Only traces of kahweol could be detected in some of them. This was expected because of the small amounts of total kahweol in robusta coffee. When the amounts of the free diterpenes were compared with the total contents of each determined
Until now only a few esters with different fatty acids have been reported (Kaufmann & Hamsagar, 1962b; Folstar et al., 1975; Folstar, 1985; Pettitt, 1987). The working group under Speer identified a number of other esters of 16-OMC (Speer, 1991, 1995) and of cafestol (Kurzrock & Speer, 1997a,b). Using the gel chromatographic system described above, the diterpene esters were isolated together with sterol esters, which could be removed by solid phase extraction on silica cartridges. One fraction was obtained for arabicas, containing the cafestol and kahweol esters; a second fraction was achieved for robustas containing the 16-Omethylcafestol esters. Subsequent analysis by RPHPLC with acetonitrile/isopropanol as eluent facilitates the determination of the individual esters. Figure 1.18 shows the chromatogram of the cafestol esters in a robusta coffee sample. Cafestol esters with fatty acids such as C14 , C16 , C18 , C18:1 , C18:2 , C18:3 , C20 , C22 and C24 were identified, as well as esters with the fatty acid C20:1 and some odd-numbered fatty acids such as C17 , C19 , C21 and C23 . These data were proven for the fatty acids with 16-O-methylcafestol and they seem to be valid for kahweol, too (Kurzrock, 1998). The individual diterpene esters were present in different concentrations in the coffee oil. The odd-
Fig. 1.18 HPLC chromatogram of cafestol fatty acid esters in a robusta coffee. Conditions: column 250 6 4 mm, Nucleosil 120-3 C18, eluent: acetonitrile/isopropanol (60:40), detection: UV 220 nm.
Chemistry I: Non-volatile Compounds: Lipids
numbered fatty acid esters were minor components, whereas the diterpenes, esterified with palmitic, linoleic, oleic, stearic, arachidic, and behenic acid, were found in larger amounts (Speer, 1991, 1995; Kurzrock & Speer, 1997a). The focus was therefore placed on these six diterpene esters, which together constitute nearly 98% of the respective diterpenes. In Table 1.25 a number of published studies on the distribution of the six esters in arabica coffees are given. The total content of these six cafestol esters ranged between 9.4 and 21.2 g/kg dry weight, corresponding to 5.2 to 11.8 g/kg cafestol in different arabica coffees. In robusta coffees, the sum was determined to be between 2.2 and 7.6 g/kg dry weight, corresponding to 1.2 to 4.2 g/kg cafestol, notably less than in the arabica coffees.
Table 1.25
C16 C18 C18:1 C18:2 C20 C22 1
39
1.16.3 Diterpenes in the lipid fraction of roasted coffees During the roasting process a number of new diterpene compounds are formed. Two decomposition products from cafestol and kahweol have been identified in roasted coffee: dehydrocafestol and dehydrokahweol (Fig. 1.19). The amounts of both compounds increase with roasting temperature, according to the contents of cafestol and kahweol in the green coffee (Speer et al., 1991c; Tewis et al., 1993; KoÈlling-Speer et al., 1997). The investigations of KoÈlling-Speer et al. (1997) revealed that there is a relationship between roasting temperature and the cafestol/dehydrocafestol ratio, for both arabica and robusta coffee in the same way (Fig. 1.20). The relationship is approximately linear; the slight deviation for higher roasting temperatures can be
Distribution (%) of diterpene esters in arabica coffees. Cafestol + kahweol (Kaufmann & Hamsagar 1962b) (n = 1)
Cafestol + kahweol (Folstar, 1985) (n = 1)
42.5 17.5 11.0 20.5 6.0 2.5
51.4 9.1 7.4 26.4 4.6 1.1
Cafestol (Kurzrock & Speer, 1997a) (n = 10) 40±49 9±11 9±15 24±30 3±6 0.6±1.2
Kahweol1 (Kurzrock, 1998) (n = 10) 46±50 8±11 8±12 25±29 3±6 0.7±1.3
Kahweol esters calculated as cafestol esters.
Fig. 1.19 Structural formulae of decomposition products of cafestol and kahweol.
Fig. 1.20 Relationship of the cafestol/ dehydrocafestol ratio to roasting temperature.
40
Coffee: Recent Developments
attributed to the formation of further cafestol decomposition products (see below). The cafestol/dehydrocafestol ratio is also related to the `roasting taste', as tests with well-trained coffee tasters for two arabica coffees showed (Fig. 1.21). Ratio values around 20 are judged to be strong roasted, values around 10 as overroasted. These results were confirmed by analyses of coffees on the German market, which mostly had ratios above 25 (KoÈlling-Speer et al., 1997; Speer et al. unpublished results). Therefore the cafestol/dehydrocafestol ratio is a suitable measure of the roasting grade of coffees. Fig. 1.22 HPLC chromatogram of a roasted arabica/robusta mixture.
tribution of the esters remains nearly the same (Kurzrock & Speer, 1997a). Kurzrock et al. (1998) demonstrated that cafestol was also dehydrated within the fatty acid esters. In model experiments in which cafestol palmitate and cafestol linoleate were heated, the corresponding dehydrocafestol esters were obtained, which were also identified in roasted coffee (Fig. 1.23). Fig. 1.21 Relationship of the cafestol/ dehydrocafestol ratio and organoleptic roasting grade.
Cafestal and kahweal (see Fig. 1.19) are two further degradation products of cafestol and kahweol which have been discovered in the unsaponifiable matter of commercial roasted coffee (Hruschka & Speer, 1997). In German coffees, levels of about 0.3 mg/g lipid have been determined; one espresso showed a significantly higher content, with 0.6 mg/g lipid. An HPLC chromatogram of a commercial roasted blend containing both arabica and robusta coffee is shown in Fig. 1.22. Decomposition of the main diterpenes during roasting is particularly noticeable for the free ones. Depending on the roasting temperature, up to 80% of the initial levels of cafestol and kahweol was lost, whereas 16-Omethylcafestol is much less affected (KoÈlling-Speer et al., 1999). The stability of the fatty acid esters of the three diterpenes during roasting is quite different. Examination of the 16-O-methylcafestol esters has shown that these are clearly stable during roasting, and in spite of different roasting temperatures, the proportional distribution for the diterpene esters is almost the same (Speer et al., 1993). In contrast, the contents of the diterpene esters of cafestol and kahweol decrease according to the roasting temperature, but the dis-
Fig. 1.23 Mass spectra of dehydrocafestol palmitate and cafestol palmitate.
Chemistry I: Non-volatile Compounds: Lipids
1.16.4 Diterpenes in coffee: health aspects Several studies have reported that drinking brewed coffee can cause an increase in serum cholesterol level. It was shown that this effect was caused by the lipids in the coffee brew, which, although barely soluble in water, were present in the brew depending on the preparation of the infusion. Initially, triglycerides were said to be responsible for this effect, however within the last few years it has been established that it is the diterpenes, especially cafestol and kahweol both in free form and as palmitate esters, which influence the serum cholesterol level (Bak & Grobbee, 1989; Weusten-Van der Wouw et al., 1994; Mensink et al., 1995; De Roos & Katan, 1999). Other diterpenes have not yet been tested. In addition, a substantial number of scientific publications have reported positive effects of diterpenes. It has been demonstrated that cafestol stimulates glutathion-S-transferase activity, through which the decomposition of chemical carcinogens, etc. accelerated (Lam et al., 1982). Other authors have reported that cafestol and kahweol protect against aflatoxin B1induced genotoxicity (Miller et al., 1993; Cavin et al., 1998; see also Chapter 8 in this book). Thus, reports on the presence of diterpenes in coffee beverages prepared by different methods have been of great interest (Ratnayake et al., 1993; Sehat et al., 1993; Urgert et al., 1995; Gross et al., 1997). Using the example of 16-O-methylcafestol esters, Sehat et al. (1993) showed that lipophile diterpene esters flow into the coffee infusion and are even detectable in instant coffee granules. The amount in the drink is decisively dependent on the method of preparation and is directly related to the amount of total lipids in the drink. With filtered coffee prepared in a household coffee machine, the amount of lipids was less than 0.2%. In contrast, when espresso coffee was prepared, between 1 and 2% of the lipids, and thereby diterpenes as well, dispersed from the finely ground espresso coffee into the drink. When coffee was prepared Scandinavian style by boiling, it was shown to contain up to 22% of the coffee fat. The proportional distribution of diterpenes in the coffee drink was nearly identical to the distribution in the roasted coffee powder. In espresso prepared from arabica coffee, levels of 1.3 mg cafestol fatty acid esters and 0.5 mg kahweol esters per 50 ml cup were determined by Kurzrock (1998), corresponding with approximately 1.5% of cafestol esters and approximately 1.0% of kahweol esters in the roasted ground coffee. These results confirm the findings for the 16-O-
41
methylcafestol esters. In addition, the decomposition products dehydrokahweol, dehydrocafestol and cafestal, as well as some esters from dehydrocafestol, have also been identified in coffee beverages.
1.17 STEROLS Coffee contains a number of sterols typical for other seed oils as well. In addition to 4-desmethylsterols, various 4-methyl- and 4,4-dimethylsterols have been identified. A summary of the results published by different authors is given in Table 1.26. The structural formulae of the main sterols are presented in Figs 1.24± 1.26. In coffee beans, sterols have been found in both free (around 40%) and esterified form (around 60%) (Nagasampagi et al., 1971; Picard et al. 1984). The total amount is determined in the unsaponifiable matter of the coffee oil as TMS-derivatives by means of GC or GC/MS. A fractionation in desmethyl, 4-methyl- and 4,4-dimethylsterols using TLC, HPLC or silica gel cartridges was often performed (Nagasampagi et al., 1971; Itoh et al., 1973a,b; Picard et al., 1984; Horstmann & Montag, 1986; Homberg & Bielefeld, 1989). The desmethylsterols represent 90% of the total sterol fraction, which ranged from 1.5% to 2.4% of the lipids (Picard et al., 1984). Nagasampagi (1971) found higher values of 5.4%. The distribution of the main desmethylsterols in different robusta and arabica coffee samples is presented in Table 1.27. The main sterol is b-sitosterol at about 50%, followed by stigmasterol and campesterol. 24-Methylenecholesterol and 5-avenasterol, occurring in much higher amounts in robusta than in arabica coffee beans, are suitable for coffee blend studies (Duplatre et al., 1984; Frega et al., 1994). However, because of their varying natural contents, they are qualified for determining robusta parts in arabica coffee mixtures only from 20% onwards (see previous section). Picard et al., (1984) separated the free sterols and the sterol fatty acid esters by means of column chromatography (aluminium oxide). When 12 robusta coffee samples were analyzed, a changed distribution for the main sterols in comparison to the total sterols was evident. For the free sterols, the order is stigmasterol > b-sitosterol > campesterol, while for the sterol esters the order is b-sitosterol > campesterol > stimasterol. Furthermore Picard et al. studied the individual fatty acids in the sterol esters, C18 , C16 and C18:1 were the main compounds, with a proportional distribution
42
Table 1.26
Coffee: Recent Developments
Sterols identified in coffee.
Sterols 4-Desmethylsterols Cholesterol* Campesterol* Stigmasterol* b-Sitosterol* 5-Avenasterol* Cholestanol* Campestanol* 24-Methylenecholesterol* Stigmastanol* = sitostanol 7-Stigmastenol* 7-Avenasterol* 7-Campesterol* 5,23-Stigmastadienol 5,24-Stigmastadienol = fucosterol Clerosterol* Brassicasterol 4-Methylsterols Citrostadienol* Cycloeucalenol* Obtusifoliol* Gramisterol* 24-Methylenelophenol 24-Ethylenelophenol 4a,24R-Dimethyl-5a-cholest-8-en-3b-ol 4a,24R-Dimethyl-5a-cholest-7-en-3b-ol 4a, 24R Methyl-5a-stigmast-8-en-3b-ol 4,4-Dimethylsterols Cycloartenol* 24-Methylenecycloartanol* Cycloartanol* Cyclobranol b-Amyrin
References N N N N N N N
I I I I I
I I
T T T T T T T T T T
P P P P P
P P
D D D D D
D D D
I N N N N
I I I
N N N N N
T T T T T T T
I I I I I
T T T
P P P P
M M M M M M M M M M M M M M M
F F F F F
S S S S S
F F
S S S S
F
S S S
P
P P
F F
S S
* Structural formula is given D = Duplatre et al. (1984), F = Frega et al. (1994), I = Itoh et al. (1973a,b), M = Mariani & Fedeli (1991), N = Nagasampagi et al. (1971), P = Picard et al. (1984), S = Speer et al. (1996), T = Tiscornia et al. (1979).
similar to that reported in triglycerides. Roasting the coffee beans hardly affected the amounts and the distribution of the sterols (Duplatre et al., 1984; Speer et al., 1996). No changes, either, have been observed after industrial steaming processes according to the Lendrich procedure (Speer et al., 1996). In order to examine the sterols in coffee infusions, a Scandinavian style coffee, an espresso and a filter coffee were analyzed. Cholesterol, campesterol, stigmasterol, b-sitosterol, stigmastanol, 5-avenasterol, 7-stigmastenol, 7-avenasterol, citrostadienol, gramisterol and cycloartenol and traces of 24-methylenecycloarte-
nol were identified and quantified in all the coffee infusions. As reported for the diterpenes, filter coffee obtained the lowest content of sterols (Fig. 1.27) (Speer et al., 1996).
1.18 TOCOPHEROLS The presence of tocopherols in coffee oil was described by Folstar et al. (1977) for the first time, a-tocopherol was clearly identified, while b- and g-tocopherol, not being separated by TLC and GC, were considered as
Chemistry I: Non-volatile Compounds: Lipids
Fig. 1.24 4-Desmethylsterols.
one group (Fig. 1.28). Cros et al. (1985) also determined total b-and g-tocopherol by HPLC. Folstar et al. (1977) found concentrations of a-tocopherol of 89 to 188 mg/kg oil, and values for b- plus g-tocopherol of 252±530 mg/kg oil. In 1988, Aoyama et al. analyzed a-, b- and g-toco-
43
Fig. 1.25
4-Methylsterols.
Fig. 1.26
4,4-Dimethylsterols.
pherols in different varieties of coffee beans. They were contained in a ratio of approximately 2:4:0.1, the total content being about 5.5 to 6.9 mg/100 g. The predominance of a-tocopherol is a prominent feature of coffee beans, in contrast to other vegetables and fruits. Ogawa et al. (1989) determined the contents of tocopherols by HPLC in 14 green coffee beans, their roasted beans and infusions, and in 38 instant coffees. The maximum of total tocopherols in the green coffee beans was 15.7 mg/100 g and the average was 11.9 mg/ 100 g. The contents of a- and b-tocopherol were 2.3 to 4.5 and 3.2 to 11.4 mg/100 g, respectively, g- and dtocopherol were not found. Roasting diminishes the content of a-tocopherol, b-tocopherol and total tocopherols to 79 to 100%, 84 to 100% and 83 to 99%, respectively. In Fig. 1.29 the HPLC chromatograms of tocopherols for a green arabica and robusta coffee ± using the method published by Coors (1984) for vegetable oils ± are presented. In the arabica coffee oil, values of 161 mg/kg a-tocopherol and 597 mg/kg b-tocopherol
44
Coffee: Recent Developments
Table 1.27 Distribution (%) of desmethylsterols in arabica and robusta coffees (30 samples) (Mariani & Fideli, 1991). Mean value Sterols Cholesterol Campesterol Stigmasterol b-Sitosterol 5-Avenasterol Campestanol 24-Methylenecholesterol Sitostanol 7-Stigmastenol 7-Avenasterol 7-Campesterol 5,23-Stigmastadienol 5,24-Stigmastadienol Clerosterol
Arabica
Robusta
Arabica
Robusta
0.2±0.4 14.7±17.0 20.5±23.8 46.7±53.8 1.6±4.1 0.2±0.6 0.0±0.4 1.4±2.8 0.9±4.5 1.2±2.1 0.4±1.2 0.2±0.5 0.0±0.4 0.2±0.8
0.1±0.3 15.5±18.8 20.0±26.7 40.6±50.7 5.1±12.6 0.1±0.3 1.5±2.4 0.5±1.2 0.1±0.8 0.2±0.6 0.1±0.6 0.1±2.0 0.0±0.3 0.5±1.0
0.3 15.8 21.9 51.6 2.7 0.4 0.2 2.0 2.2 1.5 0.6 0.3 0.1 0.5
0.2 16.9 23.1 45.4 9.1 0.2 1.9 0.8 0.2 0.4 0.2 0.5 0.0 0.7
were found; the robusta coffee oil contained 107 mg/kg a-tocopherol and 260 mg/kg b-tocopherol ± significantly higher values. By the use of GC-MS, gtocopherol has been detected in some robusta coffees (KoÈlling-Speer, unpublished data). Tocopherols have also been analysed in coffee brews. The contents of total tocopherols in coffee infusions and instant coffee solutions were determined as 0.003±0.013 and 0.001±0.013 mg/100 ml, respectively (Ogawa et al., 1989). Fig. 1.28
Structural formulae of tocopherols.
1.19 OTHER COMPOUNDS
Fig. 1.27 Contents of selected sterols in differently prepared coffee infusions.
Kaufmann & Sen Gupta (1964) identified squalene in the unsaponifiable matter of coffee oil. Furthermore, Folstar (1985) reported a number of both odd and even chain length alkanes in wax-free coffee oil as well as in coffee wax. In 1999, Kurt & Speer detected and isolated a new component with the molecular formula C19 H30 O2 . Its structure is similar to the known coffee diterpene cafestol. The most important differences are the absence of the furan ring and the location of one methyl group at the carbon atom C10 . The new component was named coffeadiol (Fig. 1.30).
Chemistry I: Non-volatile Compounds: Lipids
45
Fig. 1.31 Structural formulae of carbonic acid 5hydroxytryptamides (C-5-HT).
Fig. 1.29 HPLC chromatograms of tocopherols of green coffees. HPLC conditions: LiChrosorb Si 60, 5 mm, n-hexane/dioxan (94:6), detection: 295/330 nm.
Fig. 1.30 Structural formula of coffeadiol.
1.20 COFFEE WAX The surface of green coffee beans is covered by a thin wax layer. The wax content is generally defined as the material obtained from the unground beans by extraction with chlorinated solvents such as chloroform or dichloromethane. The amount of surface wax is between 0.2% and 0.3% of the total coffee lipids. Folstar (1985) observed that only 37% of the coffee wax is soluble in petroleum ether. Investigating the fatty acid composition, he reported a large difference between the fatty acids in the petroleum ether-soluble part of the wax and that of coffee oil. The relatively high percentage of saturated higher fatty acids found in coffee wax is significant. The first investigations into the wax composition of green arabica coffee beans were performed by Wurziger and his co-workers (Dickhaut, 1966; Harms, 1968). They isolated and identified three carbonic acid 5hydroxytryptamides (C-5-HT) (Fig. 1.31). Arachidic acid (n = 18), behenic acid (n = 20) and lignoceric acid
(n = 22) are combined with the primary amino group of 5-hydroxytryptamine. In addition, Folstar et al. (1979) reported the presence of stearic acid 5-hydroxytryptamide (n = 16), and later that of o-hydroxyarachidic acid 5-hydroxytryptamide (n = 18), and o-hydroxybehenic acid 5hydroxytryptamide (n = 20) (Folstar et al., 1980). All components were confirmed by KoÈnig & Sturm (1982). Arachidic acid and behenic acid 5-hydroxytryptamide predominate, the other amides are only minor constituents. The coffee wax and its constituents may not be digested adequately in certain susceptible individuals, and are considered responsible for their gastro-enteric reactions to coffee beverages (Lickint, 1931). Removal of the waxy layer by washing the beans with solvents or by steaming increases their wholesomeness (Behrens & Malorny, 1940; Wurziger, 1971a; Fintelmann & Haase, 1977; Corinaldesi et al. 1989). A steaming method was developed in 1933 (Lendrich et al.), and was subsequently improved several times, for example by Roselius et al. (1971). As main constituents of the wax, although poorly water-soluble, the C-5-HT were considered to be the `irritating substances' (Wurziger, 1971b; RoÈsner et al., 1971). However, Fehlau and Netter (1990), studying the influence of coffee infusions on the gastric mucosa of rats, reported that the gastric irritating effect of C-5-HT was much less than that caused by comparable coffee infusions. Several working groups developed analytical methods for determining the contents of C-5-HT in green, roasted and variously treated coffees (Culmsee, 1975; Hubert et al., 1975; Kummer & BuÈrgin, 1976; Hunziker & Miserez, 1977, 1979; Studer & Traitler, 1982; Chiacchierini & Ruggeri, 1985; Lagana et al., 1989; Battini et al., 1989; Kele & Ohmacht, 1996). The total content in green arabica coffees ranged between 500 and 2370 mg/kg, whereas in robusta coffees levels of 565±1120 mg/kg were found (Maier, 1981b). In coffee
46
stored for 30 years, the total content varied between 30 and 625 mg/kg (Wurziger, 1973). Treatment of the coffee beans, for example by polishing, dewaxing and decaffeinating, led to a substantial reduction in C-5HT (Harms & Wurziger, 1968; Hunziker & Miserez, 1979; Folstar et al., 1979, 1980; van der Stegen & Noomen, 1977). Furthermore, C-5-HT is partly decomposed by roasting (Wurziger, 1972; Hunziker & Miserez, 1979). For normal roasted coffees the contents ranged from 600 to 1000 mg/kg. Viani & Horman (1975) proposed pathways for the thermal decomposition of C-5-HT. They identified a number of alkylindoles and alkylindanes after pyrolysis of pure C22 -5-HT. Folstar et al. (1980) detected 5-hydroxyindole, 3-methyl-5-hydroxyindole as well as several n-alkanes, n-alkanenitriles and n-alkaneacidamides. Considering 5-hydroxytryptamides to be present only in the waxy layer of the coffee beans and being reduced by the processes described above, Wurziger (1971b) suggested the amount of C-5-HT should be a measure for treated coffees. Since 1973 in Switzerland, roasted C-5-HT-reduced coffees are designated `low irritating' (in German: `reizarm'), when the content of C-5-HT is lower than 400 mg/kg (Anon, 1973). According to van der Stegen (1979), coffee brew may contain up to 2.3 mg C-5-HT per liter when the beverage is prepared by percolation of untreated beans. The C-5-HT could not be detected in beverages prepared by a filtration method, or by percolating dewaxed beans. Furthermore, the C-5-HT become of interest because of its antioxidant effects (Lehmann et al., 1968; Bertholet & Hirsbrunner, 1984).
REFERENCES Aeschbach, R., Kusy, A. & Maier, H. G. (1982) Diterpenoide in Kaffee. Z. Lebensm. Unters.-Forsch., 175, 337±41. Anon (1973) Kreisschreiben des EidgenoÈss. Gesundheitsamtes Bern. Kaffee- und Tee-Markt, 23, 1. AOAC (1965) Method 14.029. In: Official Methods of Analysis of the Association of Official Analytical Chemists Handbook, 10th edn. AOAC, Washington, DC. Aoyama, M., Maruyama, T., Kanematsu, H. et al. (1988) Studies on the improvement of antioxidant effect of tocopherols. XVII. Synergistic effect of extracted components from coffee beans. Yukagaku, 37, 606±12. Bak, A.A.A. & Grobbee, D.E. (1989) The effect on serum cholesterol levels of coffee brewed by filtering or boiling. N. Engl. J. Med., 321, 1432±7. Battini, M.L., Careri, M., Casoli, A., Mangia, A. & Lugari, M.T. (1989) Determination of N-alkanoyl-5-hydroxytryptamines (C-5-HT) in coffee beans by means of HPLC and TLC. Ann. Chim., 79, 369±77.
Coffee: Recent Developments
Behrens, B. & Malorny, G. (1940) Naunyn-Schmiedebergs Arch. Pharmak. U. exp. Path., 194, 369. Bengis, R.O. & Anderson, R.J. (1932) The chemistry of the coffee bean. I. Concerning the unsaponifiable matter of the coffee bean oil. Extraction and properties of kahweol. J. Biol. Chem., 97, 99±113. Bertholet, R. & Hirsbrunner, P. (1984) Preparation of 5-hydroxytryptamine from coffee wax. Patent EP 1984±104696. Bradbury, A.G.W. & Balzer, H.H. (1999) Carboxyatractyligenin and atractyligenin glycosides in coffee. In: Proceedings of the 18th ASIC Colloquium (Helsinki), pp. 71±7. ASIC, Paris, France. Calzolari, C. & Cerma, E. (1963) Sulle sostanze grasse del caffeÁ. Riv. Ital. Sostanze Grasse, 40, 176±80. Carisano, A. & Gariboldi, L. (1964) Gaschromatographic examination of the fatty acids of coffee oil. J. Sci. Food Agric., 15, 619±22. Casal, S., Oliveira, M.B. & Ferreira, M.A. (1997) Discrimination of Coffea arabica and Coffea canephora var. robusta beans by their fatty acid composition. In: Proceedings of Euro Food Chem IX, Vol. 3, 685, Interlaken, Switzerland (eds R. AmadoÁ & R. Battaglia). Cavin, C., Holzhauser, D., Constable, A., Huggett, A.C. & Schilter, B. (1998) The coffee-specific diterpenes cafestol and kahweol protect against aflatoxin B1-induced genotoxicity through a dual mechanism. Carcinogenesis, 19, 1369±75. Chakravorty, P.N., Levin R.H., Wesner, M.M. & Reed. G. (1943b) Cafesterol III. J. Am. Chem. Soc., 65, 1325±8. Chakravorty, P.N., Wesner, M.M. & Levin R.H. (1943a) Cafesterol II. J. Am. Chem. Soc., 65, 929±32. Chassevent, F., Dalger, G., Gerwig, S. & Vincent, J.-C. (1974) Contribution aÁ l'eÂtude des Mascarocoffea. CafeÂ, Cacao, TheÂ, 18, 49±56. Chiacchierini, E. & Ruggeri, P. (1985) Dewaxed coffee: determination of carboxylic acid 5-hxdroxytryptamides (C-5-HT). Riv. Soc. Ital. Sci. Aliment., 14, 197±200. Coors, U. (1984) Bestimmung und Verteilung von Tocopherolen and Tocotrienolen in Lebensmitteln. Dissertation, University of Hamburg, Germany. Corinaldesi, R., De Giorgio, R., Stanghellini, V. et al. (1989) Effect of the removal of coffee waxes on gastric acid secretion and serum gastrin levels in healthy volunteers. Curr. Ther. Res., 46, 13±18. Cros, E., Fourny, G. & Vincent, J.C. (1985) Tocopherols in coffee. High-pressure liquid chromatographic determination. In: Proceedings of the 11th ASIC Colloquium (LomeÂ), pp. 263± 71. ASIC, Paris, France. Culmsee, O. (1975) Methode zur quantitativen Bestimmung der CarbonsaÈure-5-hydroxy-tryptamide im Kaffee. Dt. Lebensm. Rundschau, 71, 425±7. De Roos, B. & Katan, M.B. (1999) Possible mechanisms underlying the cholesterol-raising effect of the coffee diterpene cafestol. Curr. Opin. Lipid., 44, 41±5. De Roos, B., van der Weg, G., Urgert, R., van de Bovenkamp, P., Charrier, A. & Katan, M.B. (1997) Levels of cafestol, kahweol, and related diterpenoids in wild species of the coffee plant Coffea. J. Agric. Food Chem., 45, 3065±9.
Chemistry I: Non-volatile Compounds: Lipids
È ber phenolische Substanzen in Kaffee und Dickhaut, G. (1966) U deren analytische Auswertbarkeit zur Kaffeewachsbestimmung. Dissertation, University of Hamburg. DIN (1999) Method 10779. Analysis of coffee and coffee products ± determination of 16±O-methylcafestol content of roasted coffee ± HPLC-method. Deutsches Institut fuÈr Normung, Berlin. Duplatre, A., Tisse, C. & Estienne, J. (1984) Contribution a l'identification des espeÁces arabica et robusta par eÂtude de la fraction steÂrolique. Ann. Fals. Exp. Chim., 828, 259±70. Fehlau, R. & Netter, K.J. (1990) The influence of untreated and treated coffee and carboxylic acid hydroxytryptamides on the gastric mucosa of the rat. Z. Gastroenterol., 28, 234±8. Finnegan R.A. & Djerassi, C. (1960) Terpenoids XLV. Further studies on the structure and absolute configuration of cafestol. J. Am. Chem. Soc., 82, 4342±4. Fintelmann, V. & Haase, W. (1977) Unbearbeiteter und veredelter Kaffee. Vergleich der magensaÈurestimulierenden Wirkung im gekreuzten Doppelblindversuch. Z. Allg. Med., 53, 1888±95. Folstar, P. (1985) Lipids. In: Coffee Chemistry, Vol. 1, (eds R.J. Clarke & R. Macrae) pp. 203±22. Elsevier Applied Science, London. Folstar, P., Pilnik, W., de Heus, J.G. & van der Plas, H.C. (1975) The composition of fatty acids in coffee oil and wax. Lebensmittelwiss. Technol., 8, 286±8. Folstar, P., van der Plas, H.C., Pilnik, W. & De Heus, J.G. (1977) Tocopherols in the unsaponifiable matter of coffee bean oil. J. Agric. Food Chem, 25, 283±5. Folstar, P., van der Plas, H.C., Pilnik, W., Schols, H.A. & Melger, P. (1979) Liquid chromatographic analysis of Nb -alkanoylhydroxy-tryptamide (C-5-HT) in green coffee beans. J. Agric. Food Chem., 27, 12±15. Folstar, P., Schols, H.A., van der Plas, H.C., Pilnik, W., Landherr, C.A. & van Vildhuisen, A. (1980) New tryptamine derivatives isolated from wax of green coffee beans. J. Agric. Food Chem., 28, 872±4. Frega; N., Bocci, F. & Lercker, G. (1994) High resolution gas chromatographic method for determination of Robusta coffee in commercial blends. J. High Res. Chromatogr., 17, 303±7. German Society for Lipid Sciences (DGF), No. B-1b (1952), Einheitsmethoden 1950±1975. Wissenschaftliche Verlagsgesellschaft, Stuttgart. Gross, G., Jaccaud, E. & Huggett, A.C. (1997) Analysis of the content of the diterpenes cafestol and kahweol in coffee brews. Food Chem. Toxicol., 35, 547±54. Harms, U. (1968) BeitraÈge zum Vorkommen und zur Bestimmung von CarbonsaÈure-5-hydroxy-tryptamiden in Kaffeebohnen. Dissertation, University of Hamburg. Harms, U. & Wurziger, J. (1968) Carboxylic acid 5-hydroxytryptamides in coffee beans. Z. Lebensm. Unters.-Forsch., 138, 75±80. Hartmann, L., Lago, R.C.A., Tango, J.S. & Teixeira, C.G. (1968) The effect of unsaponifiable matter on the properties of coffee seed oil. J. Am. Oil. Chem. Soc., 45, 577±9. Haworth, R.D. & Johnstone, R.A.W. (1957) Cafestol. Part II. J. Chem. Soc. (London), 1492±6.
47
Homberg, E. & Bielefeld, B. (1989) Einflub von Minorbestandteilen des Unverseifbaren auf die Sterinanalyse. Fat. Sci. Technol., 91, 105±108. Horstmann, P. & Montag, A. (1986) Neue Methoden zur schnellen Isolierung von Sterinen aus Fettmatrices. Fette Seifen Anstrichm., 88, 262±4. Hruschka, A. & Speer, K. (1997) Cafestal in coffee. In: Proceedings of Euro Food Chem IX, Vol. 3, Interlaken, Switzerland (eds R. AmadoÁ & R. Battaglia), pp. 655±8. Hubert, P., Kwasny, H., Werkhoff, P. & Turner, U. (1975) Analytik von CarbonsaÈurehydroxytryptamiden in Kaffee. Z. Anal. Chem., 285, 242±50. Hunziker, H.R. & Miserez, A. (1977) Bestimmung der 5Hydroxytryptamide in Kaffee mittels Hochdruck-FluÈssigkeitschromatographie. Mitt. Geb. Lebensm. Unters. Hyg., 68, 267±74. Hunziker, H.R. & Miserez, A. (1979) Bestimmung der 5Hydroxytryptamide in Kaffee mittels Hochdruck-FluÈssigkeitschromatographie, Mitt. Geb. Lebensm. Unters. Hyg., 70, 142±52. Itoh, T., Tamura, T. & Matsumoto, T. (1973a) Sterol composition of 19 vegetable oils. J. Am. Oil Chem. Soc., 50, 122±5. Itoh, T., Tamura, T. & Matsumoto, T. (1973b) Methylsterol compositions of 19 vegetable oils. J. Am. Oil Chem. Soc., 50, 300±303. Kaufmann, H.P. & Hamsagar, R.S. (1962a) Zur Kenntnis der Lipoide der Kaffeebohne. II. Die VeraÈnderung der Lipoide bei der Kaffee-RoÈstung. Fette Seifen Anstrichm. 64, 734±8. Kaufmann, H.P. & Hamsagar, R.S. (1962b) Zur Kenntnis der Lipoide der Kaffeebohne. I. Uber FettsaÈure-Ester des Cafestols. Fette Seifen Anstrichm., 64, 206±13. Kaufmann, H.P. & Sen Gupta, A.K. (1964) Uber die Lipoide der Kaffeebohne. V. Die Triterpene und Kohlenwasserstoffe. Fette Seifen Anstrichm., 66, 461±6. Kele, M. & Ohmacht, R. (1996) Determination of serotonin released from coffee wax by liquid chromatography. J. Chromatogr., 730, 59±62. KoÈlling-Speer, I., Kurt, A., Nguyen, Thu, Speer, K. (1997) Cafestol and dehydrocafestol in roasted coffee. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 201±204: ASIC, Paris, France. KoÈlling-Speer, I. & Speer, K. (1997) Diterpenes in coffee leaves. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 150± 54. ASIC, Paris, France. KoÈlling-Speer, I. & Speer, K. 16-O-methylkahweol in Robusta coffee. In preparation. KoÈlling-Speer, I., Strohschneider, S., Speer, K. (1999) Determination of free diterpenes in green and roasted coffees. J. High Resol. Chromatogr., 22, 43±6. KoÈnig, W.A. & Sturm, R. (1982) Gas chromatography and mass spectrometry as aids in studying high-boiling coffee components. In: Proceedings of the 10th ASIC Colloquium (Salvador), pp. 271±8. ASIC, Paris, France. Kummer, P. & BuÈrgin, E. (1976) Neue Erkenntnisse zur quantitativen Bestimmung der Carbon-saÈure-5-hydroxytryptamide in Kaffee. Mitt. Geb. Lebensm. Unters. Hyg., 67, 212±15.
48
Kurt, A. & Speer, K. (1999) A new component in the lipid fraction of coffee. In: Proceedings of Euro food chem X, Vol. 3. 22±24 September, Budapest, Hungary, pp. 882±6. Kurzrock, T. (1998) CafestolfettsaÈureester im Kaffee. Dissertation, Technical University of Dresden, Germany. Kurzrock, T., KoÈlling-Speer, I. & Speer, K. (1998) Identification of dehydrocafestol fatty acid esters in coffee. In: Proceedings of the 20th International Symposium on Capillary Chromatography (eds P. Sandra & A.J. Rackstraw) No 27. Naxos Software Solutions, M. Schaefer, Schriesheim, Germany. Kurzrock, T. & Speer, K. (1997a) Fatty acid esters of cafestol. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 133±40. ASIC, Paris, France. Kurzrock. T. & Speer, K. (1997b) Identification of cafestol fatty acid esters. In: Proceedings of Euro Food Chem. IX, Vol. 3. Interlaken, Switzerland. (eds R. AmadoÁ & R. Battaglia), pp. 659±63. Lagana, A., Curini, L., De Angelis Curtis, S. & Marino, A. (1989) Rapid liquid chromatographic analysis of carboxylic acid 5hydroxytryptamides in coffee. Chromatographia, 28, 593±6. Lam; L.T.K.; Sparnis, V.L. & Wattenberg, L.W. (1982) Isolation and identification of kahweol palmitate and cafestol palmitate as active constituents of green coffee beans that enhance glutathione S-transferase activity in the mouse. Cancer Res., 42, 1193±8. Lehmann, G., Neunhoeffer, O., Roselius, W. & Vitzthum, O. (1968) Antioxidants made from green coffee beans and their use for protecting autoxidizable foods. Patent DE 1668236. Lendrich, P., Wemmering, E. & Lendrich, O. (1933) Verfahren zum Verbessern von Kaffee. DR-Patent 576.515. Lercker, G., Caboni, M.F., Bertacco, G. et al. (1996) Coffee lipid fraction I. Influence of roasting and decaffeination. Ind. Alimentari, 35, 1057±65. Lickint, F. (1931) Med. Klin., 27, 387. È bergaÈnge von Kaffeelipiden am Beispiel von LuÈtjohann, J. (1993) U freien FettsaÈuren und Triacylglycerinen in das KaffeegetraÈnk. Thesis, University of Hamburg. Maier, H.G. (1981a) Rohkaffee. In: Kaffee, pp. 21±2. Paul Parey Verlag, Berlin and Hamburg. Maier, H.G. (1981b) Rohkaffee. In: Kaffee, pp. 26±7. Paul Parey Verlag, Berlin and Hamburg. Maier, H.G. & MaÈtzel, U. (1982) Atractyligenin und seine Glykoside im Kaffee. In: Proceedings of the 10th ASIC Colloquium (Salvador), pp. 247±51. ASIC, Paris, France. Maier, H.G. & Wewetzer. H. (1978) Bestimmung von DiterpenGlykosiden im Bohnenkaffee. Z. Lebensm. Unters.-Forsch., 167, 105±107. Mariani, C. & Fedeli, E. (1991) Sterols of coffee grain of Arabica and Robusta species. Riv. It. Sost. Grasse, 68, 111±15. Mensink, R.P., Lebbink, W.J., Lobbezoo, I.E., Van der Wouv Weusten, M.P.M.E., Zock, P.L. & Katan, M.B. (1995) Diterpene composition of oils from Arabica and Robusta coffee beans and their effects on serum lipids in man. J. Int. Med., 237, 543±50. Miller, E.G., Gonzales-Sanders, A.P., Couvillion, A.M., Binnie, W.H., Sunahara, G.I. & Bertholet, R. (1993) Inhibition of oral carcinogenesis by roasted coffee beans and roasted coffee bean
Coffee: Recent Developments
fractions. In: Proceedings of the 15th ASIC Colloquium (Montpellier), pp. 420±25. ASIC, Paris, France. Nackunstz, B. & Maier, H. G. (1987) Diterpenoide im Kaffee. III. Cafestol und Kahweol. Z. Lebensm. Unters.-Forsch., 184, 494±9. Nagasampagi, B.A., Rowe, J.W., Simpson, R. & Goad, L.J. (1971) Sterols of coffee. Phytochemistry, 10, 1101±17. Obermann, H. & Spiteller, G. (1976) Die Strukturen der `KaffeeAtractyloside'. Chem. Ber., 109, 3450±61. Ogawa, M., Kamiya, C. & Iida, Y. (1989) Contents of tocopherols in coffee beans, coffee infusions and instant coffee. Nippon Shokuhin Kogyo Gakkaishi, 36, 490±94. Pettitt, B.C. Jr (1987) Identification of the diterpene esters in Arabica and Canephora coffees. J. Agric. Food Chem., 35, 549± 51. Picard, H., Guyot, B. & Vincent, J.-C. (1984) EÂtude des composeÂs steÂroliques de l'huile de cafe coffea canephora. CafeÂ, Cacao, TheÂ, 28, 47±62. Pokorny, J. & Forman, L. (1970) Pflanzenlipide, 2. Mitt. Kaffeelipide. Nahrung, 14, 631±2. Ratnayake, W.M.N., Hollywood, R., O'Grady, E. & Starvric, B. (1993) Lipid content and composition of coffee brews prepared by different methods. Food Chem. Toxicol., 31, 263±9. Roffi, J.A., Corte dos Santos, J.T.M., Busson, F. & Maigrot, M. (1971) CafeÂs verts et torrefieÂs de l'Angola. In: Proceedings of the 5th ASIC Colloquium (Lisbon), pp. 179±200. ASIC, Paris, France. Roselius, W., Vitzthum, O. & Hubert, P. (1971) Removal of undesirable irritants from raw coffee beans. Patent US 1971± 187168. RoÈsner, P., Keiner, F. & KuÈhn, U. (1971) Der Einflub von behandeltem und unbehandeltem Kaffee auf die MagensaÈuresekretion. Med. Klin., 66, 238±42. SchluÈter, S. (1992) Die Verteilung der freien und in den Triacylglycerinen gebundenen FettsaÈuren von Robusta-Kaffee. Thesis, University of Hamburg. Schweizerische Lebensmittelbuch (1973) Methode No 35A/08. Schweizerische Lebensmittelbuchkommission und EidgenoÈssisches Gesundheitsamt, Bern. Sehat, N., Montag, A. & Speer, K. (1993) Lipids in the coffee brew. In: Proceedings of the 15th ASIC Colloquium (Montpellier), pp. 583±92. ASIC, Paris, France. Speer, K. (1989) 16-O-Methylcafestol ± ein neues Diterpen im Kaffee ± Methoden zur Bestimmung des 16-O-Methylcafestols in Rohkaffee und in behandelten Kaffees. Z. Lebensm. Unters.Forsch., 189, 326±30. Speer, K. (1991) 16-O-methylcafestol ± a new diterpene in coffee; the fatty acid esters of 16-O-methylcafestol. In: Proc. Euro Food Chem. VI, Hamburg, Germany, Vol. 1, (eds W. Baltes., T. Eklund, R. Fenwick, W. Pfannhauser, A. Ruiter & H.-P. Thier) pp. 338±42. Behr's Verlag, Hamburg. Speer, K. (1995) Fatty acid esters of 16-O-methylcafestol. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 224±31. ASIC, Paris, France. Speer, K., Hruschka, A., Kurzrock, T. & KoÈlling-Speer, I. (2000) Diterpenes in coffee. In: Caffeinated Beverages, Health Benefits, Physiological Effects, and Chemistry (eds T. H. Parliament, C-T. Ho & P. Schieberle), pp. 241±51. ACS symposium series 754.
Chemistry I: Non-volatile Compounds: Lipids
Speer, K. & Mischnick, P. (1989) 16-O-Methylcafestol ± ein neues Diterpen im Kaffee ± Entdeckung und Identifizierung. Z. Lebensm. Unters.-Forsch., 189, 219±22. Speer, K. & Mischnick-LuÈbbecke, P. (1989) 16-O-Methylcafestol ± ein neues Diterpen im Kaffee. Lebensmittelchemie, 43, 43. Speer, K. & Montag, A. (1989) 16-O-Methylcafestol ± ein neues Diterpen im Kaffee ± Erste Ergebnisse: Gehalte in Roh- und RoÈstkaffees. Dtsch. Lebensm.-Rundsch., 85, 381±4. Speer, K., Tewis, R. & Montag, A. (1991a) 16-O-Methylcafestol ± a quality indicator for coffee. In: Proceedings of the 14th ASIC Colloquium (San Francisco), pp. 237±44. ASIC, Paris, France. Speer, K., Tewis, R. & Montag, A. (1991b) 16-O-Methylcafestol ± ein neues Diterpen im Kaffee ± Freies und gebundenes 16O-Methylcafestol. Z. Lebensm. Unters.-Forsch., 192, 451±4. Speer, K., Tewis, R. & Montag, A. (1991c) A new roasting component in coffee. In: Proceedings of the 14th ASIC Colloquium (San Francisco), pp. 615±21. ASIC, Paris, France. Speer, K., Sehat, N. & Montag, A. (1993) Fatty acids in coffee. In: Proceedings of the 15th ASIC Colloquium (Montpellier), pp. 583±92. ASIC, Paris, France. van der Stegen, G.H.D. (1979) The effect of dewaxing of green coffee on the coffee brew. Food Chem., 4, 23±9. van der Stegen, G. H. D. & Noomen, P. J. (1977) Mass-balance of carboxy-5-hydroxytryptamides (C-5-HT) in regular and treated coffee. Lebensmittelwiss. Technol., 10, 321±3. Streuli, H. (1970) Kaffee. In: Handbuch der Lebensmittelchemie VI (ed J. SchormuÈller), pp. 19±21. Springer Verlag, Berlin. Streuli, H., Schwab-van BuÈren, H. & Hess, P. (1966) Methodik der Fettbestimmung in Roh- und RoÈstkaffees. Mitt. Geb. Lebensm. Unters. Hyg., 57, 142±6. Studer, A. & Traitler, H. (1982) Quantitative HPTLC determination of 5-hydroxytrytamides of carboxylic acids and tryptamines in food products. J. High Resol. Chromatogr. Chromatogr. Commun., 5, 581±2. Tewis, R., Montag, A. & Speer, K. (1993) Dehydrocafestol and dehydrokahweol ± two new roasting components in coffee. In: Proceedings of the 15th ASIC Colloquium (Montpellier), pp. 880±83. ASIC, Paris, France. Tiscornia, E., Centi-Grossi, M., Tassi-Micco, C. & Evangelisti, F. (1979) Sterol fractions of coffee seeds oil (Coffea arabica L.). Riv. Ital. Sost. Grasse, 56, 283±92. Trouche, M.-D., Derbesy, M. & Estienne, J. (1997) Identification of Robusta and Arabica species on the basis of 16-O-Methylcafestol. Ann. Fals. Exp. Chim., 90, 121±32. Urgert, R., van der Weg, G., Kosmeijer-Schuil, T.G., van der Bovenkamp, P., Hovenier, R. & Katan, M.B. (1995) Levels of the cholesterol-elevating diterpenes cafestol and kahweol in various coffee brews. J. Agric. Food Chem., 43, 2167±72.
49
Viani, R. & Horman, I. (1975) Determination of trigonelline in coffee. In: Proceedings of the 7th ASIC Colloquium (Hamburg), pp. 273±8. ASIC, Paris, France. Vitzthum, O.G. (1976) Chemie und Bearbeitung des Kaffees. In: Kaffee und Coffein (ed. O. Eichler), pp. 3±64. Springer Verlag, Berlin, Heidelberg, New York. Wahlberg, I., Enzell, C.R. & Rowe, J.W. (1975) Ent-16-kauren19-ol from coffee. Phytochemistry, 14, 1677. Wajda, P. & Walczyk, D. (1978) Relationship between acid value of extracted fatty matter and age of green coffee beans. J. Sci. Food Agric., 29, 377±80. Wettstein, A., Spillmann, M. & Miescher, K. (1945) Zur Konstitution des Cafesterols 6. Mitt. Helv. Chim. Acta, 28, 1004±13. Weusten Van der Wouw, M.P.M.E., Katan, M.B., Viani, R. et al. (1994) Identity of the cholesterol-raising factor from boiled coffee and its effect on liver function enzymes. J. Lipid Res., 35, 721±33. White, D.R. (1995) Coffee adulteration and a multivariate approach to quality control. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 259±66, ASIC, Paris, France. Wilson, A.J., Petracco, M. & Illy, E. (1997) Some preliminary investigations of oil biosynthesis in the coffee fruit and its subsequent re-distribution within green and roasted beans. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 92±9. ASIC, Paris, France. Wurziger, J. (1963) L'huile du cafe vert et du cafe torreÂfieÂ. CafeÂ, Cacao, TheÂ, 7, 331±40. Wurziger, J. (1971a) Neuentdeckte Kaffee-Inhaltsstoffe. Ihre Bedeutung fuÈr die BekoÈmmlichkeit von KaffeegetraÈnken. Med. Heute, 22, 10±13. Wurziger, J. (1971b) CarbonsaÈure-5-hydroxy-tryptamide zur Beurteilung von frischen und bearbeiteten Kaffees. In: Proceedings of the 5th ASIC Colloquium (Lisbon), pp. 383±7. ASIC, Paris, France. Wurziger, J. (1972) CarbonsaÈuretryptamide oder aÈtherloÈsliche Extraktstoffe um Nachweis und zur Beurteilung von bearbeiteten bekoÈmmlichen RoÈstkaffees. Kaffee- und Tee-Markt, 22, 3±11. Wurziger, J. (1973) CarbonsaÈurehydroxytryptamide und Alkalifarbzahlen in Rohkaffees als analytische Hilfsmittel zur Beurteilung von RoÈstkaffee-Genuûwert und BekoÈmmlichkeit. In: Proceedings of the 6th ASIC Colloquium (Bogota), pp. 332±42. ASIC, Paris, France.
Chapter 2
Chemistry II: Non-volatile Compounds, Part II S. Homma Ochanomizu University Tokyo, Japan tidine were quantitatively determined in robusta and arabica beans. The sum of these minor amino acids is, on average, 2.8% of the total concentration of free amino acids for arabica green coffee, and 1.9% for robusta green coffee. Arnold and Ludwig (1996) determined amino acid content changes in green coffee beans after the processing in the post-harvest period, such as drying, fermentation and storage (1 to 3 months). Drying of freshly harvested coffee beans at different temperatures of 208C and 408C alters the content of free amino acids: the concentration of glutamic acid increased in all the samples by about 500 mg/kg dry basis (db), while aspartic acid decreased in five out of seven samples by 110 to 780 mg/kg (db). Hydrophobic amino acids such as valine, phenylalanine, leucine and isoleucine significantly increased for most of the samples by 50 mg/kg db. When the adhering pulp was removed from the coffee beans by fermentation (27 h at 20±308C, 32 h at 6±408C) before subsequent drying, the concentrations of most amino acids did not change significantly compared with drying alone. Steinhart and Luger (1995) investigated the effect of steam treatment on free and total amino acids in green coffee beans. The free amino acids decreased significantly during the steam treatment; in particular, glutamic acid, asparagine, arginine, leucine, phenylalanine, tryptophan and lysine decreased remarkably with the duration of steam treatment at 0.8 bar. The decrease in the total amount of free amino acids was greater in arabica than in robusta coffee beans. The protein-bound amino acids amounted to 95% of the original value after 1 hour of steaming and to 80±85% after 4 hours of steaming at 1008C with saturated steam. This decrease was observed in all arabica and robusta coffee beans. Therefore, the industrially steamed coffee bean is roasted with about 90% of the original protein-bound amino acids and only half of the original free amino acids.
The desirable color, aroma and taste of brewed coffee are formed by the roasting process that is applied to the green coffee beans. The major reactions involved that occur during roasting are the Maillard reaction and oxidative polymerization or degradation of phenolics compounds. This chapter refers to non-volatile components with potential to contribute to coffee brew quality, such as minor constituents and compounds with bitter-tasting, antioxidative and metal-chelating activities in green and roasted coffee beans.
2.1 AMINO ACIDS AND PROTEIN Amino acids are involved in the formation of flavor and color of coffee brew; both quantity and types of amino acids affect the intensity and quality of aroma. Since free amino acids in coffee beans are largely transformed by roasting, resulting in negligible amounts in the roasted coffee, the presence of free amino acids in coffee beans after the harvesting of coffee berries should be checked at each processing step of coffee beans.
2.1.1 Amino acids Arnold et al. (1994) developed analytical methods for the determination of free and total amino acid content in green coffee beans by extracting milled samples of green coffee beans with 5-sulphosalicylic acid solution and precolumn derivatization of amino acids in the extract with 9-fluorenylmethylchloroformate reagent. They showed that arabica and robusta coffee beans consisted of the same main and minor amino acids also reported by Macrae (1985) from the work of Thaler. The minor free amino acids such as ornithine, hydroxyproline, b-alanine, pipecolic acid and 3-methylhis50
Chemistry II: Non-volatile Compounds, Part II
2.1.2 Amino acid derivatives Minor constituents such as hydroxycinnamic acid and caffeoyl derivatives of amino acids have been isolated from green coffee beans. Caffeoyltryptophan, which has been identified in green robusta coffee samples (Morishita et al., 1987), has been detected in commercial coffee brands by mass spectroscopy (MS) and UV. According to Balyaya and Clifford (1995) caffeoylL-tryptophan and caffeoyl-L-tyrosine are only found in robusta coffees. The botanical and geographical distributions of p-coumaroyl-L-tryptophan (Murata et al., 1995), the content of which in green coffee beans is 30 mg/kg, need to be investigated. Unidentified components with characteristic absorption spectra in the range of 270±350 nm were observed by threedimensional HPLC analysis in green coffee beans, and it is generally speculated that they are caffeoyl, feruloyl, or p-coumaroyl derivatives (Clifford et al., 1989a). The systematic analyses of coffee beans provide data to assist botanical and geographical classification.
2.1.3 Protein Information relating to the protein and amino acid content of green coffee was fully reviewed by Macrae in 1985. Protein in coffee bean has been little investigated despite its involvement in flavor and color formations in the roasting process. Compared with free amino acids, protein-bound amino acids seem to be rather inert in the reactions during roasting, as shown in their rapid degradation in the roasting process. Nevertheless, recent reports suggest that proteins or peptides may contribute to the formation of aroma, bitter taste and metal-chelating compounds in coffee brews. Rogers et al. (1997) undertook biochemical and molecular characterization of the major storage protein in Coffea arabica bean endosperm by molecular genetic studies. The endosperms proteins were analyzed by two-dimensional electrophoresis and amino acid micro-sequencing. The principal bean storage protein has been characterized, and a full length 1706 bp cDNA coding for this protein has been identified. The protein bears a strong sequence homology to the 11S storage protein of soybean seeds, which supports the assumption of a storage function for this protein. This protein accounts for approximately 50% of total proteins in the endosperm, representing between 5% and 7% of coffee bean dry weight, and exists in vivo as a mature coffee flavor precursor of approximately 52 kilo daltons (kDa) (see also Chapter 11). Ludwig et al. (1995) investigated the reactivity of a
51
coffee protein isolate. The protein isolate was prepared by aqueous extraction from the matured green beans of Columbian Coffea arabica, with a yield of 45% total coffee protein. The protein isolate consists of 15% albumin and 85% globulin, without color, taste or smell. The protein isolate contained small amounts of sugars such as galactose, arabinose, rhamnose and glucose. The molecular masses of the reduced protein monomers were in the range of 20±30 kD for major bands and 30±45 kD for minor bands, by SDS PAGE. The main bands were electrofocused between pH 5 and 7. The major amino acids were glutamine/glutamic acid (28.2 mol%), glycine (8.8%) and asparagine/ aspartic acid (7.8%). The major N-terminal amino acids were glycine (36%), glutamic acid (15%) and aspartic acid (10%). After heating at 2008C, simulating coffee roasting, the protein isolate showed a yellow to light brown color with a light roasted taste and smell. Almost half of the protein (45±46%) was dissolvable in water at room temperature. Amino acid decomposition was not strong, whereas peptides with a molecular mass smaller than 10 kD were formed.
(a) Reactivity of protein Protein in green coffee beans has been regarded as being rather labile in reactions occurring on heating, compared with free amino acids. The rapid decomposition of free amino acids in green coffee beans upon roasting supports the rapid reactivity of free amino acids. Nevertheless, recent report and communications suggest possible contributions of protein or peptide to the formation of aroma, bitter taste and metal-chelating compounds in coffee brew. Several workers have investigated the reactivity of the E-amino group of lysine, and sulfhydryl and methylthio groups on side chains of amino acids in the protein, compared with those same functional groups in the free amino acids present.
Reactivity of the lysine side chain According to Hofmann et al. (1999a,b), as well as ionic condensation reactions, mechanisms involving amineassociated oxidative carbohydrate fragmentation and free radical formation also produce colored compounds by the Maillard reaction, prior to the Amadori rearrangement. They designed a series of model experiments in order to test the reactivity of the E-amino group of lysine in protein. Color development in the thermal treatment of neutral aqueous solutions of alanine and carbohydrate
52
degradation products at 958C was studied; glycolaldehyde was shown to lead to the most effective color development (Table 2.1) accompanied by intense radical formation. The radicals were also detected in heated mixtures of L-alanine and pentose or hexose, respectively, and were identified as 1,4-dialkylpyrazinium radical cations. Under the reaction conditions applied, glyoxal is formed as an early product in hexose/L-alanine mixtures prior to radical formation. Reductones then initiate radical formation upon reduction of glyoxal and/or glyoxal imines, formed upon reaction with amino acid, producing glycolaldehyde. The thermal treatment of neutral aqueous solutions of glucose and N-a-acetyl-L-lysine, a model substance of lysine side chains of proteins, generated 1,4-bis[5(acetylamino)-5-carboxyl-1-pentyl]-pyrazinium radical cations accompanied by intense browning development. The thermal treatment of neutral aqueous solutions of bovine serum albumin (0.05 mmol) and glycolaldehyde (1.25 mmol) at 958C for 5 minutes generated 1,4-bis[5-amino-5-carboxy-1-pentyl]-pyrazinium radical cation (CROSSPY) as a cross-linker leading to protein dimerization (Fig. 2.1). In order to verify the formation of CROSSPY in foods, wheat bread crust and roasted cocoa as well as coffee beans during intense non-enzymatic browning, were investigated by electron spin resonance (ESR). An intense radical was detected; it was identified as the protein-bound CROSSPY by comparison with the
Coffee: Recent Developments
Fig. 2.1 Structure of protein-cross-linking amino acid 1,4-bis(5-amino-5-carboxy-1-pentyl)pyrazinium radical cation (CROSSPY) (Hofmann, 1999b).
radical formed upon reaction of bovine serum albumin with glycolaldehyde. The E-amino group on the lysine of proteins participates in radical-associated non-enzymatic browning reactions during the thermal processing of foods.
Reactivity of sulfhydryl and methylthio groups Sulfur-containing amino acids in coffee beans are known to be involved in the formation of characteristic aroma compounds such as furan derivatives like 2furfurylthiol and 2-methyl-3-furanthiol during roasting (Grosch, 1995). Generally, the reactivity of sulfurcontaining amino acids has been considered to be
Table 2.1 Color development (CD) and radical formation in binary mixtures of L-alanine and carbohydrates and carbohydrate degradation products, respectively (Hofmann, 1999a). Carbonyl compound Glucose Xylose N-(1-deoxy-D-fructos-1-yl)-L-Alanine Glycolaldehyde Glyoxal Furan-2-carboxaldehyde Pyrrol-2-carboxaldehyde 2-Oxopropanal Butane-2,3-dione 5-(Hydroxymethyl)furan-2-carboxaldehyde Glycerinaldehyde 2-Hydroxy-3-butanone 1
CD factor1
Rel radical formation (%)
16 64 8 1024 128 1024 256 256 128 2 2 2
42 82 12 1003 43 03 03 03 NA NA NA NA
The CD factor was applied to compare the color intensities of the reaction mixtures, which were heated for 15 minutes at 958C. For EPR measurements the mixture was heated for 10 minutes at 958C. 3 For EPR measurements the mixture was heated for 2 minutes at 958C. NA: not analyzed. 2
Chemistry II: Non-volatile Compounds, Part II
greater when they are in the free form than when they are bound in protein. Rizzi (1999) compared reactivity during thermal treatment, between the free and bound forms (with Nacetyl or peptide bond) according to the amount of production of sulfur-containing compounds produced. Furfuryl alcohol or 5-methylfurfuryl alcohol was allowed to react with cysteine, methionine, and peptides with these sulfur amino acids and their N-acetyl derivatives in pH 4 acetate buffer solution at 1008C to simulate the protein reaction in the initial stage of coffee bean roasting. It was predicated that N-acetyl amino acids should be prone to net positive charge at the sulfur atom on the peptide, since they are uncharged molecules in pH 4 solution as amides. Nacetyl cysteine (AcCys) and furfuryl alcohol produced more 2-furfurylthiol (7.9% of total volatiles, TV) compared with that from free cysteine and furfuryl alcohol (trace of TV). In addition, similar reactions with 5-methylfurfuryl alcohol gave 5-methylfurfurylthiol (11% TV). Also, N-acetylmethionine with 2furfurylalcohol and 5-methylfurfuryl alcohol produced methyl furfuryl sulfide (1.4% TV) and 5-methylfurfuryl methyl sulfide (0.03% TV), respectively. As expected, the yield of methyl furfuryl sulfide was higher with N-acetylmethionine when compared with free methionine (0.05% TV). The cysteine tripeptide, glutathione, reacted with furfuryl alcohol to produce more 2-furfurylthiol (2.7% TV) than free cysteine, suggesting that protein-bound cysteine could function as a direct precursor of 2-furfurylthiol during heating of foods. The methionine dipeptide, glycyl-methionine, failed to generate methyl furfuryl sulfide from furfuryl alcohol, apparently as the value of its pK1 is greater than the 2.38 reported for methionine. A higher pK1 will lead to more net positive charge on the peptide at pH 4 and therefore reduced reactivity (nucleophilicity) at the sulfur atom. Coffee bean protein would provide biogenetically determined numbers and locations of sulfhydryl and methylthio residues, which could offer more control over aroma formation compared with reactions of free amino acids. Reducing sugars in coffee glycoprotein may be characteristically positioned during roasting to interact with basic and sulfur-containing amino acid residues on protein. Rizzi suggested that proteins might play a key role in aroma and melanoidin formation in coffee beans as well as in processed foods, because of the higher reactivity of E-amino, thiol or methylthio groups, and the higher content of protein compared to free amino acids.
53
(b) Bitter tasting componds The bitter taste of coffee brew has always interested food chemists. Since caffeine tastes bitter, it has been regarded as the major contributor to the bitter taste of coffee brew. Nevertheless, on the basis of sensory evaluation of caffeine concentration in coffee brews, the caffeine concentration only accounts for some 10±30% of the bitter taste of coffee brews (Macrae, 1985). The simple fact that decaffeinated instant coffee also tastes bitter suggests that substances other than caffeine might be contributing to the bitter taste. Other bittertasting compounds in coffee brew considered as candidates are trigonelline, polyphenolic compounds such as chlorogenic acids and melanoidin or polymeric compounds. Ginz and Engelhardt (2000) have suggested that bitter tasting compounds might be formed by roasting protein. An 80±90% protein-rich isolate was prepared from green coffee beans. It was roasted in a model roaster, and the bitter tasting hot water extract was subsequently fractionated. The presence of the cyclic peptides cyclo(Pro-Val), cyclo(Pro-Pro), cyclo(ProLeu) and cyclo(Pro-Phe) was identified. The cyclic dipeptides isolated are diketopiperazines, which are known to taste bitter in beer (Gautschi et al., 1997) and in cocoa (Pickenhagen et al., 1975). The threshold concentration of these cyclic dipeptides for bitter taste is reported as 10±50 ppm, and a synergistic effect of the cyclic peptides with theobromine on the bitter taste has been reported in cocoa (Pickenhagen et al., 1975). It is anticipated that cyclic dipeptides with a bitter taste will soon be isolated from coffee brews. The roasting of coffee beans accelerates the Maillard reaction, as well as the oxidative polymerization of chlorogenic acids. Hofman (1999) identified bittertasting bispyrrolidino- and pyrrolidinohexose reductones formed by roasting an equimolar mixture of powdered glucose and L-proline at 1808C for 15 minutes. Sensory evaluation of the dry-heated mixture revealed a strong bitter taste, as already reported by Papst et al. (1984, 1985) using heated mixtures of sucrose and proline. The development of the bitter taste by roasting was drastically reduced by adding Lcysteine to the mixture of glucose and proline. Hofmann (1999) identified acetylformoin derived from hexose as the precursor of the aminohexose reductones, which reacted more easily with cysteine to form 7-hydroxy-4a,6-dimethyl-2H,3H,4aH-furo[2,3-b]thiazine than with L-proline to form the aminohexose reductones. The addition of L-cysteine was found to block the development of the bitter taste (Fig. 2.2).
54
Coffee: Recent Developments
roasting on the production of characteristic flavors in roasted coffee beans. Compared with the caffeic acid moiety in chlorogenic acid, the quinic acid moiety might be considered rather inert for oxidative changes during the roasting process.
2.2.1 Quinic acid moiety
Fig. 2.2 Blocking of formation of bitter-tasting bispyrrolidinohexose reductone 1 and pyrrolidinohexose reductone 2, by formation of 7hydroxy-4a,6-dimethyl-2H,3H,4aH-furo[2,3-b] thiazine 4 upon reaction of acetylformoin 3 with Lcysteine (Hofmann 1999).
Fig. 2.3
(±)-Quinic acid occurs free or bound in the chlorogenic acids of green coffee beans, and the roasting makes bound (±)-quinic acid, free, producing stereoisomers and quinides. Scholz-BoÈttcher and Maier (1991) determined the isomers of quinic acid (Fig. 2.3) and quinides (Fig. 2.4) produced in roasted coffee beans of Salvador arabica. Using GC/MS, five further quinic acids and seven quinides were identified, including (+)-g-quinide.
Stereoisomers of quinic acid (from Schloz-BoÈttcher & Maier, 1991).
Hofmann (personal communication) has suggested that these aminohexose reductones may be present in roasted coffee beans. It follows that the bitter taste may be developed by the combination of caffeine, trigonelline and phenolic compounds originally present in green coffee beans with compounds formed by roasting, such as cyclic peptides and aminohexose reductones. The evidence for their presence in coffee brews and the degree of contribution to the bitter taste, are involved in their current work.
2.2 FATE OF CHLOROGENIC ACID DERIVATIVES DURING ROASTING Information relating to chlorogenic acids in green and roasted coffee beans was reviewed by Clifford in 1985. The fate of chlorogenic acids is still one of the major factors to be considered when assessing the effect of
Fig. 2.4 Quinic acid lactones (from Scholz-BoÈttcher & Maier, 1991).
Chemistry II: Non-volatile Compounds, Part II
55
g/kg green coffee dmb
The changes in the quinic acids and quinides contents during the roasting process are shown in Fig. 2.5. (+)Quinic acid slightly increases at the beginning of the roasting process and remains relatively constant up to higher degrees of roast. The amount of (+)-g-quinide increases in light to medium roasts and decreases slightly in dark roasts. In contrast to (+)-quinic acid and its quinide, all other stereoisomeric acids and lactones continuously increase with higher roasting temperature, and the increasing rate of each product differs. The amount of (+)-epi-quinic acid rises with an approximately linear increase at higher roasting degrees. (+)-epi-g-Quinide and (+)-epi-d-quinide show a strong increase at higher roasting loss. Only scyllo-quinic acid and its quinide were found in light roasted coffees, and nearly all stereoisomeric quinic acids and quinides were generated at medium degrees of roast. The higher the degree of roast, the higher the contents of the isomeric compounds. The major stereoisomers in roasted coffee are scyllo- and mesoquinic acid I for quinic acid and scyllo-d- and (+)-epig-quinide for quinides.
acid and corresponding lactones. The presence of feruloylquinic acid lactones was also reported (Wynnes et al., 1987). Both 3- and 4-caffeoylquinic acid-g-lactones were identified in roasted coffee beans by Bennat et al., (1994) (Fig. 2.6). The content of the caffeoylquinides in coffee beans after different degrees of roasting produced from green arabica coffee ranged from 1.5 to 3.5 g/kg dry matter. The formation of lactones reaches a maximum in medium roasted coffee. A higher degree of roasting reduces the contents, suggesting decomposition of chlorogenic acid lactones. Bennat et al. (1994) also considered that the content of the lactones in an instant coffee sample was very low, because of hydrolysis during the extraction process.
Fig. 2.6 Structure of 4-caffeoyl-g-quinide (from Bennat et al., 1994).
12 10 8 6 4
(+_)–quinic acid
2
(+_)–γ–quinide
0
0
2
4
6
8
10
12
14
16
ORV (%)
Fig. 2.5 Variation of the (+)-quinic acid and the (+)-g-quinide concentrations as a function of the roasting loss on a dry matter basis (ORV) (from ScholzBoÈttcher & Maier, 1991).
On the basis of observations made during the roasting of coffee beans, they proposed the possibility that quinic acids and quinides could be used as an indicator for the degree of roasting: they defined the degree of isomerization as the ratio between the sum of the isomeric quinic acids and quinides (neo-quinides as well as quinide No 1 excluded) and the sum of (+)quinic acid and its quinide. A correlation between the degree of isomerization and the roasting loss is obtained. The content of chlorogenic acids in green coffee beans decreases during the roasting, giving (-)-quinic
Schrader et al. (1996) developed an HPLC system to determine mono- and dicaffeoylquinic acids, corresponding lactones and feruloylquinic acids in roasted coffee with one chromatographic run. The levels of 3and 4-caffeoylquinic acid-g-lactones were found to be 2.1 and 1.0 g/kg dry matter, respectively. Keeping coffee brews at an elevated temperature of 808C reduced the amount of caffeoylquinic acid lactone to 60% of its initial value. The change in these chlorogenic acid lactones and quinic acid lactones is involved with the generation of a bitter, sour taste in a coffee brew after it has stood for some hours on a hot plate. The patent of Bradbury et al., (1998) refers to the hydrolysis of chlorogenic acid lactones into quinic acid and caffeic acid, as well as to the hydrolysis of quinic acid lactones. These acids contribute significantly to the increase in acidity in a coffee brew. Table 2.2 shows the change in organic acids with time in stored coffee brew. A standard coffee solution prepared from Colombian beans was stored at 608C, the development of acid levelled off after about 200 hours, and the pH dropped from about 4.9 to 4.5. The concentration of quinic acid increased by 14.8 mmol/kg while the quinic acid lactone concentration decreased by 12.2 mmol/kg, suggesting that
56
Table 2.2
Coffee: Recent Developments
Change in organic acids in stored coffee brew (from Bradbury et al., 1998).
Acid
Time (hours)
(g/kg) Quinic Acetic Glycolic Formic Malic Citric Phosphoric
0
2.5
8
24
72
120
7.8 3.15 1.14 2.0 2.09 6.6 1.44
8.7 3.6 1.29 2.10 2.19 6.9 1.50
8.7 3.6 1.23 2.13 2.16 6.9 1.53
9.0 3.6 1.25 2.19 2.40 6.9 1.59
9.9 3.9 1.23 2.22 2.22 6.9 1.71
10.8 3.9 1.32 2.28 2.19 6.9 1.83
these lactones represent the primary precursors to the acids which develop upon storage.
2.2.2 Cinnamic acid derivative moiety Chlorogenic acids in roasted coffee beans are probably first hydrolyzed to quinic and caffeic acids, which then undergo pyrolysis to form phenolic volatiles. For the cinnamic acid, decarboxylation producing vinylphenols such as 4-vinylguaiacol is a commonly reported mechanism. The quantity of 4-vinylphenols formed in roasted coffee beans is small relative to the amount of the chlorogenic acid consumed (Heinrich & Baltes, 1987a), suggesting a competitive reaction pathway. Rizzi & Boekley (1993) investigated the mechanism of the thermal decomposition of p-hydroxycinnamic acid derivatives for alternative reaction pathways. Each derivative was pyrolyzed at 2078C, and the residual amount of the derivatives and carbon dioxide evolved were determined to estimate the extent of initial vinylphenol production (Table 2.3). The ethyl acetate soluble reaction product was methylated, and analyzed Table 2.3
Thermal decomposition of cinnamic acids at 2078C (from Rizzi & Boekley, 1993).
Cinnamic acid or ring derivative Cinnamic acid 3-Hydroxy 4-Hydroxy 3,4-Dihydroxy 3,4-Dihydroxy (K salt) 3-Hydroxy-4-methoxy 4-Hydroxy-3-methoxy 4-Methoxy 1 2
by GC/MS. Cinnamic acids with p-hydroxysubstituents, such as p-coumaric, caffeic and ferulic acids, were inclined to decomposition, disappearing completely in 45 minutes. The effect of a single phydroxy group is to accelerate decomposition relative to unsubstituted cinnamic acid (48%). p-Hydroxycinnamic acids which are further substituted with hydroxyl or methoxy groups are susceptible to relatively easy decomposition, for example caffeic and ferulic acids. Since the p-hydroxy derivatives exist in equilibrium with trienone isomers, they undergo easy loss of carbon dioxide, producing a 75±97% yield of carbon dioxide, indicating initially high yields of vinylphenols (Fig. 2.7). The final decomposition products of p-hydroxycinnamic acid were largely polymeric, resulting from rapid polymerization. The major products of cinnamic acid decomposition were vinylphenol dimers, that is 1,3-bis-arylbutenes with five isomeric forms depending on position and Z/E orientation of the olefinic bond.
Common name
p-Coumaric acid Caffeic acid Hesperitic acid Ferulic acid
Decomposition1 (%)
CO2 % yield2
48 24 > 99 > 99 Ð 0 > 99 22
Ð Ð 89 75 55 Ð 97 Ð
HPLC estimate of starting material lost after 45 minutes. Per cent of theoretical based on weight of BaCO3 isolated in 160 minutes (with the exception that ferulic acid was heated for 100 minutes).
Chemistry II: Non-volatile Compounds, Part II
57
Fig. 2.7 Proposed mechanism for cinnnamic acid decarboxylation (from Rizzi & Boekley, 1993).
2.3 ANTIOXIDATIVE COMPOUNDS IN COFFEE BREW Recent interest in the coffee beverage has been focused on the antioxidative activity of a roasted coffee brew (Turesky et al., 1993). This interest is not restricted to antioxidative usage in food systems, but also relates to the function that protects cells from oxidative damage in a biological system. The antioxidative activity in a coffee brew depends on natural constituents such as phenolic compounds, as well as reacted compounds formed by roasting (see also Chapter 8 on health and safety aspects).
2.3.1 Compounds occurring naturally in green beans Morishita and Kido (1995) have reported the potent contribution of chlorogenic acids to the antioxidative activity of coffee by using a 1, 1, diphenyl-2-picryl hydazil (DPPH) radical scavenging system and superoxide anion-mediated linoleic acid peroxidation system in vitro. Ohnishi et al. (1998) showed that DPPH radical scavenging activity of caffeoyltryptophan, a minor constituent in green coffee beans, increased dose-dependently at concentrations ranging from 1 to 50 mM. Caffeoyltryptophan inhibited the formation of conjugated diene from linoleic acid with the inhibitory activity in the increasing order of caffeic acid < 5-caffeoylquinic acid < caffeoyltryptophan < dla-tocopherol. They also examined the effects of caffeoyltryptophan on the in vitro haemolysis and
peroxidation of mouse erythrocytes induced by hydrogen peroxide. Caffeoyltryptophan exhibited strong inhibitory activities, suggesting that caffeoyltryptophan may be a natural antioxidant in the human diet and may intervene in toxicological processes mediated by radical mechanisms. Nakayama (1995) has shown that caffeic acid enhanced hydroxyl radical formation in the presence of transition metal ions such as Fe3 , Cu2 and Mn2 that cause oxidative damage, while caffeic acid esters showed protective effects in the absence of the metal ions. Devasagayam et al. (1996) showed caffeine at a millimolar concentration to be an effective inhibitor of oxidative damage to rat liver microsomes induced by the hydroxy radical, peroxy radical and singlet oxygen. It was speculated that caffeine might quench these reactive species, which suggests one more positive attribute from a daily intake of caffeine. According to Stadler et al. (1996a), caffeine and related methylxanthines were changed to the corresponding C-8 hydroxylated analogues as the major products of hydroxyl radical mediated attack. Further oxidation products of caffeine were found to be the N-1-, N-3and N-7-demethylated methylxanthine analogues, theobromine, paraxanthine and theophylline, respectively. It is generally recognized that hydrogen peroxide is detectable in roasted coffee beans. Since many factors may be involved in hydrogen peroxide formation in roasted coffee, caffeic, chlorogenic and quinic acids were pyrolyzed as model systems. The pyrolyzed caffeic acid catalyzed the highest levels of hydrogen peroxide formation over time in the presence of Mn2 (Table 2.4). The novel tetraoxygenated phenylindan isomers, 1,3-cis- and 1,3-trans-tetraoxygenated phenylindans, were identified as the major products in both the caffeic acid pyrolyzate at 2308C. The combined yield of both isomers was 3.6%. The acidcatalyzed cyclization of caffeic acid also gave a com-
Table 2.4 Hydrogen peroxide formation (micromolar) of pyrolyzed chlorogenic acids over time and effect of addition of Mn2 (from Stadler et al. 1996b). Caffeic acid Incubation time (min) 0 30 60
Chlorogenic acid
Quinic acid
-Mn2
+Mn2
-Mn2
+Mn2
±Mn2
+Mn2
11 + 1.3 33 + 2 64 + 2
15 + 1.0 104 + 2 205 + 5
6 + 1.2 10 + 3 17 + 1
16 + 2.2 24 + 1 41 + 2
3.8 + 0.8 6 + 0.5 3.4 + 0.3
11 + 0.5 15 + 3 18 + 0.7
58
bined yield of 5±6% for both isomers. The significant contribution of these indans to hydrogen peroxide formation has been demonstrated (Stadler et al., 1996b). Vacuum pyrolysis of rosmarinic, chlorogenic and caffeic acids at 2288C for 15 minutes raised their antioxidative activity in a rat liver membrane assay by 4-, 11-, and 460-fold, respectively (Guillot et al., 1996). The antioxidative components, 1,3-cis- and 1,3-transtetraoxygenated phenylindan isomers, were identified only in the caffeic acid pyrolyzates. These indan isomers were 8-fold larger in antioxidative activity than BHT. The potent reducing properties of the phenylindan isomers resulted in a pro-oxidative effect at a relatively high concentration in an ethyl linoleate peroxidation assay, and promoted the hydroxylation of 20 deoxyguanosine to produce 8-oxo-20 -deoxyguanosine. Although the existence of these phenylindans has not yet been confirmed in roasted coffee beans, a comparison of non-roasted, light-roasted, and dark-roasted coffee extracts showed a positive correlation between the degree of roasting and the inhibition of lipid peroxidation in rat liver membranes. The formation mechanism for the phenyl indans suggests the importance of a free carboxylic function in the dimerization process, and it is believed that these tricyclic dimers are formed by acid- or thermally catalyzed decarboxylation of caffeic acid monomers, which rapidly condense to form the phenylindan structure. Antioxidative activity of a coffee brew depends on the presence of compounds ranging from low to macromolecular, already present in raw green beans and those formed in roasting. The active compounds are not only phenolic compounds such as caffeoyltryptophan, caffeine and others naturally occurring in green coffee beans, but also more active compounds such as melanoidin and phenylindans, which might be produced by roasting. The mechanism of antioxidative action in coffee brew is complicated, and all of the antioxidative factors seem to be involved in chelation of transition metals, radical scavenging in chain reactions, trapping of active oxygen, and so on. It would be interesting to find a positive effect of coffee brew constituents on biological systems, in the view of the interets in the health aspects of coffee intake.
2.3.2 Effect of roasting on antioxidative activity An antioxidative effect of roasted coffee on processed foods was investigated by preparing cookies containing roasted coffee bean powder or another test antioxidant
Coffee: Recent Developments
with a 200 ppm level of iron. The change in lipid autoxidation was monitored by a storage test at 408C for 12 months. The coffee bean powder showed a strong antioxidative effect according to the peroxide value of the lipid fraction, while caffeic acid and rosemary extract were slightly less effective (Ochi et al., 1997). Nicoli et al. (1997) prepared hot water extracts from coffee beans after various degrees of roasting, and evaluated the antioxidative activity of the aqueous extracts by their chain-breaking activity and oxygen consumption properties. The chain-breaking activity was measured by croicin bleaching due to the presence of peroxy radicals, and the presence of an antioxidant slowed down the rate of bleaching. The highest antioxidative properties were found in the medium-dark roasted coffee brew (Fig. 2.8).
Fig. 2.8 Oxygen scavenging properties of coffee brews expressed as percentage of oxygen uptake/min per g dry matter (DM) as a function of the roasting time (from Nicoli et al., 1997).
2.4 COLORED MACROMOLECULAR COMPOUNDS 2.4.1 Characterization of colored polymers It has been generally recognized that phenolic polymerization and the Maillard reaction are the major reactions contributing to the formation of the colored polymer, melanoidin, in roasted coffee beans. Evidence of the Maillard reaction in the roasting of coffee beans has been reviewed by a number of authors (Dart & Nursten, 1985; Ho et al., 1993; Maier, 1993; Reinec-
Chemistry II: Non-volatile Compounds, Part II
59
cius, 1995). The formation of furans, pyrazines and aldehydes as aroma constituents supports the degradation of sugars by the Maillard reaction.
(a) Chemical characterization Melanoidin in a hot-water extract from roasted coffee bean has been analyzed by gel filtration column chromatography. The molecular mass of coffee melanoidin was estimated by HPLC using a protein PAK-125 column developed with water. Pullulan was used as the most suited standard markers, which are macromolecular linear polysaccharides. The molecular mass of the coffee melanoidin ranged between 3000 and more than 100 000 daltons, according to the degree of roasting and the coffee species, and increased with longer roasting time. The high molecular weight melanoidin increased in amount in the robusta samples compared to the arabica (Steinhart et al., 1989). The aqueous extract was charged on a Sephadex G-25 column, developed with water and four fractions were separated in order of molecular mass. The fractions were further separated into three or four bands by TLC on Sephadex, and developed with a mixture of 25% ammonia and 1-propanol. Each band was hydrolyzed and analyzed for its sugar composition which varied between the bands, mannose, arabinose, galactose and glucose being the major sugars, and rhamnose the minor (Steinhart & Packert, 1993). Nevertheless, such phenolics as chlorogenic acids are degraded by roasting green coffee beans. A significant quantity of the chlorogenic acid lost during roasting remains in an uncharacterized form. Leloup et al. (1995) monitored the fate of chlorogenic acids during a medium-slow-roasting process at 2408C by kinetic analysis. 5-Caffeoylquinic acid and dicaffeoylquinic acid decreased the most rapidly. With a short roasting time, dicaffeoylquinic acids (diCQA) are partly hydrolysed into caffeoylquinic acid and the caffeic moiety. 5-Caffeoylquinic acid isomers rapidly undergo esterification with carbohydrate and protein, producing bound chlorogenic acids (Fig. 2.9). With a longer roasting time, the phenolic and quinic moieties start rapidly degrading such diverse phenolic components as 4-vinyl catechol and catechol from the phenolic moiety, and slowly degrading hydroquinone, catechol, phenol and pyrogallol from the quinic moiety (Fig. 2.10). Maier (1993) has reviewed melanoidin and the nonvolatile compounds in coffee by referring to the investigations of the analysis of pyrolyzed melanoidin in coffee brew. Heinrich and Baltes (1987b) prepared seven fractions of melanoidins from roasted Robusta
Fig. 2.9 Comparison of esterified quinic and phenolic moieties during roasting (expressed in g/100 g green coffee, dry basis) (from Leloup et al., 1995).
coffee beans. These melanoidins were degraded by Curie point pyrolysis, which was monitored by highresolution GC/MS, to about 100 products, among which were found 33 phenols.
(b) Microbiological characterization It is generally recognized that a dialyzed coffee brew and the separated polymeric fractions show an absorption spectrum, consisting of the general absorption in the visible range characteristics of a model melanoidin (with sugar±amino acid), and a similar spectrum to that of chlorogenic and caffeic acids in the ultraviolet range. Therefore, melanoidin in brewed coffee is considered to be a mixture of both sugar and phenolic type melanoidins, or copolymers of sugar and phenolic type moieties. The microbiological decolorization of brown pigments in foods has been tested in order to categorize the chemical structure of these brown pigments. The fungus Paecilomyces canadensis NC-1 that can decolorize an instant coffee solution has been isolated from a glass bottle containing instant coffee of the freeze-dried type. This glass bottle was left open for 2 weeks. The fungus decolorized the instant coffee solution by 79% under optimal conditions. The decolorized coffee solution was analyzed by gel permeation chromatography, the absorbance being detected at 500 nm. The two chromatograms for the control solution and the decolorized one are compared in Fig. 2.11, which shows that the high molecular weight fractions were decolorized (Terasawa et al., 1994). This strain also decolorized black tea. Streptomyces werraensis TT14
60
Fig. 2.10 1995).
Coffee: Recent Developments
Suggested degradation mechanism of 3,5 dicaffeoylquinic acid during roasting (from Leloup et al.,
Fig. 2.11 Decolorization of coffee by Paecilomyces canadensis NC-1 analyzed by gel permeation chromatography, ____, control; . . . . ., decolorized coffee (from Terasawa et al. 1994).
(Murata et al., 1992) and Coriolus versicolor IFO 30340 (Aoshima et al., 1985) have also been screened from soil by the decolorization rate of model melanoidin prepared from glucose and glycine. Three microorganisms, S. werraensis TT14, C. versicolor IFO 30340 and P. canadensis NC-1, have been cultured to compare the decolorization rate of model brown pigments and brown-colored foods (Terasawa et al., 1996). The resulting decolorization rates are summarized in Tables 2.5 and 2.6. These data could be categorized by the decolorization rate of the brown pigment that differed significantly (Fig. 2.12). Paecilomyces canadensis NC-1 decolorized phenol type brown pigments and is unique in comparison with the other two microorganisms. It follows that the major brown pigment in brewed coffee can be considered to be of the phenol type. The metal-chelating affinity for the brown pigments could be applied to categorize these pigments. Either Zn(II), Cu(II) or Fe(II) was charged into a chelating Sepharose 6B column, and soy sauce, instant coffee and a model melanoidins prepared from glucose and
Chemistry II: Non-volatile Compounds, Part II
61
Table 2.5 Percentage of decrease in brown color intensity after microbial cultivation of synthetic brown pigments (from Terasawa et al. 1996). Synthetic brown pigment Glc-Gly1 Xyl-Gly1 Gal-Gly1 Glc-GABA1 Glc-Lys1 Glc-Trp1 Oxidized2 Reduced2 Catechin3 Chl3 Chl-Suc4 Caramel P Caramel N
S. werraensis TT 14 42.6 + 6.4 70.4 + 7.2 40.1 + 12.6 1.2 + 4.0 72.9 + 1.6 40.5 + 14.5* 46.6 + 3.9 29.1 + 16.3 5.8 + 22.2 796.2 + 10.8 733.6 + 13.4 737.4 + 7.5* 70.8 + 4.7*
C. versicolor IFO 30340 68.1 + 1.5* 63.0 + 4.4 70.7 + 2.1* 58.7 + 4.3* 51.0 + 12.2 74.1 + 27.3 71.1 + 1.5* 58.5 + 6.3* (84.7 + 11.7)5 7140.0 + 30.0 730.1 + 11.3 14.2 + 14.3 34.4 + 5.8
P. canadensis NC-1 36.9 + 5.3 58.3 + 1.0 35.9 + 0.3 32.9 + 11.0 28.4 + 4.7* 94.5 + 0.2 35.9 + 5.2 23.8 + 13.1 95.9 + 1.0* 32.1 + 3.0* 42.9 + 2.3* 2.2 + 36.4 47.7 + 13.6
Values are shown as mean + SD. 1 Sugar and amino acid model melanoidins. 2 The Glc±Gly melanoidin was oxidized with K3 [Fe(CN)6 ] and reduced with NaBH4 . 3 Catechin and chlorogenic acid were oxidized with KI03 . 4 Chlorogenic acid and sucrose were heated at 2208C for 40 min. 5 The pigment was adsorbed on the surface of the mycelia. 6 The decrease in brown color intensity was expressed as a percentage of the decrease in A500 vs the control after microbial cultivation for 10 days at 278C for P. canadensis NC-1, 5 days at 378C for S. werraensis TT 14, and 10 days at 278C for C. versicolor IFO 30340. * Significantly different from the decolorization rate of the other two microorganisms (P 0.05).
Table 2.6 Percentage of decrease in brown color intensity after microbial cultivation of browned foods (from Terasawa et al. 1996). Browned food Cane molasses Soy sauce Miso Caramel A Caramel B Dark beer Cola Barley tea Instant coffee Black tea Worcestershire sauce A Worcestershire sauce B Worcestershire sauce C Cocoa Chocolate
S. werraensis TT 14 4.0 + 9.1 728.7 + 16.9 751.4 + 13.5 72.3 + 9.1 15.8 + 6.5 26.0 + 6.2 54.8 + 15.4 20.3 + 4.0 7100.0 + 61.4 768.9 + 16.4 741.9 + 12.3* 724.2 + 13.0 748.0 + 26.7* 76.9 + 7.3 766.2 + 31.2*
C. versicolor IFO 30340
P. canadensis NC-1
72.5 + 4.7* 66.6 + 1.6* 64.7 + 2.3* 51.8 + 2.1 36.2 + 4.3 45.9 + 8.6* 40.8 + 15.9 60.0 + 8.3* 713.2 + 11.0 7109.0 + 32.7 61.8 + 9.2 68.1 + 5.6* 69.4 + 3.1 87.7 + 4.0 79.5 + 11.4
719.0 + 11.5 18.3 + 26.3 753.0 + 16.5 14.6 + 3.0* 25.7 + 4.5 2.6 + 7.3 46.2 + 13.0 27.6 + 10.3 61.8 + 7.7* 58.1 + 8.4* 60.0 + 10.2 33.3 + 35.3 64.5 + 7.0 82.3 + 21.6 86.8 + 4.4
Values are shown as mean + SD. 1 The decrease in brown color intensity was expressed as a percentage of the decrease in A500 vs the control after microbial cultivation for 10 days at 278C for P. canadensis NC-1, 5 days at 378C for S. werraensis TT 14, and 10 days at 278C for C. versicolor IFO 30340. * Significantly different from decolorization rate of the other two microorganisms (P 0.05).
62
Coffee: Recent Developments
C. versicolor Model melanoidins Cane molasses Soy sauce Miso Dark beer Barley tea Worcestershire sauce B Model caramel N Chocolate
Glc–Lys melanoidin
S. werraensis
Commercial caramel A Cocoa Cola Xyl–Gly melanoidin
Glc–Trp melanoidin Worcestershire sauce A and C Phenol-type model pigments Instant coffee Black tea
P. canadensis
Fig. 2.12 Categorization of synthetic brown pigments and browned foods by statistical significance of microbial decolorization (from Terasawa et al., 1996).
glycine were chromatographed in these metal-chelating Sepharose columns. The brown pigment of soy sauce is typical of melanoidin formed by the reaction of a reducing sugar with amino acids and peptides. The model and soy sauce melanoidins showed weak affinity for the Fe(II) column, while the coffee pigment showed strong affinity, being eluted with EDTA. Column chromatography of instant coffee in the Fe(II)-chelating Sepharose 6B column supports the proposition that the major brown-colored components of brewed coffee are of the phenol type (Homma & Murata, 1995).
2.4.2 Characterization of the zincchelating compounds in coffee brews (a) Effect of coffee intake on the biological availability of minerals The effect of brewed coffee on mineral nutrition has recently been noticed. Brewed coffee has metalchelating activity which results in a trace element deficiency in those known geographical areas of malnutrition, where coffee is consumed in a large quantity (MuÄnoz et al., 1988). In addition, studies on non-heme iron absorption by humans (Morck et al., 1983) and by rats (Greger & Emery, 1987; Brown et al., 1990) have
been reported showing the result of a reduction in the non-heme iron availability. Using suckling rats fed on MnSO4 and polyphenol containing beverages such as tea, coffee and red and white wines, little effect on manganese absorption (91.7%) has been reported in the animal study by Fraile & Flynn (1992). Tannic acid, which was tested in the same series of animal studies, slightly reduced the absorption of manganese from 91.0% to 77.0%, suggesting that binding of manganese to the beverage polyphenols may be weaker than that to tannic acid. Mueller et al. (1997) have estimated that approximately 5% of the 19 mg dry matter of aluminum in ground coffee was transferred into the coffee brew. However, the metal-chelating activity also contributes to the antioxidative activity in food and biological systems. This activity depends on phenolics as well as on the Maillard reaction products formed by roasting coffee beans. It suggests a significant difference in the binding activity of metal ions to phenolics and the other chelators in beverages. Most of the test samples of coffee used in these biological studies were warm- or hot-water extracts that were preliminarily determined as tannin and phenolics. The modified vanillin±HCl assay (Burns, 1971) is the accepted method for determining condensed tannin in foods, based on the reaction between vanillin and the resorcinol group in flavanols and flavanoids. A disadvantage of the Folin-Ciocalteau reagent for total phenolics is that it detects all phenolic groups, including proteins and reducing substances such as ascorbic acid and reductones that are formed by the Maillard reaction. Brune et al. (1991) have developed a spectrophotometric assay to determine the iron-binding phenolic compounds in foods. An iron (III)-containing reagent is added to a dimethylformamide (50%) extract of the food sample, resulting in the development of colors due to Fe±galloyl and Fe±catechol complexes. The different absorbance maxima were separately determined at two wavelengths. Further fractionation and chemical characterization of phenolics and other active compounds with metal binding activity in brewed coffee are anticipated.
(b) Separation of the zinc-chelating compounds from instant coffee There have been a few studies on the separation of these metal-chelating compounds in roasted coffee. Asakura et al. (1990) have shown that the ligands present in instant coffee that bind zinc(II) are acidic in nature and have a molecular size of less than 5000 Da
Chemistry II: Non-volatile Compounds, Part II
by paper electrophoresis. Evidence that coffee pigments bind copper and iron in vitro has been reported by Homma et al. (1986). Tetramethylmurexide (TMM), a chelating titration reagent for Zn(II), has been used to determine the free Zn(II) in an assay system containing coffee constituents. This assay system used a pH 5.0 hexamine buffer (10 mM) with 50 mM ZnCl2 , 0.05% of a sample, and 10 mM KCl (Homma & Murata, 1995). Low molecular weight compounds such as citric acid for Zn(II) (NakamuraTakada et al., 1994) and aspartic acid for Fe(II) (Sekiguchi et al., 1994) have been isolated from instant coffee as chelators, and the Scatchard plot for the Zn(II)±citric acid complex gave 3.50 6 10ÿ8 [M] for the dissociation constant. These low molecular weight compounds were isolated as ligands because the metalbinding ability of a test sample was shown by the amount of bound zinc per gram of the sample at each fractionation step. Work on separating aspartic acid has demonstrated it to be an enantiomorphic mixture that might have been produced during roasting, the relative amount of aspartic acid to the total amount of free amino acids being much higher in instant coffee than in regular brewed coffee. It is expected that a fractionation method will be developed to separate macromolecular ligands with a dissociation constant less than or of the order of 10.ÿ6 The Zn(II)-chelating compounds have been isolated from instant coffee by coagulation with ZnCl2 , and after being dissolved in 1% ammonia they were purified by passage through successive ion-exchange (Amberlite IRA-410 and IR-120) columns followed by chromatography on cellulose columns and development in a mixture of 1% ammonia and n-propanol (Fig. 2.13). Ap is the purified chelated Zn(II) complex, which is subsequently separable into six different fractions, ApI±IV (as in Fig. 2.13), V and VI, with increasing molecular size but different zinc contents. During this separation procedure Zn(II) was released from the Zn(II)±coffee complex with a large dissociation constant, resulting in the selection of a complex with a small dissociation constant. The yield from instant coffee of the active compounds, Ap-III, that was finally separated in the cellulose column, was 0.3± 0.4%. The Ap-III was a brown amorphous powder, soluble in water. Its molecular weight was estimated to be about 48 000 by HPLC, using proteins as standard markers. The apparent dissociation constants for Zn(II) measured from a Scatchard plot were 1.82 6 10ÿ9 and 1.13 6 10ÿ7 [M], and the numbers of binding sites were 1.05 and 1.98, respectively. This active compound also showed chelating activity for
63
A470 32.0 30.0 28.0 15.0 n –PrOH:1%NH =5:2 3
Zn (ppm)
Ap–III
A470
3:2
1:1
10.0 Zn (ppm)
Ap–I–2 5.0
Ap–IV
Ap–I–1 Ap–II–1 Ap–II–2 0
100
200
2.0 1.0 300
400
500
600 Fr.
Fig. 2.13 Cellulose column chromatography of sample Ap (from Homma & Murata, 1995).
Cu(II) with apparent dissociation constants of 3.33 6 10ÿ9 and 2.67 6 10ÿ7 [M] and numbers of binding sites of 1.6 and 4.0, respectively. This active compound was dissociated into its subunits by treatment with EDTA and was polymerized by further exposure to Zn(II), resulting in the migration of the Ap-III fraction to the Ap-IV fraction which has a larger molecular mass than Ap-III on cellulose column chromatography (Homma & Murata, 1995). Iron (II)-chelating compounds have been also separated from instant coffee by the same procedure as that used for the Zn(II)±coffee complex. The yields of Ap-III (MW 36 000 Da) and Ap-IV (MW 50 000 Da) from instant coffee were 0.11% and 0.05%, respectively. The dissociation constant of Ap-III for iron(II), which was determined by the dialysis equilibrium method, was 5.56 6 10ÿ6 . The value of the dissociation constant of Ap-III for the Fe (II)±coffee complex is thus larger than that for the Zn(II)±coffee complex. Since iron was detected in the Fe(II)±coffee complex, most of the strong binding sites in the ApIII fraction had already been combined with iron, and the binding sites measured seem to have easily dissociated Fe(II) during the fractionation process. Further exposure of the Ap-III fraction to Fe(II) resulted in migration to the Ap-IV fraction on cellulose column chromatography. The results of gel permeation HPLC also support the proposal that Ap-III was converted to Ap-IV. Chelators such as EDTA, o-phenanthroline and bipyridyl did not release iron from Ap-III. EDTA was involved in the formation of complexes with ApIII, while o-phenanthroline and bipyridyl were involved in the migration from Ap-III to Ap-IV on
64
cellulose column chromatography (Homma & Murata, 1995).
(c) General properties of the zinc-chelating compound The chemical formula of the Zn(II)-chelating compound was experimentally determined to be C16 H21 O9 N3 , while its chemical composition was found to be 30.4% phenolics by colorimetry with the Folin-Denis method, using chlorogenic acid as the calibration standard, and 3% sugar and 4% amino acids. The nitrogen content of more than 10% is indicative of the involvement of protein through the Maillard reaction in the formation of the zinc-chelating polymer. If phytate is involved in the formation of the active compounds during roasting of green coffee beans, a strong contribution to chelation from the phosphorous group would be expected (McKenzie 1984). However, phosphorous was hardly detectable by the modified Bartlett method. The Zn-chelating compound Ap-III was found to be antioxidative toward linoleic acid by measuring the peroxide formed with ammonium thiocyanate-FeCl2 . The less the chelating metal in Ap-III, the greater its antioxidative properties. Hydrogenation of Ap-III reduced the antioxidative activity and Zn-chelating activity by half. Olefinic moieties such as enol and enaminol in the structure seem to have been involved in both these activities (Homma et al., 1997). The formation of metal-chelating compounds during the roasting process was monitored, and the Zn(II)chelating activity increased with increasing degree of roasting of the green beans. The chelating activity of brewed coffee was found to be greater in regular coffee than in instant coffee (Homma, 1999). The constituents of brewed coffee that formed brown compounds with Zn(II)-chelating activity were investigated by a model system. Model systems were prepared with one or mixtures of two to four combinations of chlorogenic acid, sucrose, bovine serum albumin and cellulose, and roasted at 2008C for 30 minutes. The Zn-chelating ability per gram of a sample was found to be highest in the model prepared with chlorogenic acid only, and lowest in the model using all four compounds (Homma & Murata, 1995; Homma et al., 1997).
(d) Chemical composition of the zinc-chelating compound The chemical structure of the active compound Ap-III has been investigated by characterizing the products
Coffee: Recent Developments
formed through such degradative reactions on the ApIII sample as alkaline fusion (3508C) and alkaline decomposition (2508C) in glycerol (Homma et al., 1997). Alkaline fusion of the active compound Ap-III yielded about 11% of ether-soluble compounds, the major ones being low molecular weight polyphenols such as pyrogallol (2.16%), protocatechuic acid (3.56%), catechol (2.16%), and p-hydroxybenzoic acid (0.76%). Table 2.7 shows tentatively characterized compounds in the acidic and basic fractions produced by alkaline fusion of Ap-III. The acidic fraction shows the presence of polyphenolics, benzoic acid and its derivatives, and carboxylic acids with 4±5 carbon chains. The basic fraction shows the presence of amides, which is indicative of the involvement of sugar and protein in the formation of Ap-III. Alkaline degradation in glycerol gave similar chromatographic patterns by LC-MS to those for alkaline fusion. Some peaks by LC-MS from the acidic fraction showed the connection of two adjacent benzene rings. The oxidative degradation of Ap-III with KMnO4 NaIO4 gave different HPLC patterns after a prior treatment by methylation. This shows that Ap-III contained a hydroxy group and a carboxy group which could be methylated. Degradation with NaClO2 (O'Neil & Selvendran, 1980) yielded pyrogallol and caffeic acid, which is indicative of the presence of benzene rings connected by an ether bond. The degradative reactions other than alkaline fusion gave fewer ether-soluble compounds than alkaline fusion, and similar phenolics were determined in the degradation products. This shows that the benzene rings involved in Ap-III were connected by strong bonds which alkaline fusion could release to the greatest effect to produce phenolics. Heinrich and Baltes (1987b) characterized pyrolyzed products of separated melanoidin in coffee brew by Curie point pyrolysis, monitored by high-resolution GC/MS, as described earlier. Comparing two independent analyses of degraded compounds from coffee melanoidins, five phenolics were found to be common to both the melanoidins separated on the basis of molecular size and on Zn(II)-chelation, respectively (Table 2.7): common phenolics are phenol, o-, m- and p-cresols, and 4-ethylphenol. Although a few nitrogeneous compounds were characterized, most of them, except for amino acids, were unknown. Most of the degraded compounds were tentatively characterized by libraries of GC MS and LC MS. Phenolic structures seem to predominate in the polymerization structures of coffee melanoidin. Cilliers and Singleton (1989, 1991) characterized oxidation products of caffeic acid
Chemistry II: Non-volatile Compounds, Part II
65
Table 2.7 Degradation products of coffee melanoidin by Curie point pyrolysis, alkaline fusion and alkaline degradation in glycerol, and a combination of these (from Heinrich & Baltes, 1987b; Homma et al. 1997). Curie point pyrolysis 2-Ethylphenol 3-Ethylphenol 2,3-Dimethylphenol 2,5-Dimethylphenol 2,6-Dimethylphenol 3,4-Dimethylphenol Ethylmethylphenol 3(4)-Hydroxyacetophenone 2-Hydroxyphenylacetate
Pyrocatechol 3-Methylpyrocatechol 4-Methylpyrocatechol 3(4)-Ethylpyrocatechol 3-Hydroxybenzaldehyde Hydroquinone Methylhydroquinone 4-Methylguaiacol Ethylguaiacol
Curie point pyrolysis, alkaline fusion and alkaline degradation Phenol p-Cresol 4-Ethylphenol o-Cresol Alkaline fusion and alkaline degradation Catechol Protocatechuic acid Pyrogallol Dihydroxyethylbenzene 2,4-Dimethylphenol Floroglucinol Butylated hydroxyanisol Butylated hydroxytoluene Trimethylbenzene Methylpropylbenzene Diethyl-propylbenzene 3,5-Dimethoxyacetophenone Benzoic acid 2-Methylbenzoic acid 3-Methylbenzoic acid 2,3-Dimethylbenzoic acid
m-Cresol
3-Hydroxybenzoic acid 2-Hydroxy-4-methylbenzoic acid 3-Methoxybenzoic acid Acetone Phenylbutanone Phenylbutanedione Diphenylbutanedione 1-Hydroxy-2-propanone 1-(1-Methylethoxy)-propane 2,2-Dimethyl-3-octanone 2-Ethoxybutane 1-(1-Methylethoxy)-butane Formic acid Acetic acid Propanoic acid 2-Hydoxypropanoic acid
in a model system at pH 8.5 and room temperature. The controlling factor in the rate of autoxidation was shown to be phenolate anion concentrations. The products were found to be specific oligomers of caffeic acid formed by reactions involving the side chain of at least one of the caffeic acid units. They were analogous to natural lignans and neolignans bridged with dioxane, furan, or cyclohexene between the caffeic units. The development of novel methods to cleave specific polymers of the parent phenolics from which these degraded phenolics were derived, is expected.
REFERENCES Aoshima, I., Tozawa, Y., Ohmomo, S. & Ueda, K. (1985) Production of decolorizing activity for molasses pigment by Coriolus versicolor Ps4a. Agric. Biol. Chem., 49, 2041±5.
Vinylguaiacol 4(5)-Propenylguaiacol Resorcinol Syringol Eugenol 3(4)-Hydroxybenzoic acid methylester 3-Hydroxyphenylacetate
2-Methylpropanoic acid Dimethylpropanoic acid 3-Benzenepropanoic acid 2-Methylbutanoic acid 3-Methylbutanoic acid Pentanoic acid Heptanoic acid 4-Methyl-2-pentanol 2-Methyl-cyclopentanol 5-Methyl-3-hexanol 1,2-Cyclohexanediol 2-Hexanal 2-Methoxy-3-(1-methylethyl)-pyrazine 5-Methylpyrimidine N-Ethyl-N-(1-methylethyl)2-propanamine
Arnold, U. & Ludwig, E. (1996) Analysis of free amino acids in green coffee beans. II. Changes of the amino acid content in arabica coffees in connection with post-harvest model treatment. Z. Lebensm. Unters.-Forsch., 203, 379±84. Arnold, U., Ludwig, E., Kuhn, R. & Moschwitzer, U. (1994) Analysis of free amino acids in green coffee beans. I. Determination of amino acids after precolumn derivatization using 9fluorenylmethylchloroformate. Z. Lebensm. Unters.-Forsch., 199, 22±5. Asakura, T., Nakamura Y., Inoue, N., Murata, M. & Homma, S. (1990) Characterization of zinc chelating compounds in instant coffee. Agric. Biol. Chem., 54, 855±62. Balyaya, K.I. & Clifford M.N. (1995) Individual chlorogenic acids and caffeine contents in commercial grades of wet and dry processed Indian green robusta coffee. J. Food Sci. Technol., 32, 104±8. Bennat, C., Engelhardt, U.H., Kiehne, A., Wirries, F.M. & Maier, H.G. (1994) HPLC analysis of chlorogenic acid lactones in roasted coffee. Z. Lebensm. Unters.-Forsch., 199, 17±21.
66
Bradbury, A.G.W., Balzer, H.H. & Vitzthum, O.G. (1998). Stabilization of liquid coffee by treatment with alkali. European Patent Application 0 861 596 A1. Brown, R., Klein, A., Simmons, W.K. & Hurrel, R.F. (1990) The influence of Jamaican herb teas and other polyphenol-containing beverages on iron absorption in the rat. Nutr. Res., 10, 343±53. Brune, M., Hallberg, I. & Skaanberg, A.B. (1991) Determination of iron-binding phenolic groups in foods. J. Food Sci., 56, 128±31. Burns, R.E. (1971) Method for estimation of tannin in grain sorghum. Agron. J. , 63, 511. Cilliers, J.J.L. & Singleton, V.L. (1989) Nonenzymic autoxidative phenolic browning reactions in a caffeic acid model system. J. Agric. Food Chem., 37, 890±96. Cilliers, J.J.L. & Singleton, V.L. (1991) Characterization of the products of nonenzymic autoxidative phenolic reactions in a caffeic acid model system. J. Agric. Food Chem., 39, 1298±303. Clifford, M.N. (1985) Chlorogenic acid. In: Coffee, Vol.1, Chemistry (eds R.J. Clarke & R. Macrae), pp. 153±202. Elsevier Applied Science, London and New York. Clifford, M.N., Kellard, B. & Ah-Sing, E. (1989a) Caffeoyltyrosine from green robusta coffee beans. Phytochemistry, 28, 1989±90. Clifford, M.N., Williams, T. & Bridson, D. (1989b) The chlorogenic acid and caffeine as possible taxonomic criteria in coffea and psilanthus. Phytochemistry, 28, 829±38. Dart, S.K. & Nursten, H.E. (1985) Volatile components. In: Coffee, Vol. 1, Chemistry (eds R.J. Clarke, & R. Macrae), pp. 239±51. Elsevier Applied Science, London and New York. Devasagayam, T.P.A., Kamat, J.P., Mohan, H. & Kasavan, P.C. (1996) Caffeine as an antioxidant: inhibition of lipid peroxidation induced by reactive oxygen species. Biochim. Biophys. Acta, 1282, 63±70. Fraile, A.L. & Flynn, A. (1992) The absorption of manganese from polyphenol-containing beverages in suckling rats. Int. J. Food Sci. Nutr., 43, 163±8. Gautschi, M., Schmid, J.P., Peppard, T.P., Ryan, T.P., Tuorto, R.M. & Yang, X. (1997) Chemical characterization of diketopiperazines in beer. J. Agric. Food Chem., 45, 3183±9. Ginz, M. & Engelhardt, U.H. (2000) Bitter compounds. Part I: Identification of proline-based diketopiperazines from roasted coffee proteins. J. Agric. Food Chem., 48, 3528±3532. Greger, J.L. & Emery, S.M. (1987) Mineral metabolism and bone strength of rats fed coffee and decaffeinated coffee. J. Agric. Food Chem., 35, 551±6. Grosch, W. (1995) Instrumental and sensory analysis of coffee volatiles. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 147±56. ASIC, Paris, France. Guillot, F.L., MalnoeÈ, A. & Stadler, R.H. (1996) Antioxidant properties of novel tetraoxygenated phenylindan isomers formed during thermal decomposition of caffeic acid. J. Agric. Food Chem., 44, 2503±10. Heinrich, L. & Baltes, W. (1987a) Uber die Bestimmung von Phenolen im KaffeegetraÈnk. Z. Lebensm. Unters.-Forsch., 185, 362±5. Heinrich, L. & Baltes, W. (1987b) Vorkommen von Phenolen in Kaffee-Melanoidin. Z. Lebensm. Unters.-Forsch., 185, 366±70.
Coffee: Recent Developments
Ho, C.-T., Hwang, H.-I., Yu T.-H. & Zhang, J. (1993) An overview of the Maillard reactions related to aroma generation in coffee. In: Proceedings of the 15th ASIC Colloquium (Montpellier), pp. 519±27. ASIC, Paris, France. Hofmann, T. (1999) Influence of L-cysteine on the formation of bitter-tasting aminohexose reductones from glucose and Lproline: identification of a novel furo[2,3-b]thiazine. J. Agric. Food Chem., 47, 4763±8. Hofmann, T., Bors, W. & Stettmaier, K. (1999a) Studies on radical intermediates in the early stage of the nonenzymatic browning reaction of carbohydrate and amino acids. J. Agric. Food Chem., 47, 379±90. Hofmann, T., Bors, W. & Stettmaier, K. (1999b) On the radicalassisted melanoidin formation during thermal processing of foods as well as under physiological conditions. J. Agric. Food Chem., 47, 391±6. Homma, S. (1999) Nonvolatile compounds in coffee. In: Proceedings of the 18th ASIC Colloquium (Helsinki), pp. 83±9. ASIC, Paris, France. Homma, S., Aida, K. & Fujimaki, M. (1986) Chelation of metal with brown pigments of coffee. In: Amino Carbonyl Reactions in Food and Biological Systems (eds M. Fujimaki, M. Namiki & H. Kato), Elsevier, Amsterdam, Netherlands. Homma S. & Murata, M. (1995) Characterization of metalchelating compounds in instant coffee. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 183±91. ASIC, Paris, France. Homma, S., Murata, M. & Takenaka, M. (1997) Chemical composition and characteristics of a zinc(II)-chelating fraction in instant coffee. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 114±9. ASIC, Paris, France. Leloup, V., Louvrier, A. & Liardon, R. (1995) Degradation mechanism of chlorogenic acids during roasting. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 192±8. ASIC, Paris, France. Ludwig, E., Raczek N.N. & Kurzrock, T. (1995). Contribution to composition and reactivity of coffee protein. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 359±64. ASIC, Paris, France. McKenzie, J.M. (1984) Content of phytate and minerals in instant coffee, coffee beans and coffee beverage. Nutr. Rep. Int., 29, 387±95. Macrae, R. (1985) Nitrogenous compounds. In: Coffee, Vol. 1, Chemistry (eds R.J. Clarke & R. Macrae), pp. 115±51. Elsevier Applied Science, London and New York. Maier, H.G. (1993) Status of research in the field of non-volatile coffee components. In: Proceedings of the 15th ASIC Colloquium (Montpellier), pp. 567±76. ASIC, Paris, France. Morck, T.A., Lynch, S.R. & Cook, J.D. (1983) Inhibition of food iron absorption by coffee. Am. J. Clin. Nutr., 37, 416±20. Morishita, H. & Kido, R. (1995) Antioxidant activities of chlorogenic acids. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 119±24. ASIC, Paris, France. Morishita, H., Takai, Y., Yamada, H. et al., (1987) Caffeoyltryptophan from green robusta coffee beans. Phytochemistry, 26, 1195±6. Mueller, M., Anke M. & Illing-Guenther, H. (1997) Availability
Chemistry II: Non-volatile Compounds, Part II
of aluminium from tea and coffee. Z. Lebensm. Unters.-Forsch., 205, 170±73. Munoz, L.D., Lonnerdal, B., Keen, C.L. & Dewey, K.G. (1988) Coffee intake during pregnancy and lactation in rats: maternal and pup hematological parameters and liver iron, zinc and copper concentration. Am. J. Clin. Nutr., 48, 645±51. Murata, M., Okada, H. & Homma, S. (1995) Hydroxycinnnamic acid derivatives and p-coumaroyl-(L)-tryptophan, a novel hydroxycinnamic acid derivative, from coffee bean. Biosci. Biotech. Biochem., 59, 1887±90. Murata, M., Terasawa, N. & Homma S. (1992) Screening of microorganisms to decolorize a model melanoidin and the chemical properties of a microbially treated melanoidin. Biosci. Biotech. Biochem., 56, 1182±7. Nakamura-Takada, Y., Shata, H., Minao, M. et al. (1994) Isolation of zinc-chelating compound from instant coffee by the tetramethyl murexide method. Z. Lebensm. Wiss. Technol., 27, 115±18. Nakayama, T. (1995) Protective effect of caffeic acid esters against H2 O2 -induced cell damages. Antioxidant activities of chlorogenic acids. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 119±24. ASIC, Paris, France. Nicoli, M.C., Manzocco, L. & Lerici, C.R. (1997) Antioxidant properties of coffee brews in relation to the roasting degree. Lebensm.-Wiss. u.-Technol., 30, 292±7. Ochi, T., Aoyama, M., Maruyama, T. & Niiya, I. (1997) Effects of various antioxidative substances on cookies containing iron. J. Jap. Soc. Nutr. Food Sci. (Nippon Eiyo Shokuryo Gakkaishi), 50, 231±6. Ohnishi, M., Morishita, H., Toda, S., Yase, Y. & Kido, R. (1998) Inhibition in vitro of linoleic acid peroxidation and haemolysis by caffeoyltryptophan. Phytochemistry, 47, 1215±18. O'Neil, M.A. & Selvendran, R.R. (1980) Glycoproteins from the cell wall of Phaseolus. Biochem. J., 187, 53±63. Papst, H.M.E., Ledl, F. & Belitz, H.-D. (1984) Bitterstoffe beim Erhitzen von Proline und Saccharose. Z. Lebensm. Unters.Forsch., 178, 356±60. Papst, H.M.E., Ledl, F. & Belitz, H.-D. (1985) Bitterstoffe beim Erhitzen von Saccharose, Maltose und Proline. Z. Lebensm. Unters.-Forsch., 181, 386±90. Pickenhagen, W., Dietrich, P., Keil B., Polonsky, J., Nouaille, F. & Lederer, E. (1975) Identification of the bitter principle of cocoa. Helv. Chim. Acta, 58, 1078±86. Reineccius, G.A. (1995) The Maillard reaction and coffee flavor. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 249± 57. ASIC, Paris, France. Rizzi, G.P. (1999) Formation of sulfur-containing volatiles under coffee roasting conditions In: Proceedings of the 217th ACS National Meeting in Anaheim. Division of Agriculatural and Food Chemistry No. 048. American Chemical Society. Rizzi, G.P. & Boekley, L.J. (1993) Flavor chemistry based on the thermally-induced decarboxylation of p-hydroxycinnamic acids. In: Food Flavors, Ingredients and Composition (ed. Charalambous), pp. 663±70. Elsevier Science Publishers, Amsterdam, Netherlands.
67
Rogers, W.J., Bezard, G., Deshayes, A., Petiard, V. & Marraccini, P. (1997). An 11s-type storage protein from Coffea arabica L. endosperm: biochemical characterization, promoter function and expression during grain maturation. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 161±8. ASIC, Paris, France. Scholz-BoÈttcher, B.M. & Maier, H.G. (1991) Isomers of quinic acid and quinides in roasted coffee: indicators for the degree of roast? In: Procceedings of the 14th ASIC Colloquim (San Francisco), pp. 220±29. ASIC, Paris, France. Schrader, K., Kiehne, A., Engelhardt, U.H. & Maier, H.G. (1996) Determination of chlorogenic acids with lactones in roasted coffee. J. Sci. Food Agric., 71, 392±8. Sekiguchi N., Yata M., Murata M. & Homma, S. (1994) Identification of iron-binding compound in instant coffee. Nippon Nogeikagaku Kaishi, 68, 821±7. Stadler, R.H. Richoz, J., Turesky, R.J., Welti, D.H. & Fay, L.B. (1996a) Oxidation of caffeine and related methylxanthines in ascorbate and polyphenol-driven Fenton-type oxidations. Free Rad. Res., 24, 225±40. Stadler, R.H., Welti, D.H., Staempfli, A.A. & Fay, L.B. (1996b) Thermal decomposition of caffeic acid in model systems: identification of novel tetraoxygenated phenylindan isomers and their stability in aqueous solution. J. Agric. Food Chem., 44, 898±905. Steinhart, H. & Luger A. (1995) Amino acid pattern of steam treated coffee. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 278±85. ASIC, Paris, France. Steinhart, H., Moller, A. & Kletschkus, H. (1989) New aspects in the analysis of melanoidins in coffee with liquid chromatography. In: Proceedings of the 13th ASIC Colloquium (Paipa), pp. 197±205. ASIC, Paris, France. Steinhart, H. and Packert, A. (1993) Melanoidins in coffee. Separation and characterization by different chromatographic procedures. In: Proceedings of the 15th ASIC Colloquium (Montpellier), pp. 593±600. ASIC, Paris, France. Terasawa, N., Murata, M. & Homma, S. (1994) Isolation of a fungus to decolorize coffee. Biosci. Biotech. Biochem., 58, 2093± 5. Terasawa, N., Murata, M. & Homma, S. (1996) Comparison of brown pigments in foods by microbial decolorization. J. Food Sci., 61, 669±72. Turesky, R.J., Stadler, R.H. & Leong-Morgenthaler, P.M. (1993) The pro- and antioxidative effects of coffee and its impact on health. In: Proceedings of the 15th ASIC, Colloquium (Montpellier), pp. 426±32. ASIC, Paris, France. Wynnes, K.N., Fumilari, M., Boublic, J.H., Drummer, O.H., Rae, I.D. & Funder, J.W. (1987) Isolation of opiate receptor ligands in coffee. Clin. Exper. Pharmacol. Physiol., 14, 785±90.
Chapter 3
Chemistry III: Volatile Compounds W. Grosch Deutsche ForschungsansaÈtlt fuÈr Lebensmittelchemie Garching, Germany 3.1 INTRODUCTION
Table 3.1 Volatile compounds in coffee.
Besides its stimulatory effect, coffee is appreciated and/or consumed for its pleasing aroma, which is the result of roasting. It is not surprising, therefore, that numerous investigations have been carried out to identify the volatile compounds which evoke this pleasing aroma for most people, assessable directly through the nostrils of the nose, or as the odour element of the overall flavour on drinking a brew. Reichstein and Staudinger carried out the first exhaustive research in the years 1920±30. They isolated a yellow-coloured oil from large quantities of roasted ground coffee and identified more than 29 volatile substances by the preparation of derivatives and measurements of physical constants (Reichstein & Staudinger, 1926, 1950, 1955). The authors maintained that not a single one of the compounds identified cause the coffee aroma. However, they emphasised that a highly diluted aqueous solution of 2-furfurylthiol `exhales a pleasant note indicative of coffee' (Reichstein & Staudinger, 1955). Progress in instrumental analysis, particularly highresolution gas chromatography (HRGC) and mass spectrometry, has shown that the volatile fraction of roasted coffee consists of a great multiplicity of compounds. More than 800 volatile compounds with a wide variety of functional groups have been identified (Table 3.1). In the first place this progress is due to the work of Gianturco et al. (1963, 1964, 1966), Bondarovich et al. (1967), Goldman et al. (1967), Stoll et al. (1967), Friedel et al. (1971), Vitzthum & Werkhoff (1974a,b, 1975, 1976), Tressl et al. (1978a,b, 1981), Tressl & Silwar (1981), and Silwar et al. (1987). Details of these studies have been reviewed by Dart and Nursten (1985), Flament (1989, 1991) as well as by Nijssen et al. (1996). As the lists of volatile compounds increased in length, the question arose whether all of them, or which
Class of compound
Number
Hydrocarbons Alcohols Aldehydes Ketones Carboxylic acids Esters Pyrazines Pyrroles Pyridines Other bases (e.g. quinoxalines, indoles) Sulphur compounds Furanes Phenols Oxazoles Others
80 24 37 85 28 33 86 66 20 52 100 126 49 35 20 Total
841
Source: Nijssen et al. (1996).
of them, contribute to the aroma, and can be considered as potential odorants, or odoriferous compounds, especially in coffee brews, but also from the dry roast and ground product. A first approach to combine instrumental results with sensory properties of the volatile compounds was undertaken by Rothe and Thomas (1963). They suggested that only those compounds, the concentrations of which surpass their odour thresholds are odour active in food. However, the determination of activity values (OAVs, ratio of odour threshold of a compound to its odour threshold) for all of the volatiles in coffee would be very laborious because concentration and odour threshold data must be determined for the large number of compounds listed in Table 3.1. The first practicable methods to convert the results of instrumental analysis of the volatiles into sensory 68
Chemistry III: Volatile Compounds
69
data were CHARM analysis (Acree et al. 1984; Acree, 1993) and aroma extract dilution analysis (AEDA; Schmid & Grosch, 1986; Ullrich & Grosch, 1987). In both procedures serial dilutions of an extract containing the volatile fraction of a food are analysed by gas chromatography/olfactometry (GCO). The application of AEDA by Holscher et al. (1990) and Blank et al. (1992) to ground roasted coffee and brew was the starting point for the identification and quantification of the compounds contributing to the aroma. Grosch (1998a) and Vitzthum (1999) have reviewed these new developments in coffee flavour research. The aim of this chapter is to present the characterising impact flavour compounds of raw and roasted coffee as well as those of the coffee brew. Changes in the composition of relevant odorants during storage of roasted coffee and identification of those causing off flavours are additional topics of this chapter. However, first the state of the art in the methodology of aroma analysis needs to be considered. Furthermore, reaction routes which lead to the development of important odorants in the roasting process will be discussed.
Table 3.2
3.2 METHODOLOGY Roasted coffee contains a very complex mixture of volatile compounds, the concentrations of which vary over a broad range. Because of this, the identification and quantification of the aroma-active compounds is a difficult task. As will be discussed, the procedures listed in Table 3.2 have proved to be successful.
3.2.1 Isolation of the volatile fraction The first step, preparation of a coffee extract containing the volatile compounds, has to be performed under mild conditions. The temperature, in particular, is a critical point, due to the instability of, for example, thiols and disulphides (Guth et al., 1995). In roasted coffee some of the 2-furfurylthiol and other thiols are linked by disulphide bonds to cysteine and cysteinecontaining peptides and proteins (see the section on the formation of odorants). The increase of 2-furfurylthiol, which was observed when the volatiles from a coffee brew were isolated by combined steam distillation extraction (SDE) according to Likens and Nickerson (Grosch et al., 1994), was most likely caused by a reduction in the corresponding disulphides. To avoid the formation of artefacts, the temperature during isolation of the volatiles may not exceed 508C for a longer period as in SDE. Therefore, distillation has to
Step
Outline of aroma analysis. Procedure
I
Extraction of the coffee sample with a solvent, e.g. diethyl ether; distillation of the extract in vacuum
II
Separation of the extract by high-resolution gas chromatography (HRGC) and localisation of potent odorants by aroma extract dilution analysis (AEDA) or CHARM analysis
III
Detection of highly volatile potent odorants by gas chromatography-olfactometry of static headspace samples (GCOH)
IV
Enrichment of potent odorants by separation of the volatile compounds in neutral/basic and acidic compounds, by column chromatography and by multidimensional gas chromatography (MDGC)
V
Identification of the potent odorants by comparison of their HRGC and mass spectrometric (MS) data and odour quality with the corresponding properties of authentic substances
VI
Quantification of potent odorants and calculation of their odour activity values (OAVs)
VII
Preparation of a synthetic blend of the potent odorants on the basis of the quantitative data obtained in step VI. Critical comparison of the aroma profile of the synthetic blend, denoted aroma model, with that of the original
VIII
Comparison of the overall odour of the aroma model with that of models in which one or more components are omitted (omission experiments)
be carried out in vacuum, for example by using the new technique of solvent assisted flavour evaporation (Engel et al., 1999) which has been applied to coffee brews (Mayer & Grosch, 2000).
3.2.2 Screening for potent odorants According to Table 3.2, in the next step of analysis the coffee extract is separated by HRGC and the effluent from the capillary column is examined by sniffing. This procedure is denoted gas chromatography-olfactometry (GCO). However, the number of odorants detectable by GCO depends not only on the odour thresholds of the volatile compounds, but also on the parameters that are arbitrarily selected, such as the amount of coffee sampled, the dilution of the volatile fraction by the solvent and the sample size analysed by HRGC. Consequently, one GCO run alone is usually
70
insufficient to distinguish between the potent odorants that contribute strongly to the aroma and those odorants that are only components of the background aroma or that are insignificant. In AEDA and CHARM analysis, therefore, the extract is diluted with a solvent, for example as a series of 1+1 (v/v) dilutions, and each dilution is analysed by GCO. In the case of AEDA, the result is expressed as a flavour dilution (FD) factor (Grosch, 1993), which is the ratio of the concentration of the odorant in the initial extract to its concentration in the most dilute extract in which odour was detected by GCO. Consequently, the FD factor is a relative measure and is proportional to the OAV of the compound in air (Grosch, 1994). CHARM analysis constructs chromatographic peaks, the areas of which are proportional to the amount of chemical in the extract (Acree, 1993). The primary difference between the two methods is that CHARM analysis measures the dilution value over the entire time the compounds elute, whereas AEDA simply determines the maximum dilution value detected (Acree, 1993). Holscher et al. (1990), Blank et al. (1992) and Grosch et al. (1996) performed AEDA of roasted arabica coffee. Figure 3.1 illustrates a FD-chromatogram, which was obtained in these studies, as an example. Altogether, 38 odorants with FD factors 16 were found. Among them, numbers 14, 17 and 35, smelling catty/ roasty, earthy/roasty and boiled apple-like, respectively, appeared with the highest FD factor of 2048. The highly volatile odorants are not perceivable by AEDA or CHARM analysis because they get lost during concentration of the aroma extract or are masked in the gas chromatogram by the solvent peak. To overcome this limitation, AEDA or CHARM analysis has to be completed by GCOH of decreasing headspace volumes (step III in Table 3.2). In the example presented in Table 3.3, the procedure was started with a headspace volume of 25 ml. GCOH revealed 22 odorants. Then the headspace volume was reduced in a series of steps to reveal the most potent odorants. GCOH of the 0.4 ml volume indicated only six odorants (numbers 5, 8, 9, 11, 12 and 14 in Table 3.3) and after reduction to 0.2 ml, only 2,3pentanedione (number 8) was found. According to these results, number 8 was the most potent, highly volatile odorant in this sample of medium roasted arabica coffee. As thiols might be adsorbed, the surface of the glassware used in GCOH has to be deactivated, for example by treatment with a silyl reagent (Semmelroch
Coffee: Recent Developments
Fig. 3.1 FD Chromatogram of odorants isolated from medium roasted arabica coffee (Blank et al., 1992) 1 2,3-Butanedione (1)a 2 3-Methylbutanal (2) 3 2,3-Pentanedione (1) 4 3-Methyl-2-buten-1-thiol (3) 5 2-Methyl-3-furanthiol (3) 6 2-Furfurylthiol (1) 7 2-/3-Methylbutanoic acid (1) 8 Methional (4) 9 Unknown 10 2,3,5-Trimethylthiazole (5) 11 Trimethylpyrazine (6) 12 Unknown 13 3-Mercapto-3-methyl-1-butanol (3) 14 3-Mercapto-3-methylbutyl formate (3) 15 2-Methoxy-3-isopropylpyrazine (7) 16 5-Ethyl-2,4-dimethylthiazole (5) 17 2-Ethyl-3,5-dimethylpyrazine (6) 18 Phenylacetaldehyde (8) 19 2-Ethenyl-3,5-dimethylpyrazine (9) 20 Linalool (8) 21 2,3-Diethyl-5-methylpyrazine (10) 22 3,4-Dimethyl-2-cyclopentenol-1-one (11) 23 Guaiacol (1) 24 4-Hydroxy-2,5-dimethyl-3(2H)-furanone (12) 25 3-Isobutyl-2-methoxypyrazine (13) 26 2-Ethenyl-3-ethyl-5-methylpyrazine (9) 27 6,7-Dihydro-5-methyl-5H-cyclopentrapyrazine (14) 28 (E)-2-Nonenal (15) 29 2-(or 5-)Ethyl-4-hydroxy-5-(or 2-)methyl-3(2H)-furanoneb (12) 30 3-Hydroxy-4,5-dimethyl-2(5H)-furanone (16) 31 4-Ethylguaiacol (17) 32 p-Anisaldehyde (16) 33 5-Ethyl-3-hydroxy-4-methyl-2(5H)-furanone (16) 34 4-Vinylguaiacol (1) 35 (E)-b-Damascenone (3) 36 Unknown 37 Bis(2-methyl-3-furyl)disulphide (18) 38 Vanillin (19) a The reference for the first identification in roasted coffee is numbered in brackets: 1, Reichstein & Staudinger (1926); 2, Zlatkis & Sivetz (1960); 3, Holscher et al. (1990); 4, Silwar et al. (1987); 5, Vitzthum & Werkhoff (1974a); 6, Goldman et al. (1967); 7, Becker et al. (1988); 8, Stoll et al. (1967); 9, Czerny et al. (1996); 10, Bondarovich et al. (1967); 11, Gianturco et al. (1963); 12, Tressl et al. (1987b); 13, Friedel et al. (1971); 14, Vitzthum & Werkhoff (1975); 15, Parliment et al. (1973); 16, Blank et al. (1992); 17, Gianturco et al. (1966); 18, Tressl & Silwar (1981); 19, Clements & Deatherage (1957). b Of the two tautomeric forms, only the 5-ethyl-2-methyl isomer is odour-active (BruÈele et al., 1995).
Chemistry III: Volatile Compounds
Table 3.3
71
GCOH of ground roasted arabica coffee1
No
Odorant
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
Acetaldehyde (1)5 Methanethiol (1)5 Propanal (2)5 Methylpropanal (3)5 2,3-Butanedione 3-Methylbutanal 2-Methylbutanal 2,3-Pentanedione 3-Methyl-2-buten-1-thiol 2-Methyl-3-furanthiol Methional 2-Furfurylthiol Unknown 3-Mercapto-3-methylbutyl formate 2-Ethyl-3,5-dimethylpyrazine Guaiacol 2-Ethyl-3,5-dimethylpyrazine 2,3-Diethyl-5-methylpyrazine 2-Ethenyl-3-ethyl-5-methylpyrazine 2-lsobutyl-3-methoxypyrazine Unknown (E)-b-Damascenone
Rl2
Volume3 (ml)
FD Factor4
< 500 < 500 &500 &500 580 653 662 697 822 870 906 911 986 1022 1086 1092 1107 1155 1182 1186 1225 1400
1 5 5 5 0.4 2 5 0.2 0.4 1 0.4 0.4 25 0.4 1 2 1 1 2 1 25 5
25 5 5 5 62.5 12.5 5 125 62.5 25 62.5 62.5 1 62.5 25 12.5 25 25 12.5 25 1 5
Source: Semmelroch & Grosch (1995). 1 The sample (100 mg) was placed into a vessel (volume 250 ml), sealed with a septum and then held at room temperature. 2 RI, retention index on a non-polar capillary (RTX-5). 3 Lowest headspace volume required to perceive the odorant at the sniffing port. 4 The highest headspace volume (25 ml) was equated to an FD factor of 1. The FD factor values of the other odorants were calculated on this basis. 5 Odorants numbers 1±4 were only detected by GCOH, the others also by AEDA (see Fig. 3.1). Numbers 1±4 were identified in coffee for the first time by (1) Reichstein & Staudinger (1926); (2) Prescott et al. (1937); (3) Rhoades (1958).
& Grosch, 1995). A limitation of GCOH is that polar odorants, such as 4-hydroxy-2,5-dimethyl-3(2H)-furanone, are not detected, although they are important contributors to the coffee aroma (see the section on the evaluation of key odorants). Therefore, GCOH cannot replace AEDA or CHARM analysis.
3.2.3 Enrichment and identification The screening for potent odorants is not corrected for the losses of odorants during the isolation procedure. Consequently, the identification experiments should be focused not only on compounds with the highest dilution value, but also on those perceived at lower dilutions, particularly in the 50±100-fold dilution range (Grosch, 1993). In the case of roasted coffee, the identification experiments were focused on the 38
odorants with FD factors in the range 16 to 2048 (Fig. 3.1). Altogether the chemical structures of 35 odorants were established. The list of authors (legend of Fig. 3.1) who had detected these compounds for the first time in roasted coffee reveals that six odorants (numbers 1, 3, 6, 7, 23 and 34 in Fig. 3.1) had been already identified by Reichstein & Staudinger (1926) in their classical study. Nine odorants (numbers 4, 5, 13, 14, 19, 26, 30, 33 and 35) were concealed in the gas chromatogram by large peaks of odourless volatiles, and were only detectable by GCO analysis of the coffee extract (Holscher et al., 1990; Blank et al., 1992). Therefore, they had to be enriched before their identification by HRGC-MS was successful, by a procedure such as will now be described. After separation of the acids and furanones the aroma extract was chromatrographed on a silica gel
72
column and each fraction was examined by GCO to establish the position of the analyte (Blank et al., 1992). This procedure was helpful for the identification of 3isopropyl-2-methoxypyrazine, 2,3-diethyl-5-methylpyrazine and (E)-b-damascenone. However, it was not sufficient for 3-mercapto-3-methylbutyl formate. Therefore, the fraction containing this analyte was further purified by HPLC on silica gel. It was then possible to identify this thiol on the basis of the criteria outlined in Table 3.2 (step V). As reported by Czerny et al. (1996) a special procedure was necessary for identification of the pyrazines, numbers 19 and 26 (Fig. 3.1). Multidimensional gas chromatography (MDGC; Weber et al., 1995) is a new procedure for the enrichment of trace amounts of volatile compounds. After separation of the extract on a polar precolumn, the cut of the effluent containing the analyte is cryofocused with liquid nitrogen and then transferred to the non-polar main column, which is combined with a mass spectrometer and a sniffing port. Mayer et al. (1999) analysed arabica coffees of different provenance and degree of roast using MDGC. As mentioned above, odorants are often concealed in the gas chromatogram by major volatile compounds that do not contribute to the aroma. To avoid misidentification it is necessary to compare, by GCO, the odour quality of the analyte with that of the authentic sample at approximately equal levels. Only when there is agreement in the sensorial properties, in addition to GC and MS data is the analyte, which has been perceived by GCO in the volatile fraction, correctly identified.
3.2.4 Quantification As discussed above, the results of the dilution experiments are not corrected for losses of the odorants during the isolation and concentration steps. Furthermore, in AEDA and CHARM analysis the odorants are completely volatilised and then evaluated by GCO, whereas the volatility of the aroma compounds in ground coffee and in the coffee beverage depends on their binding to non-volatile constituents, and on their solubility in water, respectively. To indicate which of the compounds revealed by the dilution experiments might be involved in the aroma, quantification of the potent odorants and calculation of their OAVs is the next step of the analytical procedure (Table 3.2, step VI). Because of the complexity of the volatile fraction of coffee, and the large differences in concentration,
Coffee: Recent Developments
volatility and reactivity of the odorants, it is not possible to quantify them precisely (error < 15%) by using conventional methods. Losses during the clean-up of the analyte as well as during adsorption in gas chromatography (Blank et al., 1992), which are often overlooked, may lead to incorrect results. However, precise quantitative measurements of the odorants can be performed by the use of stable isotopomers of the analytes as internal standards (see examples in Fig. 3.2) in the so-called `stable isotope dilution assays'. Losses occurring during isolation and purification of the analyte are corrected because the corresponding isotopomer has identical chemical and physical properties apart from a small isotope effect that can be ignored.
Fig. 3.2 Isotopomers of potent coffee odorants used as internal standards in isotope dilution assays. Position of labeling with carbon-13 (&) or deuterium (.). The number in brackets refers to the odorant (see Table 3.6) which was quantified on the basis of the standard.
The precision of stable dilution assays has been confirmed in model experiments (Schieberle & Grosch, 1987; Guth & Grosch, 1990). Although after clean-up the yield of some analytes was lower than 10%, the quantification was correct as the standards showed equal yields. In contrast to coffee brews, direct addition of internal standards to a solid coffee sample is not practicable, because errors may result from incomplete extraction. As reported by Semmelroch et al. (1995), Semmelroch & Grosch (1996) and Mayer et al. (1999),
Chemistry III: Volatile Compounds
a number of solvents have been used to extract the odorants in high yields. The extracts were than spiked with the labelled internal standards. 2-Methyl-3-furanthiol and 3-methyl-2-buten-1thiol, which belong to the potent odorants of coffee (numbers 4 and 5 in Fig. 3.1) are difficult to quantify due to the instability of the former (Hofmann et al., 1996) and the very low concentration of the latter (see Table 3.6). However, an exact determination is possible when, after addition of the corresponding labelled internal standards, the thiols are trapped by a reaction with p-hydroxymercuribenzoic acid (Darriet et al., 1995). After extraction of the derivatives with a phosphate buffer, the analytes and their standards are liberated by addition of excess cysteine and then quantified by a dynamic headspace procedure (Kerscher & Grosch, 1998; Mayer et al., 1999). With a few exceptions, the quantitative data discussed in the following sections were obtained by the application of stable isotope dilution assays.
3.2.5 Aroma models and omission experiments In the dilution experiments the odour impact of the volatiles is evaluated separately. Interactions of the odorants, which in most cases are characterised by inhibition and suppression (Acree, 1993), are abolished. Therefore, the question of which compound among the potent odorants actually contributes to the aroma has to be answered. To detect these odorants, synthetic mixtures (aroma models) were prepared on the basis of quantitative data for roasted coffee (Czerny et al., 1999) and brew (Mayer et al., in press) (step VII in Table 3.2). Then omission experiments were performed as triangle tests (step VIII in Table 3.2). The results are discussed in the section on the evaluation of key odorants.
3.3 RAW COFFEE 3.3.1 First studies A review of the literature on volatiles of green coffee led to a list of some 230 compounds (Holscher & Steinhart, 1995). To gain an insight into the odour-active volatiles, which may contribute to the characteristic `green peas' aroma, Vitzthum et al. (1976) were the first to analyse the volatile fraction of raw coffee by GCO. They identified four 3-alkyl-2-methoxypyrazines and concluded that 2-methoxy-3-isopropylpyrazine and
73
the corresponding isobutyl-derivative are involved in the aroma.
3.3.2 Potent odorants Table 3.4 summarises the results of recent studies in which the odorants of raw coffee were analysed by GCO (Holscher & Steinhart, 1995) and AEDA (Czerny & Grosch, 2000). AEDA revealed 21 odorants, of which 3-isobutyl-2-methoxypyrazine and 2-methoxy-3,5-dimethylpyrazine appeared with the highest FD factors. Nine of the odorants detected by AEDA (numbers, 1, 3, 6, 8, 10, 11 and 13±15 in Table 3.4) were also detected by Holscher & Steinhart (1995). In addition, they found phenylacetaldehyde, (E)-bdamascenone and a further five odorants (numbers 22, 23, and 25±27) which are known oxidation products of unsaturated fatty acids.
3.3.3 Content and OAVs of odorants More information about the compounds contributing to the characteristic smell of raw coffee was obtained by quantification and calculation of OAVs using odour threshold values of the compounds on cellulose (Czerny & Grosch, in press). The results shown in Table 3.5 reveal that 3-isobutyl-2-methyoxypyrazine with an OAV of 490 is the predominant odorant of raw coffee. The concentration of this pyrazine with a pealike smell (97 mg/kg) agrees with the value reported by Holscher & Steinhart (1995) and is near the range of 50 to 70 mg/kg reported by Spadone & Liardon (1988). However, the high concentrations reported for 4vinylguaiacol (2.3 to 7.5 mg/kg; Spadone & Liardon, 1988), (E)-2-nonenal (280 mg/kg) and (E)-b-damascenone (90 mg/kg; Holscher & Steinhart, 1995) are not confirmed by the data presented in Table 3.5. Most likely, the conventional quantitative methods used by Spadone & Liardon (1988) and Holscher & Steinhart (1995) are not suitable for an accurate determination of odorants numbers 8, 11 and 13 in raw coffee. 2-Methoxy-3,5-dimethylpyrazine, with an earthy odour, is the second important odorant on the basis of OAV (Table 3.5). With odour threshold values of 0.4 ng/l (water) and 6 ng/l (cellulose) it belongs to the most odour-active volatiles which have been detected in food. 3-Isopropyl-2-methoxypyrazine reached an OAV of only 23 (Table 3.5). As its odour quality is very similar to that of the isobutyl derivative, its contribution to the aroma of raw coffee in comparison to the latter might be small. However, Becker et al. (1988) found that a pea-like off-flavour of some batches of
74
Coffee: Recent Developments
Table 3.4
Potent odorants in raw coffee.
No.
Compound
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
n-Hexanal Butyric acid 2-/3-Methylbutyric acid Ethyl 2-methylbutyrate Ethyl 3-methylbutyrate Methional Pentanoic acid 1-Octen-3-one 2-Methoxy-3,5-dimethylpyrazine 2-Methoxy-3-isopropylpyrazine Linalool 3-Hydroxy-4,5-dimethyl-2(5H)-furanone (Sotolon) (Z)-2-Nonenal (E)-2-Nonenal 3-Isobutyl-2-methoxypyrazine Unknown (RI on DB-5: 1248)4 Unknown (RI on DB-5: 1259)4 4-Ethylguaiacol 4-Vinylguaiacol Vanillin Unknown (RI on FFAP: 2068)4 Nonanal (E,Z)-2,6-Nonadienal Phenylacetaldehyde (E,E)-2,4-Nonadienal (E,Z)-2,4-Decadienal (E,E)-2,4-Decadienal (E)-b-Damascenone
FD factor1
GCO2
16 16 32 256 256 64 16 16 512 128 16 64 64 128 4096 256 256 64 64 128 64
+ +++ +++ + +++ +++ +++ +++ +++
First identification3 1 2 3 1 1 3 2 3 4 5 1 4 3 6 5 4 5 6
++ +++ +++ ++ + +++ +++
6 3 5 7 3 6 6
Source: Holscher & Steinhart (1995), Czerny & Grosch (in press). 1 Flavour dilution (FD) factor. 2 Odour intensity in GCO: +, weak; ++, strong; +++, very strong (Holscher & Steinhart, 1995). 3 References: (1) Guyot et al. (1983); (2) WoÈhrmann et al.. (1997); (3) Holscher & Steinhart (1995); (4) Czerny & Grosch, (in press); (5) Vitzthum et al. (1976); (6) Spadone et al. (1990); (7) Spadone & Liardon (1988). 4 Retention index (RI) on capillary DB-5 or FFAP.
roasted East African coffee was caused by unusual high concentrations of 3-isopropyl-2-methoxypyrazine. This off-flavour, which is also called `potato taste', is most likely caused by the interaction of insects and bacteria (Bouyjou et al., 1999). The variegated coffee bug (Antestiopis orbitales) and other insects inflict wounds on the unripe coffee cherries so that methoxypyrazine-producing bacteria can penetrate them. An increase in ethyl 2-methylbutyrate and ethyl 3methylbutyrate (numbers 1 and 2 in Table 3.5) of up to 37 and 345 mg/kg, respectively, indicates an uncontrolled fermentation of raw beans (Bade-Wegner et al., 1997). In addition, cyclohexanoic acid ethylester has been detected in over-fermented beans (Bade-Wegner et al., 1997). Concentrations of 10 to 20 mg/kg were
responsible for the fruity, silage-like off-flavour due to the very low odour threshold of 0.01 mg/kg determined for the ester dissolved in a coffee brew (Bade-Wegner et al., 1997). An off-flavour reminiscent of rotten fish has been detected in immature green beans (Illy & Viani, 1995). 4-Heptenal, an autoxidation product of linolenic acid, was identified by Full et al. (1999) as a key odorant of this aroma defect. The data in Table 3.5 confirm the assumption of Vitzthum et al. (1976) that methoxypyrazines are stable during roasting. Other odorants like methional, sotolon and the phenol numbers 10 to 12 increase greatly during this process. Also, (E)-b-damascenone appears in a concentration of 255 mg/kg (Table 3.5).
Chemistry III: Volatile Compounds
75
Table 3.5 Concentrations, odour thresholds and odour activity values (OAVs) of potent odorants in raw and medium roasted arabica coffee. Concentration (mg/kg) No.
Compound
1 2 3 4 5 6 7 8 9 10 11 12 13
Ethyl 2-methylbutyrate Ethyl 3-methylbutyrate Methional 2-Methoxy-3,5-dimethylpyrazine 3-Isopropyl-2-methoxypyrazine 3-Hydroxy-4,5-dimethyl-2(5H)-furanone (Z)-2-Nonenal (E)-2-Nonenal 3-Isobutyl-2-methoxypyrazine 4-Ethylguaiacol 4-Vinylguaiacol Vanillin (E)-b-Damascenone
Raw coffee 2.4 22 22 0.5 2.3 0.7 < 0.3 12 97 21 117 82 < 0.3
Roasted coffee 3.9 14 213 1.1 2.4 1870 < 0.3 19 97 4060 39000 3290 255
Odour threshold1 OAV in raw coffee2 0.5 0.6 9 0.006 0.1 2.1 ND 15 0.2 35 80 100 0.15
4.8 37 2.4 83 23 <1 Ð <1 490 <1 1.5 <1 <2
Source: Czerny & Grosch (2000). 1 Values in mg/kg cellulose. 2 The odour activity values (OAVs) were calculated by dividing the concentration by the odour threshold values in cellulose. ND: not determined.
3.3.4 Contaminants causing off-flavour
3.4 ROASTED COFFEE
A flavour defect described as hardish, phenolic, medicinal and musty rioy may occur in Brazilian (Spadone et al., 1990) and Kenyan coffees (Holscher et al., 1995). 2,4,6-Trichloroanisole (2,4,6-TCA) was identified as a major contributor (Spadone et al., 1990), which might be formed by degradation of the fungicide prochloraz (Holscher et al., 1995). The latter authors developed a method for the quantification of 2,4,6-TCA using the structure analogue 2,3,6-TCA as internal standard. The concentration of 2,4,6-TCA ranged from 4.3 to 7.9 mg/kg in hardish green Kenyan coffee. A sample of roasted rioy Brazilian coffee was contaminated with the high level of 36.1 mg/kg. Due to an extraction yield of 20%, the brew prepared from the latter contained 0.6 mg/l. This value is about 500 times the taste threshold of 2,4,6-TCA in a coffee beverage (Spadone et al., 1990). In addition to 2,4,6-TCA, 2-methylisoborneol and geosmin were found to be responsible for a mouldy offflavour in Mexican coffees (arabica) (Cantergiani et al., 1999). Their concentrations in mouldy samples were between 0.1 and 1 mg/kg; this was 5 to 10 times higher than in the reference sample.
3.4.1 Concentration of important odorants The aroma of roasted coffee is dependent on a number of factors, including species, provenance and degree of roast. To obtain an insight into the concentration differences affecting the aroma, 28 odorants which had been screened by dilution experiments were used as indicators (Semmelroch et al., 1995; Semmelroch & Grosch, 1996; Mayer et al., 1999 and 2000). Aqueous solutions containing each of the 28 odorants in a concentration that was 50 times higher than its odour threshold were smelled by 10 assessors (Mayer, unpublished data). The compounds were then grouped according to the odour qualities listed in Table 3.6. Altogether 23 odorants were assigned to the notes sweetish/caramel-like, earthy, sulphurous/roasty and smoky/phenolic, which were used to describe the aroma profile of coffee (see Table 3.7 later). In addition, a group of fruity smelling odorants and two furanones with a spicy/seasoning-like aroma belong to the potent odorants of roasted coffee. Table 3.6 indicates that in a medium roasted arabica blend from Colombia the concentration range of the odorants ranged from 9.9 mg/kg for 3-methyl-2-buten1-thiol up to 130 mg/kg for acetaldehyde. As three
76
Coffee: Recent Developments
Table 3.6 Groups of volatile compands with similar odour qualities: concentrations in medium roasted arabica coffee blends from Colombia. Concentration (mg/kg) No.
Group/odorant
Sweetish/caramel group 1 Methylpropanal 2 2-Methylbutanal 3 3-Methylbutanal 4 2,3-Butanedione 5 2,3-Pentanedione 6 4-Hydroxy-2,5-dimethyl-3(2H)-furanone 7 5-Ethyl-4-hydroxy-2-methyl-3(2H)-furanone 8 Vanillin
Mean1 28.2 23.4 17.8 49.4 36.2 120 16.7 4.1
Variation2 24.0±32.3 20.7±26.0 17.0±18.6 48.4±50.8 34.0±39.6 112±140 16.0±17.3 3.4±4.8
Earthy group 9 2-Ethyl-3,5-dimethylpyrazine 10 2-Ethenyl-3,5-dimethylpyrazine 11 2,3-Diethyl-5-methylpyrazine 12 2-Ethenyl-3-ethyl-5-methylpyrazine 13 3-Isobutyl-2-methoxypyrazine
0.326 0.053 0.090 0.017 0.087
0.249±0.400 0.052±0.053 0.073±0.100 0.015±0.018 0.059±0.120
Sulphurous/roasty group 14 2-Furfurylthiol 15 2-Methyl-3-furanthiol 16 Methional 17 3-Mercapto-3-methylbutyl formate 18 3-Methyl-2-buten-1-thiol 19 Methanethiol 20 Dimethyl trisulphide
1.70 0.064 0.239 0.112 0.0099 4.55 0.0283
1.68±1.70 0.060±0.068 0.228±0.250 0.077±0.130 0.0082±0.013 4.4±4.7
Smoky/phenolic group 21 Guaiacol 22 4-Ethylguaiacol 23 4-Vinylguaiacol Fruity group 24 Acetaldehyde 25 Propanal 26 (E)-b-Damascenone Spicy group 27 3-Hydroxy-4,5-dimethyl-2(5H)-furanone 28 4-Ethyl-3-hydroxy-5-methyl-2(5H)-furanone
3.2 1.6 55 130 17.43 0.226 1.58 0.132
2.4±4.2 1.42±1.8 45±65 120±139 0.195±0.260 1.36±1.90 0.104±0.160
1 Mean values were calculated from the data published by Semmelroch et al. (1995), Semmelroch & Grosch (1996), Mayer et al. (1999 and 2000). 2 Lowest and highest values of the samples. 3 Only one sample was analysed.
Colombian blends were analysed, the variation of the odorant concentrations was estimated (Table 3.6). The lowest and the highest values differ at the most by 103% (2-ethenyl-3-ethyl-5-methylpyrazine) and 75% (guaiacol). In contrast, the values reported for odorants numbers 3±5, 7, 10, 12, 14±16 and 24 vary only by a maximum of 20%.
To demonstrate the influence of the coffee origin Mayer et al. (1999) compared the 28 potent odorants in medium roasted blends from Brazil, Colombia, El Salvador and Kenya as well as in the varieties Tipica from Colombia and Caturra from two regions in Ecuador. In general, the differences in the concentrations of the odorants were smaller within the blends
Chemistry III: Volatile Compounds
77
than between the blends and the varieties. For example, in the variety Tipica, the concentration of the important odorant 2-furfurylthiol was 52% higher than in a sample from Brazil which, with 1.91 mg/kg, contained the highest amount of the four blends. Holscher (1996) investigated whether potent odorants of coffee play a role in the aroma of surrogates. A large difference was found for 2-furfurylthiol, which was at least five times lower in surrogates. Furthermore, 3-isobutyl-2-methoxypyrazine, 3-methyl-2buten-1-thiol and 3-metcapto-3-methylbutyl formate were not found in surrogates.
3.4.2 Evaluation of key odorants A synthetic mixture of odorants imitated the aroma of an arabica coffee sample (Table 3.7). In particular, the very characteristic sulphurous/roasty odour note of coffee was as intense as in the real coffee sample. With the exception of 2-ethenyl-3-ethyl-5-methylpyrazine (number 12 in Table 3.6) the mixture, which was the aroma model in Table 3.7, contained the compounds in concentrations equal to those shown in Table 3.6. To replace pyrazine number 12 the concentration of 2ethyl-3, 5-dimethylpyrazine was correspondingly increased in the model, as both pyrazines agree in aroma quality and odour threshold in air (Wagner et al., 1999). Omission experiments were performed (Czerny et al., 1999). In triangle tests two samples containing the complete set of 27 odorants were singly compared with a mixture lacking in one or more components. A Table 3.7 Aroma profile of roasted ground coffee and the corresponding aroma model. Intensity1 Attribute Sweetish/caramel Earthy Sulphurous/roasty Smoky Similarity3
Coffee
Model2
1.0 1.6 2.3 1.7
1.4 1.3 2.1 1.4 2.3
Source: Czerny et al. (1999). 1 The intensity of the attributes was scored by 10 assessors on a scale of 0 (absent) to 3 (strong). 2 Altogether 27 odorants were dissolved in a sunflower oil±water mixture (1:20, v/v). 3 Similarity rating scale: 0 (no similarity) to 3 (identical with the coffee sample).
summary of the results is shown in Table 3.8. Experiment (expt) 1 indicates that omission of the pyrazines affected the aroma. Experiment 2 indicates that the aroma of the model was already changed when only the three alkylpyrazines were lacking. In corresponding experiments with an aroma model for the coffee brew the assessors did not notice the absence of 3-isobutyl-2-methoxypyrazine (Mayer et al., 2000). This leads to the conclusion that roasting of raw coffee does not only produce the pleasant aroma, but, in addition, generates odorants which mask the `green peas' note caused by the methoxypyrazine (see the section on raw coffee). The absence of furanones I and II in experiment 3, as well as furanones III and IV in experiment 4, was not recognized by a significant number of panel members (Table 3.8). However, when the four furanones were lacking (experiment 5) the aroma of the model was significantly different. In contrast, b-damascenone (experiment 6) was not significantly missed. A comparison of experiments 7 and 8 indicates that 2furfurylthiol is the outstanding odorant of the sulphurous/roasty group. Its absence in experiment 8 changed so clearly the odour of the model that 15 out of 20 answers were correct. This contrasted to experiment 7 in which the confidence limit for significance was not reached when the remaining components of the sulphurous/roasty group were omitted. Altogether, this study reveals that 2-furfurylthiol, 4vinylguaiacol, the alkylpyrazines and furanones listed in Table 3.8, as well as acetaldehyde, propanal, methylpropanal, and 2- and 3-methylbutanal, have the greatest impact on the coffee aroma. This conclusion underlines that only a small fraction of the volatiles, which are produced in the roasting process of coffee, is aroma-active, that is 4 out of 86 alkylpyrazines (Tables 3.1 and 3.8). A comparison of 80 alkylpyrazines indicated that the three listed in Table 3.8 have by far the lowest odour threshold in air (Wagner et al., 1999). This explains their role in coffee aroma. Obviously the olfactory system is so selective that only a relatively small number of compounds is perceived in the complex mixture of volatiles occurring in coffee (Grosch, 1998b).
3.4.3 Arabica versus robusta coffee The aroma profiles of arabica and robusta coffee brews are different (Vitzthum et al., 1990; Semmelroch & Grosch, 1996). The roasty, earthy and smoky/phenolic notes are more intense in the robustas and the sweetish/caramel and `green peas' notes
78
Coffee: Recent Developments
Table 3.8
Aroma of the model for roasted ground coffee as affected by the absence of compounds. Number1
Expt
Compound(s) omitted
1
2-Ethyl-3,5-dimethylpyrazine (I), 2-ethenyl-3,5-dimethylpyrazine (II), 2,3-diethyl-5-methylpyrazine (III), 3-isobutyl-2-methoxypyrazine
13*
2
Alkylpyrazines I to III as in expt 1
12*
3
4-Hydroxy-2,5-dimethyl-3(2H)-furanone (I), 5-ethyl-4-hydroxy-2-methyl-3(2H)-furanone (II)
8
4
4-Hydroxy-4,5-dimethyl-2(5H)-furanone (III), 5-ethyl-3-hydroxy-4-methyl-2(5H)-furanone (IV)
9
5
Furanones I to IV
11*
6
b-Damascenone
6
7
2-Methyl-3-furanthiol, dimethyl trisulphide, methional, 3-mercapto-3-methylbutyl formate, 3-methyl-2-buten-1-thiol, methanethiol
10
8
2-Furfurylthiol
15*
Source: Czerny et al. (1999). 1 Number of assessors who detected that the sample was different from the overall aroma (maximum 20). * Significant result (P < 0.05).
in the arabicas. Tressl et al. (1978a,b) and Heinrich & Baltes (1987) found that the content of volatile phenols is much higher in robusta than in arabica coffees. Quantification of potent odorants by stable isotope dilution assays confirmed this suggestion. As shown in Table 3.9, the concentrations of guaiacol, 4ethyl- and 4-vinylguaiacol were 9, 11 and 3 times greater, respectively, in robusta coffee. A further significant difference was found for pyrazines numbers 9 and 11 (Table 3.9) of which the robusta contained three times higher concentrations. In comparison to
Table 3.9 Major differences between Arabica and Robusta coffee in the concentrations of potent odorants. Concentration (mg/kg) No.1 Odorant 9 11 21 22 23
2-Ethyl-3, 5-dimethylpyrazine 2,3-Diethyl-5-methylpyrazine Guaiacol 4-Ethylguaiacol 4-Vinylguaiacol
Arabica2
Robusta2
0.326 0.090 3.2 1.61 55
0.940 0.310 28.2 18.1 178
Source: Semmelroch et al. (1995); Semmelroch & Grosch (1996); Mayer et al. (1999 and 2000). 1 The numbering of the odorants refers to Table 3.6. 2 Medium roasted arabica from Colombia (see Table 3.6) and robusta from Indonesia.
the arabica blend from Colombia (Table 3.6) a robusta from Indonesia contained only half the concentration of 4-hydroxy-2,5-dimethyl-3 (2H)-furanone (Semmelroch et al., 1995). However, this is not characteristic for robustas as this low value was also found in a medium roasted sample of Arabica var. caturra (Mayer et al., 1999). Vitzthum et al., (1990) assumed that the musty smelling 2-methylisoborneol (MIB) belongs to the typical odorants of robusta coffee. Therefore MIB was determined in various green and roasted robusta and arabica coffees using a stable isotope dilution assay (Bade-Wegner et al., 1993) and a conventional technique (Rouge et al., 1993). Indeed, the MIB concentration in robustas lay in the range of 120 to 430 ng/kg, whereas the MIB level of arabicas was lower than 20 ng/kg (Bade-Wegner et al., 1993). As the odour threshold of MIB in water and in coffee brew is at 2.5 and 25 ng/l, respectively, extremely low (Vitzthum et al., 1990; Blank, personal communication), MIB might contribute to the aroma of robustas. However, addition of 25 ng/l MIB to a brew prepared from 54 g/l arabica coffee caused only a mouldy/earthy off-flavour which was perceived in a robusta coffee brew (Blank, personal communication). Rouge et al. (1993) confirmed the predominance of MIB in raw robusta coffee, but in contrast to the results discussed before, no MIB was detected after roasting. The authors concluded that MIB was degraded during the heat treatment.
Chemistry III: Volatile Compounds
3.4.4 Influence of degree of roast Roasting of coffee from light to dark increases the sulphurous/roasty, earthy and smoky notes in the aroma profile. Analysis of the potent odorants (Mayer et al., 1999) suggested that the stronger sulphurous/ roasty and smoky notes of the dark sample might be caused by 2-furfurylthiol and guaiacol, which increased by 63% and 82%, respectively, when an arabica sample from Colombia was roasted from medium to dark. As exemplified in Fig. 3.3 by 2,3-butanedione and 2-ethyl3,5-dimethylpyrazine, respectively, the changes in the concentrations of the carbonyl compounds and alkylpyrazines were much smaller during roasting. The labile 4-hydroxy-2,5-dimethyl-3(2H)-furanone decreased somewhat when the roast degree increased from medium to dark (Fig. 3.3).
Fig. 3.3 Increase of potent odorants during roasting (from Mayer et al., 1999. Arabica coffee from Colombia was roasted light (&, colour value 14.9), medium (&, 12.2) and dark (&, 9.2). The odorant numbering refers to Table 3.6; 4:2,3-butanedione; 6: 4-hydroxy-2, 5-dimethyl-3(2H)-furanone; 9: 2-ethyl3, 5-dimethylpyrazine; 14: 2-furfurylthiol; 21: guaiacol.
Recently, the concentration changes of guaiacol, 4vinylguaiacol and 4-ethylguaiacol were monitored online using the combination of laser-induced resonanceenhanced multiphoton ionization and time-of-flight mass spectrometry (Dorfner et al., 1999). Thus it might be possible to develop a sensitive feedback-
79
guided system to control the roasting process using aroma-active volatiles as indicators.
3.4.5 Aroma changes during storage The aroma of coffee is not stable. Light, high temperature, moisture and oxygen accelerate the development of a stale off-flavour (Holscher & Steinhart, 1992a). Autoxidation of linoleic acid resulting in hexanal (Spadone & Liardon, 1989) as well as great losses of methanethiol, 2,3-pentanedione, methylpropanol, 2and 3-methylbutanal (Kallio et al., 1990; Holscher & Steinhart, 1992a) contribute to the decrease in aroma quality. The latter authors found that methanethiol, in contrast to the carbonyl compounds, disappeared most rapidly when roasted coffee beans were stored in the dark at room temperature with access to air. After 10 days a residue of only approximately 25% was detected in the sample. At room temperature, as well, the very important odorant 2-furfurylthiol is not stable. In ground arabica coffee the loss amounted to 80% in 40 days (Grosch et al., 1994). This result is in contrast to that obtained by Tressl and Silwar (1981), who found a strong increase in this thiol during storage of a coffee sample. However, the thiol was isolated by SDE, which may lead to cleavage of disulphides with the formation of thiols (see the section on the isolation of the volatile fraction). The aroma of roasted coffee was reported to change immediately after grinding. The intensity of the sweetish/caramel-like odour quality decreased distinctly within 15 minutes and continuously during the next 3 hours. The earthy and the smoky notes increased slowly whereas the intensity of the roasty/ sulphurous note in the aroma profile remained nearly constant (Grosch & Mayer, 2000). After a storage period of 30 minutes the concentrations of 22 potent odorants in the headspace of an arabica coffee were determined using a sensitive and precise method (Grosch & Mayer, 2000). As shown in Table 3.10, methanethiol and acetaldehyde were evaporated with the highest rates yielding losses of 66% and 45%, respectively. The loss of vanillin, 3-isobutyl2-methoxypyrazine, 2-furfurylthiol and methional (20% to 30%) was comparable to that of the malty smelling Strecker aldehydes methylpropanal and 2- and 3-methylbutanal. The four alkylpyrazines were liberated at a slower rate and the lowest rates were found for the three furanones numbers 6, 7 and 27, of which only 1% disappeared in 30 minutes. The aroma of a synthetic mixture containing 21 odorants (only number 12 was lacking) in the
80
Coffee: Recent Developments
Table 3.10 Release of potent odorants from ground roasted Colombian coffee at room temperature. No.1
Release (%)2
Group/odorant
Sweetish/caramel group 1 Methylpropanal 2 2-Methylbutanal 3 3-Methylbutanal 4 2,3-Butanedione 5 2,3-Pentanedione 6 4-Hydroxy-2,5-dimethyl-3(2H)-furanone 7 2-Ethyl-4-hydroxy-5-methyl-3(2H)-furanone 8 Vanillin
25 32 27 19 25 1.4 1.3 20
Earthy group 9 2-Ethyl-3,5-dimethylpyrazine 10 2-Ethenyl-3,5-dimethylpyrazine 11 2,3-Diethyl-5-methylpyrazine 12 2-Ethenyl-3-ethyl-5-methylpyrazine 13 3-Isobutyl-2-methoxypyrazine
12 6.6 13 13 21
Sulphurous/roasty group 14 2-Furfurylthiol 16 Methional 19 Methanethiol
23 29 66
Smoky/phenolic group 21 Guaiacol 22 4-Ethylguaiacol 23 4-Vinylguaiacol
18 8.4 4.9
Fruity group 24 Acetaldehyde 26 (E)-b-Damascenone
45 12
Spicy group 27 3-Hydroxy-4,5-dimethyl-2(5H)-furanone
1.1
Source: Grosch & Mayer (2000). 1 The numbering refers to Table 3.6. 2 Release in 30 minutes. The percentage figure is the ratio of headspace to coffee
concentrations found in the headspace matched the aroma profile of the genuine coffee sample very closely (Grosch & Mayer, 2000). This result is in agreement with those discussed in the section on the evaluation of key odorants.
3.5 COFFEE BREW 3.5.1 Extraction yield of potent odorants The aroma of the brew is different from that of ground coffee. Caramel-like, buttery and phenolic notes become more intense in the brew. AEDA, GCOH and CHARM analyses have shown that this change in the
aroma profile is not caused by the formation of new odorants (Blank et al., 1992; Semmelroch & Grosch, 1995; Deibler et al., 1998) but by a shift in the concentrations. As is to be expected, the polar odorants are preferentially extracted by hot water leading to yields greater than 75% for 2,3-butanedione, 2,3-pentanedione, furanones 6, 7 and 27, 2-ethyl-3,5-dimethylpyrazine and thiols 17 and 18 (Table 3.11). A group of eight odorants (numbers 1±3, 11, 16, 19, 21, 24 in Table 3.11) lay in the yield range of 50±75%. In the case of 2-ethenyl-3-ethyl-5-methylpyrazine, 3-isobutyl-2-methoxypyrazine, 2-furfurylthiol and bdamascenone only 25% and less of the amount present in ground coffee was found in the brew.
Chemistry III: Volatile Compounds
Table 3.11 brew.1
81
Concentration, extraction yield and odour activity values (OAV) of potent odorants from a coffee Odour threshold4 (mg/kg)
OAV5
0.7 1.9 0.4 15 30 10 1.15 25
1090 460 1430 140 50 720 700 8
Concentration (mg/kg)
Extraction yield3 (%)
Sweetish/caramel group 1 Methylpropanal 2 2-Methylbutanal 3 3-Methylbutanal 4 2,3-Butanedione 5 2,3-Pentanedione 6 4-Hydroxy-2,5-dimethyl-3(2H)-furanone 7 2-Ethyl-4-hydroxy-5-methyl-3(2H)-furanone 8 Vanillin
0.76 0.87 0.57 2.10 1.60 7.2 0.8 0.210
59 62 62 79 85 95 93 95
Earthy group 9 2-Ethyl-3, 5-dimethylpyrazine 10 2-Ethenyl-3,5-dimethylpyrazine 11 2,3-Diethyl-5-methylpyrazine 12 2-Ethenyl-3-ethyl-5-methylpyrazine 13 3-Isobutyl-2-methoxypyrazine
0.017 0.001 0.0036 0.002 0.0015
79 35 67 25 23
0.16 ND 0.09 ND 0.005
300
Sulphurous/roasty group 14 2-Furfurylthiol 15 2-Methyl-3-furanthiol 16 Methional 17 3-Mercapto-3-methylbutyl formate 18 3-Methyl-2-buten-1-thiol 19 Methanethiol
0.017 0.0011 0.010 0.0057 0.0006 0.170
19 34 74 81 85 72
0.01 0.007 0.2 0.0035 0.0003 0.2
1700 160 50 1630 2000 850
Smoky/phenolic group 21 Guaiacol 22 4-Ethylguaiacol 23 4-Vinylguaiacol
0.120 0.048 0.740
65 49 30
25 50 20
Fruity group 24 Acetaldehyde 26 (E)-b-Damascenone
4.7 0.0016
73 11
10 0.00075
Spicy group 27 3-Hydroxy-4,5-dimethyl-2(5H)-furanone
0.08
78
20
No.2
Group/odorant
110 40
50 1 40 470 2130 4
Source: Mayer et al. (2000). 1 Brew from medium roasted Colombian arabica coffee (54 g/l). 2 The numbering of the odorants refers to Table 3.6. 3 The yields were calculated by comparison of the concentration values in the brew with those of the powder (Mayer et al., in press). 4 Odour threshold of the compound dissolved in water (Rychlik et al., 1998); ND not determined. 5 OAV was calculated by dividing the concentration by the odour threshold.
Odour activity values have been calculated to evaluate which compound is most likely to elicit the given attribute in the aroma. According to Table 3.11, 3methylbutanal, methylpropanal and the two furanones 6 and 7 were the most aroma-active in the sweetish/ caramel group. Thiols 14, 17 and 18 showed the highest OAVs in the sulphurous/roasty group. Omittance experiments confirmed the contribution of 2-
furfurylthiol and 3-mercapto-3-methylbutyl formate to the aroma of the brew (Mayer et al., 2000). However, in these experiments the absence not of 3-methyl-2buten-1-thiol, but of methional was significantly perceived by the assessors, although the OAV of the latter was much lower (Table 3.11). In comparison to ground coffee the aroma impact of 2-furfurylthiol was weaker in the brew due to its low
82
extraction yield. On the other hand, methional and the formate belonged to the key odorants of the brew but not to those of ground coffee. An open question is the contribution of (E)-2nonenal to the coffee brew aroma. On the basis of the results in Table 3.5, its concentration is of the order of 1 mg/kg in the brew. According to Parliment et al. (1973) this amount is sufficient to stimulate a woody impression, which has been perceived by coffee tasting experts in brewed coffee but is absent in soluble coffee. Whether (E)-2-nonenal actually plays a role in the brew aroma needs to be studied in omission experiments analogous to those discussed above. The potent odorants of brews prepared from ground arabica and robusta coffee were compared by GCOH with those of an instant coffee brew (Semmelroch & Grosch, 1995). The aroma activity of propanal, methylpropanal, 2- and 3-methylbutanal, 2-ethyl-3,5dimethylpyrazine, guiaiacol, 2,3-diethyl-5-methylpyrazine and 2-ethenyl-3-ethyl-5-methylpyrazine agreed with that of the corresponding odorants of either the arabica or the robusta coffee brew, or with both. On the other hand, 2-furfurylthiol, 2-methyl-3-furanthiol and 3-methyl-2-buten-1-thiol were unimportant for the aroma of the instant coffee brew. Pollien et al. (1997), using another GCO approach, confirmed these results as far as they were able to identify the evaluated potent odorants.
Aroma changes during heating Heat sterilization at 1238C for 20 minutes changes the aroma of coffee (Kumazawa et al., 1998). This process, which is applied to canned coffee drinks, leads to a decrease in the roasty aroma note. Comparative AEDA before and after heating indicated a strong decrease in the character impact odorants 2-furfurylthiol, methional and 3-mercapto-3-methylbutyl formate (Kumazawa et al., 1998).
3.6 FORMATION OF ODORANTS Tressl (1989), Holscher & Steinhart (1992b), Ho et al. (1993) and Reineccius (1995) have reviewed precursors and reaction routes generating aroma compounds during roasting of coffee. They show that odorants are mainly formed by the Maillard reaction, degradation of phenolic acids and carotenes as well as by a reaction starting from phenyl alcohol. In the following studies, the formation of volatile compounds which have been identified as key odorants
Coffee: Recent Developments
of coffee is discussed. However, in most cases only liquid reaction systems have been used, that obviously differ from the conditions used for coffee roasting. Therefore the results give only a first indication of the reaction routes in coffee.
3.6.1 Mono- and dicarbonyl compounds During roasting of coffee, mono-, di- and oligosaccharides are degraded to volatile compounds which act as precursors for important odorants of the sweetish/caramel and earthy groups. 2-Oxopropanal, which has been detected by Aeschbacher et al. (1989) in roasted coffee, and its reduction product hydroxyacetone are examples. Studies on carbohydrate cleavage in the Maillard reaction (Weenen & Apeldoorn, 1996) suggested that 2-oxopropanal is formed by different reaction routes, for example from hexoses via 1- and 3-deoxyglycosones (Fig. 3.4).
Fig. 3.4 Formation of 2-oxopropanal by retroaldolisation of 1- and 3-deoxyglycosones.
2-Oxopropanal and other a-dicarbonyl compounds can initiate decarboxylative transamination of amino acids (Strecker degradation). According to the mechanism which has been clarified by Grigg & Thianpatanagul (1984), the reaction of the amino acid with 2-oxopropanal generates a 1,3-dipole (Fig. 3.5). After addition of water, the intermediate formed breaks down into aminoacetone and the Strecker aldehyde. Of the potent odorants discussed in this chapter, methylpropanal, 2- and 3-methylbutanal as well as methional are provided by the Strecker degradation of valine, isoleucine, leucine and methionine, respectively. Aminoacetone is proposed as the precursor of alkylpyrazines (see Fig. 3.9 later). However, it is unclear to what extent peptide-bound amino acids occurring in raw coffee in addition to the free forms are converted by the Strecker degradation into aldehydes. 2,3-Butanedione and 2,3-pentanedione are potent odorants of coffee (see Table 3.6), but are also of interest as precursors of 2(5H)-furanones (see below). Weenen & Apeldoorn (1996) explained the formation
Chemistry III: Volatile Compounds
83
An alternative pathway to a-dicarbonyl compounds results from fragmentation of sugars followed by intermolecular condensation (Hofmann, 1995; Weenen & Apeldoorn, 1996). As detailed in Fig. 3.6, aldol condensation of acetaldehyde with hydroxyacetaldehyde and hydroxyacetone yields 2,3-butanedione and 2,3-pentanedione, respectively (Hofmann, 1995).
3.6.2 Furanones Fig. 3.5 Decarboxylative transamination of amino acids (Strecker degradation).
Fig. 3.6 Formation of a-dicarbonyl compounds.
of 2,3-butanedione as resulting mainly from 1deoxyglycosone via isomerisation, H2 O-elimination and û-cleavage. However, model experiments using the 1-deoxyglycosone as starting material are difficult to perform because, in contrast to the 3-deoxyglycosone, it appears to be rather unstable (Weenen & Tjan, 1992).
Heating of hexoses or hexose-phosphates directly produces 4-hydroxy-2,5-dimethyl-3(2H)-furanone (HDMF) (Schieberle, 1992). The formation is thought to occur via 1-deoxyglycosone and 2,4-dihydroxy-2,5dimethyl-3(2H)-furanone (acetylformoine). A disproportionate concentration of the latter yields HDMF. This pathway (Fig. 3.7) was confirmed by experiments with [1-13 C]-glucose (Tressl et al., 1994) and by a study on the degradation of the Amadori compound N-(1-deoxy-D-fuctose-l-yl)-glycine (Blank et al., 1998), which is formed in the Maillard reaction of glucose and glycine. HDMF and 2(5)-ethyl-4-hydroxy-5(2)-methyl3(2H)-furanone (EHMF) are also generated from pentose sugars in the presence of glycine and alanine, respectively (Blank & Fay, 1996; Blank et al., 1997). Using 13 C-labelled precursors, the authors showed that the pentose carbon chain remains intact on the major reaction route. On the basis of the position of labelling in the products, they suggested that chain elongation of the 1-deoxypentosone by the Strecker aldehydes from
Fig. 3.7 Formation of 4-hydroxy-2,5-dimethyl-3(2H)-furanone (HDMF). (a) Formation of glucosylamine and rearrangement to the Amadori compound; (b) 2, 3-enolisation and allylic elimination of the amine results in 1deoxy-2, 3-hexodiulose; (c) isomerisation; (d) allylic elimination of water yields the open-chain form of acetylformoine; (e) cyclisation and reduction; (f) elimination of water with the formation of HDMF.
84
glycine (formaldehyde) or alanine (acetaldehyde), and reduction of the resulting acetylformoine-type intermediates provides HDMF and EHMF, respectively. However, there are indications that HDMF and EHMF are partly formed by sugar fragmentation. This suggestion is supported by labelling experiments (Blank & Fay, 1996) as well as by model reactions of C3 and C4 fragments that may arise during the Maillard reaction (Schieberle & Hofmann, 1996). For example, heating-induced aldol condensation of 2-hydroxypropanone and 2-oxopropanal yielded HDMF and the corresponding experiment with 2-oxobutanal provided EHMF. 2-Oxobutanal might be formed by aldol condensation of acetaldehyde and hydroxyacetaldehyde (Hofmann, 1995). Aldol condensation of sugar fragments can also explain the formation of 3-hydroxy-4,5-dimethyl2(5H)-furanone (sotolon) and 4-ethyl-3-hydroxy-5methyl-2(5H)-furanone (abhexone) (Hofmann, 1995; Hofmann & Schieberle, 1996). This has been supported by experiments indicating that 2,3-butanedione and 2,3-pentanedione act as precursors of sotolon (Fig. 3.8) and abhexone, respectively. The generation of these two a-dicarbonyl compounds is explained in Fig. 3.6.
Coffee: Recent Developments
aminoacetone, 2-aminopropanal and acetaldehyde. As displayed in Fig. 3.9, these compounds provide EDMP. DEMP could arise from aminoacetone and 1amino-2-butanone, but the formation of the latter is not clear (Amrani-Hemaimi et al., 1995).
Fig. 3.9 Formation of 2-ethyl-3, 5-dimethylpyrazine.
3.6.4 Phenols Model experiments indicated that 4-feruloyl quinic acid is the precursor of guaiacol, 4-ethyl- and 4vinylguaiacol (Tressl, 1989). As robustas contain more of this acid than arabicas, the higher level of phenolic odorants in the former (see Table 3.9) is understandable. Heating of free ferulic acid at 2008C in air yielded among others, guaiacol, 4-vinylguaiacol and vanillin (Tressl et al., 1976). The authors suggested that these odorants are formed by a radical-mediated reaction sequence (Fig. 3.10).
Fig. 3.8 Formation of 3-hydroxy-4, 5-dimethyl2(5H)-furanone.
3.6.3 Alkylpyrazines Among the most aroma-active alkylpyrazines, 2-ethyl3,5-dimethylpyrazine (EDMP) and 2,3-diethyl-5methylpyrazine (DEMP) occur in high-concentrations in coffee and other roasted foods (Grosch, 1998b). The formation of the two pyrazines has been studied at pH 5.6 by heating mixtures of sugars and amino acids at 1808C (Cerny & Grosch, 1994) as well as by labelling experiments (Amrani-Hemaimi et al., 1995). On the basis of these results, and in consideration of the suggestions of Shibamoto and Bernhard (1977) and Weenen & Tjan (1992), the following reaction route for EDMP was proposed (Cerny & Grosch, 1994). Strecker degradation of alanine by 2-oxopropanal yields
Fig. 3.10 Thermal degradation of ferulic acid to 4vinylguaiacol (1), vanillin (2) and guaiacol (3). R: radical.
3.6.5 Thiols As discussed in the section on the evaluation of key odorants, 2-furylthiol (FFT) is the outstanding odorant of the sulphur-containing fraction of roasted coffee.
Chemistry III: Volatile Compounds
Model experiments performed under controlled roasting conditions revealed that pentoses were effective and significantly more effective than hexoses as precursors of FFT (Parliament & Stahl, 1995; Grosch, 1999). Of the sulphur sources, the cysteine-containing tripeptide, glutathione, was more effective than free cysteine in the formation of FFT (Parliment & Stahl, 1995). Further model experiments confirmed that free and peptide-bound cysteine as well as arabinose, occurring as building blocks of polysaccharides (Bradbury & Halliday, 1990), are the most active precursors of FFT (Grosch, 1999). Roasting of an arabinogalactan, which was isolated from raw coffee, indicated that the low moisture content and the acidic conditions of raw coffee enhance the production of FFT. A partial hydrolysis of the polysaccharide during roasting with formation of free arabinose might be the cause for this effect (Grosch, 1999). It is well known that furfural results from thermal degradation of pentoses. As furfural produces high amounts of FFT (Parliament & Stahl, 1995; Hofmann & Schieberle, 1997) the mechanism shown in Fig. 3.11 was proposed to explain the formation of FFT. Disulphide bridges to cysteine and cysteine-containing peptides and proteins link a portion of FFT and other thiols. They were cleaved by a treatment with dithiothreitol (Grosch, 1999). Table 3.12 lists the concentrations of bound thiols determined in a sample of roasted coffee. 3-Methyl-2-buten-1-thiol and 3-mercapto-3-methylbutyl formate (MMBF) are potent odorants due to their very low odour thresholds (Holscher et al., 1992). Prenyl alcohol, of which about 0.5 mg/kg occurs in raw coffee, has been proposed as a precursor of the two thiols (Holscher et al., 1992). Hydrogen sulphide liberated from free and bound cysteine may substitute the hydroxy group of the prenyl alcohol to form 3-methyl-2-buten-1-thiol. On the other hand hydrogen sulphide may react with the double bond of the prenyl alcohol yielding 3-mercapto-
Fig. 3.11 Hypothetical mechanism suggesting the formation of 2-furfurylthiol (Hofmann & Schieberle, 1997).
85
Table 3.12 Bound thiols in medium roasted Colombian arabica coffee. Odorant 2-Furfurylthiol 3-Methyl-2-buten-1-thiol 3-Mercapto-3-methylbutanal 3-Mercapto-3-methylbutyl formate Methanethiol
Concentration (mg/kg) 914 12 800 13 1310
Source: Grosch (1999).
3-methylbutanol. As formic acid did not produce MMBF in model reactions (Holscher et al., 1992), it should be investigated whether formaldehyde, the Strecker aldehyde of glycine, can convert 3-mercapto3-methylbutanol into MMBF.
3.7 CONCLUSIONS During the last decade the character impact odorants of raw and roasted coffee have been identified and analytical methods for their accurate determination have been developed. Application of these methods has led to the identification of the odorants originating in raw coffee as well as those responsible for the difference in aroma between ground roast coffee and coffee brew, for both arabica and robusta coffee. Preliminary results indicate the dependence of the composition of key odorants on the coffee origins and degree of roast. Research needs to be extended to clarify the dependence of the formation of important odorants on the coffee plant cultivar as well as on the conditions of breeding and processing.
REFERENCES Acree, T.E. (1993) Bioassays for flavor. In: Flavor Science ± Sensible Principles and Techniques (eds T. E. Acree & R. Teranishi) pp. 1±20. American Chemical Society, Washington, DC. Acree, T.E., Barnard, J. & Cunningham, D.G. (1984) A procedure for the sensory analysis of gas chromatographic effluents. Food Chem, 14, 273±86. Aeschbacher, H.U., Wolleb, U., Loeliger, J., Spadone, J.C. & Liardon, R. (1989) Contribution of coffee aroma constitutents to the mutagenicity of coffee. Food Chem. Toxicol., 27, 227±32. Amrani-Hemaimi, M., Cerny, C. & Fay, L. B. (1995) Mechanisms of formation of alkylpyrazines in the Maillard reaction. J. Agric. Food Chem., 43, 2818±22. Bade-Wegner, H., Bendig, I., Holscher, W. & Wollmann, R.
86
(1997) Volatile compounds associated with the over-fermented flavour defect. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 176±82. ASIC, Paris, France. Bade-Wegner, H., Holscher, W. & Vitzthum, O.G. (1993) Quantification of 2-methylisoborneol in roasted coffee by GCMS. In: Proceedings of the 15th ASIC Colloquium (Montpellier), pp. 537±44. ASIC, Paris, France. Becker, R., DoÈhla, B., Nitz, S. & Vitzthum, O. G. (1988) Identification of the peasy off-flavour note in central African coffees. In: Proceedings of the 12th ASIC Colloquium (Montreux), pp. 203±15. ASIC, Paris, France. Blank, I., Devaud, S. & Fay, L.B. (1998) Study on the formation and decomposition of the Amadori compound N-(1-deoxy-Dfructose-1-yl)-glycine in Maillard model systems. In: The Maillard Reaction in Foods and Medicine (ed. J. O'Brien) pp. 43±50. The Royal Society of Chemistry, Cambridge. Blank, I. & Fay, L.B. (1996) Formation of 4-hydroxy-2,4-dimethyl-3(2H)-furanone and 4-hydroxy-2(or 5)-ethyl-5(or 2)methyl-3(2H)-furanone through Maillard reaction based on pentose sugars. J. Agric. Food. Chem., 44, 531±6. Blank, I., Fay, L.B., Lakner, F.J. & Schlosser, M. (1997) Determination of 4-hydroxy-2,5-dimethyl-3(2H)-furanone and 2(or 5)-ethyl-4-hydroxy-5(or 2)-methyl-3(2H)-furanone in pentose sugar-bases Maillard model systems by isotope dilution assays. J. Agric. Food Chem., 45, 2642±8. Blank, I., Sen, A. & Grosch, W. (1992) Potent odorants of the roasted powder and brew of Arabica coffee. Z. Lebensm. Forsch., 195, 239±45. Bondarovich, H.A., Friedel, P., Krampl, V., Renner, J.A., Shepard, F.W. & Gianturco, M.A., (1967) Volatile constituents of coffee. Pyrazines and other compounds. J. Agric. Food. Chem., 15, 1093±9. Bouyjou, B., Decazy, B. & Fourny, G. (1999) Removing the `potato taste' from Burundian Arabica. Plant. Rech. DeÂvelop., 6, 113±16. Bradbury, A.G.W. & Halliday, D.J. (1990) Chemical structures of green coffee polysaccharides. J. Agric. Food Chem., 38, 389±92. Bruche, G., Dietrich, A. & Mosandl, A. (1995) Stereoisomeric flavour compounds. LXXI: Determination of the origin of aroma-active dihydrofuranones. Z. Lebensm. Forsch., 201, 249± 52. Cantergiani, E., Brevard, H., Amado, R., Krebs, Y., Feria-Morales, A. & Yeretzian, C. (1999) Characterisation of mouldy/earthy defect in green Mexican coffee. In: Proceedings of the 18th ASIC Colloquium (Helsinki), pp. 43±9. ASIC, Paris, France. Cerny, C. & Grosch, W. (1994) Precursors of ethyldimethyl isomers and 2,3-diethyl-5-methylpyrazine formed in roasted beef. Z. Lebensm. Unters.-Forsch., 198, 210±14. Clements, R.L. & Deatherage, F.E. (1957) A chromatographic study of some of the compounds in roasted coffee. Food Res., 22, 222±32. Czerny, M. & Grosch, W. (2000) Potent odorants of roasted coffee. Their changes during roasting. J. Agric. Food. Chem. 48, 868±872. Czerny, M., Mayer, F. & Grosch, W. (1999) Sensory study on the character impact odorants of roasted Arabica coffee. J. Agric. Food Chem., 47, 695±9.
Coffee: Recent Developments
Czerny, M., Wagner, R. & Grosch, W. (1996) Detection of odoractive ethenylalkylpyrazines in roasted coffee. J. Agric. Food Chem., 44, 3268±72. Darriet, P., Tominaga, T., Lavigne, V., Boidron, J.N. & Dubourdieu, D. (1995) Identification of a powerful aromatic component of Vitis vinifera L. var. Sauvignon wines: 4-mercapto-3-methylpentan-2-one. Flavour Fragrance J., 10, 385± 92. Dart, S.K. & Nursten, H.E. (1985) Volatile components. In: Coffee. Vol. 1. Chemistry (eds R. J. Clarke & R. Macrae), pp. 223±65, Elsevier Applied Science, London. Deibler, K.D., Acree, T.E. & Lavin, E.H. (1998) Aroma analysis of coffee brew by gas chromatography-olfactometry. In: Food Flavors: Formation, Analysis and Packaging Influences (eds E. T. Contis, C.-T, Ho, C.J. Mussinan, T.H. Parliment, F. Shahidi & A.M. Spanier) pp. 69±78, Elsevier Science, Amsterdam. Dorfner, R., Zimmermann, R., Yeretzian, C. & Kettrup, A. (1999) On-line analysis of food processing gases by resonance laser mass spectrometry (REMPI-TOFMS): Coffee roasting and related applications. In: Proceeding of the 18th ASIC Colloquium (Helsinki), pp. 136±42 ASIC, Paris, France. Engel, W., Bahr, W. & Schieberle, P. (1999) Solvent Assistant Flavour Evaporation (SAFE) ± a new and versatile technique for the careful and direct isolation of aroma compounds from complex food matrices. Eur. Food Res. Technol., 209, 237±41. Flament, I. (1989) Coffee, cocoa, and tea. Food Rev. Int., 5, 317± 414. Flament, I. (1991) Coffee, cacao and tea. In: Volatile Compounds in Foods and Beverages (ed. H. Maarse) pp. 617±69, Marcel Dekker, New York. Friedel, P., Krampl, V., Radford, T., Renner, J.A., Shephard, F.W. & Gianturco, M.A. (1971) Some constituents of the aroma complex of coffee. J. Agric. Food Chem., 19, 530±32. Full, G., Lonzarich, V. & Suggi-Liverani, F. (1999) Differences in chemical composition of electronically sorted green coffee beans. In: Proceedings of the 18th ASIC Colloquium (Helsinki), pp. 35±42, ASIC, Paris, France. Gianturco, M.A., Giammarino, A.S. & Friedel, P. (1966) The volatile constituents of coffee. Nature, 210, 1358. Gianturco, M.A., Giammarino, A.S. & Friedel, P. & Flanagan, V. (1964) The volatile constituents of coffee. IV. Furanic acid pyrrolic compounds. Tetrahedron, 20, 2951±61. Gianturco, M.A., Giammarino, A.S. & Pitcher, R.G. (1963) The structures of five cyclic diketones isolated from coffee. Tetrahedron, 19, 2051±9. Goldman, I.M., Seibl, J., Flament, I. et al. (1967) Aroma research. About the aroma of coffee. II. Pyrazines and pyridines. Helv. Chim. Acta, 50, 694±705. Grigg, R. & Thianpatanagul, S. (1984) Decarboxylative transamination: mechanisms and applications to the synthesis of heterocyclic compounds. Chem. Soc. Chem. Commun., 1984, 180±81. Grosch, W. (1993) Detection of potent odorants in foods by aroma extract dilution analysis. Trends Food Sci. Technol., 4, 68±73. Grosch, W. (1994) Determination of potent odourants in foods by
Chemistry III: Volatile Compounds
aroma extract dilution analysis (AEDA) and calculation of odour activity values (OAVs). Flavour Fragrance J., 9, 147±58. Grosch, W. (1998a) Flavour of coffee. A review. Nahrung, 42, 344±50. Grosch, W. (1998b) Which compounds are preferred by the olfactory sense in heated foods? Lebensmittelchemie, 52, 143±6 (in German). Grosch, W. (1999) Key odorants of roasted coffee: Evaluation, release, formation. In: Proceedings of the 18th ASIC Colloquium (Helsinki), pp. 17±26. ASIC, Paris, France. Grosch, W., Czerny, M., Wagner, R. & Mayer, F. (1996) Studies on the aroma of roasted coffee. In: Flavour Science. Recent Developments (eds A. J. Taylor & D. S. Mottram) pp. 200±205. The Royal Society of Chemistry, Cambridge. Grosch, W. & Mayer, F. (2000) Release of odorants from roasted coffee. ACS-Symposium on Flavor Release, New Orleans, 22±26 August 1999, Series 763, pp. 430±438. American Chemical Society, Washington, DC. Grosch, W., Semmelroch, P. & Masanetz, C. (1994) Quantification of potent odorants in coffee. In: Proceedings of the 15th ASIC Colloquium (Montpellier), pp. 545±9. ASIC, Paris, France. Guth, H. & Grosch, W. (1990) Deterioration of soya-bean oil: quantification of primary flavour compounds using a stable isotope dilution assay. Lebensm. Wiss. Technol., 23, 513±22. Guth, H., Hofmann, T., Schieberle, P. & Grosch, W. (1995) Model reactions on the stability of disulfides in heated foods. J. Agric. Food Chem., 43, 2199±203. Guyot, B., Cros, E. & Vincent, J.C. (1983) Characterization and identification of the compounds of the volatile fraction of a sound green Arabica coffee and a malodorous Arabica coffee (in French). In: Proceedings of the 10th ASIC Colloquium (Salvador), pp. 33±58. ASIC, Paris, France. Heinrich, L. & Baltes, W. (1987) Determination of phenols in coffee. Z. Lebensm. Unters.-Forsch., 185, 362±5 (in German). Ho., C.-T., Hwang, H.-I., Yu, T.-H. & Zhang, J. (1993) An overview of the Maillard reactions related to aroma generation in coffee. In: Proceedings of the 15th ASIC Colloquium (Montpellier), pp. 519±27. ASIC, Paris, France. Hofmann, T. (1995) Characterisation of intense odorants in carbohydrate/cysteine model reactions and evaluation of their formation. A contribution to the Maillard reaction. Thesis, Technical University Munich (in German). Hofmann, T. & Schieberle, P. (1996) Studies on intermediates generating the flavour compounds 2-methyl-3-furanthiol, 2acetyl-2-thiazoline and sotolon by Maillard-type reactions. In: Flavour Science. Recent Developments (eds. A. J. Taylor & D. S. Mottram) pp. 182±7. The Royal Society of Chemistry, Cambridge. Hofmann, T. & Schieberle, P. (1997) Influence of the heating conditions on the formation of key odorants in cysteine/ribose reaction flavours. In: Flavour Perception. Aroma Evaluation. Proceedings of the 5th Wartburg Symposium, Eisenach (eds H.-P. Kruse & M. Rothe) pp. 345±5. UniversitaÈt Potsdam, Eigenverlag. Hofmann, T., Schieberle, P. & Grosch, W. (1996) Model studies on the oxidative stability of odor-active thiols occurring in food flavors. J. Agric. Food Chem., 44, 251±5.
87
Holscher, W. (1996) Comparison of some aroma impact compounds in roasted coffee and coffee surrogates. In: Flavour Science, Recent Developments (eds A.J. Taylor & D.S. Mottram) pp. 239±44. The Royal Society of Chemistry, Cambridge. Holscher, W., Bade-Wegner, H., Bendig, I., Wolkenhauer, P. & Vitzthum, O.G. (1995) Off-flavor elucidation in certain batches of Kenyan coffee. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 174±82. ASIC, Paris, France. Holscher, W. & Steinhart, H. (1992a) Investigation of roasted coffee freshness with an improved headspace technique. Z. Lebensm. Unters.-Forsch., 195, 33±8. Holscher, W. & Steinhart, H. (1992b) Formation pathways for primary roasted coffee aroma compounds. In: Thermally Generated Flavors (eds T. Parliment, M. Morello & R. McGorrin) pp. 207±17. ACS Symposium Series 543, American Chemical Society, Washington, DC. Holscher, H. & Steinhart, H. (1995) Aroma compounds in green coffee. In: Food Flavors ± Generation, Analysis and Process Influence (ed. G. Charalambous) pp. 785±803. Elsevier Science, Amsterdam. Holscher, W., Vitzthum, O.G. & Steinhart, H. (1990) Identification and sensorial evaluation of aroma-impact compounds in roasted Colombian coffee. CafeÂ, Cacao, TheÂ, 34, 205±12. Holscher, W., Vitzthum, O.G. & Steinhart, H. (1992) Prenyl alcohol ± source for odorants in roasted coffee. J. Agric. Food Chem., 40, 655±8. Illy, A. & Viani, R. (1995) Espresso Coffee. The Chemistry of Quality. Academic Press, London. Kallio, H., Leino, M., Koullias, K., Kallio, S. & Kaitarauta, J. (1990) Headspace of roasted ground coffees as an indicator of storage time. Food Chem., 36, 135±48. Kerscher, R. & Grosch, W. (1998) Quantification of 2-methyl-3furanthiol, 2-furfurylthiol, 3-mercapto-2-pentanone, and 2mercapto-3-pentanone in heated meat. J. Agric. Food Chem., 46, 1954±8. Kumazawa, K., Masuda, H., Nishimura, O. & Hiraishi, S. (1998) Change in coffee drink during heating. Nippon Shokuhin Kagaku Kaishi, 45, 108±13 (in Japanese). Mayer, F., Czerny, M. & Grosch, W. (1999) Influence of provenance and roast degree on the composition of potent odorants in Arabica coffees. Eur. Food Res. Technol., 209, 242±50. Mayer, F., Czerny, M. & Grosch, W. (2000) Sensory study on the character impact aroma compounds of coffee beverage. Eur. Food Res. Technol. 211, 212±276. Nijssen, L.M., Visscher, C.A., Maarse, H., Willemsens, L.C. & Boelens, M.H. (1996) Volatile Compounds in Food. Qualitative and Quantitative Data, 7th edn, pp. 72. 1±72.23. TNO Nutrition and Food Research Institute, Zeist, The Netherlands. Parliment, T.H., Clinton, W. & Scarpellino, R. (1973) trans-2Nonenal: coffee compound with novel organoleptic properties. J. Agric. Food Chem., 21, 485±7. Parliment, T.H. & Stahl, H.D. (1995) Formation of furfurylmercaptan in coffee model systems. In: Food Flavors: Generation, Analysis and Process Influence (ed. G. Charalambous) pp. 805±13, Elsevier Science, Amsterdam. Pollien, P., Krebs, Y. & Chaintreau, A. (1997) Comparison of a
88
brew and an instant coffee using a new GC-olfactometric method. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 191±6. ASIC, Paris, France. Prescott, S.C., Emerson, R.L., Woodward, R.B. & Heggie, R. (1937) The staling of coffee. II. Food Res., 2, 1±5. Reichstein, T. & Staudinger, H. (1926) A new or improved method of producing artifical coffee oil. Br. Patent 260960. Reichstein, T. & Staudinger, H. (1950) About the coffee aroma. Angew. Chem., 62, 292 (in German). Reichstein, T. & Staudinger, H. (1955) The aroma of coffee. Perfum. Essent. Oil. Rec., 46, 86±8. Reineccius, G.A. (1995) The Maillard reaction and coffee flavor. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 249± 57. ASIC, Paris, France. Rhoades, J.W. (1958) Sampling method for analysis of coffee volatiles by gas chromatography. Food Res., 23, 254±61. Rothe, M. & Thomas, B. (1963) Aroma compounds of bread. Z. Lebensm. Unters.-Forsch., 119, 302±10 (in German). Rouge, F., Gretsch, C., Christensen, K. & Liardon, R. (1993) Thermal stability of 2-methylisoborneol in Robusta coffee. In: Proceedings of the 15th ASIC Colloquium (Montpellier), pp. 866±8. ASIC, Paris, France. Rychlik, M., Schieberle, P. & Grosch, W. (1998) Compilation of Odor Thresholds, Odor Qualities and Retention Indices of Key Food Odorants. Deutsche Forschungsanstalt fuÈr Lebensmittelchemie und Institut fuÈr Lebensmittelchemie der Technischen UniversitaÈt MuÈnchen, 85748 Garching. Schieberle, P. (1992) Formation of furaneol in heat-processed foods. In: Flavour Precursors. Thermal and Enzymatic Conversions (eds R. Teranishi, G.R. Takeoka & M. GuÈntert) pp. 164± 74. ACS Symposium Series 490, American Chemical Society, Washington, DC. Schieberle, P. & Grosch, W. (1987) Quantitative analysis of aroma compounds in wheat and rye bread crusts using a stable isotope dilution assay. J. Agric. Food Chem., 35, 252±7. Schieberle, P. & Hofmann, T. (1996) Influence of the carbohydrate moiety on the formation of odour-active aroma compounds in thermally processed Maillard systems containing cysteine. In: Chemical Reactions in Food III. Proceedings of the Third Symposium on Chemical Reactions in Food, pp. 89±93. Czech Chemical Society, Division of Food and Agricultural Chemistry, Prague. Schmid, W. & Grosch, W. (1986) Identification of highly aromatic compounds from cherries (Prunus cerasus L.) (Z. Lebensm. Unters.-Forsch., 182, 407±12 (in German). Semmelroch, P. & Grosch, W. (1995) Analysis of roasted coffee powders and brews by gas chromatography-olfactometry of headspace samples. Lebensm. Wiss. Technol., 28, 310±13. Semmelroch, P. & Grosch, W. (1996) Studies on character impact odorants of coffee brews. J. Agric. Food Chem., 44, 537±43. Semmelroch, P., Laskawy, G., Blank, I. & Grosch, W. (1995) Determination of potent odorants in roasted coffee by stable isotope dilution assays. Flavour Fragrance J., 10, 1±7. Shibamoto, T. & Bernhard, R.A. (1977) Investigation of pyrazine formation pathways in glucose-ammonia model systems. Agric. Biol. Chem., 41, 143±53. Silwar, R., KamperschroÈer, H. & Tressl, R. (1987) Gas
Coffee: Recent Developments
chromatographic-mass spectrometry investigation of coffee aroma ± quantitative determination of steam-volatile aroma constituents. Chem. Mikrobiol. Technol. Lebensm., 10, 176±87 (in German). Spadone, J.C. & Liardon, R. (1988) Identification of specific volatile components in Rio coffee beans. In: Proceedings of the 12th ASIC Colloquium (Montreux). pp. 194±202. ASIC, Paris, France. Spadone, J.C. & Liardon, R. (1989) Analytical study of the evolution of coffee aroma compounds during storage. In: Proceedings of the 13th ASIC Colloquium (Paipa), pp. 145±57. ASIC, Paris, France. Spadone, J.C., Takeoka, G. & Liardon, R. (1990) Analytical investigation of Rio off-flavor in green coffee. J. Agric. Food Chem., 38, 226±33. Stoll, M., Winter, M. Gautschi, F., Flament, I. & Willhalm, B. (1967) About the aroma of coffee. I. Helv. Chim. Acta., 50, 628± 94 (in French). Tressl, R. (1989) Formation of flavor components in roasted coffee. In: Thermal Generation of Aromas (eds T.H. Parliment, R.J. McGorrin & C.-T. Ho) pp. 285±301. ACS Symposium Series 409, American Chemical Society, Washington, DC. Tressl, R., Bahri, D., KoÈppler, H. & Jensen, A. (1978b) Diphenols and caramel compounds in roasted coffee of different varieties. II. Z. Lebensm. Unters.-Forsch., 167, 111±14 (in German). Tressl, R., GruÈnewald, K.G., KoÈppler, H. & Silwar, R. (1978a) Phenols of roasted coffees of different varieties. I. Z. Lebensm. Unters.-Forsch., 167, 108±110 (in German). Tressl, R., GruÈnewald, K.G. & Silwar, R. (1981) Gas chromatographic-mass spectrometric study of N-alkyl- and N-furfurylpyrroles in roasted coffee. Chem. Microbiol. Technol. Lebensm., 7, 28±32 (in German). Tressl, R., Kersten, E., Nittka, C. & Rewicki, D. (1994) Formation of sulfur-containing flavor compounds from [13 C]labeled sugars, cysteine, and methionine. In: Sulfur compounds in Foods (eds C.J. Mussinan & M.E. Keelan) pp. 224±35. ACS Symposium Series 564, American Chemical Society, Washington, DC. Tressl, R., Kossa, T., Renner, R. & KoÈppler, H. (1976) Gas chromatographic and mass spectrometric investigation of the formation of phenols and aromatic hydrocarbons in food. Z. Lebensm. Unters.-Forsch., 162, 123±30 (in German). Tressl, R. & Silwar, R. (1981) Investigation of sulfur-containing components in roasted coffee. J. Agric. Food Chem., 29, 1078± 82. Ullrich, F. & Grosch, W. (1987) Identification of the most intense volatile flavour compounds formed during autoxidation of linoleic acid. Z. Lebensm. Unters.-Forsch., 184, 277±82. Vitzthum, O.G. (1999) Thirty years of coffee chemistry research. In: Flavor Chemistry ± Thirty Years of Progress (eds R. Teranishi, E.L. Wick & I. Hornstein) pp. 117±33. Kluwer Academic/Plenum Publishers, New York. Vitzthum, O.G., Weisemann, C., Becker, R. & KoÈhler, H.S. (1990) Identification of an aroma key compound in Robusta coffees. CafeÂ, Cacao, TheÂ., 34, 27±33. Vitzthum, O.G. & Werkhoff, P. (1974a) Newly discovered
Chemistry III: Volatile Compounds
nitrogen-containing heterocycles in coffee aroma. Z. Lebensm. Unters.-Forsch., 156, 300±307 (in German). Vitzthum, O.G. & Werkhoff, P. (1974b) Oxazoles and thiazoles in coffee aroma. J. Food Sci., 39, 1210±15. Vitzthum, O.G. & Werkhoff, P. (1975) Cycloalkylpyrazines in coffee aroma. J. Agric. Food Chem., 23, 510±16. Vizthum, O.G. & Werkhoff, P. (1976) Steam volatile aroma constituents of roasted coffee: neutral fraction. Z. Lebensm. Unters.-Forsch., 160, 277±91. Vitzthum, O.G., Werkhoff, P. & Ablanque, E. (1975) Volatile components of raw coffee (in German). In: Proceedings of the 7th ASIC Colloquium (Hamburg), pp. 115±23. ASIC, Paris, France. Wagner, R., Czerny, M., Bielohradsky, J. & Grosch, W. (1999) Structure±odour±activity relationships of alkylpyrazines. Z. Lebensm. Unters.-Forsch. A., 208, 308±16. Weber, B., Maas, B. & Mosandl, A. (1995) Stereoisomeric distribution of some chiral sulfur-containing trace components of yellow passion fruits. J. Agric. Food Chem., 42, 2438±41.
89
Weenen, H. & Apeldoorn, W. (1996) Carbohydrate cleavage in the Maillard reaction. In: Flavour Science, Recent Developments (eds A.J. Taylor & D.S. Mottram) pp. 211±16. The Royal Society of Chemistry, Cambridge. Weenen, H. & Tjan, S.B. (1992) Analysis, structure, and reactivity of 3-deoxyglucosone. In: Flavor Precursors. Thermal and Enzymatic Conversions (eds R. Teranishi, G.R. Takeoka & M. GuÈntert) pp. 217±31. ACS Symposium Series 490, American Chemical Society, Washington, DC. WoÈhrmann, R., Hojbr-Kalali, B. & Maier, H.G. (1997) Volatile minor acids in coffee. I. Contents of green and roasted coffee. Dtsch. Lebensm. Rdsch., 93, 191±4. Zlatkis, A. & Sivetz, M. (1960) Analysis of coffee volatiles by gas chromatography. Food Res., 25, 395±8.
Chapter 4
Technology I: Roasting R. Eggers and A. Pietsch Technical University Hamburg-Harburg, Germany 4.1 INTRODUCTION
4.2 ROASTING METHODS AND THEIR PARAMETERS
The aroma of green coffee beans is very weak and roasting is required to obtain the typical coffee aroma. In most cases, the coffee beans are put in contact with hot surfaces or gases to raise their temperature and thereby start complex chemical and physical changes. Clarke (1987) discussed the mechanisms and technology of coffee roasting in detail. Although there has been a long evolution of the understanding and design of roasting processes, the development has neither stopped nor slowed down. The reasons for this are the enormous market for coffee products and the complexity of the roasting mechanism. In the last decade, several publications have given new scientific insights; approximately 35 roasting patents (USA) have been filed and new roasting machines have been developed and taken into service. In this chapter, current understanding of the coffee roasting process is summarized, emphasizing newer developments since the mid-1980s.
Fig. 4.1
4.2.1 General From the technical point of view, the roasting process is complex and several parameters and processes influence each other. The main aspects are illustrated in Fig. 4.1. Roasting is induced by heat energy from the roaster. Heat from hot gases or hot metallic surfaces is transported into the green bean. The bean heats up and water begins to evaporate in an endothermic process. Differential thermal analysis has shown that exothermic reactions start at 1608C and peak at 2108C (arabica coffee) (Raemy & Lambelet, 1982). Gaseous reaction products like carbon dioxide and water vapor leave the bean. Gases entrapped in cells lead to an increase in the internal bean pressure. The bean expands and pops at a certain moment. When the desired degree of roast (color, flavor, roast mass loss) is reached, the beans are discharged from the roaster and cooled rapidly by water quenching and air cooling. In Table 4.1 some average data for green and roasted arabica coffee beans are given, based on a green bean weight of 0.15 g.
Roasting of coffee beans ± main aspects. 90
Technology I: Roasting
Table 4.1
Some average physical properties of arabica coffee beans (initial mass 0.15 g).
Green Medium roast 1
91
Mass (g)
Moisture (wt%)
Roast loss (wt%)
Dry matter loss (wt%)
Density (g/ml)
Volume (ml)
Radius1 (mm)
Porosity (Ð)
0.15 0.13
10±12 2±3
0 15±18
0 5±8
1.2±1.4 0.7±0.8
0.11±0.13 0.16±0.19
3 3.5
<0.1 0.5
Of volume equivalent sphere.
4.2.2 Conventional roasting Beans are mixed with hot gases in a horizontal rotating drum, a vertical drum with paddles or a vertical rotating bowl. Traditionally the roasting time is 8 to 12 minutes and the inlet gas temperature is of the order of magnitude of 4508C. This roasting process, carried out with high temperatures and long roasting times is abbreviated as HTLT roasting. Higher gas to coffee ratios, for example in continuous drum roasters, allow processing at lower temperatures of around 2508C and with shorter roasting times of 3 to 6 minutes.
4.2.3 Fluidized bed roasting Fludization of the beans is achieved by high velocity hot gas directed towards the beans, usually from the bottom of the roasting machine, so that the gases heat and move the floating beans simultaneously. The general advantages of fluidized bed processes are good control of the process parameters and high uniformity of the products since there are no excessive temperature peaks. Mechanical agitation is not necessary thus mechanical reliability is increased, with lower maintenance and easier cleaning. Special advantages with regard to coffee roasting are a lower loss of small coffee particles, such as broken or small beans, and the reliable removal of coffee skins from the roasting zone and therefore less smoke in the roasting area. The principle of roasting coffee beans in a fluidized bed has been known for more than 70 years, but good industrial roasting machines based on this technology took a long time to be developed. Recent developments in this field (see Section 4.6) have involved a closer look at the fluidization of coffee beans. A typical coffee bean has the shape of a half ellipsoid with well rounded edges. For ease of calculation, a spherical-shape is assumed in the following. Fluidizing properties of solid particles can be predicted by the Geldart classification, (Kunii & Levenspiel, 1991). With an average bean diameter dsphere of 6 mm and bean
densities ranging between 700 and 1400 kg/m3 , coffee beans belong to Geldart group D, indicating that they are not easily fluidized and that an enormous amount of gas is needed for fluidization of a rather shallow bed. Flow characteristics are described by the dimensionless Reynolds number, Re, defined as Re = vs dsphere / gas where vs is the superficial gas velocity and vgas is the kinematic gas viscosity. A rough estimate for the minimum superficial fluidization velocity for green beans, by approximating them to the coarse particles of Chitester, et al. (Kunii & Levenspiel, 1991) in a hot air stream at 2008C gives a value of 2.1 m/s (Re 360). During the roasting process the bean swells, its density decreases and the theoretical minimum velocity decreases correspondingly to 1.7 m/s (Re 340) for a medium roasted bean. These values illustrate that control of the gas velocity is crucial. Too low a velocity does not move the beans sufficiently and some may even burn. On the other hand, a velocity suitable to fluidize the green beans might blow swollen roasted beans out of the roaster later on. In addition, the gas temperature influences the theoretical minimum fluidization velocity: at 2508C the calculations yield 2.2 m/s (Re 310) for green and 1.8 m/s (Re 300) for medium roasted beans, values somewhat higher than the experimental data of Vincent et al. (1977). All the Re values so determined lie in the transition region between laminar and turbulent flow, and most roasting machines are unlikely to achieve full turbulent flow, with Re > 3 6 105 . According to their classification as Geldart group D material, coffee beans spout easily and roasting in a spouting bed is possible. A spouting bed does not fluidize all beans equally, but high velocity spouts of gas punch through the bed of beans and thereby transport beans to the top of the bed. Spouting requires less gas and the minimum spouting velocity depends on the bed height. Modeling of spouting beds is complicated and publications in this field are scarce. Nagaraju et al. (1997) experimented with light peaberry beans, the green beans of which have a density of
92
0.9 g/ml, and reported a minimum air:solid ratio of 1.1 to 1.5 for spouting in their set-up. They observed a maximum (0.15 m) and a minimum (0.075 m) bed depth for spouting but did not publish some important parameters for further evaluation, such as air velocity, roast loss and density of the roasted beans. A fundamental study on the fluidization and spouting parameters of coffee beans during roasting has yet to be done. Fluidized bed and spouting bed roasting facilitate roasting for short time periods with low temperature gases. Spouting increases the heat transfer but can lead to inhomogeneous roasting of the beans.
4.2.4 Fast roasting In the last 15 years fast roasting processes have become well accepted in the coffee industry. Fast roasting is performed with a high gas to bean ratio, improved heat transfer to the beans and typical roasting times of 1 to 4 minutes. Processes with roasting times below 2 minutes are called very fast or even ultrafast. Fast roasted coffee beans are different from traditionally roasted beans: they are larger, have a reduced density and greater porosity. Because of these structural differences the fast roasted beans allow better water penetration and extraction of the coffee solubles. The term `high yield' refers to those coffees which give the consumer a higher amount of soluble solids as compared to traditionally roasted coffee; less coffee is required to achieve a comparable brew strength. High yield products have become well accepted in the US market, but less so in Germany. Nevertheless, most of the coffee industry has now reduced its roasting time. To produce high yield coffee, the roasting machines have to provide intense bean movement, as fluidized bed roasters, for example, do. Processing with high temperature and short roasting time is abbreviated to HTST roasting. Maier (1985) compared the chemical composition of traditional and fast roasted coffees, using samples of similar roast color, and found that the water-soluble extract (soluble solids) increased as the roasting time decreased. The content of specific substances (such as saccharose) changed with roasting, indicating that the chemical compositions of traditional and fast roasted coffee are similar but not identical. As expected, fine grinding of coffee samples led to an increased amount of extract. Interestingly the difference in the soluble yield between traditional and fast roasted coffee was diminished by fine grinding. This supports the thesis that fast roasted coffee shows increased brew strength due to structural but not chemical changes. For a
Coffee: Recent Developments
further discussion of swelling mechanisms see Section 4.3.2. Furthermore, Maier (1985) reported that the content of caffeine in roasted coffee is independent of the speed of roasting and therefore high yield coffee leads to a reduced caffeine content in the brew, if the ratio of coffee to water is reduced as recommended by the manufacturers. Nagaraju et al. (1997) used a spouting bed for experimental fast roasting and reported that shorter roasting times with higher gas temperatures led to more soluble solids, a higher content of chlorogenic acid, less of a burnt flavour and a lower loss of volatiles. According to Bersten (1993), aroma is connected with fat and faster roasted coffee holds more fat. Nevertheless, he recommended a minimum roasting time of 3.5 minutes for arabicas and 2.5 minutes for robustas to obtain good quality aroma. There has been intense patenting activity in the field of fast roasting in the years since 1986. Brandlein et al. described a fast roasting machine to process 5000 to 12 000 lb/h green coffee in a bubbling bed, where roasting times less than 90 seconds lead to higher flavor strength (US 4 737 376; see list of patents in Table 4.4, later). Price et al. suggested the use of fluidized bed roasting to make ultrafast roasted coffee (3±120 seconds) and reported reduced bitterness and improved freshness retention during storage of the brew (US 4 988 590). Kirkpatrick et al. (US 5 160 757) and Jensen et al. (US 5 322 703) emphasized the high yield of fast roasted robusta coffees, with predrying of the beans prior to fast roasting. In order to reach a desirable brew acidity in fast roasted coffee, Gutwein et al. recommended the use of high acidity beans, such as high quality arabica or Colombian beans (US 5 721 005).
4.2.5 Detection of optimum degree of roast Automated termination of the roasting process requires a method to control roasting time and a suitable detector. In the last 15 years new methods have been suggested and patented, indicating that there is a lasting interest in further improvement. Sivetz (1991) recommended that roasting in a fluidized bed should be terminated by the end bean temperature and gave a simplified graphical relationship between bean temperature and aroma and flavor intensity. He specified end bean temperatures of 2268C for Anglo tan, 2328C for American light brown and 2388C for European brown and related these end bean temperatures with the roast loss. Figure 4.2 depicts data on the relationship between
Technology I: Roasting
93
with sensory analysis and Zimmermann et al. (1999) found characteristic concentration profiles of volatiles (headspace). A cooled photo-optical device that operates in conjunction with the inspection glass of the oven and detects the roast colour of the coffee beans was patented for industrial applications (US 4 849 625; 1989), while for small scale roasters, timers able to detect pressure (US 5 257 574; 1993) or sounds (US 4 602 147; 1986) were patented. If the method of sound detection is to be used for large batch roasters a sound distribution has to be evaluated since the beans do not pop simultaneously. Fig. 4.2 Total roast weight loss (change in moisture content/initial moisture content) depicted vs bean temperature. The variations are probably due to different beans, initial water content and temperature measurement techniques.
roast loss and bean temperature reported by different authors using different roasters. It can be seen that there are deviations, especially in the temperature range below 2208C, due to the different green beans used and their varying moisture content (Sivetz, 1991: 12%; Schenker et al., 1999; 11.1%; Da Porto et al., 1991: 9.8%; Maier, 1985: unknown; Hobbie et al., 1999: 8%) and the inclusion of water in the value roast loss. It should be noted that the terms `bean temperature' and `product temperature' are not well defined, although manufacturers of fluidized bed roasters report precise measurements from sensors in the roaster. However, the recorded values depend to a great extent on the method used and the correlation of bean temperature and roast loss or aroma is roaster specific. For the measurement of absolute bean temperatures, infrared detectors or thermosensors attached to or even intruded into the bean are required. In a recent roaster design (Burns System 90) control is achieved by measuring the outlet gas temperature instead of the bean temperature. With its exactly controlled inlet gas temperature and burner power, a more precise operation is claimed for this system. However, there is also an influence of kinetics on the relationship bean temperature ± roast loss as illustrated by the two curves from Schenker et al. (Fig. 4.2). The course of bean temperatures and masses is depicted later on (see Fig. 4.3). Recent publications suggest monitoring of the roasting process by on-line detection of certain chemical substances. For example, Hashim and Chaveron (1996) correlated methylpyrazine ratios in the bean
4.3 BEAN BEHAVIOUR DURING ROASTING 4.3.1 Bean temperature, mass and moisture Obviously the bean temperature is very important in roasting and its connection with the bean mass was shown in Fig. 4.2. The third important parameter is roasting time. Obviously roasting temperature and roasting time depend strongly on the heat transfer system and therefore the technology applied. New data on the progression of bean temperature during roasting have been published by Da Porto et al. (1991) and by Schenker et al. (1999). The latter authors chose fast fluidized bed roasting of 100 g arabica beans, a hot air flow of 0.01885 m3 /s and two modes of operation: first the so-called LTLT roasting with a hot gas temperature of 2208C and a roasting time between 9 and 12 minutes, and second HTST roasting with a gas temperature of 2608C and a roasting time between 2.6 and 3 minutes. For LTLT roasting, the bean temperature and other parameters were recorded as a function of roasting time: the temperature rose continuously from 20 to 1908C in 2 minutes (Fig. 4.3). This rapid increase was 90% of the final increase (2118C after 14 minutes). During HTST roasting the change in the slope of the temperature progression occurred around one minute at a bean temperature of 1808C, equalling 75% of the final increase at 2358C (3 minutes). The bean temperature data published by Da Porto et al. (1991) are less regular, probably due to the experimental procedure. Reported bean temperatures in their investigations with a laboratory roaster (HTLT) and Santos coffee started at 1558C after 3.3 minutes and increased up to 2458C (10 minutes). It can be seen, that in comparison with this traditional labora-
94
tory roaster the fluidized bed roaster facilitates a significantly faster increase in bean temperature. The bean moisture, initially around 12% in the green bean, diminishes during roasting. Generally, the water content falls more rapidly during the first phase of roasting than later on. Da Porto et al. (1991) published data for a Santos coffee and HTLT roasting: the moisture content decreased from 9.8 to 5.3% in 4 minutes, reduced slightly to a value of 5.1% after 6.3 minutes (marked as medium roast) and reached 2.1% after 10 minutes of roasting (Fig. 4.3). Massini et al. (1990) used the same roaster and found a moisture reduction from 11% to 3.2% for a 14 minute roast. Schenker et al. (1999) also reported moisture data for their experimental fluidized bed with LTLT roasting. During roasting the moisture reduced gradually from 11.1% to 2.3% after 6 minutes. From this point on, the slope of the moisture curve was less steep, reaching 1.5% after 10 minutes (marked as medium roast) and finally 1.3% after 14 minutes of roasting. All three publications indicate that the moisture content of beans leaving the roasting section is not zero but at least 2%. A more vigorous gas flow reduces the partial pressure of water in the gas and increases the driving force for vapor transport out of the beans. Accordingly, beans roasted in the fluidized bed are drier, although a lower gas temperature was used. During the course of roasting the bean mass reduces (i.e. the roast loss is growing). A steady increase of roast
Fig. 4.3 Course of experimental relative bean masses (triangles), relative bean volumes (circles), relative moistures (squares) and bean temperatures (diamonds) during roasting. (Relative value = value at tR/value at tR = 0.) Solid symbols represent data measured by Schenker et al. (1999) with LTLT fluidized bed roasting and open symbols represent data published by Da Porto et al. (1991) for drum roasting.
Coffee: Recent Developments
loss with roasting time has been reported. The authors of the study on the fluidized bed roasting defined a medium roast as a roast with a total mass reduction of 15% (Schenker et al., 1999). This value was attained with the LTLT process in 10 minutes and with HTST in 2.7 minutes. For LTLT roasting, the dry matter mass loss can be calculated as 6.4%. From the data of Da Porto et al. (1991) (HTLT roasting) it can be seen that 8.1 minutes were necessary to obtain a roast loss of 15.2%, equaling a dry matter loss of 8.6%. Ortola et al. (1998) published results of their experimental investigations with six coffee varieties and with varying roast gas temperatures in a sample drum roaster. Final weight loss and moisture content did not show a clear relationship with roast gas temperature.
4.3.2 Swelling and structure The mechanisms of bean swelling and puffing are not known in detail. Certainly the vaporization of bean moisture and the release of water and CO2 from the roasting reactions are the driving forces for bean expansion which act against the resistance of the structural strength of the bean material. Calculation of the internal bean pressure can be performed according to Radtke (1975) or Clarke (1987) to several (5±10) atmospheres. Brandlein et al. (1988) believed that the enhanced puffing of fast roasted beans is related to the moisture: fast roasting leads to a rapid rise in the bean temperature and thus retains a high internal bean moisture that keeps the bean more pliable during roasting. It would be very interesting to have data on the structural strength of different bean sections with varied roast degree and water content to verify this model. According to polymer science the term `glass transition temperature' is the temperature at which a polymer softens and may lose its crystal structure. Brandlein et al. (1988) reported a value of 4208F (2158C) for the glass transition temperature of coffee beans, with mannan and cellulose playing a mayor role. Small and Horrell (1993) also emphasize the influence of water on swelling, but believe that the decomposition of chlorogenic acid is important. They found that pure chlorogenic acid decomposes around 2108C and ejects CO2 sharply, which is believed essentially to cause bean swelling. Massini et al. (1990) investigated the CO2 release of roasted samples via headspace analysis and combination with bean temperature data from the same research group (Da Porto et al., 1991) shows a sharp increase in CO2 release around 2008C, a value close to the decomposition temperature of chlorogenic acid reported by Small and Horrel (1993). The latter
Technology I: Roasting
authors concluded that faster heat transport into the bean results in a more sudden decomposition of the cholorogenic acid and a greater incentive for puffing. The heat of evaporation of bean moisture hinders rapid heating and therefore the authors concluded that predrying of the beans should enhance swelling. High yield coffee with improved roast uniformity resulted from experimental roasting of predried beans (moisture below 5%). This new concept with predrying has led to two recent patents related to fast roasting (see Section 4.2.4). The relative bean volume Vrel is the bean volume at a certain roast degree divided by the volume of the green bean. Again, Schenker et al. (1999; HTST and LTLT) and Da Porto et al. (1991; HTLT) reported experimental data on the development of the bean volume (or density) during roasting. HTLT roasting showed a linear increase in Vrel ± analogous to the bean temperature ± with 1.74 for medium roast (roast loss 15.2%). During fluidized bed roasting, the increase in Vrel was not linear but weakened after the first strong increase. Values for medium roast are 1.44 (HTST) and 1.7 (LTLT). Schenker et al. (1999) measured pore volumes of arabica coffee beans with mercury intrusion porosimetry. The validity of this method for the investigation of coffee was shown by comparison with SEM micrographs and X-ray microanalysis (Schenker et al., 1998). They found an overall cumulative intruded pore volume of 130 mm3 /g for green beans, but ascribed this value to a possible artifact due to compression of coffee oil. During roasting (depicted for HTST) the pore volume increased almost linearly with the roasting time. Comparison of the pore volume of two different roasted bean samples, both with a roast loss of 15%, showed that the pore volume of HTST beans of 840 mm3 /g is 35% larger than the pore volume of LTLT beans of 630 mm3 /g. The difference between the relative bean volumes is not as large (18%) and also leads to a reduced difference in the relative porosity of the beans (12%). Bean porosity is defined as pore volume divided by bean volume and relative values, as usual, refer to the porosity of the green bean. If we assume a roasted bean weight of 0.15 g and a bean volume of 180 mm3 , the porosity of the medium roasted LTLT beans is 0.5, a value also reported by Radtke (1975) for two different roasted coffees. According to Kazi and Clifford (1985) the size of cavities in traditionally roasted coffee is in the range of 15 to 40 nm (mean value 26 nm) and in fast roasted high yield coffees in the range of 15 to 39 nm (mean value
95
24 nm). The origin of the cavities in roasted coffee is not fully understood. While some think of newly formed cavities filled with the developing gases, others use the term `excavated cell', emphasizing that the structures between the cavities are the remains of the cell walls. The detailed structure study by Schenker et al. summarized the knowledge up to 1999 and gave new insights. The pore size distribution obtained by mercury intrusion is dominated by small micropores in the range of 20 to 50 mm. The authors believe that the access for the mercury to the larger cell lumina is provided by smaller micropores in the cell walls and that the high values for the apparent pore volume of the micropores corresponded to the filling of the cell lumina. Consequently, the maximum of the pore size distribution function represented the size of these cell wall micropores. These cell wall micropores enlarged during roasting (with increasing degree of roast) and their size distribution narrowed. At equal degrees of roast (15% roast loss), the roasting conditions influenced the final cell wall pore size: for HTST beans an average apparent radius of 13.4 nm (LTLT 11.22 nm) was detected. With regard to fast roasting, the authors reported that high temperature roasted beans exhibited greater bean volume, higher cumulative pore volume and larger micropores in the cell walls as compared to low temperature roasted coffees and that these larger cell wall micropores may promote faster gas desorption, oil migration, enhanced oxygen accessibility and accelerated loss of flavor compounds. Despite these findings, the origin and structural organization of the cell wall micropore network remains unclear. For detail information on the composition of released gases, emission control, measurement of degree of roast and contents of soluble solids see Clarke (1987) and for the effects of infrared and microwave roasting see Sivetz and Desrosier (1979).
4.3.3 Decaffeinated coffee According to Radtke (1975) a cross-sectional cut of a roasted decaffeinated coffee bean shows more and considerably larger cavities than a similar roasted bean containing caffeine. The thermal treatment before the decaffeination may lead to loosening of the bean structure. The increased bean porosity reduces the average thermal conductivity of the bean and therefore decaffeinated beans need stronger roasting, as also stated by Radtke. Moreover, it is anticipated that decaffeinated roasted beans contain less fat and wax than regular beans. Since the thermal conductivity of fats is of the same order of magnitude as those for other
96
Coffee: Recent Developments
organic components in the bean (for example olive oil: 0.16 W/mK, Swern, 1979; wood: 0.17 W/mK), a reduced fat and wax content is not believed to influence the average thermal conductivity, but may influence the aroma formation and thereby, enforce stronger roasting.
4.4 HEAT AND MASS TRANSPORT 4.4.1 Complexity of the process As mentioned earlier, the physical and chemical processes during roasting are complex (Fig. 4.4). Roasting is a time- and temperature-dependent process and bean movement and water content are important. Mass and heat transfer are unsteady so that time-dependent temperature and concentration profiles form inside the bean, which have not yet been elucidated. Furthermore the local physical properties depend on the local temperature and bean composition. To visualize the effects within the bean, some speculative curves are depicted in Fig. 4.5. None of these has been measured or modeled so far, but research is in progress at the Technical University Hamburg-Harburg. As the heat flows in during the first period of roasting, the steam leaves the bean in the opposite direction. This type of heat transfer is very
Fig. 4.4
ineffective, but unavoidable. If the initial distribution of bean moisture is assumed to be even, a water content profile as depicted in Fig. 4.5 could form. Water transport starts when the local bean temperature exceeds the local vaporization temperature of water and when the vapor molecules can diffuse to the boundary of the bean. The local temperature of vaporization depends on the local pressure, amount of substances dissolved in the water (osmotic effects) and the state of the water. Vaporized water causes a rise in pressure in the cavities and an assumed pressure front moves into the bean. A reasonable temperature profile may show a slight point of inflection in the region of higher temperatures.
4.4.2 Specific heat of coffee New heat capacity data for an arabica/robusta blend were reported by Small and Horrel (1993). A ground green coffee (arabica/robusta blend) containing some moisture was investigated by DSC (differential scanning calorimetry). The measured heat capacity cp of the green coffee material increased from 2.8 kJ/kgK at 258C to a maximum of 6 kJ/kgK at 858C and declined to 2 kJ/kgK at 2108C. The temperature range above 2108C is termed the exothermic phase, and negative cp values indicate that the applied method is delivering
The roasting process: some parameters and influences.
Technology I: Roasting
97
such as the vaporization of water. Calculation leads to a theoretical vaporized water content of 0.1 wt% and less.
4.4.3 Thermal conductivity The measurement of the heat conductivity is not simple and reliable values are still lacking. In this respect, the necessity of investigating the whole beans showing inhomogeneous structures is rather problematic. Possibly measurements of temperature profiles in temperature and time regions with little interfering reactions and negligible changes of physical properties, will lead to calculated values in the near future.
4.4.4 Heat uptake of the bean
Fig. 4.5 Assumed profiles in the coffee bean.
heat capacity values including heat of reaction. During the DSC procedure the green coffee material is roasted and loses water. Small and Horrel (1993) also used the DSC roasted material to obtain values for coffee without water. In this case, they recorded only a slight temperature influence, with 2 kJ/kgK at 258C and an approximate mean value of 2.5 kJ/kgK between 258C and 2108C. At temperatures above 2108C the values declined significantly due to further roasting reactions. Comparison with the known data at 308C, which are 0.395 cal/g8C (1.65 kJ/kgK) for green arabica without moisture and 0.47 cal/g8C (1.97 kJ/kgK) for green arabica with 12% moisture (Clarke 1987, based on Raemy & Lambelet, 1982), illustrates that the newer values are significantly higher, especially in the case of the moist coffee. Elucidation of the temperature influence on the heat capacity is very welcome, although additional information on water content and degree of roast would be helpful for further understanding. The difference between the reported cp curves of green and DSC roasted coffee samples is not fully explained by the heat capacity of the moisture in the green beans. Since the heat capacity of water is of the order of 4.2 kJ/kgK, values as high as 6 have to be explained by the consideration of endothermic effects
The heat uptake of coffee beans during roasting can be calculated either from the heat requirement of technical roasters or by addition of the heat sensitive processes, like sensible and latent heats. Methods and published results were discussed in detail by Clarke (1987). In both cases the exothermic heat generated during roasting has to be estimated. Data were reported by Small and Horrel (1993), who published a diagram in which the necessary energy, E, to heat the bean from 18 to 1008 and to 2048C was plotted against the initial moisture content (Xw %) in the bean. In the depicted moisture range of 0 to 30% the relationships are linear and correspond to the equations E
1 to 100 c
KJ=kg
27 Xw
% 210
1 to 204 c
KJ=kg
28 Xw
% 450
and E
For the typical case of 12% initial bean moisture, E 1 to 1008C is approximately 530 kJ/kg and E 1 to 2048C is approximately 790 kJ/kg. Unfortunately, detailed information on coffee type and experimental method was not specified, but the title implies that the given values were based on green bean weight and included exothermic effects previously not known. The heat energy without exothermic effects, E0 results from the addition of sensible and latent heats. As discussed earlier, the final bean moisture is not zero and was chosen as 2% for the following calculations. Furthermore, an approximate moisture content of 6% was assumed for a bean temperature of 1008C. With a heat
98
Coffee: Recent Developments
capacity value of 1.9 kJ/kgK for the coffee substance without water, the theoretical E0; 1 to 1008C amounts to 234 kJ/kg and E0; 1 to 2048C to 527 kJ/kg. Surprisingly, these theoretical E0 values, based on green bean weight, are lower than the measured E values. The inverse effect would be more reasonable, since the heat from exothermic effects should lead to E < E0 . Comparison with other published values, all based on green bean weight, underlines that the values from Small and Horrel (1993) are somewhat high. Calculation with the premises listed above leads to E0;15 2108C = 520 kJ/kg, Sivetz and Desrosier (1979) calculated E0 = 710 and E = 561 kJ/kg and Vincent et al. (1977) reported that E = 364±410 kJ/kg for fluidized bed roasting. With the assumption of a 15% roast loss, a rough transformation of the data from Raemy and Lambelet (1982) leads to E = 300±500 kJ/kg. A concluding evaluation of these values is not possible, since bean end temperatures are often missing or other rough estimates are used. For example Raemy and Lambelet (1982) used coffee with an unusually low moisture content (7.5%), Vincent did not specify the moisture content at all and Sivetz and Desrosier (1979) assumed that 50% of the CO2 released from the bean delivers heat according to the reaction of carbon to carbon dioxide. The chosen amount of released CO2 with approximately 4% of green bean weight is much higher than values given by other authors, which are typically in the range 0.5±1%. Nevertheless, some tendencies are depicted in Fig. 4.6. In this context it is important to note that a simple heat balance such as E = E0 + Ereactions is only valid for an infinitely small period of time. An overall balance neglects the fact that the heat specific effects do not all occur simultaneously. For example, heat from an exothermic reaction at the end of the roasting process cannot evaporate water at the beginning.
Fig. 4.6
4.4.5 Temperature profiles in the bean Obviously the marginal zone of the bean heats up faster than the centre. The formation of a temperature profile in the bean can be assumed, and some authors have reported that extremely fast roasted coffee shows a lighter color in the centre than in the outer regions of the bean. Recently, new measurements of bean temperatures during roasting have been presented which show that the bean temperature takes a steep and linear incline and levels after a certain time. These effects were verified by our own measurements (Hobbie & Eggers, unpublished data). A single bean equipped with thermocouples was roasted in hot air. Temperature profiles across the bean have not yet been recorded, but simultaneous measurements of the temperatures at the bean surface and at the centre show a temperature difference, T, due to transient thermal effects. The temperature difference is greatest in the first minute ( 508C) and is still of the order of approximately 108C at the end of a long roasting period of 14 minutes, indicating the lasting vaporization of water (Fig. 4.7). A significant dent in both temperature curves appeared between 2 and 4 minutes. This sudden retardation of the temperature increase can be explained by rupture or abrupt expansion of the bean cavities, causing a pressure drop which in turn leads to evaporation of bean moisture. A rough calculation with heat of vaporization and a cp value for the bean of 2.5 kJ/kgK indicates that approximately 0.5 to 1% of the bean moisture evaporated additionally. Similar temperature curves from Schenker et al. (1999) do not show the above mentioned dent. This can be explained if we accept the model of Small and Horrel (1993; see Section 4.3.2). The green beans used by Schenker et al. (1999) contained more water (11.1%) and may therefore puff less abruptly. In the context of measuring or calculating temperature
Assumed tendency of quantitative effects based on green bean weight.
Technology I: Roasting
Fig. 4.7 Roasting of a single coffee bean suspended in hot air (Brazilian arabica, gas temperature 2508C, initial moisture 8%; Hobbie & Eggers, unpublished data).
profiles in the coffee bean, it should be noted that it is likely that not even the surface temperature is constant.
4.4.6 Heat transfer from gas to bean and overall heat transfer coefficient The convective heat transfer between solid spheres and gas can be calculated by the dimensionsless Nusselt equations (Bird et al., 1960), provided that the relative velocity sphere/gas, vrel , or at least the superficial gas velocity, vs , can be determined. For known Nu numbers, the heat transfer coefficient a can be calculated by Nu lgas /dbean . The thermal conductivity, l, and the
99
heat capacity, cp , of the fluid (here hot air) are influenced by the temperature. Assuming spherical coffee beans with an average and constant diameter of 6 mm, the tendency of Nu and a can be calculated for a given gas temperature. The heat transfer coefficient is plotted against the Reynolds number, Re, in Fig. 4.8. In a motionless system (vrel = 0), the heat transfer coefficient is around 14 W/m2 K and a minimum fluidization velocity (e.g. Re = 300) leads to a values of the order of 75 W/m2 K. The discussed heat transfer to single beans represents an ideal case, which is different from the heat transfer to a bed of beans. For the theoretical determination of the heat transfer to a bed of packed spheres, the bed or bulk porosity ' has to be known (Baehr and Stephan, 1998). Porosity values can be calculated from bean and bulk densities. Data from Vincent et al. (1977) lead to jgreen = 0.46 and jroasted = 0.53. An average value of 0.5 is close to the theoretical value for a cubic packing of true spheres (j = 0.48). In Fig. 4.8 the theoretical heat transfer coefficient for a packed bed with j = 0.5 is depicted. These calculated values are higher than those for single spheres and a roasting process is expected to show heat transfer values in the region between both curves. Two consequences are important. On one hand, it is possible to increase the heat transfer coefficient with higher velocities, but on the other hand the improvement is limited due to the low gradient of the Nu functions. To illustrate this, an increase in the velocity in a packed bed beyond the theoretical minimum fluidization velocity is considered. Process parameters like these can be achieved in a Burns System 90 roaster (Section
Fig. 4.8 Nusselt number (Nu) and heat transfer coefficient (d) for coffee beans assumed to be as for spheres. * Bird et al., 1960. ** Baehr et al., 1998.
100
Coffee: Recent Developments
4.6.3). Increasing the velocity by 20%, from Re 300 to 360, increases a by just 8%. In summary, the size of the theoretical heat transfer coefficient a is of the order of magnitude of 40±120 W/m2 K. The overall heat transfer coefficients include the heat transfer from gas to bean and conduction into the bean matrix. Values were reported by Sivetz and Desrosier (1979) for a large continuous drum roaster (0.5 Btu/sq ft8F = 0.24 cal/h cm2 8C = 2.84 W/m2 K) and by Nagaraju et al. (1997). The latter group investigated a spouting bed and found an overall heat transfer coefficient of 14 W/m2 K with the assumption of constant material data (cp , bean surface, bean mass). Recalculation using the crude data yields to an average overall heat transfer coefficient of about 10 W/m2 K, still indicating that the heat transfer in a spouting process is faster than in a conventional drum roaster. In order to determine whether the transfer to the bean surface or the conduction is limiting; the thermal conductivity of coffee beans has to be known. If the coffee can be assumed to behave like wood, as suggested by Sivetz and Desrosier (1979), a conductivity value of l = 413 6 10 6 cal/s cm K = 0.173 W/mK could be used. This value, combined with an average bean radius r of 0.003 m, gives l/r = 58 W/m2 K. Comparison of this value with the theoretical a leads to the conclusion that neither conduction nor the transfer to the bean surface is controlling. The very varied overall transfer coefficients that have been reported indicate that the respective importance of either factor depends on the range of the relative velocity gas/bean. In particular, the current tendency towards shorter roasting time enforces the consideration of conductive Table 4.2
effects. If it is assumed that there is no heat transport resistance at the boundary of the bean and that the surface of the bean (considered to be a sphere with a radius of 3 mm and the same l as wood) has a constant temperature of 2008C. Calculation of heat conduction into the sphere (Carslaw and Jaeger, 1989) shows that the centre reaches 1308C after 30 seconds and 1808C after one minute. It is likely that the gas filled cavities, which form during the roasting process, together with endothermic reactions and opposed mass transport will also hinder the heat flow and it is evident that conductive effects should be considered.
4.5 SOME ASPECTS ON FUTURE SCIENTIFIC RESEARCH Despite intensive research, the roasting mechanisms are not fully understood. Although many process factors have been varied and different parameters recorded, due to the enormous complexity of the subject a well-rounded model is still missing. Naturally a scientific study has to concentrate on certain aspects of the field and therefore neglect others. Nevertheless, it would be very helpful to report all available relevant information in order to combine and compare results of different studies. However, some authors do not even specify the starting material (bean type, initial moisture). On the basis of the current knowledge we propose that the parameters listed in Table 4.2 should be reported. Detailed observations of bean popping (visual observation) linked with other bean data (temperature,
Important physical values for research reports on coffee roasting. Minimum
Additional
Green beans
Bean type Average bean size Average bean mass Moisture
Average bean volume Average bean density Average bulk density
Roasted beans
Average bean size Average bean mass Moisture (give detailed method) Color Appearance
Average bean volume Average bean density Average bulk density Degassing CO2 Soluble solids
Process
Roaster type, name Roasting time Roast gas temperature Product end temperature (give detailed method)
Roast gas throughput Time of bean popping Analytic roast gas
Technology I: Roasting
density, etc.) are missing and physical parameters like swelling should be combined with new results in the field of chemical composition and reactions. Generally, more data points should be recorded in studies regarding the progression of bean properties, especially in the first 2 minutes. Mass, volume, porosity and moisture data known today do not indicate whether or not there are discontinuities. Measurement of temperature profiles within the bean could give new information on the heat and mass transfer effects, might facilitate calculation of thermal conductivity values and help to optimise roast gas temperature profiles. To improve understanding of the swelling mechanisms, determination of the structural strength and elasticity of different layers of the beans would be useful. In order to model the heat transfer in technical roasting processes, the average relative bean velocities and bed porosities have to be measured or calculated. Computation of the heat transport due to convection and conduction should be attempted in the future and the influence of temperature and the course of roasting on the local bean properties should not be neglected. To verify such a model, results could be compared with temperature profiles gained from experiments with larger spheres or half ellipsoids, which are made of a well-defined and homogeneous material. The influence of radiation between the surfaces of the heated beans has not been investigated so far. Generally, radiation increases with increasing surface temperature and should lead to a more homogeneous temperature distribution in the batch, especially during the final stage of roasting. Inclusion of the roasting reactions, the local heat of vaporisation and the resulting mass transfer will lead to a highly complex model, that will certainly require much more information on the roasting mechanisms than is known today. Regarding the whole roasting process, it should be noted that so far little information on cooling methods and their respective mechanisms has been reported.
4.6 INDUSTRIAL ROASTING EQUIPMENT Roasters have been designed along different mechanical principles, so that beans and hot gas come into contact in different ways. Most available machines move the beans with mechanical agitators, others with the assistance of the hot gas flow. Various companies supply their specialized design. Some of today's large industrial roaster manufacturers are Barth Ludwigs-
101
burg, Gothot (Probat Group), Neuhaus Neotec (Kahl Group) and Probat in Germany; Burns (Probat Group) and Wolverine Proctor & Schwartz in the USA; Tecaire in Spain; STA Impianti and Scholari in Italy and Cia Lilla in Brasil. Modern Process Equipment Inc (ILL) is the North American sales agent for Neuhaus Neotech and Equip for Coffee (Cal) for Probat. Used and rebuilt roasting plants are, for instance, supplied by Union Standard Equipment (New York, USA). In Germany, BuÈhler and Lurgi have stopped their production of roasting machines while Nepro has started with a new roasting plant design. Roasting plants existing pre-1986 are described by Clarke (1987). In the following, these roasters are summarized as `traditional roasters'. In the last decade several new roaster designs have been developed and marketed. These new designs are now briefly described, together with their special benefits.
4.6.1 Traditional roasters Traditional roasters are horizontal rotating drum roasters for batch or continuous operation, vertical fixed drum roasters with paddles and rotating bowl roasters. Just about all of these are still in operation and for sale on the market. Generally, a tendency towards a larger coffee product variety has led to smaller uniform charges and therefore the demand for large continuous roasters is decreasing. Burns no longer markets its continuous horizontal drum roaster. Due to modern controls, batch roasters can make very consistent products while being able to vary the heat input over time. Roasters for operation under pressure and for steam roasting are not marketed today, although there is patenting activity in this field.
4.6.2 Fluidized beds The first industrial size fluidized bed roaster was the Lurgi `Aerotherm' roaster, which has not been marketed since Lurgi stopped their roasting activities. A continuous fluidized bed type roaster is the `Jet zone roaster' from Wolverine Proctor & Schwartz. An oscillating conveyor moves the beans through the roaster and jets of hot gas are directed into the bed through tubes from above. Neuhaus Neotec has introduced a new rotating fluidized bed roaster (RFB). The roasting and cooling chamber are geometrically identical and their shape leads to what is called a rotating fluidized bed. Figure 4.9 illustrates the mode of action of this roaster. The gases entering from the
102
Coffee: Recent Developments
Fig. 4.9 Rotating fluidized bed roaster flow diagram: 1 Roasting chamber; 2 cooling chamber; 3 cyclone hotair recirculation system; 4 channel burner; 5 recirculation fan; 6 catalytic afterburner system; 7 product feed bin; 8 roasted coffee discharge bin; 9 cooler fan; 10 cyclone cooler system; 11 water quenching. (By courtesy of Neuhaus Neotec GmbH.)
Technology I: Roasting
bottom fluidise the beans, the shape of the chamber directs them to the opposite sloped chamber wall and they move towards the bottom again. Thereby, the RFB leads to an intensive mixing and movement of the beans. The gas flow can be well above the minimum fluidization velocity without the risk of blowing beans out of the roasting section. Batch operation allows different roasting processes with temperature profiles. At present a special version of the RFB (type RH 2010), with a wide range of gas to coffee ratios, is being developed. It is designed for large capacities, long roasting time and high density coffee. RFB plants have been built since 1984 and operate in approximately 100 larger coffee companies. In Magdeburg, Germany, the RoÈstfein Kaffee GmbH set up a new fluidized bed roaster for torrefacto coffee (with the addition of sugar in order to caramelize the coffee beans) in 1999 (MoÈrl, personal commu-
103
nication). The annual capacity is quoted as 10 000 t roasted coffee or 4000 t torrefacto coffee. Batches of 150 to 200 kg coffee are roasted at temperatures between 2308C and 3608C with a circulating roasting gas/steam atmosphere. A new design is offered by Nepro, in Germany, named the Nepro Vortex Fluidat. This `multiple zone quasi-continuous belt process' transports the beans on a perforated conveyor belt through three heating, one quenching and one air cooling zone (Fig. 4.10). Gases enter the bed from the bottom, causing variable bean movement from spouting bed to fluidization. Independently adjustable heating or cooling zones allow a temperature profile during roasting in combination with continuous processing. Nepro states that roasting of broken beans is possible. One industrial size plant has been installed so far.
Fig. 4.10 Nepro Vortex1 I: 1 Roasting zone I; 2 roasting zone II; 3 roasting zone III; 4 precooling/quenching zone; 5 cooling zone/air; 6 recirculation fan; 7 roasting belt; 8 burner; 9 rotary valve; 10 burner chamber/chaff separator combination; 11 precooling fan; 12 waterpump; 13 cooling fan; 14 cooling air inlet; 15 feed hopper; 16 feeder; 17 cooler chaff separator; 18 afterburner; 19 cooling air outlet; 20 afterburner outlet; 21 rotary valve; 22 conveyor belt; 23 exhaust fan; 24 burner; 25 chaff collector; 26 chaff discharge. (By courtesy of Nepro GmbH.)
104
4.6.3 Packed bed roasting In 1995 Burns introduced a packed bed roaster suited for conventional and high-yield, low density roasting (Anon, 1995). Although this roaster is under review at present, the principle is described in the following. A batch of beans is roasted in a conical chamber with high velocity roasting gas (Fig. 4.11). The gas enters tangentially, passes through a special pattern of louvers and whirls the beans in the chamber, creating a spinning packed bed. This packed bed rotates in a horizontal circle with a thickness of 30 to 50 mm (Anon, 1996) and all beans are believed to be forced into a predestined travel path. It is possible to increase the gas speed beyond the fluidization speed of the beans since the coffee bed is kept tightly packed by centrifugal forces. These high velocities are believed to lead to high heat transfer rates so that roasting can be performed at lower temperatures (typically 2758C to 3008C) and for short times. Burns states that these parameters reduce the loss of volatile aromatic components and `high yield' beans show slightly higher densities compared to traditional similar swollen `high yield' ones. The cooling chamber with identical geometry is positioned below the roasting chamber and the coffee flows through the plant under gravity. Burns states that the energy consumption equals 60% of their horizontal drum roaster due to shorter roasting times and lower temperatures and that roasting of broken beans is possible. An interesting detail is process control by outlet gas temperature measurement in combination with precisely controlled inlet gas temperature.
Fig. 4.11 Burns packed bed roaster, System 90 type. Schematic drawing, showing a conical roasting or cooling chamber with gas entering tangentially.
4.6.4 Roasting with heated cooling gas Probat patented a new roaster arrangement in which essentially all of the cooling waste air is heated and fed to the roaster. This leads to an increased roasting gas throughput and a lower roasting gas temperature, in
Coffee: Recent Developments
the range of 3008C to 3608C. Stated advantages of this new roaster are more uniform roasting, reduced emissions, increased thermal efficiency and a reduced structural outlay. The roasting machine uses heat exchangers combined with the burner to recover the heat of the roast gas. Smoke from the combustion does not come into direct contact with the beans. At present, the first industrial size roaster of this new design is under construction. Simplified flow sheets and a schematic drawing of the burner encapsulated by heat exchangers are given in the corresponding US patent 5 718 164 (1998).
4.6.5 Technical data and capacities The above mentioned types of roasting plants are supplied in different sizes. Data on several currently available machines (from manufacturers' brochures) are given in Table 4.3.
4.6.6 Roaster patents 1986±99 During the period 1986±99 patenting was intense, often directed at small-scale household or coffee shop roasting equipment. These patents deal with fluidized bed roasting and methods to reduce the level of pollutants in the exhaust gases vented from the roaster. US patents connected with industrial applications are listed in Table 4.4. In the field of industrial roasting, most US patents focus on fast roasting. The effects, benefits and drawbacks of fast roasting processes have been discussed in detail in sections 4.2.4 and 4.3. Three US patents deal with the gas circulation. A method to avoid contact between food and the smoke of the combustion roasting, with ambient air heated via heat exchangers, was patented by Farina (US 5 372 833). Besides a basic schematic flow diagram, a variation with hot gas recirculation from the roaster to the combustion chamber is given. In 1999 Felip patented the wellknown hot gas recirculation in combination with a secondary combustion for the exhaust stream (US 5 928 697). Available information on the new patented Probat roaster (US 5 718 164), which uses the heated cooling gas for roasting, has been discussed in section 4.6.4. The method of conveying the beans through the process is also an objective of patents. A continuous `revolving drum roaster' was patented by Pera (US 4 924 765) in 1990. Several drums containing the coffee rotate in a larger rotating drum such that the coffee passes through the successive processes of heating, roasting, cooling and discharging. This design seems to
Technology I: Roasting
Table 4.3
105
Data on some industrial roasting machines, capacity 240 to 5000 kg/h.
Roaster (manufacturer, type)
Roasting time (min)
Coffee capacity (kg/h)
(lb/h)
(kg/batch)
Horizontal drum, batch operation Burns (USA) ± Thermalo 23, 24 Burns (USA) ± Thermalo Series 4200 3.5±15 LEOGAP (Brazill)±2000 Compacto LEOGAP (Brazil) ± RaÂpido Cia Lilla (Brazil) ± CG/CO Cia Lilla (Brazil) ± COA OPS (Italy) ± MPS Probat (Germany) ± G 8±13 Probat (Germany) ± CN 12 Probat (Germany) ± R 8±10 Probat (Germany) ± Excelsior 5±12 STA Impianti (Italy) ± M3 10±15 STA Impianti (Italy) ± Futura 10±15 Tecaire (Spain) TNA Tecaire (Spain) TNA F
545/907 544/1179 540/900/1260 900/1500/2100 625/1250/2500 400/800/1600 120±1200 240/280 240/480/1200/2000 1000/1500/2000/3000 480 260±2880 240±1920 700/1500/3000 700/1000
1200/2000 1200/2600 1191/1984/2778 1984/3307/4630 1375/2750/5500 880/1760/3520 264±2642 528/617 528/1058/2646/4409 2205/3307/4409/6614 1057 573±6342 528±4228 1543/3307/6614 1543/2205
280
For torrefacto coffee: Probat (Germany) ± CN Tecaire (Spain) TTA Tecaire (Spain) Complet
180/360/900/1500 420/700/1200 600/1000/1500/2500
367/794/1984/3307 926/1543/2646 1323/2205/3307/5512
140/240/400 120/200/300/500
2500/3500/4000 2500/3500/4500
5512/7716/8819 5512/7716/9921
Ð Ð
15±17
Horizontal drum, continuous operation Probat (Germany) ± RC 1.5±8 Neuhaus Neotec (Germany) ± C 1.5±6 or 5±12
180/300/420 180/300/420 125/250/500 120/240/480
60±480 60±480 140/300/500 60±90/80±100
Rotating bowl Probat (Germany) ± RZ
2.5±9.5
2500/3500/4000/5000
5512/7716/8819/11023
Ð
Vertical fixed drum with paddles Gothot (Germany) ± RN Gothot (Germany) ± RT
3±8 3±9
1000/2000/3000/4000 1000/2000/3000/4000
2205/4409/6614/8819 2205/4409/6614/8819
Ð Ð
Packed bed roasting Burns (USA) ± System 90
1±5
500/1000/3000/4000
1100/2205/6614/8819
Ð
300±3000 200±1500
660±6600 440±3300
15±300 20±90
240
529
16±17
500±6000 453±5443
1101±13211 1000±12000
Ð Ð
Fluidlzed bed roasters, batch operation Neuhaus Neotec (Germany) ± RFB 1.5±8 Neuhaus Neotec (Germany) ± RFB G Neuhaus Neotec (Germany) ± RH 2010 Leogap (Brazil) Turbo 3500 3.5±4.5 Fluldized bed roasters, continuous operation Nepro Vortex (Germany) ± Fluidat 1±15 Wolverine Proctor & Schwartz (USA) 1±3 ± Jetzone
be rather complicated to the authors. Neuhaus Neotec secured its new RFB design in German patent DE 3116723 C2 from 1984 and Nepro has applied for two US patents (09/335 247 entitled `Apparatus and
method for the thermal treatment of granular material' and 09/427 975 `Perforated bottom plate for the production of fluidized bed') relating to the Vortex Fluidat roaster.
106
Table 4.4
Coffee: Recent Developments
US patents in the field of industrial coffee roasting 1986±99.
Patent number
Title
Year
Inventor
Assignee
6 000 144
Method for heating and cooling food products
1999
Bussmann, P. et al.
Ð
5 972 409
Soluble instant coffee prepared from extract obtained from green coffee
1999
Liu, R. T.-S. et al.
Nestec S.A.
5 928 697
Purification of roaster gases
1999
Felip, A.
Nestec S. A.
5 721 005
Fast roasted coffee providing increased brew strength and darker cup color with desirable brew acidity
1998
Gutwein, R.W. et al.
Procter & Gamble
5 718 164
Arrangement for roasting vegetable bulk material, such as coffee beans
1998
Finken, H. et al.
Probat
5 681 607
Process for roasting coffee beans with steam
1997
Maki, Y. et al.
General Foods
5 372 833
Roasting system and method
1994
Farina, S.
Petroncini SPA
5 368 875
Method of manufacturing rich-flavored roasted coffee beans and ground roasted coffee
1994
Hibi, H. et al.
Nagoyasei-raku Co
5 322 703
High-yield roasted coffee with balanced flavor
1994
Jensen, M.R. et al.
Procter & Gamble
5 160 757
Process for making reduced density coffee
1992
Kirkpatrick, S.J. et al.
Procter & Gamble
5 019 413
Process for improving the quality of robusta coffee
1991
Becker, R. et al.
Jacobs Suchard AG
4 988 590
Ultra fast roasted coffee
1991
Price, S.E. et al.
Procter & Gamble
4 985 271
Process for treating coffee beans to make a bettertasting coffee
1991
Neilson, D.H. et al.
Procter & Gamble
4 924 765
Equipment for roasting coffee, hazelnuts, peanuts and similar commodities
1990
Pera, B.
Ð
4 849 625
Device, applicable to ovens, for monitoring the color of coffee and similar commodities during the course of roast
1989
Camerinie Porzi, P.C.
Officine Vittoria SpA
4 737 376
Coffee roasting method
1988
Brandlein et al.
General Foods
Roasting without hot gases can be performed according to two further patents. It is known that coffee beans can be roasted by bringing them into contact with heated, solid materials in granular form. Bussmann suggested using the granular material (preferably silica zeolites) for roasting and cooling in a combined process (US 6 000 144). For soluble coffees the typical sequence of roasting followed by percolation can be reversed. According to US patent 5 972 409, dried green coffee extract can be roasted (`heat caramelized') with twin-screw extruders. The green powder is heated to temperatures of 1308C to 2408C for up to 5 minutes. In accordance with the known influence of pressure on roasting, the authors prefer an extruder pressure of 10 bars. Two different green coffee extract fractions are treated separately with adapted process parameters.
Patents focused on automated termination have been discussed in section 4.2.5. Flavor improvement is the aim of other patents. Hibi stated that rapid cooling of the beans to 7178C or lower leads to a substantial delay in flavor deterioration (US 5 368 875). Data on CO2 -development influenced by cooling temperature and cooling speed are given. Other patented ways to produce a richer flavor include the treatment of partially roasted beans with an alkaline solution (US 4 985 271) or of green robusta beans with steam at 1358C to 1408C (US 5 019 413). Steam roasting is in the focus of a patent by Maki (US 5 681 607 and WO 95/20325). Initiation of hydrolysis under steam leads to improved flavour. The problem of too high acidity is avoided with a two stage steam roasting process: a pressure roasting
Technology I: Roasting
(6.5 to 20 bar) followed by roasting under atmospheric pressure conditions.
REFERENCES Anon (1995) Roasting revolution. Coffee Cocoa Int., 2, 38. Anon (1996) Packed bed vs. fluid bed. Tea Coffee Trade J, 168. Baehr, H.D. & Stephan, K. (1998) Heat and Mass-Transfer Springer, Berlin. Bersten, I. (1993) Coffee floats Tea sinks ± Through History and Technology to a Complete Understanding. Helian Books, Sydney. Bird, R.B., Steward, W.E. & Lightfoot, E.N. (1960) Transport Phenomena. John Wiley & Sons, New York. Carslaw, H.S. & Jaeger, J.C. (reprinted 1989) Conduction of Heat in Solids. Clarendon Press, Oxford. Chitester et al. (1984) quoted from: Kunii, D. & Levenspiel, O. (1991) Fluidization Engineering. Butterworth±Heinemann, Boston. Clarke, R.J. (1987) Roasting and grinding. In: Coffee, Vol. 2, Technology (eds R.J. Clarke & R. Macrae). Elsevier Applied Science, Barking. Da Porto, C., Nicoli, M.C., Severini, C., Sensidoni, A. & Lerici, C.R. (1991) Study on physical and physiochemical changes in coffee beans during roasting. Note 2. Ital. J. Food Sci., 197± 207. Hashim, L. & Chaveron, H. (1996) Use of methylpyrazine ratios to monitor the coffee roasting. Food Res. Int., 28 (6), 619±23. Kazi, T. & Clifford, M.N. (1985) Comparison of physical and chemical characteristics of `high yield' and `regular' coffees. In: Proceedings of the 11th ASIC Colloquium, LomeÂ, pp. 297±308. ASIC, Paris, France. Maier, H.G. (1985) Zur Zusammensetzung kurzzeitgeroÈsteter Kaffees. Lebensmittelchem. Gerichtl. Chem., 39, 25±9. Massini, R., Nicoli, M.C., CassaraÁ, A. & Lerici, C.R. (1990) Study on physical and physiochemical changes in coffee beans during roasting. Note 1. Ital. J. Food Sci., 123±30. Nagaraju, V.D., Murthy, C.T., Ramalaksshmi, K. & Srinivasa Rao, P.N. (1997) Studies on roasting of coffee beans in a spouted bed. J. Food Eng., 31, 263±70.
107
Ortola, M.D., LondonÄo, L., GutieÂrrez, C.L. & Chiralt, A. (1998) Influence of roasting temperature on physiochemical properties of different coffees. Food Sci. Technol. Int., 4, 59±66. Radtke, R. (1975) Das Problem der CO2 -Desorption von RoÈstkaffee unter dem Gesichtspunkt einer neuen Packstoffentwicklung. In: Proceedings of the 7th ASIC Colloquium, pp. 323±33. ASIC, Paris, France. Raemy, A. & Lambelet, P. (1982) A calorimetric study of selfheating in coffee and chicory. J. Food Technol., 17, 451±60. Schenker, S., Handschin, S., Frey, B., Perren, R. & Escher, F. (1998) Verification of mercury intrusion into coffee beans by scanning electron microscopy and X-ray microanalysis. Scanning J. of Scanning Microscopies, 273. Schenker, S., Handschin, S., Frey, B., Perren, R. & Escher, F. (1999) Structural properties of coffee beans as influenced by roasting conditions. In: Proceedings of the 18th ASIC Colloquium (Helsinki), pp. 127±135. ASIC, Paris, France. Sivetz, M. (1991) Growth in use of automated fluid bed roasting of coffee beans. In: Proceedings of the 14th ASIC Colloquium (San Francisco), pp. 313±18. ASIC, Paris, France. Sivetz, M. & Desrosier, N.W. (1979) Coffee Technology. AVI, Westport, Connecticut. Small, L.E. & Horrel, R.S. (1993) High yield coffee technology. In: Proceedings of the 15th ASIC Colloquium (Montpellier), pp. 719±26. ASIC, Paris, France. Swern, D. (ed.) (1979) Bailey's Industrial Oil and Fat Products, Vol. 1, 4th edn, p. 210. John Wiley & Sons, New York. Vincent, J.-C., Arjona, J.-L., Rios, G., Gibert, H. & Roche, G. (1977) TorreÂfaction du cafe en couche fluidiseÂe gazeuse. In: Proceedings of the 8th ASIC Colloquium (Abidjan), pp. 217±26. ASIC, Paris, France. Zimmermann, R., Heger, H.J., Yeretzian, C., Nagel, H. & Boesel, U. (1999) Application of laser ionization mass spectrometry for on-line monitoring of volatiles in the headspace of food products: roasting and brewing of coffee. Rapid Com. Mass Spectrom., 10, 1975±9.
Chapter 5
Technology II: Decaffeination of Coffee W. Heilmann Bremen, Germany 5.1 INTRODUCTION Caffeine is a physiologically active component in coffee, which has been studied intensively. It was discovered by Runge in 1820; and the chemical name of this purine is 1.3.7 trimethylaxanthine (see Fig. 5.1). It consists of needle-shaped crystals with a melting point of 2368C. Coffee beans contain between 0.8 and 2.8% caffeine, depending on species and origin, and it contributes to 10 to 30% of the often desired bitterness of the coffee beverage. There are only negligible losses of caffeine in the roasting process. Caffeine is mainly used in soft drinks, but also together with other active agents in remedies for headaches, cardiac insufficiency, migraine or disorder of the respiratory centre. Updated information on the physiological effects are described in detail, in Chapter 8. O H3C
O
N
CH3
N
CH3
Fig. 5.1
. Swelling the raw beans with water in order to solubilize the caffeine±potassium chlorogenate complex, and to make caffeine available for extraction. . Extracting the caffeine from the beans with a solvent. . Steam stripping to remove all solvent residues from the beans (if applied). . Regenerating adsorbents (if applied). . Drying the decaffeinated coffee beans to their initial moisture content. The different decaffeination procedures can be classified into three major groups:
N
N
maximum caffeine concentration of 0.1% related to the dry mass; in the US, it means less than 3% of the amount initially present in the beans. Principally all decaffeination processes consist of five steps:
Caffeine (1.3.7 trimethylxanthine).
In order to minimize the `negative' physiological effects and still keep the desirable attributes of a coffee beverage, quite a number of decaffeination processes have been developed. In order to minimize flavour and aroma losses, the commercial decaffeination of coffee is at present carried out on the green coffee beans before roasting (although the patent literature also outlines more complex processes for decaffeination of roasted coffee extract) (Lack et al., 1993). `Decaffeinated coffee' means in the EU countries a
(1) Decaffeination with chemical solvents such as methylenchloride or ethyl acetate was first successfully applied by Roselius in `Kaffee HAG'. Although these solvents are approved under the rules of all food legislations, some experts have raised questions about the safety of residual solvents in the coffee. The counter argument is that effective steps are taken to remove residues almost totally and one would need to drink a million cups or more before the solvent shows up in the blood. (2) Water decaffeination (first mentioned in 1941 by Berry and Walters), uses green coffee extract with equilibrium quantities of non-caffeine soluble solids and removal of caffeine from the extract with dichlormethane in liquid±liquid extraction. The application of preloaded activated carbon for the caffeine adsorption instead of solvent extraction was the basis for the `Swiss water process' (Fischer, 1979). (3) The selective solubility of caffeine in moistened supercritical carbondioxide was first applied by
108
Technology II: Decaffeination of Coffee
HAG, Germany, and is based on a patent of `Studiengesellschaft Kohle in MuÈhlheim' (Zosel, 1965) referring to the solvent-like properties of supercritical gases (`destraction'). All these fundamental procedures have been described in the corresponding patents and scientific literature (Katz, 1987; Clarke, 1988). The development of caffeine-free coffee plants is reported in Chapter 11. The main object of this chapter is to inform about new developments in all relevant areas in the past 15 years, concentrating on: . the three basic decaffeination processes (with new solvents, process steps, etc.); . systems for caffeine recovery from activated carbon; . economical aspects.
5.2 SOLVENT DECAFFEINATION In spite of many new developments, the original method of extraction of caffeine from prewetted green coffee beans with solvents is still the most preferred process. Due to its relatively low investment and operating costs and the high coffee quality, more than 50% of the worldwide capacity is based on solvent decaffeination. More than 30 solvents have been tested and described in the literature of the past 15 years, but dichloromethane (DCM) and ethylacetate still predominate representing about 98% of all solvent processes. The effects of the solvents used on health and environment have been totally minimized. The amount of solvent required to decaffeinate 1 t of green coffee is less than 10 kg. A maximum of 2 mg/kg of solvent residues is allowed in decaffeinated roasted coffee; in optimized plants a level of 0.3 mg/kg is achieved. Maximum concentrations in the waste water (0.01 ml/l) and the working environment (0.015 g/m3 ) have to follow the local legislation. Dichloromethane is not mentioned in the list of `ozone depleting substances' (Montreal Protocol). Use of DCM will be covered by the VOC Directive (Directive on the limitation of the emissions of volatile organic compounds due to the use of organic solvents in certain processes and industrial installations), which is in a state of implementation (Reference: OJEC 85, 29.03.1999). Member states must adopt national measures before April 2001. For the time being, neither DCM nor the coffee industry, in general, is specifically mentioned, but it clearly will be. It is expected that
109
emissions of VOCs will be limited to a `technically and economically feasible' value. Improvements have been made in process technology, mainly in changing from batch operation to semicontinuous processes. Morrison et al. (1984) described an improved countercurrent extraction process for accelerated decaffeination. Turbulent flow of the extracting solvent is used to obtain a caffeine extraction time of only 3 to 5 hours. Use of multiple vessels in countercurrent operation (Fig. 5.2) and the reduced extraction time decrease the loss of non-caffeine solubles, resulting also in an improved quality of the decaffeinated coffee. Another invention relates to a process whereby extraction of caffeine and substances that are potentially detrimental to health, from green coffee, is carried out by use of sufficient acid to remove chlorogenic compounds from the coffee (Van der Stegen, 1985). If the typical organic solvent is mixed with an acid (such as acetic, formic or citric acid, up to 25%) and water (up to 15%) in the decaffeination process, a `mild' coffee is produced in which caffeine, wax and chlorogenic acid compounds are removed simultaneously. Another process was described by Katz (1984), in which green coffee beans are decaffeinated at a moisture content near that employed for shipping and storage. The green beans are mixed with dimethyl sulphoxide, an aprotic solvent (boiling point 1898C). It appears that the dimethyl sulfoxide is capable of breaking up the complex between potassium chlorogenate and caffeine even in the absence of high water concentration. To obtain an effective rate of decaffeination, the concentration of caffeine within the dimethyl sulfoxide should be maintained at a low level (0.05±0.3 g/kg solution). As other coffee components, such as sugars, are dissolved to varying degrees by dimethyl sulfoxide, it is desirable to maintain the concentration of these other soluble materials near their point of saturation. Caffeine can be adsorbed out of the solution with preloaded activated carbon. Zeller and Saleeb (1999) report experiments which used caffeic acid (3,4-dihydroxycinnamic acid) in order to increase the batch decaffeination of methylene chloride, ethyl acetate and soybean oil. Mixing these non-aqueous solvents with a suspension of caffeic acid crystals in water causes rapid crystallization of a caffeine±caffeic acid complex at room temperature. The extent of decaffeination obtained is much greater than when the same solvents are mixed with water only (Fig. 5.3). First results indicate that this technique may reduce water use, increase caffeine selectivity and reduce coffee flavour losses in downstream operation.
110
Coffee: Recent Developments
22 20
21 23
11
51
10
9 N
8 N
7 N
6 N
5
4
3
2
1
N
102
13
104
50
103
106
N
28 25
Fig. 5.2 The accelerated decaffeination process (Morrison et al., 1984) 1 vessel with cleaned decaffeinated beans, 2,3,4,5 solvent removal, 6,7,8,9,10 in decaffeination, 11 fresh beans, 13 discharge lorry, 20,21 blending bins, 23 mixer for prewetting, 25 fresh solvent, 28 solvent saturator, 50 steam, 51 to vacuum source, 104 solvent discharge, 106 solvent recovery system.
100
2:1 Water:MC volume
90 % Decaffeination (2:1 CA:caffeine)
Another improved roast and ground coffee extract decaffenation method has been published (Jones et al., 1985). Before the production of dry instant coffee, the extract is mixed with a halogenated solvent (e.g. DCM) in a liquid±liquid extraction. Caffeine and some coffee solids are transferred into the solvent phase, which is then mixed with water in a second liquid±liquid extraction step. Most of the non-caffeine coffee solids are absorbed in water, which can be recycled into the decaffeinated extract; the caffeine has thereafter to be separated from the solvent by distillation. Another decaffeination method for green and roast coffee extracts has been developed (Kaleda et al., 1986). The caffeine-containing extract solution is mixed with caffeic acid crystals in the presence of water. The caffeine and the caffeic acid form an insoluble caffeine/ caffeic acid complex. This complex can be separated from the extract using techniques like filtration or centrifugation.
Water + CA
80 70 60 50 40 30 20 Water only 10 0
0
1
2
3
4
5
6
7
8
9
10
% Caffeine in MeCl2 (wt/wt)
Fig. 5.3 Extent of decaffeination of MeCl2 (Dichloromethane) by water and water plus caffeic acid (CA) (Zeller & Saleeb, 1999).
5.3 WATER DECAFFEINATION The main reason for replacing organic solvents by water as extraction agent was the anticipation of DCM being banned by the food authorities. But in spite of several attacks in different media, DCM is still approved as a food solvent by relevant organizations.
Technology II: Decaffeination of Coffee
Although the water process is much more expensive than solvent decaffeination, it developed to a procedure that is now used in almost 22% of all decaffeination plants, based on the claim that it is a `natural' process. In the first stage it was only an `indirect solvent process' because the caffeine-containing liquid green coffee extract had been treated in a liquid±liquid extraction step with organic solvents for decaffeination and then recycled to the next extraction step (Berry, 1943). Only by the introduction of caffeine adsorption on to activated carbon or an ion exchange resin did it become a pure water process. Originally, fresh water was used for decaffeination of the swollen beans. After removal of the caffeine the solution was concentrated and readsorbed on the predried decaffeinated beans from which it had been obtained, thus replacing non-caffeine solids otherwise lost. Losses of coffee solids are usually higher with adsorption processes than with solvent processes. But losses may be reduced by precharging the adsorbent with substances having the same molecular structure and size (e.g. glucose or saccharose), the degree of loading being adjusted so that an adsorption equilibrium exists (Fischer & Kummer, 1979). General Foods published a patent describing the preloading of the activated carbon with ethyl cellulose. The activated carbon is first contacted with ethyl cellulose solution and then dried before application in the liquid green coffee extract (Hinman & Saleeb, 1985). Douwe Egberts demonstrated, in their patent, that the caffeine adsorption can be optimized by using an activated carbon with a specific adsorption capacity for caffeine of at least 100 g/kg carbon and a specific selectivity (load ratio of caffeine/non-caffeine solids) of at least 0.2 combined with a countercurrent `droplet extraction'. This process can be optimized so that the caffeine can be removed ± completely, if necessary ± only by a special combination of adsorption and desorption steps. An adsorption agent has to be used that adsorbs caffeine preferentially from the aqueous equilibrium extract; the caffeine itself is obtained in a practically pure solution (Mooiweer, 1982). Jacobs Suchard (Heilmann, 1991) developed a modified `Secoffex' decaffeination process with an acid-washed `hard coal' carbon, preloaded with sugar to effect higher selectivity for caffeine. The adsorptive capacity is 11 g caffeine per 100 g activated carbon at 808C. The use of a coconut-based carbon has reduced abrasive action when handled during continuous process operations. The extraction of the water swollen beans takes place with a caffeine-free green coffee extract in equilibrium with the beans' solid
111
concentration (Fig. 5.4). A new carousel extractor with rotating sections of beans has been introduced that provides a higher capacity for the beans, better quality and greater bacteriological safety; the closed system prevents the loss of volatile compounds. The caffeineladen equilibrium extract is then fed through activated carbon (also in a carousel extractor) and back through the beans. Reactivation is accomplished with a threestage fluidized bed reactor; caffeine cannot be recovered. An imprint polymer has been proposed as an alternative to active carbon and microporous resins for decaffeination of coffee (Hay et al., 1995). The polymer has non-covalent recognition sites for caffeine that can selectively remove caffeine from an aqueous extract derived from green or roasted coffee beans. The caffeine imprint polymer is formed by polymerising a monomer in the presence of caffeine. Such polymers can remove more than 99% of free caffeine without substantial loss of coffee flavours. A cartridge containing a caffeine imprint polymer can be applied. Nestle proposed the adsorption of caffeine from a liquid green coffee extract by a non-ionic, microporous resin with an average pore size of 6±40 AÊ, and a specific surface of 400 m2 /g (Blanc & Margolis, 1981). These resins have a great affinity for caffeine and the simultaneously adsorbed non-caffeine coffee solids are only of the same order of magnitude as caffeine. The regeneration of the resin can be achieved by means of hot polar solvents (in particular water±ethanol mixtures). Another process involves adsorbing the caffeine in the extract, using activated carbon fibres as the adsorbent (Sipos & Jones, 1994). It has been found that activated carbon fibres are very selective in respect of caffeine, and a complicated preloading with sugar is not required. Additionally, the carbon fibres can be loaded with caffeine up to about 40% by weight. The process can be used with both aqueous green coffee extracts and aqueous roasted coffee extracts, and also with aqueous tea extracts and aqueous caffeine-containing extracts from other vegetable sources. Jacobs Suchard (Bunselmeyer & Culmsee, 1991) demonstrated in pilot plant tests, that it is possible to replace activated carbon by a molecular sieve (in particular a Y-zeolite) to adsorb selectively the caffeine out of a liquid green coffee extract. The pelletized adsorbent (first mentioned by Izod, 1979, as `UHP-Y') is not preloaded and is arranged in fixed columns. After the adsorption step, the zeolite is washed with water without desorbing the caffeine, the coffee solids containing water can be used for presoaking the green
112
Coffee: Recent Developments
Decaffeinated green coffee
Green coffee
Water
Carousel extractor
Carousel adsorber
Sugar preloading
Reactivated active carbon Bell dryer
Caffeine loaded active carbon
Reactivation
Fig. 5.4
Secoffex equilibrium process (Heilmann, 1991).
beans. In the next step the caffeine is desorbed from the zeolite with a water/alcohol solution and finally cleaned with steam and water. The alcohol is separated and recycled, in a distillation column, and the aqueous caffeine solution is sold for refining. A useful arrangement is to have a plurality of zeolite-filled columns in a ring circuit where the columns undergo each step successively. Another method to separate caffeine from an aqueous green or roasted coffee extract has been developed by General Foods (Zeller et al., 1983). Caffeic acid, a caffeine-complexing compound that is native to coffee, particularly roasted coffee, is combined with the coffee extract to form an insoluble, colloidal caffeic acid/ caffeine complex. Crystals of the caffeic acid/caffeine complex are then grown and subsequently separated from the decaffeinated coffee extract ± in particular by centrifugation. Recovery can be achieved by dissolving the complex crystals in lower boiling alcohols or ethyl acetate, thus breaking the complex and permitting the separate recovery of caffeic acid and caffeine. Caffeine adsorption on amorphous silica has been patented by Grace & Co. (Welsh, 1986). The caffeine adsorbents have the versatility to be used with a variety of commercial caffeine extraction systems, whether
aqueous or non-aqueous. Dehydroxylated, hydrophobic silicas are preferred for use in aqueous systems, while hydroxylated hydrophilic silicas are preferred for use in non-aqueous systems. The adsorption step takes place ideally in a continuous flow packed bed process; a multiple column configuration is best. Following the adsorption step, the caffeine-depleted solution is removed and the adsorbed caffeine is separated from the silica by elution with water at higher temperatures. The regenerated, dried silica may be recycled for further use, and the caffeine may be purified and crystallized by conventional methods. The Dow Chemical Company developed a process in which the caffeine-containing aqueous green coffee extract comes into contact with an effective amount of an adsorbent resin (Dawson-Ekeland & Stringfield, 1991). The adsorbent resin is made from a gel copolymer of a monovinyl aromatic monomer and a crosslinking monomer, where the gel copolymer has been post-cross-linked in the swollen state in the presence of a Friedel±Crafts catalyst. This adsorbent resin removes the caffeine without removing a substantial amount of the chlorogenic acid, so that the resulting coffee tastes similar to regular coffee. The resin can be used in the form of beads, pellets or any other form and the
Technology II: Decaffeination of Coffee
5.4 SUPERCRITICAL CO2 DECAFFEINATION As mentioned before, there has been a growing disquiet among consumers about the use of synthetic, mainly chlorinated chemical solvents in the food industry. This has led to the development of alternative processes using innocuous solvents of natural origin (Lack & Seidlitz, 1993). One of those new technologies was the application of supercritical CO2 for the extraction of caffeine from green coffee beans. Even if discussions about health risks in connection with the use of DCM have been practically closed, the CO2 decaffeination process is still used for about 20% of the worldwide capacity. The supercritical fluid extraction concept was first recognized in 1879. It had been discovered that solid compounds could be dissolved in supercritical fluids (Hannay & Hogarth, 1879). Almost all food processing applications of supercritical fluid technology employ CO2 as the solvent. Dense CO2 is not only a powerful solvent for a wide range of compounds of interest in food processing, but it is relatively inert, inexpensive, non-toxic, recyclable, non-flammable, readily available in high purity and leaves no residues. In addition, it should be emphasized that the use of CO2 does not produce any additional effect on the earth's ozone layer or aggravate the `greenhouse effect', because commercial CO2 is obtained as a byproduct from fermentation processes rather than by the combustion of fossil fuels (Moyler, 1993). With a critical point at 31.18C and 7.58 MPa (75.8 bars) (Fig. 5.5), near critical and supercritical CO2 can be used at temperatures and pressures which are relatively safe, convenient and particularly appropriate for the extraction of heat-labile compounds (Palmer & Ting, 1995). The solubility of a compound in a supercritical fluid is dependent on the density of the solvent, as well as on the physicochemical affinity of the solute for the solvent. Dissolved compounds can be recovered either by simply decreasing the pressure or increasing the temperature
Pressure (bar) Melting line 1200
1100
1000
900 800
350
700 300 250
Solid
adsorption itself is preferably executed continuously in a packed column. The desorption of the caffeine from the resin in the column after saturation can be accomplished using a straight solvent or a water/ organic solvent mixture at elevated temperatures. Prior to contact with the regeneration effluent the column has to be washed with an effective amount of water; after regeneration the packed column is backwashed for reuse, and to remove particulate contaminants.
113
Supercritical fluid
600
500
200
400 300
150
200
100 CP
P = 73.8 bar
100
50
Liquid TP
0
ling Gas B o i li n e
–60 –40 –20 0 10 20 30 40 50 60 70 80 90 –50 –30 –10
Temperature ( C)
Sublimation line
T 31.06 C
Fig. 5.5 Pressure/temperature diagram of CO2 with density as the third dimension (g/l) (Martin, 1982).
in order to decrease density, or by adsorption of the compound on an appropriate adsorbent. As a disadvantage of this technology, however, one has to recognize the high initial investment and maintenance costs due to the high pressure operation. The costs are the most important issues in the commercialization. The application of using compressed CO2 for decaffeination was first described by `Studiengesellschaft Kohle' in the mid-1960s (Zosel, 1965). The process conditions proposed are 70±908C and 160±220 bar. Three possible methods were suggested: (1) The moistened green beans are mixed with a stream of CO2 in a pressure vessel. The caffeine diffuses from the beans into the CO2 which passes into a washing tower, where the caffeine is absorbed in water. After 10 hours recycling, almost all caffeine is dissolved in the washing water, from which it can be separated by distillation. (2) Using the extraction conditions outlined in the
114
first proposal, the caffeine is removed from the CO2 fluid by passing this through a bed of activated carbon in which the caffeine is adsorbed. (3) A mixture of green coffee beans and activated carbon pellets is introduced into a pressure vessel, in which CO2 is brought to 908C and 220 bar. In 5 hours, the caffeine diffuses from the bean through the CO2 directly on to the activated carbon. After extraction, the coffee beans are separated from the activated carbon pellets by a vibrating sieve. On the basis of these three methods, there have been quite a number of inventions, claiming process improvements in terms of economy or quality. Some of these developments will be described in the following. Based on Zosel's second process, the CO2 decaffeination of green moistened coffee beans has been used commercially by the HAG Company since 1979 in Germany, wherein quantitative extraction of caffeine is achieved with moist supercritical CO2 (Vitzthum & Hubert, 1975). One modified example of Zosel's first process is the `Schoeller±Bleckmann Plant' (Lack et al., 1989). Moistening to the optimum water content of between 40% and 45% is effected by contact with steam, water or a combination of the two. The plant is designed with four high pressure extraction columns of which three are in operation in series, while the remaining one is off-line while the decaffeinated beans are discharged and it is recharged with fresh beans. After repressurising, this extractor will be reconnected to the CO2 cycle and another one will be disconnected. The CO2 containing dissolved caffeine flows into an adsorption column where it is decaffeinated by a countercurrent flow of fresh water. If a decaffeination rate of more than 97% is necessary (as in the US), then, after the washing column the gas stream has to pass an active carbon adsorption column. The application of a cascade of extractors in series guarantees a reasonably constant concentration of caffeine in the wash water. Schoeller±Bleckmann (Seidlitz & Lack, 1989) also described a process in which green unmoistened coffee beans are dampened in the high pressure reactor with supersaturated (1±2% by weight H2 O) supercritical CO2 . The decaffeination takes place layer by layer throughout the reactor at about 35% bean moisture, which is said to be optimal for the decaffeination process. The water ± originating from the caffeine scrubber ± is taken for supersaturation and contains coffee oil and coffee flavour. It is also optionally used for a second extraction step of the beans with water, countercurrent to the CO2 flow.
Coffee: Recent Developments
The Buse Company published a patent (Coenen & Ben-Nasr, 1987) on an `intermittent pressure system' in which the moistened beans are held under CO2 atmosphere in one of three or more extractors for a few minutes up to a few hours at 250 bar and 608C. The pressure is then quickly released which leads to an expulsion of the water±caffeine solution out of the cells on to the bean surface. After repressurising the reactor to 250 bar/608C, CO2 is circulated through the coffee bed and the decaffeination is completed. The caffeine is absorbed in water in a washing tower and separated by water evaporation. The wet beans are finally treated in a centrifuge, reducing the residual caffeine solution and predrying the beans. This system has been improved by the same inventors (Ben-Nasr & Coenen, 1990). Instead of supercritical CO2 , they use a CO2 -saturated, caffeinefree, green coffee extract for the decaffeination at pressures up to 300 bar and temperatures up to 1108C for from 1 up to several hours. After rapid release of the pressure, the beans are washed with the liquid for up to 2 hours. The CO2 -saturated, caffeine-containing, green coffee extract is then decaffeinated in a washing tower with supercritical CO2 . From the CO2 , caffeine is separated in a water absorption tower. Experiments have also been published which used supercritical nitrous oxide, instead of CO2 , as a solvent for decaffeination (Brunner, 1987). It has higher solvent power than CO2 , due to a higher density for a given temperature and pressure conditions, than CO2 , and a relatively low critical temperature (36.58C) (Fig. 5.6). Nitrous oxide can decompose uncontrolled, if handled inappropriately: yet in the temperature range suitable for decaffeination, nitrous oxide can be safely handled, if ignition sources are avoided. So far, no large scale technical application is known. Quantitative information on the adsorption capacity from activated charcoal, which is dependent on the specific surface area, has been reported (Gabel, et al., 1987). With the calculation of a `caffeine specific surface area' and experimental analysis of the heat of adsorption (which decreases with increasing mass of caffeine, that is with increasing surface coverage), the adsorption capacity can be predicted. For application in coffee producing countries, the Colombian Coffee Federation (Quijano-Rico, 1987) proposed the decaffeination of fresh green coffee beans as a tailored modern technology. After harvest, wet preparation and deparchment, the green beans can be decaffeinated. Avoidance of the traditional moistening and drying steps leads to reduced operation costs and improved cup quality. The decaffeination itself can be
Technology II: Decaffeination of Coffee
800
115
80 C N2O 100 C 80 C
600 Density (kg/m3)
CO2 100 C 400
200
0
100
200
300
400
Pressure (bar)
Fig. 5.6 Density pressure diagram of CO2 and N2O (Brunner, 1987).
realized with supercritical CO2 , using a temperature programmed extraction (a continuous increase from 608C to 858C, thus reducing the `thermal stress') (Toro, 1985). The first semi-continously operating process was developed by General Foods (Katz et al., 1990). The essentially caffeine-free supercritical CO2 is continuously fed to one end of a vertical cylindrical extraction vessel (Fig. 5.7) containing green coffee, and the caffeine-laden CO2 is continuously withdrawn from the opposite end. The moisturized coffee beans are charged under pressure (300 bar) to the extraction vessel via a big ball valve and lock hopper arrangement (Katz, 1989). The decaffeinated beans at the bottom of the vessel are discharged to the bottom lock hopper by opening the ball valves one after another. The valves are then closed and the extraction continues. The caffeine is washed from the CO2 stream in a countercurrent water wash column, taking advantage here of a favorable distribution coefficient for caffeine to partition into the aqueous phase. The rich caffeine-laden water from the absorber is concentrated by reverse osmosis to obtain caffeine of 97% or greater purity and Green coffee beans in Lock hopper
Ball valves
Water column (to remove caffeine)
Coffee extractor
Caffeine rich CO2
3
Caffeine lean CO2
Green coffee beans
CO2 and caffeine out
Caffeine free water
Extraction vessel
CO2 in Decaffeinated green coffee (sent to roasting)
(a)
Fig. 5.7
7 Caffeine rich water (to membrane system for caffeine concentration and water recycle)
Ball valves
(b)
Lock hopper Decaffeinated coffee out
General Foods' process of continuous decaffeination with supercritical CO2 (Krukonis et al., 1993).
116
Coffee: Recent Developments
a permeate containing acidic non-caffeine solids which are added to the process to improve yield and increase the rate of extraction. An improved CO2 decaffeination process design is based upon pilot plant data for commercial-scale coffee decaffeination using the Liquid Carbonic Corporation's proprietary supercritical CO2 extraction process (Linnig et al., 1991). It is a four-column high pressure extraction vessel design, of which three columns are continuously passed through countercurrently by CO2 at 14±35 MPa and 70±1308C for 6±12 hours, and the fourth column is discharged then refilled (Fig. 5.8). The rich CO2 fluid, laden with caffeine, is then washed with water at reduced temperature (15±508C) and pressure (5±10 MPA). The aqueous stream containing crude caffeine is sold for further processing (caffeine recovery and refining). Using primarily equipmentbased estimating techniques, they developed a detailed cost estimate for three capacities (base case: 10 000 t/a, Case: 5000 t/a, Case III: 20 000 t/a). A summary of the cost evaluation is shown in Table 5.1. A special apparatus for extracting certain substances from natural products (such as caffeine from tea or coffee) by means of supercritical CO2 has been proposed by Uhde GmbH (Theissing et al., 1991). The design includes a cylindrical high pressure vessel with annular cylindrical perforated baskets, which contain the natural product and the adsorbent wherein the fluid flows from the outer cylindrical area to the inner one (Fig. 5.9). The pressure drop and the risk of clogging
are reduced, the radial increasing velocity improves the mass transfer in the adsorber and the equipment is extremely compact. A method for producing decaffeinated green coffee beans after pretreatment with an aqueous acid solution has been described by the Liquid Carbonic Corporations (Kazlas et al., 1991). The acidification, preferably with 1.5±2% citric acid, compensates for the slight loss of acidity which occurs during supercritical decaffeination, thus improving the flavour and aroma of the coffee beverage. The acidifying treatment can be executed in combination with the moistening step before decaffeination. The decaffeination of coffee beans has been taken as an example for the application of supercritical fluid extraction with CO2 (McCoy, 1993). Decaffeination was measured as a function of CO2 flow rate, temperature and pressure. The rate of decaffeination increased with both temperature and pressure. The mathematical model describes the external and intraparticle diffusion resistances and the distribution of caffeine between water and CO2 . The partition coefficient for caffeine distributed between water and supercritical CO2 depends on temperature and pressure; soaking the raw beans in water prior to decaffeination enhances the rate of decaffeination. INTUS, a Berlin located institute (Roethe et al., 1994), published a new decaffeination system based on supercritical CO2 under conventional conditions, but
Process flow diagram Separator
Green beans
Extraction vessels Heat exchanger
Feeding system
Heat exchanger CO2 tank
Receiver Wetting system
Caffeine solution Injection pump Heater
Surge tank Bean dryer
Heat exchanger
CO2 take-up Heat exchanger
Main solvent pump
Decaffeinated beans
Fig. 5.8
Liquid Carbonic's supercritical CO2 extraction process (Linnig et al., 1991).
Technology II: Decaffeination of Coffee
117
Carbon dioxide decaffeination process: overall project economics (in 103 $/year).
Table 5.1
Base Case Utilities Labour and supervision Maintenance Taxes and insurance Plant overhead Annual capital recovery factor Total
2630 425 880 330 85 4400 8750
Total, cent/lb Total cost includes a reasonable return on investment.
37.7
Case II
Case III
1410 425 650 240 85 3240 6050
4990 425 1380 520 85 6900 14300
52.1
30.8
Source: Linnig et al., 1991.
with the addition of a liquid water stream on top of the moistened green coffee layer in the high pressure vessel. The amount of water used is 1±4 kg per kg green coffee (dry matter); if applied as continuous flow in the water saturated CO2 atmosphere, it does improve the mass transfer of caffeine from the bean surface to the fluid and reduces the decaffeination time by about 50%. If applied in pulses of about 9 minutes, with one
4 11 CO2 10
CO2
8 14 1 2
6
12
7
3
Fig. 5.9 Uhde's cylindrical high pressure CO2 decaffeination vessel (Theissing et al., 1991): 1 High pressure vessel; 2 cylindrical shell; 3 bottom; 4 cover; 6 basket; 8 vessel wall porous lining; 10 CO2 inlet nozzle; 11 CO2 discharge nozzle; 12 cylindrical inner space; 14 annular outer space.
to four pulses per hour, then even a high percentage of chlorogenic acid (sometimes producing gastric irritant substances) is removed and can be absorbed in the CO2 washing column in water. HAG proposed the replacement of activated carbon by ion-exchangers in order to adsorb the caffeine out of the supercritical CO2 -stream (Hubert & Vitzthum 1981). It was found that strong acidic cation exchangers are considerably more selective than activated carbon, that they can be regenerated with aqueous salt solutions or mineral acids and that they are pressure resistant. To recover the caffeine, the aqueous regenerant solutions are either concentrated until the caffeine crystallizes out or they are subjected to a liquid±liquid extraction, for example with methylene chloride. In Japan, laboratory tests have been published in which activated carbon was replaced by a zeolite membrane to separate the caffeine from the supercritical fluid (Tokunaga et al. 1997). The zeolite was coated as a thin layer on the surface of a tubular alumina membrane by hydrothermal synthesis. The system has turned out to have high thermal and pressure resistance and a good caffeine separation performance. The caffeine recovery was not reported. An application of the supercritical CO2 decaffeination process to roasted coffee (or tea leaves) has been patented by HAG (Gehring, 1984). In the first step, the aroma components are extracted by dry supercritical CO2 and in the second step, the wetted roast coffee is decaffeinated under the usual conditions with CO2 . Thereafter, the water soluble components are extracted from the decaffeinated roast coffee. The coffee extract is mixed with the aroma-containing supercritical dry CO2 , out of which the aroma components condense into the liquid after decompression; the extract can then be freeze dried.
118
Coffee: Recent Developments
5.5 LIQUID CO2 DECAFFEINATION So far, all decaffeination processes with CO2 referred to operate under supercritcal conditions (above 318C, 73.8 bar). Despite the fact that the extraction rate in supercritical CO2 is much higher due to the higher diffusivity in this state (the substances to be extracted are more soluble), liquid CO2 is preferred to supercritical, if lower temperatures and pressures than those critical for CO2 are sufficient. In 1989, the Hermsen Company in Germany obtained a patent describing a decaffeination process with liquid CO2 (Hermsen & Sirtl, 1989). The moisture content of the green beans has to be 45% to 55% and the CO2 has to be saturated with water, then decaffeination can be realised at very low temperatures (20±258C) and at pressures between 65 and 70 bar (Fig. 5.10). The separation of caffeine takes place by decompression below 60 bar in a separator; the caffeine±water mixture is reported to be very clean. The extraction time is about 60 hours, but due to the extremely low temperature the quality of the decaffeinated beans is said to be close to that of nontreated coffee. Practically all sugar inversions, Maillard and thermal decomposition reactions are avoided, particularly if careful low temperature moisturization and drying conditions complete the scheme. The decaffeination of roasted ground coffee with liquid CO2 has been published in a patent from HAG (Gehring et al., 1989). Moistened roast coffee is put in a pressure vessel which is connected in a cycle with another pressure vessel filled with a strong acid ion
2
1
3
7 9
4 6 8
5
Fig. 5.10 Decaffeination process with liquid CO2: 1 Extraction vessel; 2 separator; 3 condenser; 4 collection vessel; 5 pump; 6 heat exchanger; 7 CO2 storage; 8 saturation tank; 9 expansion valve.
exchanger. To minimize the extraction of coffee aroma components, the temperature of the circulating CO2 is kept between 15 and 308C and the pressure between 50 and 80 bar. After 2 to 3 hours the decaffeination is completed; the caffeine is almost selectively adsorbed in the ion exchanger. Cleaning of the ion exchanger and recovery of the caffeine can be achieved by a desorption step with supercritical CO2 . Remaining roast coffee components in the CO2 stream can be separated at the end by CO2 evaporation and aroma adsorption on the roasted coffee. The perceived quality of the coffee decaffeinated in such a manner is said to be equivalent to non-treated coffee in terms of flavour and aroma.
5.6 DECAFFEINATION WITH FATTY MATERIAL Liquid, water-immiscible fatty materials can also be utilized as caffeine solvents. They are usually composed of esters of fatty acids (mainly glycerol ester), but safflower oil, soy bean oil, corn or peanut oil and coffee oil are all advantageously employed in coffee decaffeination because they are also edible (Pagliaro et al., 1976). Fatty materials can be applied in the extraction of caffeine-containing liquid green coffee extracts, as described in the water decaffeination method. The best result is achieved in a multistage countercurrent extraction with high oil/extract ratios at about 308C. Green coffee beans have to be moistened with up to 40±60% water before they are treated with the fatty material. To achieve maximum efficiency of caffeine removal, decaffeination of beans should be carried out at 908C to 1208C. The regeneration of the fatty material can be achieved by liquid±liquid extraction with water. Separation of fatty material from the green beans can be achieved by steaming. Less thermal degradation and better preservation of the green bean quality can be achieved by bringing the beans into contact with a substantially laminar flow of fatty material (Proudly & Symbolik, 1988). The rate of decaffeination is increased, thus decreasing the residence time for achieving a desired degree of decaffeination (with draw-off ratios preferably below 10 (fatty material) to 1 (coffee beans)). Recycling a portion of the recovered fatty material to the extraction vessel will improve the efficiency.
Technology II: Decaffeination of Coffee
119
5.7 SPECIAL DEVELOPMENTS
surface with water will rejuvenate it over an extended period. Another method for in-home decaffeination has been proposed in which the consumable coffee extract, resulting from any brewing process, is treated with an inorganic silicate or aluminate material, which has been conditioned to act as a high capacity adsorbent (Lehrer, 1996). The adsorbent (preferably bentonite) is placed in pockets between two layers of a conical filter paper of a filter brewing system. Substantial quantities of caffeine are said to be removed by this method.
5.7.1 In-home decaffeination A process for decaffeinating freshly brewed coffee at home has been invented by Crose and Waldman (1995). The decaffeination system is incorporated into a coffee maker and is based on the generation of high voltage electrostatic field, delivered by a dry cell battery. This field is employed to attract and draw the caffeine molecules from the liquid coffee ideally transverse to the liquid flow, and to retain them, for example on a polystyrene sulfonic acid resin. The liquid passes the slots between the electrodes in the shape of an inverted cone, one of them coated with the resin (Fig. 5.11). After a brewing cycle the unit has to be disassembled, cleaned and reassembled for the next cycle. It is anticipated that washing the ionic resin
4
32 10 42 38
36
12
46 30 14
Coffee with a desired and constant amount of caffeine has been proposed by the ECOFE Corporation (Nufert & Fowkes, 1997). Because the caffeine concentration in different types and brands of coffee is subject to variability and not disclosed to consumers, they cannot control their caffeine intake. In order to reduce the physiological side effects of `health-deleterious noncaffeine components' in the beverage for heavy drinkers, coffee with an enhanced caffeine content (recaffeinated coffee) can be prepared by mixing roast and ground coffee with ground caffeine crystals or by spraying a liquid caffeine solution on to ground coffee.
5.8 CAFFEINE RECOVERY FROM ACTIVATED CARBON
28
22
5.7.2 Coffee with adjusted caffeine content
40 44 34
Fig. 5.11 Method for home decaffeination of a liquid (Crose & Waldman, 1995): 4 Receiver; 10 separator; 12 assembly for generating an electrostatic field; 14 collector; 22 funnel shaped base; 28,30 spaced apart opposite members of disks; 32 passageway; 34 entry port; 36 exit port; 38,40 opposite electrodes; 42 substrate; 44 caffeine receiver; 46 power supply.
Caffeine is an essential raw material for the pharmaceutical, food and related industries. It can be obtained by decaffeination of coffee or tea or a synthetic form can be manufactured (Lack & Seidlitz 1993). After 1970, the production of natural caffeine decreased due to the increasing adoption of CO2 based decaffeination processes (and later water processes) of a type which did not produce reusable caffeine. In these processes, the caffeine was adsorbed onto activated carbon. The recovery of caffeine from this adsorbent was not economic, and it was destroyed during regeneration of the adsorbent at temperatures between 6008C and 8008C. For that reason, about 75% of the total volume of approximately 10 000 t/year is now produced synthetically. (Using symmetrical dimethyl-urea, a high yield of theophylline is produced. This can be transformed to caffeine by methylisation). If it is based on petrochemicals, the synthetic caffeine can be differentiated from the `natural' substance by the use of 14 C isotopes. The price at which natural caffeine sells depends on
120
the price of synthetic caffeine and also on the demand. It rose from ca 20 Deutsche Marks in 1989 to ca 24 Deutsche Marks/kg in 1995. This additional income increases the profitability of those decaffeination plants which produce saleable caffeine. As only slightly more than 50% of all decaffeination plants are equipped with caffeine recovery systems (mainly organic solvent and CO2 -water absorption processes), it is obvious that a recovery system of caffeine from activated carbon is of considerable importance. The corresponding process development is reflected in many patent applications: . General Foods (Katz & Proscia, 1981) described a process for recovering caffeine from activated carbon by treating the carbon with an organic acid or an alcohol. Acetic acid and its azeotropes are especially preferred. However, with glacial acetic acid, operational and safety problems arose. Recovery yield is said to be below 73%. . Another process permits the separate recovery of caffeine and non-caffeine coffee solids adsorbed on activated carbon (Katz & Proscia, 1985). In the first stage the carbon is treated with an aqueous base solution (e.g. potassium carbonate) in order to remove the non-caffeine solids at about 608C. In the second stage, the carbon is treated with an aqueous acidic solution (e.g. acetic acid) at 858C in order to remove very clean caffeine. The carbon is then flushed with water and steam before being reused. Non-caffeine coffee solids are separated from the basic solution and recycled; caffeine is separated from the acidic solution and sold. . The application of an aqueous solution of either ethylene or propylene carbonate for at least 50% caffeine recovery from activated carbon, has been proposed by General Foods (Karmiol et al., 1984). The optimal carbonate concentration is a compromise between selectivity (better with higher concentrations) and dissolving capacity (decreasing with increasing concentration). A concentration of about 20% at temperatures close to 1008C, works well. The caffeine may be recovered from the solution by precipitation. The addition of a concentrated salt solution (e.g. potassium carbonate) alters the polarity of the caffeine-containing solution, causing the caffeine to precipitate out. The caffeine can then be separated by filtration or centrifugation. . General Foods disclosed that caffeine can be desorbed by means of aqueous acetic acid solutions of about 70% at above 1008C with about 70 to 80% yield. The caffeine can easily be separated by eva-
Coffee: Recent Developments
.
.
.
.
.
poration, the carbon has to be washed and steamed to remove residual acetic acid (Katz & Proscia, 1985). HAG proposed the application of formic acid which permits shorter treatment times, lower treatment temperatures and the use of a lower concentration of acid compared to acetic acid (Vitzthum et al., 1983). Separation of the caffeine takes place by distillation of the acid. A yield of 78% has been achieved. HAG developed a process for recovering caffeine with simultaneous regeneration of the carbon, in which the caffeine-loaded carbon is treated with hot water at a pressure of at least 86 bar, preferably 200 bar, and at a temperature of at least 3008C for 1 to 3 hours (Gehring et al., 1985). More rigid conditions in time, temperature and pressure run the risk of excessive decomposition of the caffeine. The activated carbon can be reused immediately without requiring additional treatment. The adsorption capacity decreases slightly from one recycle to the next; after many cycles the carbon has to be reactivated at high temperatures. The caffeine (yield about 65%) can be recovered from the aqueous solution by evaporation or membrane separation. Another invention relates to a process for recovering caffeine from caffeine-loaded activated carbon with acids such as benzoic, acetic, dichloroacetic or lactic, as such or in admixture (Kaper et al., 1987). The concentration of the acids used is at least 50%, the amount of solvent per amount of carbon is considerably lower than the amount required according to the state of the art. The caffeine can be removed from the solution, for example by crystallization. The best recovery yield (89.1%) has been achieved with benzoic acid at temperatures of approximately 1508C. Another process to recover caffeine has been developed by Douwe Egberts (Kaper, 1987). The caffeine-laden carbon is treated with a mixture containing at least 65% by weight of acetic acid and at least 2% citric acid. This mixture leads to lower inflammability, and particularly good extraction efficiency. The extraction process should be carried out at a temperature of about 1508C. A 5 to 10-fold excess of extracting agent relative to the carbon is sufficient to achieve a recovery of more than 90%. The caffeine can be recovered from the resulting solution by crystallisation. Sara Lee/Douwe Egberts found that activated carbon from a water decaffeination process can be regenerated using simple chemical compounds
Technology II: Decaffeination of Coffee
121
which possess recovery properties which are superior in terms of speed and energy requirements to those described before (Noomen & Putten 1993). Methylethylketone, ethyl actetate, dichloromethane, or a mixture of the first two, desorb up to 63% of the caffeine load under relatively mild process conditions. Dichlormethane proved to be the best in terms of yield and economy, without substantially removing other soluble coffee solids from the carbon. The carbon has to be steam treated after regeneration to remove any remaining solvents. The caffeine can be separated by evaporation of the recovery agent. After an evaluation of all the recovery systems, the conclusion is that obviously a number of disadvantages limit their broad commercial application: . The claim of a `chemical free' decaffeination process is hardly defensible, when chemical solvents are used for caffeine recovery. . A two-step process is necessary, involving use of a solvent, to produce the crude caffeine. . Solvent is lost when the activated carbon is cleaned and the caffeine is separated from the solvent.
Spent carbon + caffeine
. Applied solvents are volatile and often flammable. . The carbon has to be cleaned and regenerated before further application. The recently developed `direct caffeine desorption process' (Heilmann, 1997) avoids use of all chemicals. It is based on two patents from HAG (Wilkens, 1986; Sipos & Jones, 1994). Here the spent, caffeine-laden carbon is treated in a three-stage fluidised bed reactor (Fig. 5.12). In the upper layer, the desorption of caffeine takes place at about 3608C, directly into a sweeping gas flow. This clean sweeping gas, an inert gas from combustion, almost free from oxygen, enters the lowest stage at temperatures of approximately 8008C and reactivates the carbon. In the middle stage, non-desorbed caffeine and organic impurities are decomposed. After reactivation, the carbon is water quenched and recycled to the decaffeination plant. The caffeine-laden sweeping gas is treated in a scrubber/ quench system, the caffeine is absorbed in water, and subsequently crystallized. The resultant crystalline crude caffeine for sale (approximately 60% of the amount entering the system) contains less than 5% impurities and about 35% water.
Caffeine recovery Exhaust gas + caffeine
Quench 70 C Scrubber 60 C
To flue gas combustion /heat recovery
Filter cake dewatering/ drying
Desorption/ reactivation Crystalliser brine 4–6 C
Oven
Activated carbon quench/ reactivate silo
Fig. 5.12 Direct recovery of caffeine from activated carbon (Heilmann, 1997).
122
Table 5.2
Coffee: Recent Developments
Worldwide decaffeination capacities1. MC
EA
SCO2
Water
LCO2
Oil
Total
Europe North America Central/South America Asia
132 9 18 9
45 70 18 Ð
67 40 Ð Ð
98 18 5 Ð
20 Ð Ð Ð
Ð Ð 9 Ð
362 137 50 9
Total
168
133
107
121
20
9
558
1 In 1000 t, based on `usual split' between arabica and robusta. MC = methylene dichloride, EA = ethyl acetate, SCO2 = supercritical carbon dioxide, LCO2 = liquid carbon dioxide, Oil = fatty material.
5.9 ECONOMIC ASPECTS Due to an increase in consumption of decaffeinated coffee in the 1970s and 1980s there have been installed decaffeination plants of very large capacities, mainly from roast coffee companies. The installation followed the main areas in which decaffeinated coffee is consumed, namely central Europe and the United States. Approximately 50% of the worldwide capacities are located in Germany and France (Table 5.2). With reference to the percentage distribution of the decaffeination systems, the majority still apply solvents such as dichloromethane and ethyl acetate; only 21.7% switched to the `Swiss or French water process' (Table 5.3). This is also a consequence of profitability. The total non-material decaffeination costs are in the following order of magnitude: Solvents: Supercritical CO2 : Water:
30±40 US cents/kg 55±65 US cents/kg 80±100 US cents/kg
Whereas the consumption of decaffeinated coffee in the Western industrial countries is declining, this is compensated for by an increasing demand in Eastern Europe (Table 5.4). In total, the worldwide decaffeinated coffee business represented, in 1998, a volume of 382 000 t with a sales value of US $5 billion Table 5.5).
Table 5.3 methods.
Percentage distribution of decaffeination
Methylene dichloride Ethyl acetate Water Supercritical carbon dioxide Liquid carbon dioxide Fatty materials
30.1% 23.8% 21.7% 19.2% 3.6% 1.6%
Working on the basis of the split in the world coffee production of 34% robusta to 66% arabica coffee, and taking average caffeine values as 1% for robusta and 2% for arabica, then a theoretical annual caffeine production worldwide could be 6340 t, if all plants were to achieve a caffeine recovery of 100%. Based on a realistic yield of 80% on average, then the total amount of naturally sourced caffeine could be slightly more than 5000 t/year. As only 2500 t/year of natural caffeine is available in the market, it may be supposed that only 50% of all capacities are equipped with a caffeine recovery system. The actual worldwide caffeine demand is about 10 000 t/year, consequently the production of the synTable 5.4 Coffee sales (in 1000 tonnes) in Western Europe, Eastern Europe and the world, by product sector. 1994
1998
% Growth 1994/8
Western Europe Decaffeinated Standard
97.4 1174.6
93.9 1151.9
73.6 71.9
Total
1272.0
1245.8
72.1
Eastern Europe Decaffeinated Standard
3.3 199.2
9.6 254.5
186.9 27.8
Total
202.5
264.1
30.4
World Decaffeinated Standard
303.6 3293.3
381.6 3469.7
25.7 5.4
Total
3596.9
3851.3
7.1
Source: Euromonitor. Notes: Regional and world totals may not sum due to rounding growth rates calculated from unrounded data.
Technology II: Decaffeination of Coffee
123
Table 5.5 The value of coffee sales (in US $ million) in Western Europe, Eastern Europe and the world, by product sector. 1994
1998
% Growth 1994/8
Western Europe Decaffeinated Standard
1181.2 11463.1
1124.5 11316.0
74.8 71.3
Total
12644.3
12440.6
71.6
Eastern Europe Decaffeinated Standard
53.0 2287.7
199.6 3800.8
276.2 66.1
Total
2340.7
4000.4
342.3
World Decaffeinated Standard
3544.2 32345.5
4926.0 34825.3
39.0 7.7
Total
35889.6
39751.3
10.8
Source: Euromonitor. Notes: Regional and world totals may not sum due to rounding growth rates calculated from unrounded data.
thetic material is 7500 t/year. This production is dominated by a small number of US and German manufacturers; not less than 4000 t are produced by Boehringer Ingelheim in Germany and Mexico. The predominant application of caffeine is in soft drinks; about 25% is needed for the production of medicines (Fig. 5.13).
Medicines 25%
Industrial 1%
Beverages 74%
Fig. 5.13 Caffeine application worldwide.
REFERENCES Ben-Nasr, H. & Coenen, H. (1990) EP 0439710 B1, Buse GmbH. Berry, N.E. & Walters, R.H. (1943) Process of decaffeinating coffee. US Patent 2 309 092. Blanc M. & Margolis, G. (1981) EP 0049357, NestleÂ. Brunner, G. (1987) Decaffeination of raw coffee by means of compressed nitrous oxide. In: Proceedings of the 12th ASIC Colloquium (Montreux) pp. 294±305. ASIC, Paris, France. Bunselmeyer, D., Culmsee, O. & Heilmann, W. (1991) EP 0523268, Jacobs Suchard. Clarke, R.J. (1988) Patenting of coffee inventions. In: Coffee, Vol. 6. Commercial and Technico-Legal Aspects, (eds R.J. Clarke & R. Macrae) pp. 145±176. Elsevier Applied Science, Barking, UK. Coenen, H. & Ben-Nasr H. (1987) EP 0482675 A2, Buse GmbH. Crose, J.R. & Waldman, A.A. (1995) US Patent 5 503 724, IMSCO. Dawson-Ekeland, K.R. & Stringfield, T.R. (1991) EP 0432960 A2, Dow Chemical Company. Fischer, A. & Kummer, P. (1979) EP 008 398, Coffex. Gabel, P.W., Sarge, S. & Camenga, H.K. (1987) Search for optimal adsorbents for decaffeination processes by calorimetric investigation. In: Proceedings of the 12th ASIC Colloquium (Montreux) pp. 306±312. ASIC, Paris, France. Gehring, M. (1984) EP 0151202, HAG. Gehring, M., Barthels, M. & Wienges, H.R. (1985) US Patent 4 506 072, HAG. Gehring, M., Vitzthum, O. & Wienges, H. (1989) Patent DE 3303679 C2, HAG. Hannay, J.B. & Hogarth, J. (1879). On the solubility of solids in gases. Proc. Roy. Soc. London, 29, 324. Hay, P., Leigh, D. & Liardon, R. (1995) EP 0776607 A1, NestleÂ. Heilmann, W. (1991) A modified Secoffex process for green bean decaffeination. In: Proceedings of the 14th ASIC Colloquium (San Francisco) pp. 349±356. ASIC, Paris, France. Heilmann, W. (1997) Caffeine recovery from activated carbon. In: Proceedings of the 17th ASIC Colloquium (Nairobi) pp. 254±60. ASIC, Paris, France. Hermsen, M. & Sirtl, W. (1989) EP 0316694, Hermsen GMbH. Hinman, D.C. & Saleeb, F.Z. (1984) EP 0140629, General, Foods. Hubert, P. & Vitzthum, O.G. (1981) US Patent 4 411 923, HAG. Izod, T.P.J. (1979) EP 0013451 A1, Union Carbide. Jones, G.V., Musto, J.A. & Meinhold, J.F. (1985) EP 0159829, General Foods. Kaleda, W.W., Saleeb, F.Z. & Zeller, B.L. (1986) US Patent 4 467 634. General Foods. Kaper, L. (1987) EP 0259905 A1, Douwe Egberts. Kaper, L., Klamer, R. & Noomen, J.P. (1987) EP 0251364 B1, Douwe Egberts. Karmiol, M.H., Hickernell, G.L. & Hall, B.J. (1984) US Patent 4 443 601, General Foods. Katz, S.N. (1984) US Patent 4 472 443, General Foods. Katz, S.N. (1987) Decaffeination of coffee. In: Coffee, Vol. 2, Technology, (eds R.J. Clarke & R. Macrae) pp. 59±72. Elsevier Applied Science, Barking.
124
Katz, S.N. (1989) US Patent 4 820 537, General Foods. Katz, S.N., Spence, J., O'Brian, M.J. et al., (1990) US Patent 4 911 941, General Foods. Katz, S.N. & Proscia, G.E. (1981) US Patent 4 298 736, General Foods. Katz, S.N. & Proscia, G.E. (1985) US Patent 4 513 136, General Foods. Katz, S.N., Proscia, G.E. & Clisura, G.L. (1985) US Patent 4 548 827, General Foods. Kazlas, B.J., Novak, R.A. & Raymond, R.J. (1991) Patent WO92/ 03061, EP-547119, Liquid Carbonic Corporation. Krukonis, V.J., Gallagher-Wetmore, P.M. & Coffey, M.P. (1993) Food processing with supercritical fluids: fact and fiction. In: Science for the Food Industry of the 21st Century, ATL Press, USA. Lack, E. & Seidlitz, H. (1993) Commercial scale decaffeination of coffee and tea using supercritical CO2 . In: Extraction of Nature Products using Near Critical Solvents, (eds M.B. King & T.R. Bott), pp. 101±39. Blackie, Glasgow. Lack, E., Seidlitz, H. & Toro, P. (1989) Decaffeination of coffee samples by CO2 extraction. In: Proceedings of the 13th ASIC Colloquium (Paipa) pp. 236±245. ASIC, Paris, France. Lehrer, R. (1996) Patent Application WO 97/07686. Leigh, D., Hay, P. & Liardon, R. (1995) EP 0776607, NestleÂ. Linnig, D.A., Leyers, W.E. & Novak, R.A. (1991) Decaffeination with supercritical carbon dioxide. In: Proceedings of the 14th ASIC Colloquium (San Francisco), pp. 357±64. ASIC, Paris, France. Martin, H. (1982) Selective extraction of caffeine from green coffee beans and the application of similar processes on other natural products. In: Proceedings of the 10th ASIC Colloquium (Salvador) pp. 21±28. ASIC, Paris, France. McCoy, B.J. (1993) Rate processes in supercritical fluid extraction. In: Proceedings of the 6th International Congress on Engineering and Food, Chiba, Japan. Vol. 2, Blackie, Glasgow. Mooiweer, G.D. (1982) EP 0078088, Douwe Egberts. Mooiweer, G.D. (1983) EP 011375, Douwe Egberts. Morrison, L.R. & Phillips, J.H. (1984) US Patent, 4 474 821, Proctor and Gamble. Moyler, D.A. (1993) Decaffeination of coffee. In: Extraction of Natural Products Using Near Critical-Point Solvents (eds M.B. King & T.R. Bott), pp. 140±78. Chapman and Hall, Glasgow.
Coffee: Recent Developments
Noomen, P.J. & Putten, C.V. (1993) EP 0612744, A1, Sara Lee Douwe Egberts, The Netherlands. Noronha, T. (2000) Coffee: world, Western Europe and Eastern Europe sales volume and value. Euromonitor, Feb. Nufert, T. & Fowkes, S. (1997) Patent Application WO 98/ 07330, Ecofe Corporation. Pagliaro, F.A., Franklin, J.G. & Gasser, R.J. (1976) US Patent 4 465 699, NestleÂ. Palmer, M.V. & Ting, S.S. (1995) Applications for supercritical fluid technology in food processing. Food Chem., 52, 345±52. Proudly, J.C. & Symbolik, W.S. (1988) US Patent 4 837 038, NestleÂ. Quijano-Rico, M. (1987) New ways of industrial coffee processing. In: Proceedings of the 12th ASIC Colloquium (Montreux) pp. 187±193. ASIC, Paris, France. Roethe, K.P., Roethe, A., Suckow, M., Mothes, S. & Stackfleth, M. (1994) Patent WO94/26125, Method for depleting green coffee of caffeine and chlorogenic acids. INTUS, Berlin. Seidlitz, H. & Lack, E. (1989) UK Patent GB 2235121 A, Schoeller±Bleckmann. Sipos, S. & Jones, G.V. (1986) US Patent 5 702 747, Kraft Foods. Sipos, S. & Jones, G.V. (1994) EP 0666033 A1, Kraft Foods. van der Stegen, G. (1985) EP 0158381, Douwe Egberts. Theissing, P., Saamer, P. & KoÈrner, J.-P. (1991) US Patent 5 153 015, Uhde GmbH. Tokunaga, Y., Fujii, T. & Nakamura, K. (1997) Separation of caffeine from supercritical carbon dioxide with a zeolite membrane. Biosci. Biotech. Biochem., 61, 1024±1026. Toro, P. (1985) DE 3445502 A1, Cafe Toro. Vitzthum, O. & Hubert, P. (1975) US Patent 3 879 569, Genera Foods. Vitzthum, O., Werkhoff, P. & Gehring, M. (1983) EP 0129609, HAG. Welsh, W.A. (1986) EP 0173297, Grace & Co. Wilkens, J. (1986) DE 3511129 A1, HAG. Zeller, B.L., Kaleda, W.W. & Saleeb, F.Z. (1983) US Patent 4 521 438, General Foods. Zeller, B.L. & Saleeb, F.Z. (1999) Decaffeination of non-aqueous solvents using caffeic acid. In: Proceedings of the 18th ASIC Colloquium (Helsinki), pp. 168±72. ASIC, Paris, France. Zosel, K. (1965) Studiengesellschaft Kohle. Chem. Abstr., 63, 110456.
Chapter 6
Technology III: Instant Coffee R.J. Clarke Consultant, Chichester, UK
6.1 INTRODUCTION 6.1.1 Instant coffee in the market place The production and sales of instant coffee have increased markedly in the last ten years or so, in most countries. Thus in the UK, the retail sales of instant coffee were estimated to be some £520 m in value in 1993 in the UK, according to official data. In the UK also, there has been a noticeable increase in the number of different brands (and sizes) available; thus in 1987 (Clarke & Macrae, 1988), were listed some 17 different brands (excluding private label brands) covering spraydried powdered, granules, freeze-dried and decaffeinated products, whilst in 1999, the same supermarket shelves show some new introductions (five new product brands) and a few withdrawals. In the 1987 listing, prices per 100 g net contents were given; thus, Nestle Alta Rica freeze dried at £2.59, which price appears not unexpectedly lower than that in 1999 (i.e. £3.29 noted, July 1999). Brands once marketed from General Foods Ltd, Banbury, are now from Kraft Foods (with Headquarters in Cheltenham, Glos, England), following change of ownership in 1984, with some new brand names, like `Kenco'. So-called off-shore soluble product importation, largely from Brazil, has continued; whilst liquid coffee extracts have appeared, thus a product sold from chill cabinets by Douwe Egberts; whilst in Japan, various canned liquid coffee drink products are on the market (see Chapter 7).
6.1.2 New technology A comprehensive account of the processes and equipment involved in instant coffee technology, up to 1987, was provided in the chapters of Coffee, Vol. II Technology (Clarke, 1987a); with some updating in sections under `coffee' in the Encyclopaedia of Food Technology (Clarke, 1993a). There was also a review in
Ullman’s Encyclopaedia of Technology (Viani, 1986). A review of quality control methods for green, roasted and instant coffee was published (Clarke, 1987b); and on coffee shelf life (Clarke, 1993b). It is to be expected that there will have been numerous developments in the technology of instant coffee manufacture, over the last ten years, as reflected in the many new patents issued, some 35 European Patent applications. A number of papers on different aspects of instant coffee processing have been published in the Proceedings of ASIC Colloquia 1987±99, as will be described under the different sections of processing below; with two overall descriptive papers (Sylla, 1989; Grùnlund, 1995).
6.1.3 The legacy of Professor H.A.C. Thijssen As evidenced in their numerous papers, Thijssen and his colleagues at the Eindhoven Institute of Technology, during the late 1960s and 1970s, into the 1980s, mainly discussed in Coffee, Vol. II (Clarke, 1987a), were powerful interpreters of and ingenious experimentalists in the relevant chemical engineering operations of evaporation and drying of coffee extracts (also of other food products). Sadly, Hans Thijssen died in 1986, after which there was held a well-attended Memorial Symposium in November 1987 in Eindhoven. The title of this Symposium was `Preconcentration and Drying of Foodstuffs', being appropriate to Thijssen's main areas of achievement, that is, in freeze and thermal concentration, and the `selective diffusion concept' in drying. The papers then delivered were subsequently published (Bruin, 1988). Another chemical engineer, who also helped pioneer the application of chemical engineering principles to food processing operations in general, was Professor Marcel Loncin, of Belgium, latterly of Karlsruhe University, who died in 1994, and to whom a memorial tribute was paid at the 7th International Congress on
125
126
Coffee: Recent Developments
Engineering and Food in April 1997 at Brighton, England, (the Proceedings have been published: Jowitt, 1977). Loncin himself had delivered a keynote lecture to ASIC in 1977 entitled `Food Engineering and Coffee'.
6.1.4 Legislation and standardization No major changes at least in European legislation affecting instant coffee have been made since those last reported (Clarke, 1988). However, a new Directive, 1999/4/EC, 22 February 1999 was introduced, with some modifications of 80/232/EEC mainly related to labelling and prescribed weights, to be brought into force before 13 September 2000. New analytical test methods for instant coffee from the International Standards Organization (ISO) are described in Appendix 1 to this book.
6.2 PROCESSING Instant coffee processing as described (Clarke, 1987a) consists of a number of successively performed operations, as follows: (1) (2) (3) (4)
grading, storage, blending of green coffees; roasting and grinding; extraction; drying, by either (a) spray (with either powdered or granule formation); or (b) freeze-drying (to granules only).
To which are optionally added, now most usually in sophisticated and quality manufacture (3a), procedures for pre-concentration prior to drying, which may be either freeze concentration, or thermal evaporation with separate volatile aroma compound handling; and (4a), procedures for `aromatising' the finished product, whether powdered or granules. (5) Packing into jars/tins. Finally, total factory processing will include spent grounds disposal. All these products may be derived from decaffeinated green coffees, where the latest developments in decaffeinating procedures are described in Chapter 5.
6.2.1 General The object of all processing operations for instant coffee manufacture is two-fold:
(1) The optimization of the type and amount of coffeesoluble solids in the finished product, mainly complex carbohydrate, but other non-volatile substances responsible for taste characteristics detected by the tongue, and for stimulant effects on the body. Some of these components will be almost entirely extracted in a first `fresh' extraction stage (atmospheric), such as caffeine/mineral substances, whilst others such as carbohydrates and melanoidins will be increasingly extracted in subsequent stages, in a continuum of composition, although splitting off of some monosaccharide substances may or will be allowed to occur. (2) The optimization of the type and amount of coffee volatile substances extracted and subsequently retained in the final product, responsible for the total aroma element of the final flavour obtained in the made-up product, or headspace odour over the still dry product, detected by the olfactory sensors in the back of the nose. Fresh roast coffee volatiles are mainly extracted in the first stage, dependent upon the ratio of water to coffee used and other factors; but may only be completed by use of subsequent autoclaving stages, where some additional coffee volatile compounds will be generated rather than just released. Whilst there is also a continuum of flavour obtained across the extraction stages, the autoclaving stages tend to produce disproportionate amounts of some undesirable volatile compounds, such as furfural, whereas `fresh brew' coffee flavour is to be preferred. However, recent compositional studies (see Chapter 3) indicate that fresh brew by household brewing methods may only extract up to 80% of those actually present in the roast (and ground) coffee as determined by exhaustive solvent extraction methods; furthermore, some compounds in brews will be extracted to a lesser or greater extent than others. Extraction methods in instant coffee manufacture should, however, obtain all the available volatile compounds present, together with some additional or increased amounts of compounds resulting from an extension of roasting, but under wet conditions, in the autoclaving stages. There will also be differences of volatile substances extracted according to the blend (arabica/robusta) of roast coffee used. Recent developments to be described, indicate the use of particular processes, which seek to separate out or include selected volatile compounds. Whatever is extracted, however, the final aroma characteristics, of the finished product, will depend more importantly upon the choice of drying method, and operating procedures.
Technology III: Instant Coffee
A recent paper (Steinhaeser et al. 1999) describes some model experiments in the extraction of roast and ground coffee during the production of instant coffee, with special reference to the volatile compounds present.
6.2.2 Roasting/grinding The methods and equipment for roasting coffee beans are the same as for producing roasted coffee (either whole beans, or subsequently ground) destined for retail sale in closed packages. Developments in this area are separately described in Chapter 4 of this book, entirely devoted to this subject. Similarly, methods and equipment for grinding roasted coffee beans, in which no particular innovations have been reported, apart from further design features to exclude air/oxygen during the grinding itself.
6.2.3 Extraction The use of battery column percolators, incorporating both the so-called `atmospheric' extraction (water temperatures around 1008C), and subsequent `autoclaved' extraction (water temperatures up to around 1708C, under hydraulic pressure) appear to continue to be the most popular system of manufacture. However, a number of modifying features in the basic operation have also continued to be recommended. Thus, the advantage of pre-wetting the roast and ground coffee in a separate vessel before charging a percolation column is emphasised by Niro A/S, the Danish manufacturers of instant coffee processing equipment. Air/carbon dioxide in the roast and ground coffee charges need also to be removed, to assist the aqueous extraction and minimise foaming ± when the application of a vacuum during column filling is suggested. It should be noted that these gases will contain useful amounts of coffee volatile compounds, which should also be retained by suitable trapping means, collected and added later in the process to dried product. Niro A/S now also recommend the physical separation of the two extraction stages; firstly, an extraction in the fresh coffee stage at temperatures below 1208C, and taking as fast as 10±15 minutes to give an `aroma-rich' extract; secondly, an autoclaving extraction on the partially extracted coffee with pressured water at 1808C, flowing at velocities up to 2.5 times higher than in conventional systems, resulting in a shorter extraction time. The two streams of extract are then handled separately (see freeze concentration and thermal concentration sections), and then
127
combined for drying. It is claimed that the total overall extraction time is reduced from the conventional 200 minutes to approximately 90 minutes. An especially designed type of extraction column is available (Grùnlund, 1995). The optimum size, shape and number of percolators in a battery, whether in relation to solubles or volatile compound extraction has been discussed (Clarke, 1987a). According to EP O489 401 (1990), (Koch, K. & Vitzthum, O.G, Kraft Jacobs Suchard), shorter columns, doubling of their number and use of a fine grind size of roast coffee is to be favoured, along with particular aroma handling arrangements, for improved roast coffee brew flavour. Use of high water±coffee ratios, especially in the fresh stage is, however, an important factor in this effect. The optimum extraction of soluble solids also features in two new patent applications. The use of higher temperatures than in a conventional countercurrent battery, such as up to 2208C, features in a three-stage extraction procedure in a Nestle patent, EP 0826 308 (1998). Three separate stages are described as taking place at: (1) 80 to 1608C; (2) 160 to 1908C with a `secondary' aqueous liquid, the grounds are then drained, and thermally hydrolysed at 160 to 2208C for 1 to 15 minutes; (3) 170 to 1958C, in which soluble solids are further extracted from the hydrolysed grounds using a `tertiary' extraction liquid (more water). A soluble coffee product is obtained, which has 30% carbohydrates, comprising 4% monosaccharides, > 10% oligosacclarides and 19% polysaccharides, which have a weighted average of > 2000 units (Dalton) with a polydispersity > 3. A similar multi-stage approach is evidenced in an earlier General Foods Corporation patent (USP 4 798 930A, 1989 = EP 0336 837 (1990) proposing a hydrolysis of partially extracted R and G coffee to give increased yield of soluble solids with minimal degradation of flavour. This hydrolysis is to be performed in a separate fixed bed reactor at a temperature of 380 to 4508F (193 to 2238C) in which the extraction time is short, some 7 to 45 minutes, using some six times the weight of water to weight of grounds. The use of very high extraction temperatures is not practical in a conventional battery, for various reasons (time and temperature control and insoluble product formation). The extra yield is derived from a controlled solubilization of the insoluble mannan, which soluble mannan oligosaccharides (amorphous) are, however, notoriously unstable, and readily form sediments. Such mannan deposits from commercial instant coffee brews held at temperatures of 608C and above, by crystal-
128
Coffee: Recent Developments
lization, are described in the literature (Bradbury & Atkins 1997). In view of the ever present economic significance of the percentage amount of soluble solids taken out of roasted (or equivalent green) coffee, some further data on yields obtainable in a batch two-stage process (with temperatures up to 908C and 1808C, respectively) have been given by Noyes (1995). A different approach in the autoclaving stage of extraction is that by the use of enzymes to solubilize some of the complex carbohydrates in roast coffee, thus the use of immobilized b-mannanase, Nestle EP 0676 145 AI (1995) although other enzymes have been previously described.
The different kinds of extraction method that can be used for soluble coffee manufacture, are classified in Fig. 6.1, whilst those for household/catering brewing are included for comparison.
6.2.4 Freeze concentration of extracts Freeze concentration is now a well-established means of concentrating extracts of roast coffee, whether they be used on the total extract from a percolation battery, or on only an aroma-rich extract from a fresh atmospheric stage. The process is particularly favoured for the latter, with its known high retention capability
(a) Industrial scale (soluble coffee)
Infusion/separation
Open tanks with agitators
Closed tanks/reactors with/without agitators
Percolation/separation
Column batteries Multi-stage intermittent
Separation by external centrifuges Multi-stage operation Separate temperature regimes
Continuous screw Countercurrent
Built-in filtering Temperature profiling
(b) Household catering scale (fresh brewed coffee)
Infusion/separation of loose R and G (rarely in bags) in a vessel with independent filtering
Decoction extraction with boiling water followed by decantation/ straining e.g. Swedish KOK coffee, grounds not separated e.g. Turkish coffee
Fig. 6.1
Steeping extraction with water below 100 C with/without agitation, followed by decantation/ straining/plunger filter
Percolation/separation of loose R and G (occasionally in bags) held in a fixed bed with built-in filter
Once through under gravity, e.g. drip pot, automatic filter
Once through under pressure
Pressure in a closed vessel water temperature above 100 C e.g. Moka express
Methods of extraction of R and G coffee with water or aqueous liquid.
Multi-pass under pressure, e.g. 'Percolator', 'Cona' (2 passes)
Pump pressure in a closed vessel water temperature below 100 C true espresso
Technology III: Instant Coffee
129
for volatile compounds as in the Niro recommended process already described. It has been known also that the conditions of freeze concentration can have a marked influence on the colour/aroma content of a coffee extract subsequently freeze-dried; for example, slow freezing gives rise to a dark-coloured, freeze-dried product. Colour is also determined by foaming conditions on a semi-frozen (`slush') extract. The various factors involved have, again, been described in recent papers (Dithmer, 1995, Jansen & Van Pelt, 1987). The value of fast freezing (without concentration) prior to freeze drying has been described in a Unilever patent, EP O256 567 A2 (1988).
6.2.5 Thermal concentration and volatile compound recovery Evaporation is also a well-established means of concentration, again whether on a total extract from a percolation battery, or from an individual stage. Such evaporation, as is now well known, will strip off volatile aromatic compounds according to the percentage of water evaporated off. With separate autoclaved extracts as already described, such loss may be deemed advantageous, when the volatile compounds may be largely heat generated compounds during the autoclaving process, of a caramelised or furfural nature. With aroma-rich extracts as from a fresh coffee stage, any thermal evaporation designed to increase soluble solids concentration must be conducted in two stages, a stripping stage in a single-effect evaporator unit (or steam/gas stripping unit) where the required volatile compounds are stripped off, condensed and held for subsequent re-incorporation. Typical evaporation is 10% of the water content. This subject is further discussed in Section 3, Physical properties of volatile compounds. A second stage is bulk evaporation of the stripped extract, for example to a remaining 50 to 55% (w/w) water content, conducted in multipleeffect evaporators for fuel economy. The main types of evaporating equipment used are the plate and falling film evaporators. Stripping can be accomplished alternatively; thus a comparatively recent innovation is the spinning cone column (SCC) developed by the Australian chemical engineer, Andrew Craig, and marketed by Flavourtech (Fig. 6.2 and 6.3). It has now found wide usage in the food industry, especially for fruit juices, milk products and instant coffee extracts (Casimir & Huntington, 1970; Casimir & Craig, 1990). It is based upon gas stripping, rather than evaporation. An SCC consists of a vertical stainless steel cylinder in which an inert stripping gas removes, under
Fig. 6.2 Mechanical layout of a spinning cone column (by courtesy of Flavourtech).
vacuum, a vapour stream of volatile compounds from a liquid/slurry. The largest model is approximately 1 m in diameter, and 5 m in height, capable of processing up to 10 000 litres/hour or more. The cylinder contains two series of inverted cones, fixed cones are attached to the inside wall, whilst a central rotating shaft carries another series of cones, parallel to the fixed cones in such a way that they alternate vertically: one fixed, one rotating. A liquid product, such as coffee extract, is fed to the top of the column, where it encounters the surfaces of the cones, until it reaches the bottom of the column; gas for stripping (steam under vacuum) enters the bottom of the column, countercurrently and picks up the volatile compounds and the vapours emerge from the top of the column to be condensed to a liquid volatile compound concentrate. There is a good liquid/
Fig. 6.3 Cross-section of a cone set within a spinning cone column (by courtesy of Flavourtech).
130
Coffee: Recent Developments
gas mass transfer of aroma compounds assisted by the turbulent flow, partly induced by fins on the underside of the rotating cones. In this system, the level of actual water evaporation from the extract will be small, and approximate to a packed or plate distillation column, where the mathematics of the operation can be studied (Bomben et al. 1973). Whilst the Alfa Level Centritherm has a rotating heated surface, it is an evaporator. The claimed advantages of this stripper include small liquid hold-ups and residence times, high separation efficiency and the ability to handle highly viscous liquids. The fluid mechanics of such a column in operation, especially under flooding conditions, has been recently studied (Moy et al. 1997). Volatile compounds can also be recovered, prior to any processing, from the roast and ground coffee by steaming methods (Clarke, 1987a). Vacuum steaming appears to have its attractions, so that EP O227 262A1 (1987) describes a method of partial re-incorporation of such volatile compounds before final spray drying.
6.2.6 Volatile compound handling There have been a number of further variants described in patents, to make selections of volatile compounds arising from the extraction stages, particularly where higher temperatures are employed. As with most patents for coffee identified, there may also be GB, German and USA issuance of the same application. Table 6.1 below gives a listing. USP 4 900 575 describes the processing of volatile compounds arising from an autoclaving stage, through a distillation column passing the upper off-take stream Table 6.1
through an absorbent column (non-polar microporous absorbent). Collection takes place to maximise the content of diacetyl/acetaldehyde, and to leave furfural.
6.2.7 Reverse osmosis This technique for concentration has shown little applicability for use in instant coffee processing except for rather dilute solution streams, primarily due to potential membrane fouling due to the presence of some high molecular weight substances. There is also the problem of retention of low molecular weight volatile compounds. A recent investigation in Japan (Imura & Danno, 1995) showed that weak extracts (5% w/w) could be concentrated to about 15% in half the time, if reverse osmosis was preceded by filtration through a ceramic micro filter (having 0.8 mm pore size).
6.2.8 Spray drying and agglomeration It is now widely accepted that in spray drying, in order to have a good retention of volatile aroma compounds, the coffee extract must have a soluble solids concentration of at least 50% w/w. This level of concentration can only be achieved by thermal methods of bulk evaporation, to which, however, a separately prepared aroma concentrate (obtained by the stripping and handling methods already described) can be added with only a small diminution of percentage soluble solids concentration. Similarly, an aroma-rich first stage extract after freeze concentration (say to 35% w/ w) can be added to a thermally concentrated `second' or `third' stage extract. The thoroughly mixed streams, in each case, can then be spray dried, with or without
Some patents for different methods of coffee volatile compound handling/recovery.
Country
Patent or application No
EP
0227 263A1
EP
Issue date
Applicant
Remarks
1987
GF Ltd
Flash stripping
0220 889 A2
1987
GF Ltd
Addition of fines
EP
0847 699 A1
1998
Ajinomoto GF
Removal of undesirables
EP
0240 754 A2
1987
Nestle´
Use of coffee oil
EP
0234 338 A1
1987
GF Ltd
Flash stripping
USP
4 900 575
1990
Kraft General Foods
Hydrolysis, volatiles selection
USP
5 225 223
1993
Jacobs Suchard
Volatiles handling
Technology III: Instant Coffee
simultaneous foaming of the extract on passage to the spray nozzle, for bulk density/colour control. As a means of concentration, freeze concentration on a complete coffee extract would only be used prior to freeze drying. There have been a number of further investigations into establishing conditions for maximum aroma retention in spray drying. In particular, the concept of and practice of `low temperature' drying was developed by joint work between a Brazilian instant coffee (actually Japanese owned) and a Japanese milk products company (Bassoli et al., 1993; Ohtani et al., 1995). In this process, a single spray nozzle is used, but where the spray droplets on formation are enveloped by warm air currents in addition to the conventional hot air supply from the top of the tower (although 130 to 1808C rather than 200 to 3008C stated usual). This warm air supply is claimed to give a more even downward flow of the total air, reducing swirling close to the nozzle and deposition/adherence of powder at the top sides of the tower. Data were presented to show the very favourable aroma retention (especially of low boiling point compounds) of the low temperature spray dried products (with different hot air temperatures), in fact close to, if not better than a corresponding freeze dried product. The extracts had generally been freezeconcentrated, to either 30% or 43% w/w soluble content; or thermally concentrated to 43%. Taste testing was also used to confirm the quality of the products, in particular the reduction of caramelised flavour notes in the lower temperature samples. Further work was carried out including carbon dioxide gas foaming for control of bulk density, and additionally found not to affect aroma retention percentage values. Studies have continued on the sources and reasons for the volatile losses which can occur during spraying. These are usefully reviewed by King (1988), based upon the experimental work of himself and colleagues, thus Frey (1984) and Papadekis (1987). The losses from the emergent sheet from a pressurised nozzle, and during actual spray droplet formation already known are important. King also emphasises the important aspect of the influence of hot air mixing in the nozzle zone, and discusses the overall flow dynamics of the air within a drier. There is a strong pumping effect on the overall pattern; so that with a sufficiently high nozzle pressure the volumetric flow of air drawn in by the spray exceeds the feed flow of air. Whilst the concentration of soluble solids in the coffee extract being sprayed is most important, a number of other improvements in operation can be made to enable selective diffusion within the droplets as soon as
131
possible, and so secure maximum retention of volatile compounds. There will still be differentials in percentage retained for different volatile compounds, determined primarily by their different diffusion coefficients; although for given spray drying conditions, an estimate can be made for different compounds, when that for a marker compound is known (Clarke, 1990). Bjernstad et al. (1988) in experimental work have also demonstrated the effect on aroma retention by initial rates of drying, thus by drop velocities and air mass flow. Yamamoto (1995) carried out interesting experimental work in a sophisticated set-up for determining volatile retention during drying of single suspended droplets, with discussion and review. The Niro company continues to offer the conventional tall spray towers with pressure nozzles for coffee spray drying to powers; but now has also the multistage fluidized spray drier for more granular products. This comprises an integrated fluid bed with a separate warm air supply at the bottom of the chamber, to cause fluidization of the semi-moist powder entrainment of fines. Agglomerated particles (300 to 500 mm) of suitable size and strength are removed from the fluidizing bed (Grùnlund, 1995). Where granules of a much larger particle size are required, a separate agglomerating unit is used, in which the spray dried particles are fed to a rotating disc in a chamber fitted with either a water/mist supply or steam. The granules formed are then fed to a separate fluid bed drying/cooling unit. The production of granular particles or granules by an entirely new approach features in a Nestle patent, USP 5 750 178 (1998), in which a molten mass of soluble coffee (with about 15% moisture) with added volatile aroma compounds is extruded to form a hard glass, which is then ground to particles of required size. The foaming of coffee extract before spray drying is used to prepare a `soluble espresso coffee', USP 5 882 717 (1999), Kraft Foods Inc. The instantization of water soluble powdered foods, in general, by a compaction and subsequent steam jet agglomeration, is described by Hogekemp and Schubert (1997).
6.2.9 Freeze drying Freeze drying continues to be a widely used method of drying coffee extracts, especially those designed for `premium' quality products. The scientific literature does not reveal any new development. The patent on the freeze drying of coffee extracts
132
Coffee: Recent Developments
from Suwerlack, in Germany, should be noted: DE 195 19 129 (1996). Freeze drying is favoured for its capacity for retaining volatile compounds; further experimental data have been provided by Pardo et al. (1999), based on some six selected volatile compounds under different freezing and freeze drying conditions.
6.2.10 Aromatisation This term is usually applied to a process, whereby essentially headspace coffee aroma volatiles are made available by `plating' coffee aroma oil, prepared by expression methods from roast coffee, or other sources, onto the soluble coffee, usually at the packing stage. Such oils have also been used as vehicles for more flavour generating volatile compounds, from other coffee sources. The use of commercial electronic noses was studied to assess the presence, or otherwise, of aroma oil `plated' onto instant coffee samples, i.e. in-jar aroma (Gretsch et al., 1997). Two types were studied; a metal oxide sensor (Fox System) was shown to provide good discrimination based on the detection of sulfur compounds. Further information on expression methods (e.g. Anderson expellers) appears from time to time in the literature, for example Falla et al. (1989).
6.2.11 Spent grounds disposal The most economic method of disposal of the wet spent grounds from coffee percolation batteries after the required coffee soluble solids have been taken, continues to exercise the attention of coffee technologists. An interesting paper (Stahl & Turek, 1991) shows how this waste can be considered as a chemical feedstock for the production of higher value products, specifically D-mannose and, upon reduction, D-mannitol, which is a valuable product in the pharmaceutical and food industries. Spent grounds (dependent upon Table 6.2
the original degree of extraction of soluble solids), are now known to contain about 15% cellulose, 25% mannan and 5% arabino-galactan (dry basis). These are carbohydrate polymers, which can be hydrolysed to their basic monosaccharides by aqueous sulphuric acid and in a suitable reactor at elevated temperatures. Conditions were found for the optimum and selective production of mannose, whilst the conversion to mannitol is a conventional industrial process. The patents based on this process are USP 4 484 012 (1984), EP O178 357B (1988) and USP 4 508 745. Table 6.2 lists patents for obtaining other components from spent grounds.
6.2.12 Grading, storage and blending of green coffee The optimal selection of green coffees for use in instant coffee manufacture continues to be important to the manufacturer/consumer. Whilst mainly directed towards roast and ground coffee products, some recent methods of green coffee `improvement', such as steaming, are listed in Table 6.3.
6.2.13 Liquid extracts There has been a continued interest in the stability of fresh coffee extracts, as these may also be considered as `instant' or `ready-to-drink' products. Stability is closely related to temperature of storage; long term stability seems only possible at temperatures as low as 7208C (Severini et al., 1991), primarily due to the development of acidity (lower pHs) and other changes. Changes may be due to micro-organisms, oxidation, reactions between components in aqueous solution and the hydrolysis of quinides to further quinic acid, all dependent also upon the particular coffee extract considered. Prevention of micro-organism growth by pasteurisation in itself is not adequate, nor by use of very high pressure, for example up to 7000 bar
Some patents for extracting useful components from spent coffee grounds.
Country
Patent or application No
Date
Applicant
Remarks
EP
0239 730 A1
1987
Nestle´
Preparation of cafestol
EP
0223 982 A2
1987
Ð
Anti-foaming agents
EP
0819 385 A1
1988
Nestle´
Deterpenation of coffee wastes
Technology III: Instant Coffee
Table 6.3
133
Some patents for the improvement of green coffee.
Country
Patent or application No
Date
EP
0735 631
1997
EP
0478 839 A1
EP
Applicant
Remarks
Darboven GmbH
Steaming
1992
Asama (Japan)
Addition of tannins
0282 762 A2
1988
Jacobs Suchard
Increasing acidity
EP
0282 345 A2
1988
Compack, Hungary
Robusta coffee
EP
0271 957 A2
1988
Proctor & Gamble
Alkali treatment
(Severini et al., 1995). The use of antioxidants has recently been patented (NestleÂ, EP 0934 702, 1999) with claimed success against oxidative reaction (see also Chapter 1). A commercial concentrated liquid coffee extract has recently been marketed (Cafinesse, Douwe-Egberts), with storage below 58C recommended for not longer than one month.
6.3 PHYSICAL PROPERTIES OF VOLATILE COMPOUNDS 6.3.1 Important physical properties in relation to instant coffee processing The seminal studies by Thijssen showed the particular significance of two physical properties of volatile compounds in instant coffee processing (Clarke, 1987a). These were (1)
a1 j;w , the relative volatility of the particular volatile compound of j in aqueous solution at infinite dilution w referring to pure water as in aroma concentrates. In practice, the solution of interest may contain soluble coffee solids, so that w becomes ws.
Table 6.4
(2) Similarly Dj;w the liquid molecular diffusivity or diffusion coefficient of component j, which is a complex property, and is often considered as a ratio Dj /Dw in a given solution, where Dw is the diffusion coefficient for water. The use of a1j and Dj values in equations that have been derived to describe typical instant coffee processing operations, such as evaporation/stripping and spray drying, have been described in the many publications of Thijssen, and reviewed (Clarke, 1990), where the factors influencing the values such as temperature and solubles/acid content of the solution, are considered. Of the 800 or more volatile compounds in roast coffee and in coffee extract, even now, values of these properties are only known directly for very few, although values are predictable (though not published) for a greater number. The determination of these two physical properties in these ways is shown in Table 6.4. A new and convenient static headspace method for determining air±water partition coefficients Kaÿw (also air±coffee oil and oil±water coefficients for the same compounds) has been described (Gretsch et al., 1995), and results given for a number of coffee volatile compounds of interest. Thus for acetaldehyde (ethanal) and ethyl acetate
Determination of physical properties.
Property
Direct Measurement
Predictive from
a1 j;w (dimensionless)
By GC for partition coefficient (air±water)
(1) Vapour pressure + (2) activity coefficient from (a) solution aqueous solubility, or (b) Pierrotti correlation
Dj (m2 s±1 units)
Various direct methods
Molal volume and Wilke±Chang equation
134
Coffee: Recent Developments
(ethyl ethanoate) in very dilute solutions (10 to 2000 ppm), values of Kj;aÿw were determined at 308, 408 and 608C with a precision range +10±30%, and were found to be consistent with results published by others at lower temperatures, from the expected linear relationship between ln K and I/T (absolute temperature, K). Values for a number of other compounds were determined, categorised as high and medium volatiles; but it was stated that very low partition coefficients could not be assessed by this method. Values of Kj;aÿw are simply related to a 1 j;w values (i.e. at infinite dilution), which is the physical property preferred for chemical engineering calculation purposes thus, ÿ6 P sw a1 j;aÿw K j;aÿw =0:97 10
Table 6.5
where P sw is the vapour pressure of water at the same temperature. Thus at 608C, at which temperature aroma volatile stripping might be conducted (water vapour pressure 149 mm) the following data are presented. In Table 6.5, the low relative volatilities for the pyrazines is of interest, when it is also probable that they will be even lower in coffee extracts, which are acid, due to cationization; lower molecular weight pyrazines are also very soluble in water. Buttery et al. (1969) determined the partition coefficients of both 2methyl and 2-ethyl pyrazine, but at 258C when their values were reported as follows (water vapour pressure, 27.3 mm Hg): K a1 jw ÿ5 2-Methyl pyrazine 9 10 3.4 3.8 2-Ethyl pyrazine 10 10ÿ5
Relative volatilities of some organic compounds. Partition coefficient1 Kj;aÿw at 608C
Relative volatility2 a1 j;w
2.91 6 10±2
200
Ethanol (acetaldehyde)
1.60 6 10±2
110
2-Methyl propanol (isobutanol)
1.64 6 10±2
113
3-Methyl butanol (iso pentanol)
6.56610±2
453
Dimethyl disulfide
1.52 6 10±1
1050
Medium volatiles 2-Methyl pyrazine
2.80 6 10±3
19.3
2-Ethyl pyrazine
±3
5.20 6 10
35.9
2, 6 Dimethyl pyrazine
7.92 6 10±4
5.4
±4
4.2
Volatile compound (j) High volatiles Ethyl ethanoate (ethyl acetate)
Trimethyl pyrazine
6.05 6 10
2, 3, -Butadione (diacetyl)
±3
7.71 6 10
53
2, 3 Pentadione (acetylacetone)
1.33 6 10±2
92
Furfural
1.13 6 10±3
78
±2
147
Ethanethiol (ethyl mercaptan) 1 2
Data from Gretsch et al. (1995). Calculated from partition coefficient.
2.13 6 10
Technology III: Instant Coffee
135
These figures also show the known increase in the relative volatility (in infinitely dilute solutions) of compounds in a homologous series, though they have increasing molecular weight and boiling point (of the pure liquid component), i.e. compare 2,3 butadione with 2,3 pentadione, and similarly 2-methyl propanol, with 2-methyl butanol. Both partition coefficients (directly determined) and relative volatility values (derived) will vary with temperature, as shown in Table 6.6. It can be seen that partition coefficients are decreasing quite markedly with a lower temperature of the aqueous solution, of significance in tasting. Relative volatility values, of more interest in processing (evaporation etc.) tend to increase, although not markedly. As will be noted later, neither of these two compounds are, however, of great significance in roast coffee/instant coffee. The data for partition coefficients above shows good straight line relationships between ln K and temperature I/T (K). Both relative volatilities and partition coefficients are dependent upon so-called activity coefficients, (g1 j;w ), so that 1 s s a1 j;w gj;w P j =P w
and K 0:97 P sj 10ÿ6 g1 j;w
compounds, a the volatile compound and w water at a given temperature. Thermodynamically @ln g1 j j =R ÿAh @
1=T j is the partial molal heat solution of the where Ah component. This means that when t 6 log g is constant, the heat of solution is negative as with ethyl acetate and acetaldehyde, though it cannot be clearly demonstrated with the above data. g1 j;w is, in fact, separately related to the water solubility of the compound j in water w, so that the higher value of g1 j;w , the lower is the saturation water solubility. Any further data on solubility are to be welcomed for predictive relative volatility values by relatively simple methods, although do not seem to be forthcoming, as indicated in Tables 6.7 and 6.8. Values of g1 j;w are of separate use in coffee extraction or brewing studies since percentage yield of volatile compounds is a function of the water/coffee ratio and 1/g1 j;w . Similarly, further data for diffusion coefficient values, directly determined, appear to be scarce, although predictive methods can be applied, of use in spray and freeze-drying studies, where the percentage retention of a given compound will be a function of 1/D, all other conditions being equal.
where P sj and P sw are the vapour pressures of the pure
Table 6.6
Variation of relative volatility with temperature. Vapour pressure Ps (mm Hg)
Temp (8C)
1/T (K) 6 103
K 6 102
7ln K
For ethyl acetate (j) and water (w) 80 2.83 6.00 60 3.00 2.91 40 3.19 1.11 25 3.35 0.46 3.35 0.63 3.35 0.52
2.81 3.53 4.50 5.51 5.00 5.25
For acetaldehyde and water 60 3.00 40 3.19 30 3.30
1.6 0.7 0.42
3.53 4.50 5.40
25
0.27
5.65
3.35
j Ð 436 281 93 93 93
w
a1 j;w
Data source
355 149 55 27.3 27.3 27.3
174 199 208 256 237 196
Gretsch et al. (1995) Gretsch et al. (1995) Gretsch et al. (1995) Chandrasekan & King (1972) Kolb et al. (1992) Extrapolated from Gretsch et al. (1995)
149 55 32
110 131 136
Gretsch et al. (1995) Gretsch et al. (1995) Gretsch et al. (1995)
102
Buttery et al. (1969)
27.3
136
Coffee: Recent Developments
Table 6.7 Physical properties of roasted coffee volatile compounds deemed to be of especial coffee flavour importance. Solubility Substance name (synonym)
Formula
Molecular weight
Boiling point1 (8C)
Wt % of soln
Parts per 100 of water
2-Furfurylthiol, or furylmethanethiol (furfuryl mercaptan)
C5H6OS
114
160 84/65 mm
NAA
NAA
4-Hydroxy-2,5 methyl-3(2H)-furanone (furaneol)
C6H8O3
128
Ð
S/Sol
Ð
2-Methyl-butanal
C5H10O
86
Ð
Ð
Ð
3-Methyl-butanal (isovaleraldehyde)
C5H10O
86
93
NAA
S/Sol
2-Methoxy-phenol (guaiacol)
C7H8O2
124
205 163/200 mm
1.87
1.9 at 158C
4-Vinyl-guaiacol
C9H10O2
150
204/100 mm
NAA
NAA
2,3 Butadione
C4H6O2
86
88
Ð
Ð
2-Methyl-3-oxa-8-thiabicyclo [3,3,0]-1,4 octadiene, or 4'5'-dihydro-2-methylthiopheno [3,4-b]furan or (kahweofuran)
C7H8OS
140
105/20 mm
NAA
NAA
C6H6ON2
122
80
Ð
Ð
2-Methyl-3-furanthiol
C5H6OS
114
160 84/65 mm
Ð
Ð
2,3,5, trimethylpyrazine
C7H10N2
122
61/62/35 mm
Sol in water
Ð
Ð
Ð
Ð
Ð
Ð
C5H8O2
100
108
7%
1:15
2-Acetyl pyrazine
(E)-b-Damascenone Pentan-2, 3-dione 3-Mercapto-3-methylbutyl formate
C6H1302S
149
Ð
Ð
Ð
2-Ethyl, -3, 5-dimethylpyrazine
C8H12N2
136
65/8 mm
Ð
Ð
3 Isobutyl-2-methoxypyrazine
C9H14ON2
166
Ð
Ð
Ð
3-Hydroxy -4,5-dimethyl-2 (5H)furanone (sotolon)
C6H8O3
128
Ð
Ð
Ð
5-Ethyl-4-hydroxy-2 methyl-3(2H)furanone
C6H8O3
128
Ð
Ð
Ð
3-Methyl propanal
C4H8O
72
Ð
Ð
Ð
Methanethiol (methyl mercaptan)
Ch4S
48
6
Ð
2.3
Compounds from Silvar et al. (1987), Grosch (1995, 1999), Holscher (1991). 1 At 760 mm, unless otherwise indicated. NAA or Ð, data `not apparently available'. S/Sol slightly soluble.
Technology III: Instant Coffee
Table 6.8 flavour.
137
Physical properties of roasted or green coffee volatile compounds considered undesirable to Solubility
Substance name (synonym)
Formula
Molecular weight
Boiling point (8C)
Wt % of soln
Parts per 100 of water
Furfural or 2-furfuryl aldehyde
C5H6O2
96
181.7 121.8/200
8.3
9.1 at 208C
2-Methyl isoborneol
C11H20O
168
Over 200
Ð
Ð
2, 4, 6-tri-chloroanisole (TCA-Rio flavour)
C7H6OCl3
211
240
Ð
Ð
Ethyl 2-methyl butanoate (ethyl ester of methyl butanoic acid)
C7H13O2
129
Ð
Ð
Ð
Ethyl 3-methyl butanoate
C7H13O2
129
Ð
Ð
Ð
Ethyl carboxy cyclohexane (ethyl ester of cyclohexane carboxylic acid)
C9H16O2
156
Ð
Ð
C9H14ON2
166
Ð
Ð
2-Methoxy-3-isopropyl-pyrazine
Ð
Compounds from Liardon et al. (1989), Bade-Wegner et al. (1993, 1995, 1997).
6.3.2 Tables of physical properties Compilations of some of the relevant physical properties of the various coffee volatile compounds (i.e. boiling points at different pressures, solubility data where available) have been published (Clarke, 1986, 1990, 1991). The latter compilation related specifically to sulfur compounds such as the large numbers of thiazoles, thiophenes and alkylthiols present in roasted coffee, although few of these are now singly recognised as being significant flavour components. Table 6.7 relates, however, to components which are deemed to be highly significant (Grosch, 1995; Chapter 3 in this book). To provide, therefore, for instant coffees which also have these flavourful compounds present, full knowledge of their physical properties in relation to extraction, general volatile compound handling and drying is very desirable. Recently, attention has been directed towards certain compounds known to be highly undesirable which are found in green coffee and which may reach instant coffee. A selection is given in Table 6.8
REFERENCES Bade-Wegner, H., Bendig, I., Holscher, W., Wolkenhauer, P. & Vitzthum, O.G. (1995) Off flavour elucidation in certain
patches of Kenya Coffee. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 174±82. ASIC, Paris, France. Bade-Wegner, H., Bendig, I., Holscher, R. & Wollmann, R. (1997) Volatile compounds associated with the over-fermented effect. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 178±82. ASIC, Paris, France. Bade-Wegner, H., Holscher, W. & Vitzthum. O.G. (1993) Quantification of 2-methyl isoborneol in roasted coffee by GCMS. In: Proceedings of the 15th ASIC Colloquium (Montpellier), pp. 537±43. ASIC, Paris, France. Bassoli, D.G., Suni, A.P., Akeshi, V. & Castro, A.S. (1993) Instant coffee with natural aroma by spray drying. In: Proceedings of the 15th ASIC Colloquium (Montpellier), pp. 712±18. ASIC, Paris, France. Bjernstad, A., Pedersen, A. & Schwartzberg, H. (1988) Drop velocities and air incorporation in nozzle atomiser dryers. In: Process Technology Proceedings, Vol. 5, Preconcentration and Drying of Food Materials, (ed. S. Bruin). Elsevier, Amsterdam. Bomben, J.L., Bruin, S., Thijssen, H.A.C. & Merson, R.M. (1973) Aroma recovery and retention. In: Advances in Food Research, pp. 2±111, Academic Press, New York. Bradbury, A.G.W. & Atkins, E.D.T. (1997) Factors affecting mannan solubility in roast coffee extracts. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 128±32. ASIC, Paris, France. Bruin, S., (Ed.) (1988) Process Technology Proceedings, Vol. 5, Preconcentration and Drying of Food Materials. Elsevier, Amsterdam. Buttery, R.G., Gudagni, D.E. & Ling, L.C. (1969) Volatities of
138
aldehydes, ketones and esters in dilute water solution. J. Agric. Food Chem., 17, 385±9. Cale, K. & Imura, N. (1993) Recovery of beneficial coffee aromas from thermal hydrolysates. In: Proceedings of the 15th ASIC Colloquium (Montpellier), pp. 685±93. ASIC, Paris, France. Casimir, D.J. & Huntington, J.N. (1970) In: Proceedings of the XV International Federation Fruit Juice Procedures Conference, Berne, Switzerland. Casimir, D.J. & Craig, A.J.McA. (1990) Flavour recovery using the Australian spinning cone/column. In: Engineering and Food, (eds W. E. Spies & H. Schubert), pp. 106±77. Elsevier Applied Science, Barking. Chandrasekan, S.K. & King, C.J. (1972) A. I. Ch. E. J., 18 513± 20. Clarke, R.J. (1986) The flavour of coffee. In: Developments in Food Science 3B. Food flavours, Part B, (eds I.D. Morton & A.J. Macleod), pp 1±47. Elsevier, Amsterdam. Clarke, R.J. (1987a) In: Coffee, Vol. II Technology, (eds R.J. Clarke & R. Macrae), pp. 35±58 (Grading), pp. 73±108 (Roasting and grinding), pp. 109±46 (Extraction), pp. 147±200 (Drying) and pp. 201±20 (Packing). Elsevier Applied Science, Barking. Clarke, R.J. (1987b) Coffee technology. In: Quality Control in The Food Industry, (ed. S.M. Herschdoefer), pp. 161±92. Academic Press, London. Clarke, R.J. (1988) International standardization. In: Coffee, Vol. 6, Commercial and Technico-Legal Aspects, (eds R.J. Clarke & R. Macrae), pp. 105±42. Elsevier Applied Science, Barking. Clarke, R.J. (1990) Physical properties of the volatile compounds of coffee. CafeÂ, Cacao, TheÁ, XXXIV (No. 4) 285±94. Clarke, R.J. (1991) Physical properties of the volatile compounds of roasted coffee. In: Proceedings of the 14th ASIC Colloquium (San Francisco), pp. 331±8. ASIC, Paris, France. Clarke, R.J. (1993a) Instant coffee. In: Encyclopaedia of Food Science, Food Technology and Nutrition (ed. R. Macrae), pp. 1126±31. Academic Press, London. Clarke, R.J. (1993b) Coffee. In: Shelf Life Studies of Foods and Beverages, Developments in Food Science, No. 33, pp. 801±19. Elsevier, Amsterdam. Clarke, R.J. & Macrae, R. (eds) (1988) Appendix I. In: Coffee, Vol. 6, Commercial and Technico-Legal Aspects, p. 202. Elsevier Applied Science, Barking. Dithmer, L. (1995) New developments in foaming and freezedrying. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 463±9. ASIC, Paris, France. Falla, L., Ospima, J. & Posada, E. (1989) Determinacion de las condiciones de extraccion de aceite de cafe per extrusion. In: Proceedings of the 13th ASIC Colloquium (Paipa), pp. 232±5. ASIC, Paris, France. Frey, D.D. (1984) Doctoral dissertation, University of California, Berkeley. Gretsch, C., Delame, J., Toury, A., Visani, P. & Liardon, R. (1997) Detection of aroma above a coffee powder: limits and perspectives of electronic sensors. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 183±90. ASIC, Paris, France. Gretsch, C., Grandjean, G., Haering, M., Liardon, K. & Westfall, S. (1995) Determination of the partition coefficients of coffee
Coffee: Recent Developments
volatiles using static headspace. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 326±31. ASIC, Paris, France. Grosch, W. (1995) Instrumental and sensory analysis of coffee volatilities. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 147±56. ASIC, Paris, France. Grosch, W. (1999) Key odorants of roasted coffee. In: Proceedings of the 18th ASIC Colloquium (Helsinki). ASIC, Paris, France. Grùnlund, M. (1995) Recent trends in soluble coffee. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 457±62. ASIC, Paris, France. Hogekemp, S. & Schubert, H. (1997) Instantization of water soluble powdered foods by compaction and subsequent steam jet agglomeration. In: Engineering and Food at ICEF 7, (ed. R. Jowitt), pp. 32±7. Academic Press, Sheffield. Holscher, W. (1991) Thesis, University of Hamburg. Imura, M. & Danno, S. (1995) Effect of extract pre-treatment via micro filtration on the concentration of coffee extract via reverse osmosis. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 470±77. ASIC, Paris, France. Jansen, H.A. & Van Pelt, W.H.J.M. (1987) Freeze concentration economics and applications. In: Process Technology Proceedings, Vol. 5, Preconcentration and Drying of Food Materials, (ed. S. Bruin), pp. 77±86. Elsevier, Amsterdam. Jowitt, R. (ed.) (1997) Marcel Loncin±In Memoriam. In: Engineering and Food at ICEF7, (ed. R. Jowitt), pp. 33±42. Academic Press, Sheffield. King, G.J. (1988) Spray drying of food liquids and volatiles retention. In: Process Technology Proceedings, Vol. 5, Preconcentration and Drying of Food Materials, (ed. S. Bruin), pp. 147±62. Elsevier, Amsterdam. Kolb, B., Bichler, C. & E., Welter, C. (1992) Chromatographia, 34, pp. 235±240. Liardon, R., Spadone, J.C., Braendlen, N. & Dentan, E. (1989) Multi-disciplinary study of Rio flavour in Brazilian green coffee. In: Proceedings of the 13th ASIC Colloquium. (Paipa) pp. 117±26. ASIC, Paris, France. Moy, S., Shah, N. & Pyle, D.L. (1997) Flood and efficiency studies on a spinning cone column. In: Engineering and Food at IECF, (ed. R. Jowitt), pp. 133±6. Academic Press, Sheffield. Noyes, R.M. (1995) Relative extraction yields of green coffee. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 309±16. ASIC, Paris, France. Ohtani, N., Takchaski, K., Yamura Y. et al. (1995) Spray drying instant coffee product at low temperatures. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 447±56. ASIC, Paris, France. Papadekis, S.E. (1987) Dissertation at the University of California, Berkley. Pardo, J.M., Mottram, D.S. & Niranjan, K. (1999) Relation between volatile retention and movement of the ice front during freeze drying of coffee. In: Proceedings of the 18th ASIC Colloquium (Helsinki), pp. 150±58. ASIC, Paris, France. Severini, C., Nicoli, H.C., Dalla Rosa, M. & Lerici, C.R. (1991) Effect of some extraction conditions on brewing and stability of coffee beverage. In: Proceedings of the 14th ASIC Colloquium (San Francisco), pp. 649±56. ASIC, Paris, France.
Technology III: Instant Coffee
Severini, C., Nicoli, M.C., Romani, S. & Pinnavaia, G.C. (1995) Use of high pressure treatment for stabilizing coffee brew during storage. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 498±500. ASIC, Paris, France. Silvar, R., Kamperschroer, A. & Tressl, R. (1987) Gaschromatographisch massenspecktrometrische Untersuchungen des Rostkoffeearomes. Chemie, Mikrobiologie, Technologie der Lebensmittel (Nuremburg), 10, 176±87. Stahl, H. & Turek, E. (1991) Acid hydrolysis of spent grounds to produce D-mannose and D-mannitol. In: Proceedings of the 14th ASIC Colloquium (San Francisco), pp. 339±48. ASIC, Paris, France.
139
Steinhaeser, V., Oestrich-Jansen, S. & Baltes, W. (1999) Model experiments in the extraction of roast and ground coffee during the production of instant coffee. Deutsches Lebensmittel±Rundshau, 95, pp. 257±62. Sylla, K.J. (1989) Processing of instant coffee. In: Proceedings of the 13th ASIC Colloquium (Paipa), pp. 219±25. ASIC, Paris, France. Viani, R. (1986) Coffee. In: Ullman’s Encyclopaedia of Technology. Yamamoto, S. (1995) Aroma and enzyme retention during drying. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 457±62. ASIC, Paris, France.
Chapter 7
Technology IV: Beverage Preparation: Brewing Trends for the New Millennium M. Petracco illycaffeÁ, Trieste, Italy 7.1 INTRODUCTION `Once upon a time' is the usual incipit of the very many legends about the discovery of coffee as a food. Sometimes, they tell the story of shepherds consuming directly the seeds of some Coffea plant, by chewing either raw cherries or cooked, generally roasted, beans (Burton 1860). For those of our readers who have tried to munch a bean, it is hard to imagine that the success of coffee ± today the most traded food commodity ± could have ever occurred, if persisting in those primitive habits. On the one hand, it seems logical that the appeal of coffee to those early `food science pioneers' derived essentially from their experiencing an arousal condition that proved to be beneficial to their activities (Ellis, 1998). In other words, the primordial reason for coffee consumption must have been the, by now well documented, physiological effects of caffeine in the human organism (Viani, 1988). On the other hand, it is pretty obvious why coffee has become so popular, up to achieving the position of the second most largely consumed beverage after water: it is a matter of flavor, or better still of overall sensory impact. Whilst people like the flavor of coffee, they do not like the disturbing sensory feeling of chomping and swallowing hard particles deriving from a bean. This fact makes beverage preparation a key step for enjoying the benefits of this commodity, and sometimes for transforming it into a specialty. From a chemical point of view, a beverage can be defined as a liquid system with some nutritional factors. Water is, of course, the primary liquid that forms the matrix of all beverages, due to the basic physiological requirement of maintaining a balanced water content in bodily tissues. Nutritive substances may be present within the matrix as solutes ± i.e. in molecular form ± and/or as dispersed phases, such as an emulsion of other immiscible liquids, a foam-forming effervescence of gaseous material and suspended solid particles.
Coffee beverages are no exception in this respect; one classification criterion is to group them according to their content of hetero-phases, ranging from smooth pure solutions (e.g. drip filter coffee) to emulsions (e.g. Nordic boiled coffee) to thick suspensions (e.g. Turkish style brew). Effervescence, when present, terminates inevitably with the surfacing of gas bubbles and possibly with the build-up of a foam layer (e.g. espresso). To make things more complex, a number of additives can be attached to the basic beverage, which are sometimes dispersed in a homogeneous phase (e.g. sugar), and sometimes produce more or less composite multi-phase beverages (e.g. milk in lattes and cappuccinos). The operations implemented in order to convey coffee nutrients into the beverage are generically called brewing, but may be more rigorously defined as solid± liquid extraction (Pictet 1987). The latter is a unit operation much used in chemical engineering and in food technology in particular, where it is mostly practised as a discontinuous (batch) process. Its application to any coffee, either green or roasted, either whole bean or ground powder, leads to an aqueous liquor that may be used as such to prepare a beverage or constitute an intermediate semi-manufactured product. An example of this industrial extraction is instant coffee manufacture (Clarke, 1987a), where the liquor derived from coarsely ground roast coffee is later submitted to spray or freeze drying to yield the well known soluble powder (or granules), herald of the era of convenience foods. The total redissolution of instant coffee, which can be made in warm ± not even necessarily hot ± water, permits us to enjoy a basic coffee cup with no frills, when proper extraction apparatus is not available. There are, to the author's knowledge, no claims about brewing green coffee bean material to produce a beverage, even if the active physiological component, caffeine, can be effectively extracted by hot water from
140
Technology IV: Beverage Preparation
triturated raw seeds. The grassy, astringent taste of such a brew is surely a deterrent against any commercial tentative in that direction. This chapter will deal, therefore, just with brewing techniques that start from roasted coffee beans, almost invariably through a grinding step useful to increase the amount of solid surface exposed to water (Clarke, 1987b). It must, however, be noted that the size of a single serving, or cup, is enormously variable in different cultures, ranging from 15 ml of concentrated espresso in Sicily to over 250 ml in the USA, and can derive from the brewing of a roast and ground coffee amount as little as 5 g, up to 15 g or more. Therefore, any figure quoted in quantitative assessments of coffee beverage properties or consumption must be taken with due caution. For the purpose of preparing a beverage, several brewing techniques have been developed and, in some cases, brought close to perfection during the centuries of coffee history. Most of them are linked to local traditions and are therefore better known by their geographical denomination than by the description of the method itself. Anyway, in the global market of the end of the millennium hardly any one of those practices misses a commercial utilization, with consequent production of tools and machines that in some cases have become objects of fashion and met with wide distribution. Much higher is the number of devices conceived for domestic rather than for professional usage, because, in home brewing, two aspects foster the implementation of fanciful ideas: the small scale of the apparatus, usually designed to produce 1 to 12 cups per run, and the limited life expectancy of the appliance, which makes it possible to manufacture it in an inexpensive way. Professional catering machines have usually become, on the contrary, objects of no compromise, since the success of the specialist coffee brewer ± call him maestro, or plainly bartender or waiter ± depends to a large extent on careful design and construction, as well on easy maintenance and proper servicing, of the brewing equipment. Coffee machines are, nowadays, complex pieces of advanced engineering, mostly equipped with sophisticated computers on board controlling every function and sometimes making their operations possible even to the end user, the customer, with full satisfaction. To conclude this introduction, a short account must be given of some phenomena, which are little related to the conventional meaning of coffee brewing: let us call them `modified coffee beverages'. Besides the wellestablished habit of pouring some milk or cream in the coffee cup, two innovative practices have recently
141
surfaced: the artificial flavoring of coffee products and the industrial concoction and preservation of coffee drinks. The addition of foreign flavors to a product ± the coffee cup ± that was starting to be seen in the 1980s as somehow old-fashioned, currently constitutes a profitable business in the USA. The last millennium decade ± the 1990s ± saw, as well as the increasing popularity of the fashionable habit of brewing the coffee cup in the face of the guest, also some success of its antithesis. Among younger generations, a can of coffee-based beverage is, nowadays, often replacing other established soft drinks. But the preservation of coffee beverages poses serious hygienic problems: while green coffee can be stored for months with little precaution, and whereas roast and ground coffee is rather trouble-free as regards microbiological attack, all coffee beverages are an attractive culture broth for bacteria and must therefore be sterilized and sealed to maintain their edibility.
7.2 EXTRACTION METHODS The intimate contact of water with roasted coffee solids is the cardinal requirement for producing a coffee beverage: chemical engineers call this unit operation `leaching', a term which originally referred to percolation of liquid through a fixed bed, but is now used to mean solid±liquid extraction generally. By definition (Miller et al. 1984), leaching is the removal of a soluble fraction, in the form of a solution, from an insoluble, permeable solid phase with which it is associated. Because of its variety of applications to several industries, this operation is also known by a number of other names, such as lixiviation, percolation, infusion, decoction, washing and maceration. Two steps are always involved in solid±liquid extraction: contact of solid and solvent to effect mass transfer of solubles to the solvent, and separation of the resulting solution from the residual solid: this latter step is usually done by filtration, or even centrifugation, but some favoured brews use just the less effective decantation. The mechanism of the first step is favored by increased surface per unit volume of solids to be leached and by decreased radial distances that must be traversed within the solids, both of which are assisted by decreased particle size. It is therefore rather unlikely that any brew is obtained by soaking of whole roasted coffee beans, which are very impermeable. The grinding operation will not be dealt with here; the reader is therefore referred to the relevant specialist literature (Petracco & Marega, 1991, Petracco 1995a).
142
Particle size, albeit the most obvious to the customer, is just one of the several independent variables influencing coffee extraction, a dynamic process; the other main ones (not to mention the intrinsic factors linked to coffee bean quality, like agronomic and roasting parameters) are solid/water ratio, contact time and water temperature. More variables enter the picture if considering special extraction cases like espresso, in which water energy deriving from pump pressure is of paramount importance. The principal dependent variable that can be objectively measured (along with the obviously all important sensory ones, which are largely based on subjective evaluation) is brewing yield, namely the ratio between the mass of the coffee material that passes into the cup and the total coffee material used (the balance to be disposed as spent grounds). It is worth mentioning right away that yield is a different concept from beverage concentration, or strength, measured in grammes of extracted matter per litre of beverage, and can be set independently, if a different brewing formula (coffee/water ratio) is chosen. Extraction yields may range between 14% and 30%, with darker roasting degrees procuring higher yields. Brewing temperature exerts a predominant effect on yield variability, whilst both contact time and grinding degree have a limited influence and brewing formula only a marginal one, as shown in Fig. 7.1, adapted from Nicoli et al. (1990). A naive way of thinking could suggest that the more one extracts from the purchased material, the better: nothing is falser, quality-wise. Since the many chemical species identified in roasted coffee (more than 1800 so far) exhibit different extraction rates, it is logical that extraction yield relates to sensory quality too, leading to the two celebrated converse brewing errors: under- and overextraction. The former is due to lower yields (deriving from a variety of reasons such as low water temperature, short contact time or a coarse grind),
Coffee: Recent Developments
when the most soluble substances ± the acidic and sweet ones ± are predominantly driven into the cup, producing the so-called underextracted beverage. The latter happens when higher yields are obtained, thanks to the opposite settings, by forcing more substances ± the bitter and astringent ones ± into the cup. It is therefore evident that brewing yield should be adjusted according to personal taste, in agreement with the nature of the beans used (origin, blend, roasting). The general principle ruling extraction is Fick's law of diffusion, which can be written in the following form (Barbanti & Nicoli 1996): s k T=Z A=x
C ÿ c y in which s is quantity of solute diffusing from a solid particle surrounded by a liquid, where: k = constant, depending on molecular factors T = absolute temperature Z = viscosity of the liquid, f(T) A = layer cross-section around the particle x = layer depth C = solute concentration in the solid c = solute concentration in the liquid y = contact time. Practical mass transfer equations of the extraction process have been established for the design of industrial soluble coffee plants; for a review of some of them, see Clarke (1987a, b). Most household coffee makers (and many catering ones too), however, are designed on an empirical basis. They may be grouped according to various principles: continuous/ discontinuous process, with/without recirculation, temperature below/above 1008C, pressureless/overpressure (Peters, 1991; Clarke, 1986; Pictet, 1987; Cammenga et al. 1997). Their classification will be made here from a qualitative perspective; the grouping criterion chosen takes into account both manner and
Fig. 7.1 Extraction yield as a function of the main brewing variables: (a) temperature, (b) grinding, (c) brewing formula (constant blend).
Technology IV: Beverage Preparation
time of coffee/water contact, but with no attempt to suggest optimum recipes, or even ranges for the relevant variables. An exception will be made for the espresso brewing method, which falls within the specific competence of the author (Petracco, 1989, 1995b; Petracco & Suggi, 1993).
7.2.1 Decoction methods One can speak of decoction (from the Latin verb decoquo, decocis, decoxi, decoctum, -ere, 38, translated as `to overcook'; Campanini & Carboni, 1995a), every time that a partially soluble solid is kept in contact with a given amount of water, at an appropriate temperature, not necessarily coincident with the boiling point, for a considerable amount of time, while allowing the concentration of solubles in the liquid to increase throughout the operation. Due to the law of mass action, the extraction rate decreases as the concentration increases, making an overly prolonged decoction ineffective and, as a side effect, possibly unfavorable to flavor and taste purposes due to volatile losses and hydrolytic changes. The beverage concentration (strength) increases with the time allowed for contact, before the grounds are separated from the liquid, also due to hydrolysis of the insoluble part of the former. Higher temperature favors, of course, higher extraction yields and rates. In the early years of coffee brewing, the main rule was to put ground coffee into water and make sure the water came to a boil. This was more a health principle than a taste clue, because boiling an often-contaminated drinking water was the only way to keep from getting sick. Later on, however, it was discovered that the taste of coffee was much better if the water was added to the grounds after it had been boiled.
(a) Boiled coffee The most basic coffee brewing method of all, boiled coffee, used to be the favorite hot beverage preparation in Nordic countries like Norway and Finland. It consists in putting coarsely ground coffee into water in a pot, or jug, and allowing it to warm up to boiling point on a stove, thus eliminating any need for thermometers. The resulting beverage is poured into a cup or mug after allowing some settling of the floating grounds, sometimes by the aid of a strainer. In order to have a hot beverage always ready, the remaining liquid is often left in the kettle, with the spent grounds, on top of the stove for longer periods of time, causing severe
143
hydrolytic changes in the grounds and possible further solubilization of poorly soluble compounds. Experimental studies indicate that boiled coffee may increase blood cholesterol, the reason for this being the considerable amount of suspended solids ingested along with the liquid: they contain an insoluble diterpene fraction, that mainly consists of cafestol, which has been shown to act as a blood cholesterol upregulator (Urgert, 1997). A simple filtering step through filter paper is enough to remove this fraction.
(b) Turkish coffee A brew not peculiar to Turks (Petropoulos 1979), it is served in all Mediterranean countries from Slovenia to Morocco. Its name encompasses several coffee brewing styles, including Greek coffee and Israeli `mud coffee', and only minor variations are to be seen among different cultures. Two distinctive tools are used for its preparation: the long brass grinder and the copper pot called an ibrik. The grinder is designed to produce a very fine powder, most of which is made up of fragments of roasted seed tissue where the cellular structure has been disrupted. These broken coffee cells are unable to float, as their density is higher than 1.46 (sinks in trichloroethylene); this helps in forcing the spent grounds to settle down, while at the same time allowing a strong extraction due to the increased exposed surface. The ibrik, a small truncated-conical and long-handled pot, is loaded with impalpable ground coffee, and usually sugar as well, then filled with cold water and placed on an open flame. After reaching the ebullition temperature, revealed by vigorous bubbling and foaming, the pot is removed from the flame and allowed to cool down a bit, then replaced on the flame to a second, and thereafter a third boiling. The ibrik's content is gently poured into a cup, which cannot prevent a considerable part of the floating grounds from entering the cup. This beverage must therefore be sipped with care in order to minimise the rather unpleasant feeling of a mouth filled with impalpable coffee grounds.
(c) Percolator coffee A favourite old-time brewing method in the United States, the percolator ± called informally `perk', perhaps for assonance with the energising properties of coffee ± is a continuous refluxing machine, in which the beverage is recycled as long as heat is applied (Peet & Thye 1954). The `secret' lies in the vessel's inner
144
construction, wherein a mobile chamber is situated below a funnel-shaped bell. When the water contained therein boils, a tiny steam pressure difference impels water upwards through the funnel's neck. The flow ends by gushing over a strainer filled with ground coffee, from which the `fresh' beverage trickles down into the base vessel, where it re-starts the boiling process over and over again. It is easy to understand how this repeated circulation enriches the liquid more and more with all extractable substances, exploiting the coffee grounds completely by drawing the less soluble (and, unfortunately, less pleasant) coffee components into the beverage. The resultant harsh and astringent taste, along with a loss of aroma's volatile compounds, has made the popularity of this brewing method decline during the last decades.
(d) Vacuum coffee This utilizes an elegant thermodynamic apparatus working between two temperatures obtained in different moments, the initial high one being derived from an open flame ± typically a spirit burner ± and the final low one being simply close to room temperature. A commercial model is called Cona, which is available as an electrically heated model too (Ehrenkranz & Inman, 1948). The device consists of a funnel-shaped glass flask, with a perforated screen, placed on top of a second one: ground coffee is put upon the screen and water in the bottom flask. The assembly of the two vessels is heated up to boiling, when steam pressure forces hot water up the neck of the funnel to invade the upper vessel and to mix with ground coffee, starting the extraction. After the heat source is removed, the assembly is allowed to cool down until the pressure in the lower flask has subsided, perhaps not to a real vacuum, but low enough to allow the upper liquid to flow down through the same funnel neck. Coffee may be served directly from the lower bowl.
7.2.2 Infusion methods A popular belief considers infusion as an act of allowing a pulverised material to steep in hot (but not boiling) water for a while, it is commonly set against decoction, seen as a term restricted to simmering procedures executed on the fire. From a chemical engineering standpoint, however, there is no major difference between an operation conducted at a constant tem-
Coffee: Recent Developments
perature of, say, 1008C or at a temperature decreasing from, say, 1008C to 808C: both are solution processes carried on under a given thermal treatment. The distinction is basically a quantitative one. In the coffee context, whilst decoction may be defined as a batch operation requiring a long contact between solids and water, one should speak of infusion every time hot water flows through a bed of partially soluble material, allowing just a short contact time to every elementary volume of clean liquid. This meaning stems from the Latin verb infundo, -is, infudi, infusum, ere 38, which may be translated as `to pour over' (Campanini & Carboni, 1995b). Another term that is often confounded with infusion is percolation: once more, a merely quantitative distinction allows one to speak about infusion as long as water trickles through a loose bed of coarsely ground coffee, without the need to employ mechanical forces. Percolation starts where no flow can be obtained through a denser bed, due to the phenomena of capillary and of colloidal imbibition, without the use of external power. Interestingly, in percolation processes some new variables appear, such as bed geometry and water flow. More on this will be found in Section 7.2.3. As an instance of infusion in the strict sense, pouring water on top of chopped tea leaves contained in a strainer that allowed the liquid to drip into a cup underneath used to be the main way to prepare a cup of tea before the appearance of paper teabags (sachets), which made the operation simpler and cleaner. The latter preparation is indeed a decoction at decreasing temperature, and could be properly called steeping, instead of infusion. However, in the real world things are seldom either black or white: thus, a certain amount of decoction must be present even in the strictest infusion process in order to allow a suitable head of liquid to build up and supply, by gravity, the pressure drop needed for seepage through the ground coffee and the filtering surface itself. Due to the shorter contact time of this `washing out' procedure, infused coffee beverages are generally sensorily assessed as milder than the decocted ones, often enhancing coffee's acidity and flavour.
(a) Filter coffee The product of the main brewing system now generally used in Northern Europe, filtered coffee is a clean, transparent beverage. The typical set-up consists of a simple device, where a paper filter is placed in a plastic cone-shaped holder. Medium ground coffee is put into
Technology IV: Beverage Preparation
the filter and the holder placed on top of a glass jug. Boiled water is then poured into the filter and allowed to seep through (Van Zante, 1968). The invention ± or at least the commercial implementation ± of this method, also known as drip coffee, is attributed to Mrs Melitta Bentz, as early as 1908 (Bentz, 1908). Numerous filter shapes and materials are available: besides the disposable paper ones, cloth filters can be used, but need proper maintenance for hygienic reasons. Also micro-perforated gold filters have become fashionable, since they are obviously enough, not disposable; care must be applied in keeping them clean without breaking their quite fragile structure. Many automatic or semi-automatic machines are available on the market to make preparation as easy as possible; some of them, electrically operated, produce hot water that actually drips at a controllable rate on the ground coffee, sticking to the proper infusion pattern: these only should deserve the name of `drip coffee' machines. Most of them include a heating plate on which the jug rests: by letting the brewed coffee dwell there for a long time, more chemical reactions can occur, which make the coffee taste change in a generally unpleasant direction.
(b) Napoletana coffee One of the simplest brewing methods by infusion, the macchinetta napoletana (also known as `flip drip pot') was already an Italian tradition in the first years of 1800. It uses gravity to make hot water percolate through a bed of medium-coarse ground coffee. A pot, on top of which a perforated bushel contains the ground coffee, is filled with water and heated indirectly by flames. After reaching the boiling point, the macchinetta is removed from the heat and turned upside down, allowing hot water to drip through the coffee into the second (upper) half of the device, which receptacle is also used to serve the beverage through a spout. The main mechanical difference between drip filter and Napoletana methods is that in the latter the ground coffee is immobilised between two filtering perforated plates, preventing the movement of the granules in the water: what actually occurs is water flowing through the coffee, not the coffee swimming within the water. The beverage is a rather strong one ± depending on the coffee/water ratio ± but, because the ground coffee undergoes some steam heating during all the time spent to heat water up to boiling point, a certain scalding occurs, which produces a bitter flavor.
145
7.2.3 Pressure methods Pressure is a well-known notion of general physics: it defines a force exerted perpendicularly over a unit of area. In the domain of fluid dynamics, pressure refers to an intensive type property, defined instant by instant in any point of space taken up by a fluid; it can be represented by means of a scalar field. The existence of a pressure field within a fluid yields a potential energy (Bernoulli's piezometric energy) that can be easily transformed into kinetic energy, thus giving speed to elementary masses of fluid. It is beyond dispute that every percolation method requires some pressure (at least several millimeters of water column height) as a driving force to overcome the head loss required for obtaining a flow through ground coffee beds; nevertheless, a higher relative pressure (one or two orders of magnitude) is needed when dealing with finer grinds. Even higher pressure, up to 10 atm (corresponding to 10 330 cm H2 O) must be used when trying to pass water through a compacted bed of finely ground coffee, usually called `cake' as in the espresso method. The energy expended during this kind of operation produces interesting effects, like driving of micron-size solid particles or oil droplets into the cup. This may change the beverage's properties dramatically, enhancing the sensory character often referred to as `body', as further explained in the subsequent sections.
(a) Plunger Also called, for some reason, French press (probably after a popular device called La CafetieÁre), this brewing method is particularly popular in restaurants, where it makes it possible to accomplish the brewing ceremony on the customer's very table. A very simple set-up is needed: a cylindrical glass beaker equipped with a freely moving piston (plunger), whose perforated base is lined with a very fine wire mesh. After removing the plunger, a spoonful of coffee (medium to coarse ground) is poured into the beaker and boiling water is added. These ingredients should stand in the beaker for a couple of minutes (decoction lurking again!), before the plunger and mesh filter are pushed down slowly (several seconds), allowing the sediment-free liquid to pass through. The beverage produced is a foamy, full-bodied turbid one, mostly due to fine suspended particles, but also to oil droplets. The maximum pressure exerted may be calculated by dividing the muscular force used to press the plunger downwards (somewhere between 2
146
and 10 kg) by the plate cross-section (from 20 to 200 cm2 ), ranging from 0.01 up to 0.5 relative atmospheres. It depends much on the fineness of grinding, up to the extreme situation of too finely ground, when no beverage passes through the filter no matter how much force is applied on the plunger!
(b) Moka By this trademark denomination is commercially known an inexpensive brewing machine, often misnamed stove-top espresso. Pressure is applied to the water by simply boiling it in an autoclave-type aluminum or steel kettle on an open flame, and forcing it through an immersed pipe across the coffee cake formed in a strainer basket. The beverage is conveyed through appropriate tubing into an upper vessel, screwed and sealed by a rubber gasket to the base kettle. Contact time depends mainly on the cake hydraulic resistance, which in turn hangs on three factors: the amount of ground coffee used, its particle size distribution (grinding fineness) and the tamping force exerted on it, whether deliberately or just by forcing down a heap of coffee while fastening the top part. The end of the brewing operation is usually announced by noisy steam flowing from the upper tube, to indicate water depletion. To reach a pressure level suitable for flowing across not very coarse grounds, the strict thermodynamic relationship between water pressure and temperature demands a very high temperature, well above 1008C: this process causes substances that are normally insoluble to be extracted, leaving in the cup a harsh bitter flavour often described as `burnt'. Unlike the regular espresso method, its pressure is far from the one met in actual espresso machines: around 9 atm, which makes its ability to produce the distinctive oil emulsion implausible (see later). Pressure in the Moka must not exceed 0.5 relative atmospheres (corresponding to 1108C), not only to prevent an excessive rise in water temperature, which would produce an unpleasantly bitter beverage, but for safety reasons too: actually, a release valve is present in the vessel to assure that no dangerous pressure can build up on filter clogging. The main feature shared by Moka and espresso, however, is not pressure, but the fact that water wets the grounds once through, increasing the extraction yield by fresh solvent power and at the same time minimising the overextraction of astringent compounds. The thermal balance of such a device is somewhat flimsy, being affected by several variables not easy to
Coffee: Recent Developments
control. In addition to the above-mentioned hydraulic resistance, these are water quantity and initial temperature, flame intensity and position below the vessel, and upper lid open or closed.
(c) Espresso Although originally an Italian speciality, the popularity of espresso has spread enormously, especially in Latin European countries, and in recent years it has been on the increase in other markets too, above all in the USA and Japan. For this reason it will be dealt with here in a more detailed way. Espresso success, besides being a phenomenon of fashion, seems to be based on the greater sensory satisfaction it gives to the consumer when compared with coffees prepared by other methods. It is not easy to obtain a quality espresso cup: its very strength ± the ability to concentrate aroma ± is also its weak point because, while enhancing the qualities, it shows up the defects of the raw material at the same time. Even the preparation of the cup, which does not forgive mistakes, gives good results only if it is carried out under the best conditions.
Definition of espresso As a product of Italian culture exported into different environments, espresso has been sometimes overadapted to local habits, or misinterpreted in several ways. For instance, many people (and many roasters too, apparently) seem to believe that espresso is a roasting degree, instead of a brewing method. It therefore appeared helpful to the author (Petracco, 1995c) to set three necessary benchmarks, deriving from the history of this beverage.
Espresso as a lifestyle: it is extemporaneous A commonly accepted meaning in Italian (as well as in other European languages) of the word `espresso' (express) or `espressamente' (expressly) carries the connotation of `made for a specific purpose, on the instant, on an explicit order'. This leads us to the concept of `extemporaneous', that is the impromptu nature of the preparation. Such terms must be understood not in the sense of improvisation, for espresso preparation is an operation that requires accuracy and care as will be seen later, but in the sense of `prepared on the instant, on express order'. This idea can be condensed effectively in the following motto: `it is not the espresso coffee that awaits the consumer, but it is the consumer who waits for his cup of coffee'.
Technology IV: Beverage Preparation
As a direct consequence the espresso beverage is not meant to be kept fresh beyond a certain limit of time: only the lapse of time necessary to serve it, add sugar and drink it. This must rule out the practice of preparing the coffee in advance in order to meet foreseeable heavy and concentrated orders and, ever more, any industrial type of preparation that involves any sort of storage. The reason for such a strict condition ± apart from the obvious pleasure that a customer feels when his order is carried out on the instant so that he can follow with his own eyes his very own cup of coffee being prepared ± is that a lag between preparation and consumption causes considerable variations that spoil the beverage's external appearance and organoleptic characteristics. The most evident degenerative change is the collapse of the foam, caused by dehydration: this results in an increase in the foam's stickiness and a decrease of its fluidity, with a consequent appearance of spots that are not covered by a foamy layer. At the end, only traces of dried foam remain sticking to the cup wall, above the free surface of a uniform liquid. The variation of the sensory characteristics that most interests the consumer is the increase in the perception of acidity with the passing of time: it increases regardless of whether the cooling of the coffee occurs concurrently. When the variety of chemical species found in solution is considered, later on, it is no surprise that some of them may continue to react mutually in the cup. One of the most predictable reactions is hydrolysis of the weak organic acids, which could provide an explanation for the modification of organoleptic sensations (Maier, 1987).
Espresso as a brewing method: it requires pressure From this first, very broad definition of espresso: `a peculiar way of enjoying a common habit, the coffee cup', historical evolution has conferred on it a more specific connotation, linked to the development of particular equipment suitable to fulfil the requirement of `extemporaneousness'. At the beginning of the twentieth century, the need to supply a coffee cup within seconds, not even minutes, from a customer's request led to higher water pressure to be resorted to as a means to speed up extraction. Water was heated in a sealed kettle up to its boiling point, and pressure was applied to it by the steam in equilibrium. This ingenious technique allowed an easy water pressure generation without any pumping device needed, but with a serious drawback: the higher the pressure required, the higher the temperature of supplied water. The fixed
147
pressure/temperature relationship at boiling point is summarised in Table 7.1 extracted from the Mollier graph (Liley et al., 1984). Table 7.1 Pressure/temperature relationship of boiling water. Absolute pressure (atm)
Temperature (8C)
1 1.2 1.5 2 3 5 10
100 104 110 120 133 152 180
Extracting coffee with pressurised boiling water, at temperatures certainly above 1008C, is a sure way to overextract, i.e. to transfer into the cup exceedingly high amounts of the less soluble substances present in roast and ground coffee, which unfortunately happen to have a bitter astringent taste and yield an unpleasant burnt-tasting beverage. A step forward was made when a separation was accomplished between the heating water contained in a kettle and the water actually used to brew coffee: it facilitated the production of water that was hot, but not boiling. Pressure can be created by a lever piston acting as a multiplier of the bartender forearm's force, or by an electric pump, much simpler to operate (this is the system most utilised nowadays). In those ways pressures as high as 10 atm (i.e. approximately 10 kilograms of force per square centimetre, which for practical purposes is equivalent to 10 bars) can be obtained, able to produce a very thick foam layer on top of the beverage. This brief survey of the outfit's evolution (see the section on the espresso mechine, later) highlights the importance of the method employed to prepare an espresso beverage. The modus operandi itself should therefore be outlined and embodied into a more detailed definition of espresso. The key role in the espresso method is played by pressure. This beverage draws a number of its characteristics from the transformation of pressure energy into kinetic energy, and from the further transformation of the latter partly into surface-type potential energy and partly into heat. Nutritive substances found in roast and ground coffee are transferred into the cup mainly due to the solvent
148
properties of water, facilitated by the high water temperature, usually not far from its boiling point. However, among the numerous chemical species found in roasted coffee (over 1800 have been identified) many are insoluble in hot water and usually do not end up in the coffee beverage. This suggests that a different type of energy input, deriving from pressure, can substantially modify the behaviour of the raw material (ground coffee) undergoing extraction. Consequently, the resulting beverage in the cup possesses many characteristics that differ from those found in coffee products that are brewed exclusively by the thermochemical energy produced by hot water. The result is most evident on examining both the external appearance and the organoleptic character of the beverage. Such an examination clearly reveals a number of peculiarities of the brewed cup that appear in other preparation systems only to a negligible degree: . The aspect displays a particularly evident trait: the liquid is crowned with an abundant layer of compact foam. A macroscopic examination of the liquid further reveals the liquid's opacity, caused by a dispersed phase made up of very small oil droplets in emulsion (Petracco, 1989). . The organoleptic characteristics, namely those perceptible to the senses of taste and olfaction, are unique. They will be described later on, in Section 7.3.2; for the meanwhile, let us limit ourselves to observing that the non-hydrosoluble substances found in an espresso cup, but absent in coffees prepared by other brewing systems, produce the particular effect on sensory perception called `body'. Anyone who has ever tasted a cup of espresso knows this very well: the mouth seems to be filled up with a rich, creamy layer of `coffee taste', whose memory lingers for a while (up to a quarter of an hour) after the beverage has been swallowed. To bestow upon the beverage these characteristics, the espresso method cannot therefore be separated from the application of pressure to a resistant coffee layer, better called a coffee cake. This need of pressure as a crucial defining element is endorsed by the Latin etymology of espresso, coming from the verb exprimo, -is, expressi, expressum, -ere, 38 (Campanini & Carboni, 1995c). A literal rendering of the word would sound like `pressed out', and foreshadows pressure as the prime mover of espresso. One additional rule, slightly less self-evident, must be observed in order to obtain a true espresso coffee. It concerns brewing equipment about a balance of
Coffee: Recent Developments
energies: the power exerted to develop pressure must be transformed into a work of forces, and consequently into kinetic energy, inside the ground coffee bed. It would be possible to extract a brew without these conditions, but under conditions of high static pressure; in this case, however, the static pressure would not perform any work on the coffee particles found inside the cake, as depressurization would probably take place downstream of the cake. The same happens when using devices capable of applying kinetic energy to a brew that is no longer in contact with its original ground coffee. Examples of such devices are stirring propellers, choke nozzles and sprayers: they give kinetic energy to the brew, which would otherwise be `smooth', with the purpose of developing a mass of foam that renders the external appearance of the beverage similar, but only apparently, to that of an authentic espresso. The operation of applying water pressure on a coffee cake and using it up within the cake to generate kinetic energy is from now on termed `percolation' in its strictest sense. The qualitative definition of espresso brewing method sounds like the following: espresso method is a beverage preparation technique based on pressure induced percolation of a limited amount of hot water through a ground coffee cake, where the energy of water pressure is spent within the cake itself. It should be noted that, every time the word `coffee' was mentioned, it has been taken for granted that espresso coffee cannot but derive from the raw material `roast and ground coffee', namely from the roasted seeds of plants belonging to the botanical genus Coffea, ground either just before brewing (by a small-scale process) or in advance (by an industrial process, with an appropriate packaging system for its storage and preservation). Nevertheless, as far as espresso method and espresso brewing equipment are concerned, it would be perfectly possible to employ them for extracting solubles from different kinds of other ground products (so-called coffee surrogates), provided that those products possess a structure allowing pressure energy to be used up within their bed. Of course, it is unlikely that such a characteristic beverage as `true-coffee espresso' could be obtained, since the complex inner structure of roast coffee beans (described in the preceding chapters) confers on the beverage a quite unmistakable identity. Furthermore, espresso is usually appreciated as a natural beverage, coming directly from nature: therefore, its raw material should never contain any sort of
Technology IV: Beverage Preparation
additives, such as flavoring chemicals or effervescencepromoting substances that stimulate the formation of a foam layer.
Espresso extraction: it must be rapid Up to now, no quantification of the espresso beverage and espresso method characteristics has been submitted. No doubt that, since `taste is a matter of taste', everyone must be free to set his own variables by fixing cup volume, beverage concentration and so on. There is, anyway, a well established tradition, born in Italy but spread nowadays all over the world, whose elements may well constitute a third-level definition useful to fix the limits of what will be named from now on `Italian espresso'. The focal point in this stricter meaning is the rapidity of percolation, 30 seconds being an optimum. The reason for this limitation comes from the complexity of the chemicals in the coffee: as already seen, a long waiting time in the cup brings about variations in the espresso's characteristics. This is even more true when the ground coffee cake undergoes a long extraction, because conditions of reaction prevail even more drastically in the cake and consequently are most likely to modify the quality of the product. Mention has already been made of the greater extraction power exerted by espresso conditions on ground coffee because of its greater, and more diversified, energy input. This yields a greater number, and concentration, of substances capable of passing into the cup. Moreover, some of these substances are not hydrophilic and therefore not completely soluble: submitted to extraction conditions, they may be found in the solid or liquid state, and their partial solubility may depend to a large extent on the thermodynamic equilibrium of solubilization reactions, but also on their own kinetics. In other words, it is possible to assume that the longer water flows through ground coffee, the greater the percentage quantity of less soluble substances is found within the dry residue determined in the beverage. The less soluble substances seem to be the least tasteful, as can be experienced by testing separately the beverage fraction that percolates from the ground coffee cake after the prescribed 30 seconds of extraction. Hence, it is possible to deduce a general pragmatic rule that all espresso bartenders know very well and apply in their operations: never exceed an extraction time of half a minute, otherwise unpleasant substances are dragged into the cup compromising the overall balance of taste and flavour reached in the cup.
149
Such a detrimental performance, also called overextraction, is one of the major causes of loss of pleasantness of an espresso coffee reported by espresso operators. It should be kept in mind that espresso brewing conditions are quite drastic, due to the high amount of energy involved. They lead to excellent results of correctly used but can spoil the expected outcome if the outfit is overdriven, and please note that exceeding the extraction time is the easiest error to be made. Underextraction, although a lesser mistake, all the same should not be understated, for percolation times shorter than 15 seconds furnish a weak and acidic unbalanced beverage.
Quantitative espresso definition Without entering into too much detail about the variables affecting espresso quality (for a more complete review see Petracco, 1989, 1995b), the overall definition given below can be useful to frame espresso coffee preparation. This definition is still very broad indeed. It focuses on the raw material and on the preparation method needed to supply a peculiar product, the espresso beverage. However, no attempt is made by the definition to indicate either what the essential characteristics of an espresso beverage are, or even what a good espresso is, for it behaves as a necessary but not sufficient condition. To fulfil such more challenging requirements, an entire book has been written about the need of excellence in every step leading to espresso: from the plant through agronomy and processing to the roasted coffee beans, and more through grinding and percolation to the beverage cup (Illy & Viani, 1995).
Quantitative espresso definition Italian espresso is a beverage prepared on request from roasted and ground coffee beans, by means of hot water pressure applied for a short time to a compact roast and ground coffee cake by a percolation machine, to obtain a small cup of a concentrated foamy elixir. The variables' ranges are: 6.5 + 1.5 g for ground coffee portion 90 + 58C for water temperature 9 + 2 BAR for inlet water pressure 30 + 5 s for percolation time
150
The espresso machine Preparing an espresso coffee requires a particular type of equipment, suited to contain a portion of ground coffee in the form of a compact cake inside a container that allows the beverage to drip out while retaining particles, and able to impart to the water a temperature close to 1008C and a pressure customarily fixed at 9 relative atmospheres. Such equipment, called an `espresso coffee machine' is available on the market in many models, ranging from compact light ones for occasional home usage to the professional computerised beverage-preparing units. The thermo-hydraulic circuit of a typical espresso machine is shown in Fig. 7.2.
Fig. 7.2 Thermo-hydraulic circuit of a typical espresso machine.
From a methodological perspective, the major leap towards modern espresso preparation was the perception that a system was needed to separate the generation of pressure from that of temperature, in order to limit the drawback of over-extraction deriving from temperatures above 1008C. After the compressed-airdriven machine patented by Francesco Illy, whose unpractical technology was probably ahead of its time (Illy, 1935), a lever system was developed: it consisted in a spring that had to be wound up by the espresso bartender by the force of his muscles, and that then released energy to the water contained in a pressurization chamber. The performance of this system ± now used only in some traditionalist coffee houses ± depends on the operator's manual ability in properly dosing out and applying muscle force and time, rendering espresso preparation some sort of an `art' with uncertain results. Electricity facilitated the automation of espresso machines, through the introduction, as a source of pressure, of a pump driven by an electric motor of a couple of hundred watts of power. Professional
Coffee: Recent Developments
pumping devices consist of a finned impeller, driven by centrifugal force, and an adjustable bypass valve that places in circulation a part of delivery water back into the suction. This power `dissipation' is legitimate, because it enables the pump head (back pressure) to be kept constant, free from the variability of the hydraulic resistance (coffee cake) through which the water flow has to pass. Home usage machines, where neither a large volume of space nor the cost of a centrifugal pump is justifiable, utilize instead a vibration pump. It is a type of volumetric pump with a small piston set in a vibrating reciprocating motion by an electromagnet with a return spring. These types of pumps exhibit a very high head (up to 30 atmospheres), which heavily depends on the back hydraulic resistance. As a consequence, the beverage in the cup is strongly influenced by the amount of ground coffee weight used, by the degree of grinding and by the amount of tamping employed. This indicates once again that controlling the machine in preparing a coffee actually resembles an art. Electric resistance heating, too, has led to great innovation: it has permitted the manufacture of direct heat exchangers. As a matter of fact, the classical heating system is based on a low pressure kettle boiler, which supplies hot water for tea and other infusions and saturated steam for cappuccino steamed milk. In addition, the kettle functions as a reservoir for heat generated by an immersion resistance or by a gas burner. The water for espresso percolation runs through a completely different circuit: the water is driven by the pump through a coil immersed in boiling water inside the boiler, where it reaches dynamically the desired temperature (many degrees lower than the temperature in the boiler). This system, particularly cumbersome and expensive, is used in all professional machines and only in the best home-use machines, where it permits a considerable production of steam suitable for cappuccino preparation. The extraction chamber consists in an upper block, onto which a filter-holder cup is inserted (like a bayonet): the ground coffee portion is poured and compacted into this filter. Since the block protrudes externally from the machine body, it is advisable to heat it by a small calibrated flow of boiler water passing through a built-in tube. In more compact models, the boiler is replaced by a direct heat exchanger. It consists in a metal block containing an immersion electrical resistance fed by a bimetallic thermostat or, in more modern models, by an electronic power supply driven by a thermocouple sensor. Cold water is forced by the pump to pass through a
Technology IV: Beverage Preparation
sinuous circuit inside the block and then directly through the coffee cake beneath. A further essential element of the equipment is the pre-treatment system of drinking water, as water ± generally with a high calcium and magnesium content ± is usually drawn directly from the city water supply system. Any system of heating such `hard' water must take into account the formation of alkaline-earth carbonate deposits on heated surfaces and the consequent harmful fall in the heat exchange efficiency (Cammenga & Zielasko, 1997). An operation most commonly performed to prevent malfunctioning in professional machines is water softening. It consists in forcing the water to pass through a bed of ion-exchange resins that capture calcium ions found in the water and replace them with sodium ions, whose salts are completely soluble. Water softeners need to be regenerated regularly by washing with sodium chloride, which is a vital operation for ensuring efficient water conditioning and a long service life of the espresso machine. Domestic machines, which cannot generally afford softening apparatus, must rely on washing cycles with mild acid solutions, to be carried out periodically as a maintenance routine. Finally, the effect of water acidity and hardness on espresso cake may affect extraction (Fond, 1995). More details about water influence on espresso will be found in a later section.
7.3 BEVERAGE CHARACTERIZATION 7.3.1 Physical and chemical characteristics (a) Physical characterization Coffee beverages used to be typically associated with such vague generic features as `dark', `steamy' and `aromatic'. Beyond that label, very little scientific, indeed, some structural analysis is needed in order to characterise them adequately. The most conspicuous physical aspect to be examined is homogeneity. The presence of heterophases dispersed within the liquid matrix of the brew is of high importance for the sensory character of the drink: as mentioned in the introduction, it is the emulsion of lipid droplets that imparts espresso its peculiar texture and after-taste, and the suspended particulate matter makes the thick mouthfeel essential to the appeal of Turkish coffee. Anyhow, these details can withstand some delay before analytical examination; in contrast there is an aspect that can undergo rapid evolution, and even disappear in a short
151
time: the presence of a separate layer of foam on top of the liquid.
Supernatant foam As for mineral waters, the first laboratory checklist item must be performed immediately at the source, which in our case is the brewing device. In fact all variables that are prone to fast modification after coming out of the `source' must be determined as soon as possible. As with spring water, the property that varies a lot after emerging at the surface is ± along with temperature of course ± the quantity of dissolved gases (O2 , CO2 , N2 ) that are suddenly released due to the lower pressure encountered. In the case of some effervescent coffee brews, such as espresso and plunger, the same applies because dissolved gases, mainly CO2 , are quickly released in the cup. They tend to form a foam layer by the help of surfactant substances present in the brew, so producing a composite beverage where two very different and easily separable constituents are present: supernatant foam and underlying liquid. Foam is by itself a biphasic system, composed of gas bubbles framed by liquid films (called lamellae) chemically constituted of a solution of surfactant in water (Petracco et al., 1999). These films tend to set in a configuration of two layers of surface-active molecules facing the exterior, with water molecules between them. The high tension of the film allows its peculiar geometry: a bubble if insulated, or a honeycomb-like structure if many bubbles are grown close together. An abundant presence of foam can be considered as a freshness indicator for the ground coffee, which has not yet released all of its CO2 derived from roasting. In some preparations, like drip coffee, the formation of excessive foam may adversely affect brewing by prolonging the filtering time: the solution to this is to carefully pre-wet the ground coffee. Coffee foam has a fairly short life: this fact compels its measurement of quantity soon after brewing. Another characteristic of foam, which is of special importance in espresso, is persistence: it should survive at least a couple of minutes before breaking and leaving a first uncovered black spot on the surface of the beverage. The factor that makes foam disappear after a period is drainage: a phenomenon that causes entrapped water to flow out of the films, leaving at the end only surfactant molecules to bear the stress of the structure. Due to their limited elasticity, the films eventually fail, releasing the entrapped gas and redissolving into the main liquid. The third property of the
152
Coffee: Recent Developments
foam is its solidity, namely its capability to bear for some instants the weight of a spoonful of sugar.
Underlying liquid The bulk properties of the liquid beverage that are commonly considered as relevant to a consumer's enjoyment are density, viscosity, surface tension and refractive index. Before these are discussed, some words must be spent to elucidate the complex, multiphase nature of the coffee beverage.
Dispersed phases The liquid part of a coffee beverage ± over which the foam floats ± is still a complex system, with the simultaneous presence of a matrix formed by a solution of salts, acids, sugars and specific molecules like caffeine, and several dispersed phases, whose properties must be taken into account along with the properties of the liquid as a bulk in order to explain some peculiar characteristics of coffee beverages. The main dispersed phases are: emulsion (formed by oil droplets), suspension (formed by solid particles) and effervescence (formed by gas bubbles). The presence of heterophases in unfiltered coffee brews may be macroscopically evident, in as much as the beverage is opaque, but to demonstrate this feature some form of microscopical investigation is needed. Different techniques have been employed to point out the discontinuities of the matrix, the most useful ones being light microscopy, scanning electron microscopy (SEM) and transmission electron microscopy (TEM). To make the best of light microscopy, some form of highlighting of the particles is needed: interference techniques are quite useful to evidence bubbles, while to examine oil droplets specific staining media for lipids (e.g. Oil Red O and osmium tetroxide) have proven to be ideal (Heathcock, 1988, personal communication). Dispersed phases differ in their stability with time: effervescence is a phenomenon that can be observed during the maximum lapse of approximately one minute, while emulsion and suspension behave like stable systems for days or longer. Therefore the problems in studying them are quite different: effervescence must be tackled as soon as possible by some form of `freezing', while emulsion is difficult to break in order to collect the oil phase for analysis. The relative amount of stable particles, namely lipid droplets and solid particles, is strongly skewed in favour of droplets of micron size in espresso, as shown in Fig. 7.3. Conversely, in Turkish coffee there is a preponderance of sub-micron-sized cell wall fragments.
Fig. 7.3 Micrograph of lipid droplets and solid particles as espresso dispersed phases.
Density Density is little influenced by the material extracted from coffee. In espresso, only differences in the second decimal digit, if compared with pure water, have been reported (Petracco, 1989). A partial justification for this could be the fact that coffee lipids dispersed as an emulsion are lighter than water, so decreasing the overall density. Higher density is of course imparted to the beverage by the addition of sugar, which can raise it up to values like 1.08. A density contribution to the perception of the sensory characteristic `body' could be foreseen by the striking body difference experienced when sipping a cup of coffee edulcorated with synthetic sweeteners like saccharin or aspartame instead of sucrose (ordinary table sugar): having a dosage much smaller than sucrose, they add sweetness with little or no density increase.
Viscosity All coffee beverages (with the possible exception of cappuccino) exhibit Newtonian behaviour, in which the viscosity is independent of the rate of shear. For espresso, viscosity is considerably higher than that displayed by pure water: the reported values (see Table 7.2) are approximately double, at usual consumption temperatures. This beverage property seems somehow influenced by the presence of dispersed phases, as has been argued by comparing a correctly prepared cup with a poorly prepared one. The viscosity increase associated with body, as high as 33% as displayed in Table 7.2, has been correlated with the amount of lipid
Technology IV: Beverage Preparation
153
Table 7.2 Viscosity of high- and low-body espresso, compared with that of water (after Petracco, 1989). Temperature (8C)
45 30
Viscosity (mPa s) Pure water
Low-body espresso
High-body espresso
0.61 0.81
1.32 1.51
1.64 2.01
droplets in emulsion: this sounds logical considering that emulsions usually display higher viscosities than pure solutions.
Surface tension Surface tension is a property of interfaces, due to the tendency of molecules to diffuse from one material to another: pure water±air interfaces present a tension of 73 6 10ÿ3 N/m at 208C. Its lower value in coffee beverages is linked to the presence of a particular class of materials, called surfactants, whose molecule is a long apolar chain ending with a polar head. They gather at the liquid±air interface ± and at other interfaces as well ± building up monomolecular layers. Their action helps to form and stabilize foam and emulsion; they may also somehow influence the fluid's behaviour during the percolation through narrow interstices of the packed coffee bed, where capillarity plays an important role. The chemical nature of natural coffee surfactants has not been clarified so far: several classes of complex molecules, like glycolipids or glycoproteins, might exhibit such behavior (Nunes et al., 1997). Apparently, in espresso beverages there are two classes of compounds responsible for two different properties: foamability seems to be influenced by a polysaccharidic fraction, while foam persistency depends more on a proteic fraction (Petracco et al., 1999). An additional interesting effect of surface tension, besides foam and emulsion promotion, is that it helps the beverage to penetrate into the tongue's taste buds, thus enhancing our sensory response.
Refractive index Refractive index measurement is a technique used mainly to evaluate solute concentrations in transparent liquids. Coffee polyphasicity makes this method less accurate, and scarcely convenient to predict the body character. In comparison with the pure water index (1.333 at 208C), filtered brews exhibit values from 1.333 to 1.338, while espresso beverages display values
ranging from 1.338 to 1.342. A question may arise as to whether refractive index is influenced mainly by the solutes or by the dispersed phases as well: measures of refractivity on a filtered aliquot of espresso exhibit no major difference.
(b) Chemical composition The obvious statement that any beverage is composed by some nutrients dispersed in water suggests the chemical nature of the extract should be described separately, i.e. the nature of the components of roasted coffee that are driven into the drink, and of the extractant. The importance and complexity of the latter ± water, of course ± should not be overlooked, even if it is apparently made up of just one chemical species in contrast with the over 1800 present in coffee material.
Coffee extract Total solids The first characteristic of coffee beverages to be dealt with is definitely overall concentration, often perceived by lay consumers as `strength'. In chemical terms, it is associated with the total solids content of the drink. Since this property is determined by drying the liquid in an oven to constant weight, not only suspended solid material is taken into account but emulsified lipids and solutes as well. Of course, the total solids concentration in the beverage is dictated by the brewing formula, namely the coffee/water ratio, but it varies greatly also depending on both roasting degree and percolation temperature: darker roasting produces more solubles, and higher temperature extracts them more efficiently (Nicoli et al., 1991). Values from as low as 10 g/l up to 60 g/l may be encountered. Table 7.3 shows typical total solids contents of beverages issued by different brewing methods, along with their ponderal extraction yields. The fraction of true solubles can be measured by filtering the liquid, resulting typically in more than 90% of total solids, if brewing methods using just a decantation separation step (like Turkish and boiled) are excepted. Some data are shown in Table 7.4. The chemical composition of total coffee solids can vary considerably, depending on the raw material variety, roasting process and brewing method. Table 7.5 may be regarded as a typical proximate analysis of the soluble part of them, where the lipids are neglected.
154
Coffee: Recent Developments
Table 7.3 Brew concentration and extraction yield by different brewing methods (after Peters, 1991). Brewing method1 Boiled Percolator Filter Napoletana Plunger Moka Espresso
Total solids in brew (g/l)
Extraction yield (%)
13.0 10.9 13.0 26.9 14.2 41.1 52.5
26.9 25.5 30.4 29.6 23.9 31.9 24.2
1 Coffee/water ratio and brewing temperature according to method, constant blend.
Table 7.4 Insoluble solid content (fines) in different brewing methods (after Peters, 1991). Brewing method Boiled Percolator Filter Napoletana Plunger Moka
Fines in brew (g/l)
Fines/total solids ratio (%)
2.04 0.22 0.07 1.71 1.06 1.10
15.7 2.0 0.5 6.3 7.5 2.7
Table 7.5 Chemical composition by mass of coffee solids (from Clinton, 1984). Class Caffeine Chlorogenic acid Reducing sugars Other carbohydrates Peptides Potassium Other minerals Acids Trigonelline
Percentage (%) 8.25 18.50 1.45 19.90 6.00 10.00 13.60 17.30 5.15
Lipids The oily substances present in raw coffee (Folstar, 1985), which seem not to be very affected by the roasting process, are only poorly drawn out into the cup by most brewing methods. A substantial amount of lipids may be found in cups prepared by unfiltered brewing methods, but just as a consequence of the
presence of non-defatted particulate material. However, the existence of an emulsified lipid fraction has been confirmed by microscopy (Heathcock, 1988), and can be quantified by liquid±liquid solvent extraction. Probably due to the sensitive application of this separation method, varying amounts have been reported in both espresso and non-espresso coffee beverages (Petracco, 1989; Peters, 1991; Sehat et al., 1993). In espresso, a figure such as 10% of the total solids may be considered as typical, accounting for something less than the 10% of total lipids present in roast and ground coffee. Paper filters are most effective at retaining oil droplets, allowing only around 10 mg/l in the brew. The unsaponifiable fraction of beverage lipids seems to be somehow lower than the one found in roast and ground coffee (Ratnayake et al., 1993), suggesting a preferential extraction of more polar compounds like triglycerides. Diterpenes, which account for some 0.6% in roasted Arabica beans, are extracted to lesser extents into the brews (Urgert, 1997).
Acids Acidity of coffee beverages is an important feature in terms of their appreciation by the consumer (Woodman, 1985). The presence of acetic, formic, malic, and lactic acids has been detected in the brew, along with quinic and chlorogenic acids (Peters, 1991; Severini et al., 1993). The content of the latter ones is of course smaller when using dark roasted coffee, in which they have largely disappeared (Clifford, 1985). Also a mineral acid ± phosphoric ± has been found in brews (Maier, 1987), possibly deriving from thermal degradation of phytic acid, the phosphoric ester of inositol. The easiest way to assess it is by pH measurement obtained using electrode-type instruments. Values ranging from 5.2 to 5.8 have been reported, manifesting a clear dependence on the roasting degree (DallaRosa et al., 1986a, 1986b) and on the extraction time (Nicoli et al., 1987). Unfortunately, sensory perceived acidity seems not always correlated with pH, perhaps due to the partial dissociation of weak organic acids and to their not fully understood interaction with the taste buds of the tongue (Maier, 1987). (See also Chapter 1 of this book.)
Carbohydrates This broad class of compounds is present in coffee seeds, as in all plant materials, both as simple sugars and as polysaccharides (Trugo, 1985). Upon roasting, heavy reactions and transformations happen, changing the balance between soluble and insoluble carbohy-
Technology IV: Beverage Preparation
drates: further detailed information and references are given in Chapter 1. While only the former are relevant to the beverage, the brewing method may influence the hydrolytic degradation of insoluble carbohydrate polymers, adding to the solubles content. Typically, only tiny amounts of monosaccharides can be found in the brew, but no sucrose (the sweet disaccharide well known as common sugar) is present, in as much as it is already transformed into bitter tasting Maillard compounds on roasting (Trugo & Macrae, 1985). Considerable research has been done on coffee carbohydrates (Leloup et al., 1997 and references therein), but few data on their content in coffee beverages are available: in espresso, a typical figure for soluble carbohydrates is 8 g/l, corresponding to some 15% of total solids (Petracco, 1989). Carbohydrates are responsible for most of the caloric content of a coffee cup, of the order of magnitude of a couple of kcal/cup. From a nutritional standpoint, this is negligible if compared to the calories supplied by the sucrose that is commonly added to the beverage (one spoon containing about 5 g corresponds to 20 kcal) (Macrae, 1988).
Nitrogenous compounds Nitrogen-containing compounds are present in coffee beverages in the form of transformed proteic material, grouped under the broad name of melanoidins, and caffeine, while trigonelline has largely disappeared during roasting to yield volatile compounds adding to aroma (Macrae, 1985). The way to analyse this complex class of compounds is by determining total elemental nitrogen, subtracting the caffeine nitrogen and multiplying by the standard number 6.25 to supply a conventional proteic content value. Typical values are around 1.4 mg/ml for total N, corresponding to approximately 4 mg/ml of conventional proteins. Caffeine is the most studied coffee component, due to its physiological activity (Dews, 1984; Gilbert, 1992; Garattini, 1993; Debry 1994; Spiller, 1998). Caffeine's solubility in water displays a marked dependence on temperature, with a more than 30-fold increase from standard conditions to atmospheric boiling temperature, where almost quantitative dissolution takes place provided enough time is allowed to cope with the kinetics (Spiro & Page, 1984; Spiro & Selwood, 1984; Spiro & Hunter, 1985; Spiro et al., 1989; Spiro, 1993). Also water hardness seems to influence caffeine extraction to a certain extent (Cammenga & Eligehausen, 1993). Different brewing methods somehow influence the
155
caffeine content of the beverage. Table 7.6 suggests that techniques where no fresh water is added, tend to extract less. The caffeine present in espresso coffee has not been investigated by many authors. Some data suggest that caffeine extraction from coffee grounds by espresso percolation is not complete, with a yield usually within the range 75 to 85%. This is due to the short time allowed by the espresso method to extract caffeine out from the cellular structure, as demonstrated by fractionated extraction curves (Fig. 7.4). Therefore caffeine concentrations in the beverage vary from 1.2 mg/ml up to 4 mg/ml, depending on cup size and blend composition. Corresponding cup contents range from 60 mg of caffeine (for pure arabica blends) up to double: 120 mg (for pure robusta blends). Table 7.6 Caffeine concentration and extraction yield by different brewing methods (Peters, 1991). Brewing method Filter Moka Napoletana Plunger Percolator Boiled
Caffeine in brew (g/l)
Caffeine yield (%)
0.67 2.36 1.35 0.69 0.58 0.57
100 92 98 81 95 89
Fig. 7.4 Cumulative caffeine content in time fractions of an espresso coffee cup (after Petracco, 1989).
Minerals The presence of minerals in coffee brews is low, as can be inferred from ash contents that never exceed 7 g/l, but it is not negligible when compared to salt contents of around 2 g/l for the hardest mineral waters.
156
Nevertheless, minerals constitute about 15% of total solids and, since potassium forms a substantial part of them, they might be considered as a beneficial contribution to human diet. It is also possible that the potassium present in the brew interacts with other taste components by modulating the type or intensity of their taste stimulation. More research is needed to understand fully the effect of potassium on the brew.
Water The second most important ingredient of the cup, apart from coffee itself, is water, constituting more than 95% of any coffee beverage. However self-evident this statement may seem, the role of water in coffee preparation must be taken into due consideration, otherwise it would be senseless to devote the utmost care and attention to only one of the two ingredients, while neglecting the other. Yet, this regularly occurs, partly because the correct choice of the water is left to the end user and partly because the end user ± bartender, waiter, housewife ± has an extremely limited influence over the choice of the water. In fact, very few people brew coffee using mineral water, whose characteristics are stated on the label and are supposed to be constant. Most consumers rely instead on the quality of tap water, and only sometimes treat tap water with simple purification devices. It must be stressed that the quality of water coming from public waterworks is perfectly suitable for human consumption. This makes further purification or disinfecting treatments unnecessary, as they are already adequately undertaken by the public water supply board. However, treatment of drinking water may turn out to be necessary in two critical cases: one is linked to sensory perception and the other concerns the use of the brewing equipment.
Chlorination In the first case, the beverage must be free from any unpleasant foreign flavours left over in the water by generalised disinfecting treatments, such as chlorination. Salts that release free chlorine (hypochlorites) lend the beverage their peculiar taste, which is perceptible when the initial concentration of Cl2 exceeds 0.5 mg/l. A system most commonly used for removing chlorine taste involves passing the liquid through a bed of granules of activated carbon, which has the ability to adsorb many odor-bearing substances. A drawback in the use of activated carbon is that it reaches saturation without showing any apparent signs: consequently, it is
Coffee: Recent Developments
necessary to regenerate, or to replace, the bed at previously established periods.
Hardness The second critical case is related to water hardness, namely to the water's calcium and magnesium content. On heating, these cations produce insoluble salts (mainly carbonates, but also sulphates and silicates). These salts tend to precipitate in the form of compact plaques, especially on the heated surfaces. The plaques, by forming a coating on the surfaces, detrimentally affect the heat exchange coefficient. This leads to unpleasant white sediment haloes on the heated parts of pots and similar, and to quite serious technical inconveniences in the espresso machines, in particular a reduction in heat transfer effectiveness and a possible consequent failure in the electric resistances due to overheating. Potable water may be nearly always defined as hard, that is having a hardness that exceeds 15 French degrees, which corresponds to a concentration of 150 mg/l of CaCO3 . In some waterworks in Italy, water hardness may reach a peak of 37 French degrees. This is not a reason for concern because the hardness for drinking water recommended by Italian law (DPR 236, 24/5/88) ranges from 15 to 50 French degrees; it is noteworthy to remember that even many mineral waters exceed 100 French degrees. Only very few waterworks supply very soft water at the lowest recorded hardness of 2 French degrees. The direct use of public waterworks' drinking water in both professional and home espresso machines produces intolerable deposits in a short period. Home-use machine manufacturers advise removing these deposits by periodic washing with weak acid solutions (citric or formic acid, even vinegar, will do). In professional machines, deposits are prevented from forming by fitting the machine upstream with purification systems that act on calcium and magnesium ions, and on other heavy metal ions nearly as well. There are two types of purification systems: softeners or demineralizers.
Softeners Softeners consist of a bed of ion exchange resins, which hold back some of the calcium ions and replace them with sodium ions (which do not affect hardness, nor form deposits) until a tolerable hardness of about 8 French degrees is reached, as recommended by espresso machine manufacturers. The resins must be periodically regenerated by washing with NaCl (common table salt) so as to restore their effectiveness. Softeners entail
Technology IV: Beverage Preparation
two problems. Functionally, localized channels form in the bed over some time, thus diminishing the effect of ion exchange (this can be avoided by stirring the resins from time to time). Hygienically, the danger of a possible proliferation of micro-organisms in the bed is averted by using beds that contain a bacteriostatic additive. In case of treating very hard water, the quantity of sodium added may be substantial: in addition to being perceptible to taste, Na makes the water alkaline with a further effect on extraction.
Demineralizers
Demineralizers instead eliminate Ca ions without introducing any foreign metallic ions. One type of demineralizer employs separate beds of resins, fitted in series, or one single bed in which resins have been mixed but can be separated by floating for regeneration. One resin is of the so-called cationic type, which captures cations, Ca and Mg included, replacing them with H ; the other resin is of the anionic type and captures the negative ions (Clÿ , HCO3 ÿ , SO4 ÿÿ and others) by replacing them with hydroxyls, which neutralise hydrogen ions and form H2 O. These types of resins must be regenerated separately with HCl and with NaOH, respectively. A more modern type of demineralizer is based on the principle of reverse osmosis. Under this principle, pressure energy is employed to filter water through a semi-permeable membrane capable of keeping back ions and of letting the smaller H2 O molecules through. The system eliminates the excluded salts by discharging an aliquot of the concentrated solution that has been so produced. With the use of a suitable membrane or of a graded bypass, water with a desired degree of hardness is obtained.
Influence on beverage The influence of hardness on espresso beverage quality mainly consists in an increase in percolation time by 15% or more when very soft water (below 2 French degrees) is used; only slight differences occur when the water hardness is raised above 8 French degrees up to typical waterworks values. Hence, it is not recommended to soften or demineralize water below the former value, otherwise a coarser grind should be used to compensate the increase in percolation time. There is no precise knowledge on the influence of the pH of water, a measure of the water's acidity or alkalinity. It is one more factor affecting percolation time, but to a less predictable extent and probably in association with hardness. One possible explanation of the influence of
157
hardness and pH on percolation resorts to the effect of calcium ions and of acidity on the foam-producing and emulsifying properties of natural surfactants found in roasted coffee. It is not clear, however, whether the increase in percolation time is to be put down to a greater foam and emulsion viscosity in relation to pure water, or to the interstitial precipitation of calcareous soaps or other substances sensitive to pH.
7.3.2 Organoleptic characteristics Organoleptic means `learnt through perceptions', that is by direct information coming from our organs of sense: eyes, ears, skin, nose or mouth. Its implementation, called sensory analysis, is used to describe and evaluate subjective characteristics such as taste, flavor, odor, color, appearance, and related factors of foods and beverages (Heath, 1988). A cup of coffee is a powerful stimulator of our senses: the rich color, the intense aroma, the strong and long lasting taste give pleasure to the eyes, the nose, the tongue. Only the sense of hearing seems to be extraneous to this multi-sensory experience (if just disregarding the call of a puffing Moka machine, ready to be emptied out). It is remarkable that such a deep pleasure can be transmitted by just a cup of hot liquid, with almost no calories attached. In this cup, the meeting occurs of the amazing complexity of coffee chemical composition with the no lesser complexity of the sensory perception system.
(a) Vision As far as the vision is concerned, the main attribute of the cup is the aspect of foam, particularly important in Turkish and in espresso brews. What confers the beautiful color to the foam are tiny gas bubbles surrounded by viscous liquid, where some infinitesimal particles of cell walls float, producing a sort of `tiger skin' effect. Consumers used to ascribe great importance to the color and the texture of foam, and they are right in as much as a perfect foam is the signature of a perfect preparation. Any error in grinding or in the percolation phase, as temperature or level of extraction mistakes, is immediately denounced by the color, the texture and the persistence of the foam. Foam also acts like an aroma-sealing lid, because the volatile compounds responsible for the odor of the espresso beverage have a high vapor pressure and would easily escape from the hot liquid: the foam is trapping them for our pleasure, if the coffee used to brew the cup is fine, or for our torment if the coffee is poor or
158
defective. The foam acts like an amplifier, enhancing virtues and vices.
(b) Degustation In order to systematise the topic of organoleptic attributes, it is rational to demarcate the boundaries between taste and after-taste, odor and flavor, body and astringency: all of which are typical manifestations of human ability to judge aliments.
Taste Taste is perceived by the homonymous sense, whose primary elements, called buds, and numbering about 500, are located on the surface of the tongue. The existence of specialized buds able to perceive mainly one of four major gustation sensations ± acidic, bitter, sweet and salty ± has been assumed for a long time. It has been postulated that sweet-sensitive buds are concentrated on the tip of the tongue, bitter buds are to be found on its back part, while salty and acidic sensations come mostly from the tongue's sides. This theory has been recently challenged (Bartoshuk, 1993). Acidity, a fundamental property of filter coffee, is not much appreciated in espresso because its high concentration amplifies this character giving an unbalanced feeling. Fine washed coffees, having frequently a strong acidic tone, are appreciated for the preparation of filter coffee since they will withstand hydrolysis and keep their quality during the time elapsing between preparation and consumption (Feria-Morales, 1989). In an espresso blend they will help with the aroma, but a blend of pure washed coffees will taste excessively acidic and lack body. Sun-dried coffees will, in contrast enhance both body and balance. The roasting intensity can also be used to change the bitterness vs acidity ratio; nevertheless washed coffees roasted too dark will develop a pungent, burnt character usually judged as unpleasant. Another objectionable acidic taste is the consequence of very fast roasting: it resembles a metallic tone, probably due to the presence of a residue of chlorogenic acid that has not been involved, for lack of time, in the reactions leading to the formation of flavour components. Sweetness is a character that everybody appreciates, and is positively correlated with value. The contrary holds for bitterness, which is negatively correlated with price. A fine espresso should, however, taste bittersweet with an initial slightly acidic note, should display strong body and intense aroma and should be pleasantly persistent.
Coffee: Recent Developments
Consumers in different countries demonstrate marked differences in the appreciation of the ratio of bitterness to acidity. Southern populations usually prefer a cup of coffee with a bitter dominance and give great value to the body, while northern people prefer a balanced taste and consider an excess of bitterness a defect. This is especially true if related to the concentration of the espresso cup ± very small and concentrated in the south, where it is preferred without milk; more diluted and with the frequent addition of milk or dairy cream in the north. Unfortunately, the preparation of an espresso cup of more than 50 ml without overextracting is very difficult: this is the reason why northern European consumers frequently receive a cup in the dilution they require, but tasting bitter and woody and sometimes astringent, because the large volume drew out a lot of less soluble and less pleasant components. Espresso may be seen as a fractional extraction, and not all that is soluble is agreeable: one must know when to stop, if one likes quality.
After-taste If compared to other coffee brews, the rheology of espresso is very peculiar. Its strong concentration is responsible for its high viscosity, whereas its content of natural surfactants lowers surface tension. These apparently contrasting characters promote the intense taste, but also the long-lasting after-taste (and afterflavor): those sensations that are felt some time after having swallowed and emptied the mouth. Espresso soaks the fine texture of the surface of our tongue, colouring it. The liquid is then trapped in the taste buds, where emulsified oils will strongly adhere to some receptors, slowly releasing the dissolved volatile compounds, allowing them to be perceived for a while (up to 15 minutes) after the beverage has been swallowed. The fact just described may also help to reduce the perception of bitterness. An anecdote relates the following experiment: an expert cuptesters' panel was presented with a quinine solution, whose bitterness was rated 100. Then 1% of a large polysaccharide molecule having a colloidal behaviour, carboxymethylcellulose, was added to the bitter solution and, to everybody's surprise, the bitterness was rated 40. The conclusion was that colloids are able to reduce the perception of bitterness by receptor blockade. The definition of colloid is a complex one, but ± for the purpose of this book ± it is enough to think of dispersed solid, or liquid, particles with dimensions less than 10 mm. The
Technology IV: Beverage Preparation
droplets of the emulsified oil present in an espresso with high body have a diameter of less than that size, as mentioned earlier. This theory explains a further phenomenon as well: the fact that by diluting a concentrated espresso it is perceived as bitterer. This seems strange, for dilution usually decreases the intensity of any taste. But it is enough to suggest that the molecules responsible for the bitter taste are present in the liquid in the many billions, and the oil droplets at several orders of magnitude less: the probability of an oil droplet meeting and inhibiting a bitter receptor in a taste bud is adequate only if the concentration is strong. Diluting the beverage increases the probability of a bitter molecule interacting with a receptor, therefore making diluted espresso taste more bitter.
(c) Olfaction The olfactory sense perceives the presence of odorous volatile molecules ± the ensemble of the ones given off by coffee is, in this context, called aroma ± by means of thousands of receptors located in the inner membranes of the nose. To stimulate them, molecules must be conveyed there by air flow, either during inhalation or while breathing out. The former of the two sensations originated this way is known as odor, the latter . . . oops! In the author's mother tongue (Italian), as in most of other European languages, a specific word for the latter notion is sadly missing. Thanks to the official language of science (English), a brilliant candidate may be found in the word `flavor', which is the super-national definition proposed here ± in a slightly more circumscribed meaning, if compared with past terminology (Heath, 1988) ± for the sensation originated this way. The personal conception of the author, apologising to the International Standard Organisation for the limited agreement with their norm ISO/TC34±5492/1, is that both odor and flavor pertain to the sphere of olfaction, the sense of smell.
Odor Odor is the definition of olfactory sensation caused by inhaling. Whereas the aroma of freshly ground coffee is a powerful stimulus, coffee beverages do not release a lot of volatile compounds in the environment that surrounds the cup, and this holds for espresso too, as already explained in discussion of the foam's role in trapping aroma. In contrast, a lot of volatile molecules of aroma are released within the mouth after drinking and reach the
159
olfactory receptors in the nose by retro-diffusion, namely the movement of gaseous molecules from the mouth through the pharynx up to the nose. A simple experiment that can prove the existence of this phenomenon is just munching and swallowing a bite of apple while pinching one's nostrils. The missing sensation is exactly what should be called `flavor'.
Flavour This is one of the most controversial organoleptic characters of coffee (Wrigley, 1988). Both intensity and quality of this `olfactory component of taste' are important, especially where espresso is concerned. More than 800 volatile species have been detected so far in coffee headspace, and many of them are liposoluble. Since the espresso method is able to extract a substantial quantity of coffee oil by dispersing it in small droplets, it probably also transports into the cup a lot of the oilsoluble volatile molecules imprisoned within the lipidic phase of roast coffee during roasting and storage. After brewing, these gaseous molecules tend to escape the liquid and eventually reach the nasal receptors. This complex mechanics accounts for the qualitative difference between the flavours of espresso and filter coffee: besides the evident difference of concentrations, there is a distinct aroma spectrum as well. Flavor descriptive terms are usually borrowed from common life experience, and recall the world of flowers, fruits and various fresh or baked foods. An unlucky case is the need to describe negative flavours, or taints, where frequent and specific terms are `stinker': a sort of rotten vegetable tang, `mouldy': reminiscent of mildew-infected fruits and `peasy': reminiscent of pea pods. A consumer-oriented model list has been developed by the International Coffee Organization (ICO 1991), but every expert cuptesters' panel usually sticks to its own terminology for more complete classification.
(d) Mouthfeel Mouthfeel is a tactile sensation perceived by buccal mucous membranes, along with the thermal response due to the beverage's temperature. Most of its nature is related to small movements of the tongue against palate and gums, which apply a shear stress to the liquid, performing a sort of rheological measurement of viscosity and texture. Body, an attribute mainly applicable to espresso coffee, is felt this way. Conversely, astringency is related to a chemical phenomenon: precipitation of saliva's proteins, which is due to
160
specific phenolic compounds present in some beverages.
Body It is sometimes claimed that body is a peculiarity of robusta espresso brews, and, as a matter of fact, those beverages `fill the mouth'. It is difficult to believe that oil droplets of colloidal size are present in higher amounts in robusta than in arabica espresso, given that the lipid content of roast and ground robusta is much lower than arabica's (some 9% as opposed to 13%). Some cuptesting experiments indicated that freshly percolated robusta espresso has actually some more body than arabica's, but only as a first sensation: keeping the sip in one's mouth, after some seconds arabica's body is still present while robusta's decreases. What's more, an inversion of assessment takes place when the beverage is allowed to rest for 2 minutes or so, with arabica again displaying stronger body. It has been postulated that the reason for this strange behaviour could lie in rapidly disappearing gas bubbles of colloidal size ± possibly more evident in robusta ± behaving as a viscosity enhancer (Petracco, 1989).
Astringency Another tactile sensation, astringency, is always considered as negative, since an astringent coffee beverage reminds one of medicine and is a very unpleasant experience. This defect has been related to the presence of immature beans containing dicaffeoylquinic (dichlorogenic) acids, which are astringent to mucous membranes via precipitation of soluble proteins from the saliva (Ohiokpehai et al., 1982). Deplorably immature beans can be found more and more frequently, because green coffee producers seem to pay less attention to the quality of their crops, due to the low price of recent periods.
7.4 MODIFIED COFFEE BEVERAGES 7.4.1 Coffee–milk admixtures Disregarding the old adage that deems milk just to be poured into poor tasting cups of coffee, in order to correct or mask their objectionable flavor (milk fat emulsion does modify the aroma balance), a chapter on coffee beverages cannot ignore the enormous number of coffee-with-milk cups that are consumed every day. The nutritional importance of milk in everybody's diet cannot be overstated: a complete feed for newborns
Coffee: Recent Developments
by itself, it helps adults to keep in good health mainly via its calcium content. Therefore, blending coffee drinks with milk may be seen as a good way to increase milk consumption and foster calcium intake, to the benefit of bone strength. Conversely, it must be remembered that the caloric content of milk is rather high (60 to 80 kcal/100 ml for whole milk, around 40 for skimmed), and may contribute seriously to the total energy intake, exceeding by far the caloric effect of coffee when they are mixed together. A plethora of recipes are present in various cultures, ranging from a drop of coffee in a glass of milk, just to add some flavour, to a drop of milk to discolour a coffee cup. Milk adds to appearance, to texture and to aftertaste persistence because of its fat content, which contributes mouthfeel and flavor too, and helps to distribute fat-soluble flavor molecules of coffee, principally when used in espresso drinks. It may be put in cold or hot, and some people even question the priority: should milk be added to coffee, or coffee to milk? Perhaps the most attractive marriage between the two products happens when milk is added in the form of foam, originating the endless family of cappuccinolike drinks. The chemistry of milk foam formation is at least as complex as the coffee one, ensuing from casein± lipid interactions mediated by phospholipids (Goff & Hill, 1993). Both the protein and fat contents in milk are critical to foam development: skimmed milk contains the greatest percentage of protein, and foams better than low-fat or whole milk, while the fat content helps to keep the foam stable. The freshest milk ± straight from the cow ± is seldom used, for hygienic reasons. Pasteurized milk is favoured because it can be kept up to 14 days in refrigerated storage with no noticeable change in foaming properties. It can be produced either by the conventional process (heated to 648C for 30 minutes) or by the HTST process (728C for 15 seconds) (Hinrichs & Kessler, 1995). Aseptically-processed milk (the so-called UHT, heated up to 1448C for 4 seconds) can be kept sealed at room temperature for several months. From a practical standpoint, it has many advantages: less refrigeration is needed in a business that usually exists in limited space and less time and lower temperature are needed for heating an already room temperature product. However, this product is not widely used by bartenders. The rule of thumb to making a good cappuccino is `thirds': first, make a standard espresso shot in a larger cup, where espresso should take about one-third of its volume, then add a third of hot liquid milk and a third
Technology IV: Beverage Preparation
of steamed, frothed milk. Any variation on the above prescription is permitted, producing beverages that have been baptized with fancy, often exotic-sounding names like cafe au lait, cafe creÂme, macchiato, latte, frappuccino and mochaccino, expanding in this way the coffee lingo tremendously. Several pitfalls are present in cappuccino preparation: off-flavor water can taint the steamed milk; damaged or unclean steaming tools can add a burnt or scorched flavor, lengthy steaming can produce milk temperatures above 708C, where it gets a cooked flavor. Typically, more than 75% fresh (unsteamed) milk should be present when producing the foam. Failure to drain steam lines before frothing can add a lot of water to milk and results in loss of mouthfeel and creaminess, as well as in decreased foaming capacity. A final crucial warning: unlike roast and ground coffee, milk is a `perfect' food, not just for people but for micro-organisms too, so when handled incorrectly it is subject to spoilage. Keeping it refrigerated as long as possible and maintaining all relevant tubing and vessels clean is a must.
7.4.2 Canned coffee beverages The pleasure of a freshly brewed cup of coffee is such that is difficult to imagine anybody refusing it, in favour of an analogous, but preserved, beverage. There are, nevertheless, situations in which it is not possible or convenient to have on hand, or properly operational, the right brewing equipment. The worst solution to this should be resorting to pre-brewed coffee beverages kept waiting, either warmed up again or ± even worse ± kept simmering for hours (not even microwave reheating is a good idea). Apart from taste deterioration due to the chemical changes caused by hydrolysis (Feria-Morales, 1989; Nicoli et al., 1989, 1991; DallaRosa et al., 1990), a substantial microbiological spoiling risk exists in allowing such a nutritious `broth', as coffee beverages are, to tarry in an unprotected environment. Both moulds and bacteria coming from the atmospheric dust can easily colonise such a carbohydraterich solution of high water activity. Several classic preserving techniques (sterilization, pasteurization, etc.), developed for use in various popular drinks such as milk and juices, have been borrowed by the coffee industry and adapted. With the help of stabilizing techniques to minimize sedimentation (Severini et al., 1995), they have facilitated the production of a range of coffee drinks, which may be canned, bottled or packed in paper containers.
161
These drinks, not so popular in western countries for reasons linked to the traditional image of coffee as a freshly brewed beverage, have encountered an enormous success in Japan. The overall influence of western food in that country also brought the practice of drinking coffee, which is now such a common beverage it makes Japan the fourth largest (per capita) coffee drinking nation. The creativity of a local industrialist, Tadao Ueshima, led in the early 1970s to the development of a process to stabilize and bottle a sweetened mix of coffee and milk, which quickly became the most popular relevant drink. Most of Japanese canned coffee is sold (either warmed at 608C or chilled) from vending machines, 2 million of which have been installed ± mainly outdoors ± in that country. Approximately 2.5 billion litres per year are consumed locally, making some 100 cans per year per capita (Nakanishi, 1998). The prolonged heat applied to cans sold as warm causes some taste deterioration, attributed to degradation products of caffeic acid (Yamada et al., 1997). Coffee drinks in containers are categorized in Japan into three varieties depending on the percentage of coffee content: `coffee', `coffee drink' and `soft drink that contains coffee'. According to a fair competition agreement, the term `coffee' refers to a drink with more than 5 g of coffee, extracted or concentrated from preroasted green coffee beans, while `coffee drinks' are defined as drinks that use coffee beans as their ingredients, and `soft drinks that contain coffee' are based on coffee drinks with added sugar, dairy products, emulsified oil and fat, and other edible items (JETRO, 1996).
7.4.3 Flavored coffee beverages `Whoever invented ``flavored coffee'' should be shot. Several times.' This extremist vox populi echoes a common refusal by coffee purists to adulterate what should be a freshly roasted agricultural product. But the firing squad should travel quite far back in time and in space to reach the very cradle of Coffea arabica, Ethiopia, where coffee drinkers used to add spices to their brew several centuries ago. This habit still persists there, and also in the neighbouring Arabian countries, where flavoring a cup of coffee with pungent cardamom seeds is routine. Even where the artificial-flavor-reinforced coffees were born, the USA, in the beginning they were created to tempt non-coffee drinkers to try something new instead of sodas. Nowadays, a considerable part of coffee sales in the USA ± up to 15% of the total ± is
162
flavored with one of the top ten fragrances, namely Vanilla Nut, Irish Cream, Chocolate, Vanilla, Macadamia Nut, Amaretto, Chocolate Almond, Coconut, Cinnamon and Chocolate Raspberry Creme: this momentous marketing phenomenon must therefore be covered here with some general notes. Essentially, there are two types of flavor: artificial and natural. To be considered natural, flavors must be made from raw materials that are processed physically or by fermentation. The coffee cherry itself undergoes a certain amount of fermentation, caused by indigenous yeasts and bacteria, during the post-harvest process. Some of the intrinsic coffee flavor is produced that way, and unfortunately taints and off-flavors too happen to be produced by fermentation allowed to run in inappropriate conditions. A vast range of vegetal material might be used to change coffee taste: an infamous example is the lemon peel used to abate some bitterness in coffees that have been roasted too dark, in a stray search of misunderstood espresso character. Within the artificial category, there are two types: nature-identical and entirely artificial. Nature-identical flavouring means that the flavors have been duplicated by a chemical process from natural flavors so that the end product has exactly the same molecular structure as the natural flavor. Nevertheless, nature-identical flavors must, by USA law, be labelled as `artificial'. While flavor formulas are extremely complex ± up to 80 ingredients may be used for one additive ± the process itself of flavoring coffee is not a complicated one. The key operation is a thorough mixing of the liquid additive with the whole roasted beans, or of the powdered flavoring agent with ground coffee. The latter agent is often formulated by binding a flavour combination to a carbohydrate carrier. Coffee varieties that will round out the particular flavor are chosen: mild arabicas are strongly preferred. The roasting may be slightly darker than customary, to deepen the coffee flavor. The flavoring liquid, or powder, is added once the beans have cooled to approximately room temperature. The average commercial roaster formulates its flavors to be added at 2% dosage by weight; only exceptionally flavored coffees are the result of dosages up to 4%. Flavored coffee can be brewed by any traditional method, provided it does not apply extreme temperature for an extended time, so evaporating most of the added volatile content. Of course, variables like dilution, presence of emulsion and admixture with sugar or milk are likely to change the sensory response to the flavoring agent, sometimes resulting in an unexpected outcome.
Coffee: Recent Developments
New creative ways to flavor a coffee beverage in the cup are under development, as for instance flavorcoated stirrers that release their active component while agitating the beverage, or biscuits that melt partially in the cup, adding their specific scent. As a conclusion: while `everything goes' nowadays in food commerce (after all, we flavor virtually everything else we eat or drink, don't we?), it is debatable whether one of the most powerful aroma reactors and containers, the coffee bean, needs any artificial fortification.
REFERENCES Barbanti, D. & Nicoli, M.C. (1996) Estrazione e stabilitaÁ della bevanda caffeÁ: aspetti chimici e tecnologici. Tecnol. Aliment., 1/ 96, p. 62. Bartoshuk, L.M. (1993) The biological basis of food perception and acceptance. Food Qual. Pref., 4, 21. Bentz, M. (1908) Kaffeefilter mit nach unten gewoÈlbten, mit einem Abflubloch versehenen Boden. German patent No. 347896/13 July 1908. Burton, R.F. (1860) The Lake Regions of Central Africa (reprinted 1961). Horizon Press. Cammenga, H.K., Eggers, R., Hinz, T., Steer, A. & Waldmann, C. (1997) Extraction in coffee-processing and brewing. In: Proceedings of the 17th ASIC Colloquium (Nairobi) pp. 216± 219. ASIC, Paris, France. Cammenga, H.K. & Eligehausen, S. (1993) Solubilities of caffeine, theophylline and theobromine in water, and the density of caffeine solutions. In: Proceedings of the 15th ASIC Colloquium. (Montpellier), p. 734. ASIC, Paris, France. Cammenga, H.K. & Zielasko, B. (1997) Kinetics and development of boiler scale formation in commercial coffee brewing machines. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 284±289. ASIC, Paris, France. Campanini, G. & Carboni, G. (1995a) Vocabolario Latino ± Italiano, p. 339. Paravia Torino, Italy. Campanini, G. & Carboni, G. (1995b) Vocabolario Latino± Italiano, p. 767. Paravia Torino, Italy. Campanini, G. & Carboni, G. (1995c) Vocabolario Latino± Italiano, p. 524. Paravia Torino, Italy. Clarke, R.J. (1986) The flavour of coffee. In: Food Flavours: 3B The Flavour of Beverages (eds A.J. MacLeod & I.D. Morton), pp. 1±48. Elsevier, Amsterdam. Clarke, R.J. (1987a) Extraction. In: Coffee Vol. 2 Technology (eds R.J. Clarke & R. Macrae), pp. 109±146. Elsevier Applied Science, London. Clarke, R.J. (1987b) Roasting and grinding. In: Coffee Vol. 2 Technology (eds R.J. Clarke & R. Macrae), pp. 73±108. Elsevier Applied Science, London. Clifford, M.N. (1985) Chlorogenic acids. In: Coffee Vol. 1 Chemistry, (eds R.J. Clarke & R. Macrae), pp. 153±202. Elsevier Applied Science, London.
Technology IV: Beverage Preparation
Clinton, W.P. (1984) The chemistry of coffee. In: Banbury Report 17. Coffee and Health, (eds B. Macmahon & T. Sugimura), p. 3. Cold Spring Harbour Laboratory. DallaRosa, M., Barbanti, D. & Lerici, C.R. (1990) Changes in coffee brews in relation to storage temperature. J. Sci. Food Agric., 50, 227±35. DallaRosa, M., Barbanti, D. & Nicoli, M.C. (1986a) Produzione di caffeÁ tostato ad alta resa. Nota 2. QualitaÁ della bevanda di estrazione. Ind. Aliment., 7±8, p. 537. DallaRosa, M., Nicoli, M.C. & Lerici, R.C. (1986b) Caratteristiche qualitative del caffeÁ espresso in relazione alle modalitaÁ di preparazione. Ind. Aliment., 9, p. 629±33. Debry, G. (1994) Coffee and Health. John Libbey Eurotext, Paris, France. Dews, P.B. (1984) Caffeine Perspective from Recent Research, Springer-Verlag, Berlin. Ehrenkranz, F. & Inman, L. (1948) Equipment in the Home: Appliances, Wiring and Lighting, Kitchen Planning, p. 83. Harper and Brothers Publishing. Ellis, J. (1998) The History of Coffee. Cafepress (www.cafepress.pair.com). Feria-Morales, A. (1989) Effect of holding-time on sensory quality of brewed coffee. Food Qual. Pref., 1, 87. Folstar, P. (1985) Lipids. In: Coffee Vol. 1 Chemistry, (eds R.J. Clarke & R. Macrae), pp. 203±22. Elsevier Applied Science, London. Fond, O. (1995) Effect of water and coffee acidity on extraction. Dynamics of coffee bed compaction in espresso type extraction. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 413± 42. ASIC, Paris, France. Garattini, S. (ed.) (1993) Caffeine, Coffee and Health, Raven Press. Gilbert, R. J. (1992) Caffeine: the most popular stimulant. In: The Encyclopedia of Psychoactive Drugs. Series 1, (ed. S.H. Snyder). Chelsea House Publications. Goff, H.D. & Hill, A.R. (1993) Dairy chemistry and physics. In: Dairy Science and Technology Handbook, Vol. 1, Principles and Properties, (ed. Y.H. Hui), Chapter 1. VCH Publishers Heath, B. (1988) The physiology of flavours. In: Coffee Vol. 3 Physiology, (eds R.J. Clarke & R. Macrae), pp. 141±70. Elsevier Applied Science, London. Heathcock, J. (1988) Espresso Microscopy and Image Analysis, private communication to the author. Hinrichs, J. & Kessler, A.G. (1995) Thermal processing of milk ± processes and equipment: In: Heat-induced Changes in Milk, (ed. P.F. Fox), 2nd edn. International Dairy Federation, Brussels. ICO (1991) Sensory Evaluation of Coffee, p. 74. International Coffee Organization, London. Illy, F. (1935) Apparecchio per la rapida ed automatica preparazione dell’infusione di caffeÁ. . . . Italian patent No. 333293/26 December 1935. Illy, A. & Viani, R. (eds) (1995) Espresso Coffee: The Chemistry of Quality, Academic Press, London. JETRO (1996) Market Survey on Coffee in Japan. Japan External Trade Organization Leloup, V., DeMichieli, J.H. & Liardon, R. (1997) Characterisation of oligosaccharides in coffee extracts. In: Proceedings of
163
the 17th ASIC Colloquium (Nairobi), pp. 120±27. ASIC, Paris, France. Liley, P.E., Reid, R.C. & Buck, E. (1984) Physical and chemical data. In: Perry’s Chemical Engineers’ Handbook, (eds R.H. Perry & D.W. Green), pp. 3±46. McGraw-Hill, New York. Macrae, R. (1985) Nitrogenous components. In: Coffee Vol. 1 Chemistry, (eds R.J. Clarke & R. Macrae), pp. 115±52. Elsevier Applied Science, London. Macrae, R. (1988) Nutritional factors. In: Coffee Vol. 3 Physiology, (eds R.J. Clarke & R. Macrae), pp. 125±40. Elsevier Applied Science, London. Maier, H.G. (1987) The acids of coffee. In: Proceedings of the 12th ASIC Colloquium (Montreux), pp. 229±37. ASIC, Paris, France. Miller, S.A. and other authors (1984) Liquid±solid systems. In: Perry’s Chemical Engineers’ Handbook Sixth Edition, (eds. R.H. Perry & D.W. Green), pp. 19±48. McGraw-Hill. Nakanishi, N. (1998) Canned coffee is Japan's most popular drink. Indian Express Newspaper, 16 April. Nicoli, M.C., DallaRosa, M. & Lerici, R.C. (1987) Caratteristiche chimiche dell'estratto di caffeÁ: Nota I. Cinetica di estrazione della caffeina e delle sostanze solide. Ind. Aliment., 5, 467. Nicoli, M.C., DallaRosa, M. & Lerici, R.C. (1990) Influence of some processing conditions on solid±liquid extraction of coffee. Lebensmittel-Wissen. Technol., 23, 386±9. Nicoli, M.C., DallaRosa, M. Lerici, R.C. & Bonora, R. (1989) Caratteristiche chimiche dell'estratto di caffeÁ: Nota III. Cinetica di invecchiamento ed influenza di alcuni interventi tecnologici sulla stabilitaÁ della bevanda. Ind. Aliment., 28, 706± 10. Nicoli, M.C. Severini, C. DallaRosa, M. & Lerici, C.R. (1991) Effect of some extraction conditions on brewing and stability of coffee beverage. In: Proceedings of the 14th ASIC Colloquium (San Francisco), pp. 649±53. ASIC, Paris, France. Nunes, F.M., Coimbra, M.A., Duarte, A.C. & Delgadillo, I. (1997) Foamability, foam stability, and chemical composition of espresso coffee as affected by the degree of roast. J. Agric. Food Chem. 45, 3238±43. Ohiokpehai, O., Brumen, G. & Clifford, M.N. (1982) The chlorogenic acids content of some peculiar green coffee beans and the implications for beverage quality. In: Proceedings of the 10th ASIC Colloquium (Salvador), pp. 177±86. ASIC, Paris, France. Peet, L.J. & Thye, L.S. (1954) Household Equipment, 4th edn, p. 83. John Wiley & Sons, New York. Peters, A. (1991) Brewing makes the difference. In: Proceedings of the 14th ASIC Colloquium (San Francisco), pp. 97±106. ASIC, Paris, France. Petracco, M. (1989) Physico-chemical and structural characterisation of espresso coffee brew. In: Proceedings of the 13th ASIC Colloquium (Paipa), pp. 246±61. ASIC, Paris, France. Petracco, M. (1995a) Grinding. In: Espresso Coffee: The Chemistry of Quality, (eds A. Illy & R. Viani), Chapter 6, p. 122. Academic Press, London. Petracco, M. (1995b) Percolation. In: Espresso Coffee: The Chemistry of Quality, (eds A. Illy & R. Viani), Chapter 8, p. 155. Academic Press, London.
164
Petracco, M. (1995c) Definition of espresso. In: Espresso Coffee: The Chemistry of Quality, (eds A. Illy & R. Viani), Chapter 2, p. 5. Academic Press, London. Petracco, M. & Marega, G. (1991) Coffee grinding dynamics: a new approach by computer simulation. In: Proceedings of the 14th ASIC Colloquium (San Francisco), pp. 319±30. ASIC, Paris, France. Petracco, M., Navarini, L., Abatangelo, A., Gombac, V. D'Agnolo, E. & Zanetti, F. (1999) Isolation and characterization of a foaming fraction from hot water extracts of roasted coffee. In: Proceedings of the 18th ASIC Colloquium (Helsinki), pp. 95± 105. ASIC, Paris, France. Petracco, M. & Suggi L.F. (1993) Espresso coffee brewing dynamics: development of mathematical and computational models. Proceedings of the 15th ASIC Colloquium (Montpellier), pp. 702±11. ASIC, Paris, France. Petropoulos, I. (1979) O tourkikos kafes en Elladi [The Turkish Coffee in Greece]. Ellenika Grammata Ekdoseis. Pictet, G. (1987) Home and catering brewing of coffee. In: Coffee Vol. 2 Technology (eds R.J. Clarke & R. Macrae), pp. 221±56. Elsevier Applied Science, London. Ratnayake, W.M.N., Hollywood, R., O'Grady, E. & Stavric, B. (1993) Lipid content and composition of coffee-brews prepared by different methods. Food Chem. Toxicol., 31, 263. Sehat, N., Montag, A. & Speer, K. (1993) Lipids in the coffee brew. In: Proceedings of the 15th ASIC Colloquium (Montpellier), pp. 869±72. ASIC, Paris, France. Severini, C., Nicoli, M.C., Romani, S. & Pinnavaia, G.G. (1995) Use of high pressure treatment for stabilizing coffee brew during storage. In: Proceeding of the 16th ASIC Colloquium (Kyoto), pp. 498±500. ASIC, Paris, France. Severini, C., Pinnavaia, G.G., Pizzirani, S., Nicoli, M.C. & Lerici, R.C. (1993) EÂtude des changements chimiques dans le cafe sous forme de boisson pendant l'extraction et la conservation. In: Proceedings of the 15th ASIC Colloquium (Montpellier) pp. 601±606. ASIC, Paris, France. Spiller, G. (ed.) (1998) Caffeine, CRC Press, Boca Raton, Florida.
Coffee: Recent Developments
Spiro, M. (1993) Modelling the aqueous extraction of soluble substances from ground roasted coffee. J. Sci. Food Agric., 61, 371. Spiro, M. & Hunter, J.E. (1985) The kinetics and mechanism of caffeine infusion from coffee: the effect of roasting. J. Sci. Food Agric., 36, 871±6. Spiro, M. & Page, C.M. (1984) The kinetics and mechanism of caffeine infusion from coffee: hydrodynamic aspect. In: J. Sci. Food Agric., 35, 925±30. Spiro, M. & Selwood, R.M. (1984) The kinetics and mechanism of caffeine infusion from coffee: the effect of particle size. In: J. Sci. Food Agric., 35, 915±24. Spiro, M., Toumi, R. & Kandiah, M. (1989) The kinetics and mechanism of caffeine infusion from coffee: the hindrance factor in intra-bean diffusion. J. Sci. Food Agric., 46, 349±356. Trugo, L.C. (1985) Carbohydrates. In: Coffee Vol. I Chemistry, (eds R.J. Clarke & R. Macrae), pp. 83±114. Elsevier Applied Science, London. Trugo, L.C. & Macrae, R. (1985) The use of the mass detector for sugar analysis of coffee products. In: Proceedings of the 11th ASIC Colloquium (LomeÂ), pp. 245±51. ASIC, Paris, France. Ueshima, T. (1975) Process for producing canned coffee. Japanese Patent n. 59-34571, January 1975. Urgert, R. (1997) Health effects of unfiltered coffee. Thesis, Agricultural University of Wageningen, The Netherlands. Van Zante, H.J. (1968) Household Equipment Principles, p. 358. Prentice Hall, New York. Viani, R. (1988) Physiologically active substances in coffee. In: Coffee Vol. 3 Physiology, (eds R.J. Clarke & R. Macrae) pp. 1± 31. Elsevier Applied Science, London. Woodman, J.S. (1985) Carboxylic acids. In: Coffee Vol. 1 Chemistry, (eds R.J. Clarke & R. Macrae) pp. 266±89. Elsevier Applied Science, London. Wrigley, G. (1988) Coffee, p. 491. Longman Scientific & Technical, London. Yamada, M., Komatsu, S. & Shirasu, Y. (1997) Changes in components of canned coffee beverage stored at high temperature. In: Proceedings of the 17th ASIC Colloquium (Nairobi) pp. 205±10. ASIC, Paris, France.
Chapter 8
Health Effects and Safety Considerations BenoõÃt Schilter, Christophe Cavin, Angelika Tritscher and Anne Constable Food Safety Group Nestle Research Center Lausanne, Switzerland 8.1 INTRODUCTION It is well established that foods and food components may significantly impact human health. Until recently, the overall safety and wholesomeness of foods have been established based exclusively on nutritional, microbiological and toxicological considerations. Recommendations have been mainly aimed at preventing nutrient deficiency and at avoiding microbiological and chemical intoxications. Nowadays, it has been increasingly acknowledged that under specific conditions, certain food constituents could provide consumers with physiological benefits beyond basic nutritional functions. Therefore, the overall evaluation of the impact of foods and food constituents on human health should consider all possible aspects, including potential nutritional, beneficial and adverse effects. The present chapter applies such an approach to coffee and relevant coffee components. As with any food plant, coffee is a complex chemical mixture. It is composed of over 1000 different chemicals. Theoretically, it can be anticipated that according to the doses involved, some may possess biological activities that could be considered potentially adverse to health or, conversely, beneficial. Although coffee has a long history of human food use of over 1000 years, until recently most of the studies on its health effects have focused on potential adverse and toxic effects. More than 100 diseases have been alleged to be caused or exacerbated by coffee consumption (Leviton et al., 1994). Among others, issues have concerned hypertension, cardiovascular disease, cancer, spontaneous abortion, delayed conception, low birth weight and osteoporosis. Despite a vast amount of research, evidence to support a direct link of coffee with these diseases has been limited and inconsistent. However, although not yet proven, there is an increasing body of
scientific literature suggesting potentially beneficial health effects of coffee and several of its constituents. For example, positive effects on performance, and protection against some types of cancers, liver disease and radiation-induced tissue damage have been documented. Coffee contains substantial amounts of antioxidants and this may explain some of its potentially beneficial activities, although several other important putatively advantageous active components have also been identified. The aim of this review is to assemble relevant scientific information on the impact of coffee consumption on human health. Potential adverse and beneficial effects are both addressed. Overall, the scientific information available supports the safety of moderate coffee consumption and reveals possible beneficial effects which deserve further attention.
8.2 OBJECTIVES AND SCOPE The biological and health effects of coffee have been extensively investigated in various animal and in in vitro model systems as well as in humans. A lot of literature addressing many different types of physiological and health effects is currently available. The objective of this chapter is to provide a summarised overview of the current knowledge and understanding on selected topics considered of public health importance. For more exhaustive information, the reader can consult several very thorough reviews which are cited and used as references in the present chapter. Priority has been given to information derived from human studies. Data from experimental models investigating mechanisms of action are presented only where relevant to discuss the plausibility of hypotheses drawn from human studies. Both potential adverse and beneficial effects are discussed.
165
166
8.3 COFFEE CONSUMPTION Coffee is consumed by a large proportion of the human population (about 70 to 80%). The type of coffee preparation used, the size of a standard cup and the amount consumed may vary significantly on both a geographical and an individual basis. In the present chapter, we have classified coffee consumption as occasional (< 1 cup/day), low (1 to 3 cups a day), moderate (3 to 5 cups/day) and high (> 5 cups/day). An average caffeine concentration of 60 to 85 mg of caffeine per cup has been assumed for instant and roasted and ground coffees, respectively.
8.4 COFFEE AND CANCER 8.4.1 Human data Numerous epidemiological studies have focused on the relationship between coffee consumption and cancer incidence at various sites. Thorough reviews on coffee and cancer have been published recently (World Health Organisation International Agency for Research on Cancer (WHO/IARC, 1991; Nehling & Debry, 1996). Overall, there is no conclusive evidence that coffee drinking represents a significant risk for the development of cancer in humans. For example, in a large study of almost 43 000 people conducted in Norway (Stensvold & Jacobsen, 1994), a country where the per capita coffee consumption is among the largest in the world, there was no association between coffee drinking and the overall risk of cancer.
(a) Cancers of the sexual organs Breast cancer The review of seven case studies by IARC (WHO/ IARC, 1991) did not reveal any association between breast cancer risk and the consumption of coffee. Recent studies further support that coffee intake is not related to breast cancer (Folsom et al., 1993; Smith et al., 1994; Tavani et al., 1998) in both pre- and postmenopausal women.
Ovarian cancer A weak positive association between coffee consumption and ovarian cancer has been seen in case-control studies. Only one reported a dose±response relationship and in most of the studies, the effect was not significant (Nehling & Debry, 1996; WHO/IARC,
Coffee: Recent Developments
1991). In its review IARC concluded that the evidence for an association between coffee drinking and ovarian cancer is inadequate, although the data indicate a marginal but significant increase in relative risk (WHO/IARC, 1991). In another review of the same body of literature, it was concluded that coffee consumption is unlikely to increase the risk of ovarian cancer (Leviton, 1990). A more recent study did not identify any association between coffee consumption and ovarian cancer (Polychronopoulou et al., 1993).
Prostate cancer Recent studies have not found any association between coffee consumption and prostate cancer (Jain et al., 1998; Hsieh et al., 1999).
(b) Cancers of the urinary tract Kidney, urinary tract The aetiology of renal cancer is still largely undefined (Nehling & Debry, 1996; Tavani & La Vechia, 1997). Available information indicates that renal cancer is unlikely to be associated with coffee consumption (WHO/IARC, 1991; Nehling & Debry, 1996; Tavani & La Vechia, 1997). A similar conclusion can be drawn for cancers of the urinary tract (WHO/IARC, 1991; Nehling & Debry, 1996).
Bladder cancer The epidemiological data regarding coffee consumption and bladder cancer have been equivocal. Many studies have suggested the possibility that coffee consumption could be a risk factor for bladder cancer, although many others found no such correlation (WHO/IARC, 1991; Nehling & Debry, 1996; Bruemmer et al., 1997; Donato et al., 1997; Probert et al., 1998). The association between coffee consumption and bladder cancer is often weak and only a few studies reported a dose±response relationship (WHO/IARC, 1991; Nehling & Debry, 1996). In its review of 1991, WHO/IARC concluded that the data are consistent with a weak positive relationship between coffee consumption and occurrence of bladder cancer (WHO/ IARC, 1991). It is important to note, however, that strong confounding factors such as smoking, dietary habits and occupation are well known for bladder cancer and it is recognised that they could have significant impacts on the data regarding coffee (Viscoli et al., 1993; Nehling & Debry, 1996). From two comprehensive overviews (Viscoli et al., 1993; Nehling &
Health Effects and Safety Considerations
Debry, 1996), it appears that coffee consumption is unlikely to be an important risk factor for bladder cancer in humans.
(c) Cancers of the gastrointestinal tract Oesophageal and gastric cancers The relationship between coffee consumption and cancers of mouth, pharynx, oesophagus and stomach has been addressed in several studies. Overall, no association was found (WHO/IARC, 1991; Nehling & Debry, 1996), in fact, some protective effects have been sometimes suggested (Inoue et al., 1998).
Pancreatic cancer Numerous studies have examined the potential link between coffee drinking and the risk of pancreatic cancer. Although most of the studies do not support any association, some have raised the possibility of a weak increase in pancreatic cancer for heavy coffee drinkers (WHO/IARC, 1991). Further studies and more recent analysis of the aetiological factors for pancreatic cancer have revealed that the weak effects of coffee, if any, are likely to be related to confounding factors such as smoking and therefore coffee consumption is not considered to represent a significant risk factor for cancer of the pancreas (Nehling & Debry, 1996; Silverman et al., 1998; Weiderpass et al., 1998). In a recent case control study, a U-shaped relationship was found between the level of coffee consumption and the risk for pancreatic cancer (Nishi et al., 1996). These authors concluded that as compared to abstinence, small to moderate amounts of coffee might prevent pancreatic cancer, whereas large amounts may increase the risk of developing this disease. The U-shaped dose±response effect was confirmed through a meta-analysis involving 14 studies published between 1981 and 1993 (Nishi et al., 1996). The lowest relative risk of pancreatic cancer was found at low consumption levels ranging from one to four cups a day.
Colorectal cancer The relationship between coffee consumption and the incidence of colorectal cancer has been addressed in numerous studies conducted in different geographical areas (WHO/IARC, 1991; Nehling & Debry, 1996). The data are inconsistent, but many case-control studies have revealed an inverse (protective) association between coffee drinking and the risk of colorectal
167
cancer. In its review of 1991, WHO/IARC concluded that although it is not possible to exclude bias and confounding as the source of the apparent inverse association, the collective evidence is compatible with a protective effect. A meta-analysis of coffee consumption and risk of colorectal cancer was published recently (Giovannucci, 1998). The results from 12 case-control studies showed an inverse association between coffee consumption and risk of colorectal cancer, while five cohort studies did not support a positive or negative link. Although definitive conclusions cannot be drawn because of inconsistencies between case-control and cohort studies, this metaanalysis strongly suggests a lower risk of colorectal cancer associated with a substantial consumption of coffee (> 4 cups a day).
8.4.2 Experimental data (a) Mutagenic and antimutagenic effects Many studies have addressed the mutagenic effects of coffee, using various biological test systems such as bacteria, yeast, fungi, mammalian cells and whole animal (WHO/IARC, 1991; Nehling & Debry, 1994a). Overall, relatively high concentrations of coffee have been shown to be slightly mutagenic in in vitro systems such as bacteria, fungi and mammalian cells. In bacterial assays coffee is particularly mutagenic to strains sensitive to oxidative mutagens (electrophiles). Most of the mutagenicity of coffee is abolished by the addition of exogenous detoxification systems such as liver extracts, catalase or peroxidases, implying that hydrogen peroxide (H2 O2 ) plays a key role in mediating coffee genotoxicity (Nagao et al., 1986). The formation of H2 O2 and the pro-oxidant activity of coffee in vitro has been attributed to polyphenolic thermal degradation products of chlorogenic and caffeic acid which reduce atmospheric oxygen in the presence of transition metals. The health significance of the mutagenic activity of coffee should be interpreted with caution since the in vitro assays used do not reflect adequately conditions present in physiological situations. In particular, the oxygen tension and the concentration of iron, two major players in the production of H2 O2 , are much higher in experimental assays than in the body. Furthermore, organisms possess efficient oxidant detoxifying mechanisms as well as repair systems. In this context, it is important to note that in contrast to the results of the in vitro studies, in vivo experiments in
168
rodents have not shown any evidence of mutagenicity (Nehling & Debry, 1994a). Depending upon the end-point measured, the mechanism of oxidation and the concentration range of the compound tested, dietary phenolic compounds can act either as pro-oxidants or antioxidants. The antioxidant activity of coffee has been demonstrated in in vitro systems (Stadler et al., 1994, 1995). It was shown that instant coffee and its polyphenolics which catalyse H2 O2 formation and mutagenicity also exhibit potent antioxidant and antimutagenic activity as evidenced by the protective effect of coffee against t-butylhydroperoxide-challenged cells (Stadler et al., 1994). Other in vitro studies have documented that coffee or polyphenolic-rich coffee fractions protect against the mutagenicity of several carcinogenic compounds such as heterocyclic amines (Obana et al., 1986) or nitrosating agents (Stich et al., 1982) as well as counteract the effects of UV-radiation (Obana et al., 1986). Studies conducted in in vivo test systems confirm the antimutagenic effects of coffee. For example, instant and roasted ground coffees were reported to protect mice against the genotoxic actions of various carcinogenic chemicals (Abraham, 1991). In summary, based on an overall review of the in vitro and in vivo mutagenicity data of coffee, and taking into consideration the mechanisms involved, it appears that in the amounts usually consumed by humans, coffee is unlikely to produce any genetic damage (Nehling & Debry, 1994a; Nehling & Debry, 1996). The possibility of protective, antimutagenic effects has gained experimental support.
(b) Experimental carcinogenicity data The carcinogenic potential of coffee has been investigated in several long-term animal bioassays. Feeding high levels of coffee as part of the diet did not produce tumours in either rats or mice (Nehling & Debry, 1996). On the contrary, some studies reported that instant coffee resulted in a decreased incidence of spontaneous tumours (Stadler et al., 1990). Other studies have shown that coffee or coffee constituents protect against the action of well-known carcinogens such as nitrosamines (Nishikawa et al., 1986) or 1,2dimethylhydrazine (Gershbein, 1994). Several studies have demonstrated that green as well as roasted coffees inhibit the development of 7,12-dimethylbenz[a]anthracene-induced carcinogenesis at various tissue sites in different animal cancer models (Huggett et al., 1997).
Coffee: Recent Developments
(c) Mechanistic information A number of coffee components have been identified as being potentially responsible for the chemoprotective effects of coffee. As discussed above, several coffee constituents have been shown to possess strong antioxidant properties resulting in significant antimutagenic activity (Stadler et al., 1994, 1995; Abraham, 1991). Among others, caffeine, polyphenols including chlorogenic acid derivatives and degradation products such as caffeic acid and phenylindans as well as melanoidins have been documented to exhibit antioxidant activities. There is increasing evidence that oxidative damage may be involved in various pathological processes including cancer and that antioxidants may be protective. Therefore antioxidant activity could be a key mechanism involved in the chemoprotective effects of coffee on cancer development. Other potential mechanisms of chemoprotection have emerged from experimental investigations. For example, the coffee-specific diterpenes cafestol and kahweol (C + K) have been reported to be anticarcinogenic in several laboratory animals (Huggett et al., 1997). Experimental evidence has indicated that this protective activity may be related to the ability of C + K to induce detoxifying enzymes such as glutathione S-transferases (Schilter et al., 1996; Huggett et al., 1997). Recently it has been suggested that besides a stimulation of detoxification processes, a reduction of carcinogen activation could also play an important role in the chemoprotective effects of C + K. With respect to the hepatocarcinogen aflatoxin B1 (AFB1), C + K was shown to decrease the expression of AFB1-activating cytochrome P450s in the rat liver and to strongly induce glutathione S-transferase subunit Yc2 which efficiently detoxifies aflatoxin 8,9epoxide, the most genotoxic metabolite of aflatoxin B1 (Cavin et al., 1998). Further studies are necessary to address the significance of these effects with regards to human chemoprotection.
8.4.3 Conclusions There is still debate on the potential impact of coffee drinking on human cancer. Based on the available data, there is currently no conclusive evidence that moderate coffee consumption (up to five cups a day) represents a risk for the development of cancer. The American Cancer Society did not find any evidence to recommend against moderate coffee intake (American Cancer Society, 1996). The most consistent effect observed in epidemiological studies is a potential, but not yet
Health Effects and Safety Considerations
demonstrated, protection against colorectal cancer and maybe other types of cancers. Experimental data and mechanistic information are compatible with such a possibility.
8.5 COFFEE AND CARDIOVASCULAR DISEASE The potential association between coffee consumption and cardiovascular disease has been highly debated and is not yet totally clarified. In humans, studies have focused on the potential link between coffee consumption and recognised endpoints of cardiovascular disease such as myocardial infarction and arrhythmias. Many reports have also investigated the possible effects of coffee on known cardiovascular risk factors such as hypertension, elevated blood cholesterol, and more recently, increased blood homocysteine.
8.5.1 Myocardial infarction or coronary death During the past two decades, reports addressing the potential relationship between coffee consumption and heart disease have provided conflicting results suggesting positive, negative or no effects. Part of the reported negative effects of coffee on cardiovascular disease can be attributed to the consumption of boiled coffee which is known to increase blood cholesterol (see Section 8.5.4). In addition, some of the data have been difficult to interpret because of the presence of important confounding factors which were often not taken into account, particularly in the earlier studies. Among the possible confounders are very strong risk factors for heart disease such as stress, cigarette smoking, alcohol consumption, dietary habits and sedentary lifestyle. Over time, methodologies have improved and adjustments for major confounding factors have been introduced. However, heavy coffee consumption seems to be directly associated with the lifestyle risk factors of cardiovascular disease (Debry, 1994) and therefore, it is always difficult to conclude that observed effects are specifically related to coffee drinking and not to residual confounding factors. The link between coffee consumption and myocardial infarction has been assessed in many prospective surveys (cohort studies) as well as in casecontrol studies. Most of the cohort studies did not find any correlation between moderate coffee drinking and myocardial infarction, while the absence of a link to heavy coffee consumption (more than five cups a day)
169
is less clear (Debry, 1994). In a meta-analysis including 11 prospective studies and involving a total of 143 030 people, it was concluded that there is no association between coffee consumption (at any level) and coronary heart disease (Myers, 1992). In this metaanalysis, case-control studies were excluded due to the potential for the introduction of serious uncontrolled bias such as problems in selecting appropriate controls and in estimating coffee intake. The conclusion was more ambiguous in another report presenting a meta-analysis of 22 studies (8 casecontrol and 14 cohort studies) of coffee use and myocardial infarction or coronary death (Greenland, 1993). Although from the case-control studies an increased risk is suggested, a more heterogenous trend was obtained from the cohort studies. Many studies involved in this analysis failed to account for strong confounding factors such as cigarette smoking. Based on this meta-analysis, the author concluded that an increased risk of coronary heart disease is unlikely at five cups of coffee a day, but cannot be ruled out at ten cups a day. Over the past few years, the potential association between coffee drinking and myocardial infarction has still been a matter of debate. For example, a study found a higher risk of myocardial infarction in women consuming more than five cups of coffee a day (Palmer et al., 1995). However, most of the new studies do not provide any evidence of a link between coffee and myocardial infarction. The Scottish Heart Health Study found the prevalence of coronary heart disease to be highest among those who abstain from coffee drinking and the lowest amongst those who drink five or more cups a day (Brown et al., 1993). In a follow-up study, a small benefit of coffee consumption was still observed among men (Woodward & Tunstall-Pedoe, 1999). In the US Nurses Study involving over 80 000 women, after adjustment for important cardiovascular risk factors such as smoking and age, there was no evidence for an association between coffee consumption and risk of coronary heart disease (Willett et al., 1996). In 1990, Tverdal et al. reported that coffee intake was related to death from coronary heart disease in men and women. After 6 years' follow-up, a slight increased risk was found only in subjects who drank nine or more cups a day (Stensvold et al., 1996). In a recent case-control study, neither caffeinated nor decaffeinated coffee was associated with the risk of myocardial infarction, even for those drinking more than four cups a day (Sesso et al., 1999). In summary, there is no evidence supporting a health relevant association between moderate coffee
170
consumption (up to five cups a day) and the occurrence of myocardial infarction or coronary death. A slight increased risk associated with higher coffee consumption cannot be ruled out, although data have to be interpreted with caution since heavy coffee drinking may just be an indicator of a high risk lifestyle for cardiovascular disease. A different conclusion has to be drawn for boiled coffee, which has been associated with an elevated risk for cardiovascular disease.
8.5.2 Arrhythmias Experimental, epidemiological and clinical studies have addressed the possible effects of coffee on heart rate. Because of its known pharmacological activity on cardiac tissue, caffeine has been the focus of many studies on arrhythmia. Studies include normal subjects, patients with pre-existing arrhythmias and patients with a recent history of myocardial infarction. Overall, the results of the available studies are ambiguous, although most of them suggest that usual, moderate amounts of coffee or caffeine do not affect cardiac rhythm. In a review on this subject (Myers, 1991), it was judged that caffeine ingestion at levels equivalent of up to five or six cups of coffee a day does not affect the severity or frequency of cardiac arrhythmias in healthy subjects, patients with coronary heart disease or persons with known ventricular ectopy. Such a conclusion was further confirmed in an epidemiological study of 128 934 people (Klatsky et al., 1993) which did not find any influence of coffee consumption on death attributed to cardiac arrhythmias. Further studies have confirmed that moderate caffeine is unlikely to affect heart rate in both normal people and patients with heart disease (Newby et al., 1996; Arciero et al., 1998; Daniels et al., 1998; Myers, 1998).
8.5.3 Caffeine and blood pressure The potential effects of coffee on blood pressure are still a matter of controversy and debate (Debry, 1994; Jee et al., 1999). It has been shown in animal models and in humans that caffeine can interfere with purinergic receptors and can therefore antagonise the vasodilatating effect of adenosine (Debry, 1994). This pharmacological effect increases peripheral vascular resistance and may therefore induce hypertension. Although less plausible, a caffeine-dependent stimulation of the sympathetic nervous system activity resulting in increased plasma norepinephrine has also been proposed as a possible trigger for high blood pressure (Debry, 1994).
Coffee: Recent Developments
The effects of caffeine on blood pressure have been extensively investigated in humans through various experimental designs. Several different types of clinical studies have been conducted involving acute caffeine dosing in the presence or absence of stress, or chronic exposure to caffeine. Data on both normotensive and hypertensive human populations are available. In addition, several epidemiological studies have addressed the relationship between coffee consumption and blood pressure in the general population.
(a) Acute dosing The effects of acute caffeine/coffee ingestion on blood pressure have been reviewed (Green et al., 1996; Myers, 1988, 1998). Either no effects or a small and transient rise within the first few hours following the dosing has been observed. Increases in blood pressure (up to 10±15 mm Hg) were found, mostly in studies involving caffeine-naõÈve individuals, where caffeine was restricted for variable periods of time before dosing. In habitual caffeine users, little or no change was found. It is well documented that although caffeine may raise blood pressure after a period of abstention, tolerance then develops with multiple exposure and blood pressure returns to the baseline level in 2 to 3 days. Abstinence from caffeine for periods as short as 24 hours may lead to a partial loss of tolerance to caffeine.
(b) Acute dosing and stress Physical and mental stress is known to increase blood pressure. The possible potentiating effects of acute caffeine dosing on the blood pressure increases induced by various types of stress have been investigated in many studies. In a review of 27 studies, it was concluded that, overall, stress plus acute dosing of caffeine cause small increases in blood pressure in caffeinenaõÈve individuals (Green et al., 1996). In the research setting, these effects have been found to be additive (Green et al., 1996; Myers, 1998).
(c) Chronic exposure The studies on repeated/chronic exposure to caffeine are in agreement with those using acute dosing (Green et al., 1996; Myers, 1988, 1998). Many of them did not find any effects on blood pressure while some reported a small increase. When an increase was found, the magnitude of the effect was much less than in the acute dosing studies (Myers, 1998). In many of the repeated/
Health Effects and Safety Considerations
chronic exposure studies, a complete or partial tolerance to caffeine was induced (Myers, 1998). In a recent meta-analysis of 11 controlled clinical trials in which the effects of long-term coffee drinking on blood pressure was assessed, a small increase of 2.4 and 1.2 mm Hg were found, respectively, for systolic and diastolic pressure (Jee et al., 1999). Compared to other factors known to affect blood pressure on a daily basis, the clinical significance of the increases, if any, resulting from caffeine ingestion has been considered to be minimal (Green et al., 1996; Myers, 1998). It is considered that 24-hour ambulatory blood pressure is the most accurate measure of blood pressure status. Recent studies applying this approach have provided inconsistent data with increases, decreases or no effect on blood pressure (Green et al., 1996; Myers, 1998). Overall, these studies support previous results suggesting that chronic caffeine exposure has modest or no effects on blood pressure (Myers, 1998).
(d) Studies in hypertensive people The potential effects of caffeine or coffee drinking on blood pressure have also been investigated in hypertensive subjects. For example, 2 weeks of caffeine use versus placebo were compared in treated hypertensive patients (Eggertsen et al., 1993). No effect of caffeine was found on ambulatory blood pressure. Similarly, in untreated patients with borderline hypertension, caffeine use over 2 weeks had no effect on ambulatory blood pressure (MacDonald et al., 1991).
(e) Epidemiological studies The potential link between coffee consumption and blood pressure in the general population has been addressed in several epidemiological studies. The results are variable and inconsistent (Green et al., 1996; Myers, 1988, 1998) and they suffer from methodological limitations. Reports have indicated no association, positive associations, and inverse relationships with systolic and/or diastolic blood pressure (Green et al., 1996). One study found a curvilinear association (Stensvold et al., 1989), with abstainers and high users (more than nine cups) showing no difference in blood pressure, but with those taking one to four cups per day showing a slight rise.
(f) Summary Single doses of caffeine corresponding to several cups of coffee produce a small increase in blood pressure,
171
mainly in caffeine-naõÈve individuals. The increases found are considered to be within the physiological range which could be observed during common activities such as conversation (Myers, 1998). Caffeine tolerance develops after 1 to 3 days of repeated exposure. Epidemiological surveys have provided inconsistent results which preclude drawing definitive conclusions. Overall, they do not support a major role of caffeine in inducing hypertension. Stress and psychological tension, which are known to increase blood pressure, may be correlated with higher coffee consumption. Therefore coffee may not be the cause of the effects observed in some studies.
8.5.4 Serum cholesterol Population studies regarding the influence of coffee consumption on serum cholesterol have provided conflicting results. For example, in a review article, it was found that in approximately two-thirds of the studies evaluated, coffee consumption was associated with an increase in serum cholesterol concentration (Thelle et al., 1987). In only some studies was this effect dose-dependent. Further analysis of available information revealed that the correlation between coffee and serum cholesterol was mostly restricted to studies conducted in Scandinavia, but was much less consistent in studies from USA or other European regions (Ugert & Katan, 1997). High consumption of boiled coffee (decanted without filtering), a brew particular to Scandinavian countries, has been clearly associated with elevated levels of serum cholesterol (Ugert & Katan, 1997). In this region, a relation between boiled coffee consumption and coronary heart disease was also identified (Tverdal et al., 1990). In Scandinavia, a substantial percentage of the decline in serum cholesterol over the years has been attributed to the switch from boiled to filtered coffee (Tuomilehto & Pietinen, 1991), leading to a reduction in cardiovascular disease (Tverdal et al., 1990; Johansson et al., 1996; Stensvold et al., 1996). Subsequent epidemiological and controlled clinical studies have further confirmed that the hypercholesterolaemic effect of coffee was dependent on the method of preparation of the coffee brew. For example, in contrast to boiled coffee, the consumption of filtered coffee has no significant effect on serum cholesterol levels (Van Dusseldorp et al., 1991) whereas Turkish style coffee appears to increase serum cholesterol (Kark et al., 1985). A series of clinical trials has found that the hypercholesterolaemic agents are present in the lipid
172
fraction of boiled coffee. In addition, the major causative factors were identified as the diterpenes cafestol and kahweol, which are present mainly as fatty acid esters (Weusten-Van der Wouw et al., 1994; Ugert & Katan, 1997). These investigations demonstrated a dose response and reversible effect of cafestol and kahweol on increasing serum cholesterol concentrations. Chemical analysis of the diterpenes in different coffee brews has shown that Scandinavian-type boiled coffee, Turkish and cafetiere coffee contain the highest amounts of cafestol and kahweol, while instant and filtered coffee contain negligible amounts and espresso intermediate amounts (Ugert et al., 1995; Ugert & Katan, 1997). In summary, only substantial amounts of coffee containing high levels of the diterpenes cafestol and kahweol, such as boiled coffee, have been shown to consistently raise serum cholesterol levels. Filtering the brew or using other types of coffee prevent any hypercholesterolaemic effect.
8.5.5 Serum homocysteine For more than 20 years, moderately elevated serum concentrations of total homocysteine have been correlated with an increased risk of atherosclerotic cardiovascular disease and peripheral vascular disease (Meleady & Graham, 1999). Recently, it has been suggested that the association may be causal (Meleady & Graham, 1999). The mechanisms involved are still unclear. Experimental evidence has indicated that homocysteine may promote vascular damage through oxidative stress (Meleady & Graham, 1999). Only a few studies have explored the lifestyle factors determining homocysteine blood concentrations. The most important factors identified up to now have been a low intake of fruits, vegetables, folic acid, vitamin B12 and vitamin B6 . Cigarette smoking and age also play an important role. Recently, three studies have suggested a potential link between heavy coffee consumption and higher plasma total homocysteine (Nygard et al., 1997a; 1998; Oshaug et al., 1998; Stolzenberg-Solomon et al., 1999). For example, Nygard et al. (1997a, 1998) reported a direct and dose-dependent association between coffee consumption and blood homocysteine in a population in Norway. This effect was particularly pronounced in subjects drinking nine or more cups of coffee a day. No effect was found with decaffeinated coffee or caffeinated tea. In contrast, no significant association was found between coffee consumption or caffeine and total blood homocysteine in a sample of the
Coffee: Recent Developments
atherosclerotic risk in communities study (Nieto & Comstock, 1997). One controlled intervention clinical study is available where several sources of bias present in previous studies were eliminated. It was observed that the consumption of 1 litre of strong unfiltered boiled coffee every day for 2 weeks was associated with a 10% increase in mean plasma total homocysteine concentration (Grubben et al., 2000). However, this extreme coffee intake may affect diet composition and other factors which may influence plasma homocysteine (Vollset et al., 2000). The association between coffee consumption and total blood homocysteine is an unexpected finding and deserves further confirmation. Furthermore, there are no plausible mechanisms proposed yet. Since several other lifestyle factors are likely to play a major role, it is still unclear whether the coffee effect found in the observational studies is real or whether it results from residual confounders such as smoking or other unmeasured or unidentified factors. The authors of the Norwegian study acknowledged the possibility that residual confounders with vitamins, especially folate, could be responsible for their finding (Nygard et al., 1997b). Furthermore, the health significance of the coffee effects is difficult to interpret. In the Norwegian study (Nygard et al., 1997a), it was observed that elevated plasma homocysteine levels were correlated with coffee intake mainly in people with low to intermediate homocysteinemia. Therefore the link between the coffee-dependent increase in homocysteine and overall cardiovascular risk within the general population may not be straightforward to establish. In summary, a slight increase in blood homocysteine in heavy coffee drinkers has been shown in several studies. The direct implication of coffee and the health significance of such an effect have still to be demonstrated.
8.5.6 Conclusions The potential relationship between coffee drinking and cardiovascular disease has been the subject of much debate and investigation. It is important to note that there are many other dietary and lifestyle factors which are known to have a greater impact on cardiovascular disease. Some of these factors are associated with coffee consumption and may explain some of the coffee effects reported. Overall, except for brews rich in diterpenes such as boiled coffee, the data show that moderate coffee consumption is not a causal factor in the development of cardiovascular disease.
Health Effects and Safety Considerations
8.6 COFFEE AND BONE HEALTH Osteoporosis is a chronic degenerative bone disease that affects mainly, but not exclusively, postmenopausal women, in which demineralisation (calcium loss) of bones leads to an increased likelihood of fracture. It has a complex aetiology that includes genetic, physiological and environmental contributors. Among factors, oestrogen deficiency, smoking, heavy alcohol consumption, lack of exercise, obesity and inadequate nutrition are believed to play significant roles in the development of this disease. Of the nutritional factors, low calcium intake throughout life is believed to be most important, although low intakes of other minerals and poor vitamin D status have also been implicated. During recent decades the number of osteoporotic fractures has increased in industrialised countries. Since this increment cannot be solely explained by an increased life expectancy, other aetiological factors have been intensively examined, particularly those related to nutrition and lifestyle. Experimental data obtained in animals and humans have suggested that caffeine may affect calcium metabolism. The potential role of caffeine, mainly through coffee consumption, as a contributing factor for bone loss in humans has received a lot of attention. In recent years numerous studies have reported on caffeine consumption as a possible risk factor for osteoporosis.
8.6.1 Calcium metabolism Caffeine has been shown to increase the urinary excretion of calcium in experimental animals (Debry, 1994). In humans, several studies have also suggested that caffeine may negatively influence calcium balance, particularly in women, but the data are inconsistent (Debry, 1994). An initial study reported a small but significant negative effect of caffeine intake on calcium economy in 168 premenopausal women (Heaney & Recker, 1982). This effect was, however, no longer significant when dietary calcium intake was considered. Other studies indicated that caffeine induces a significant acute calcium diuresis (Massey & Wise, 1984; Debry, 1994; Heaney, 1998). However, subsequent investigation suggested that the increase in calcium excretion was followed by a reduction in excretion, resulting in a net negative effect on calcium balance lower than previously thought (Kynast-Gales & Massey, 1994). The effect of caffeine on calcium metabolism was recently addressed in a double-blind, randomised,
173
placebo-controlled, cross-over metabolic study. The administration of 400 mg of caffeine over 19 days did not produce any effect on total 24 hours calcium loss (Heaney & Recker, 1994; Barger-Lux & Heaney, 1995). However, a small negative balance effect was detected due to a slight decrease in calcium absorption efficiency. Importantly, the calcium intake of the women enrolled in these studies was significantly lower than current recommendations. Therefore, these data indicate that caffeine may lead to a small negative calcium balance when dietary calcium intake is inadequate. This has been confirmed in another study showing that caffeine only produced observable effects on calcium metabolism in women consuming less than 600 mg of calcium daily (Massey et al., 1994). New dietary reference intakes for calcium intakes in adults range from 1000 to 1300 mg/day according to lifestage. Overall, the magnitude of the caffeine effect on calcium balance is low and it has been estimated that it could be offset simply by the addition of 1 to 2 tablespoons of milk to a cup of coffee (Barger-Lux & Heaney, 1995). In summary, it is currently thought that at standard recommended calcium intake, caffeine is unlikely to have harmful effects on calcium bone economy. In the most recent US RDA, it was stated that the available evidence does not warrant a specific calcium intake recommendation for people with a different caffeine intake.
8.6.2 Osteoporosis Studies addressing the potential link between caffeine consumption and the risk of osteoporosis have given contradictory results (Debry, 1994; Heaney, 1998). Comparison and interpretation of the studies are complicated by the variety of both bone-related measurements (e.g. fracture risk, bone density, bone mass) and caffeine intake estimations. Furthermore, in many studies, variables known to affect bone loss such as calcium intake, smoking, body weight, physical activity and hormone replacement therapy were not or could not be adequately controlled for. This aspect is of particular importance for calcium intake, which is likely to be inversely correlated with caffeine exposure (Heaney, 1998). Although some epidemiological studies have suggested that caffeine may slightly increase the risk of fracture or may decrease bone density, the majority of the reports available failed to find any effects from caffeine (Debry, 1994; Heaney, 1998). A review of 23
174
observational studies indicated that five showed a negative effect of caffeine on bone health, two a partial effect and 16 showed no effect (Heaney, 1998). Two studies proposed a negative effect only for women whose dietary intake of calcium is below the recommendations (Barrett-Connor et al., 1994; Harris & Dawson-Hughes, 1994). In a recent study which was designed to minimise confounding variables no association could be found between caffeine intake and bone loss in 138 healthy postmenopausal women (Lloyd et al., 1997). The potential effects of caffeine on bone health have also been evaluated in women who are still in the period of bone gain. Caffeine did not affect the rate of gain in spinal bones in women of 30 years age or less (Packard & Recker, 1996). Another study revealed that dietary caffeine intake at levels presently consumed by American teenage women was not correlated with total bone mineral gain or hip bone density at age 18 (Lloyd et al., 1998).
8.6.3 Conclusions Based on current available literature, there is no evidence that moderate caffeine intake through coffee consumption has a harmful effect on bone health in normal healthy individuals ingesting recommended levels of calcium. Caffeine and coffee consumption are therefore unlikely to be important risk factors for osteoporosis. Furthermore, data indicate that the small effects of caffeine in women with calcium deficiency may be counteracted by increased calcium intake, for example through milk.
8.7 REPRODUCTIVE AND DEVELOPMENTAL POTENTIALS OF COFFEE AND CAFFEINE Since caffeine was shown to be teratogenic in animal models, safety concerns were raised regarding coffee drinking during pregnancy. It is well documented that caffeine metabolism is slower in pregnant women, resulting in longer and possibly higher exposures. Consequently, there have been a number of animal and human studies addressing the potential effects of coffee/caffeine on various reproductive and developmental outcomes such as teratogenicity (congenital malformations), neurodevelopment, low birth weight (growth retardation and prematurity), spontaneous abortion (miscarriage) and fertility parameters.
Coffee: Recent Developments
8.7.1 Congenital malformations (a) Human data Several epidemiological studies have examined the association between caffeine ingestion and congenital malformations (Nehling & Debry, 1994b; Brent, 1998). Most of these studies do not support any link between caffeine intake and teratogenicity (Nehling & Debry, 1994b; Brent, 1998). In a review of 14 publications addressing the relationships between caffeine or coffee consumption and congenital malformations, only three mentioned potential teratogenic effects (Nehling & Debry, 1994b), while the other 11 failed to provide evidence for an association. Jacobson et al. (1981) reported three cases of extrodactyly in children born from mothers consuming high amounts of coffee (8 to 25 cups/day). Unless an increased incidence of such a malformation is observed and confirmed in other controlled, large-scale epidemiological studies, this report cannot be appropriately interpreted. In a Japanese study, the rate of many different types of congenital malformations, including chromosomal abnormalities, was found to be about twice as high in the coffee drinkers than in non-drinkers (Furuhashi et al., 1985), suggesting that coffee may have teratogenic and mutagenic effects. This outcome is surprising since most of the well-documented teratogens are known to produce a specific pattern of teratogenicity and not a wide variety of different malformations. In addition there is no evidence that coffee is mutagenic in vivo. Hidden bias has been considered to be the most probable explanation for these data (Narod et al., 1991). In the third study, it was suggested that drinking more than eight cups of coffee a day during pregnancy was weakly associated with an increased frequency of congenital malformations (BorleÂe et al., 1978). However, this study involved only a small number of cases and did not account for important confounding factors such as tobacco consumption. Furthermore, the statistical analysis was questionable (Nehling & Debry, 1994b). Overall, there is no evidence to implicate moderate coffee or caffeine consumption in the aetiology of human congenital malformations.
(b) Animal data Contrary to the human data, dose-dependent teratogenic effects have been observed in various animal models including mice, rats, rabbits and monkeys (Nehling & Debry, 1994b; Brent, 1998). Most of these effects were observed with very high doses of caffeine,
Health Effects and Safety Considerations
which resulted in maternal toxicity. Such doses are not achievable through normal coffee consumption in humans. In addition, the mode of administration of caffeine plays a major role on the final outcome (Nehling & Debry, 1994b). Teratogenic effects in animals have been principally observed in studies using a daily administration of caffeine as single high doses (injection, subcutaneous or gavage) resulting in high plasma concentrations. Exposing the animals through multiple fractionated administrations or dietary feeding generally failed to produce any effects or required much higher doses to be active. The experimental designs of most of the animal studies showing teratogenic effects do not reflect the pattern of human caffeine exposure through coffee consumption. The relevance of these data to humans is therefore difficult to evaluate.
8.7.2 Neurodevelopmental effects It is well known that caffeine is a stimulant because of its neuropharmacological properties. Neuropharmacological agents may produce subtle neurochemical or behavioural effects on developing organisms at doses which do not induce any overt toxicity. Therefore the potential neurodevelopmental effects resulting from either pre- or postnatal exposure to caffeine have been investigated in both humans and animal models (Nehling & Debry, 1994b; 1994c).
(a) Human data Limited information is available regarding the potential influence of caffeine intake by pregnant women on the function of the newborn nervous system. The consequences on neurodevelopment resulting from maternal caffeine consumption during pregnancy were investigated by following about 500 children from birth to 7 years of age. The prenatal caffeine exposure did not influence neurobehavioral outcomes and the suckling reflex in the first 2 days of life (Barr & Streissguth, 1991) and no effects on cognitive and motor development could be observed at 8 months of age (Streissguth et al., 1980; Barr et al., 1984). In addition, no effects on intelligence quotient at 4 and 7 years or on motor ability at 4 years and on vigilance at 7 years were found (Barr & Streissguth, 1991). In one study, no physiological or neurobehavioral disturbances were observed in infants who had measurable caffeine plasma concentrations at birth (Dumas et al., 1982) although in another, increased visual arousal and nervousness were linked to salivary
175
caffeine concentration in 1 to 2-day-old babies (Emory et al., 1988). A withdrawal type syndrome has been observed in babies from mothers exposed to very high levels of caffeine (Nehling & Debry, 1994b). In addition, some studies have suggested a possible role of caffeine on respiratory control function and neonatal apnea (Toubas et al., 1986) and therefore the issue of caffeine exposure and sudden infant death syndrome has been raised. In a recent retrospective epidemiological study, an increased risk for sudden infant death syndrome was associated with heavy maternal caffeine ingestion (Ford et al., 1998) although no dose response was found. Since such a link has not been described before, further analysis is required, taking into consideration all possible known risk factors which could play a confounding role. In this context it is important to note that caffeine has been used safely in clinical settings to treat apnea in premature babies (Nehling & Debry, 1994c).
(b) Animal data The administration of moderate to high doses of caffeine or coffee to female rodents during gestation has been documented to significantly alter the brain neurochemistry and composition of the neonatal pups (Nehling & Debry, 1994b; Brent, 1998). In addition, pre- and perinatal caffeine exposure have been shown to affect sleep control and behaviour in the offspring. For example, in the offspring of dams exposed to moderate to high doses of caffeine during gestation and/or lactation, increased locomotion and spontaneous activities were observed later in life (Nehling & Debry, 1994b). Effects on learning abilities have also been found. In general, long-term behavioural disturbances in rodents have only been observed at high levels of maternal caffeine exposure, which induce other effects such as delayed growth. The doses involved are unlikely to be achievable in the human situation.
8.7.3 Low birth weight, growth retardation and prematurity Low birth weight is often defined as < 2500 g. Low birth weight may be the result of a shortened gestational period (prematurity) or the consequence of intrauterine growth retardation resulting in a `small for age' infant. Many medical, social and lifestyle factors are known to influence birth weight, some of them being directly correlated with coffee consumption. Therefore the interpretation of the literature in this
176
field is particularly difficult. Maternal caffeine or coffee ingestion as a risk factor for delivering low birth weight infants has been extensively investigated. Generally, there appears to be no relationship between caffeine intake during pregnancy and the probability of premature delivery in humans (Hinds et al., 1996). The epidemiological literature on maternal coffee or caffeine intake during gestation and the risk factor for low birth weight or `small for age' babies is conflicting. Some studies did not find any evidence of association while others found a direct correlation (Brent, 1998; Narod et al., 1991; Nehling & Debry, 1994b). In studies showing a link, effects were observed at relatively high dose of caffeine, often > 300 mg/day, although some studies suggested effects at lower levels of exposure. The standard reported decrease in birth weights ranged from 70 to 121 g (Narod et al., 1991). More recent studies have also provided contradictory data. For example, in a study conducted in Brazil, the proportion of mothers who delivered babies with intrauterine growth retardation increased according to coffee consumption (Rondo et al., 1996). In a Japanese study, coffee was not identified as a risk factor for low birth weight (Maruoka et al., 1998). The likelihood for a low birth weight baby to suffer from further health problems later in life is thought to be dependent upon the cause of the low birth weight. There is no evidence reported on potential long-term adverse consequences resulting from coffee- or caffeine-induced low birth weight. The inconsistency of the literature regarding maternal caffeine ingestion and low birth weight has been thought to result from the difficulty of establishing small effects or their absence. Furthermore, the direct relation between coffee intake and other agents documented to adversely affect fetal growth such as smoking and alcohol consumption complicates the interpretation of the data in this area of research. Smoking is thought to induce foetal hypoxia through an increase in blood carboxyhemoglobin and a reduction in placental blood flow (Nehling & Debry, 1994c). These effects trigger an adaptive tissue response. Caffeine, by blocking adenosine receptors, could inhibit the normal adaptive cellular response to hypoxia and therefore may potentiate the effects of smoking. Reports have suggested that the effects of caffeine on birth weight were stronger in or even restricted to smokers (Nehling & Debry, 1994b,c). In a recent prospective study (Cook et al., 1996) the relationship of fetal growth to caffeine intake and blood caffeine concentrations during pregnancy was investigated. In
Coffee: Recent Developments
smokers, caffeine consumption was inversely related to birth weight. Smokers were found to consume more caffeine than non-smokers. However, blood caffeine concentrations were lower in smokers than in nonsmokers due to the stimulation of caffeine metabolism by tobacco consumption. No relation was found between blood caffeine concentrations during pregnancy and birth weight. These data support the key role played by cigarette smoking as a confounder in the studies addressing the association between prenatal caffeine exposure and birth weight.
8.7.4 Spontaneous abortion The relationship between maternal coffee or caffeine ingestion in pregnancy and the risk of spontaneous abortion has been extensively investigated. The data are conflicting. Several studies have suggested an association while others did not observe any effects (Narod et al., 1991; Nehling & Debry, 1994b,c; Hinds et al., 1996; Brent, 1998). Recent studies found either no effect or a slight increase in spontaneous abortion associated with coffee or caffeine consumption. In one study, neither total estimated caffeine nor individual caffeinated beverage consumption during the first trimester of pregnancy was associated with an appreciable increase in risk for spontaneous abortion (Fenster et al., 1997). Another study found a modest increased risk of clinically recognised spontaneous abortion when caffeine intake exceeds 300 mg per day (Dlugosz et al., 1996). Data have been reported on the association between maternal serum paraxanthine, the primary caffeine metabolite, and the risk of spontaneous abortion (Klebanoff et al., 1999). It was found that only extremely high serum paraxanthine concentrations corresponding to a consumption of more than six cups/day were associated with spontaneous abortion. Confounding factors and bias may have played an important role in the association between caffeine or coffee consumption and spontaneous abortion reported in some articles. For example, it is known that nausea in pregnancy is associated with food aversion and is likely to result in a reduction of coffee or caffeine consumption. Furthermore, it is documented that nausea is associated with a decrease in spontaneous abortion (Stein & Susser, 1991). Therefore, it was postulated that a pregnancy with a higher probability of a viable outcome might increase nausea and in consequence decrease caffeine ingestion. Based on this hypothesis, it appears that studies addressing fetal loss and caffeine which do not account for nausea are likely to over-
Health Effects and Safety Considerations
estimate the risk of caffeine exposure. Most of the studies available do not have any information on nausea incidence. One study reported that the risk of spontaneous abortion for heavy caffeine consumers varied according to whether there was nausea during pregnancy (Fenster et al., 1991).
8.7.5 Fertility Animal studies on the effects of caffeine on fertility and reproduction in animals are limited. High doses have been shown to produce an increased time to pregnancy in rodents, suggesting a possible effect of coffee or caffeine on delaying fertility. The literature regarding coffee or caffeine consumption and human fertility is controversial and inconsistent. Some studies observed a reduction in fertility associated with coffee intake, sometimes in a dose-dependent way (Wilcox et al., 1988), others found no effects, even in heavy drinkers (Narod et al., 1991; Nehling & Debry, 1994c; Bolumar et al., 1997). In a large study conducted in Denmark involving 10 886 women (Olsen et al., 1991), a delayed time of conception (subfecundity) was found, but only in smokers consuming high doses of coffee (> 8 cups). In a recent American study, it was found that high levels of caffeine consumption (> 300 mg/day) may result in delayed conception in women who do not smoke cigarettes (Stanton & Gray, 1995). In a recent study involving a large random European population, subfecundity was associated with a high caffeine intake (> 500 mg/day) in fertile women (Bolumar et al., 1997). This effect was stronger in smokers. At the highest level of intake, the time leading to the first pregnancy was increased by 11%. In this study, several but not all important confounding factors were adjusted for and the caffeine exposure determinations were relatively accurate. Based on their data, the authors stated that caffeine, among others, can be considered a weak risk factor that probably reduces fecundity by a certain fraction, but without being a sufficient cause of infertility. Delayed conception is relatively common and many factors, including exercise, stress, nutrition, lifestyle and social influences, may be involved. In many studies, most of these confounding factors were not or could not be adjusted for. In addition, the question of the mechanism involved in the potential effects of coffee or caffeine on fertility is not answered. In conclusion, there is no solid evidence linking moderate coffee consumption and adverse effects on fertility parameters.
177
8.7.6 Conclusions Caffeine, mostly through coffee consumption, has been implicated in several types of developmental and reproductive adverse events. Based on the literature, it appears that caffeine at levels lower than 300 mg/day is unlikely to produce in humans any effects on reproductive and developmental health outcomes. The information currently available does not allow the accurate estimate of the effects of higher levels of exposure.
8.8 EMERGING BENEFICIAL HEALTH EFFECTS Coffee has recently been shown to have positive health effects. Although not yet proven and requiring substantial confirmation through well controlled epidemiological studies, taking into account all possible bias and confounders, these emerging beneficial effects were considered worth discussing.
8.8.1 Neuroactivity Coffee consumption has been perceived to have a positive influence on human behaviour. Caffeine appears to be the key coffee constituent responsible for these effects. For example, caffeine has been documented to increase alertness, to improve performance on vigilance tasks and to reduce fatigue (Smith, 1998). A beneficial effect which has emerged in the area of coffee-related neuroactivity is the potential preventive influence of caffeine on suicide and depression. A strong inverse association has been reported between coffee intake and risk of suicide in a prospective study involving 128 933 people (Klatsky et al., 1993). A dosedependent relationship was observed and those consuming more than six cups of coffee per day showed a 5-fold lower risk of suicide than nonconsumers. Another prospective study showed the relative risk of suicide in women consuming two to three cups of coffee per day to be about 3-fold lower than in non-coffee drinkers (Kawachi et al., 1996). This association could be spurious if depressed patients avoid caffeine either spontaneously or through the advice of health professionals. However, the neuroactive property of caffeine could explain a preventive effect of caffeine. Compared to placebo, experimental administration of caffeine has been shown to increase the subjective feelings of well-being, social disposition, self-confidence, energy and motivation at work
178
(Griffiths et al., 1990). In a psychiatric setting, reports suggest that patients' experience improved mood (Furlong, 1975) and decreased irritability (Stephenson, 1977) following administration of caffeine. Although no population-based prospective studies on coffee or caffeine intake and depression have been reported, a cross-sectional study of Japanese medical students found that high intake of caffeine was associated with fewer depressive symptoms among female but not male students (Mino et al., 1990).
8.8.2 Chemoprotection Epidemiological evidence indicates that coffee consumption may be protective against certain types of cancers such as colon cancer (Giovannucci, 1998). Experimental data suggest that this chemoprotective effect could be related to the presence in coffee of both strong antioxidants and components able to stimulate chemical detoxification processes (see Section 8.4). Since these coffee components act through relatively general mechanisms, other chemoprotective effects can be expected. Recently, several epidemiological studies have observed that coffee consumption significantly reduced the risk of developing liver disease (cirrhosis) induced by alcohol ingestion (Klatsky et al., 1992, 1993; Corrao et al., 1994). In the first report on that topic, it was observed that persons who drank four or more cups of coffee a day had a 5-times lower risk of developing alcoholic cirrhosis than non-coffee drinkers (Klatsky et al., 1992). In a recent study, the beneficial effect of coffee on alcoholic cirrhosis was further confirmed (Collaborative GESIA group, in press). In addition, this study addressed the joint action of coffee consumption and hepatic viral risk factors of cirrhosis on the resulting risk of developing the disease. Coffee was found to antagonise the promoting effects of hepatitis B and C infection on cirrhosis development, suggesting a protective effect of coffee on non-alcoholic cirrhosis. Further work is required to demonstrate clearly this chemopreventive effect. For instance, it has to be clarified whether the inverse association between coffee intake and cirrhosis observed in epidemiological studies is real or whether it is a consequence of coffee aversion in patients developing severe cirrhosis. Furthermore, the mechanism of action has to be established. A plausible mechanism refers to the presence in coffee of potentially protective factors including antioxidants and detoxification stimulating agents. Evidence excludes caffeine as playing a key role
Coffee: Recent Developments
(Collaborative GESIA Group, in press; Klatsky et al., 1992). Coffee consumption has been repeatedly found in clinical and epidemiological studies to reduce the levels of serum g-glutamyltransferase, a marker of hepatobiliary diseases (Nilssen & Forde, 1994; Ugert & Katan, 1997; Tanaka et al., 1998). This effect further supports a potentially more general protective effect of coffee on the liver. The coffee-specific diterpenes cafestol and kahweol have recently been shown to reduce serum levels of g-glutamyltransferase and to modulate the levels of other common indicators of liver function (Ugert et al., 1995; Ugert & Katan, 1997) and detoxification (Schilter et al., 1996; Cavin et al., 1998). Other recent data suggest that coffee and caffeine may possess broader chemoprotective properties. The potential development of late radiation-related tissue complications is an important dose-limiting consideration in clinical radiotherapy treatment to control tumours. An epidemiological study has found that patients consuming caffeine-containing beverages such as coffee at the time of their radiotherapy against cervical cancer had significantly decreased incidence of severe late radiation injury (Stelzer et al., 1994).
8.9 COFFEE CONSUMPTION – SAFETY CONSIDERATIONS Traditional foods are thought to be safe on the basis of long-term experience, even though these foods may contain inherent toxicants. With respect to toxic potential, food is presumed safe unless a significant risk has been identified in humans. However, it is also acknowledged that the innocuity of food is not strictly demonstrated without the provision of a fully documented history of safe use in humans based on specific data. For most traditional foods, such data are not available. Coffee has been consumed for over 1000 years by many human beings. There has been no evidence that coffee intake is associated with clearly identified adverse health effects. Therefore coffee should be considered as a traditional food with a long history of safe use. In contrast to most other traditional foods, coffee and coffee components have been the subject of many extensive scientific investigations in both animal models and humans. Epidemiological studies addressing the potential adverse effects of moderate coffee consumption (3 to 5 cups/day) on various key health outcomes including cancers, cardiovascular disease, osteoporosis and developmental effects have been largely inconsistent.
Health Effects and Safety Considerations
Where present, the effects were usually weak and not dose-dependent. Plausible mechanisms are often missing. Based on this literature, it appears that, in general, moderate coffee consumption ranging from 3 to 5 cups per day is unlikely to be of any health concern. While most of the human data converge to show that moderate coffee consumption is safe, the information presently available does not allow accurate evaluation of the risk associated with higher levels of consumption. It is important to note that in the situation of high intake, residual confounding factors may significantly bias the data. Many studies have indirectly seen that heavy coffee consumption was directly related to lifestyles known to be important risk factors for vascular diseases, malignancies and developmental adverse effects. A recent study (Leviton et al., 1994) indicated that heavy coffee drinkers were more likely to smoke and less likely to take vitamin supplements and to consume a healthy diet (high vegetable, high vitamin, high fibre, low fat). The authors of this study proposed that heavy coffee drinkers may be at increased risk for a number of diseases not because of coffee consumption per se, but because of other aspects of their lives. Assuming an average caffeine concentration of 60 to 85 mg per cup of instant and roasted and ground coffees, respectively (Barone & Roberts, 1996), moderate coffee consumption as defined above corresponds to a caffeine exposure ranging from 180 to 425 mg. This is in the range of what has been considered safe by many authors cited in the present paper. Some developmental studies have sometimes suggested slight effects at lower doses, so it is prudent to advise pregnant women to stay at the lower level of the safe range in order to account for the remaining uncertainty. Safe levels of exposure to the cholesterol-raising diterpenes cafestol and kahweol have not been officially established. However, based on an average cafestol concentration of 1 mg/cup (0 to 3.1 mg/cup) and using clinical data on the effects of the diterpenes on blood cholesterol, the consumption of five cups per day of espresso coffee has been considered to have negligible hypercholesterolemic effects (Ugert et al., 1995). Comparison of this figure used as a safe level of exposure with cafestol occurrence data (Ugert et al., 1995; Ugert & Katan, 1997) reveals that, except for boiled Turkish and French press coffee, up to five cups of coffee a day are unlikely to have any appreciable effects on blood cholesterol.
179
8.10 CONCLUSIONS The potential health effects of coffee consumption have been extensively investigated in animal models and in human studies. Overall, the available information indicates that moderate consumption, corresponding to three to five cups of average strength coffee per day, is safe for human health. The data do not allow an accurate evaluation of the potential risk at higher consumption levels.
REFERENCES Abraham, S.K. (1991) Inhibitory effects of coffee on the genotoxicity of carcinogens in mice. Mut. Res., 262, 109±14. American Cancer Society 1996 Advisory Committee on Diet, Nutrition, and Cancer Prevention (1996) Guidelines on diet, nutrition, and cancer prevention: reducing the risk of cancer with healthy food choices and physical activity. Cancer J. Clin., 41, 334±7. Arciero, P.J., Gardner, A.W., Benowitz, N.L. & Poehlman, E.T. (1998) Relationship of blood pressure, heart rate and behavioral mood state to norepinephrine kinetics in younger and older men following caffeine ingestion. Eur. J. Clin. Nutr., 52, 805± 12. Barger-Lux, M.J. & Heaney, R.P. (1995) Caffeine and the calcium economy revisited. Osteoporosis Int., 5, 97±102. Barone, J.J. & Roberts, H.R. (1996) Caffeine consumption. Food Chem. Tox., 34, 119±29. Barr, H.M. & Streissguth, A.P. (1991) Caffeine use during pregnancy and child outcome: a 7-year prospective study. Neurotoxicol. Pharmacol., 13, 441±8. Barr, H.M., Streissguth, A.P., Martin, D.C. & Herman, C.S. (1984) Infant size at 8 months of age: relationship to maternal use of alcohol, nicotine, and caffeine during pregnancy. Pediatrics, 74, 336±41. Barrett-Connor, E., Chang, J.C. & Edelstein, S.L. (1994) Coffeeassociated osteoporosis offset by daily milk consumption. J. Am. Med. Assoc., 271, 280±83. Bolumar, F., Olsen, J., Rebagliato, M., Bisanti, L. & the European study group on infertility and subfecundity (1997) Caffeine intake and delayed conception: a European multicenter study on infertility and subfecundity. Am. J. Epidemiol., 145, 324±34. BorleÂe, I., Lechat, M.F., Bouckert, A. & Mission, C. (1978) Le cafeÂ, facteur de risque pendant la grossesse? Louvain MeÂd., 97, 279±84. Brent, R.L. (1998) A Systematic Evaluation of the Reproductive Risks of Caffeine. International Life Sciences Institute (ILSI) North America Publishers, Washington, DC. Brown, C.A., Bolton-Smith, C. & Woodward, M. (1993) Coffee and tea consumption and the prevalence of coronary heart disease in men and women: results from the Scottish heart health study. J. Epidemiol. Com. Health, 47, 171±5. Bruemmer B., White E., Vaughan, T.L. & Cheney, C.L. (1997)
180
Fluid intake and the incidence of bladder cancer among middle-aged men and women in a three-county area of western Washinton. Nutr. Cancer, 29, 163±8. Cavin, C., HolzhaÈuser D., Constable A., Huggett, A.C. & Schilter B. (1998) The coffee-specific diterpenes cafestol and kahweol protect against aflatoxin B1 -induced genotoxicity through a dual mechanism. Carcinogenesis, 19, 1369±75. Collaborative GESIA (Epidemiologic Group of the Italian Society of Alcohology) group for the Italian study on liver cirrhosis determinants (2000) Lifetime caffeine intake and the risk of liver cirrhosis: results from the Italian SIDECIR project. (in press). Cook, D.G., Peacock, J.L., Feyerabend, C. et al. (1996) Relation of caffeine intake and blood caffeine concentrations during pregnancy to fetal growth: prospective population based study. Br. Med. J., 313, 1358±62. Corrao, G., Lepore, A.R., Torchio, P. et al. (1994) The effect of drinking coffee and smoking cigarettes on the risk of cirrhosis associated with alcohol consumption. A case-control study. Provincial group for the study of chronic liver disease. Eur. J. Epidemiol., 10, 657±64. Daniels, J.W., MoleÂ, P.A., Shaffrath, J.D. & Stebbins, C.L. (1998) Effects of caffeine on blood pressure, heart rate, and forearm blood flow during dynamic leg exercise. J. Appl. Physiol., 85, 154±9. Debry, G. (1994) Coffee and Health, Gerard Debry, Edition John Libbey, Eurotext, Paris, France. Dlugosz, L., Belanger, K., Hellenbrand, K., Holford, T.R., Leaderer, B. & Bracken, M.B. (1996) Maternal caffeine consumption and spontaneous abortion: a prospective cohort study. Epidemiology, 7, 250±55. Donato, F., Boffetta, P., Fazioli, R., Aulenti, V., Gelatti, U. & Porru, S. (1997) Bladder cancer, tobacco smoking, coffee and alcohol drinking in Brescia, northern Italy. Eur. J. Epidemiol., 13, 795±800. Dumas, M., Gouyon, J.B., Tenenbaum, D., Michiels, Y., Escousse, A. & Alison, M. (1982) Systematic determination of caffeine plasma concentrations at birth in preterm and fullterm infants. Dev. Pharmacol. Ther., 4 (Suppl. 1), 182±6. Eggertsen, R., Andreasson, A., Hedner, T., Karlberg, B.E. & Hansson, L. (1993) Effect of coffee on ambulatory blood pressure in patients with treated hypertension. J. Intern. Med., 233, 351±5. Emory, E.K., Konopka, S. Hronsky, S., Tuggey, R. & Dave R. (1988) Salivary caffeine and neonatal behavior: assay modification and functional significance. Psychopharmacology, 94, 64±8. Fenster, L., Eskenazi, B., Windham, G.C. & Swan, S.H. (1991) Association of caffeine consumption during pregnancy and spontaneous abortion. Epidemiology, 2, 168±74. Fenster, L., Hubbard, A.E., Swan, S.H. et al. (1997) Caffeinated beverages, decaffeinated coffee, and spontaneous abortion. Epidemiology, 8, 515±23. Folsom, A.R., McKenzie, D.R., Disgard K.M., Kushi, L.H. & Sellers, T.A. (1993) No association between caffeine intake and post-menopausal cancer incidence in the Iowa women's health study. Am. J. Epidemiol., 138, 380±83.
Coffee: Recent Developments
Ford R.P.K., Schluter, P.J., Mitchell, E.A., Taylor, B.J., Scragg, R. & Stewart, A.W. (1998) Heavy caffeine intake in pregnancy and sudden infant death syndrome. New Zealand cot death study group. Arch. Dis. Childhood, 78, 9±13. Furlong, F.W. (1975) Possible psychiatric significance of excessive coffee consumption. Can. J. Psychiatry, 20, 577±83. Furuhashi, N., Sato, S., Suzuki, M., Hiruta, M., Tanaka, T. & Takahashi, T. (1985) Effects of caffeine ingestion during pregnancy. Gynecol. Obst. Invest., 19, 187±91. Gershbein, L.L. (1994) Action of dietary trypsin, pressed coffee oil, silymarin and iron salt on 1,2-dimethylhydrazine tumorigenesis by gavage. Anticancer Res., 14, 1113±16. Giovannucci, E. (1998) Meta-analysis of coffee consumption and risk of colorectal cancer. Am. J. Epidemiol., 147, 1043±52. Green, P.J., Kirby, R. & Suls, J. (1996) The effects of caffeine on blood pressure and heart rate: a review. An. Behav. Med., 18, 201±16. Greenland, S. (1993) A meta-analysis of coffee, myocardial infarction, and coronary death. Epidemiology, 4, 366±74. Griffiths, R.R., Evans, S.M. & Heishman, J. (1990) Low-dose caffeine discrimination in humans. J. Pharmacol. Exp. Ther., 2556, 1123±32. Grubben, M.J., Boers, G.H., Blom, H.J. et al. (2000) Unfiltered coffee increases plasma homocysteine concentrations in healthy volunteers: a randomized trial. Am. J. Clin. Nutr., 71, 480±84. Harris, S.S. & Dawson-Hughes, B. (1994) Caffeine and bone loss in healthy postmenopausal women. Am. J. Clin. Nutr., 60, 573± 8. Heaney, R.P. (1998) Effects of caffeine on bone and calcium economy, International Life Sciences Institute (ILSI) North America Publishers, Washington, DC. Heaney, R.P. & Recker, R.R. (1982) Effects of nitrogen, phosphorus, and caffeine on calcium balance in women. J. Lab. Clin. Med., 99, 46±55. Heaney R.P. & Recker, R.R. (1982) Determinants of endogenous fecal calcium in healthy women. J. Bone Min. Res., 9, 1621±7. Hinds, T.S., West, W.L., Knight, E.M. & Harland, B.F. (1996) The effect of caffeine on pregnancy outcome variables. Nutr. Rev., 54, 203±7. Hsieh, C.C., Thanos, A., Mitropoulos, D., Deliveliotis, C., Mantzoros C.S. & Trichopoulos, D. (1999) Risk factors for prostate cancer: a case-control study in Greece. Int. J. Cancer, 80, 699±703. Huggett, A.C., Cavin, C., HolzhaÈuser, D. & Schilter, B. (1997) Effect of coffee components on glutathione S-transferases: a potential mechanism for anticarcinogenic effects. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 43±50. ASIC, Paris, France. Inoue, M., Tajima, K., Hirose, K. et al. (1998) Tea and coffee consumption and the risk of digestive tract cancers: data from a comparative case-referent study in Japan. Cancer Causes Contr., 9, 209±16. Jacobson, M.F., Goldman, A.S. & Syme, R.H. (1981) Coffee and birth defects. Lancet, 1, 1415±16. Jain, M.G., Hislop, G.T., Howe, G., Burch, J.D. & Ghadirian, P. (1998) Alcohol and other beverage use and prostate cancer risk among Canadian men. Int. J. Cancer, 78, 707±11.
Health Effects and Safety Considerations
Jee, S.H., He, J., Whelton, P.K., Suh, I. & Klag, M.J. (1999) The effects of chronic coffee drinking on blood pressure: a metaanalysis of controlled clinical trials. Hypertension, 33, 647±52. Johansson, L., Drevon, C.A. & Aa Bjorneboe, G.-E. (1996) The Norwegian diet during the last years in relation to coronary heart disease. Eur. J. Clin. Nutr., 50, 277±83. Kark, J.D., Friedlander, Y., Kaufmann, N.A. & Stein, Y. (1985) Coffee, tea, and plasma cholesterol: the Jerusalem lipid research clinic prevalence study. Br. Med. J., 291, 699±704. Kawachi, I., Willet, W.C., Colditz, G.A., Stampfer, M.J. & Speizer, F.E. (1996) A prospective study of coffee drinking and suicide in women. Arch. Intern. Med., 156, 521±5. Klatsky, A. & Amstrong M. (1992) Alcohol, smoking, coffee and cirrhosis. Am. J. Epidemiol., 136, 1248±57. Klatsky, A.L., Amstrong, M.A. & Friedman, G.D. (1993) Coffee, tea, and mortality. Ann. Epidemiol., 3, 375±81. Klebanoff, M.A., Levine, R.J., Dimonian, R., Clemens, J.D. & Wilkins, D.G. (1999) Maternal serum paraxanthine, a caffeine metabolite, and the risk of spontaneous abortion. N. Engl. J. Med., 341, 1639±44. Kynast-Gales, S.A. & Massey, L.K. (1994) Effect of caffeine on circadian excretion of urinary calcium and magnesium. J. Am. Coll. Nutr., 13, 467±72. Leviton A. (1990) Methylxanthine consumption and the risk of ovarian malignancy. Cancer Lett., 51, 91±101. Leviton, A., Pagano, M., Allred, E.N. & El Lozy, M., (1994) Why those who drink the most coffee appear to be at increased risk of disease. A modest proposal. Ecol. Food Nutr., 31, 285±93. Lloyd, T., Rollings, N., Eggli, D.F., Kieselhorst, K. & Chinvilli, V.M. (1997) Dietary caffeine intake and bone status of postmenopausal women. Am. J. Clin. Nutr., 65, 1826±30. Lloyd, T., Rollings, N.J., Kieselhorst, K., Eggli, D.F. & Mauger, E. (1998) Dietary caffeine intake is not correlated with adolescent bone gain. J. Am. Coll. Nutr., 17, 454±7. MacDonald, T.M., Sharpe, K., Fowler, G. et al. (1991) Caffeine restriction: effect on mild hypertension. Br. Med. J., 303, 1235±8. Maruoka, K., Yagi, M., Akawaza, K., Kinukawa, N., Ueda, K. & Nose, Y. (1998) Risk factors for low birthweight Japanese infants. Acta Paedriatica, 87, 304±9. Massey, L.K., Bergman, E.A., Wise, K.J. & Sherrard, D.J. (1994) Interactions between dietary caffeine and calcium on calcium and bone metabolism in older women. J. Am. Coll. Nutr., 13, 592±6. Massey, L.K. & Wise, K.J. (1984) The effect of dietary caffeine on urinary excretion of calcium, magnesium, sodium and potassium in healthy young females. Nutr. Res., 4, 43±50. Meleady, R. & Graham, I. (1999) Plasma homocysteine as a cardiovascular risk factor: causal, consequential, or of no consequence. Nutr. Rev., 57, 299±305. Mino, Y., Yatsuda, N., Fujimura, T. & Ohara, H. (1990) Caffeine consumption and depressive symptomatology among medical students. Jpn. J. Alcohol Drug Depend., 25, 486±96. Myers, M.G. (1988) Effects of caffeine on blood pressure. Arch. Intern. Med., 148, 1189±93. Myers, M.G. (1991) Caffeine and cardiac arrhythmias. Ann. Intern. Med., 114, 147±50.
181
Myers, M.G. (1992) Coffee and coronary heart disease. Arch. Intern. Med., 152, 1767±72. Myers, M.G. (1998) Cardiovascular effects of caffeine. International Life Sciences Institute (ILSI) North America Publishers, Washington, DC. Nagao, M., Fujita, Y., Wakabayashi, K., Nukaya, H., Kosuge, T. & Sugimura, T. (1986) Mutagens in coffee and other beverages. Environ. Health Perspect., 67, 89±91. Narod, S.A., de SanjoseÂ, S. & Victoria, C. (1991) Coffee during pregnancy: a reproductive hazard? Am. J. Obst. Gynecol., 164, 1109±14. Nehling, A. & Debry, G. (1994a) Potential genotoxic, mutagenic and antimutagenic effects of coffee: a review. Mut. Res., 317, 145±62. Nehling, A. & Debry, G. (1994b) Potential teratogenic and neurodevelopmental consequences of coffee and caffeine exposure: a review on human and animal data. Neurotoxicol. Teratol., 16, 531±43. Nehling, A. & Debry, G. (1994c) Effets du cafe et de la cafeÂine sur la fertiliteÂ, la reproduction, la lactation et le deÂveloppement. J. Gynecol. Obstet. Biol. Reprod., 23, 241±56. Nehling A. & Debry G. (1996) Coffee and cancer: a review of human and animal data. World Rev. Nutr. Diet, 79, 185±221. Newby, D.E., Neilson, J.M., Jarvies, D.R. & Boon, N.A. (1996) Caffeine restriction has no role in the management of patients with symptomatic idiopathic ventricular premature beats. Heart, 76, 355±7. Nieto, F.J. & Comstock, G.W. (1997) Coffee consumption and plasma homocysteine: results from the atherosclerosis risk in communities study. Am. J. Clin. Nutr., 66, 1475±6. Nilssen, O. & Forde, O.H. (1994) Seven-year longitudinal population study of change in gamma-GT: the Tromso study. Am. J. Epidemiol., 139, 787±92. Nishi, M., Ohba, S., Hirata, K. & Miyake, H. (1996) Doseresponse relationship between coffee and the risk of pancreas cancer. Jpn. J. Oncol., 26, 42±8. Nishikawa, A., Tanaka, T. & Mori, H. (1986) An inhibitory effect of coffee on nitrosamine-hepatocarcinogenesis with aminopyrine and sodium nitrite in rats. J. Nutr. Growth Cancer, 3, 161±6. Nygard, O., Refsum H., Ueland, P.-M. et al. (1997a) Coffee consumption and plasma total homocysteine: the hordaland homocyteine study. Am. J. Clin. Nutr., 65, 136±43. Nygard, O., Refsum, H., Ueland, P.M. & Vollset, S.E. (1998) Major lifestyle determinants of plasma total homocysteine distribution: the Hordaland homocysteine study. Am. J. Clin. Nutr., 67, 263±70. Nygard, O., Vollset, S.E., Refsum, H. & Ueland, P.M. (1997b) Reply to F.J. Nieto et al. Am. J. Clin. Nutr., 66, 1476±7. Obana, H., Nakamura, S. & Tanaka, T. (1986) Suppressive effects of coffee on the SOS responses induced by UV and chemical mutagens. Mut. Res., 175, 47±50. Olsen, J., Overvad, K. & Frische, G. (1991) Coffee consumption, birthweight, and reproductive failures. Epidemiology, 2, 370±74. Oshaug, A., Bugge, K.H. & Refsum H. (1998) Diet, an independent determinant for plasma total homocysteine. A cross sectional study of Norwegian workers on platforms in the North Sea. Eur. J. Clin. Nutr., 52, 7±11.
182
Packard, P.T. & Recker, R.R. (1996) Caffeine does not affect the rate of gain in spine bone in young women. Osteoporosis Int., 6, 149±52. Palmer, J.R., Rosenberg, L., Sowmya, R. & Shapiro, S. (1995) Coffee consumption and myocardial infarction in women. Am. J. Epidemiol., 141, 724±31. Polychronopoulou, A., Tzonou, A., Hsieh, C. C. et al. (1993) Reproductive variables, tobacco, ethanol, coffee and somatometry as risk factors for ovarian cancer. Int. J. Cancer, 55, 402± 407. Probert, J.L., Persad, R.A., Greenwood, R.P., Gillatt, D.A. & Smith, P.J.B. (1998) Epidemiology of transitional cell carcinoma of the bladder: profile of an urban population in the south-west of England. Br. J. Urol., 82, 660±66. Rondo, P.H.C., Rodrigues, L.C. & Tomkins, A.M. (1996) Coffee consumption and intrauterine growth retardation in Brazil. Eur. J. Clin. Nutr., 50, 705±709. Schilter, B., Perrin, I., Cavin, C. & Huggett, A.C. (1996) Placental gluthatione S-transferase (GST-P) induction as a potential mechanism for the anti-carcinogenic effect of the coffeespecific components cafestol and kahweol. Carcinogenesis, 17, 2377±84. Sesso, H.D., Gaziano, J.M., Buring, J.E. & Hennekens, C.H. (1999). Coffee and tea intake and the risk of myocardial infarction. Am. J. Epidemiol., 149, 162±7. Silverman, D.T., Swanson, C.A., Gridley, G. et al. (1998) Dietary and nutritional factors and pancreatic cancer: a case-control study based on direct interviews. J. Nat. Cancer Inst., 90, 1710±19. Smith, A. (1998) Effects of caffeine on human behavior. International Life Sciences Institute (ILSI) North America Publishers, Washington, DC. Smith, S.J., Deacon, J.M. & Chilvers, C.E.D. (1994) Alcohol, smoking, passive smoking and caffeine in relation to breast cancer in young women. Br. J. Cancer, 70, 112±19. Stadler, R., Bexter, A., WuÈrzner, H.P. & H. LuginbuÈhl (1990) A carcinogenicity study of infant coffee in swiss mice. Food Chem. Tox., 28, 829±30. Stadler, R.H., Richoz, J., Turesky, R.J., Welti, D.H. & Fay, L.B. (1995) Oxidation of caffeine and related methylxanthines in ascorbate and polyphenol-driven Fenton-type oxidations. Free Rad. Res., 24, 225±40. Stadler, R.H., Turesky, R.J., MuÈller, O., Markovic, J. & LeongMorgenthaler P.-M. (1994) The inhibitory effects of coffee on radical-mediated oxidation and mutagenicity. Mut. Res., 308, 177±90. Stanton, C.K. & Gray, R.H. (1995) Effects of caffeine consumption on delayed conception. Am. J. Epidemiol., 142, 1322± 9. Stein, Z. & Susser, M. (1991) Miscarriage, caffeine and the epiphenomena of pregnancy: the causal model. Epidemiology, 2, 163±7. Stelzer, K.J., Koh, W.J., Kurtz, H., Greer, B.E. & Griffin, T.W. (1994) Caffeine consumption is associated with decreased severe late toxicity after radiation to the pelvis. Int. J. Radiat. Oncol. Biol. Phys., 30, 411±17. Stensvold, I. & Jacobsen, B.J. (1994) Coffee and cancer: a
Coffee: Recent Developments
prospective study of 43 000 Norwegian men and women. Cancer Causes Contr., 5, 401±408. Stensvold, I., Tverdal, A. & Foss, O.P. (1989) The effects of coffee on blood lipids and blood pressure: results from a Norwegian cross-sectional study, men and women, 40±42 years. J. Clin. Epidemiol., 42, 877±84. Stensvold, I., Tverdal, A. & Jacobsen, B. (1996) Cohort study of coffee intake and death fron coronary heart disease over 12 years. Br. Med. J., 312, 544±5. Stephenson, P.E. (1977) Physiologic and psychotropic effects of caffeine in man. J. Am. Dietetic. Assoc., 71, 240±44. Stich, H.F., Risin, M.P. & Bryson, L. (1982) Inhibition of mutagenicity of a model nitrosation reaction by naturally occuring phenolics, coffee and tea. Mut. Res., 259, 307±24. Stolzenberg-Solomon, R.Z., Miller E.R., Maguire M.G., Selhub, J. & Appel, L.J. (1999) Association of dietary protein intake and coffee consumption with serum homocysteine concentrations in an older population. Am. J. Clin. Nutr., 69, 467±75. Streissguth, A.P., Barr, H.M., Martin, D.C. & Herman, C.S. (1980) Effects of maternal alcohol, nicotine, and caffeine use during pregnancy on infant mental and motor development at eight months. Alcoholism Clin. Exp. Res., 4, 152±64. Tanaka, K., Tokunaga, S., Kono, S. et al. (1998) Coffee consumption and decreased serum gamma-glutamyltransferase and aminotransferase activities among male alcohol drinkers. Int. J. Epidemiol., 27, 438±43. Tavani, A. & La Vechia, C. (1997) Epidemiology of renal-cell carcinoma. J. Nephrol., 10, 93±106. Tavani, A., Pregnolato, A., La Vecchia, C., Favero, A. & Franceschi, S. (1998) Coffee consumption and the risk of breast cancer. Eur. J. Cancer Prevent., 7, 77±82. Thelle, D.S., Heyden, S. & Fodor, J.G. (1987) Coffee and cholesterol in epidemiological and experimental studies. Artherosclerosis, 67, 97±103. Toubas, P.L., Duke, J.C., McCaffree, M.A., Mattice, C.D., Bendell, D. & Orr, W.C. (1986) Effects of maternal smoking and caffeine habits on infantile apnea: a retrospective study. Pediatrics, 78, 159±63. Tuomilehto, J. & Pietinen, P. (1991) Coffee and cardiovascular disease. Cardiovasc. Risk Factors, 1, 165±73. Tverdal, A., Stensvold, I., Solvoll, K., Foss, O.P., Lund-Larsen, P. & Bjartweit, K. (1990) Coffee consumption and death from coronary heart disease and mortality in middle-aged Norwegian men and women. Br. Med. J., 300, 566±9. Ugert, R. & Katan, M.B. (1997) The cholesterol-raising factor from coffee beans. Ann. Rev. Nutr., 17, 305±24. Ugert, R., Van der weg, G., Kosmeijer-Schuil, T.G., Van Bovenkamp, P., Hovenier, R. & Katan, M. (1995) Levels of the cholesterol-elevating diterpenes cafestol and kahweol in various coffee brews. J. Agr. Food Chem., 43, 2167±72. Van Dusseldorp, M., Katan, M., Van Vliet, T., Demacker, P.N. M. & Stalenhoef, A. (1991) Cholesterol-raising factor from boiled coffee does not pass a paper filter. Arterioscler. Thromb., 11, 586±93. Viscoli, C.M., Lachs, M.S. & Horwitz, R.I. (1993) Bladder cancer and coffee drinking: a summary of case-control research. Lancet, 341, 1432±7.
Health Effects and Safety Considerations
Vollset, S.E., Nygard, O., Refsum, H. & Ueland P.M. (2000) Coffee and homocysteine. Am. Clin. Nutr., 71, 403±404. Weiderpass, E., Partanen, T., Kaaks, R. et al. (1998) Occurrence, trends and environmental etiology of pancreatic cancer. Scand. J. Work Environ. Health, 24, 165±74. Weusten-Van der Wouw, M.P.M.E., Katan, M., Viani, R. et al. (1994) Identity of the cholesterol-raising factor from boiled coffee and its effects on liver function enzymes. J. Lipids Res., 35, 721±33. Wilcox, A., Weinberg, C. & Baird, D. (1988) Caffeinated beverages and decreased fertility. Lancet, 2, 1453±6. Willett, W.C., Stampfer, M.J., Manson, J.E. et al. (1996) Coffee consumption and coronary heart disease in women, a ten-year follow-up. J. Am. Med. Assoc., 276, 458±62.
183
Woodward M. and Tunstall-Pedoe, H. (1999) Coffee and tea consumption in the Scottish heart health study follow-up: conflicting relations with coronary risk factors, coronary disease, and all cause mortality. J. Epidemiol. Community Health, 53, 481±7. World Health Organisation International Agency for Research on Cancer (WHO/IARC) (1991) Coffee, Tea, Mate, Methylxanthines and Methylglyoxal. In: IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, 51, 47±206. IARC, Lyon, France.
Chapter 9
Agronomy I: Coffee Breeding Practices Herbert A.M. Van der Vossen Plant Breeding & Seed Consultant Venhuizen, the Netherlands 9.1 INTRODUCTION 9.1.1 World production increase World coffee production ± about 70% arabica (Coffea arabica) and 30% robusta (C. canephora) coffees ± continues to show large annual fluctuations, but has generally increased by about 14% over the past 15 years: from 5.2 million tonnes per year averaged over the years 1980±84, to 5.9 million over the period 1995± 9 (ICO, 1990±99). There was a very low production of 4.8 million t in 1986±7 and a record of 6.4 million t in the 1998±9 crop year. As usual, the tremendous variation in annual coffee output by the leading producer Brazil ± from 19% to 33% of world coffee production, mainly as a result of recurring abiotic calamities (frosts and droughts) ± had a major impact on fluctuations in world coffee supplies and market prices. About 61% of the world coffee was produced in Latin America, 18% in Africa and 21% in Asia during the 1998±9 crop year. Of particular interest is the accelerated expansion of robusta coffee production in Vietnam and Indonesia to more than 400 000 t. Brazil produced in that year some 300 000 t robusta coffee in addition to its 1.8 million t arabica crop and is rapidly overtaking the traditionally leading robusta producers Ivory Coast and Uganda. India has almost doubled its annual production during the last decade and may soon reach an annual output of 300 000 t of high quality arabica (40%) and robusta (60%) coffees.
9.1.2 Selection and breeding before 1985 Reviews on the history and progress of selection and breeding for arabica and robusta coffees until the mid1980s have been presented by Van der Vossen (1985), Carvalho (1988), Charrier & Berthaud (1988) and Wrigley (1988). Bettencourt & Rodrigues (1988) produced a review specifically on disease resistance
breeding and Cambrony (1988) on interspecific hybridisation in coffee. The following summary of major advances in breeding and variety development, as reviewed by these authors, may serve as a useful background to the remaining paragraphs of this chapter. Much of the world coffee is still produced by traditional cultivars released some five to eight decades ago from relatively simple selection and breeding programmes and generally multiplied by seed. Cultivars of the self-pollinating arabica coffee are truebreeding lines from single-plant selections in growers' fields, or from progenies of simple crosses and backcrosses; while those of the outbreeding robusta are open-pollinated cultivars produced from selected seedling and bi- or polyclonal gardens. Clonal robusta cultivars have found limited application so far, except in plantation coffee in Indonesia and the Ivory Coast, largely because the logistics of mass propagation and distribution are too complex and expensive for smallholder production systems, which dominate coffee production. Yield, plant vigour and quality have been the main selection criteria in both coffee types, but host resistance to the destructive coffee leaf rust (CLR) disease has already been given high priority in arabica coffee breeding in India since the 1920s. Breeding programmes with systematic crossing designs and statistically laid out field trials implemented during the last 30 years have provided opportunities for biometrical genetic analyses of yield and other agronomic characters. There is considerable evidence of predominantly additive genetic variance for almost all components of yield, quality and other quantitative traits, except sometimes the complex factor yield itself, both for arabica as well as robusta coffee. This should facilitate the estimation of parental breeding values from relatively simple progeny testing and increase selection progress. Hybrid vigour for yield noticed in crosses between parents of different origins
184
Agronomy I: Coffee Breeding Practices
appears to be the result of accumulation of complementary polygenes dispersed over subpopulations. Some breeding programmes in robusta coffee (e.g. in the Ivory Coast) have already adopted methods of reciprocal recurrent selection with distinct subpopulations to increase chances of producing genotypes superior in yield, quality and other important traits. In arabica coffee, disease resistance is a breeding objective of the highest priority. Efforts to obtain durable resistance to CLR have had a long history of initial successes followed by disappointments because of the repeated appearance of new virulent races of the rust fungus, but some lines of the cultivar Catimor (selected from crosses between Caturra and Hibrido de Timor) have shown complete resistance in most countries. These results were obtained by breeding plans normally applied to self-pollinating crops, including recombination crosses followed by backcrossing, inbreeding and pedigree selection. A similar plan was initially also applied in a breeding programme in Kenya to obtain resistance to coffee berry disease (CBD), which turned out to be controlled by a few major genes but nevertheless also durable. The change of breeding strategies to produce F1 hybrid (seed) cultivars instead of clones or true breeding lines was partly inspired by the confirmation of transgressive hybrid vigour in genetically divergent crosses in arabica coffee. Other advantages were chances of earlier introduction of cultivars with resistance to both CBD and CLR, as well as several other desirable agronomic characters. Interspecific hybridization has played a significant role in coffee, such as crosses between arabica and robusta coffee with the objective of introgressing disease resistance into arabica (for example the cultivar Icatu in Brazil) or improved liquor quality into robusta coffees (for example the variety Arabusta in the Ivory Coast). Other examples of interspecific hybridization leading to successful cultivars in arabica and in robusta coffee can be found in India.
9.1.3 New developments Modern arabica cultivars with higher yield potential and resistance to important diseases (CLR and CBD) have started to replace traditional varieties on a large scale in several countries: for example Catimor and Sarchimor types of cultivars in Colombia, Brazil, Central American countries and India, Icatu in Brazil, Java in Cameroon, Ruiru II in Kenya and Ababuna in Ethiopia (the latter two being F1 hybrids). In robusta
185
coffee the release of new cultivars from advanced selection programmes is taking place more gradually, such as the BP and SA clones in Indonesia, the BR (seed) cultivars in India, the IF clones in the Ivory Coast and the cultivar Apoata in Brazil. The high expectations of the Arabusta programme in the Ivory Coast have not been fulfilled, because of persistent problems of genetic instability and low fertility. On the other hand, the `C 6 R' variety of India, which arose from a cross between C. congensis and C. canephora, has proved to be a success as a productive and stable robusta coffee with superior bean and liquor characteristics. The breakdown of resistance to CLR in Catimor lines in India, the sudden reappearance of a wilt disease (tracheomycosis) in robusta coffee in DR Congo and Uganda, increasing nematode problems in arabica coffee in Central America and the arrival of the coffee berry borer in Colombia (1988) and India (1990) are just a few examples of new challenges to national coffee industries, which need to be met also by innovative selection and breeding programmes. Plant biotechnology has evolved, particularly during the past decade, into an applied science providing powerful additional tools for plant breeding with the potential of increasing selection efficiency and creating new approaches to hitherto unattainable objectives. This so-called molecular breeding has basically two main applications of plant biotechnology: molecular markers and transgenic plants. In coffee, molecular marker technology has already been implemented in germplasm characterisation and management, detecting genetically divergent breeding subpopulations (for example to predict hybrid vigour), establishing gene introgression from related species and molecular marker-assisted selection (Lashermes et al., 1996a, 1997a; Charrier & Eskes, 1997). Generally, successful genetic transformation is still limited to characters controlled by major genes for which gene isolation and transfer is relatively easy. Techniques of regenerating plants from in vitro micropropagation and somatic embryogenesis are by now well established for various coffee species and transgenic coffee plants have been produced already, for example with insect resistance and with caffeinefree beans (see Chapters 10 and 11). However, lack of adequate legislation in some countries for proprietary rights and biosafety, as well as a negative public perception of biotechnology, can be temporary obstacles to the introduction and unrestricted cultivation of transgenic coffee cultivars. While formerly, breeding programmes were often
186
carried out in relative isolation at the various coffee research centres, the 1990s saw increased networking between a number of them and also with renowned agricultural research centres in Europe and the USA to implement collaborative research projects on genetic resources, resistance breeding and plant biotechnology.
9.2 GENETIC RESOURCES 9.2.1 World collections The first systematic efforts to collect Coffea arabica germplasm by an FAO mission to Ethiopia in 1964 had the real intention of international collaboration and the resulting 623 accessions were distributed worldwide to several coffee research centres (Meyer et al., 1968). A second expedition to the south-western highlands of Ethiopia mounted by the IRD (ex ORSTOM) in 1966 produced another 70 new accessions of C. arabica. The IRD (ex ORSTOM), sometimes in collaboration with IPGRI, made several collecting expeditions for a large number of other Coffea species between 1960 and 1985 to important centres of genetic diversity in Guinea, the Ivory Coast, Central Africa, Cameroon, Congo, East Africa and Madagascar. Most of these accessions are maintained as base field collections in the Ivory Coast and Madagascar (Berthaud & Charrier, 1988). More recent efforts of coffee germplasm collection include a mission to Yemen for C. arabica (Eskes & Mukred, 1989) and one to north-west Tanzania for C. canephora (Nyange & Marandu, 1997). There is a considerable amount of ex situ germplasm of C. arabica collected and maintained in Ethiopia since 1966 (Bellachev, 1997). Altogether about 100 species (taxa) of the genus Coffea have been identified so far (Bridson & Verdcourt, 1988). They are without exception indigenous to the forests of tropical Africa and all are diploid (2n = 22) species (C. canephora, C. congensis, C. liberica, C. eugenioides, C. stenophylla, C. racemosa, C. zanguebariae, etc. and some 50 taxa belonging to the section Mascarocoffea) except the allotetraploid (2n = 44) C. arabica, which has its origin in the highland forests of south-west Ethiopia. Progressive crop improvement requires easy access to intra- and interspecific genetic variation. The monophyletic origin of all Coffea species and general absence of strong interspecific crossing barriers (Charrier & Berthaud, 1985; Charrier & Eskes, 1997) provide opportunities to exploit as well the genetic variation of several other species for the purpose of
Coffee: Recent Developments
introgressing agronomically and biochemically interesting characters into the two species of commercial value, C. arabica and C. canephora. The three main base collections in Ethiopia, the Ivory Coast and Madagascar are, therefore, of vital importance to coffee breeding in general and should receive adequate support for maintenance and further systematic exploration for new germplasm by an internationally recognised network organization for the conservation and study of coffee genetic resources with the participation of all major coffee producing countries. IPGRI in conjunction with IRD, CIRAD and the African Coffee Research Network (ACRN) has taken initiatives in that direction (Guarino et al., 1995; Ngategize, 1997), but free exchange of coffee germplasm between producing countries will continue to be restricted until the formal establishment of a network, such as the one already existing for cocoa (Eskes et al., 1998). Hamon et al. (1998) proposed strategies to manage such large germplasm collections more efficiently and increase their accessibility by identification of core collections, which are representative of specifically desired genetic diversity without undue duplication. The identification procedure involves the application of so-called principal component score strategies. Dulloo et al. (1998) have described strategies for in situ conservation of Coffea species. In addition to the three earlier mentioned base collections, many research centres in coffee-producing countries maintain duplicate field collections of cultivated and wild coffee germplasm as genetic resources for their breeding programmes. Table 9.1 presents an overview of location, type and size of major field collections of coffee germplasm in the world. Accessions often contain several genotypes, especially when originally collected as a seed sample from wild coffee trees. Published reports on coffee germplasm collections sometimes give rather inflated numbers, because the distinction between an accession and a genotype is not always clearly made. In Table 9.1 adjustments have been made for number of accessions, whenever such information was available.
9.2.2 Species relationships Various recent studies of genetic diversity and phylogenetic relationships within the genus Coffea have applied molecular marker technologies to chloroplast and nuclear DNA extracted from several coffee species (Orozco-Castillo et al., 1994, 1996; Lashermes et al., 1995, 1996b, 1997a, 1999a; Cros et al., 1998). These molecular markers have advantages over the
Agronomy I: Coffee Breeding Practices
Table 9.1
187
Field collections of coffee genetic resources. Germplasm of
Country
Organisation
Base collections Ethlopia BCRI/E EIAR/JARC
Location
Coffea arabica
Choche Jimma, Gera
6 6 6
Coffea canephora
Other species
Mascarocoffea
Total number of accessions 800 2000
Ivory Coast
CNRA
Divo Man
6
Madagascar
FOFIFA
Kianjavato Ilaka-Est Sahambavy
6
Work collections Costa Rica CATIE
Turrialba
6
6
[24]
6 6
1200 200 6
[7]
800 1200 350 1700
Brazil
IAC
Campinas
6
6
[9]
310
Colombia
CENICAFE
Chinchina
6
6
[10]
980
Kenya
CRF
Ruiru
6
[5]
500
Tanzania
TARO
Lyamungu
6
6
[10]
300
Cameroon
IRAD
Foumbot, Santa Nkoumvane
6
6
[3]
100 50
Indonesia
ICCRI
Jember, etc.
6
6
6
1000
India
CCRI
Chickmagalur
6
6
[18]
360
Note: [24] = number of other species. Sources: Dulloo et al. (1998); IPGRI Directory of coffee germplasm collections (1999). 6 = available at location.
morphological and biochemical characters measured in conventional taxonomic analyses, because they are more polymorphic and unaffected by environmental influences. Some of the main conclusions can be summarized as follows: . The results based on molecular markers produce dendrograms of relationships of coffee species very similar to those of conventional taxonomic studies. They confirm a fairly recent African origin of the genus Coffea and subsequent ecological differentiation into numerous species, which are clustered in groups of genetic relationships corresponding to geographic regions (Table 9.2). However, the process of differentiation has not yet progressed into a stage of strong genetic barriers, as is shown for instance by successful interspecific hybridization between the Mascarocoffea and other groups of mainland Africa. The earlier taxonomic
classifications of the genus Coffea into sections and subsections have, therefore, been abolished. . There is molecular evidence for fairly close genetic relations between the genera Coffea and Psilanthus, as distinguished by Bridson & Verdcourt (1988) to suggest taxonomic revision into a single genus, Coffea. This is supported also by Couturon et al. (1998), who achieved viable hybrids from crosses between P. ebracteolatus (at the tetraploid level) and C. arabica. . Species very closely related or identical to C. eugenioides and to C. canephora (or C. congensis) are indeed the most likely maternal and paternal progenitors, respectively, of the allotetraploid C. arabica. Raina et al. (1998) arrived at similar conclusions in a cytogenetic study of C. arabica using genomic and fluorescent in situ hybridization techniques. Segregation analysis with molecular (RFLP) markers confirmed earlier cytogenic evi-
188
Table 9.2
Coffee: Recent Developments
Geographic origin of some Coffea species in Africa.
Species C. arabica C. eugenioides C. canephora C. congensis C. liberica C. humilis C. stenophylla C. brevipes C. racemosa C. salvatrix C. zanguebariae C. fadenil Mascarocoffea (50 species)
West
G 6 6 6
Central I (Atlantic)
Central II
C 6 6 6
6 C 6 6
Ethiopla
East
Madagascar
6
6
6 6 6 6
6
Note: G = Guinean, C = Congolese subpopulations of C. canephora. Adapted from Berthaud & Charrier (1988). 6 = available at location.
dence for regular disomic meiotic behaviour in C. arabica, probably under the control of pairing regulating genes (Lashermes et al., 2000a). . The existence of two subgroups of partial genetic differentiation within germplasm of C. arabica was established by an analysis with molecular (RAPD) markers (Lashermes et al., 1996c) and also by a multivariate analysis of phenotypic characters (Montagnon & Bouharmont, 1996): (a) all accessions collected in Ethiopia, west of the Rift Valley (Kaffa, Illubabor, Wollega) and (b) accessions collected east of the Rift Valley (Sidamo, Hararge) and the cultivated varieties. In this perspective it would appear that the coffee cultivated in the Yemen, from where almost all cultivated varieties of C. arabica derive, had its origin in Ethiopia east of the Rift Valley. . The existence of genetically diverse subpopulations (Congolese and Guinean) has also been confirmed within C. canephora by means of characterization with morphological, biochemical (isozymes) and molecular markers (Berthaud, 1985; Leroy et al., 1993). . Genetic variation of C. arabica populations is much enhanced by introgressive breeding with C. canephora genotypes. The arabica-like variety Hibrido de Timor is derived from a natural cross between C. arabica and C. canephora. Molecular studies with AFLP technology indicated that the genetic variation in Catimor and Sarchimor lines (derivatives
of crosses with Hibrido de Timor) was almost double that observed in traditional arabica cultivars or accessions from Ethiopia (Lashermes et al., 2000b).
9.2.3 Conservation In coffee there have been no alternatives so far to ex situ field collections for long-term germplasm conservation, because coffee seeds are recalcitrant and conventional methods of seed storage cannot extend viability beyond 2 to 3 years (Van der Vossen, 1985). Field collections require expensive resources of land, qualified staff and upkeep, while there is also a risk of losing valuable germplasm due to diseases and pests, as well as to poor adaptation of certain species to the local environment. By applying slow-growth conditions to in vitro cultured explants (zygotic embryos, apical meristem or nodal cuttings) and repeated subculturing, Dussert et al. (1997a) were able to conserve a core collection of coffee germplasm for about 3 years. However, there appeared to be some risk of genetic drift (random loss of genetic variability), since accessions differed considerably in their survival rates after several subcultures. Such methods of in vitro conservation have great advantages for germplasm distribution (less volume during shipping and simple quarantine procedures), but appear unsuitable for long-term germplasm conservation.
Agronomy I: Coffee Breeding Practices
Encouraging results of cryopreservation techniques (storage under liquid nitrogen at -1968C) applied to coffee seeds have been reported recently (Dussert et al., 1997b). Expanding on methods developed by Abdelnour-Esquivel et al. (1992) for zygotic coffee embryos, Dussert et al. (1997c) succeeded in retrieving high rates of viable embryos from cryopreserved whole seeds of C. arabica. Rather low rates of normal seedling development after initially high seed germination indicated damage to the endosperm during the freezing and thawing stages of the process. However, in vitro cultured embryos, excised from the seeds after cryopreservation, survived and produced a very high rate of normal seedlings. Specific conditions of seed dehydration and precooling before and rapid thawing after cryopreservation, as well as the Coffea species, are all important factors influencing the retrieval rate of viable seedlings after storage (Dussert et al., 1998, 1999). Much additional experimental work may still be needed to solve remaining problems, but there is little doubt at this stage that cryopreservation opens interesting perspectives for long-term conservation of coffee germplasm in seed gene banks.
9.3 BREEDING
189
Table 9.3
Major selection criteria in coffee.
Criteria
Priorities Arabica
Robusta
3 3 3 3
3 3 3 2
Quality Bean size and shape Liquor quality Caffeine content
3 3 1
1 2 2
Host resistance to diseases Coffee leaf rust Coffee berry disease (Africa only) Other diseases
3 3 1
1 Ð 1
Host resistance to pests Nematodes Leaf miners Coffee berry borer Stem borers
3 2 2 2
1 1 2 1
Drought tolerance
1
1
Productivity Yield: kg per plant and per hectare Yield stability Plant vigour Compact plant type (short internodes)
Note: 1 = low, 3 = high breeding priority.
9.3.1 General objectives and strategies Arabica and robusta coffee breeding programmes have the same main objective of developing new cultivars, which have the potential of yielding optimum economic returns to coffee growers. An overview of selection criteria applied in coffee breeding (Table 9.3) indicates equal importance for factors of productivity in both species, but higher priority for bean size and liquor quality, as well as for host resistance to major diseases and pests in arabica coffee. Variations in the circumstances of climate, soil, biotic and abiotic stresses, cropping systems, socio-economic factors, market dynamics and consumer preferences further define priorities of selection criteria applied in specific programmes. Methods applied in breeding and variety propagation depend primarily on the mating systems of arabica (inbreeding) and robusta (outbreeding) coffee. Outlines of coffee breeding schemes have been discussed in detail elsewhere (Van der Vossen, 1985; Charrier & Berthaud, 1988) and Table 9.4 presents just a summary of actual methods implemented in various coffee research centres, together with examples of released cultivars. Four basic methods of breeding and selection
can be distinguished in each of the two species. These are listed in order of increasing complexity from line or mass selection to intra-and interspecific hybridization, the application depending on breeding objectives and intended output.
9.3.2 Productivity Some coffee breeding centres now emphasize hybrid varieties as the best strategy for further and more rapid increases of plant productivity. In arabica coffee, 30 to 60% heterosis in yield over the better parent has been observed in Ethiopia (Ameha, 1990), Cameroon (Cilas et al., 1998) and Central America (Bertrand et al., 1997), which confirms earlier reports from Kenya (Walyaro, 1983) and India (Srinivasan & Vishveshwara, 1978). Coffee hybrids were also found to have greater yield stability over location and time (fewer genotype 6 environment interaction effects). Chances of substantial (transgressive) hybrid vigour are increased by combining parents selected from genetically diverse subpopulations, such as crosses between
Varieties, clones
Varieties, clones (distinction of two sub-populations)
Arabica variety, tertaploid robusta genotypes C. congensis accession, robusta genotypes
(6) Family and clonal selection
(7) Reciprocal recurrent selection
(8a) Interspecific hybridization (arabica 6 robusta), family and clonal selection
(8b) Interspecific hybridization (C. congensis 6 robusta), backcrossing to robusta and family selection
Local or introduced varieties and accessions
Arabica varieties, tetraploid/diploid robusta genotypes
(4) Interspecific hybridization (arabica 6 robusta), backgrossing and pedigree selection
Robusta (5) Mass selection (individual plants)
Composite hybrid F1 hybrid F1 clone
Crossing and selfing
Varieties/ accessions, pedigrees of crosses
(3) Intraspecific F1 hybrids
Crossing, backcrossing, sibmating
Crossing and OP
Bi-parental crossing for inter-group combining ability tests and intra-group recombination; + doubled heploids
OP (half-sib families)
OP (open pollination)
Crossing and selfing
Line
Crossing and selfing
Varieties
(2) Pedigree selection after hybridization (sometimes also backgrossing)
OP variety
Synthetic hybrids (poly-clonal gardens) clone
Seed
Cuttings
Seed
C 6 variety (India)
Arabusta (Ivory Coast)
In progress (Ivory Coast, France)
Cuttings or somatic embryogenesis Seed F1 hybrid
In progress (Ivory Coast, France)
Seed
Synthetic hybrids (bi-clonal gardens) clone
BR sel 2 (India), SA and BP selections (Indonesia) IF 126, 202, 461 clones (Ivory Coast) BP39, BP42 (Indonesia)
Apoata (Brazil), S274 (India) Nemaya (C. America)
Icatu (Brazil), S2828 (India)
Ruiru II (Kenya) Ababuna (Ethiopia) in progress: Catimor 6 Et (C. America)
Catuai, Tupi (Brazil), Catimor Sarchimor (Costa Rica), S795 (India) Colombia (Colombia)
Caturra (Brazil), Kents (India) SL28 (Kenya), Java (Cameroon)
Examples
Seed
Seed
Seed
Seed (hand pollination) som. embryogenesis
Seed
Seed
Popagation by
Synthetic variety (bi-, polyclonal gardens) clone
OP variety
Line
Line
Selfing
Variety
Output
Arabica (1) Pure line selection
Breeding system
Source populations
Method
Table 9.4 Summary of methods applied in coffee breeding.
Agronomy I: Coffee Breeding Practices
common cultivars and Ethiopian accessions (Lashermes et al., 1996c). In robusta coffee, experimental evidence for marked hybrid vigour for yield in progenies from interpopulation crosses has been reported from the large breeding programme based on methods of reciprocal recurrent selection implemented in the Ivory Coast since 1985 (Leroy et al., 1993, 1994, 1997; Montagnon et al., 1998a, 1998b). Yields of more than 40% above those of the best commercial clones were recorded in some progenies of intergroup (Congolese 6 Guinean) biparental crosses. Genetic variance for (early) yield and most other characters was considerable and mainly due to general combining ability (additive gene effects) in these trials, as well as in breeding trials including crosses with doubled haploids (Lashermes et al., 1994a). The presence of sufficient (predominantly additive) genetic variance and the possibility of exploiting transgressive hybrid vigour of interpopulation crosses provide ample opportunity for considerable selection progress for higher yields per tree in arabica and robusta coffee. Plant vigour was found to be highly correlated with yields in the aforementioned selection trials. This may have a physiological background, since coffee fruits are strong assimilate-accepting sinks requiring each at least 20 cm2 leaf area (half an arabica leaf) for support without affecting vegetative growth (Cannell, 1985). Vigorous trees will have a high rate of new shoot and leaf production to sustain a heavy crop. Coffee yields per unit area can be increased considerably by closespaced planting systems. However, strong plant vigour and large canopies will lead to early between-tree competition for light and, consequently, to reduced flower initiation and yields. The character short internode length of the compact arabica variety Caturra (single dominant gene Ct, reducing internode length by about 50%) has provided the opportunity of combining vigour with high density planting to increase productivity per hectare while avoiding early yield reduction due to mutual shading. It has been widely applied in variety development in arabica coffee (Table 9.4: cvs Catuai, Catimor, Colombia). F1 hybrids with one parent of the CtCt genotype will also show the required compact growth (Table 9.4: Ruiru II; Catimor 6 Et). The recent confirmation of `dwarf' mutants with short internodes (Kumar et al., 1994) in robusta coffee and also in the C 6 R (ex cross between C. congensis and C. canephora) variety (Srinivasan, 1996) opens the way for similar developments in robusta variety development.
191
9.3.3 Quality Selection for bean size and cup quality has generally received much attention in arabica coffee breeding, particularly in countries producing mild (washed) coffee types, because the quality of new disease resistant cultivars should be at least equal to that of the traditional cultivars in order to uphold the country's reputation and position in the world coffee market. This was obviously achieved in Kenya with the CBD and CLR resistant hybrid cultivar Ruiru II (Njoroge et al., 1990) and in Colombia with the CLR resistant cultivar Colombia (Moreno et al., 1995), as judged by international coffee tasting panels. Most components of coffee quality show considerable (additive) genetic variation, but they are also affected by environmental factors (Walyaro, 1997). Rigorous standardization of pre- and post-harvest practices, bean grading and cup tasting applied in these two breeding programmes contributed to increased selection progress and helped to overcome the initially negative effects on quality due to introgression of disease resistances from exotic germplasm. Verification of the quality of new cultivars in widely different environments (for example climate, altitude and shade) appears necessary because of significant genotype 6 environment interaction effects reported for bean size and some components of liquor quality (Mawardi & Hulupi, 1995; Guyot et al., 1996; Agwanda et al., 1997). The CLR resistant Icatu cultivars released in 1992 in Brazil are also similar in bean size and liquor quality to traditional cultivars according to Brazilian quality standards for unwashed coffees (Fazuoli et al., 1999). This was achieved by a long-term breeding programme of interspecific hybridisation (robusta 6 arabica) started in 1950 and followed by repeated backcrossing to arabica and pedigree selection (Carvalho, 1988). Moschetto et al., (1996) have presented the results of studies on genetic variation in bean size and cup quality for robusta coffee grown in the Ivory Coast. Conclusions on the inheritance of quality characters are similar to those made for arabica coffee. Genotypes of the Congolese group within C. canephora were generally better in cup quality than those of the Guinean group, meaning milder body and acidity, lower bitterness and fewer undesirable aromas, but of course still far below the standards of an average arabica coffee. On the other hand, the cup quality of some Arabusta and Congusta coffees came close to arabica. The Arabusta programme had to be abandoned because of low and irregular yields (Charmetant et al., 1991; Yapo, 1995). However, the Congusta material appears to hold much promise to
192
improving cup quality for robusta-like coffee production, as was shown by the C 6 R variety developed in India from a `Congusta' hybrid backcrossed once to robusta coffee followed by full-sib family selection (Srinivasan, 1996; Srinivasan et al., 1999). This variety combines quality close to arabica coffee with compact growth, good productivity and adaptation to low altitudes. Caffeine content (about twice as high in robusta as in arabica coffee) is a quantitative character with a high heritability (Montagnon et al., 1998c). Barre et al. (1998) deduced from interspecific crosses between the caffeine-free C. pseudozanguebariae and C. liberica var dewevrei that presence or absence of caffeine is under control of one major gene with the double recessive genotype conditioning absence. Unfortunately, absence of caffeine is linked to the presence of a heteroside diterpine causing bitterness, which is also under the control of one major (codominant) gene. Molecular markers for these characteristics could accelerate the search for a genotype lacking both caffeine and the bitter taste, as an alternative to transgenic caffeine-free coffee plants proposed by Moisyadi et al., (1999; see also Chapter 11). On the other hand, coffee is drunk mainly for its stimulating properties derived from the caffeine it contains and moderate coffee drinking does not pose health hazards to most people (see Chapter 8). Caffeine-free cultivars may, therefore, attain only limited prominence in coffee cultivation considering the demand for decaffeinated coffees, which is small (10% of total coffee consumption) and unlikely to increase much in the foreseeable future.
9.3.4 Resistance to coffee leaf rust Coffee leaf rust (Hemileia vastatrix) has spread to all other coffee-producing countries between 1970 (Brazil) and 1986 (PNG and Jamaica), reaching the Central American countries after 1976 and Colombia in 1983 (Carvalho et al., 1989). The relevance of durable host resistance to CLR can be deduced from its economic damage to world arabica coffee production, which has been estimated at US$1±2 billion per year due to crop losses (20 to 25%) and the need to apply cultural and chemical control measures (10% of production costs). For an authoritative review on research for host resistance and breeding of CLR-resistant arabica cultivars reference is made to Eskes (1989). Resistance to CLR is conditioned primarily by a number of major (SH) genes and coffee genotypes are classified in resistance groups according to their interaction with physiological races of the rust pathogen. Some
Coffee: Recent Developments
genotypes with relevance to resistance breeding in arabica coffee and also used as differentials, to test the virulence of rust races, are presented in Table 9.5. The host resistance of Catimor lines, which is based on major genes (SH6±SH9 + ?) originating from C. canephora, has continued to provide adequate protection against CLR epidemics during the past 15 years in countries where arabica coffee is grown under relatively cool climatic conditions (high altitudes) and the number of physiological races present usually remains limited, for example in Colombia, Central American countries, Kenya, Tanzania and PNG. Race II is usually the first one to appear and represents 58% of all isolates tested for virulence from 32 countries, followed by race I (14%), III (9%) and XV (4%). Races II, I and XV had been isolated on susceptible arabica cultivars in Brazil by 1974. However, several more races were subsequently found in the selection fields of the IAC (Instituto AgronoÃmico at Campinas), which emphasised the necessity of developing new cultivars with A type resistance (Carvalho et al., 1989). India has had a long history of arabica coffee breeding dominated by the repeated occurrence of new physiological races of CLR (Carvalho et al., 1989), probably due to the warmer and wetter climatic conditions in major coffee growing areas. CLR infection was noticed on Catimor (Cauvery) trees within a few years after the onset of large-scale planting in 1985 (Srinivasan et al., 1999). Some trees remained free of infection, others were severely infected and proved to be susceptible even to race II, but a number of them had rather mild symptoms of rust infection. By 1993 nine new physiological races had been identified, including races with complex virulence capable of overcoming all four resistance genes (SH6±SH9) but generally showing lower aggressiveness. This has brought the total number of races to 39, of which 30 are present in India (Rodrigues et al., 1993). Cauvery appears to be a mix of A type (resistant to all races), different R type resistance groups (which could have given rise to new virulence of the pathogen), and also some segregants of the E group. Inoculation tests at the Coffee Research Centre (CIFC-Oeiras) in Portugal confirmed that none of these new races could infect Catimor lines with A type resistance, which would indicate the presence of yet other, still unidentified SH genes. Rust infection on Catimor plants in other Asian countries (for example the Philippines) and even in Colombia has been reported recently, but the virulence of the rust isolates still remains to be confirmed (VaÂrzea & Rodrigues, personal communication) On the other hand, Indian arabica derivatives of
Agronomy I: Coffee Breeding Practices
Table 9.5
193
Important differentials for the identification of races of coffee leaf rust. Differential
Host resistance
Group
Variety/cross
Clone
Genotype
Origin of genes
A R R-1 R-2 R-3 R-4
HDT (Hibrido de Timor) HOT (Hbrido de Timor) M. Novo 6 HW 26 M. Novo 6 HW 26 M. Novo 6 HW 26 Caturra 6 HdT 1343/269 S12Kaffa S12Kaffa S12Kaffa S4Agaro S288-23 S353-4/5 Kents KP532-31 Geisha Dilla & Alghe Bourbon
832/1 1343/269 H420/10 H420/2 H419/20 H440/7 635/2 134/4 635/3 110/5 33/1 34/13 32/1 1006/10 87/1 128/2 63/1
SH 5.6.7.8.9 + ? SH 6 Sh 5.6.7.9 SH 5.8 SH 5.6.9 SH 5.6 SH 4 SH 1.4 SH 1.4.5 SH 4.5 SH 3.5 SH 2.3.5 SH 2.5 SH 1.2.5 SH 1.5 SH 1 SH 5
C. canephora C. canephora C. canephora C. canephora C. canephora C. canephora C. arabica ex Ethiopia C. arabica ex Ethiopia C. arabica ex Ethiopia C. arabica ex Ethiopia C. liberica ex India C. liberica ex India C. arabica ex India C. arabica ex Tanzania C. arabica ex Ethiopia C. arabica ex Ethiopia C. arabica
I W J G H D L C E
Note: HW 26 = Caturra 6 HdT 832/1. Adapted from Bettencourt & Rodrigues (1988).
Devamachi (a spontaneous robusta 6 arabica hybrid similar to Hibrido de timor) and a selections like S2828, which was developed from interspecific (robusta 6 arabica) hybridization followed by backcrossing to arabica and pedigree selection, have shown continued high field resistance to CLR in combination with good yields and satisfactory quality (Srinivasan et al., 1999). The nature of the resistance has still to be confirmed, but could be similar to the Catimor-derived (for example Tupi, Obata) and Icatu cultivars from Brazil, of which certain selections appear to have durable resistance to CLR based on major as well as minor genes (Carvalho et al. 1989; Carvalho & Fazuoli, 1993; Fazuoli et al., 1999). Castillo & Alvarado (1997) also found incomplete resistance to CLR in a Catimor line. Nevertheless, incomplete resistance to CLR in certain Catimor and Icatu lines was shown to be race-specific and, therefore, unlikely to provide durable resistance (Eskes et al., 1990). Molecular markers linked to SH and other genes conditioning race-specific and non-specific resistance to CLR should increase selection efficiency for durable resistance in arabica coffee, particularly in regard to gene pyramiding (accumulating several resistance genes in one genotype). Initial work based on RAPD markers shows the existence of considerable polymorphisms between some rust differentials (Santa Ram
& Sreenath, 1999) and further work in collaboration with national and international institutes on molecular marker-assisted selection in CLR (and other important characters) may commence soon (Sreenath & Naidu, 1999). Resistance to CLR in robusta coffee is usually a secondary character of selection and based on individual plant, clone or family scores for field infection. Robustas of the Congolese group are generally much more resistant than those of the Guinean group (Montagnon et al., 1994; Leroy et al., 1997).
9.3.5 Resistance to coffee berry disease Coffee berry disease is caused by the fungus Colletotrichum coffeanum, renamed C. kahawae (Waller et al., 1993). It can be a devastating anthracnose of developing berries in arabica coffee in Africa and is particularly serious at high altitudes (Van der Graaff, 1992; Masaba & Waller, 1992). Coffee berry disease may cause crop losses of 50 to 80% in years favourable to a severe disease epidemic (prolonged wet and cool weather). Control by frequent fungicide sprays is expensive (30 to 40% of total production costs), not always effective and usually beyond the means of the smallholder coffee growers. Economic damage to arabica coffee production in Africa due to CBD alone (crop losses plus costs
194
of control) is estimated at about US $ 300 to 500 million per year. Breeding programmes initiated some 30 years ago in Kenya and Ethiopia have been successful in developing new cultivars with high and apparently durable resistance to CBD (Van der Vossen, 1997). In Kenya more than 10 000 ha have been planted so far with the composite hybrid Ruiru II, which combines resistance to CBD and CLR with compact growth, high yields and quality similar to the standard Kenyan cultivars. In Ethiopia, farmers' acceptance of CBD-resistant hybrids like Ababuna is higher than of the earlier released lines, because of better agronomic performance, but data on actual area planted to CBDresistant cultivars are unavailable. A breeding programme in Tanzania, which is making use of CBD resistance found in Hibrido de Timor (clone CIFC 1343) and Rume Sudan, has reached the stage of multi-locational testing of clones derived from multiple crosses and the first release of CBD (and CLR) resistant clonal cultivars has been envisaged within a few years time (Nyange et al., 1999). However, these cultivars may not yet meet the cup quality standards of typical Tanzanian mild arabica coffees. Selection work in Cameroon, which concentrated on screening a large number of arabica varieties and also accessions of Ethiopian (Et) origin for agronomic characteristics and disease resistance, produced the CBD resistant cultivar Java by 1980 (Bouharmont, 1994). A subsequent breeding programme indicated considerable hybrid vigour for crosses involving some Et accessions as one parent, but also the poor combining ability of cultivar Java (Cilas et al., 1998). Host resistance to CBD appears to be mainly conditioned by three major genes (dominant R, codominant T and recessive k genes) according to evidence produced in Kenya, although breeders in Ethiopia claim additive effects of several recessive genes instead (Walyaro, 1997; Bellachev, 1997). Part of this discrepancy could be explained by the differences in germplasm and also the inoculation tests used in the inheritance studies. In Ethiopia the inheritance studies were mainly based on results from inoculation tests on detached berries. Gichuru et al., (1999) confirmed the absence of histological and chemical resistance mechanisms in detached berries, commonly expressed in attached berries or hypocotyl stems of CBD resistant varieties. Support for the hypothesis of major gene resistance has come from the recent application of molecular marker technology by Agwanda et al., (1997) and Cristancho (1999), who found closely linked RAPD markers for the T gene conditioning CBD
Coffee: Recent Developments
resistance in progenitor Hibrido de Timor and its derivatives such as Catimor. Only HdT1343 (progenitor of Colombian Catimor lines) carries the T gene, while Catimor lines derived from HdT832 (for example CIFC-Oeiras and Brazil) are all CBD susceptible (Van der Vossen, 1997). Rovelli et al., (1999) reported the detection of polymorphic microsatellites in arabica coffee, which offers prospects of developing useful molecular markers for other CBD resistance genes. This would enable breeders to reconfirm the genetic basis for CBD resistance and to ensure effective host resistance by gene pyramiding. It would also enable breeders outside Africa to verify or introgress CBD resistance in their breeding stock as pre-emptive action in case CBD inadvertently becomes a pathogen in their region. This strategy was followed for CLR by the Colombian breeders several years before it arrived in their country (Castillo, 1989). Reports of the existence of race-specific interactions between the CBD pathogen and arabica genotypes by Rodrigues et al., (1992) and VaÂrzea et al., (1999) could not be confirmed in extensive studies by Bella Manga et al., (1997) and Bella Manga (1999), which included a large number of isolates of C. kahawae collected from CBD susceptible and resistant arabica cultivars and accessions in several African countries. The genetic diversity of pathogen populations, as evaluated by VCG (vegetative compatibility groups) and RAPD molecular markers, was relatively narrow, but it was possible to distinguish two subgroups, one from East Africa and one from the Cameroon. Certain isolates from the Cameroon were also more pathogenic than those from East Africa. However, pathogenicity tests revealed insignificant pathogen±host interactions despite considerable variation in aggressiveness of isolates and in levels of host resistance. Omondi et al., (1997), in a study with CBD isolates from Kenya, produced comparable results. Variation in aggressiveness among CBD isolates was also found in Ethiopia (Derso, 1999). It can be concluded that the available host resistance is not (yet) threatened by race specification in the CBD pathogen, but levels of host resistance required for adequate crop protection may vary between different geographical areas.
9.3.6 Resistance to other diseases Several other fungal and bacterial diseases may affect coffee (Wrigley, 1988; Anon, 1997), but very few of these have been targeted in breeding programmes, notwithstanding considerable economic damage in certain coffee producing countries, because useful host
Agronomy I: Coffee Breeding Practices
resistance could not be detected in available coffee germplasm. For instance, black rot (Koleroga noxia) is the second most important disease after CLR in India (Bhat et al., 1995), coffee leaf scorch caused by the bacterium Xylella fastidiosa has become a problem in some coffee regions in Brazil (Beretta et al., 1996) and bacterial blight (Pseudomonas syringae pv garcae) can be severe in a few areas in Kenya (Kairu, 1997). Fusarium wilt disease or tracheomycosis (Fusarium xylarioides) has been causing severe losses of robusta coffee in north-eastern DR Congo and south-western Uganda since the 1980s (Flood & Brayford, 1997; Birikunzira & Hakiza, 1997). The exact cause of this reemergence is still unknown, but it was noted that especially old and rather neglected coffee plots were severely affected. Some of the Ugandan robusta clonal cultivars have remained resistant and renewed selection for host resistance within robusta germplasm could, therefore, be rewarding.
9.3.7 Resistance to nematodes Root-knot (Meloidogyne spp.) and root-lesion (Pratylenchus spp.) nematodes can cause considerable economic damage to arabica coffee in Brazil (Carvalho, 1988), Central America (Anzueto et al., 1991), India (Anon, 1997) and Indonesia (Mawardi & Soenaryo, 1988). They are usually a minor problem in East Africa, provided a build-up of nematodes in nurseries is avoided (Mitchell, 1988). Host resistance to both types of endoparasitic nematodes is present in germplasm of robusta coffee and selections have been widely used as rootstock for arabica cultivars in problem areas, such as Nemaya in Central America (Anzueto et al., 1991) and Apoata in Brazil (Carvalho & Fazuoli, 1993). More recently, Apoata has been recommended also as a suitable cultivar for expansion of robusta coffee production in the State of Sao Paulo, on account of its good yield potential and resistance to CLR (MedinaFilho et al., 1999). According to research work in Costa Rica, Guatemala and El Salvador two groups of Meloidogyne species can be distinguished: (a) those forming egg masses inside the roots and causing intensive gall formation but less root destruction, such as M. exigua and M. arabicida, and (b) those forming egg masses outside the roots and causing fewer and smaller galls but high root damage, such as M. incognita, M. arenaria and M. javanica (Bertrand et al., 1995). Host resistance to the first group was found in a Catimor line from Colombia, while a few arabica accessions from Ethiopia showed resistance to the second group of nematodes. The
195
resistance in each case appeared to be conditioned by one or two dominant major genes. Resistance to both groups simultaneously was found in several robusta plants (major and minor genes). All arabica germplasm is susceptible to Pratylenchus species, but some robusta genotypes are tolerant. Coffee fields in Central America are often infested with root-knot and root-lesion nematodes at the same time and the two populations are antagonistically related. Introduction of a cultivar resistant to one type only would run the risk of an escalation of the other (Bertrand et al., 1998). It is, therefore, essential to rely on broad spectrum resistance to parasitic nematodes, initially by using suitable robusta rootstock (for example Nemaya) or in the long term by developing new arabica (hybrid) cultivars with resistance to major Meloidogyne as well as Pratylenchus species. A recently started collaborative research project aims at developing molecular markers linked to resistance genes to enhance selection for such nematode resistance (Lashermes et al., 1999a).
9.3.8 Resistance to insect pests Several hundred insect species have been described as minor or major coffee pests (Wrigley, 1988). Integrated pest management (IPM) ± early warning systems in combination with cultural, biological and chemical control ± has been successfully applied to a number of important coffee pests (Bardner, 1985). The arrival of the coffee berry borer (Hypothenemus hampei) in Colombia in 1988 (Bustillo et al., 1995) and in India (Bheemaya et al., 1996; Anon, 1997) gave a new impetus to the development of effective methods of IPM for the control of this most damaging insect pest in coffee, based on specific parasitoids and entomopathogens. The identification and synthesis of a male sex pheromone of the coffee white stem borer (Xylotrechus quadripes) offers promising perspectives of biological control of this important pest in arabica coffee in India (Hall et al., 1998; Jayarama et al., 1998). Host resistance to the leaf miner Perileucoptera coffeella, a severe coffee pest in Brazil, was found in the species Coffea stenophylla and C. racemosa, but only the resistance of the latter was successfully introgressed into arabica coffee (Carvalho, 1988). This resistance is conditioned by two complementary dominant genes (Guerreiro Filho et al., 1999). So far, no other cases of useful host resistance to important insect pests have been detected in Coffea germplasm. However, the successful regeneration of transgenic coffee plants expressing resistance to leaf miners (P. coffeella and Leucoptera spp. based on Bt genes (Leroy et al., 1999;
196
see Chapter 11) could be the start of molecular breeding for resistance to important coffee pests, especially the endocarpic insects (Guerreiro Filho et al., 1998).
9.3.9 Drought tolerance Arabica coffee is generally more tolerant to water stress than robusta, at least partly as the result of a more extensive and deeper root system. However, there are also large differences in drought tolerance between genotypes of the same species. Some of the East-African cultivars (for example SL28) appear to be the best genotypes available within arabica germplasm, because of an exceptionally well developed root system, outstanding plant vigour and an ability to retain their leaves under water stress (Van der Vossen & Browning, 1978). Such genotypes must have evolved in a long process of domestication, lasting some 12 to 15 centuries, from shade-adapted trees occurring in the understorey of the highland forests in Ethiopia to the dry and unshaded conditions first in Yemen and eventually in East Africa. Wilting of coffee plants during a prolonged dry spell in Kenya always started much later in plots of SL28, than in plots planted with introduced arabica cultivars or accessions of Ethiopian origin. In Brazil, cultivars like Caturra and Mundo Novo were also found to be more tolerant to drought than Ethiopian germplasm (Carvalho, 1988). Selection for more drought tolerant robusta coffee should emphasise depth and extent of root system, as well as leaf retention under stress conditions. The Indian C 6 R cultivar appears to have better drought tolerance, but it is unlikely that robusta-like genotypes would be found with better drought tolerance than arabicas. Carvalho (1988) mentions C. racemosa as a good progenitor for drought tolerance.
9.4 PROPAGATION OF NEW CULTIVARS 9.4.1 Seeds Propagation by seeds continues to be the preferred practice for new coffee cultivars in most countries (Table 9.4). When the output of a breeding programme consists of pure lines, as in the case of arabica coffee, spatially isolated seed gardens are established for lowcost seed multiplication and distribution. Examples are Brazil (Fazuoli et al., 1999), India (Srinivasan, 1996) and the Cameroon (Bouharmont, 1994). In Colombia, a
Coffee: Recent Developments
number of Catimor lines are multiplied in separate seed gardens and the cultivar Colombia consists of a synthetic seed mix from selected lines (Moreno, 1994). Seed multiplication of F1 hybrid arabica cultivars is a logistically complex operation involving hand pollination of previously emasculated and bagged flowers (Opile & Agwanda, 1993). Experience with the Kenyan hybrid cultivar Ruiru II shows that large-scale seed multiplication is technically feasible and cheaper than clonal propagation. However, the national output would probably be improved considerably by decentralisation into smaller seed production units (Van der Vossen, 1997). Male sterility conditioned by one recessive gene has been detected in arabica accessions of Ethiopian origin (Mazzafera et al., 1989; Dufour et al., 1997), which provides opportunities of reducing costs of seed production. This gene could become a suitable object of molecular breeding (MAS and even gene transformation) to obtain male-sterile female parents for seed production within a much shorter period of time than would be required with conventional methods of introgressive breeding. Seeds of robusta coffee cultivars are produced in strictly isolated seed gardens planted with selected seedling populations (for example cultivars Apoata, Nemaya and S274). Synthetic hybrid seeds require gardens planted with (preferably two) clones of known combining ability (Charmetant et al., 1990; Montagnon et al., 1998a). Self-incompatibility ensures cross-pollination. Such `clone hybrids' are not yet very uniform due to heterozygosity of the parent clones, but genetically uniform robusta hybrid seeds may eventually be realised with parents developed from doubled haploids (Lashermes et al., 1994b).
9.4.2 Clonal propagation Conventional methods of clonal propagation are about ten times more expensive than multiplication by seed (Montagnon et al., 1998a). Clonal robusta cultivars found limited application in large-scale coffee plantations, where the increased yield potential can be fully exploited and the higher initial costs are quickly recovered (Charrier & Berthaud, 1988). Clonal propagation of hybrid arabica cultivars is even more demanding of logistic and technical resources, due to slower rates of multiplication and hardening-off problems in the cooler and dryer environments of arabica coffee cultivation. Of all the in vitro regeneration systems tried out in coffee (Carneiro, 1997) the induction of high frequency somatic embryogenesis in a liquid medium (Zamarripa
Agronomy I: Coffee Breeding Practices
et al., 1991; Berthouly & Michaux-FerrieÁre, 1996) appears to be very promising for efficient mass propagation. It is being implemented in arabica coffee in Central America to multiply Catimor 6 Et hybrids as an alternative to hybrid seed production (Etienne et al., 1997a, b) and in Uganda to multiply elite robusta clones (Berthouly et al., 1995). Coffee plants raised from rooted cuttings tend to be shallower rooting, in the absence of a tap root, than seedlings and consequently are less tolerant of prolonged spells of dry weather. A major advantage of plants raised from somatic embryogenesis is their similarity to seedlings with respect to the root system. Deshayes et al., (1999) claim that costs of producing robusta coffee plants through somatic embryogenesis are comparable to conventional rooted cuttings, in other words, still considerably more expensive than hybrid seed production.
ABBREVIATIONS ACRN AFLP BCRI/E CATIE CBD CCRI CENICAFE CIFC CIRAD CLR CNRA CRF FOFIFA IAC IAR/JARC ICO ICCRI IPGRI IPM IRAD IRD
African Coffee Research Network Amplified fragment length polymorphism Biodiversity Conservation and Research Institute, Ethiopia Centro AgronoÂmico Tropical de Investigacion y EnsenÄanza, Costa Rica Coffee berry disease Central Coffee Research Institute, India Centro Nacional de Investigaciones de CafeÂ, Colombia Centro d'InvestigacËao das Ferrugens do Cafeiero, Portugal Centre de CoopeÂration Internationale en Recherche Agronomique pour le DeÂveloppement, France Coffee leaf rust Centre National de Recherche Agronomique, CoÃte d'lvoire Coffee Research Foundation, Kenya Centre National de Recherche Agronomique AppliqueÂe au DeÂveloppement Rural, Madagascar Instituto AgronoÃmico de Campinas, Brazil Institute of Agricultural Research/Jimma Agricultural Research Centre, Ethiopia International Coffee Organization, London Indonesian Coffee and Cocoa Research Institute International Plant Genetic Resources Institute, Rome Integrated pest management Institut de Recherche Agronomique et DeÂveloppement, Cameroun Institut de Recherche pour le DeÂveloppement (ex ORSTOM), France
197
RAPD RFLP TARO VCG
Random amplified polymorphic DNA Restriction fragment length polymorphisms Tanzanian Agricultural Research Organization Vegetative incompatibility
REFERENCES Abdelnour-Esquivel, A., Vallalobos, V. & Engelmann, F. (1992) Cryopreservation of zygotic embryos of Coffea spp. Cryo-Letts, 13, 297±302. Agwanda, C.O., Baradat, P., Cilas, C. & Charrier, A. (1997) Genotype-by-environment interaction and its implications on selection for improved quality in arabica coffee (Coffea arabica). In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 424± 9. ASIC, Paris, France. Agwanda, C.O., Lashermes, P., Trouslot, P., Combes, M.C. & Charrier, A. (1998) Identification of RAPD markers for resistance to coffee berry disease, Colletotrichum kahawae, in arabica coffee. Euphytica, 97, 241±8. Ameha, M. (1990) Heterosis and arabica coffee breeding in Ethiopia. Plant Breed. Abstr. 60, 594±8. Anon (1997) Coffee Guide. Coffee Board of India, Central Coffee Research Institute, Chikmagalur District, Karnataka, India. Anzueto, F., Eskes, A.B., Sarah, J.L. & Decazy, B. (1991) Recherche de la reÂsistance aÁ Meloidogyne sp. dans une collection de Coffe arabica. In: Proceedings of the 14th ASIC Colloquium (San Francisco), pp. 534±543. ASIC, Paris, France. Bardner, R. (1985) Pest control. In: Coffee: Botany, Biochemistry and Production of Beans and Beverage (eds M. N. Clifford & K. C. Willson), pp. 208±218. Croom Helm, London, New York and Sydney. Barre, P., Akaffou, S., Louarn, J., Charrier, A., Hamon, S. & Noirot, M. (1998) Inheritance of caffeine and heteroside contents in an interspecific cross between a cultivated coffee species Coffea liberica var dewevrei and a wild species caffeinefree C. pseudozanguebariae. Theoret. App. Genet., 96, 306±11. Bellachev, B. (1997) Arabica coffee breeding in Ethiopia: a review. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 406± 14. ASIC, Paris, France. Bella Manga (1999) Etude de la diversite de Colletotrichum kahawae responsable de I’anthracnose des baies et caracteÂrisation de la reÂsistance du cafeÂier Arabica aÁ cet agent pathogeÁne. TheÁse Docteur en Sciences, Universite Montpellier II, Sciences et Technique du Langedoc, Montpellier, France. Bella Manga, Bieysse, D., Mouen Bedimo, J.A., Akalay, I., Bompard, E. & Berry, D. (1997) Observations sur la diversite de la population de Colletotrichum kahawae agent de I'anthracnose des baies du cafeÂier arabica: implications pour I'ameÂlioration geÂneÂtique. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 604±12. ASIC Paris, France. Beretta, M.J.G., Harakawa, R., Chagas, C.M. et al. (1996) First report of Xylella fastidiosa in coffee. Plant Dis. St Paul, 80, 821±6. Berthaud, J. (1985) Propositions pour une nouvelle strateÂgie d'ameÂlioration des cafeÂiers de I'espeÁce C. canephora, baseÂe sur les reÂsultats de I'analyse des populations sylvestres. In: Pro-
198
ceedings of the 11th ASIC Colloquium (LomeÂ), pp. 445±52. ASIC, Paris, France. Berthaud, J. & Charrier, A. (1988) Genetic resources of Coffea. In: Coffee Vol. 4 Agronomy (eds R.J. Clarke & R. Macrae), pp. 1±42. Elsevier Applied Science, London and New York. Berthouly, M., Alvaro, D., Carasco, C. & Duris, D. (1995) A technology transfer operation: a commercial Coffea canephora micro-propagation laboratory in Uganda. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 743±4. ASIC, Paris, France. Berthouly, M. & Michaux-FerrieÁre, N. (1996) High frequency somatic embryogenesis in Coffea canephora. Plant Cell Tiss. Org. Cult., 44, 169±76. Bertrand, B., Aguilar, G., Santacreo, R. et al. (1997) Comportement d'hybrides F1 de Coffea arabica pour la vigueur, la production et la fertilite en AmeÂrique Centrale. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 415±23. ASIC, Paris, France. Bertrand, B., Anzueto, F., Pena, M. X., Anthony, F. & Eskes, A. B. (1995) Genetic improvement for resistance to root-knot nematodes (Meloidogyne spp) in Central America. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 630±36. ASIC, Paris, France. Bertrand, B., Cilas, C., HerveÂ, G., Anthony, F., Etienne, H. & Villain, L. (1998) Relations entre les populations des neÂmatodes Meloidogyne exigua et Pratylenchys sp. dans les racines de Coffea arabica au Costa Rica. Plant. Rech. DeÂvelop., 5, 279±84. Bettencourt, A.J. & Rodrigues, C.J. (1988) Principles and practice of coffee breeding for resistance to rust and other diseases. In: Coffee Vol. 4 Agronomy (eds R.J. Clarke & R. Macrae), pp. 199± 234. Elsevier Applied Science, London and New York. Bhat, S.S., Daivasikamani, S. & Naidu, R. (1995) Tips on effective management of black rot disease in coffee. Ind. Coffee, 59, 3±5. Bheemaya, M.M., Sreedharan, K. & Dhruvakumar, H.K. (1996) Coffee berry borer: the dreaded pest on coffee. Ind. Coffee, 60, 3±5. Birikunzira, J.B. & Hakiza, J. (1997) The status of coffee wilt disease (tracheomycosis) and strategies for its control in Uganda. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 766±70. ASIC, Paris, France. Bouharmont, P. (1994) La varieÂte Java: un cafeÂier Arabica seÂlectionne au Cameroun. Plant. Rech. DeÂvelop., 1, 38±45. Bridson, D.M. & Verdcourt, B. (1988) Rubiaceae (Part 2). In: Flora of Tropical East Africa (ed. R.M. Polhill), pp. 703±23. Balkema, Rotterdam and Brookfield. Bustillo, A.E., Villalba, D., Orozco, J., Benavides, P., Reyes, I.C. & Chaves, B. (1995) Integrated pest management to control the coffee berry borer, Hypothenemus hampei, in Colombia. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 671±80. ASIC, Paris, France. Cambrony, H.R. (1988) Arabusta and other interspecific fertile hybrids. In: Coffee Vol. 4 Agronomy (eds R.J. Clarke & R. Macrae), pp. 263±91. Elsevier Applied Science, London and New York. Cannell, M.G.R. (1985) Physiology of the coffee crop. In: Coffee: Botany, Biochemistry and Production of Beans and Beverage (eds
Coffee: Recent Developments
M.N. Clifford & K.C. Willson), pp. 108±24. Croom Helm, London, New York and Sydney. Carneiro, M.F. (1997) Coffee biotechnology and its application in genetic transformation. Euphytica, 96, 167±72. Carvalho, A. (1988) Principles and practice of coffee plant breeding for productivity and quality factors: Coffea arabica. In: Coffee Vol. 4 Agronomy (eds R.J. Clarke & R. Macrae), pp. 129±65. Elsevier Applied Science, London and New York. Carvalho, A., Eskes, A.B., Castillo, J. et al. (1989) Breeding programmes. In: Coffee Rust: Epidemiology, Resistance and Management (eds A.C. Kushalappa & A.B. Eskes), pp. 293± 335. CRC Press, Boca Raton, Florida. Carvalho, A. & Fazuoli, L.C. (1993) CafeÂ. In: O Melhoramento de Plantas no Instituto AgronoÃmico, Vol. 1 (eds A.M.C. Furlani & G.P. VieÂgas), pp. 29±76. Instituto AgronoÃmico, Campinas-SP, Brazil. Castillo, J. (1989) Breeding for rust resistance in Colombia. In: Coffee Rust: Epidemiology, Resistance and Management (eds A.C. Kushalappa & A.B. Eskes), pp. 307±16. CRC Press, Boca Raton, Florida. Castillo, J. & Alvarado, A.G. (1997) Resistencia incompleta de genotipos de cafe a la roya bajo condiciones de campo en la regioÂn centrale de Colombia. CenicafeÂ, 48, 40±58. Charmetant, P., Leroy, T., Bontems, S. & Delsol, E. (1990) Evaluation d'hybrides de Coffee canephora produits en champs semenciers en CoÃte d'lvoire. CafeÂ, Cacao, TheÂ, 34, 257±64. Charmetant, P., Le Pierres, D. & Yapo, A. (1991) Evaluation d'hybrides arabusta F1 (cafeÂiers diploõÈdes doubleÂs 6 Coffea arabica) en CoÃte d'lvoire. In: Proceedings of the 14th ASIC Colloquium (San Francisco), pp. 422±30. ASIC, Paris, France. Charrier, A. & Berthaud, J. (1985) Botanical classification of coffee. In: Coffee: Botany, Biochemistry and Production of Beans and Beverage (eds M.N. Clifford & K.C. Willson), pp. 13±47. Croom Helm, London, New York and Sydney. Charrier, A. & Berthaud, J. (1988) Principles and methods of coffee plant breeding: Coffea Canephora Pierre. In: Coffee Vol. 4 Agronomy (eds R.J. Clarke & R. Macrae), pp. 167±97. Elsevier Applied Science, London and New York. Charrier, A. & Eskes, A.B. (1997) Les cafeÂiers. In: L’AmeÂlioration des Plantes Tropicales (eds A. Charrier, M. Jackot, S. Hamon & D. Nicolas), pp. 171±96. CIRAD/ORSTOM, Montpellier, France. Cilas, C., Bouharmont, P., Boccara, M., Eskes, A.B. & Baradat, Ph. (1998) Prediction of genetic value for coffee production in Coffea arabica from a half-diallel with lines and hybrids. Euphytica, 104, 49±59. Couturon, E., Lashermes, P. & Charrier, A. (1998) First intergeneric hybrids. (Psilanthus ebracteolatus Hiern 6 Coffea arabica L.) in coffee trees. Can. J. Bot., 76, 542±6. Cristancho, M. (1999) Genetic diversity of the CENICAFE germplasm collection. Third International Seminar on Biotechnology in the Coffee Agroindustry, Londrina, Brazil. Cros, J., Trouslot, P., Anthony, F., Hamon, S. & Charrier, A. (1998) Phylogenetic analysis of chloroplast DNA variation in Coffea L. Mol. Phylogen. Evol., 9, 109±117. Derso, E. (1999) Variation in aggressiveness among Colletotrichum
Agronomy I: Coffee Breeding Practices
kahawae isolates in Ethiopia. In: Proceedings of the 18th ASIC Colloquium (Helsinki), ASIC, Paris, France. Deshayes, A., Ducos, J.P., Gianforcaro, M., Florin, B. & PeÂtiard, V. (1999) A technically and economically attractive way to propagate elite coffee cultivars: in vitro somatic embryogenesis. In: Proceedings of the 18th ASIC Colloquium (Helsinki), pp. 295±304. ASIC, Paris, France. Dufour, M., Anthony, F., Bertrand, B. & Eskes, A.B. (1997) Identification de cafeÂiers maÃle-steÂriles de Coffea arabica au CATIE, Costa Rica. Plant. Rech. DeÂvelop., 4, 401±407. Dulloo, M.E., Guarino, L., Engelmann, F. et al. (1998) Complementary conservation strategies for the genus Coffea: a case study of Mascarene Coffea species. Genet. Res. Crop Evol., 45, 565±79. Dussert, S., Chabrillange, N., Engelmann, F., Anthony, F. & Hamon, S. (1997b) Cryopreservation of coffee (Coffea arabica L.) seeds: importance of the precooling temperature. CryoLett, 18, 269±76. Dussert, S., Chabrillange, N., Engelmann, F., Anthony, F., Louarn, J. & Hamon, S. (1998) Cryopreservation of seeds of four coffee species (Coffea arabica, C. costafructa, C. racemosa and C. sessiflora): importance of water content and cooling rate. Seed Sci. Res., 8, 9±15. Dussert, S., Chabrillange, N., Engelmann, F., Anthony, F., ReÂcalt, C. & Hamon, S. (1997a) Variability in storage response within a coffee (Coffea spp.) core collection. Plant Cell Rep., 16, 344±8. Dussert, S., Chabrillange, N., Engelmann, F. & Hamon, S. (1999) Quantitative estimation of seed desiccation sensitivity using a quantal response model: application to nine species of the genus Coffea L. Seed Science Res., 9, 135±44. Dussert, S., Engelmann, F., Chabrillange, N., Anthony, F., Noirot, M. & Hamon, S. (1997c) In vitro conservation of coffee (Coffea spp.) germplasm, In: Conservation of Plant Genetic Resources in vitro Vol. 1: General Aspects, (eds M.K. Razdan & E.C. Cocking), pp. 287±305. Science Publishers, USA. Eskes, A.B. (1989) Resistance. In: Coffee Rust: Epidemiology, Resistance and Management (eds A.C. Kushalappa & A.B. Eskes), pp. 171±291. CRC Press, Boca Raton, Florida. Eskes, A.B. & Mukred, A. (1989) Coffee survey in PDR Yemen. In: Proceedings of the 13th ASIC Colloquium (Paipa), pp. 582± 9. ASIC, Paris, France. Eskes, A.B., Engels, J. & Lass, T. (1998) The CFC/ICCO/PGRI project: a new initiative on cocoa germplasm utilization and conservation. Plant., Rech., DeÂvelop., 5, 412±17. Eskes, A.B., Hoogstraten, J.G.J., Toma-Braghini, M. & A Carvalho (1990) Race-specificity and inheritance of incomplete reistance to coffee leaf rust in some Icatu progenies and derivative of Hibrido de Timor. Euphytica, 47, 11±19. Etienne, H., Bertrand, B., Anthony, F., CoÃte, F. & Berthouly, M. (1997a) L'embryogeneÁse somatique: un outil pour l'ameÂlioration geÂneÂtique du cafeÂier. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 457±65. ASIC, Paris, France. Etienne, H., Solano, W., Pereira, A., Bertrand, B. & Berthouly, M. (1997b) Protocole d'acclimatation de plantules de cafeÂiers produits in vitro. Plant. Rech. DeÂvelop., 4, 304±309. Fazuoli, L.C., Medina-Filho, H.P., Guerreiro Filho, O., Con-
199
calves, W., Silvarolla, M.B. & Lima, M.M.A. (1999) Coffee cultivars in Brazil. In: Proceedings of the 18th ASIC Colloquium (Helsinki), pp. 396±404. ASIC, Paris, France. Flood, J. & Brayford, D. (1997) Re-emergence of Fusarium wilt of coffee in Africa. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 621±8. ASIC, Paris, France. Gichuru, E.K., Kingori, P.N. & Masaba, D.M. (1999) Histochemical differences during infection of Coffea arabica varieties by Colletotrichum kahawae isolates. In: Proceedings of the 18th ASIC Colloquium (Helsinki), pp. 477±479. ASIC, Paris, France. Guarino, L., Charrier, A. & Opile, W.R. (1995) A plan of action for the ACRN programme on plant genetic resources. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 626±9. ASIC, Paris, France. Guerreiro Filho, O., Denoit, P., Pefercen, M., Decazy, B., Eskes, A.B. & Frutos, R. (1998) Susceptibility to the coffee leaf miner (Perileucoptera spp.) to Bacillus thuringiensis d-endotoxins: a model for transgenic perennial crops resistant to endocarpic insects. Curr. Microbiol., 36, 175±9. Guerreiro Filho, O., Silvarolla, M.B. & Eskes, A.B. (1999) Expression and mode of inheritance of resistance in coffee to leaf miner Perileucoptera coffeella. Euphytica, 105, 7±15. Guyot, B., Gueule, D., Manez, J.C., Perriot, J.J., Giron, J. & Villain, L. (1996) Influence de l'altitude et de l'ombrage sur la qualite des cafeÂs Arabica. Plant., Rech. DeÂvelop., 3, 272±83. Hall, D.R., Cork, A., Phythian, S., Sumathi, C. et al. (1998) Studies on the male sex pheromone of the coffee white stemborer, Xylotrechus quadripes Chevrolat (Coleoptera: Cerambycidae). In: Second International Symposium on Insect Pheromones. Book of Abstracts, pp. 50±52. WICC±IAC, Wageningen, the Netherlands. Hamon, S., Dussert, S., Deu, M. et al. (1998) Effects of quantitative and qualitative principal component score strategies on the structure of coffee, rice, rubber tree and sorghum core collections. Genet. Sel. Evol., 30(Suppl 1), S236±S258. ICO (1990±99) Annual Statistics on Coffee Production and International Trade. International Coffee Organization, London, UK. Jayarama, M.G., Venkatesh, M., D'Souza, V., Naidu, R., Hall, D.R. & Cork, A. (1998) Sex pheromone of coffee white stem borer for monitoring and control is on the anvil. Ind. Coffee, 62, 15±16. Kairu, G.M. (1997) Biochemical and pathogenic differences between Kenyan and Brazilian isolates of Pseudomonas syringae pv garcae. Plant Pathol., 46, 239±46. Kumar, A., Srinivasan, C.S. & Nataraj, T. (1994) A preliminary note on the occurrence of dwarf mutants in robusta coffee (Coffea canephora). J. Coffee Res., 24, 41±5. Lashermes, P., Agwanda, C.O., Anthony, F., Combes, M.C., Trouslot, P. & Charrier, A. (1997b) Molecular marker-assisted selection: a powerful approach for coffee improvement. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 474±80. ASIC, Paris, France. Lashermes, P., Andrzejewski, S., Bertrand, B. et al. (2000b) Molecular analysis of introgressive breeding in coffee (Coffea arabica L.). Theoret. Appl. Genet., 100, 139±46.
200
Lashermes, P., Anzueto, F., Bertrand, B., Graziosi, G. & Anthony, F. (1999a) Sustainable improvement of nematode resistance in coffee cultivars (Coffea arabica L.) of Central America: enhanced use of genetic resources by the development of marker-facilitated selection programmes. In: Proceedings of the 18th ASIC Colloquium (Helsinki). ASIC, Paris, France. Abstract available. Lashermes, P., Combes, M.C., Cros, J., Trouslot, P., Anthony, F. & Charrier, A. (1995) Origin and genetic diversity of Coffea arabica L. based on DNA molecular markers. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 528±35. ASIC, Paris, France. Lashermes, P., Combes, M.C., Robert, J. et al. (1999b) Molecular characterization and origin of the Coffea arabica L. genome. Mol. Gen. Genet., 261, 259±66. Lashermes, P., Combes, M.C., Trouslot, P. & Charrier, A. (1997a) Phylogenetic relationships of coffee tree species (Coffea L.) as inferred from ITS sequences of nuclear ribosomal DNA. Theoret. Appl. Genet., 94, 947±55. Lashermes, P., Couturon, E. & Charrier, A. (1994a) Combining ability of doubled haploids in Coffea canephora P. Plant Breed., 112, 330±37. Lashermes, P., Couturon, E. & Charrier, A. (1994b) Doubled haploids of Coffea canephora: development, fertility and agronomic characteristics. Euphytica, 74, 149±57. Lashermes, P., Couturon, E., Moreau, N., Pailard, M. & Louarn, J. (1996a) Inheritance and genetic mapping of selfincompatibility in Coffea canephora. Theoret. Appl. Genet., 93, 458±62. Lashermes, P., Cros, J., Combes, M.C. et al. (1996b) Inheritance and restriction fragment length polymorphism of chloroplast DNA in the genus Coffea L. Theoret. Appl. Genet., 93, 626±32. Lashermes, P., Paczek, V., Trouslot, P., Combes, M.C., Couturon, E. & Charrier, A. (2000a) Single-locus inheritance in the allotetraploid Coffea arabica L. and interspecific hybrid C. arabica 6 C. canephora. J. Heredity, 91, 81±5. Lashermes, P., Trouslot, P., Anthony, F., Combes, M.C. & Charrier, A. (1996c) Genetic diversity for RAPD markers between cultivated and wild accessions of Coffea arabica. Euphytica, 87, 59±64. Leroy, T., Henry, A.M., Philippe, R. et al. (1999) Genetically modified coffee trees for resistance to coffee leaf miner. Analysis of gene expression, insect resistance and agronomic value. In: Proceedings of the 18th ASIC Colloquium (Helsinki), pp. 332±8. ASIC, Paris, France. Leroy, T., Montagnon, C., Charrier, A. & Eskes, A.B. (1993) Reciprocal recurrent selection applied to Coffea canephora Pierre. I. Characterization and evaluation of breeding populations and value of intergroup hybrids. Euphytica, 67, 113±25. Leroy, T., Montagnon, C., Cilas, C., Charrier, A. & Eskes, A.B. (1994) Reciprocal recurrent selection applied to Coffea canephora Pierre. II. Estimation of genetic parameters. Euphytica, 74, 121±8. Leroy, T., Montagnon, C., Cilas, C., Yapo, A., Charrier, A. & Eskes, A.B. (1997) Reciprocal recurrent selection applied to Coffea canephora Pierre. III. Genetic gains and results of first cycle intergroup crosses. Euphytica, 95, 347±54.
Coffee: Recent Developments
Masaba, D.M. & Waller, J.M. (1992) Coffee berry disease. In: Colletotrichum±Biology, Pathology and Control (eds J.A. Baily & M.J. Jeger), pp. 237±49. CAB International, Wellingword, UK. Mawardi, S. & Hulupi, R. (1995) Genotype-by-environment interaction of bean characteristics in arabica coffee. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 637±44. ASIC, Paris, France. Mawardi, S. & Soenaryo (1988) The present status of arabica coffee breeding in Indonesia. In: Proceedings of the International Seminar on Coffee Technology, pp. 63±73. Chiang Mai University, Faculty of Agriculture, Thailand. Mazzafera, P., Eskes, A.B., Parvals, J.P. & Carvalho, A. (1989) SteÂrilite maÃle deÂtecteÂe chez C. arabica et C. canephora au BreÂsil. In: Proceedings of the 13th ASIC Colloquium (Paipa), pp. 466± 73. ASIC, Paris, France. Medina-Filho, H.P., Fazuoli, L.C., Guerreiro-Filho, O. et al. (1999) Increasing robusta production in Brazil: the potential of 200 thousand hectares in Sao Paulo State. In: Proceedings of the 18th ASIC Colloquium (Helsinki) pp. 390±95. ASIC, Paris, France. Meyer, A.J.T., Fernie, L.M., Narasimhaswami, R.L., Monaco, L.C. & Greathead, D.J. (1968) FAO Coffee Mission to Ethiopia 1964±65. FAO, Rome. Mitchell, H.W. (1988) Cultivation of the arabica coffee tree. In: Coffee Vol. 4 Agronomy (eds R.J. Clarke & R. Macrae), pp. 43± 90. Elsevier Applied Science, London and New York. Moisyadi, S., Neupane, K.R. & Stiles, J.I. (1999) Cloning and characterization of xanthosine-N7 -methyltransferase, the first enzyme of the caffeine biosynthesic pathway. In: Proceedings of the 18th ASIC Colloquium (Helsinki), pp. 327±331. ASIC, Paris, France. Montagnon, C. & Bouharmont, P. (1996) Multivariate analysis of phenotypic diversity of Coffea arabica. Genet. Res. Crop Evol., 43, 221±7. Montagnon, C., Guyot, B., Cilas, C. & Leroy, T. (1998c) Genetic parameters of several biochemical compounds from green coffee, Coffea canephora. Plant Breeding, 117, 576±8. Montagnon, C., Leroy, T. & Eskes A.B. (1998a) AmeÂlioration varieÂtale de Coffea canephora I. CriteÁres et meÂthodes de seÂlection. Plant. Rech. DeÂvelop, 5, 18±33. Montagnon, C., Leroy, T. & Eskes A.B. (1998b) AmeÂlioration varieÂtale de Coffea canephora. II. Les programmes de seÂlection et leurs reÂsultats. Plant. Rech. DeÂvelop., 5, 89±98. Montagnon, C., Leroy, T., KeÂbeÂ, I. & Eskes, A.B. (1994) Importance de la rouille orangeÂe et facteurs impliqueÂs dans l'eÂvaluation de la reÂsistance au champs de Coffea canephora en CoÃte d'Ivoire. CafeÂ, Cacao, TheÂ, 38, 103±12. Moreno, G. (1994) ContribucioÂn del mejoramiento geneÂtico al desarrollo de la caficultura colombiana. Revista Innovacion y Ciencia, CENICAFE, Colombia, 3, 1±6. Moreno, G., Moreno, E. & Cadena, G. (1995) Bean characteristics and cup quality of the Colombia variety (Coffea arabica) as judged by international tasting panels. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 574±8. ASIC, Paris, France. Moschetto, D., Montagnon, C., Guyot, B., Perriot, J.J., Leroy, T.
Agronomy I: Coffee Breeding Practices
& Eskes, A.B. (1996) Studies on the effect of genotype on cup quality of Coffea canephora. Trop. Sci., 36, 18±31. Ngategize, P.K. (1997) The African Coffee Research Network: prospects and challenges into the next millennium. In: Proceedings of the 17th ASIC Colliqum (Nairobi), pp. 36±5. ASIC, Paris, France. Njoroge, S.M., Morales, A.F., Kari, P.E. & Owuor, J.B.O. (1990) Comparative evaluation of the flavour qualities of Ruiru II and SL28 cultivars of Kenya arabica coffee. Kenya Coffee, 55, 843± 9. Nyange, N.E., Kipokola, T.P., Mtenga, D.J., Kilambo, D.J., Swai, F.B. & Charmetant, P. (1999) Creation and selection of Coffea arabica hybrids in Tanzania. In: Proceedings of the 18th ASIC Colloquium (Helsinki), pp. 356±62. ASIC, Paris, France. Nyange, N.E. & Marandu, E.F. (1997) Improvement of Coffea canephora germplasm in Tanzania: exploration and collection of new robusta material from farmers' plots. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 502±505. ASIC, Paris, France. Omondi, C.O., Hindorf, H., Welz, H.G., Saucke, D., Ayiecho, P.O. & Mwang'ombe, A.W. (1997) Genetic diversity among isolates of Colletotrichum kahawae causing coffee berry disease. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 800± 804. ASIC, Paris, France. Opile, W.R. & Agwanda, C.O. (1993) Propagation and distribution of cultivar Ruiru Il: a review. Kenya Coffee, 58, 1496±508. Orozco-Castillo, C., Chalmers, K.J., Powell, W. & Waugh, R. (1996) RAPD and organelle specific PCR re-affirms taxonomic relationships within the genus Coffea. Plant Cell Rep., 15, 337± 41. Orozco-Castillo, C., Chalmers, K.J., Waugh, R. & Powell, W. (1994) Detection of genetic diversity and selective gene introgression in coffee using RAPD markers. Theoret. Appl. Genet., 87, 934±40. Raina, S.N., Mukai, Y. & Yamamoto, M. (1998). In situ hybridization identifies the diploid. progenitor species of Coffea arabica (Rubiaceae). Theoret. Appl. Genet., 97, 1204±209. Rodrigues Jr, C.J., VaÂrzea, V.M., Godinho, I.L., Palma, S. & Rato, R.C. (1993) New physiologic races of Hemileia vastatrix. In: Proceedings of the 15th ASIC Colloquium (Montpellier), pp. 318±321. ASIC, Paris, France. Rodrigues Jr, C.J., VaÂrzea, V.M. & Medeiros, E.F. (1992) Evidence for the existence of physiological races of Colletotrichum coffeanum Noack sensu Hindorf. Kenya Coffee, 57, 1417±20. Rovelli, P., Mettulio, R., Antony, F., Anzueto, F., Lashermes, P. & Graziosi, G. (1999) Polymorphic microsatellites in Coffea arabica. In: Proceedings of the 18th ASIC Colloquium (Helsinki), pp. 344±347. ASIC, Paris, France. Santa Ram, A. & Sreenath, H.L. (1999) Genetic fingerprinting of coffee leaf rust differentials with RAPD markers. In: Proceedings of the 3rd International Seminar on Biotechnology in the Coffee Agro-industry, Londrina, Brazil (in press). Sreenath, H.L. & Naidu, R. (1999) Coffee biotechnology research in India ± potential progress and future thrust areas. In: Proceedings of the 18th ASIC Colloquium (Helsinki), pp. 281± 94. ASIC, Paris, France.
201
Srinivasan, C.S. (1996). Review: current status and future thrust areas of research on varietal improvement and horticultural aspects of coffee. J. Coffee Res., 26, 1±16. Srinivasan, C.S., Prakash, N.S., Padma Jyothi, D., Sureshkumar, V.B. & Subbalakshmi, V. (1999) Genetic improvement of coffee in India. In: Proceedings of the 3rd International Seminar on Biotechnology in the Coffee Agroindustry, Londrina, Brazil (in press). Srinivasan, C.S. & Vishveshwara, S. (1978). Heterosis and stability for yield in arabica coffee. Ind. J. Genet. Plant Breed., 38, 416±20. Van der Graaff, N.A. (1992) Coffee berry disease. In: Plant Diseases of International Importance Vol. IV: Diseases of Sugar, Forest and Plantation Crops (eds A.N. Mukhopadhyay, J. Kumar, U.S. Sing & H.S. Chaube), pp. 202±30. Prentice Hall, New York. Van der Vossen, H.A.M. (1985) Coffee selection and breeding. In: Coffee: Botany, Biochemistry and Production of Beans and Beverage (eds M.N. Clifford & K.C. Willson), pp. 48±96. Croom Helm, London, New York and Sydney. Van der Vossen, H.A.M. (1997) Quality aspects in arabica coffee breeding programmes in Africa. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 430±38. ASIC, Paris, France. Van der Vossen, H.A.M. & Browning, G. (1978) Prospects of selecting genotypes of Coffea arabica which do not require tonic sprays of fungicide for increased leaf retention and yield. J. Horticult. Sci., 53, 225±33. VaÂrzea, V.M.P., Rodrigues Jr, C.J., Silva, M.C., Pedro, J.P. & Marques, D.V. (1999) High virulence of a Colletotrichum kahawae isolate from Cameroon as compared with other isolates from other regions. Presented at the 18th ASIC Colloquium (Helsinki). Abstract only available. Waller, J.W., Bridge, P.D., Black, R. & Hakiza, G. (1993) Characterization of the coffee berry disease pathogen, Colletotrichum kahawae Sp. Nov. Mycol. Res., 97, 989±94. Walyaro, D.J. (1983) Considerations in breeding for improved yield and quality in arabica coffee (Coffea arabica L). PhD thesis, Agricultural University of Wageningen. Walyaro, D.J. (1997) Breeding for disease and pest resistance and improved quality in coffee. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 391±405. ASIC, Paris, France. Wrigley, G. (1988) Coffee. Tropical Agriculture Series, Longman Scientific & Technical, Harlow, UK. Yapo, A. (1995) AmeÂlõÂoration qualitative de Coffea canephora Pierre par hybridation interspeÂcifique: exploitation d'un nouveau scheÂma de seÂlection chez les arabusta. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 655±62. ASIC, Paris, France. Zamarripa, A., Ducos, J.P., Tessereau, H., Bollon, H., Eskes, A.B. & PeÂtiard, V. (1991) DeÂveloppement d'un proceÂde de multiplication en masse du cafeÂier par embryogeneÁse somatique en milieu liquide. In: Proceedings of the 14th ASIC Colloquium (San Francisco), pp. 392±402. ASIC, Paris, France.
Chapter 10
Agronomy II: Developmental and Cell Biology M.R. Sondahl Fitolink Corporation Mount Laurel, USA T.W. Baumann Institute of Plant Biology University of Zurich, Switzerland 10.1 OVERVIEW The genus Coffea was proposed by Linnaeus in 1735, who later described the species Coffea arabica in 1753, presently known as the variety Typica. Coffea belongs to the Rubiaceae family, which includes more than 500 genera and about 800 species (Bridson & Verdcourt 1988). The Genus Coffea has about 100 species (Charrier & Berthaud 1985), but commercial production relies only on two species, C. arabica and C. canephora, which represent about 70% and 30% of the total coffee market, respectively. Arabica coffee is an isolated species in the genus Coffea because of its amphidiploid and self-pollinating nature, which makes it difficult to incorporate traits from other non-cultivated coffee species. Robusta coffee is a highly self-incompatible species with heterozygous seeds. Both species would greatly benefit from new technologies being developed at the cellular and molecular levels. This chapter deals with the development of cell biology methods and their application to coffee improvement and germplasm preservation. The fluctuation of purine alkaloids during leaf and fruit developmental stages is presented, suggesting a possible evolutionary defense mechanism for coffee species. The synthesis of caffeine and chlorogenic acid, and related regulatory factors are discussed for solid and liquid cell cultures. New advances on cell cultures methods based on embryogenic cell systems for propagation and genetic improvement will be presented. Much progress has been made in the mass production of somatic embryos in bioreactor vessels
and several test cases of scale-up programs are being reported. Finally, the opportunity of capturing in vitro variation for the development of new cultivars is exemplified in this chapter. The enhancement of our knowledge on organ differentiation and its metabolism, and the control of reliable methods for in vitro culture and plant regeneration, is essential for devising new processes for improving coffee plants and its beverage.
10.2 ORGAN DEVELOPMENT AND THE ALLOCATION OF DEFENSE COMPOUNDS 10.2.1 Introduction Studies on the development of coffee organs such as the root, fruit, leaf and flower are, if compared to other `crops', scarce and preferentially concentrate on arabica. Due to technical reasons these investigations were, and still are, largely performed under greenhouse or otherwise controlled conditions. However, despite the `artificial' environment, the studies may render a good and reliable insight into the processes of both flower and fruit development including seed maturation, since the developmental program of these organs yielding finally the essential dispersal unit, the so-called diaspore (Van der Pijl, 1982), remains little or unaffected by external factors. Conversely, the growth of the vegetative plant parts, i.e. shoots, leaves and roots, is subjected to drastic alterations caused by all kind of biotic and abiotic factors. In a simplified manner one
202
Agronomy II: Developmental and Cell Biology
203
may consider generative development as an inert, but stringently controlled process orientated to dispersal and conservation of the entire species, whereas vegetative development is rather a dynamic and plastic event allied to protection and survival of the individual. This view has to be kept in mind when discussing below the organ-related allocation of chemical defense compounds such as purine alkaloids and chlorogenic acids.
10.2.2 The leaf The arabica leaf attains (under greenhouse or phytotron conditions) the full expansion and maximum dry weight after 30 to 35 days (MoÈsli Waldhauser et al., 1997) or 50 to 60 days (Frischknecht et al., 1982), respectively. This means that within 4 to 5 weeks after emergence, the surface growth of the still soft and glossy leaf blade is completed (see Fig. 10.1), but it takes another 2 to 3 weeks until the lamina has gained its final rigidity. Therefore, we can assign several stages or transitions to leaf development, which are characterized not only by means of morphological, but also physiological and phytochemical changes: (1) the quiescent bud (B1 in Fig. 10.2), (2) emergence from the bud (B2 to B4 in Fig. 10.2), (3) lamina expansion and mechanical strengthening, and (4) senescence.
Fig. 10.1 Leaf development of C. arabica. After 30±35 days (ca 5 weeks) leaf expansion is completed. During this period the fresh weight increases by a factor of 360.
(1)
In the quiescent bud (B1 in Fig. 10.2), the apical meristem together with the paired leaf primordia is covered by two firm stipules. Additionally, between the primordia and the stipules there is a resinous layer, of which remainders may still be attached to the leaf apices when emerged from the bud. With respect to purine alkaloids all structures, the primordial leaflets, the resin layer, and the stipules, exhibit a wide variation of
Fig. 10.2 Emergence of coffee (C. arabica) leaflets from the bud (taken from Frischknecht et al., 1986.
concentration from bud to bud (Fig. 10.3), indicating only moderate significance of chemical defense in favor of mechanical protection (Frischknecht et al., 1986). (2) Emergence from the bud is depicted in Fig. 10.2 (B2 to B4). The leaf pair develops to stage 4 within a few days, pushing apart the stipules, and with the leaves still tightly associated to each other at this stage. During this developmental process, the concentration of purine alkaloids markedly increases (up to 4% at stage B3) while their coefficient of variation decreases likewise, for example the latter is three times lower for caffeine at stage B3 than at stage B1. Clearly, caffeine biosynthesis is strongly accelerated during leaflet emergence and has been reported to reach a maximum rate of 17 000 mg dayÿ1 gÿ1 at stage B3 (Frischknecht et al., 1986). The considerably lower variation coefficient signifies that chemical defense by purine alkaloids has become a very important, stringent factor at the moment of emergence when the leaflets lack the mechanical protection by the stipules and have a very soft texture. Additionally, their dietary value for predators is very high at this stage. (3) During the next following expansion stage of the leaf blade, the rate of caffeine synthesis falls
204
Coffee: Recent Developments
Fig. 10.3 Alkaloid content (theobromine and caffeine) of seven individual buds. Segments give the relative parts of stipules, resin layer and leaflets (after Frischknecht et al., 1986).
exponentially, reaching 16 mg dayÿ1 gÿ1 when the leaf is fully grown with respect to leaf area and photosynthetic capacity (Frischknecht et al., 1982). In parallel, the nutritional value per unit leaf area drops. The activity of the enzymes mediating the last steps of caffeine synthesis has been investigated during leaf expansion. The biosynthesis of caffeine in coffee most likely starts by the methylation of XMP (xanthosine monophosphate) at position N7 carried out by the first N-methyltransferase (NMT). After removal of the phosphoribose moiety, the resulting 7methylxanthine is further methylated via theobromine to caffeine by the second and third NMT respectively (Schulthess et al., 1996). For a comprehensive review on caffeine metabolism, see Ashihara & Crozier (1999). The activities of the second and third NMT are presented in dependence of leaf fresh weight in Fig. 10.4. As expected from the above-mentioned sharp increase of purine alkaloids, they both show a very high peak activity when the leaflets have completed their emergence from the protective bud. Thereafter, the activities decrease rapidly. The relative caffeine content drops as a consequence of
Fig. 10.4 Time course of NMT activities during leaf development in C. arabica. Leaves (5 to 70, depending on leaf size) of each stage were pooled and extracted. The youngest stage harvested had a fresh weight of ca 5 mg, the oldest of 1900 mg. The latter corresponded to the fully expanded, but still glossy and soft leaf, 30 to 35 days old, as shown in Fig. 10.1. (a) Activity per g leaf fresh weight; inset: purine alkaloid contents related to fresh weight (b) Total activity (pkat) per leaf; inset: absolute amounts of the purine alkaloids theobromine and caffeine per leaf. The data presented in the insets are from a separate experiment with 11 leaf classes and not from the leaf series that was used for the determination of the enzyme activities. (Taken from MoÈsli Waldhauser et al., 1997.)
`dilution by growth'. However, the absolute amount of caffeine increases steadily because of low enzyme activities persisting throughout the entire period of leaf expansion (MoÈsli Waldhauser et al., 1997). At the end of leaf expansion, the net photosynthesis (NPS) has attained a maximum rate which remains stable further on, while dark respiration gradually falls and thus the two events
Agronomy II: Developmental and Cell Biology
(4)
in combination confer to an optimum dry matter production after 50 to 60 days of development. (Frischknecht et al., 1982. Note: erroneously, in this publication the dry weights listed in Table 1 are too small by a factor of ten!) To our knowledge, the related mechanical stabilization of the coffee leaf has not been investigated in depth. Certainly, the walls of the mesophyll cells will get thicker by the deposition of cellulose and the vascular system is fortified by phenolics. Moreover, the appearance of the upper surface of the lamina turns from glossy into dull, indicating a change preferably including a thickening of the waxy layer and/or cuticula. We should finally mention that during leaf development the chlorogenic acids are allocated in parallel to the purine alkaloids (Aerts & Baumann 1994). Similar to the NMTs catalyzing caffeine biosynthesis (see Fig. 10.4), the activity of the key enzyme of phenylpropane synthesis, phenylalanine ammonia lyase (PAL), is very high in the young leaflets and decreases during the further expansion (Aerts & Baumann, 1994). The concerted formation of both the alkaloids (mainly caffeine) and chlorogenic acids (mainly 5-caffeoylquinic acid; 5-CQA) has a physiological significance: caffeine, which easily permeates through all kind of biological barriers, is physico-chemically complexed by 5CQA and thus, compartmented in the cell vacuole in order to avoid autotoxicity (MoÈsli Waldhauser & Baumann, 1996). However, these processes eventually result in a coffee leaf in which these phytochemicals are not evenly distributed in the lamina. Preliminary investigations (Wenger & Baumann, unpublished data) revealed that both chlorogenic acids and purine alkaloids are considerably enriched at the leaf margin and sharply decrease in concentration towards the mid-vein. Conceivably, this `phytochemical leaf architecture' has an ecological significance: the leaf margin, a preferential site of insect attack, is particularly well furnished with these defense compounds. This insight could be of the upmost importance in modern breeding if phytochemical leaf architecture is genetically based. Senescence of the coffee leaf is not yet well investigated. The pioneering studies of Weevers (1907) may shed some light on the behavior of caffeine during aging: in his studies the adult leaf accumulated a maximum total amount which subsequently decreased by 30 to 50% in the old but still green leaf. He collected the leaves from
205
plants growing in the `tropical greenhouse' in Amsterdam, and grouped them into `very young', `young', `adult' and `old'. Because of the long time span between the leaf classes `adult' and `old', their possibly divergent `life histories' (e.g. with respect to the light regime and other environmental factors), and merely degradation or export of caffeine may be responsible for the differences. Additionally, he tried to investigate the `caffeine status' in the leaf at shedding and recognized that only naturally aged leaves became caffeine-free, this in contrast to infected (e.g. Hemileia) leaves which still contained caffeine when shed. Despite these findings, it remains extremely difficult to tackle experimentally the problem of caffeine disappearance from the senescing arabica leaf. For instance, no metabolites were detected by Ashihara et al., (1996) after feeding ring-labeled caffeine to mature coffee leaves, except a trace of CO2 (0.03% of the applied radioactivity). Similarly, leaves still attached to the plant and fed with doubly labeled caffeine export only about 1% of the applied activity into the other leaves within one week (Baumann & Wanner, 1972). However, the recovered activity was about one fourth of the applied activity, indicating some catabolism in old leaves as it was found before in the classical investigations by Kalberer (1964, 1965). After feeding either [2-14 C], [8-14 C], [1-methyl-14 C], or [7-methyl-14 C] caffeine to aging arabica coffee leaves he always detected, besides allantoin and CO2 , the same three unknown degradation products which we speculatively classify as methylated ureides. The pathway from caffeine to these unknown compounds or to allantoin also remained obscure in more recent studies (Mazzafera et al., 1994; Ashihara et al., 1996). So far, neither a caffeine demethylase activity could be measured nor radioactivity in uric acid after feeding labeled caffeine could be detected in Arabica leaves (VitoÂria & Mazzafera 1999, and references cited therein). In conclusion, coffee leaf senescence and mobilization of related compounds certainly needs further investigation.
10.2.3 The fruit Fruit development in coffee, which covers the time between anthesis and full ripening, takes between 2 to 3 (e.g. C. racemosa) and 14 (e.g. C. liberica) months, depending on the species, genotype, climate and cultivation. The species of economic value, C. arabica
206
and C. canephora, require 6 to 8 and 9 to 11 months for maturation, respectively (for related references see Guerreiro Filho, 1992). Almost 30 years ago, we performed a time-consuming investigation of fruit development in arabica coffee, whereas growth and alkaloid content were followed over a period of more than 7 months in both the pericarp and seed (Keller et al., 1971, 1972). Fruit development was divided into 11 stages (Fig. 10.5), as they are characterized in the legend, whereas at stage 1 (38 mg fresh wt) the separation into pericarp and seed tissue was not yet practicable. As depicted by the solid line, the dry wt of the pericarp exhibits a biphasic course due to the final ripening process (stages 8 to 11). The dry weight of the seeds increases gradually to reach a maximum value at stage 8 with an already tough endosperm texture. The time courses of the absolute amount of caffeine in the two fruit tissues is most remarkable. In the seed, it parallels more or less the dry weight curve, meaning that the relative caffeine content remains unchanged (> 1%) over the entire period of seed development.
Fig. 10.5 Absolute caffeine contents and dry weights during fruit development of C. arabica. (After Keller et al., 1972. Stages 1 to 11 (fresh wt in mg) are characterized as follows: 1 (38), separation into pericarp and seed tissue not possible, 1 to 2 weeks; 2 (240) green, 2 to 3 weeks; 3 (400) green, 3 weeks; 4 (800) green, 4 weeks; 5 (1200) green, 5 weeks; 6 (1180) green, endocarp hard, 2 to 3 months; 7 (1080) green, 4 months; 8 (1600) light-green/olive, mesocarp slightly fleshy, endosperm tough, 5 to 6 months; 9 (2180) exocarp partially reddish, mesocarp very fleshy, endosperm very tough, 5 to 6 months; 10 (2160) exocarp bright red, mesocarp very fleshy, endosperm very tough, 6 months; 11 (1800) exocarp dark red, mesocarp slightly dry, endosperm very tough, 7 to 8 months.
Coffee: Recent Developments
Unfortunately, we did not differentiate between the transient perisperm and the real endosperm as they are characteristic of coffee bean development (Carvalho et al., 1969). In the pericarp, however, the allocation of total caffeine stops early, that is already at stage 5, and hardly surpasses the total of 1 mg. Because the dry weight subsequently increases considerably, caffeine is diluted during fruit ripening to reach a final concentration of ca 0.24% (dry wt) in the fleshy pericarp. This is in contrast to stages 1 and 2 with a caffeine content of 2.2 and 1.7% respectively, distinctly higher than of a ripe arabica coffee bean. Interestingly, the cessation of further caffeine accumulation in the pericarp coincides with the formation of a hard endocarpic tissue (completed at stage 6), indicating a shift from chemical to mechanical defense. Clearly, the biochemical changes of the pericarpic tissues during ripening are directed to diaspore dispersal by animals (zoochory): the tough endocarp has to protect the seed from digesting enzyme activities in the gut of the frugivores such as birds or mammals, and the fleshy, sugar-containing (Urbaneja et al., 1996) mesocarp low in caffeine acts as a reward, while the vivid coloration, due to anthocyanins (Barboza & Ramirez-Martinez, 1991), of the exocarp is to attract the dispersing animal. By application of doubly-labeled caffeine to the epidermal layer of the pericarp of young coffee fruits attached to the tree, a considerable alkaloid transport into the endosperm tissue could be shown (Baumann & Wanner, 1972). The ability of the endosperm to form caffeine is indirectly testified by the fact that the very young, `liquid endosperm' of C. arabica served as an ideal source for the isolation of N-methyltransferases (Mazzafera et al., 1994; Gillies et al., 1995) catalyzing the last steps in caffeine biosynthesis. Moreover, endosperm tissue of over 6-month-old fruits, i.e. ca stage 8 in Fig. 10.5, exhibited distinct methyltransferase activity to synthesize theobromine and caffeine (Mazzafera et al., 1994). However, the ability of the coffee endosperm to form caffeine de novo has to our knowledge never been tested. There is one report dealing with radioactive precursor feeding to arabica endosperm, however, the authors (Keller et al., 1972) did not realize at that time that they actually fed the preceding, transient perisperm as one may conclude from the related fruit fresh weight of only 500 mg (about 25 days old). Nevertheless, the results were most intriguing: de novo caffeine synthesis in the perisperm is about 2.6 times higher than in the pericarp as determined by the incorporation of 14 CO2 in the presence of light. Conversely, methylation as estimated from the incorporation of radioactivity of [methyl-14 C]
Agronomy II: Developmental and Cell Biology
methionine in the light is much higher (256) in the pericarp than in the perisperm. In the pericarp, light increases the methylation by a factor of 10 as compared to the dark condition. Unfortunately, the influence of light on caffeine formation (methylation) in the slightly greenish perisperm was not tested. (The background to light-dependent stimulation of caffeine biosynthesis will be discussed in the next section). It would be of the upmost importance to know the fate not only of the purine alkaloids but also of other secondary compounds such as chlorogenic acids allocated to the perisperm. Are they conserved and transported to the developing endosperm? Has the large fraction of dicaffeoylquinic acids present in the perisperm of arabica (Schulthess & Baumann, unpublished data) a significance similar to the occurrence of cyanogenic diglucosides in the seeds of many plant species (Selmar et al., 1988), namely to facilitate apoplastic transport by avoiding enzymatic degradation as required during seed development and/or germination? There are many other developmental key processes eventually leading to the lovely coffee bean that were completely neglected by science despite the high economic value of this product. However, very recently a remarkable investigation regarding the changes of various components during robusta and arabica seed development has been undertaken (Rogers et al., 1999). The authors followed various parameters such as sugars, polyols, organic acids and a few anorganic anions. Strikingly, the perisperm, which when fully developed fills the entire cavity later kept by the endosperm, is not only a `placeholder' but most likely also the full provider of sugars and organic acids for the endosperm. This view is essential for modern coffee breeding and implies that the maternal tissue (perisperm) controls not only the seed size but also the quantity and, to some extent, also the quality of the latter coffee bean (endosperm), this all reminding of the parental conflict theory (Grossniklaus et al., 1998).
10.3 PURINE ALKALOID FORMATION IN COFFEE CELL CULTURES 10.3.1 Introduction Even though the worldwide demand for caffeine in soft drinks has increased greatly in recent years, its production by plant tissue culture will not be economic because of the related costs. Nevertheless, caffeine formation by coffee cells deserves special attention, since this system has become a `standard of excellence'
207
in designing and testing plant bioreactors for the production of secondary plant substances, as reviewed by Prenosil et al., (1987). The reasons are as follows: coffee cell cultures are easily established; they grow well on an artificial medium; they readily produce secondary compounds such as purine alkaloids and chlorogenic acids; the former diffuse into the culture medium allowing their direct determination, for example by HPLC, and thus a quick estimate of the alkaloid productivity and the product formation can be rapidly influenced by various factors, whereby light is experimentally the most prominent. This last reason points to an additional value of tissue culture exemplified below, namely its frequent use instead of difficult-to-handle plant organs and to facilitate many of the biochemical and physiological investigations. The coffee species used in these studies was preferentially arabica, though canephora is equally suitable or even superior with respect to alkaloid formation (Baumann & Frischknecht, 1982). However, the genetical homogeneity of arabica is an excellent prerequisite for such tissue culture investigations. So far, growth of purine alkaloid formation by cultured tissue of genera other than Coffea has been found, with the exception of Paullinia cupana (Baumann & Frischknecht, 1982), to be distinctly smaller, for example Camellia sinensis (Ogutuga & Northcote, 1970; Baumann & Frischknecht, 1982; Shervington et al., 1998) and Theobroma cacao (Baumann & Frischknecht, 1982; Gurney et al., 1992). Some of the older tissue culture work has been reviewed by Baumann & Frischknecht (1988a, b).
10.3.2 Callus culture Caffeine formation by callus cultures was first reported almost 30 years ago (Keller et al., 1971, 1972). The authors derived primary cultures from arabica coffee fruit transsects and revealed that after 4 to 5 weeks of cultivation 95 to 98% of the caffeine was present in the solid agar medium, most likely due to diffusion. Since the amount of caffeine detected in the whole culture increased by a factor of 6 and the biomass only by a factor of 2, it was hypothesized by Keller et al., (1972) that in vivo the caffeine synthesis is inhibited by product formation, and that in vitro the diffusion of caffeine into the medium diminished the inhibitory effect. Subsequently, the productivity of primary as well as of subcultures derived from arabica stem segments was investigated in detail by Frischknecht et al., (1977). Caffeine formation paralleled the increase of callus dry weight suggesting a metabolic connection of
208
purine alkaloid synthesis to growth processes. Moreover, tissue caffeine concentrations of 900 to 1000 mg/ ml (ca 5 mM) and higher inhibited both callus growth and caffeine formation. The cultures grew in the dark which may explain why a caffeine concentration (5 mM) much lower than, for example, in a young coffee leaf (up to 50±60 mM) was inhibitory. In situ autotoxicity of caffeine is avoided by vacuolar complexation with chlorogenic acids mainly formed in the light (see below).
10.3.3 Suspension culture Suspension-cultured coffee cells were first established by Townsley's group (Townsley, 1974; Van de Voort & Townsley 1974, 1975; Buckland & Townsley 1975), which recognized the potential of these cultures to maintain the biosynthetic capacity for compounds unique to the parent plant. Besides purine alkaloids the formation of sterols, fatty acids, chlorogenic acids and coffee aromatics was reported. Subsequently, the growth of coffee cells in suspension was optimized and alkaloid production thoroughly investigated (Frischknecht & Baumann, 1980). The latter was shown to be highest towards the end of the growth phase when the cells stop dividing and start to expand. This is in accordance with maximum alkaloid formation in the young coffee leaflets (see Section 10.2.2), when the cells expand rapidly. Additionally, the biotransformation capacity was tested using radioactively labeled theobromine (Frischknecht & Baumann, 1980). Finally, the related N-methyltransferase activities were measured during culture growth and, most interestingly, were found to be highest during the (mitotic) growth phase. At this stage the cells produce a short supply of purine rings. Later on, when the primary metabolism is decreasing, a surplus of purine metabolites meets with comparably low methyltransferase activities (Baumann et al., 1983). In short, purine alkaloid metabolism in suspension-cultured cells is per se not well coordinated, but can be optimized by several means as outlined below. It was found that productivity correlated with the selected cell aggregate type: the culture of the small aggregate type forms less purine alkaloid than the large aggregate type (Frischknecht & Baumann, 1980). This phenomenon led to the idea to entrap and immobilize coffee cells in alginate beads to form artificial aggregates under controlled conditions (Haldimann & Brodelius, 1987). Indeed, the immobilization resulted in a considerable increase in alkaloid production. It is difficult to explain why the formation of large cell
Coffee: Recent Developments
aggregates or immobilization should stimulate caffeine biosynthesis. One explanation is based on the stress concept, which postulates a highly modulating effect of external factors on secondary plant metabolism considered to have an ecological significance so as to improve adaptation to unfavorable conditions. Therefore, environmental stress conditions such as high temperature, UV radiation, low water potential and wounding (phytophagy) are expected to enhance accumulation of qualitative defense substances such as alkaloids, cardenolides, glucosinolates, and others. Similarly, the above-mentioned large cell aggregates may, in contrast to single cells or small aggregates, have suffered from a nutrient stress since the transport of the medium components into the core of the aggregate was impeded. Stress was first introduced into tissue culture in 1985 by Frischknecht and Baumann. They used suspension-cultured cells of arabica coffee which were exposed to either high light intensity and/or high salt (NaCl). The former was most effective and induced a 100-fold stimulation of caffeine synthesis, i.e. 450 mg/ l, in the small aggregate type culture. The formation of purine alkaloids was also enhanced by the application of ethephon (Cho et al., 1988; Schulthess & Baumann, 1995) or adenine (Schulthess & Baumann, 1995). The combination of both applied to dark-grown suspension cultures resulted in an 11-fold increase (Schulthess & Baumann, 1995). In a photoperiod, as compared to the control culture in the dark, caffeine formation was stimulated by a factor of 21, which was not additionally increased by the above-mentioned stimuli. Conversely, the combination of photoperiod and ethephon led to a drastic reduction of ca 50 to 60% in the formation of both caffeine and chlorogenic acid. It was concluded that caffeine formation is dependent on chlorogenic acid accumulation. If the latter is impaired, deficient caffeine complexation results in the inhibition of the purine alkaloid biosynthesis. Baumann & RoÈhrig (1989) visualized the vacuolar localization of the chlorogenic acids in arabica suspension-cultured cells. They found that due to complex formation, caffeine is intracellularly accumulated to a certain extent, which depended on the chlorogenic acid (5-CQA) concentration in the cells. The physico-chemical and metabolic interdependence between purine alkaloids and chlorogenic acids, which is valid also for the living plant, was investigated in detail by MoÈsli Waldhauser and Baumann (1996) using suspension-cultured cells of C. arabica. By means of various conditions such as the addition of a photoperiod or methyljasmonate (both stimulating the
Agronomy II: Developmental and Cell Biology
synthesis of caffeine and chlorogenic acids), the application of a potent inhibitor of PAL, the key enzyme in the phenolic pathway eventually leading to chlorogenic acids, and of exogenous caffeine, they created metabolic situations shedding light on the above-mentioned interdependence: compartmentation of caffeine (and also that of theobromine) is highly correlated to the concentration of chlorogenic acids and relies exclusively on the physical chemistry of the complex; moreover, there is a regulatory connection between the complex partners, possibly guided by the cytoplasmic caffeine concentration. Since experimental inhibition of the chlorogenic acid synthesis drastically inhibited caffeine biosynthesis, the authors came to the conclusion that lowering the bean caffeine content by means of genetic engineering could be achieved by changing the expression not only of the caffeine, but also of the chlorogenic acid pathway.
10.4 NEW ADVANCES IN CELL AND ORGAN CULTURE 10.4.1 Brief review of the literature Detailed literature reviews of pioneer work on coffee tissue culture have been already published by Sondahl et al., (1984), Sondahl & Loh (1988), Dublin (1991) and Sondahl & Lauritis (1992). The ability to induce large quantities of somatic embryos, subsequent germination of these non-sexual embryos and recovery of normal coffee plants are techniques of paramount importance for multiple applications in coffee improvement programs. Some key reports that led to this development include the pioneer work with robusta shoot cultures (Staritsky, 1970), high-frequency embryogenesis from mature leaf explants of arabica (Sondahl & Sharp, 1977), production of somatic embryos from leaves of arabusta hybrids in auxin-free medium (Dublin, 1981), and somatic embryogenesis from young leaves of arabica (Yasuda et al., 1985). All the above protocols were based on in vitro solid cultures developed during a 20-year period, but in the early 1990s, progress was made with embryogenesis in liquid cultures. A liquid culture protocol for a highly synchronized somatic embryo production, based on a modified version of the two-step somatic embryogenesis method of Sondahl and Sharp (1977), was published by Neuenschwander and Baumann (1992). Large numbers of robusta somatic embryos were produced in 3-liter bioreactor cultures by Zamarripa et al.,
209
(1991a) soon after, followed by the work of Noriega and Sondahl (1993) with arabica embryos using a 5-liter bioreactor system. Using a unique apparatus for a temporary immersion culture, a protocol for the development of coffee plantlets was reported by Berthouly et al., (1995a). These solid and liquid medium protocols for coffee somatic embryogenesis have provided the key for a series of applications in coffee improvement programs such as micropropagation, gene transfer, in vitro mutagenesis and selection, germplasm preservation (cryopreservation) and biochemical studies. Small clumps of cells can be indefinitely maintained in liquid suspension medium through periodic subcultures (3-day intervals). Cultures of such undifferentiated cells are useful for metabolic studies of aromaproducing compounds (Townsley, 1974), lipid synthesis (Van de Voort & Townsley, 1975), purine alkaloids (Neuenschwander & Baumann 1991) and other studies. Liquid cultures of specialized cell lines, like embryogenic tissue, are used either for protoplast isolation, gene transfer, or for mass propagation. Protoplasts are single cells without a cell wall, maintained in a high-osmoticum medium. Initial work with coffee leaf protoplasts led to production of microcolonies (Sondahl et al., 1980, Orozco & Schieder 1982) but later, somatic embryos were reported from protoplasts isolated from in vitro robusta embryos by Schoepke et al., (1987). More efficient protocols for coffee protoplast isolation and regeneration were reported when protoplasts were isolated from embryogenic cells growing in liquid media (Acuna & Pena, 1991; Spiral & Petiard, 1991). Protoplast culture is an ideal system for synthesis of somatic hybrids between distant related species and for gene transfer via DNA uptake by these naked cells.
10.4.2 New advances (a) Somatic embryogenesis Acuna (1993) presented new data for the production of embryogenic tissue (ET) in two selected genotypes and two culture media. The F5.305 line in combination with NAR 12 medium resulted in the induction of ET in 93% of the explants after 2 months in culture without any subculture. The most suitable explant was soft young leaves from new suckers after pruning. The NAR 14 medium consisted of 14 strength of Murashighe & Skoog (1962) MS macro salts, 12 strength MS micro salts, B-5 organic constituents; sucrose (30 g/l) and 2ip (1 mg/l).
210
Scanning electron microscopy (SEM) studies on coffee embryogenic tissues and early stages of embryo differentiation have been previously reported by Sondahl et al., (1979) and Nakamura et al., (1992). More recently, an interesting SEM study was made by Tahara et al., (1995) using three types of coffee calli (C. arabica), one embryogenic callus (EC) and two nonembryogenic calli (yellow callus, NYC; white callus, NWC), maintained on MS medium with 2,4-D (10 mM). EC was composed of yellow, spherical cytoplasm-rich cells, uniform in size; NWC displayed elongated or swollen translucent cells; NYC had cells similar in appearance to EC cells, but more dispersed. It was observed that EC, in the presence of 2,4-D, was incapable of regenerating somatic embryos, but after transfer to a 2,4-D-free medium, globular stage embryos arose after 2 to 3 weeks. The authors reported that EC preserved its embryogenic potential when maintained in 10 mM 2,4-D medium for 6 years. These data demonstrate that 2,4-D is able to induce the proliferation of embryogenic-competent cells, but its presence inhibits the regeneration process leading to somatic embryo production. The effect of asparagine on coffee somatic embryo induction has been demonstrated by Nishibata et al., (1995). Embryogenic cell lines of C. arabica were maintained on 10 mM 2,4-D medium (growth medium) and were regularly transferred to a 5 mM 2-iP medium (regeneration medium) for the production of somatic embryos. The addition of asparagine (10 mM) to the regeneration medium promoted embryogenesis, while the addition of glutamine, glutamate or aspartate strongly inhibited somatic embryogenesis. Moreover, the addition of asparagine (10 mM) to 2,4-D growth medium was able to induce somatic embryos and inhibit further cell proliferation. The effects of plant growth regulators on somatic embryogenesis of leaf cultures of C. canephora were reported by Hatanaka et al., (1995). It was demonstrated that cytokinin (5 mM) was essential for the formation of somatic embryos in robusta leaf cultures and that 2-iP was the most effective cytokinin source. It was described that when only half of the leaf discs were immersed vertically into the solid medium, embryos developed only on the cut edges of the discs that were in contact with the medium. On the other hand, each auxin tested (IAA, IBA, NAA, 4-FA, 2,4-D) inhibited somatic embryogenesis in proportion to its concentration. The authors also evaluated the effect of ethylene and found that at a concentration of 12 ml/l somatic embryogenesis was promoted, but at 6 ml/l ethylene had no effect and at 24 ml/l ethylene had an inhibitory effect.
Coffee: Recent Developments
Culture conditions for induction of somatic embryogenesis in arabica and robusta tissues have been reported by Yasuda et al., (1995). Using young leaf explants, both species produced somatic embryos on A3 cytokinin-only medium (5 mM 2-iP or BA), but there was a different reaction in culture according to the genotype. In robusta cultures, somatic embryos arose immediately on the cut edges of young leaf explants in contact with the medium; if auxin was added, embryo formation was inhibited. In arabica cultures, embryogenic calli were induced only after a long time (16 weeks). If the embryogenic tissue was transferred to 10 mM 2,4-D medium, white nonembryogenic calli and yellow embryogenic calli were produced. The yellow calli could be maintained in continuous proliferation, retaining their ability to form somatic embryos for more than 4 years, upon transfer to a cytokinin or auxin-free medium. The induction of somatic embryogenesis was tested with ten F1 hybrids made between commercial arabica cultivars and wild genotypes from Ethiopia using young leaf explants (Etienne et al., 1997). Embryogenic cells were produced after 6 months on solid cultures, multiplied in Petri dishes, and transferred to 125 ml Erlenmeyers to establish embryogenic cell suspensions at 100 rpm and 278C, with subculture intervals of 10 weeks. Young somatic embryos were transferred to RITA1 vessels under periodic immersion technique for embryo germination and plantlet development. Plantlets with one pair of leaves and a tap root were obtained after 3 to 4 months of cultivation in RITA vessels. A genotypic differential response to somatic embryogenesis was observed among the F1 hybrids. In case of a high-embryogenic material (Family 1/hybrid 1), up to 9000 plantlets were obtained per RITA vessel, but in the case of a low-embryogenic hybrid, only 750 to 1000 plantlets were obtained per vessel. The authors emphasized the suitability of this method for largescale propagation of F1 hybrids. A critical study on `direct or low' somatic embryogenesis induction from arabica leaf explants was presented by Loyola-Vargas et al., (1999). Using soft leaves from in vitro plantlets on Yasuda et al.'s (1985) medium, somatic embryos were observed directly from mesophyll cells of the explants after 21 days. No embryogenic tissue (friable calli, embryogenic calli) were observed in these cultures. Single isolated embryos were transferred to germination conditions and more than 700 plants were produced under greenhouse conditions. No morphological differences were observed among regenerated plants, suggesting that there is no visible somaclonal variation in this
Agronomy II: Developmental and Cell Biology
coffee cloned population. The authors studied the effect of nitrogen on coffee somatic embryogenesis and suggested that total levels of 4 to 9 mM give a maximum response. The optimum ratio of nitrogen sources should be 1 NO3 :2 NH4 for a maximum response.
(b) Cryopreservation of embryos Coffee germplasm has to be maintained under field conditions due to the short life of viable seeds and the difficulty of applying long-term conservation techniques to coffee seeds. Cryopreservation of somatic embryos under liquid nitrogen (71968C) may offer one alternative for back up preservation of valuable germplasm. The ability of arabica zygotic embryos and robusta somatic embryos to withstand freezing into liquid nitrogen was evaluated by several freezing methods by Florin et al., (1993). Zygotic embryos could be cryopreserved after a controlled drying under 43% relative humidity (RH) at 188C; after thawing, normal development of zygotic embryos was observed. With the same method, or with a simple method based on a sucrose pretreatment, followed by prefreezing at 208C, regrowth of somatic embryos was observed via a secondary embryogenesis process. Normal plants were obtained from C. arabica (cultivars Catuai and Caturra) and from C. canephora and arabusta hybrid. In subsequent work, Florin et al., (1995) evaluated three preservation techniques for robusta somatic embryos. It was found that hydrated embryos can be preserved at 208C for 1 to 2 months. Partially dehydrated embryos could be stored in liquid nitrogen for an indefinite period of time, and such frozen embryos were able to develop into plantlets similar to controls. It was also reported that coffee embryos could be dehydrated and stored at 15 to 248C under 43% RH for at least 1 month. The survival rates of alginate-coated robusta somatic embryos before and after freezing in liquid nitrogen were reported by Hatanaka et al., (1995). It was found that the critical dehydration was 13% and that below this level the embryos suffered desiccation injury. Unfrozen embryos had 77% recovery and frozen embryos a maximum of 66% survival rate. It was reported that more than half of the revived somatic embryos after cryopreservation developed shoots and roots directly (no callus or secondary embryos) within 50 days of thawing, and similar results were observed after 8 months of cryopreservation. To avoid the long process of inducing somatic embryos for cryopreservation purposes, Dussert et al.,
211
(1997) proposed freezing the entire seeds and then excised the zygotic embryos after thawing. Seeds must be dehydrated from the original 0.5 down to 0.2 g H2 O gÿ1 dry weight (dw), surface sterilized, and slowly precooled to 7508C prior to immersion in liquid nitrogen. After thawing, 70% of the excised embryos were viable and could develop into normal plantlets (only 30% for germination of intact seeds).
(c) Protoplast culture Successful reports on coffee protoplast isolation and plantlet regeneration have been achieved by the use of embryogenic cells (Acuna & Pena 1991; Spiral & Petiard 1991). A simple protoplast protocol using only one medium for cell wall regeneration, microcolony formation and regeneration has been described by Yasuda et al., (1995). Protoplasts were isolated from embryogenic tissues of arabica coffee and cultivated in an A3 medium supplemented with 10% coconut water, mannitol (0.3 M) and 5 mM BA. The first cell division was observed after 3 weeks, when liquid A3 medium without mannitol was added. After 2 months of culture, somatic embryos were present and they developed into normal plantlets after subsequent subculture.
(d) Transformation and regeneration To perform a gene transfer to modify the genetic makeup of a plant species, one must successfully incorporate a gene cassette (desirable gene plus introns, promotor and terminating sequences) in a plant cell and subsequently recover a modified plant from such a single modified cell. Transformation deals with the techniques of `gene insertion' and regeneration refers to in vitro processes leading to the recovery of viable, positive transformed plants. The success rate of gene transfer varies with the in vitro cell culture system and the method utilized for gene insertion. Usually, the target cells are active proliferating embryogenic cells (embryogenic tissue) which would have the cell wall removed if the gene transfer method requires protoplast cultures. Basically, there are three main methods for gene transfer: (a) DNA uptake by protoplasts; (b) accelerated particles coated with DNA; and (c) cocultivation with Agrobacterium tumefaciens (or rhizogenes). The first two methods are physical methods and the third one is a biological method, mediated by the bacterial vector, which has the natural ability to penetrate plant cells and transfer DNA segments to the nucleus. Positive transgenic events (cells carrying foreign DNA) are usually at the 1 to 5% rate and from
212
these cells, normal plants must be recovered. Transformation/regeneration protocols rely heavily on the in vitro regeneration protocol to recover a few positive transgenic cells. In practical terms, a gene transfer project must recover about 100 positive transgenic plants in order to make the selection of the best individuals. Positive transgenic plants may carry the new DNA make-up but it is not expressed, or it is poorly expressed. Sometimes, positive transgenic plants are in vitro or somatic mutated individuals and so they must be screened out at the plant level. Detailed information on molecular biology and gene transfer studies on coffee is covered in Chapter 11 in this book. Here, we will briefly review the in vitro methods of transformation/regeneration using coffee cultures. Two gene transfer methods have been used with coffee solid cultures. Barton et al., (1991) electroporated coffee protoplasts in an attempt to transfer a genetic marker (NPT II) for protocol development. The authors reported the subsequent recovery of callus and embryos from these experiments, but confirmation of the presence of an NPT II marker incorporated at the plant level has not been provided. Spiral & Petiard (1993) reported transformation/ regeneration in three genotypes (robusta, arabica (cv Red Catuai) and arabusta #1307) using Agrobacterium rhizogenes charged with GUS and NPT II gene markers and 35-S and NOS promotors. Coffee somatic embryos were co-cultivated with A. rhizogenes for 1 hour and transferred to an embryogenic medium for 11 days. After this period, embryos were subcultured to a medium containing 200 ml of cefaloridine to kill the bacteria. After 3 weeks, callus, hairy roots and embryos were observed. The authors reported a high frequency of positive transformation (+10%). Roots with a positive GUS reaction were subcultured on embryogenic medium and after 4 weeks, fresh somatic embryos were produced. After 2 months of culturing these rootderived embryos, coffee plantlets were recovered which allowed for the positive evaluation of gene integration (GUS and NP II). The integration of several copies of the same gene at the plantlet level was reported. The stability of integrated DNA and the presence of multi-gene copies would be traced to subsequent generations. Sugiyama et al., (1995) reported positive transformation/regeneration in arabica tissues using the wild type of Agrobacterium rhizogenes strain IF 14554. Cotyledon fragments produced callus (48%) and hairy roots (39%). Hypocotyl tissues formed only callus (95%) and leaf explants produced small numbers of hairy roots and callus. Callus tissues were non-
Coffee: Recent Developments
embryogenic, but small number of somatic embryos were recovered from hairy roots after 6 months in culture in a 2 mM 2-iP medium. After these rootderived embryos were transferred to hormone-free medium, plantlets were obtained. Positive transgenic plantlets were cloned by nodal culture and the resulting plantlets continued to express the same transformed phenotype of the donor tissues.
(e) In vitro selection studies In vitro selection using either solid or liquid cultures in the presence of phytotoxins offers the possibility of recovering cell lines and plantlets resistant to pathogen, if there is a correlation between toxin resistance and in vitro resistance. This in vitro selection system explores the natural variability existing in somatic cells and/or variability induced during the in vitro culture conditions. In several cases, a positive correlation between toxin resistance in vitro and in vivo pathogen resistance has been demonstrated (Hartman et al., 1984; Hammerschlag, 1990). Coffee berry disease (CBD) is caused by the fungus Colletotrichum coffeanum and constitutes a very serious limitation for coffee production in many African countries with losses of up to 50% if not controlled by frequent fungicide spraying (Griffiths et al., 1971) (see also chapter 9). Nyange et al., (1993) reported that partially-purified culture filtrates from C. coffeanum were used against crushed calli in liquid and with cell suspension of arabica materials (N 39 and Timor Hybrid). Growth and viability of susceptible cultures were significantly reduced by the presence of C. coffeanum filtrate and several somatic embryos and plantlets were recovered from the selected calli. The resulting plants will be tested in vivo to confirm (or not) the resistance reaction to CBD.
10.5 COFFEE SCALE-UP BY MICROPROPAGATION Seed is the most common form of plant propagation and it should be used when the species is autogamous (homozygous), or when reliable production of hybrid seeds can be made in large quantities. Vegetative propagation applies when the plant species does not produce seeds, or hybrid seeds cannot be commercially utilized. Rooting and grafting are the most common methods of vegetative propagation. Micropropagation is utilized when (a) it is difficult to apply traditional
Agronomy II: Developmental and Cell Biology
propagation methods; (b) it is important to start from `disease-free' planting materials; or (c) there is a need to produce very large numbers of plants in a short period of time. Micropropagation can be achieved by different in vitro multiplication methods: (a) growth of preexisting axillary buds; (b) production of shoots via organogenesis; or (c) plantlets production from somatic embryogenesis. The great majority of commercial micropropagation protocols are based on axillary bud multiplication in solid medium (Kurtz et al., 1991). Micropropagation via organogenesis has been of limited use (Litz & Gray, 1992), either because of the lack of specific protocols or because high rates of somatic variation are suspected. Somatic embryogenesis is currently receiving a great deal of attention, due to its enormous multiplication rates (reduction of unit costs) and the relative genetic stability of the resulting plants under greenhouse and field evaluations (Jones & Hughes, 1989; Sondahl et al., 1999; Ducos et al., 1999). Indeed, somatic embryogenesis is a highly attractive propagation method for perennial species and also for tropical plants that carry elevated levels of phenolic compounds, which inhibit rooting and grafting. Arabica coffee is a self-pollinated species and so 98% of homozygous individuals can be obtained after six consecutive generations (ca 24 years of selfing and selection). Robusta is an out-crossing species and so fields propagated by seeds are composed of segregant individuals. Yield and other desirable characteristics would improve drastically in robusta plantations established from propagated elite individuals. One can easily evaluate the commercial advantages of cloning elite individual plants so bypassing the long time required for selfing and selecting homozygous seed donors. Multiple clone lines would be used at one time and would thus help preserve heterozygosity and plasticity in coffee plantations. Besides the establishment of commercial fields, coffee micropropagation could set up `seed orchard areas' for the production of multi-line bulked seeds, or to multiply segregating parental lines for the synthesis of intervarietal hybrid seeds. Coffee can be propagated by grafting, axillary bud development (nodal cultures) or by the `direct' embryogenesis pathway. All these methods are suitable for a limited cloned population due to its low multiplication rate. In this section we will focus on the use of `high-frequency (or indirect)' somatic embryogenesis for large-scale propagation of elite plants to establish commercial plantations with a population size of 100 000 to 1 000 000 plants per year.
213
10.5.1 Mass production of somatic embryos Many protocols are available for production of coffee somatic embryos but all of them are experimental in nature and not suitable for large-scale propagation. Embryogenic tissues from somatic cells of elite plants should be produced on solid cultures and maintained viable through a periodic subculture regime. Mass production of somatic embryos must be set up in liquid embryogenic suspension cultures growing in Erlenmeyer or bioreactor vessels.
(a) Erlenmeyer cultures Zamarripa et al., (1991a) made a very detailed report on the conditions to establish and maintain embryogenic suspensions of coffee cells. Embryogenesis was induced in solid media from leaf explants cultivated in Dublin (1984) medium in the dark and later, transferred to light on Yasuda (1985) medium containing 1.0 mg/l BA. Yellow, friable, embryogenic tissues were transferred to 50 ml Erlenmeyer flasks containing 20 ml of a modified Yasuda medium (Zamarripa et al., 1991a). After initial growth, suspensions are transferred to 100 ml and later to 250 ml Erlenmeyer flasks, under 100 rpm at 238C. The density of the initial inoculum is important in this liquid suspension establishment and it should be at least equal to 10 g fw/ liter. The total time to establish stable embryogenic suspension is about 8 months. Liquid cultures are maintained through a 21-day subculture regime. At each subculture, the biomass is collected in a nylon filter (mesh 50 mm) and transferred to a fresh 250 ml Erlenmeyer flask containing 100 ml of liquid medium to keep a final density of 10 g fw per liter. These coffee suspension cultures consist of cellular aggregates of 430 mm diameter in Catuai, 630 mm in arabusta and 760 to 940 mm in robusta. Once the embryogenic suspensions were well established (Zamarripa et al., 1991a), somatic embryos were induced by transferring cellular aggregates to a fresh Dublin (1984) medium. In this phase, the dilution of the initial density to 0.1±0.2 g fw/liter is critical. The lower the density, the better the development of the somatic embryos. After 6 weeks, the embryo concentration reaches a plateau of 240 000 embryos/liter. After 6 weeks, about 90 000 torpedo shape embryos can be found. This behavior is similar to all coffee genotypes studied. However, the embryo formation is not synchronized since one can observe somatic embryos at all stages of differentiation (globular, heart and torpedo
214
shape). The authors noted that this embryo production may vary with the culture conditions like inoculum density, periodicity of fresh medium addition, rate of agitation, and light/dark conditions. Weekly addition of fresh medium is better than biweekly for torpedo embryo production. After 8 weeks of embryo differentiation in liquid medium, embryos were transferred to solid medium (Zamarripa et al., 1991a) containing 0.225 mg/l BA. Following another 4 weeks, embryos were subcultured to a cytokinin-free medium and the first pair of leaves were observed within 8 additional weeks. The conversion rate of embryos to plantlets was 50 to 70% after 12 weeks in BA-free medium. Plantlets with 2 to 3 pairs of leaves were transferred to a greenhouse for hardening and subsequent growth.
(b) Bioreactor cultures Robusta embryogenic suspension cells (clone R2) were charged at the rate of 1.0 mg fw/liter to a 3-liter liquid medium, inside a stirred Setric SGI bioreactor apparatus, running at 60 rpm, with aeration at 0.04 air volume/medium volume per minute at 268C (Zamarripa et al., 1991b). Under this system, embryogenic tissue proliferates up to a yield of 200 000 embryos per liter on day 49 of culture. From this total embryo population, about 20% (40 000/liter) were torpedoshaped embryos. The conversion rates of somatic embryos derived from the bioreactor were similar to the ones produced in Erlenmeyer flasks (50 to 70%) which translates into 60 000 plantlets/bioreactor every 2 months. The success rate during hardening in the greenhouse was described at 80 to 95% after 5 months. This means that a 3-liter bioreactor culture of robusta embryogenic cells can yield ca 48 000 cloned plants every 2 months which is equivalent to 19 ha of a coffee plantation at a planting density of 2500 plants/ha. The potential of this process for scaling up elite robusta plants can be realized if the bioreactor is charged with higher working volumes and multiple bioreactor units run simultaneously. An additional study on the critical parameters for bioreactor mass production of robusta (and arabusta) coffee was presented by Ducos et al., (1993). A Setric SGI model SET4CV stirrer bioreactor was charged with an embryogenic cell suspension at rate of 0.5 g fw/ liter, operating at a 3-liter working volume, renewed every week. Agitation was kept at the lowest level of 50 rpm until day 21 and then, increased slowly to 100 rpm. Air flow was maintained at the lowest level just to maintain dissolved oxygen (DO2 ) above the
Coffee: Recent Developments
critical level. Specific oxygen uptake and the specific production rates of CO2 and ethylene decreased as a function of culture time. The increases in CO2 and ethylene were linked with the increase in the aeration rate, and this finding is in contrast with Erlenmeyer flask cultures. Production of embryos began on day 21 and it was completed on day 58. Ducos et al., (1993) reported a maximum yield of 180 000 embryos/liter for robusta cultures. The embryo±plantlet conversion rate observed in robusta was 47%, similar to control Erlenmeyer flasks. For arabusta, a population of 160 000 embryos/liter was achieved, and a conversion rate of 37% was observed for bioreactor-derived embryos. This rate was higher than the Erlenmeyer flask rate (20%). Bioreactor mass production of arabica somatic embryos has been reported by Noriega and Sondahl (1993). Friable embryogenic tissue (FET) was obtained according to Sondahl's protocols (Sondahl & Sharp 1977; Sondahl et al., 1984), using mature leaf explants of C. arabica cv Red Catuai, cultured on conditioning medium (MSI) for 6 weeks and then transferred to the induction medium (MSII). Friable embryogenic tissue colonies were isolated after 4 to 6 months of secondary culture, and this tissue was maintained on solid medium by periodic subcultures for ca 3 years. In a preliminary study, FET cultures were inoculated at low density in a bioreactor vessel where a 20-fold proliferation increase was observed (Sondahl & Noriega, 1992). After this multiplication phase, torpedo-stage embryos were observed at low frequency and the bioreactor was kept running without medium exchange for 2 months. At the end of this period, the suspension differentiated entirely into embryos, which were then cultivated into a liquid `maturation medium' for 4 weeks. Mature embryos were plated onto solid germination medium producing normal plantlets. This experiment revealed a yield of 12 500 embryos per 1.0 g fw inoculum of FET cells. In a subsequent study of arabica mass production of embryos in bioreactors, Noriega & Sondahl (1993) used embryogenic cell suspension cultures of Red Catuai, initiated in MSII medium using 125 or 250 Erlenmeyer flasks in a rotary shaker at 100 rpm at 258C in the dark. Fresh MSII medium was added twice a week. The suspension cultures were established after a period of 3 months and after that, cultures were split every 3 weeks, maintaining the packed cell volume (PCV) of about 5 to 10 ml/100 ml. These suspensions contained only FET cells with cluster of 0.5 to 1.0 mm in diameter. These FET suspension cultures were used for
Agronomy II: Developmental and Cell Biology
bioreactor inoculation (Noriega & Sondahl, 1993). A 5-liter magnetic stirring bioreactor vessel was used, running at 70 to 120 rpm and kept at 258C in the dark. Clusters of FET were charged to liquid MSII, which was replaced every week for the next 5 weeks. Cell density was maintained at 1 to 5 PCV/liter inside the bioreactor by removing the excess of tissue at the time of medium exchange. At the end of the fifth week, no more fresh medium was added to force the culture to enter into a rest period (to stop the multiplication phase). After another 4 weeks, fresh `developing medium' (DM) was added. After an additional 5 weeks, somatic embryos were harvested from the bioreactor and plated on solid agar medium for germination. Summarizing this protocol, FET cells were allowed a multiplication phase of 5 weeks, a resting phase of 4 weeks (embryo differentiation) and an additional 5 weeks on developing medium. At the end of production (3.5 months), a population of 45 000 embryos/5 liter bioreactor was estimated. The total embryo population was reduced by the periodic FET cell removal during the multiplication phase. Samples of the somatic embryo population produced by this bioreactor method revealed about 25% of torpedo, 45% of heart and 30% of globular embryos. Debris of FET cells could still be seen inside the bioreactor, which is an indication that the bioreactor differentiation process was not fully completed at the time of opening. Plated embryos began chlorophyll development after 1 to 2 weeks under light and fully germinated embryos were observed at the tenth week on solid medium (Noriega & Sondahl, 1993).
(c) Periodic immersion cultures An autoclavable filtration unit (500 ml capacity) was modified to facilitate periodic immersion flushes of liquid medium and it has been tested for coffee propagation via axillary buds and somatic embryogenesis (Berthouly et al., 1995a). A glass tube is installed to connect the upper and lower compartments. A fine screen is placed in the bottom of the upper part to hold cultivating tissues. The system is charged with explant in the top part and fresh medium in the bottom section, which is connected to a small air pump, controlled by an electric clock. When the pump is turned on, the air passes through a 0.22 mm filter to maintain sterility and enters the lower section; as the pressure builds, the liquid medium is suspended to the upper section. When the pump turns off, the liquid returns to the lower section by gravity. Using coffee orthotropic nodal segments in the
215
upper section, the authors were able to induce six to seven axillary buds to develop within 5 to 6 weeks, in contrast with solid medium cultures that take about 12 weeks. It was reported that time and frequency of periodic immersion were critical factors for optimizing the final results. Again, the protocol must be adjusted according to the genotype being cultivated. For arabica nodal cultures, the best conditions were four pulses of 15 minutes per 24 hours in the presence of 1.0 mg/ 1 BA. Prolonged immersion times would lead to vitrification. For robusta nodal cultures, four pulses of 1 minute per 24 hour period should be used in the presence of 0.1 mg/l BA. Under these conditions, an average of 6.8 shoots and 7.2 shoots were recorded for arabica and robusta cultures, respectively. The same periodic immersion system (called RITA) was charged with FET and after 40 days it produced the same amount of fresh weight as control cultures in the Erlenmeyer flasks. Immersion time and frequency were also critical for coffee embryogenic cultures. The optimum conditions for somatic cell proliferation and embryo regeneration employed four pulses of 15 minutes per 24 hours.
10.5.2 Applications Case 1: Bourbon LC cloned field To evaluate the stability of coffee somatic embryos produced via solid and liquid media, an experiment field was established using C. arabica cv Bourbon LC line B. Young plants derived from in vitro cultures were shipped from the New Jersey laboratory to Brazil and seedlings of the same line were produced as control plants at the local nursery (Sondahl et al., 1999). Friable embryogenic tissues were produced from leaf explants of adult plants according to the two-stage method of Sondahl & Sharp (1977). Liquid suspension cultures of FET were established and then propagated into bioreactor vessels as described by Noriega & Sondahl (1993). Two bioreactor models were tested: a 5-liter, magnetic-stirring Ono model and a 7-liter, blade-stirring Aplikon model. Somatic embryos from bioreactor cultures and from solid agar cultures were allowed to germinate and the resulting plantlets were transferred to 35 6 145 mm tubettes and placed in a greenhouse for hardening. Stage 4 plantlets were introduced in Brazil with the assistance of the Quarantine Service (Cenargen), and later transferred to a local coffee farm. Seeds of the same Bourbon LC line B were germinated in sand beds and later transferred to 35 6 145 mm tubettes to complete their development.
216
Coffee: Recent Developments
Planting was carried out in randomized plots in a coffee farm at 2.0 6 1.0 m spacing, during the period of 2 to 4 March 1996. This experimental plot consisted of the following materials: 220 bioreactor-derived plants, 230 solid medium plants and 500 seed-derived plants. Morphological evaluations were made after the first harvest (1998) and second harvest (1999) which can be summarized as follows: Bioreactor plants = 01 Solid medium plants = 02 Seed-derived plants = 02 01
variegated (01/220 = 0.4%) broad leaf type (02/230 = 0.9%) murta type angustifolia (03/500 = 1.0%)
After 4 years under field conditions and at the second crop, very few differences among coffee plants (as shown above) could be seen. This field test plot demonstrates that coffee plants derived from somatic embryos, produced either by solid or liquid bioreactor cultures, can be used for micropropagation. For practical purposes, all micropropagated plants evaluated can be considered to be similar to each other. Since the plants show a very low rate of variation (less than 1%), it can be concluded that the process is safe for largescale multiplication for the coffee variety tested. The frequency of in vitro variability is highly dependent on the genotype, and so each elite plant or hybrid selected for micropropagation must be field tested before large plantation areas are established.
Case 2: Nestle fields of Robusta clones Robusta is a self-incompatible species, and so vegetative propagation must be performed in order to maintain the genetic potential of selected plants. Before entering into a full scale planting program, several aspects have to be evaluated: (a) regeneration capacity of embryogenic cell lines for each selected genotype; (b) logistics of production and distribution of cloned plants; (c) cost of production of plants produced by somatic embryogenesis; and (d) true-to-type status of regenerated plants (Ducos et al., 1999). Five elite robusta plants were selected on the basis of their agronomic traits for this micropropagation evaluation. Embryogenic cell lines were isolated from young leaf explants after 8 months on solid cultures. Liquid embryogenic suspensions were established after 2 months and subcultured every 2 weeks at a density of 10 g fw/liter. Somatic embryos were induced by decreasing the inoculum density from 10 to 1.0 g fw/l and by transfer to a full strength MS medium. After 2 to 3 months, embryos were plated onto germination solid medium. The authors observed that the
embryogenic suspension cell lines decrease their embryo-producing capacity after 6 months of continuous culture. So it has been recommended to maintain the embryogenic cell lines under solid medium and only transfer to liquid suspension according to the mass production schedule (3-monthold suspensions are ideal). Based on these assumptions, the following calculations were made by Ducos et al., (1999): (a) 500 explants produce 1.0 g FET; (b) 60 g FET are generated after 3 months of liquid multiplication; (c) 1.0 g fw of 3-month-old FET liquid culture produces 56 000 plantlets. In conclusion, about 3.0 million plantlets can be produced from an initial inoculation of 500 leaf segments, that is sufficient for planting ca 1800 ha of robusta fields. Nestle laboratory sent in vitro torpedo shaped somatic embryos to a collaborating facility in the Philippines (Department of Plant Industry) from which about 70 000 plants were recovered. The conversion of embryo to plantlets in the recipient country was four to five times lower than in France, and so it was recommended to ship ex vitro acclimatized plantlets to local nursery facilities. The average recovery of ex vitro plantlets is currently 37%. This group made initial cost calculations and derived production costs of US $0.169 per somatic-embryoderived plant versus a cost of US $0.158 for a robusta cutting. Based on 1600 plants/ha plantations, this difference is only US $18.4 per hectare, which is not significant considering the several benefits offered by mass production of cloning operations. These five robusta clones are being field tested in five coffee-producing countries (4000 plants/location): Philippines, Thailand, Mexico, Nigeria and Brazil. Based on visual inspection of 8000 plants under field conditions in the Philippines, all micropropagated robusta plants have normal vegetative aspects and are developing normal flowers and fruits 2 years after planting. The first test field in the Philippines produced 1 t green coffee at its second harvest. These field tests will continue to evaluate long-term growth and production capacity in addition to comparing in vitro-derived plants with plants originated by other vegetative methods such as cuttings and microcuttings.
Case 3: Cloned fields of F1 Arabica hybrids A cooperative program between coffee-producing countries of Central America (Promecafe), the CATIE research center in Costa Rica, and a French Cooperation Consortium (Cirad, Orstom, Mae) has embarked on an interesting program for synthesis of F1 hybrids
Agronomy II: Developmental and Cell Biology
between high-yielding arabica varieties (Caturra, Catuai, Catimor, Sharchimor) and wild arabica genotypes from Ethiopia and Sudan (Etienne et al., 1997). The goal is to combine superior cup quality with yield and disease resistance. This hybridization program was initiated in 1992 and now F1 hybrids are being evaluated under field conditions and final selections should be made by year 2003 (Etienne et al., 1999). A pilot scale micropropagation of 10 F1 hybrids is already under way, using the RITA periodic immersion technique described above (Berthouly et al., 1995a). Differences in the embryogenesis capacity among different F1 hybrids have been reported, but the amount of FET recovered from any F1 hybrid is sufficient for multiplication in liquid for large-scale clone production of any single hybrid (Etienne et al., 1997). The current data reveal a production of 7500 to 15 000 embryoderived plants per gram of embryogenic suspension culture. The authors have successfully tested `direct sowing', that is transfer of early stages of cotyledonary somatic embryos to trays containing artificial soil under controlled greenhouse conditions. A total of 20 000 plants from these F1 hybrids was produced and these plants are being used to establish test field plots in four Central American countries (Etienne et al., 1999). The objective is to evaluate the performance of the embryo-derived plants under distinct farming conditions. It is reported that among 4000 vitroplants under field and nursery conditions so far evaluated, no somaclonal variation has been observed (Etienne et al., 1999).
Case 4: Uganda Robusta cloning program The Uganda Ministry of Agriculture through the Farming System Support Programme sponsored by the European Union has launched a project for largescale propagation and distribution to farmers of six selected robusta clones. Propagation will be accomplished by the cutting process and by in vitro methods (Berthouly et al., 1995a). A local cloning facility of 182 m2 was constructed in Uganda, equipped with two temperature controlled culture rooms and a nearby greenhouse. The micropropagation effort will utilize the periodic immersion technique (Berthouly et al., 1995a) for cloning via axillary bud development (microcuttings) and via somatic embryogenesis. A total of 2000 RITA vessels are being installed and the expectation is to produce 600 000 plants/year from microcuttings and 2.0 to 2.5 million plants/year via somatic embryogenesis (Berthouly et al., 1995b).
217
10.6 SOMACLONAL VARIATION AND NEW BREEDING LINES 10.6.1 Definitions and examples Variation among plants regenerated from in vitro cultures was first described for plants from tobacco callus cultures by Butenko et al., (1967). However, this in vitro variability was not clearly recognized and defined until the review made by Larkin and Scowcroft (1981). Somaclonal variants can appear when an explant (any plant part) is subjected to a tissue culture cycle. This cycle includes establishment of a dedifferentiated cell or tissue culture under defined conditions and the subsequent regeneration of plants (Hammerschlag, 1992). This phenomenon was further defined to include in vitro variability from cultivated haploid cells and named `gametoclonal variation' (Evans et al., 1984). Somaclonal variation is the expression of the naturally occurring variability of plant cells, or the result of in vitro induced variability of cells following plant regeneration (Larkin & Scowcroft, 1981; Evans & Sharp, 1986). Most of this spontaneous variability from in vitro plants is associated with chromosome alterations such as breakage, translocation, deletions, aneuploidy, polyploidy and somatic crossing-over. In addition, somaclonal variation can also have a single gene origin, for example a point mutation, alteration in gene copy number, activation of transposon elements and variation in DNA methylation (Karp et al., 1982; McCoy et al., 1982; Orton, 1983; Phillips et al., 1990). Somaclonal variation is an excellent method for shortening breeding programs, since it can provide access to genetic variability within existing cultivars (Evans & Sharp, 1986). Somaclones carry few genetic alterations and so the overall genetic integrity of the original commercial cultivar is preserved. In the case of coffee, no variability has been observed beyond the diversity already known for the arabica species (Sondahl & Bragin, 1991). Somaclonal variation has contributed to the release of improved varieties of sexually (tobacco, tomato, rapeseed, corn, blackberry, celery, coffee) and non-sexually propagated species (potato, sweet potato, sugarcane) (Evans, 1988; Hammerschlag, 1992; Sondahl & Lauritis, 1992).
10.6.2 Coffee somaclonal variation program Several agronomically important coffee genotypes were used for a somaclone variation study (Sondahl & Bragin, 1991; Sondahl & Lauritis, 1992). Since
218
somaclonal variation frequency varies with genotype and culture procedure, a wide array of genotypes were used in this program, including tall stature varieties (Yellow Bourbon, Mundo Novo, Icatu and Aramosa) and short stature varieties (Red and Yellow Catuai, Caturra, Catimor, Laurina and other genotypes). Tissue culture was initiated from mature leaf explants following the Sondahl and Sharp (1977) protocol and donor plants were maintained in a greenhouse collection. Plantlets were recovered from both the `lowfrequency pathway (LFSE)' and the `high-frequency pathway (HFSE)'. Plantlets were hardened in a greenhouse and then shipped to a coffee nursery in Brazil, with the assistance of local quarantine service. In vitro-derived plants were transferred to the field as they reached transplanting size, and different experimental fields were made according to the donor variety. Seed-derived control plants for each coffee variety being studied were planted in each experimental field. Planting was in a coffee farm in Cajuru SP, Brazil at 218 latitude south, at 1040 m altitude, with a spacing of 3.5 6 2.0 m. Normal coffee fertilization and disease control practices were used in this field. A total of 14 948 in vitro-derived plants were established in the field, representing nine different coffee genotypes. Screening was done at the Ro generation during the first and second crops. The most interesting variant forms were studied in the next generation by establishing progeny fields. The overall variability found in this in vitro-derived coffee population was 10%, but variability was highly genotype-dependent, for instance: 30.6% for Yellow Bourbon and only 3.3% for Red Catuai. The most common mutation was for fruit color (42.35% yellow to red) followed by change in plant stature (3.8% tall to short). Based on 7772 in vitro plants evaluated, the frequency of variability was similar for plants originated from HFSE (or indirect embryogenesis; 12%) and LFSE (or direct embryogenesis; 10.4%). These data demonstrate that both HFSE and LFSE could be used for micropropagation programs, since there is no enhanced variability among plants from the HFSE pathway, as initially suspected. More detailed information about this program can be found in Sondahl and Bragin (1991) and Sondahl and Lauritis (1992). Many interesting variants have been selected from this program and their progenies are being studied in subsequent generations. The most interesting mutations are being carried on by standard breeding methods aiming for the release of new varieties in the future. Emphasis is being made in selecting superior cup qualities associated with desirable agronomic traits.
Coffee: Recent Developments
Three main breeding populations have been derived from this tissue culture program: Laurina somaclones, Icatu somaclones and Aramosa somaclones. Some characteristics of segregating individuals of the Icatu and Aramosa populations were reported by Sondahl et al., (1997). Other small populations are also being studied, such as short-stature mutants of Mundo Novo and Yellow Bourbon. Another interesting population is based in one Maragogype mutant plant that was derived from Yellow Catuai leaf cultures. Second and third generations of this somaclone are showing segregation for the typical Maragogype phenotype, normal Catuai and an intermediate phenotype with short stature but very large beans.
10.6.3 Commercialization of new varieties Laurina is a natural mutation from Red Bourbon plants found in Reunion Island by the mid-1800s. These mutated plants had small leaves, thin lateral branches, short stature and elongated fruits and beans, and it has been referred to in the literature as `Laurina', `Leroy', `Bourbon Pointu' and `Smyrna' coffee (Raoul, 1897; Boutilly, 1900; Coste, 1955). This Laurina sport was immediately introduced into commercial plantations in Africa (and transferred to South America) because of its drought resistance and superior beverage properties (Raoul, 1897; Krug et al., 1954). It was not until much later (the mid-1950s) that studies of coffee collections reported that Laurina plants had a natural 50% reduced caffeine concentration (Lopes, 1973). More recently, Baumann et al., (1998) explained that the reduction in caffeine content in Laurina is due to a reduced synthetic activity. Among more than 800 in vitro-derived plants of Laurina, 15 elite plants were selected at the R0 generation in June 1991. These selected plants were clearly more vigorous than sister plants and donor controls, as demonstrated by greater leaf area, lateral branches, plant height, plant diameter, and superior yield. Seeds of these selected plants were taken to establish a separate experimental field (4 ha in size) to evaluate the performance of each somaclone line. A total of 360 plants per line was established in a coffee farm in Brazil, in March 1992, in nine random replicated blocks plus control plants (seed-derived donor Laurina) and border lines, at a spacing of 3.5 6 1.5 m, with two plants per hill. Growth pattern, yield and caffeine content were monitored for each of these 15 lines during the first 5 years under field conditions (three successive crops). It was observed
Agronomy II: Developmental and Cell Biology
that the caffeine content was stable and equal to the donor plants, the growth pattern was stable for all lines (no segregation) and the yield from the top five lines was twice as high as for the control plants. Yield evaluation continued up to the sixth harvest (1999 crop) in the experimental field, thus confirming the initial selection of the top five lines for superior yield and reduced biannual cycle. A third generation of selected lines was established in a semi-commercial plot design of 25 ha in size. Seeds from the top five highyielding lines are being scaled under the name of `Bourbon'. At the time the first round of selection of elite Laurina somaclones was completed, filings for patent protection were made on this discovery. A utility patent was awarded in the USA under Patent No 5 436 395 on 25 July 1995 (Sondahl et al., 1995). The Bourbon LC is the first example of a patent awarded for a coffee variety and it is also the first case of the release of a coffee variety derived from natural variability, isolated from somatic embryo cultures. Bourbon LC is the only naturally reduced (50%) caffeine variety being produced in commercial quantities. It is a product that should capture the interest of `coffee lovers' since it enables the consumer to drink twice as many cups per day before reaching his/her body caffeine threshold level.
10.7 SUMMARY The ontogeny of leaf and fruit formation in arabica coffee plants and the importance of the purine alkaloid tissue fluctuation as a defense mechanism for survival in tropical and sub-tropical environments have been discussed. The most recent advances in cell and organ culture have been reviewed. Enormous progress has been made in the induction and regeneration of somatic embryos from both arabica and robusta species, as well as in the maintenance of embryogenic-competent cell lines. These achievements are key for future progress in protoplast work and gene transfer programs. Greater understanding of the control of the somatic embryogenesis process, coupled with the development of bioreactor and periodic immersion culture techniques, has led to the development of reliable methods for mass coffee propagation. Vegetative propagation of a perennial species like coffee can bring flexibility for introducing new genotypes into production. The ability to shift planting material rapidly will enable farmers to enjoy the
219
agronomic advantages of elite plants, but it will also bring the opportunity for the coffee industry to work more closely with the green coffee production sector to address industry and marketing needs. Green coffee quality is controlled by genetics, environment, and farm processing and mass propagation offers the opportunity to control the first of these variables. Vegetative propagation of heterozygous elite plants and the use of multiple clone lines at one time will assure the preservation of heterozygosity and plasticity to environmental changes within coffee plantations. Micropropagation via somatic embryogenesis is the technology that has the scale to satisfy commercial plantations and to be competitive in cost with other propagation methods. Current field evaluation of somatic embryo-derived coffee plants, for both arabica and robusta, provides confidence that clonal fidelity is very high, with minimum or no somatic variation on the genotypes tested so far. First practical applications of natural variability at the cellular level are being released for production. New and improved varieties will be developed through cell biology techniques to address production constraints at the farm level, and at the same time to permit the adoption of more sustainable coffee production. These techniques are based on spontaneous somatic mutations. Plant regeneration techniques are just the tool to uncap such variability at the cellular level, and so the genetic make-up of the new mutant type selected is very similar to the original one. This similarity with the original genetic make-up of donor plants helps to hasten the release of new varieties, since the other characteristics are kept constant. Progress is still needed in the area of anther or microspore culture leading to the recovery of double homozygous plants. This technique has already been mastered with other plant species, and thus it could assist the introgression of genes from wild coffee species to cultivated species, speeding up coffee breeding programs. Another area that deserves more attention is germplasm preservation. Man is altering the natural habitats where wild coffee species have naturally evolved, and in consequence much valuable germplasm is in danger of destruction. Live coffee germplasm collections must continue to be supported, but efforts should be made to improve the techniques of long-term coffee seed and embryo preservation.
220
Coffee: Recent Developments
ABREVIATIONS BA or 6BA 2,4-D 2-iP KIN NAA PAL
6-Benzylaminopurine 2,4-Dichlorophenoxyacetic acid 2-Isopentenyladenine Kinetin 1-Naphthaleneacetic acid Phenylalanine ammonia lyase
REFERENCES Acuna, J.R. & Pena, M. (1991) Plant regeneration from protoplasts of embryogenic cell suspensions of Coffea arabica L. cv. Caturra. Plant Cell Rep., 10, 345±8. Acuna, M.E.A. (1993) Somatic embryogenesis induced by culture on single media in coffee plants from crosses of Coffea arabica by Timor Hybrid. In: Proceedings of the 15th ASIC Colloquium (Montpellier). pp. 82±88. ASIC, Paris, France. Aerts, R.J. & Baumann, T.W. (1994) Distribution and utilization of chlorogenic acid in Coffea seedlings. J. Exp. Botany, 45, 497±503. Ashihara, H. & Crozier, A. (1999) Biosynthesis and metabolism of caffeine and related purine alkaloids in plants. Adv. Bot. Res., 30, 117±205. Ashihara, H., Monteiro, A.M., Moritz, T., Gillies, M.F. & Crozier, A. (1996) Catabolism of caffeine and related purine alkaloids in leaves of Coffea arabica L. Planta, 198, 334±9. Barboza, C.A. & Ramirez-Martinez, J.R. (1991) Antocianinas en pulpa de cafe del cultivar Bourbon rojo. In: Proceedings of the 14th ASIC Colloquium (San Francisco), pp. 272±6. ASIC, Paris, France. Barton, C.R., Adams, T.L. & Zarowitz, M.A. (1991) Stable transformation of foreign DNA into Coffea arabica plants. In: Proceedings of the 14th ASIC Colloquium (San Francisco), pp. 460±64. ASIC, Paris, France. Baumann, T.W. & Frischknecht, P.M. (1982) Biosynthesis and biodegradation of purine alkaloids in tissue cultures. Proceedings of fifth International Congress on Plant Tissue and Cell Culture (ed. A. Fujiwara), pp. 365±366. Tokyo, Japan. Assoc. Plant Tissue Culture. Baumann, T.W. & Frischknecht, P.M. (1988a) Caffeine: production by plant (Coffea spp.) cell cultures. In: Biotechnology in Agriculture and Forestry, Vol. 4 Medicinal and Aromatic Plants I, (ed. Y.P.S. Bajaj) pp. 264±281. Springer-Verlag, Berlin, Heidelberg. Baumann, T.W. & Frischknecht, P.M. (1988b) Purines. In: Cell Culture & Somatic Cell Genetics of Plants, Vol. 5 Phytochemicals in Plant Cell Cultures (eds F. Constabel & I. K. Vasil), pp. 403± 17. Academic Press, San Diego. Baumann, T.W., & RoÈhrig, L. (1989) Formation and intracellular accumulation of caffeine and chlorogenic acid in suspension cultures of Coffea arabica. Phytochemistry 28, 2667±9. Baumann, T.W., Koetz, R. & Morath, P. (1983) N-methyltransferase activities in suspension cultures of Coffea arabica L. Plant Cell Rep. 2, 33±5.
Baumann, T.W., Sondahl, M.R., MoÈsli Waldhauser, I.S. & Kretschmar, J.A. (1998) Non-destructive analysis of natural variability in bean caffeine content of Laurina coffee. Phytochemistry, 49, 1569±73. Baumann, T.W. & Wanner, H. (1972) Untersuchungen uÈber den Transport von Kaffein in der Kaffeepflanze Coffea arabica. Planta, 108, 11±20. Berthouly, M., Alvard, D., Carasco, C. & Duris, D. (1995b). A technology transfer operation: a commercial Coffea canephora micropropagation laboratory in Uganda. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 743±44. ASIC, Paris, France. Berthouly, M., Dufour, M., Alvard, D., Carasco, C., Alemanno, L. & Teisson, C. (1995a) Coffee micropropagation in a liquid medium using the temporary immersion technique. In: Proceedings of the 16th ASIC Colloquium (Kyoto) pp. 514±19. ASIC, Paris, France. Bridson, D.M. & Verdcourt, B. (1988) Coffee. In: Flora of Tropical East Africa ± Rubiaceae (Part 2) Balkema (ed.) R.M. Polhill. Rotterdam. Boutilly, V. (1900) CafeÂier de Liberia. Sa Culture et sa Manipulation. Bibliotheque de la Revue des Cultures Coloniales. A, pp. 1±3. Challamel, Paris. Buckland, E. & Townsley, P.M. (1975) Coffee cell suspension cultures. Caffeine and chlorogenic acid content. J. Inst. Can. Sci. Technol. Aliment, 8, 164±5. Butenko, R.G., Shemina, Z.B. & Frolova, L.Y. (1967) Induced organogenesis and characteristics of plants produced in tobacco tissue culture. Genetika, 3, 29±39. Carvalho, A.C., Ferwerda, F.P., Frahm-Leliveld, J.A., Medina, D.M., Mendes, A.I.T. & Monaco, L.C. (1969) Coffee. Reprinted from Outlines of Perennial Crop Breeding in the Tropics (eds F.P. Ferwerda & F. Wit) pp. 189±241. H. Veenmann & Zonen N.V., Wageningen, The Netherlands. Charrier, A. & Berthaud, J. (1985) Coffee. In: Botany, Biochemistry and Production of Beans and Beverage (eds M.N. Clifford & K.C. Wilson), pp. 13±47. American Edition, Westport, Connecticut. Cho, G.H., Kim, D.I., Pedersen, H. & Chin, C.K. (1988) Ethephon enhancement of secondary metabolite synthesis in plantcell cultures. Biotech. Progress 4, 184±8. Coste, R. (1955) Les cafeÂiers et le Cafe dans le Monde, Vol. 1, Chapter II, pp. 15±31. Editions Larose, Paris. Dublin, P. (1981) Embryogenese somatique directe sur fragments de feuilles de cafeier Arabusta. CafeÂ, Cacao, TheÂ, 25, 237±41. Paris. Dublin, P. (1984) Techniques de reproduction vegetative in vitro et ameÂlioration genetique chez les cafeÂiers cultives. CafeÂ, Cacao TheÂ, 28, 231±43. IFCC, Paris. Dublin, P. (1991) Les techniques modernes de reproduction asexueÂe. Impacts sur 1'ameÂlioration genetique des cafeÂiers. In: Proceedings of the 14th ASIC Colloquium (San Francisco), pp. 365±77. ASIC, Paris, France. Ducos, J.P., Gianforcaro, M., Florin, B., Petiard, V. & Deshayeves, A. (1999) Canephora (Robusta) clones: in vitro somatic embryogenesis. In: Proceedings of the 18th ASIC Colloquium (Helsinki), 295±301. ASIC, Paris, France.
Agronomy II: Developmental and Cell Biology
Ducos, J.P., Zamarripa, A., Eskes, A.B. & Petiard, V. (1993) Production of somatic embryos of coffee in a bioreactor. In: Proceedings of the 15th ASIC Colloquium (Montpellier), pp. 89± 96. ASIC, Paris, France. Dussert, S., Chabrillange, N., Engelmann, F., Anthony, F., Hamon, S. & Lashermes, P. (1997) Cryopreservation of coffee (Coffea arabica) seeds. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 466±73. ASIC, Paris, France. Etienne, H., Barry-Etienne, D., Vasquez, N. & Berthouly, M. (1999) Aportes de la biotecnologia al mejoramiento genetico del cafeÂ: el ejemplo de la multiplicacion por embrigenesis somatica de hibridos F1 en America Central. In: (eds B. Bertrand & B. Rapidel), Desafios de Caficultura en Centroamerica, pp. 457±93. IICA, San Jose. Etienne, H., Bertrand, B., Anthony, F., Cote, F. & Berthouly, M. (1997) L'embryogenese somatique: un outil pour l'ameÂlioration genetique du cafeÂier. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 457±65. ASIC, Paris, France. Evans, D.A. (1988) Applications of somaclonal variation. In: Biotechnology in Agriculture. pp. 203±223. Alan R. Liss, Inc. Evans, D.A. & Sharp, W.R. (1986) Somaclonal variation in agriculture. Biotechnology, 4, 428±532. Evans, D.A., Sharp, W.R. & Medina-Filho, H.P. (1984) Somaclonal and gametoclonal variation. Am. J. Botany, 71, 759±74. Florin, B., Ducos, J.P., Firmin, L. et al. (1995) Preservation of coffee somatic embryos through desiccation and cryopreservation. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 542±7. ASIC, Paris, France. Florin, B., Tessereau, H. & Petiard, V. (1993) Conservation aÁ long terme des ressources genetiques de cafeeier par cryoconservation d'embryons zygotiques et somatiques et de cultures embryogenes. In: Proceedings of the 15th ASIC Colloquium (Montpellier), pp. 106±14. ASIC, Paris, France. Frischknecht, P.M. & Baumann, T.W. (1980) The pattern of purine alkaloid formation in suspension cultures of Coffea arabica. Planta medica, 40, 245±9. Frischknecht, P.M. & Baumann; T.W. (1985) Stress induced formation of purine alkaloids in plant tissue culture of Coffea arabica. Phytochemistry, 24, 2255±7. Frischknecht, P.M., Baumann, T.W. & Wanner, H. (1977) Tissue culture of Coffea arabica. Growth and caffeine formation. Planta med., 31, 344±50. Frischknecht, P.M., Eller, B.M. & Baumann, T.W. (1982) Purine alkaloid formation and CO2 gas exchange in dependence of development and of environmental factors in leaves of Coffea arabica. Planta 156, 295±301. Frischknecht, P.M., Ulmer-Dufek, J. & Baumann, T.W. (1986) Purine alkaloid formation in buds and developing leaflets of Coffea arabica: expression of an optimal defence strategy? Phytochemistry, 25, 613±16. Gillies, F., Jenkins, G.I., Ashihara, H. & Crozier, A. (1995) In vitro biosynthesis of caffeine: the stability of N-methyltransferase activity in cell-free preparations from liquid endosperm of Coffea arabica. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 599±605. ASIC, Paris, France. Griffiths, E., Gibbs, I.N. & Waller, J.M. (1971) Control of coffee berry disease. Ann. Appl. Biol. 67, 45±54.
221
Grossniklaus, U., Vielle-Calzada, J.P., Hoeppner, M.A. & Gagliano, W.B. (1998) Maternal control of embryogenesis by medea, a polycomb group gene in Arabidopsis. Science, 280, 446±50. Guerreiro Filho, O. (1992) Coffea racemosa Lour. Une revue. CafeÂ, Cacao, TheÂ, 36, 171±86. Paris. Gurney, K.A., Evans, L.V. & Robinson, D.S. (1992) Purine alkaloid production and accumulation in cocoa callus and suspension cultures. J. Exp. Bot., 43, 769±75. Haldimann, D. & Brodelius, P. (1987) Redirecting cellular metabolism by immobilization of cultured plant cells ± a model study with Coffea arabica. Phytochemistry, 26, 1431±4. Hartman, C.L., McCoy, T.J. & Knous, T.R. (1984) Selection of alfalfa (Medicago Sativa) cell lines and regeneration of plants resistant to the toxin(s) produced by Fusarium oxysporum f. sp. midicaginis. Plant Sci. Lett., 34, 183±94. Hammerschlag, F.A. (1990) Resistance response of plants regenerated from peach callus to Xanthomonas campestri pv. Pruni. J. Am. Soc. Hort. Sci. 115, 1034±7. Hammerschlag, F.A. (1992) Somaclonal variation. In: Biotechnology of Perennial Fruit Crops (eds F.A. Hammerschlag & R.E. Litz), pp. 35±55. CAB Intl., UK. Hatanaka, T., Azuma, T., Uchida, N. & Yasuda, T. (1995) Effect of plant hormones on somatic embryogenesis of Coffea canephora. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 790±97. ASIC, Paris, France. Jones, L.H. & Hughes, W.H. (1989) Oil Palm (Elaeis guineensis Jacq.). In: Biotechnology in Agriculture and Forestry, Vol. 5, Trees II (ed. Y.P.S. Bajaj), pp. 176±202. Springer-Verlag, Heidelberg. Kalberer, P. (1964) Untersuchungen zum Abbau des Kaffeins in den BlaÈttern von Coffea arabica. Thesis, University of Zurich. Kalberer, P. (1965) Breakdown of caffeine in the leaves of Coffea arabica L. Nature, 205, 597±8. Karp, A., Nelson, R.S., Thomas, E. & Bright, S.W.J. (1982) Chromosome variation in protoplast-derived potato plants. Theoret. Appl. Genet., 63, 265±72. Keller, H.R., Wanner, H. & Baumann, T.W. (1971) Koffeinsynthese und Gewebekultur bei Coffea arabica. Verhandl. Schweiz. Naturforsch. Ges.: 88±90. Keller, H., Wanner, H. & Baumann, T.W. (1972) Kaffeinsynthese in FruÈchten und Gewebekulturen von Coffea arabica. Planta, 108, 339±50. Krug, C.A., Carvalho, A. & Antunes Filho, H. (1954) Genetica de Coffea. XXI ± Hereditariedade dos caracteristicos de Coffea arabica var. Laurina (Smeathman) DC. Bol. Tech. Div. Exper. Pesq., Inst. Agronomico, Campinas, Brazil, 13, 247±55. Kurtz, S.L., Hartman, R.D. & Chu, I.Y.E. (1991). Current methods of commercial micropropagation. In: Scale-up and Automation in Plant Propagation (ed. I.K. Vasil), pp. 3±34. Academic Press, New York. Larkin, P.J. & Scowcroft, W.R. (1981) Somaclonal variation ± a novel source of variability from cell cultures for plant improvement. Theoret Appl. Genet. 60, 197±214. Litz, R.E. & Gray, D.J. (1992) Organogenesis and somatic embryogenesis. In: Biotechnology of Perennial Fruit Crops (eds F.A. Hammerschlag & R.E. Litz), pp. 3±34. CAB Intl., UK.
222
Lopes, M.H. (1973) Teor em cafeina de cafes espontaneos de MocËambique. In: Proceedings of the 5th ASIC Colloquium (Lisbon), pp. 63±9. ASIC, Paris, France. Loyola-Vargas, V.M., Fuentes-Cerda, C.F.J., MonforteGonzales, M., Mendez-Zeel, M., Rojas-Herrera, R. & Mijangos-Cortes, J. (1999) Coffee tissue culture as a new model for the study of somaclonal variation. In: Proceedings of the 18th ASIC Colloquium (Helsinki), pp. 302±307. ASIC, Paris, France. McCoy, T.J., Phillips, R.L. & Rhines, H.W. (1982). Cytogenic analysis of plants regenerated from oat (Avena sativa) tissue culture: high frequency of partial chromosome loss. Can. J. Genet. Cytol., 24, 37±50. Mazzafera, P., Crozier, A. & Sandberg, G. (1994) Studies on the metabolic control of caffeine turnover in developing endosperms and leaves of Coffea arabica and C. dewevrei. J. Agric. Food Chem., 42, 1423±7. Mazzafera, P., Wingsle, G., Olsson, O. & Sandberg, G. (1994) Sadenosyl-L-methionine:theobromine 1-N-methyltransferase, an enzyme catalyzing the synthesis of caffeine in coffee. Phytochemistry, 37, 1577±84. MoÈsli Waldhauser, S.S. & Baumann, T.W. (1996) Compartmentation of caffeine and related purine alkaloids depends exclusively on the physical chemistry of their vacuolar complex formation with chlorogenic acids. Phytochemistry, 42, 985±96. MoÈsli Waldhauser, S.S., Kretschmar, J.A. & Baumann, T.W. (1997) N-methyltransferase activity in caffeine biosynthesis: biochemical characterization and time course during leaf development of Coffea arabica. Phytochemistry, 44, 853±9. Murashighe, T. & Skoog, F. (1962) A revised medium for rapid growth and bioassays with tobacco tissue cultures. Physiol. Plant., 15, 473±97. Nakamura, T., Taniguchi, T. & Maeda, E. (1992) Studies on somatic embryogenesis of coffee by scanning electron microscope. Jap. J. Crop Sci., 61, 476±86. Neuenschwander, B. & Baumann, T.W. (1991) Purine alkaloid formation during somatic embryo development of Coffea arabica. In: Proceedings of the 14th ASIC Colloquium (San Francisco), pp. 595±600. ASIC, Paris, France. Neuenschwander, B. & Baumann, T.W. (1992) A novel type of somatic embryogenesis in Coffea arabica. Plant Cell Rep., 10, 608±612. Nishibata, T., Azuma, T., Uchida, N., Yasuda, T. and Yamaguchi, T. (1995) Amino acids on somatic embryogenesis in Coffea arabica. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 839±44. ASIC, Paris, France. Noriega, C. & Sondahl, M.R. (1993) Arabica coffee micropropagation through somatic embryogenesis via bioreactors. In: Proceedings of the 15th ASIC Colloquium (Montpellier), pp. 73±81. ASIC, Paris, France. Nyange, N.E., McNicol, R.J., Williamson, B. & Lyon, G.D. (1993) In vitro selection of Coffea arabica callus and cell suspensions for resistance to phytotoxic culture filtrates from Colletotrichum coffeanum. In: Proceedings of the 15th ASIC Colloquium (Montpellier), pp. 123±31. ASIC, Paris, France. Ogutuga, D.B.A. & Northcote, D.H. (1970) Caffeine formation in tea callus tissue. J. Exp. Bot., 21, 258±73.
Coffee: Recent Developments
Orozco, F.J. & Schieder, D. (1982) Aislamento y cultivo de protoplastos a partir de hojas de cafeÂ. Cenicafe, 33, 129±36. Orton, T.J. (1983) Experimental approaches to the study of somaclonal variation. Plant Mol. Rep., 1, 67±76. Phillips, R.L., Kaeppler, S.M. & Peschke, V.M. (1990) Do we understand somaclonal variation? In: Progress in Plant Cellular and Molecular Biology (eds H.J.J. Nijkamp et al.), pp. 131±41. Kluwer Academic Publishers, Dordrecht, The Netherlands. Prenosil, J.E., Hegglin, M., Baumann, T.W. et al. (1987) Purine alkaloid producing cell cultures: fundamental aspects and possible applications in biotechnology. Enzyme Microb. Technol., 9, 450±58. Raoul, E. (1897) Culture du cafeÂier. In: Manuel des Cultures Tropicales, Vol I, Part 1, (eds E. Raoul & P. Sagot), pp. 86±95; 167±239. A. Challamel, Paris. Rogers, W.J., Michaux, S., Bastin, M. & Bucheli, P. (1999) Changes to the content of sugars, sugar alcohols, myo-inositol, carboxylic acids and inorganic anions in developing grains from different varieties of Robusta (Coffea canephora) and Arabica (C. arabica) coffees. Plant Sci., 149, 115±23. Schoepke, C. Muller, L.E. & Kohlenbach, H.W. (1987) Somatic embryogenesis and regeneration of plantlets in protoplast cultures from somatic embryos of coffee (Coffea canephora P. ex Fr.). Plant Cell, Tiss. Org. Cult., 8, 243±8. Schulthess, B.H. & Baumann, T.W. (1995) Stimulation of caffeine biosynthesis in suspension-cultured coffee cells and the in situ existence of 7-methylxanthosine. Phytochemistry, 38, 1381±6. Schulthess, B.H., Morath, P. & Baumann, T.W. (1996) Caffeine biosynthesis starts with the metabolically-channelled formation of 7-methyl-XMP ± a new hypothesis. Phytochemistry, 41, 169±75. Selmar, D., Lieberei, R. & Biehl, B. (1988). Mobilization and utilization of cyanogenic glycosides. The linustatin pathway. Plant. Physiol., 86, 711±16. Shervington, A., Shervington, L.A., Afifi, F. & El-Omari, M.A. (1998) Caffeine and theobromine formation by tissue cultures of Camellia sinensis. Phytochemistry, 47, 1535±6. Sondahl, M.R. & Bragin, A. (1991) Somaclonal variation as a breeding tool for coffee improvement. In: Proceedings of the 14th ASIC Colloquium (San Francisco), 701±10. ASIC, Paris, France. Sondahl, M.R., Chapman, M.S. & Sharp, W.R. (1980) Protoplast liberation, cell wall construction, and callus proliferation in Coffea arabica L. callus tissues. Turrialba, 30, 161±5. Sondahl, M.R. & Lauritis, J.A. (1992) Coffee. In: Biotechnology of Perennial Fruit Crops, (eds F.A. Hammerschlag & R.E. Litz), pp. 401±20. CAB International, UK. Sondahl, M.R. & Loh, H.T. (1988) Coffee biotechnology. In: Coffee Vol. 4 Agronomy (eds R.J. Clarke & R. Macrae), pp. 235± 62. Elsevier Applied Science, London. Sondahl, M.R., Nakamura, T., Medina-Filho, H.P., Carvalho, A., Fazuoli, L.C. & Costa, W.M. (1984) Coffee. In: Handbook of Plant Cell Culture, Crop Species, Vol. 3 (eds P.V. Ammirato et al.), pp. 564±90. MacMillan, New York. Sondahl, M.R. & Noriega, C. (1992) Coffee somatic embryogenesis in liquid cultures. In: Proc. Intl Symp. in-vitro Culture
Agronomy II: Developmental and Cell Biology
Hort. Breeding, p. 75. International Horticulture Society, Baltimore, MD. Sondahl, M.R., Petracco, M. & Zambolim, L. (1997) Breeding for qualitative traits in Arabica coffee. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 447±56. ASIC, Paris, France. Sondahl, M.R., Romig, W.R. & Bragin, A. (1995) Induction and selection of somaclonal variation in coffee. US Patent Office, No. 5 436 395. Sondahl, M.R., Salisbury, J.L. & Sharp, W.R. (1979) SEM characterization of embryogenic tissue and globular embryos during high frequency somatic embryogenesis in coffee callus cells. Z. Pflanzenphysiol., 94, 185±8. Sondahl, M.R. & Sharp, W.R. (1977) High frequency induction of somatic embryos in cultured leaf explants of Coffea arabica L. Z. Pflanzenphysiol., 81, 395±408. Sondahl, M.R. & Sondahl, C.N. & Goncalves, W. (1999) Custo comparativo de diferentes tecnicas de clonagem. In: III SIBAC Symposium Londrina, Brazil (in press). Spiral, J. & Petiard, V. (1991) Protoplast culture and regeneration in coffee species. In: Proceedings of the 14th ASIC Colloquium (San Francisco), pp. 383±91. ASIC, Paris, France. Spiral, J. & Petiard, V. (1993) DeÂveloppement d'une meÂthode de transformation appliqueÂe aÁ differentes espeÂces de cafeÂier et reÂgeÂneÂration de plantules transgeÂniques. In: Proceedings of the 15th ASIC Colloquium (Montpellier), pp. 115±22. ASIC, Paris, France. Staritsky, G. (1970) Embryoid formation in callus cultures of coffee. Acta. Bort. Neerl., 19, 509±514. Sugiyama, M., Matsuoka, C. & Takagi, T. (1995) Transformation of coffee with Agrobacterium rhizogenes. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 853±9. ASIC, Paris, France. Tahara, M. Nakanishi, T., Yasuda, T. & Yamaguchi, T. (1995). Histological and biological aspects in somatic embryogenesis of Coffea arabica. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 860±67. ASIC, Paris, France. Townsley, P.M. (1974) Production of coffee from plant cell suspension cultures. J. Inst. Can. Sci. Technol. Aliment., 7, 79±81.
223
Urbaneja, G., Ferrer, J., Paez, G., Arenas, L. & Colina, G. (1996) Acid hydrolysis and carbohydrates characterization of coffee pulp. Renewable Energy, 9, 1041±4. Van de Voort, F. & Townsley, P.M. (1974) A gas chromatographic comparison of the fatty acids of the green coffee bean, Coffea arabica, and the submerged coffee cell culture. J. Inst. Can. Sci. Technol. Aliment., 7, 82±5. Van de Voort, F. & Townsley, P.M. (1975) A comparison of the unsaponifiable lipids isolated from coffee cell cultures and from green coffee beans. J. Inst. Can. Sci. Technol. Aliment., 8, 199± 201. Van der Pijl, L. (1982) Principles of Dispersal in Higher Plants. Springer-Verlag, Berlin. Weevers, T. (1907) Die Funktion der Xanthinderivate im Pflanzenstoffwechsel. Arch. Neerl. Sci. IIIB, 5, 111±95. VitoÂria, A.P. & Mazzafera, P. (1999) Xanthine degradation and related enzyme activities in leaves and fruits of two Coffea species differing in caffeine catabolism. J. Agric. Food Chem., 47, 1851±5. Yasuda, T., Fujii, Y. & Yamaguchi, T. (1985) Embryogenic callus induction from Coffea arabica leaf explants by benzyladenine. Plant Cell Physiol., 26, 595±7. Yasuda, T., Tahara, M., Hatanaka, T., Nishibata, T. & Yamaguchi, T. (1995) Clonal propagation through somatic embryogenesis of Coffea species. In: Proceedings of the 16th ASIC Colloquium (Kyoto), pp. 537±41. ASIC, Paris, France. Zamarripa, A., Ducos, J.P., Bollon, H., Dufour, M. & Petiard, V. (1991b) Production d'embryons somatiques de cafeÂier en milieu liquide: effets densite d'inoculation et renouvellement du milieu. CafeÂ, Cacao, TheÂ, 35, 223±44. Paris, France. Zamarripa, A., Ducos, J.P., Tessereau, H., Bollon, H., Eskes, A.B. & Petiard, V. (1991a) DeÂveloppement d'un proceÂde de multiplication en masse du cafeÂier par embryogenese somatique en milieu liquide. In: Proceedings of the 14th ASIC Colloquium (San Francisco), pp. 392±402. ASIC, Paris, France.
Chapter 11
Agronomy III: Molecular Biology John I. Stiles Integrated Coffee Technologies, Inc. Honolulu, Hawaii USA 11.1 INTRODUCTION Despite the importance of coffee as a commodity in world trade and as a major component of foreign exchange for many producing countries, the application of molecular biology and biotechnology to coffee has lagged behind many other crops. The first coffee gene sequences were not entered into GeneBank until 1994 and still only a handful of complete coding sequences are known. Even today, considering its economic importance, only a few laboratories are working on coffee molecular biology and biotechnology. Undoubtedly, this is due in part to the development of these techniques in research centers located predominantly in northern temperate zones, where coffee is not commercially grown, and in part to the perennial nature of coffee which makes it less attractive as a model organism and less attractive to life science companies that rely on the sale of seeds of high volume annuals, such as corn, soybeans and cotton, for revenue. As will be discussed below, many of the techniques that are standard with temperate crops are still difficult to apply to coffee. However, with the continued development and widespread dissemination of molecular techniques and biotechnology, the application of this technology to coffee will continue to increase. Coffee farmers face many challenges in producing an abundant and high quality crop. Biotic stresses, especially from insects and fungal pathogens, are particular problems. There appears to be no naturally occurring resistance gene for certain important insect pathogens such as coffee berry borer (Hypothenemus hampei). Although biological control mechanisms are being developed, coffee berry borer is still considered the most economically important insect pathogen of coffee and effective resistance introduced through biotechnology would have a major effect on the lives of coffee farmers in many areas of the world.
While there are known resistance genes to some fungal diseases of coffee, the long-term nature of the breeding cycle and the need for durable resistance, due to the perennial nature of coffee, make biotechnology an important potential tool for sustainable coffee production. The development of resistant cultivars is particularly difficult due to the long breeding time (25 to 30 years) and the need for not only agronomic characteristics, but also cup quality for acceptance of new varieties. The ability of biotechnology to move natural coffee resistance genes and non-coffee resistance genes into established cultivars and to pyramid resistance genes will save considerable time and provide sustainable and high quality coffee production. Abiotic stresses, such as freezing and drought, also represent significant problems in the coffee industry. For example, the periodic freezes experienced in certain production areas of Brazil not only cause severe economic loss to the farmers involved, but also disrupt commodity markets affecting importers, roasters, and consumers. The introduction of even a minimal increase in frost tolerance through molecular techniques would be of great importance in alleviating the disruption that a freeze can cause the coffee industry. Finally, cup quality is increasingly important as a result of the growth of the specialty coffee industry. Biotechnology will undoubtedly play a progressively more important role in assuring cup quality by reducing defects resulting from contamination by microorganisms and insects. However, it may also play a more direct role by modifying the chemicals present in the green bean. For example, coffee grown without caffeine would negate the need for chemical decaffeination and the resulting decrease in coffee quality. Although the application of molecular biology and biotechnology is still in its infancy, the current progress will be reviewed and prospects for the coffee industry discussed.
224
Agronomy III: Molecular Biology
11.2 COFFEE GENES The complete sequences of only a few coffee genes have been reported. And, even fewer genes have been characterized to any extent. Most of the nucleic acid sequences that have been reported are partial sequences, used only for phylogenetic studies. However, this is now changing rapidly. The first coffee gene to be isolated was a-galactosidase from coffee seeds (beans). Coffee bean a-galactosidase can cleave the terminal a1,3-linked galactose residues from the surface of the B-type blood group, converting it to O-group serology and permitting it to be used as a `universal donor' for transfusion therapy (Goldstein et al., 1982; Goldstein, 1989). The idea was to clone a-galactosidase from coffee seed and then produce it in high quantities in a microbial or cell culture system to obtain enough enzyme for commercial use. Zhu and Goldstein (1994) purified agalactosidase from dried green coffee beans and obtained a partial amino acid sequence using Nterminal sequencing and sequencing of CNBr fragments. Part of the cDNA was obtained by polymerase chain reaction (PCR) amplification of a segment of the gene using cDNA made from total seed mRNA and primers constructed using the partial amino acid sequence. The 50 and 30 ends of the cDNA were obtained using the RACE technique (Zhu & Goldstein, 1994). Coffee a-galactosidase shows about 80% similarity to guar (Cyamopsis tetragonoloba) a-galactosidase, previously isolated by Overbeeke et al., (1989), even though the guar enzyme principally cleaves a 1,6 glycoside linkages (Guiseppin et al., 1993), whereas the coffee enzyme cleaves mainly a 1,3 and a 1,4 linkages. The coffee a-galactosidase shows more than 50% homology to a number of other a-galactosidases from diverse organisms including human, yeast and Aspergillus niger, although there is little homology to agalactosidases from prokaryotic organisms (Zhu & Goldstein, 1994). The identity of the coffee seed agalactosidase was proven by insertion of the coffee cDNA into a baculovirus expression system and identification of a-galactosidase activity in transformed but not non-transformed insect cells. Perhaps the best-characterized coffee genes are those encoding the 11S seed storage protein. AcunÄa et al., (1999) and Rogers et al., (1999) have recently published detailed investigations on the structure and sequence of coffee 11S proteins and cDNAs. Marraccini et al., (1999) have cloned a complete 11S seed storage protein gene and carried out promoter analysis in transgenic tobacco plants. The 11S seed storage
225
protein is the most abundant protein in coffee seeds and is found principally in the storage vacuole of endosperm cells. The coffee seed 11S storage protein is similar to other legumin-like seed storage proteins. In the absence of reducing agents, the 11S protein has an apparent molecular mass of about 55 kDa and can be converted by reducing agents into two polypeptides of about 32 to 33 kDa (designated a-subunit) and 20 to 24 kDa (designated b-subunit) (Rogers et al., 1999; AcunÄa et al., 1999). Figure 11.1 shows an SDS polyacrylamide gel separation of seed proteins under reduced and non-reduced conditions. Microsequencing of seed proteins separated by two-dimensional gel electrophoresis indicates that there is heterogeneity among both the a subunit and the b subunit of the 11S proteins (Rogers et al., 1999). This heterogeneity was found in C. arabica, a tetraploid, and C. canephora, a diploid, that is thought to be one of the parents of C. arabica (Lashermes et al., 1999). Furthermore, comparison of the deduced amino acid sequences from three cDNA sequences revealed a number of amino
Fig. 11.1 SDS polyacrylamide gel electrophoresis of coffee seed (bean) proteins under reducing and nonreducing conditions. P indicates the major species of the mature 11S seed storage protein under nonreducing conditions where the a chain and b chains are attached by a disulfide bridge. a and b indicate the a chain and b chains separated by reduction of the disulfide bridge. Reproduced from Rogers et al. (1999) with permission.
226
acid substitutions as well as deletions/insertions, indicating that the 11S seed storage proteins are probably members of a gene family, as is the case with seed storage proteins in other plants, although the existence of a multigene family cannot, at this time, be confirmed by Southern blotting (Rogers et al., 1999; AcunÄa et al., 1999). As with other seed storage proteins, regulation of expression is at the transcriptional level. Rogers et al., (1999) found a large accumulation of 11S mRNA concomitant with the increase of 11S protein in developing coffee seeds. The 11S mRNA is absent in developing seeds until about 15 weeks after flowering. The message accumulates rapidly and remains high between 18 and 27 weeks after flowering, the time that 11S protein accumulation is at its peak (Fig. 11.2). Message levels decrease after week 27 after flowering, as does further accumulation of the 11S protein.
Fig. 11.2 Accumulation of 11S seed storage protein and mRNA during seed development. Reproduced from Rogers et al. (1999) with permission.
Synthesis and processing of the coffee 11S seed storage protein are similar to that of other legumin-like seed storage proteins. Marraccini et al., (1999) cloned a genomic copy of the 11S protein using the inverse polymerase chain reaction. The DNA sequence of this gene is consistent with it being the gene encoding the csp1 cDNA that they had previously isolated (Rogers et al., 1999). The gene contains three introns, two of 111 base pairs, and one of 79 base pairs. The introns are located at exactly the same positions as those in the other known legumin genes. The presumed full-length csp1 cDNA predicts an mRNA with a 32 base 50 untranslated leader region, a 1476 base open reading frame encoding a 492 amino acid protein (55 kDa), and
Coffee: Recent Developments
a 30 untranslated region of 195 bases. Comparison of the predicted protein sequence in the cDNA and the N-terminal sequences determined from the isolated protein predicts a signal sequence of 26 amino acids at the N-terminus. The predicted protein encoded by the cDNA contains a sequence NGLEET. This is identical to a highly conserved cleavage site found in 11S storage proteins from other plants. Cleavage occurs between the N and the G. The functionality of this site in coffee 11S proteins is confirmed by the occurrence of the GLEET sequence at the N-terminus of two different b-chains purified by Rogers et al., (1999), using twodimensional gel electrophoresis, and also by AcunÄa et al., (1999) using SDS polyacrylamide gel electrophoresis and N-terminal sequencing. Figure 11.3 shows the sequence of synthesis and processing of the 11S protein. The 55 kDa preproprotein is the initial product. Cleavage of the Nterminus at the sequence E/QPRL 26 or 27 amino acids from the ATG translational initiation codon, depending on the gene family member, was determined by N-terminal sequencing of the a-peptide (AcunÄa et al., 1999; Rogers et al., 1999). This is consistent with other 11S storage proteins. N-terminal sequencing indicates that the conserved NGLEET sequence directs the cleavage of the pre-protein into an acidic peptide (designated a chain by Rogers et al., 1999, and used in this chapter) and a basic peptide (designated b chain by Rogers et al., 1999, and used in this chapter). In other 11S storage proteins the two chains are held together by a disulfide bridge between two conserved cysteines. AcunÄa et al., (1999) predict, based on the analogy to other legumins, that the disulfide bridge involves C112 and C307. They also predict an internal a chain disulfide bridge between C36 and C69 based on analogy to other legumins. Overall, the coffee 11S seed storage protein appears to be a fairly typical legumintype seed storage protein. Marraccini et al., (1999) cloned a coffee 11S storage genomic gene using the inverse polymerase chain reaction (IPCR). They obtained about 1 kb of the promoter region and about 0.9 kb of the 30 region in addition to the coding region. Sequence analysis of the promoter region indicated several motifs that occur in other seed storage protein genes that are responsible for both temporal and spatial regulation. The sequence TGTAAAG appears 757 bp and 181 bp upstream of the ATG translational initiation site. This sequence is similar to the endosperm motif TGTAAAGT found in barley and wheat glutenins, pea legumin, maize zein and barley hordein promoters (see Marraccini et al., 1999 for references). Marraccini et al., (1999) also
Agronomy III: Molecular Biology
227
Fig. 11.3 Structure of the 11S seed storage protein gene and processing steps to give the mature 11S seed storage protein. The position and identity of conserved motifs in the promoter are shown: . endosperm-like motif; ^ GCN4-like motif; $ TGAC-like motif; $ soybean box; ^ E-box motif; * RY repreats. See text for full description. Pre-mRNA is shown with position of three introns indicated. Pre-proprotein is shown with signal sequence cleaved at the PQPRL site. a and b chains are produced by cleavage at the NGLEET consensus cleavage site. The arrows indicate the exact point of cleavage.
reported several other potential motifs including a `GCN4-like motif' similar to that found in the barley C-hordein promoter at positions 7742 and 7181 (with respect to the translational initiation site). The coffee sequences are TGAGTC and TGAGT, respectively, and the GCN4 motif is ATGA(C/ G)TCAT. There is also a TGAC-like motif at position 7326. This motif has been shown to be essential for pea lectin seed-specific expression (de Pater et al., 1993). There are two `soybean boxes' at positions 7248 and 742. This sequence has been found to be essential for proper expression of several soybean genes (Goldberg et al., 1989). Marraccini et al., (1999) reported the existence of sequences similar to a number of other motifs including the E-box of phaseolin, the RY repeat regions of the legumin box, and the AT-rich enhancer motif of soybean b-conglycinin. However, transcriptional regulation consensus sequences by their nature, are not exact and are generally fairly short. Thus, similar sequences can often be found in DNA sequences of reasonable length and confirmation by other means, such as promoter deletion analysis, is required to adequately address their significance. Marraccini et al., (1999) fused the upstream sequences of the coffee 11S promoter to the uidA gene (GUS gene) and transformed tobacco to address this question. GUS expression was measured in seed and leaf tissue of the transformed tobacco to assess the level and specificity of expression. Four different promoter constructs were analyzed, along with a control with no promoter, and a positive control with constitutive expression driven by the 35S promoter. Fusion of the 11S promoter to the GUS coding
sequence was immediately after the region coding for the fifth amino acid of the coffee 11S protein in all cases. The four promoter constructs used contained, with respect to the ATG translational initiation codon, 945 bp, 695 bp, 445 bp, and 245 bp of the promoter region. The level of expression in tobacco seeds was relatively high and not significantly different between the 945 bp and 695 bp promoters. There was a decrease when the promoter was shortened to 445 bp and a further decrease when shortened to 245 bp. None of the constructs showed detectable levels of expression in leaves, and all constructs, including the 245 bp promoter that showed the lowest level of activity, were significantly higher than the 35S promoter. Since the 945 bp and 695 bp promoters were not different in either strength or specificity it can be concluded that no essential sequences lie in the 250 bp between positions 7695 and 7945. There are several potential consensus sequences in this 250 bp region, but additional copies of all of these also occur within the 7695 region. Although it is difficult to show in a statistically rigorous manner, there appears to be a drop of about 50% in promoter strength by deletion of the region between 7445 and 7695. This region is relatively devoid of consensus sequences except for the only occurrence of the AT-rich enhancer motif and an E-box sequence. It is tempting to speculate that the drop of 50% in promoter strength is due to the removal of the putative AT-rich enhancer. However, a more detailed analysis is necessary to confirm the importance of this motif. Promoter strength, but not specificity, is further reduced by removal of the sequences between positions
228
7245 and 7454. This region contains two putative RY-motifs, an endosperm motif, and a soybean box that overlaps the 7254 site. From the initial analysis presented by Marraccini et al., (1999) it is not possible to assign specific roles to specific motifs; however, their analysis does point the way for future work to uncover the exact details of the coffee 11S promoter. A common problem with promoter analysis by transformation, and present in the data of Marraccini et al., (1999), is the relatively large deviation from one transformant to another. This is most likely a result of `position effects' resulting from the random integration that occurs with current plant transformation technology. However, until technology for site-specific transformation has been perfected this is a technical limitation that one must live with. The seed specificity of the 11S storage protein promoter may make it a useful tool for coffee biotechnology. By varying the length of the promoter it should be possible to impart the desired level of expression in a tissue-specific manner to coffee seeds (beans). This technology might be used for expression of disease-resistant genes specifically in the seed at the desired amount. It could also be used to express genes that affect quality factors involved in cup quality or soluble solid, important to the soluble coffee market. Neupane et al., (1999) have cloned two genes involved in ethylene biosynthesis in ripening coffee fruit. Fruit-expressed 1-amino-cyclopropane-1carboxylic acid (ACC) synthase and ACC oxidase were isolated from a cDNA library constructed using mRNA from ripening coffee fruits. Degenerate deoxyoligonucleotide primers were synthesized based on the amino acid sequence of highly conserved regions of known ACC syntheses and ACC oxidases. These primers were used in reverse transcriptase polymerase chain reactions (RPCR) to synthesize a portion of the ACC synthase and ACC oxidase genes using cDNA synthesized from ripening fruit mRNA as the template. The RPCR products were then used as probes to screen a cDNA library constructed from mRNA isolated from ripening coffee fruits to obtain full (or near full) length cDNAs of both genes. The largest ACC synthase cDNA isolated was 2040 bp in length. It contained an open reading frame of 488 amino acids. The deduced amino acid sequence of this cDNA is between 51% and 68% identical to other ACC syntheses and also contains all of the highly conserved regions. The largest ACC oxidase cDNA is 1320 bp with a 318 amino acid open reading frame that is between 50% and 83% identical to other ACC oxidases.
Coffee: Recent Developments
Although it has been known for some time that coffee fruit past a certain stage of development will ripen in response to ethylene, it has yet to be demonstrated that coffee fruit exhibits a climacteric. The data of Neupane et al., (1999) indicate that coffee is a climacteric fruit based on the accumulation pattern of both ACC synthase and ACC oxidase mRNA during fruit ripening. Figure 11.4 is a northern blot showing the accumulation of ACC synthase and ACC oxidase mRNA in coffee fruit at various stages of development. Messenger RNA was isolated from immature green, mature green, approximately 25% red, 50% red, 75% red and 100% red coffee fruits, separated by gel electrophoresis, blotted to a membrane and simultaneously probed with ACC synthase and ACC oxidase cDNAs. Both ACC synthase and ACC oxidase mRNAs accumulate during fruit ripening in a manner consistent with that of a climacteric fruit (Neupane et al., 1999). Conclusive proof that coffee is a climacteric fruit will come from observing the effect of inhibition of ethylene biosynthesis. These experiments are in progress.
Fig. 11.4 Northern blot of seed (bean) total RNA isolated at the developmental stages indicated. The blot was simultaneously probed with radioactively labeled cDNA for fruit-expressed ACC synthase and ACC oxidase.
An area of considerable interest in coffee molecular biology and biotechnology is caffeine biosynthesis. Figure 11.5 shows a consensus caffeine biosynthetic pathway that has developed over the years from radiolabeled feeding experiments and a limited amount of biochemical studies (see Crozier et al., 1997). The first step unique to the caffeine biosynthetic pathway is the methylation of xanthosine at the N7 position by xanthosine-N7 -methyl transferase. It has also been reported that XMP can serve as the initial substrate (Schulthess et al., 1996). After cleavage of the ribose to
Agronomy III: Molecular Biology
Fig. 11.5 Caffeine biosynthetic pathway. Initial substrate is xanthosine or perhaps xanthosine monophosphate. After methylation at the N7 position, the ribose is cleaved to form 7-methylxanthine. Two subsequent methylation reactions produce caffeine (trimethylxanthine).
form 7-methylxanthine, two additional methylations occur to form caffeine. The only caffeine biosynthetic pathway gene isolated to date is the gene encoding xanthosine-N7 -methyl transferase from coffee (Moisyadi et al., 1998, 1999). The gene was cloned using the `classical' biotechnology approach of purifying the enzyme, obtaining a partial amino acid sequence and back translating this amino acid sequence to obtain a degenerate oligonucleotide sequence. Degenerate PCR primers were synthesized based on the amino acid sequences obtained and a portion of the gene was synthesized using PCR. This partial gene sequence was then used to obtain the entire coding region by screening a cDNA library constructed from young leaf tissue mRNA. The most pure XMT preparations, when analyzed by two-dimensional gel electrophoresis, contain four peptides that separate into groups of two peptides of about 41 kDa and two peptides of about 40 kDa. One peptide of each size class has a charge very similar to a peptide of the other size class (Fig. 11.6) (Moisyadi et al., 1998, 1999). There are also two higher molecular weight proteins present in the most purified preparations. Partial amino acid sequencing of peptide fragments from these proteins identifies them as known `housekeeping' enzymes (Moisyadi & Stiles, unpublished data). Two cDNA clones have been completely sequenced. The largest of these contains an open reading frame encoding a protein of 41 kDa (Fig. 11.7). There is moderate similarity to an Arabidopsis cDNA that
229
Fig. 11.6 Two-dimensional gel showing the most purified fraction of xanthosine-N7-methyltransferase. The most purified xanthosine-N7-methyltransferase preparation has four major peptides that have slightly different isoelectric points and are slightly different in size.
contains an open reading frame encoding a protein of unknown function and to some putative Arabidopsis GDSL-motif lipase/hydrolases. The coffee XMT also contains sequences similar to motifs determined to be involved in adenosine and s-adenosylmethionine binding sites in certain prokaryotic modification enzymes that methylate the N6 position of adenine residues in restriction/modification sites in DNA. One
Fig. 11.7 Amino acid sequence of xanthosine-N7methyltransferase from coffee. The sequence YPPY (light shading) is similar to a conserved motif (D,S,N)PPY found in most N6-adenine methyltransferases and is most likely involved in the active site of the enzyme. The dark shaded sequence is similar to those found in N6-adenine methyltransferases and has been identified as the s-adenosylmethionine binding site.
230
such conserved motif, (D,S,N)PPY, appears in most N6 methyltransferases (Timinskas et al., 1995). Mutation of the D or the Y of the DPPY sequence found in the EcoRV adenine-N6 -methyltransferase greatly decreased enzymatic activity without affecting s-adenosylmethionine binding. A related sequence, YPPY, is found starting at position 60 in the coffee XMT sequence. A second sequence starting at position 250 in XMT shows some similarity to the s-adenosylmethionine binding site described by Roth et al., (1998). However, these consensus sites and the match between them and the XMT gene and the differences in the structures of adenine and xanthine make conclusions based on these similarities difficult. In an attempt to prove the identity of the coffee XMT gene, transgenic tobacco plants were generated that expressed the XMT cDNA sequence under control of the 35S promoter. Transgenic tobacco plants expressing the coffee XMT gene were selected by northern blot analysis. Proteins were extracted from young leaves of these plants and assayed directly or partially purified and assayed. Figure 11.8 shows that transgenic tobacco expressing the coffee XMT cDNA has detectable levels of XMT activity and that the specific activity increases with purification. Coffee plants with the XMT cDNA in antisense are currently being characterized and it is anticipated that caffeine-free arabica coffee, at least to the extent of
Fig. 11.8 Expression of the coffee xanthosine-N7methyltransferase in tobacco. The coffee xanthosineN7-methyltransferase cDNA was expressed in tobacco under control of the CaMV 35S promoter. Total proteins were extracted from young leaves of normal tobacco plants or from those expressing the xanthosine-N7-methyltransferase mRNA and assayed directly or purified by hydrophobic interaction chromatography (HIC) or both HIC and Cibacron blue F3GA (BioRad) (Affi-blue).
Coffee: Recent Developments
currently available decaffeination process, will be produced.
11.3 TRANSFORMATION SYSTEMS FOR COFFEE A number of different transformation systems have been reported for arabica and robusta coffee, although there are only a few reports of recovery of whole plants with stable integration of DNA. Transformation has been reported using the biolistic method (gene gun) (Van Boxtel et al., 1995), DNA electroporation using protoplasts (Barton et al., 1991), and various Agrobacterium systems (Spiral & Petiard 1991; Ocampo & Manzanera, 1991; GreÁzes et al., 1993; Spiral et al., 1993; Leroy et al., 1997), however, there are few reports of stable transformation of whole plants. The first report of stable transformation of coffee plants used a protoplast system and electroporation (Barton et al., 1991). Barton and co-workers reported the recovery of plantlets after selection on kanamycin for transformed protoplasts and regenerating tissue. Unfortunately, the root systems of the regenerated transgenic plants were not well developed and plants capable of flowering were not produced. This work was done on a single not well-defined arabica and it is not known if this technology will transfer to all or most other varieties. Protoplast regeneration is known to be quite genotype-dependent. Although many different systems have been investigated, most of the current work has utilized the standard Agrobacterium tumefaciens system, despite the relatively low rate of transgenic plant production obtained to date. The most advanced transformation work is that describing the production of coffee expressing the Bacillus thuringiensis CryIA(c) gene. The CryIA(c) gene product is an insecticidal protein that is toxic to certain insects including the coffee leaf miner, Perileucoptera coffeela (Guerreiro Filho et al., 1998). Transgenic coffee plants containing the CryIA(c) gene were produced using both the Agrobacterium rhizogenes and Agrobacterium tumefaciens systems (Leroy et al., 1997). Although higher rates of transformation were initially found using the A. rhizogenes system, the `hairy root' phenotype could not be suppressed and led to plants with unacceptable agronomic traits including lack of flowering. These problems resulted in further work using the A. tumefaciens system. Spiral and co-workers found that the NPTII kanamycin resistance gene was not effective as a selection system when coffee somatic embryos were
Agronomy III: Molecular Biology
transformed. However, they reported that the gene for resistance to the herbicide chlorsulfuron was effective (Spiral et al., 1999). The initial frequency of transgenic plants produced is considerably below that of many other plant transformation systems. Using the A. tumefaciens system only, about 0.4% of the somatic embryos infected gave transgenic plantlets (Leroy et al., 1997). However, improvements in the selection system have recently been reported. Leroy and coworkers found between 30 and 80% of embryogenic calli transformed with the chlorsulfuron resistance gene and the GUS gene were GUS positive after 10 to 12 months of selection on chlorsulfuron (Leroy et al., 2000). As yet, only limited DNA analysis has been presented; nonetheless, this represents a potentially significant advance in transformation efficiency. However, the results are still quite variable and show a genotype dependence (see below) (Leroy et al., 2000). The work of Leroy and co-workers is also the first, and thus far only, report of stable transformation and regeneration of coffee plants with a useful agronomic trait, insect resistance. As previously mentioned, the cryIA(c) gene was inserted into robusta (C. canephora) and two different arabica (C. arabica) coffee varieties, a Catimor (8661±4) and an F1 hybrid (Et29 6 Ca5). The Catimor and robusta varieties transformed at approximately the same efficiency, but the F1 hybrid produced for fewer transgenic plants. Of the 23 transgenic plants examined for the cryIA(c) protein using an antibody, 18 plants had detectable levels. Preliminary bioassay data indicate that at least some of the transgenic plants that express cryIA(c) also have significant decreases in their overall bioassay score and in the number of P. coffeela pupae detected on the leaves. These plants are intended for field trials starting in 2000.
11.4 PROSPECTS Despite the current controversy in some parts of the world, molecular biology and biotechnology will play an increasingly important role in the improvement of coffee cultivars. These technologies will have benefits for both farmers and consumers. Coffee is a perennial crop grown in a tropical environment. Like other tropical crops, coffee is under severe pressure from fungal and insect pathogens. Biotechnology can have a significant impact on the ability of farmers to produce high quality coffee, perhaps with unique properties, while preserving the ecosystem by reducing chemical use.
231
A potential application of biotechnology is resistance to the coffee berry borer (Hypothenemus hampei). There is no known resistance in coffee to H. hampei and, because a significant portion of its life cycle is in the berry, insecticides are often not highly effective. Reported losses can range up to 96% (Nyambo & Masaba, 1997). A biotechnology-based control would be effective, relatively inexpensive compared to chemical treatments, and would reduce chemical contamination of the ecosystem. This situation is analogous to the use of Bt-maize in the United States to control European corn borer. Under significant disease pressure, yields increased by as much as 14%, while pesticide usage decreased (Gianessi, 1999). Broad-spectrum fungal resistance such as that imparted by certain hydrolytic enzymes, either alone or in combination, could play an important role in coffee disease resistance. Pathogenic fungi result in significant losses and include diseases such as coffee berry diseases (Colletotrichum kahawae), coffee leaf rust (Hemileia vastatrix), and a number of other diseases such as those caused by Fusarium and Micena. Although there are naturally occurring resistance genes for many of the fungal diseases, additional mechanisms of resistance would be of great value. Since coffee is a perennial plant with a long breeding cycle, breakdown of an existing resistance gene is especially significant as the development of new resistant lines takes many years. In coffee this is a particular problem in that cultivars must not only carry an effective resistance gene, but also have acceptable cup quality and even have the cup characteristics expected for a coffee from the country of origin. This makes breeding of new lines especially complex and time consuming. A biotechnology approach that could `pyramid' additional resistance genes into existing cultivars that are accepted by the industry would save considerable time and avoid potential acceptance problems. A second problem with fungal contamination of coffee beans is the production of fungal toxins such as ochratoxin. Ochratoxin is now a serious concern, especially in the European community. Vega and Mercadier (1999) have shown that insects such as H. hampei can act as a vector for fungi such as Aspergillus flavus and A. ochraceus that can produce aflatoxins in infected coffee beans. The introduction of resistance genes to H. hampei and/or the introduction of broadbased fungal resistance genes should decrease the instances of contaminated coffee beans. Biotechnology can also decrease susceptibility to environmental stresses. Abnormal weather, especially freezes, can cause severe disruptions in coffee markets,
232
affecting both farmers and roasters. Since coffee is grown in areas where freezing weather is infrequent and generally of short duration, coffee may be an ideal crop for engineered freeze protection. There are multiple changes that occur in plants during cold acclimation to stabilize membranes, thought to be the initial site of cold/freezing damage (Tomashow, 1999). These changes include alteration in lipid composition, accumulation of sucrose and other sugars and synthesis of certain proteins that help to stabilize membranes such as the late embryogenesis abundant proteins (LEA proteins) and highly hydrophilic proteins such as the COR156 protein of Arabidopsis. Intervention using biotechnology to produce one or more of these protective factors could give coffee the ability to withstand most low temperature stresses that are likely to be encountered in the present growing range. The use of genetic transformation technology to manipulate quality and flavor traits is still problematic. Most quality traits are not yet understood in detail great enough to identify specific genetic changes that will result in specific flavor or quality characteristics, although it may be possible to remove some defects. One potential defect found in most robusta coffees at varying levels is methylisoborneol (MIB). Although it is not yet known whether MIB is synthesized in the plant or an associated microbe, biotechnology could be used to eliminate MIB production, increasing the quality of robusta coffees. A considerable amount of criticism of genetically engineered plants has resulted from the use of foreign genes used as selectable markers, especially markers that are antibiotic resistance genes. At this time most genetically engineered plants carry a foreign gene that imparts a selection that is used during the transformation process. This is needed since the DNA uptake process is not efficient and many untransformed cells remain mixed with the transformed cells. The gene most commonly used for selectable markers is the neomycin phosphotransferase type II (NptII) gene. This gene gives transformed cells the ability to grow in the presence of kanamycin and related antibiotics, whereas growth of normal plant cells is inhibited. Genes that give resistance to herbicides such as chlorsulfuron have also been used. Although a considerable body of scientific evidence indicates that these genes are not a danger to the environment or to consumers, they have nonetheless met with considerable resistance in certain countries. New technologies based on different selection markers or even the elimination of such genes are under development and promise alternatives to the present situation.
Coffee: Recent Developments
One type of new selection system is based on giving the transformed cells the ability to grow on carbon sources that normal plant cells cannot utilize. One such system uses the phosphate-6-mannose isomerase gene to give plant cells the ability to grow on mannose as the sole carbon source (Joersbo et al., 1998). This system is reportedly up to 5-fold more efficient as a selection system than existing markers such as NptII. Selection systems such as this do not impart a trait that would give the plants a selective advantage in the field (as would a herbicide resistance gene) and do not involve antibiotic resistance genes, so most of the objections to current selection genes are met. New more efficient transformation systems such as the use of pollen may alleviate the need for selectable markers. Smith et al., (1994) have reported a pollen transformation system that uses electroporation to introduce DNA into pollen to give transgenic plants at a rate of up to 44% when this pollen was used to pollinate tobacco flowers. With transformation rates this high, selectable markers would not be needed. Technologies that can impart site-specific genetic changes may also alleviate the need for the introduction of foreign DNA. These technologies could be-used to inactivate an existing gene to turn off production of some product such as caffeine or MIB, or perhaps even alter the gene product to impart some new property such as alternation of an existing disease resistance gene to give resistance to a new pathogen or a broader spectrum of pathogens. Two different approaches show promise for site-specific genetic changes. Two groups have recently demonstrated the use of chimeric RNA/DNA oligonucleotides to selectively and specifically mutate plant genes, at least in model systems. Zhu et al., (1999) were able to use chimeric oligonucleotides to mutate the maize acetohydroxyacid synthase gene to confer resistance to either imidazolinone or sulfonylurea herbicides. They were also able to restore activity to a mutant green fluorescent protein gene. The chimeric oligonucleotides were designed to change a single base in the plant DNA that changed a single amino acid in the acetohydroxyacid synthase protein and resulted in herbicide resistance. In one experiment, 34 out of 40 events characterized had the expected base change. The conversion frequency was generally in the range of about 1 6 10ÿ4 . However, there were also a number of unexpected changes in the target gene. This system appears to work by inducing mismatch repair and it may be that correction of the mismatch required to fix the mutation is not exact. Beetham et al., (1999) reported similar results in mutating the tobacco acetohydroxyacid synthase gene
Agronomy III: Molecular Biology
using essentially the same system. Although the frequency of the events is moderate, as compared to site-specific homologous recombination, it would be difficult to use this system at present to make changes that did not impart some type of selection. An additional limitation is that the mutations produced have been limited to single base changes in single-copy genes. To use this technology to change the characteristics of a gene might require several rounds of mutation if multiple changes in a single gene were required. Also, plant genes often exist as multi-gene families. Mutation of all members of a family would require multiple rounds of mutation. Nonetheless, the demonstration of the potential of chimeric oligonucleotides to mutate plant genes in a site-specific manner is an important advance. A second approach to site-specific genetic changes is enhanced homologous recombination. Homologous recombination has been used in yeast and some animal systems to make specific genetic changes in specific genes. However, except for some model systems, the efficiency of homologous recombination in plants has been too low for practical use. The use of proteins, such as the E. coli RecA protein, involved in the recombination process to enhance the rate of sitespecific recombination, has been reported (Pati et al., 1997). An advantage of this system, if adaptable to plants, is that it should allow more extensive changes than those reported using chimeric oligonucleotides. However, changes to all members of a gene family or duplicated genes in a tetraploid, such as C. arabica, would still require multiple rounds of mutagenesis. The prospects are excellent for applying molecular biology and biotechnology to the improvement of agronomic and quality traits in coffee. The field trials now being initiated to test coffee transformed with the Bt protein to give resistance to leaf miner should demonstrate the potential of biotechnology to alleviate significant agronomic problems. Transgenic caffeinefree coffee now undergoing the final stages of development should demonstrate the potential for quality improvement using biotechnology. Coffee, like other perennial tropical crops, is under significant disease pressure that often requires considerable input of chemicals to obtain satisfactory yields. Many coffee farmers do not have the cash needed for effective chemical control. Furthermore, large-scale use of chemicals leads to environmental damage. Biotechnology can play an important role in improving yields while decreasing environmental damage and in adding traits of value to coffee. This should stabilize and increase the profitability of coffee farming and
233
improve the standard of living of many people involved in coffee production.
REFERENCES AcunÄa, R., BassuÈner, R., Beilinson, V. et al. (1999) Coffee seeds contain 11S storage proteins. Physiol. Plant., 105, 122±31. Barton, C.R., Adams, T.L. & Zarowitz, M.A. (1991) Stable transformation of foreign DNA into Coffea arabica plants. In: Proceedings of the 14th ASIC Colloquium (San Francisco), pp. 460±64. ASIC, Paris, France. Beetham, P.R., Kipp, P.B., Sawycky, X.L., Arntzen, C.J. & May, G.D. (1999) A tool for functional plant genomics: chimeric RNA/DNA oligonucleotides cause in vitro gene-specific mutations. Proc. Natl. Acad. Sci. USA, 96, 9774±9778. Crozier, A., Baumann, T.W., Ashihara, H., Suzuki, T. & Waller, G.R. (1997) Pathways involved in the biosynthesis and catabolism of caffeine in Coffea and Camellia. ASIC, Paris, France. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 106± 13. Gianessi, L. (1999) Agricultural Biotechnology: Insect Control Benefits. National Center for Food and Agricultural Policy, Washington, DC. Goldberg, R.B., Baker, S.J. & Perez-Grau, L. (1989) Regulation of gene expression during plant embryogenesis. Cell, 56, 149± 60. Goldstein, J. (1989) Conversion of ABO blood groups. Transfusion Med. Rev., 3, 206±12. Goldstein, J., Siviglia, G., Hurst, R., Lenny, L. & Reich, L. (1982) Group B erythrocytes enzymatically converted to group O survive normally in A, B and O individuals. Science, 215, 168±70. GreÁzes, J., Thomasset, B. & Thomas, D. (1993) Coffea arabica protoplast culture: transformation assays. In: Proceedings of the 15th ASIC Colloquium (Montpellier), pp. 745±7. ASIC, Paris, France. Guerreiro Filho, O., Denolf, P., Peforoen, M, Decazy, B., Eskes, A.B. & Frutos, R. (1998) Susceptibility of coffee leaf miner (Perileucoptera spp) to Bacillus thuringiensis d-endotoxins: a model for transgenic perennial crops resistant to endocarpic insects. Curr. Microbiol., 36, 175±9. Guiseppin, M.L., Almkerk, J.W., Heistek, J.C. & Verrips, C.T. (1993) Comparative study on the production of guar a-galactosidase by Saccharomyces cerevisiae SU50B and Hansenula polymorpha 8/2 in continuous cultures. Appl. Environ. Microbiol., 59, 52±9. Joersbo, M., Donaldson, I., Kreiberg, J., Petersin, S.G., Brunsted, J. & Okkels, F.T. (1998) Analysis of mannose selection used for transformation of sugar beet. Mol. Breed., 4, 111±17. Lashermes, P., Combes, M.-C., Robert, J. et al. (1999) Molecular characterisation and origin of the Coffea arabica L. genome. Mol. Gen. Genet. 261, 259±66. Leroy, T., Henry, A.-M., Royer, M. et al. (2000) Genetically modified coffee plants expressing the Bacillus thuringiensis cry 1Ac gene for resistance to leaf miner. Plant Cell Rep. (in press). Leroy, T., Paillard, M., Royer, M. et al. (1997). Introduction de
234
geÁnes d'inteÂreÃt agronomique dans I'espeÁce Coffea canephora Pierre par transformation avec Agrobacterium sp. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 439±46. ASIC, Paris, France. Marraccini, P., Deshayes, A., PeÂtiard, V. & Rogers, W.J. (1999) Molecular cloning of the complete 11S seed storage protein gene of Coffea arabica and promoter analysis in transgenic tobacco plants. Plant Physiol. Biochem., 37, 273±82. Moisyadi, S., Neupane, K.R. & Stiles, J.I. (1998) Cloning and characterization of a cDNA encoding xanthosine-N7 -methyltransferase from coffee (Coffea arabica). Acta Hort., 461, 367± 77. Moisyadi, S., Neupane, K.R. & Stiles, J.I. (1999). Cloning and characterization of xanthosine-N 7 -methyltransferase, the first enzyme of the caffeine biosynthetic pathway. In: Proceedings of the 18th ASIC Colloquium (Helsinki), pp. 327±31. ASIC, Paris, France. Neupane, K.R., Moisyadi, S. & Stiles, J.I. (1999) Cloning and characterization of fruit-expressed ACC synthase and ACC oxidase from coffee. In: Proceedings of the 18th ASIC Colloquium (Helsinki), pp. 322±6. ASIC, Paris, France. Nyambo, B.T. & Masaba, D.M. (1997) Integrated pest management in coffee: needs, limitations and opportunities. In: Proceedings of the 17th ASIC Colloquium (Nairobi), pp. 629±38. ASIC, Paris, France. Ocampo, C.A. & Manzanera, L.M. (1991) Advances in genetic manipulation of the coffee plant. In: Proceedings of the 14th ASIC Colloquium (San Francisco), pp. 378±82. ASIC, Paris, France. Overbeeke, N., Fellinger, A.J., Toonen, M.Y., Van Wassenaar, D. & Verrips, C.T. (1989) Cloning and nucleotide sequence of the a-galactosidase cDNA from Cyamopsis tetragonoloba (guar). Plant Mol. Biol., 13, 541±50. de Pater, S., Pham, K., Chua, N.H., Memelink, J. & Kijne, J. (1993) A 22-bp fragment of the pea lectin promoter containing essential TGAC-like motifs confers seed-specific gene expression. Plant Cell, 5, 877±86. Pati, S., Mirkin, S., Feuerstein, B. & Zarling, D. (1997) Sequence-specific DNA targeting. Encyclopedia of Cancer, Vol. III, pp. 1601±25. Rogers, W.J., BeÂard, Deshayes, A., Meyer, I., PeÂtiard, V. & Marraccini, P. (1999) Biochemical and molecular characterization and expression of the 11S-type storage protein from Coffea arabica endosperm. Plant Physiol. Biochem., 37, 261±72.
Coffee: Recent Developments
Roth, M., Helm-Kruse, S., Friedrich, T. & Jeltsch, A. (1998) Functional roles of conserved amino acid residues in DNA methyltransferases investigated by site-directed mutagenesis of the EcoRV adenine-N6 -methyltransferase. J. Biol. Chem., 273, 17333±42. Schulthess, B.H., Morath, P. & Baumann, T.W. (1996) Caffeine biosynthesis starts with the metabolically channelled formation of 7-methyl-XMP ± a new hypothesis. Phytochemistry, 41, 169± 75. Smith, C.R., Saunders, J.A., Van Wert, S., Cheng, J. & Matthews, B.F. (1994) Expression of GUS and CAT activities using electrotransformed pollen. Plant Sci., 104, 49±58. Spiral, J., Leroy, T., Paillard, M. & PeÂtiard, V. (1999) Transgenic coffee (Coffea species). Biotechnol. Agric. Forestry, 44, 55±76. Spiral, J. & PeÂtiard, V. (1991) Protoplast culture and regeneration in Coffea species. In: Proceedings of the 14th ASIC Colloquium (San Francisco), pp. 383±91. ASIC, Paris, France. Spiral, J., Thierry, C., Paillard, M., PeÂtiard, V. (1993) Obtention de plantules de Coffea canephora Pierre transformeÂse par Agrobacterium rhizogenes. C. R. Acad. Sci. Paris, t316, SeÂrie III, 1±6. Timinskas, A., Butkus, V. & Janulaitis, A. (1995) Sequence motifs characteristic for DNA [cytosine-N4] and DNA [adenine-N6 ] methyltransferases. Classification of all DNA methyltransferases. Gene, 157, 3±11. Tomashow, M.F. (1999) Plant cold acclimation: freezing tolerance genes and regulatory mechanisms. Ann. Rev. Plant Physiol. Plant Mol. Biol., 50, 571±99. Van Boxtel, J., Berthouly, M., Carasco, C., Dufour, M. & Eskes, A. (1995) Transient expression of b-glucuronidase following biolistic delivery of foreign DNA into coffee tissues. Plant Cell Rep., 14, 748±52. Vega, F.E. & Mercadier, G. (1999) The coffee berry borer and associated fungi. Presented at the 18th ASIC Colloquium (Helsinki). Zhu, A. & Goldstein, J. (1994) Cloning and functional expression of a cDNA encoding coffee bean a-galactosidase. Gene, 140, 227±31. Zhu, T., Peterson, D.J., Tagliani, L., St Clair, G., Baszczynski, C.L. & Bowen, B. (1999) Targeted manipulation of maize genes in vivo using chimeric RNA/DNA oligonucleotides. Proc. Natl. Acad. Sci. USA, 96, 8768±73.
Appendix 1
International Standards Organization (ISO) R.J. Clarke Formerly Chairman, ISO/TC34/SC15 The International Standards Organization, through its sub-committees SC15 of ISO/TC34, has continued to be concerned with the initiation and development of standards for coffee and coffee products. Since its inception in 1963, some 24 different standards have been issued, which can be categorised under the following headings, with dates of the most recent versions.
1.1 GLOSSARY RELATING TO COFFEE AND ITS PRODUCTS ISO 3509±1989, 3rd edn (1st edn 1977) BS 5456±1989
1.3 INSTANT COFFEE (SAMPLING PROCEDURES) Method of sampling from cases with liners ISO 6670±1983 = BS 6379±1984 under revision.
1.4 METHODS OF TEST (CHEMICAL OR PHYSICAL) Moisture content Green coffee Determination of moisture content Reference method: ISO 1446±1978 2nd edn = BS 5752±1979 Part 1 Routine method: ISO 1447±1983 2nd edn = BS 5752±1984 Part 2
1.2 GREEN COFFEE (GUIDES AND SAMPLING PROCEDURES) Guide to storage and transport ISO 8455±1986 = BS 6827±1987
Determination of loss in mass at 1058C Routine method: ISO 6673±1983 = BS 5752±1984 Part 7
Guide to specifying ISO 9166±1992 = BS 7601±1992
Guide to defects ISO 10470±1993 = BS 7683±1993
Roasted and ground coffee Determination of moisture content by Karl Fischer method
Method of sampling
Reference method: ISO 11187±1994 = BS 5752±1995 Part 13
ISO 4072±1982 = BS 6379±1983 Part 1
Specification of coffee tryer ISO 6666±1983 = BS 6379±1984 Part 3
Determination of moisture content (loss in mass at 1038C) Routine method: ISO 11294±1994 = BS 5752±1995 Part 14
Method of preparation for use in sensory analysis ISO 6608±1991 = BS 6379±1991 Part 4 235
236
Coffee: Recent Developments
Instant coffee
Green and roasted coffee
Determination of loss in mass at 708C (under reduced pressure)
Determination of a free flow bulk density of whole beans, by a free-flow method
ISO 3726±1983 = BS 5752±1984 Part 4
ISO 6669±1995 = BS 5752±1996 Part 16
Caffeine content
Instant coffee
Green coffee (also roasted and instant)
Particle size analysis
Determination of caffeine content (Levene method) Reference method: ISO 4052±1983 = BS 5752±1984. Part 3
Determination of caffeine content (HPLC method) Routine method: ISO 10095±1992 = BS 5752±1992 Part 12.
Other chemical content Instant coffee Determination of free and total carbohydrate contents by high performance anion exchange chromatography (HPAE) ISO 11292±1997 (corrected edition) = BS 5752±1995 Part 15
Visual and physical characteristics Green coffee Visual, olfactory examination and determination of foreign matter and defects ISO 4149±1980 = BS 5752±1980 Part 4
Size analysis (manual sieving) ISO 4150±1991 (2nd edn) = BS 5752±1991 Part 5
Determination of proportion of insect damaged beans ISO 6667±1985 = BS 5752±1986 Part 8
ISO 7532±1985 = BS 5752±1986 Part 10
Determination of free flow and compacted bulk densities ISO 8460±1987 = BS 5752±1987 Part 11
1.5 GENERAL COMMENTS All these standards (analytical methods, guides, or glossaries) are available in both English and French versions from the ISO in Geneva; while most will also have been published, practically verbatim, as National Standards in the national language by the respective National Standards bodies. Of particular interest are those most recently issued, which seek to use the most modern analytical methodology, for example Karl Fischer's methods for the measurement of true moisture content, ISO 11817 (1994), and also planned for green coffee. Similarly, high performance liquid chromatography is recommended now for caffeine determination (ISO 10095± 1992); and the even more sophisticated HPAE chromatography for carbohydrate determination (ISO 11292±1997). While many coffee standards for green coffee are for field use, close to harvesting and processing, it is also felt that so-called developing countries do now also have access to modern analytical laboratory techniques. A particular feature of the guides available is the comprehensive standard for defects in green coffee, issued 1993, ISO 10470, which should attract continued attention It is designed to develop full agreement and as much accuracy as possible in definitions, causes and effects of all the different kinds of defective beans that can be encountered in wet processed and dry processed arabica coffee, and dry processed robusta. It was decided that only qualitative statements can be made about the influence of particular defective beans, on brew flavour (after roasting), on account of the differing national perceptions, as reflected in marketing specifications.
Appendix 1
Increased harmonisation in this area is desirable, thus, for example, use of weighing methods rather than counting. So far SC15 has decided that it should not be involved in setting actual specifications for any aspect of green, roasted or instant coffee, but rather provide glossary guides and methods of test (by ring testing, where possible) for all chemical and physical characteristics of concern to the coffee trade. A description of the activities of ISO/TC34/SC15 was given in the ISO Bulletin, 1995, pages 13±15, by the former Chairman, R.J. Clarke, and of the ISO generally, in food standardisation, pages 8±12. A meeting of SC15, under the new Chairman, Dr R. Viani, was held in Paris, December 1999, to develop future strategy and activity. A question is often asked about the respective roles of the ISO and of Codex
237
Alimentarius (of the FAO/WHO) in food standardisation in general, and coffee in particular. Their activities can be said to be in parallel, and the two organizations in liaision. The scope of the latter, however, is primarily regulatory (for example in health/nutritional issues), as befits its governmental basis; whereas the former is non-governmental, dealing with the standardization of those issues of concern to international trade (for example, test methods, terminology, sampling, etc.), which it has done since 1947. Codex Alimentarius decided, some time ago, not to involve itself in coffee or instant coffee standardization; although the regulatory aspects in Europe are covered by European Directives in force from the relevant Council of Ministers.
Appendix 2
International Coffee Organization (ICO) C.P.R. Dubois Head of Operations, ICO 2.1 THE INTERNATIONAL COFFEE AGREEMENT 1994 2.1.1 Background The International Coffee Agreement 1994 was the fifth long-term agreement since 1962. The process of negotiation was long-drawn out since the initial aim had been to secure an Agreement with mechanisms for price stabilization, as in the past. This proved politically impossible and the Agreement which was eventually negotiated and entered into force on 1 October 1994 had no regulatory economic clauses, concentrating instead on maintaining the International Coffee Organization (ICO) and promoting international cooperation on coffee by other means. Mr Celsius A. Lodder, a Brazilian national, was appointed Executive Director. Mr. Lodder had worked in a variety of senior positions in the Brazilian civil service, including in coffee, and was also a professor of economics.
2.1.2 Priorities In line with the general objectives of the 1994 Agreement, the International Coffee Council approved a programme of action in May 1995 which identified four priority areas: expanding membership; reviewing statistical services and developing new analytical documents and information services on the coffee market; developing the Organization's capacity to sponsor projects for financing by the Common Fund for Commodities (CFC); and undertaking studies and surveys.
2.1.3 Coffee development projects As the designated International Commodity Body for coffee at the Common Fund for Commodities (CFC), the Organization is able to assist producing countries with projects to improve production and combat pests. Six major projects valued at over US $31 million were approved between 1995 and 1999. Funding was prin-
cipally secured from the Common Fund for Commodities, but significant co-funding from other bodies, such as the European Union and bilateral donor agencies, has been achieved. Funds were mainly invested in coffee-producing countries and were provided in the form of grants rather than loans. Areas covered included improvement of quality to secure added value, combat of pests and diseases, and improvement of marketing structures. The innovative aspect of this programme has been to get away from the usual one-country project approach to address issues of concern to coffee in a number of regions and environments, and to establish new techniques and methodologies relevant to a number of other countries.
2.1.4 Promotional activity With limited funds and modest budgets, using resources remaining from the Promotion Fund established under the 1987 and 1983 Agreements, the ICO has sought to enhance the image of coffee drinking and to increase the consumption of coffee in two of the world's largest markets, China and Russia. ICO generic promotion was only one of several factors that influenced consumption, but was widely perceived in China and Russia to have been beneficial. Activities included high profile Vanessa Mae concert promotions; production and dissemination of educational materials, including a new `Coffee story' booklet to create greater awareness of coffee; developing annual coffee festivals; and a programme of media briefings, supported by tastings and demonstrations, in order to educate journalists about the benefits of coffee. The development of strong and positive relationships with leading coffee companies, and the increasing extent of private sector participation in ICO promotions, were among the most important achievements. The educational elements of the promotion such as the `Coffee story' booklets, the programme of coffee tasting and the media briefings should continue to bear fruit in terms of creating a favourable image of `coffee culture'
238
Appendix 2
with a positive impact on consumption over the long term.
2.1.5 Involvement of the private sector
239
indirectly to several thousand man-years of research project work. This should greatly enhance the effectiveness of future coffee research expenditure, increase income from coffee exports, and improve productivity.
The creation in 1997 of the Coffee Industry and Trade Associations Forum (CITAF) established a consultative mechanism which allowed private sector concerns (such as statistics, legislation and the environment) to be addressed, albeit informally, at regular meetings during the Board and Council sessions. The work of the private sector was subsequently strengthened through two important new initiatives: a new Private Sector Consultative Board (PSCB) to advise the Council and Executive Board on ICA matters, established in July 1999 and composed of eight representatives of the private sector from exporting countries and eight from importing countries, and a regular World Coffee Conference to bring government and private sector representatives together to discuss matters of common concern to the world coffee industry. The first world coffee conference held by the ICO will take place in May 2001, and will be chaired by Mr Jorge CaÂrdenas GutieÂrrez, General Manager of the FederacioÂn Nacional de Cafeteros de Colombia.
2.1.8 Economic studies and publications
2.1.6 Statistics and information
In July 1999 the International Coffee Council, at its 78th (Special) Session, adopted Resolution number 384 providing for the extension for 2 years from 1 October 1999 of the International Coffee Agreement 1994. The Resolution also indicated that the Council would take measures as soon as possible to strengthen the involvement of the private coffee sector in the work of the Organization, to promote consumption of coffee, and to improve the Organization's system of statistics. In addition, a Negotiating Group would be established to draft the text of a new International Coffee Agreement by 30 September 2000, thus giving a full year for such a text to be ratified by Member countries. A new Negotiating Group, chaired by Mr Arnoldo LoÂpez Echandi, President, Instituto Costarricense del CafeÂ, Costa Rica, began discussions at the end of 1999 with a remit of drafting the text of a new International Coffee Agreement by 30 September 2000, which could enter into force on 1 October 2001.
A new Statistics Committee was established by the Council in 1999, open to all Members of the Organization, to representatives of the private sector and to experts in the area of coffee statistics. Its remit is to ensure that the Organization's statistical services continue to be strengthened and adapted to changing conditions. The 1994 Agreement also saw the launch of iCoffee (www.iCoffee.com), a joint ICO/Dow Jones initiative, the world's first major subscription Internet service totally dedicated to the global coffee industry, and the establishment of the Organization's Web site (www.ico.org), visited on average by 650 users throughout the world each day.
2.1.7 Global research network on coffee A global research network was implemented in 2000, following preliminary research and a feasibility study. ICO Members currently invest around 1500 man-years every year in coffee-related research projects. However, dissemination of research results is often slow and incomplete. The global research network should, within a few years, provide a database linking Members
Since 1994, studies have been undertaken on key economic and environmental aspects of the coffee market by members of the Organization's technical staff and published as documents. They include regular reviews of the market situation; four studies on price determination and volatility; a study on organic coffee; a study on the impact of `El NinÄo' on coffee production; and a study on sustainability and its relevance for the coffee sector. Other studies are currently in progress or planned on the economics of speciality coffees, forecasting models, coffee and biodiversity, and coffee, trade and the environment. In addition, 13 country coffee profiles had been published by the end of 1999, with nine more planned for 2000. They contain an upto-date and comprehensive account of the coffee sector in the countries concerned.
2.1.9 Towards a new Agreement in 2001
2.2 CONCLUSIONS At the end of the coffee year 1998/99 the Organization had 63 Member countries, 45 exporting and 18 importing Members. However, there is little doubt that the increasing involvement of the private sector in the
240
Coffee: Recent Developments
work of the Organization reflects the shift in emphasis of government involvement in world coffee trade from marketing to establishing regulatory frameworks related to protection of consumers and the environment, and to facilitate the development of market-oriented trading environments. The new PSCB had, by January 2000, already identified three priority areas: the need to disseminate positive news about coffee, the need to promote sustainable development and the need to
ensure that regulations to protect public safety are appropriate to the specific conditions of coffee. The ICO, as an international forum for all stakeholders in the world coffee economy, is uniquely placed to address these challenges. We have attempted in this narrative to tell the latest history of the Organization in words (also see References). For those interested in statistics, we have added two tables (Tables A.I and A.2, Section 2.3).
2.3 STATISTICAL INFORMATION Table A.1
Volume, value and unit value of exports to all destinations, calendar years 1964 to 1998.
Year
Exports to all destinations Volume (million bags)
1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998
46.2 43.2 49.4 50.7 54.1 54.9 52.8 53.8 58.1 62.8 55.0 58.6 60.0 48.2 57.3 64.3 60.2 60.5 64.5 66.3 68.6 71.4 64.5 72.0 65.8 75.9 80.6 75.8 78.2 75.0 70.5 67.6 77.5 79.8 79.1
Value in Unit value current terms in current (million terms (US US $) cents/lb) 2306 2114 2309 2191 2368 2404 3018 2688 3221 4294 4200 4254 8395 12524 11235 12411 11778 8087 9014 9243 10680 10831 14309 9589 9437 8683 6866 6501 5326 5689 10125 11611 9993 12871 11321
37.7 37.0 35.3 32.7 33.1 33.1 43.2 37.8 41.9 51.7 57.7 54.9 105.7 196.6 148.2 145.9 147.8 101.0 105.6 105.5 117.7 114.8 167.7 100.8 108.4 86.5 64.4 64.9 51.5 57.4 108.6 129.9 97.5 122.0 108.3
Values in constant 1990 terms1
Index 1964 = 100
Value of exports (million US $)
Unit value (US cents/ lb)
Volume
Value in constant terms
10480 9609 10037 9528 10295 10017 12074 9954 11106 12630 10244 9249 17861 24557 19371 18805 15916 11720 13256 14219 16953 17192 18827 11281 10258 9542 6866 6501 5171 5865 10227 10555 9428 13134 11917
171.4 168.0 153.5 142.2 143.8 138.0 172.9 140.0 144.5 152.0 140.8 119.3 225.0 385.4 255.5 221.0 199.7 146.4 155.3 162.2 186.8 182.1 220.6 118.5 117.8 95.0 64.4 64.9 50.0 59.1 109.7 118.1 92.0 124.5 113.9
100 94 107 110 117 119 114 116 126 136 119 127 130 104 124 139 130 131 140 143 148 154 140 156 142 164 174 164 169 162 152 146 168 173 171
100 92 96 91 98 96 115 95 106 121 98 88 170 234 185 179 152 112 126 136 162 164 180 108 98 91 66 62 49 56 98 101 90 125 114
Unit value UN in constant index terms (1990 = 100) 100 98 90 83 84 81 101 82 84 89 82 70 131 225 149 129 117 85 91 95 109 106 129 69 69 55 38 38 29 35 64 69 54 73 66
22 22 23 23 23 24 25 27 29 34 41 46 47 51 58 66 74 69 68 65 63 63 76 85 92 91 100 100 103 97 99 110 106 98 95
Source: ICO database. 1 Value in current terms deflated by the UN index of unit values of exports of manufactured goods from developed market economies.
Appendix 2
Table A.2
241
Prices of coffee, 1965 to 1999 (US cents/lb).
Year
Other mild arabicas
Robustas
Composite indicator price
Colombian mild (New York)
Brazilian naturals (New York)
1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999
45.08 42.12 39.20 39.33 39.78 52.01 44.99 50.33 62.30 65.84 65.41 142.75 234.67 162.82 173.53 154.20 128.23 140.05 132.05 144.64 146.05 194.69 113.62 137.60 108.25 89.46 84.98 64.04 70.76 150.04 151.15 122.21 189.06 135.23 103.90
31.07 33.53 33.52 33.86 33.11 41.44 42.27 45.19 49.88 58.68 61.05 127.62 223.76 147.48 165.47 147.15 102.61 109.94 123.90 137.75 120.14 147.16 101.99 94.31 75.09 53.60 48.62 42.66 52.50 118.87 125.68 81.92 78.75 82.67 67.53
40.37 39.61 37.22 37.36 38.71 50.52 44.66 50.41 62.16 67.95 71.73 141.96 229.21 155.15 169.50 150.67 115.42 125.00 127.98 141.19 133.10 170.93 107.81 115.96 91.67 71.53 66.80 53.35 61.63 134.45 138.42 102.07 133.91 108.95 85.72
48.00 47.35 41.61 42.42 44.44 56.66 49.01 56.70 72.52 77.81 81.31 157.72 240.21 185.20 183.41 178.82 145.33 148.60 141.61 147.33 155.87 220.04 123.45
43.58 40.56 37.72 37.36 40.90 55.80 44.71 52.52 69.20 73.34 82.57 149.48 308.04 165.29 178.47 208.79 179.55 143.68 142.75 149.65 151.76 231.19 106.37 121.84 98.76 82.97 72.91 56.49 66.58 143.24 145.95 119.77 166.80 121.81 88.84
REFERENCE International Standards Organization (1988) In Coffee Volume 6: Commercial and Technical-Legal Aspects (eds R.J. Clarke and R. Macrae), pp. 29±54. Elsevier Applied Science, London.
107.14 96.53 89.76 67.97 75.79 157.27 158.33 131.23 198.92 142.83 116.45
Appendix 3
Units and Numerals 3.1 UNITS 3.1.1 SI base units Quantity Length Mass Time Thermodynamic temperature Amount of substance Electric current Luminous intensity
Unit name metre kilogram second kelvin mole ampere candela
Unit symbol m kg s
Dimensions [L] [M] [T]
K mol A cd
[y] [N] [I] [Iv]
3.1.2 Some SI derived units used in engineering (a) Units with special names Quantity Frequency Energy, work, quantity of heat Force Pressure Power
Name
Unit symbol
hertz
Hz
Symbol expressed in base units s±1
joule newton pascal (newtons per square metre) watt (joules per second)
J N Pa
kg m2 s±2 kg m s±2 kg m±1 s±2
W
kg m2 s±3
(b) Examples without special names Physical quantity Density Heat capacity Heat transfer coefficient Thermal conductivity Velocity Viscosity (dynamic) (kinematic)
SI unit kilograms per cubic metre joules per kilogram per kelvin watts per square metre kelvin watts per metre kelvin metres per second
Unit symbol kg m±3 J kg±1 K±1 W m±2 K±1 W m±1 K±1 m s±1
pascal second square metres per second
Pa s m2 s±1
NB These unit symbols may alternatively be expressed using a solidus, e.g. W m±1 K±1 = W/m K.
242
Appendix 3
243
3.1.3 Some prefixes for SI units Multiplication factor 1012 109 106 103 102 10 10±1 10±2 10±3 10±6 10±9
Prefix tera giga mega kilo hecto deka deci centi milli micro nano
Symbol T G M k h da d c m m n
NB The use of prefixes hecto-, deka-, deci-, and centi- is not recommended except for SI unit multiples for area and volume. The litre is an acceptable derived SI unit (equalling 1 6 10±3 cubic metres, or 1 cubic decimetre, or 1 dm3), to which the above prefixes are commonly applied; thus, the millilitre (1/1000 or 10±3 litre), the centilitre (1/1000 or 10±2 litre) and the decilitre (1/10 or 10±1 litre). The millilitre is equivalent in older use to the cubic centimetre (cm3 or cc). In weight and mass measurement, these prefixes are used in relation to the gramme, e.g. the microgramme (mg) is 10±6 g. The common use of the millimetre, etc. in linear measurement is entirely consistent with basic SI units. The micron (often cited as m, but more correctly as mm) is 1 6 10±6 m. The aÊngstroÈm is equal to 10±10 m; it is preferable to replace this unit with the nanometre (i.e. 1 AÊ = 0.1 nm).
3.1.4 Some conversions of SI and non-SI units Type of unit Linear Area Volume
Mass Density
Flow rate
from inch (in) foot (ft) miles square foot (ft2) acre cubic foot (ft3) litre gallon (US) gallon (British) pound (lb) gallons (British) ounce (oz) pound per cubic foot
cubic foot per minute (cfm) gallon (British or Imperial) per hour (igph)
To convert
to metre (m) metre (m) kilometres square metre (m2) hectare cubic metre (m3) cubic metre gallon (Imperial or British) cubic metre kilogram (kg) per mile litres per kilometre kilogram kilograms per cubic metre (or grams per litre) grams per cubic centimetre (or per millilitre) cubic metre per second cubic metre per second
Multiply by 0.025 4 0.304 80 1.6211 0.092 90 0.404 69 0.028 32 10±3 0.833 4.546 6 10±3 0.453 6 2.80 28.35 6 10±3 16.02 0.01602 4.72 6 10±4 1.263 6 10±6
244
Type of unit Heat and power
Coffee: Recent Developments
from Btu (BThU) calories (cal) thermochemical international table watt (W) horsepower (hp) (550 ft lbf per second Btu per pound
Heat transfer coefficient
Btu per square foot per hour per 8F calorie per square metre per second per 8C
Pressure
Pound-force per square inch (psig, gauge; psig, absolute pressure) psig inHg (328F) mmHg (08C) mmHg absolute bar
Thermal conductivity
Viscosity (dynamic)
atmosphere (= 760 mmHg or &30 inHg &14.7 psi abs. = 0.00 psig Btu per hour square foot per 8F/in Btu per hour per square foot per 8F/ft calorie per second per square centimetre per 8C/cm (= calorie per second per centimetre per 8C)
}
gram per centimetre per second (or dyne second per square centimetre) poise centipoise gram per centimetre per second centipoise
To convert calories joule (J)
to
joule per second watt (W) per kilogram {calorie kilocalorie per kilogram
calorie per square metre per second per 8C watt per square metre per kelvin kilogram per square centimetre pascal (Pa) kilopascal (kPa) megapascal (MPa) bar pascal pascal torr pascal megapascal pascal megapascal
{
calorie per second per square centimetre per 8C/cm watt per square centimetre per kelvin/cm (= watt per centimetre per kelvin) watt per metre per kelvin poise (P)
Multiply by 252 4.184 4.187 equivalent 7.457 6 102 555 0.555 1.356 4.187 0.070 31 6.895 6.895 0.006 895 0.06895 3.386 6 103 1.333 6 102 equivalent 1.0 6 105 0.1 1.013 6 105 0.101 3
}
3.447 6 10±4 4.13 6 10±3 4.187
4.187 6 102 equivalent
centipoise (cP) pound (mass) per foot per hour pascal second
102 2.42
millipascal second (m Pa s)
equivalent
0.1
*
Appendix 3
Type of unit Viscosity (kinematic) Diffusivity (diffusion coefficient) Force Mass transfer flux
245
To convert
from square centimetre per second stokes square centimetre per second
stokes
dyne pound-force gram per square centimetre per second
Multiply by
to
equivalent
square metre per second square metre per second
1.0 6 10±4 1 6 10±4
newton (N) newton (N) gram per square metre per second
1.0 6 10±5 4.448 1 6 104
References Kirk-Othmer, Concise Encyclopaedia of Chemical Technology, John Wiley, New York, 1985.
3.2 NUMERALS – CARDINAL Greeka
Latinb
eis, mia, en duo treis, tria tessares, tessara pente hex hepta
unus, una, unum; I duo, duae, duo; II tres, tria; III quattuor; IV quinque; V sex; VI septem; VII
okte ennea deka eicosi pentekonta hekaton
octo; VIII novem; IX decem; X viginti, xx quinquaginta; L centum; C
pentakosioi kilioi
quingenti; D mille; M
Direct English, and Prefixc,d one [uni-, L.] two [duo-, L. and Gk] three [tri, L. and Gk] four [quadri-, L.] five [quinqu(e)(i)-, L.; pent(a), Gk] six [hex(a)-, Gk; sex(i)-, L.] seven [hepta-, Gk; sept(em)(i)-, L.] eight [oct(a)(o)-, L. and Gk] nine [nan-, Gk] ten [dec(a)-, Gk; deci- (tenth), L.] twenty [eicosa-Ck] fifty [Ð] hundred [hecto-, Gk; centi(hundredth), L.] five hundred thousand [kilo-, Gk; milli(thousandth), L.]
a Greek nouns in nominative singular form, together with the indefinite form, where this is different, Latinised characters used, with e (long = Z; e (short) = e; o (short) = o; o (long) = o; kh = w; ph = f. Greek u = u (upsilon) is the Latin y. b Latin nouns given in nominative singular form with their m., f. or n. genders; where applicable. c Greek prefixes are only usable in combination with Greek-based words, and Latin prefixes with Latin words. d In general, Greek prefixes are used as multiples (e.g. kilo-) and Latin prefixes for fractions (e.g. milli-).
Reproduced with permission from Clarke & Macrae (1987, 1988) by Kluwer Academic Publishers.
Index Note: page references in italic are to figures and tables. abortion, spontaneous 176±7 acetic acid brews 154 caffeine recovery processes 120 formation mechanisms 23 green coffee 19 roast coffee 19±20, 21, 22 storage effects 26, 56 acidity 27±30, 55±6, 154, 158 acids bean swelling 94±5 brews 55±6, 154 caffeine recovery processes 120 effects of roasting 22±5, 29±30, 54±6, 59, 158 formation mechanisms 13, 23±6 green coffee 18±19, 23, 25, 26, 54±5, 59 roast coffee 19±23 sensory characteristics 27±30, 55, 147, 154, 158 solvent decaffeination 109 storage effects 26, 55±6 suspension-cultured cells 208±9 volatile 26±7, 28 activated carbon caffeine recovery 119±21 chlorinated water 156 activity coefficients, volatile compounds 135 adsorption processes, decaffeination 111, 112±13 adulterants see contaminants aflatoxins 168, 231 after-taste, brews 158±9 agronomy see breeding practices; cell biology; molecular biology air±water partition coefficients 133±4, 135 alcoholic cirrhosis 178 aldehydes aroma analysis methodology 70, 71 brews 81, 82 green coffee 73, 74 odorant formation 82, 83±4, 85 physical properties 133±4, 135, 136, 137 roast coffee 71, 76, 77, 80 storage effects 79 alkaloid formation 203, 205, 207±9 alkanes 44
alkylpyrazines aroma analysis methodology 70, 71, 72 brews 80, 81, 82 formation 82, 84 green coffee 73±4, 75 optimum roast detection 93 physical properties 134±5, 136, 137 roast coffee 75, 76, 77, 78, 79, 80 storage effects 79 amino acids 13±14, 50±1 11S seed storage proteins 225±6 carbohydrate breakdown products 13±14, 52 Maillard products 13, 14 odorant formation 82, 83, 84, 85 protein-bound 50, 51, 52±3 ripening coffee fruits 228 Strecker degradation 82, 83, 84 1-amino-cyclopropane-1-carboxylic acid (ACC) oxidase 228 1-amino-cyclopropane-1-carboxylic acid (ACC) synthase 228 aminohexose reductones 53±4 antibiotic resistance genes 232 antimutagenic effects, coffee 167±8 antioxidative compounds 57±8, 168 arabinogalactan coffee fibre 15 extract viscosity 14 green coffee 3±5 odorant formation 85 roast coffee 7, 9, 14 soluble coffee 13, 14 spent coffee grounds 132 arabinose green coffee 2, 5 roast coffee 7, 8, 9, 14 soluble coffee 8, 9, 12 aroma, brews 77±8, 80±2, 126, 159 aroma-active compounds see volatile compounds aroma analysis methodology 69±73 aroma extract dilution analysis (AEDA) 69, 70, 72, 73 aroma formation proteins 51, 53 roasting processes 79, 92 see also volatile compounds
246
Index
aroma retention, instant coffee 82, 126, 129, 131 aromatization, instant coffee 132 arrhythmias 170 artificial-flavour-reinforced coffees 141, 161±2 astringency, brews 160 atractylglycosides 37 green coffee 2±3, 4 roast coffee 6±7 bean defects, standards 236±7 bean development 205±7 bean mass 94, 96±100, 101 bean moisture international standards 235±6 and roasting 94±5, 96±7, 101 bean porosity 95, 99, 101 bean swelling mechanisms 92, 94±5 bean temperatures, roasting 92±5, 96, 97, 98±9, 101 bean volume 95, 101 beverage characterization acids 154 carbohydrates 154±5 chemical 153±7 density 152 dispersed phases 152 foam 151±2 lipids 154 minerals 155±6 nitrogenous compounds 155 organoleptic 157±60 physical 151±3 refractive index 153 surface tension 153 total solids 153±4 viscosity 152±3 water 151, 156±7 beverage preparation 140±3 decoction 143±4 and health 143, 171±2 infusion 144±5 pressure 145±51 bioreactor cultures 214±16 biotechnology 185, 224 11S storage protein promoter 228 caffeine biosynthesis 228±30, 233 coffee transformation systems 230±1, 232±3 disease resistance 231 environmental stresses 231±2 flavour traits 232 fruit development 228 fungal toxins 231 pest resistance 231
247
prospects 231±3 quality traits 232 selectable markers 232 site-specific changes 232±3 birth weight, and maternal caffeine 175±6 bispyrrolidinohexose reductones 53±4 bitterness brews 158±9 protein 53±4 black rot 195 bladder cancer 166±7 blood pressure 170±1 body, brews 160 boiled coffee 143, 154, 155, 171±2 bone health 173±4 bottled coffee drinks 161 Bourbon LC variety 215±16, 219 breast cancer 166 breeding practices 184 before 1985 184±5 biotechnology 185 clonal propagation 196±7 for disease resistance 192±5 for drought tolerance 196 genetic resources 186±9 germplasm conservation 188±9 for insect resistance 195±6 methods 189, 190 for nematode resistance 195 new cultivar propagation 196±7 new developments 185±6 objectives 189 for productivity 189, 191 for quality 191±2 seed propagation 196, 212±13 species relationships 186±8 world germplasm collections 186 world production levels 184 see also cell biology; molecular biology brewing yield 142 brews acidity 55±6, 154 aromas 77±8, 80±2, 126, 159 in-home decaffeination 119 international standards 236±7 modified 141, 160±2 preparation methods 140±51, 171±2 see also beverage characterization bubbling beds 92 2,3-butanedione 70, 71, 76, 79 brews 80, 81 odorant formation 82±3, 84
248
cafestal 40, 41 cafestol 37, 38, 39±40, 41 and health 168, 172, 178, 179 cafetieÂre (plunger) coffee 145±6, 154, 155, 172 caffeic acid antioxidation 57±8, 168 decaffeination 109, 110, 112 zinc-chelating compounds 64±5 caffeine 108 antioxidative effects 57, 168 applications 123 biosynthesis 203±4, 206±9, 228±30 and bitterness 53 and blood pressure 170±1 and bone health 173±4 brews 154, 155 callus cultures 207±8 demand for 122±3 international standards 236 and leaf development 203±4, 205 molecular biology 228±30 and neuroactivity 177±8 and pregnancy 174±7 recaffeinated coffee 119 recovery processes 119±21 removal see decaffeination processes and roasting speed 92 safe consumption levels 179 caffeine-free coffee 230, 232, 233 caffeoyltryptophan 51, 57 calcium, bone health 173±4 callus cultures 207±8 caloric content, brews 155, 160 cancers 15, 166±9, 178 canned coffee drinks 141, 161 cappucino 150, 160±1 caramelization 13 heat caramelization equipment 106 caramel odours brews 81 roast coffee 75, 76, 77, 78, 79, 80 carbohydrates 1 brews 154±5 colour development 51±2 extract viscosity 14±15 functional properties 14±15 green coffee 1±6 high molecular weight green coffee 1, 3±6 roast coffee 7±8 soluble coffee 12±13 see also arabinogalactan; mannan
Index
low molecular weight green coffee 1±3 roast coffee 4, 6±7, 8, 13 soluble coffee 8±12 odorant formation 82±5 reactions on roasting 6±7, 9, 13±14, 23, 51±2 roast coffee 4, 6±8, 13±14 sedimentation 15 soluble coffee 8±13, 14±15, 127 spent coffee grounds 132 carbon, activated caffeine recovery 119±21 chlorinated water 156 carbon dioxide decaffeination 108±9, 113±18, 119, 122 carbon fibres, decaffeination 111 carbonic acid 5-hydroxytryptamides (C-5-HT) 45±6 carbonyl compounds, aromas 79, 82±3 carboxyatractylglycosides 2±3, 4 carcinogenicity, coffee 41, 168 cardiac arrhythmias 170 cardiovascular disease 169±72 cavities, roast coffee 95, 100 cell biology 202 alkaloid formation 203, 205, 207±9 embryo cryopreservation 211 gene regeneration 211, 212, 219 gene transfer 211±12, 219 in vitro selection studies 212 micropropagation 212±17, 219 organ development 202±7 protoplast culture 209, 211, 219 somaclonal variants 217±19 somatic embryogenesis 209±11, 213±17, 219 cellulose green coffee 3, 6 soluble coffee 13 CHARM analysis 69, 70, 72 chemoprotection, from coffee 178 chimeric oligonucleotides 232±3 chlorination, water 156 chlorogenic acids (CGA) antioxidative compounds 57±8 bean swelling 94±5 brews 154 coffee acidity 28±9, 55±6, 158 decaffeination 109, 112 effects of roasting 54±6, 59, 158 green coffee content 18 and health 168 and leaf development 205 as quinic acid precursor 23, 25 roast coffee content 19±20, 22, 24
Index
suspension-cultured cells 208±9 cholesterol 42, 43 boiled coffee 143, 171±2 coffee consumption levels 179 serum levels 41, 171±2 cigarette smoking and fertility 177 in pregnancy 176 cinnamic acid 56, 57 cirrhosis 178 citric acid caffeine recovery processes 120 coffee acidity 29, 56 formation mechanisms 25 green coffee content 18, 19 roast coffee content 19±20, 21, 22, 24 storage effects 26, 56 clonal propagation 196±7, 213±17, 219 coffeadiol 44 coffee berry borer 195, 224, 231 coffee berry disease (CBD) 185, 193±4, 212, 231 Coffee Industry and Trade Associations Forum (CITAF) 239 coffee leaf rust (CLR) 184, 185, 192±3, 231 coffee leaf scorch 195 coffee machines 141, 142±4, 145±6, 150±1, 156 coffee oil 33±4 see also lipids coffee white stem borer 195 Colletotrichum coffeanum (C. kahawae) 193, 212, 231 colloids, and after-taste 158±9 colorectal cancer 15, 167, 178 colour development instant coffee 129 protein reactivity 51±2 coloured macromolecular compounds polymers 58±62 melanoidin 13±14, 53, 58±9, 60, 62 zinc-chelating 62±5 Cona coffee 144 conductivity, thermal, beans 97, 100, 101 congenital malformations 174±5 contaminants causing off-flavour 75 coffee extraction process 10, 11±12 international standards 236 cooling gas roasters 104 cooling methods, roasters 104, 106 coronary heart disease 169±70 CROSSPY 14, 52 cryopreservation, germplasm 189, 211 cyclic peptides 53, 54
249
cyclohexanoic acid ethylester 74, 137 L-cysteine 53±4 û-damascenone aroma analysis methodology 70, 71 brews 80, 81 green coffee 73, 74, 75 physical properties 136 roast coffee 71, 75, 76, 77, 78, 80 decaffeinated coffee 108 bean behaviour 95±6 economics 122±3 production see decaffeination processes decaffeination processes 108±9 caffeine recovery 119±21 economics 122±3 fatty materials 118, 122 in-home 119 liquid CO2 118, 122 solvent 108, 109±10, 122 supercritical CO2 108±9, 113±17, 122 water 108, 110±13, 120±1, 122 decoction, brew preparation 143±4 degustation, brews 158±9 dehydrocafestol 39±40, 41 dehydrokahweol 39, 41 demineralization and bone health 173 water 157 density, brews 152 dephosphorylation 6 desorption processes, decaffeination 111±12, 113, 120, 121 developmental biology (plant) 202±7 developmental outcomes (human) 174±7 development projects 238 dichloromethane (DCM) 109, 110, 122 diketopiperazines 53 dimethyl sulphoxide 109 1,1,diphenyl-2-picryl hydazil (DPPH) 57 disaccharides see carbohydrates, low molecular weight disease resistance breeding for 192±5 molecular biology 224, 231, 232 dispersed phases, brews 152 diterpenes 33, 34, 36±41 boiled coffee 143 and health 41, 168, 172, 178, 179 see also atractylglycosides drip (filter) coffee 144±5, 154, 171, 172 drought tolerance, breeding for 196
250
Index
drum roasters 101, 104±5, 105
extracts, liquid coffee 132±3
earthy odours brews 81 roast coffee 75, 76, 77, 78, 79, 80 economic studies 239 embryo cryopreservation 189, 211 embryogenesis, somatic 209±11, 213±17, 219 environmental stresses biotechnology 231±2 drought tolerance 196 enzymes caffeine synthesis 204 chlorogenic acid synthesis 205 coffee extraction processes 14±15, 128 detoxifying 168 Erlenmeyer cultures 213±14 espresso coffee 146 acid content determination 23, 24 and cappucino 150, 160±1 chemical characteristics 154, 155 definitions 146, 148, 149 foam 15, 151±2, 157±8 and health 41, 172 as lifestyle 146±7 machines 150±1, 156 physical characteristics 151, 152, 153 pressure in preparation 147±9 quantitative definition 149 rapidity of extraction 149 sensory characteristics 147, 157±9, 160 water hardness 151, 156, 157 esters decaffeination processes 118 diterpene fatty acid 33, 34, 38±9, 40, 41 sterol 34, 41±2 ethyl 2-methylbutyrate 74, 75 ethyl 3-methylbutyrate 74, 75 ethyl acetate decaffeination 109, 122 physical properties 133±4, 135 ethylene biosynthesis, fruit ripening 228 evaporation, instant coffee processing 129±30 export values 240 export volumes 240 extraction methods, brew preparation 143±51 extraction processes carbohydrates 9, 10, 13, 14±15 chlorogenic acids 55 coffee fibre 15 instant coffee 127±8 spent coffee grounds 132
F1 arabica hybrids, clonal propagation 216±17 fast roasting processes 92, 95, 98, 104±6 fats see lipids fatty acids 34±6, 38±9, 40, 41, 45 decaffeination processes 118 fertility, and caffeine 177 fibre 15 Fick's law of diffusion 142 filter coffee 144±5, 154, 155, 158, 171, 172 flavour, brews 159, 236±7 flavoured coffee drinks 141, 161±2 fluidized bed roasting 91±2, 93 bean behaviour 93±4 bean volume 95 equipment 101±3, 105 foam 151±2 stability 15 visual importance 157±8 foreign matter see contaminants formic acid brews 154 caffeine recovery processes 120 coffee acidity 29, 56 formation mechanisms 23 green coffee content 19 roast coffee content 19±20, 21, 22 storage effects 26, 56 free fatty acids (FFA) 35±6 free radicals 14 freeze concentration 128 freeze drying 131±2 freezing weather 231±2 French press (plunger; cafetieÂre) coffee 145±6, 154, 155, 172 friable embryogenic tissue (FET) cultures 214±15, 216 fructose green coffee 2 roast coffee 6, 13 soluble coffee 9, 10, 12 fruit development 205±7, 228 fruity odours brews 81 roast coffee 75, 76, 80 fungal toxins 168, 231 fungi coffee berry disease 193, 212, 231 coloured polymer characterization 59±60, 61 resistance genes 224, 231 fungicides coffee berry disease 193±4
Index
contamination with 75 furanones aroma analysis methodology 71, 72 brews 80, 81 formation 82, 83±4 green coffee 75 physical properties 136 roast coffee 75, 76, 77, 78, 79, 80 furfural arabinogalactan scission 14 odorant formation 85 fusarium wilt disease 195, 231 Fusarium xylarioides 195, 231 galactomannan foam stability 15 roast coffee 7 use of term 6 see also mannans galactose green coffee content 3±4, 5, 6 roast coffee content 7, 8, 9, 14 soluble coffee content 8, 9, 12±13 a-galactosidase 225 gas chromatography-olfactometry (GCO) 69±72, 73, 74 gas circulation, roasters 104 gastric cancers 167 gastrointestinal tract cancers 15, 167, 178 reactions to wax 45 gene regeneration 211, 212, 219 gene sequences 225±30 genetic resources, for breeding 186±9 gene transfers 211±12, 219 gene transformation systems 230±1, 232±3 geosmin 75 germplasm collections 186, 219 germplasm conservation 188±9, 211, 219 global research network 239 glucose green coffee 2, 5 roast coffee 6, 13, 14 soluble coffee 9, 12 glycolic acid coffee acidity 29, 56 formation mechanisms 23 roast coffee content 20, 21, 22, 24 storage effects 26, 56 glycosides 37 green coffee 2±3, 4 roast coffee 6±7
251
Greek coffee 143 green coffee amino acids 50, 51 antioxidative compounds 57±8 and brew astringency 160 brewing 140±1 carbohydrates 1±6 contaminants 75 content of odorants 73±5 decaffeination 109, 111±17, 118 heat capacity 96±7, 98 for instant coffee 132, 133, 137 international standards 235, 236 lipids 33, 34, 35±6, 37, 43±4, 45±6 OAVs of odorants 73±5 organic acids effects of roasting 54±5, 59 formation mechanisms 23, 25, 26 quantitative data 18±19 physical values 91, 100±1 potent odorants 73 protein 51 volatile compounds 73±5, 137 zinc-chelating compound 64 grinding processes, instant coffee 127 ground coffee, standards 235, 236 guaiacols aroma analysis methodology 70, 71 brews 81, 82 green coffee 74, 75 physical properties 136 roast coffee 75, 76, 77, 78, 79, 80 health considerations 165 antioxidative compounds 57, 58, 168 benefits of coffee 177±8 birth defects 174±6 boiled coffee 143, 171±2 bones 173±4 cancers 15, 166±9, 178 cardiovascular disease 169±72 chemoprotection 178 coffee fibre 15 coffee wax 45 diterpenes 41, 168, 172, 178, 179 neuroactivity 177±8 neurodevelopment 175 preserved coffee drinks 141, 161 solvent decaffeination 109 zinc-chelating compounds 62 heat capacity, beans 96±7, 98 heat caramelization equipment 106
252
heat conductivity, beans 97, 100, 101 heat sterilization, and aroma 82 heat transfer, gas±bean 99±100 heat transfer coefficients 99±100 heat uptake, beans 97±8 Hemileia vastatrix 192, 231 hepatitis infections 178 4-heptenal 74 herbicide resistance genes 232 high performance anion-exchange chromatography (HPAEC) 10±11, 12 home decaffeination 119 homocysteine, serum levels 172 homologous recombination, genetic changes 233 husk adulteration 10, 12 hydrogen peroxide, formation 57±8 hydrolysis coffee extraction process 127 carbohydrates 9, 10, 13, 14±15, 127 chlorogenic acids 55 fibre 15 spent coffee grounds 132 hydrolyzing enzymes 14±15, 128 hypercholesterolaemia 171±2 hypertension 171 Hypothenemus hampei 195, 224, 231 ibrik 143 industrial roasting equipment 101±7 infusion, brew preparation 144±5 in-home decaffeination 119 inositol phosphates (IPs) coffee acidity 29, 154 green coffee content 2 phosphoric acid formation 25±6 roast coffee content 6 insect damage international standards 236 odour activity values 74 insect resistance 195±6, 224, 231 instant coffee 125±6 agglomeration 130±1 aroma of brews 82, 126 aromatization 132 brew preparation 140 carbohydrates 8±15, 127, 132 extraction processes 127±8 freeze concentration 128 freeze drying 131±2 green coffee for 132, 137 and health 41, 172 international standards 126, 235, 236
Index
legislation 126 liquid extracts 132±3 Loncin's role 125±6 organic acid determination 23, 24 reverse osmosis 130 sales 125 spent grounds disposal 132 spray drying 130±1 technology 125 thermal concentration 129±30 Thijssen's legacy 125 volatile compounds extraction 126±7 freeze drying 131±2 handling 130 physical properties 133±7 recovery 129±30 zinc-chelating compounds 62±4 integrated pest management (IPM) 195 International Coffee Agreement (1994) 238±9 International Coffee Organization (ICO) 238±41 International Standards Organization (ISO) 11, 126, 235±7 in vitro selection studies 212 ion exchange chromatography 1 iron-chelating compounds 62, 63 isotope dilution assays 72 Israeli `mud' coffee 143 kahweal 40 kahweol 37, 38, 39, 40, 41 and health 168, 172, 178, 179 Koleroga noxia 195 lactic acid brews 154 coffee acidity 29 formation mechanisms 23 roast coffee content 19, 21, 22, 24 storage effects 26 lactones, chlorogenic acid 55±6 Laurina somaclones 218±19 `leaching', beverage preparation 141 leaf development 203±5 leaf miners 195±6 Leucoptera spp. 195±6 lipids 33 alkanes 44 boiled coffee 143, 171±2 brews 154 brew viscosity 152±3 coffeadiol 44
Index
decaffeinated coffee 95±6 decaffeination processes 118, 122 determination methods 33±4 diterpenes 34, 36±41, 143, 168, 172, 178, 179 fatty acids 34±6, 38±9, 40, 41, 45 and health 41, 45, 168, 172, 178, 179 isolation for analysis 34 squalene 44 sterols 34, 41±2, 143, 171±2, 179 tocopherols 42±4 wax 33, 45±6 liquid CO2 decaffeination 118, 122 liquid coffee extracts 132±3 liver disease 178 Loncin, M. 125±6 low birth weight infants 175±6 lysine 51±2 Maillard reaction 13, 14, 58±9, 62, 64 malic acid brews 154 coffee acidity 29, 56 formation mechanisms 25 green coffee content 18, 19 roast coffee content 19±20, 21, 22, 24 storage effects 26, 56 mannans green coffee 3, 5±6 roast coffee 7±8, 9 soluble coffee 12±13, 14±15, 127±8 spent coffee grounds 132 mannitol roast coffee 6 soluble coffee 10, 12, 13 spent coffee grounds 132 mannose green coffee 2, 5 roast coffee 7, 8, 9, 14 soluble coffee 8, 9, 12±13 spent coffee grounds 132 mass transfer, brew preparation 142 mass transport, during roasting 96±100, 101 measurement units 242±5 melanoidin antioxidant activity 168 brews 155 characterization 58±9, 60, 62, 64 formation 13±14, 53, 58 Meloidogyne spp. 195 metal-chelating compounds 62±5 2-/3-methylbutanoic acid 27 2-/3-methylbutyric acid 27, 74
253
16-O-methylcafestol (16-OMC) 37±8, 40, 41 2-methylisoborneol (MIB) 75, 78, 137 16-O-methylkahweol 37 methylthio groups 52±3 microbiological characterization, coloured polymers 59±62 micropores, cell walls 95 microporous resins, decaffeination 111 micropropagation 212±17, 219 milk±coffee admixtures 160±1 mineral-chelating compounds 62±5 minerals bone health 173±4 brews 154, 155±6 miscarriage (spontaneous abortion) 176±7 modified coffee beverages 141, 160±2 moisture content international standards 235±6 roasting beans 94±5, 96±7, 101 Moka coffee 146, 154, 155 molecular biology 185, 224 11S storage protein promoter 226±8 caffeine biosynthesis 228±30, 233 coffee genes 225±30 coffee transformation systems 230±1, 232±3 disease resistance 231 environmental stresses 231±2 flavour traits 232 fruit development 228 fungal toxins 231 pest resistance 231 prospects 231±3 quality traits 232 selectable markers 232 site-specific changes 232±3 monosaccharides see carbohydrates, low molecular weight mouldy flavours 75, 78 mouthfeel, brews 159±60 multidimensional gas chromatography (MDGC) 72 mutagenic effects, coffee 167±8 myocardial infarction 169±70 Napoletana coffee 145, 154, 155 nematode resistance 195 Nepro Vortex Fluidat 103, 105 NestleÂ, clonal propagation 216 neuroactivity, and caffeine 177±8 neurodevelopment, and caffeine 175 nitrogenous compounds, brews 155 see also caffeine; melanoidin nitrous oxide, decaffeination 114
254
(E)-2-nonenal aroma analysis methodology 70 brews 82 green coffee 74, 75 roast coffee 75 non-volatile compounds acids 18±30, 54±6, 59 amino acids 13±14, 50±1, 52±3 antioxidative 57±8 carbohydrates 1±15 coloured macromolecular 58±65 lipids 33±46 protein 51±4 Nusselt equations 99 ochratoxin 231 odorants see volatile compounds odour, brews 77±8, 80±2, 126, 159 odour activity values (OAVs) 68±9 green coffee 73±5 oesophageal cancers 167 oil, coffee 33±4 see also lipids olfaction 159, 236 oligonucleotides, chimeric 232±3 oral cancers 167 organic acids bean swelling 94±5 brews 55±6, 154 caffeine recovery processes 120 effects of roasting 22±5, 29±30, 54±6, 59, 158 formation mechanisms 13, 23±5 green coffee 18±19, 23, 25, 26, 54±5, 59 roast coffee 19±23 roast kinetics 29±30 sensory characteristics 27±30, 55, 147, 154, 158 solvent decaffeination 109 and storage 26, 55±6 suspension-cultured cells 208±9 volatile 26±7, 28 organic solvents, decaffeination 109 organoleptic characteristics see sensory characteristics osmosis, reverse 130 osteoporosis 173±4 ovarian cancer 166 2-oxopropanal 82, 84 packed bed roasters 104, 105 pancreatic cancer 167 partition coefficients 133±4, 135 pectin, green coffee 6 percolation 144
Index
espresso coffee 148, 149, 157 and water hardness 157 percolator coffee 143±4, 154, 155 Perileucoptera coffeella 195 periodic immersion cultures 215 pest resistance 195±6, 224, 231 pharyngeal cancers 167 phases, brews 152 phenolic compounds antioxidant activity 168 melanoidin characterization 58, 59, 60, 64 metal binding 62, 64±5 phenolic odours 75, 76, 78, 80 brews 81 phenols formation 84 physical properties 136 roast coffee 78 see also guaiacols phenylindans 58, 168 phosphoric acid coffee acidity 29, 56, 154 formation mechanisms 25±6 green coffee content 18 roast coffee content 20, 21, 22 and storage 56 phosphorylation 6 phytic acid (IP6) coffee acidity 29, 154 phosphoric acid formation 25±6 plunger (cafetieÂre) coffee 145±6, 154, 155, 172 pollen transformation system 232 polyalcohols 2±3 see also atractylglycosides polysaccharides see carbohydrates, high molecular weight porosity, bean 95, 99, 101 potassium, brews 156 Pratylenchus spp. 195 predrying processes 92, 95 pregnancy 174±7 premature births 175±6 preserved coffee drinks 141, 161 pressure methods, brew preparation 145±51 see also espresso coffee prices, coffee 241 promotional activities 238±9 propagation new cultivars 196±7 using micropropagation 212±17, 219 prostate cancer 166 protein 51
Index
bitter tasting compounds 53±4 brews 154 carbohydrate breakdown products 13±14 Maillard products 13, 14 reactivity 51±3 seed storage genes 225±6 zinc-chelating compounds 64 protoplast culture 209, 211, 219 Pseudomonas syringae pv garcae 195 psicose 2, 6 purine alkaloid formation 207±9 pyrazines aroma analysis methodology 70, 71, 72 brews 80, 81, 82 formation 82, 84 green coffee 73±4, 75 optimum roast detection 93 physical properties 134±5, 136, 137 roast coffee 75, 76, 77, 78, 79, 80 storage effects 79 pyrrolidinohexose reductones 53±4 pyruvic acid, roast coffee content 19 quinic acid antioxidative compounds 57±8 brews 154 coffee acidity 29, 55, 56 effects of roasting 54±6, 59 formation mechanisms 23, 25 green coffee 18, 19 roast coffee 19, 20, 21, 22 storage effects 26, 56 quinides 54±5 radiotherapy 178 raw coffee see green coffee recaffeinated coffee 119 refractive index, brews 153 renal cancer 166 reproduction, coffee effects 174±7 research network 239 resins, decaffeination 111, 112±13 reverse osmosis 130 rhamnose green coffee 2, 6 roast coffee 14 soluble coffee 12 roast coffee antioxidative compounds 57±8 arabica aroma profiles 77±8 aroma changes in storage 79±80 brewing techniques 141±51
255
carbohydrates 4, 6±8, 13 cavities 95, 100 degree of roast 79 international standards 235, 236 lipids 33, 34 diterpenes 37, 39±40 fatty acids 34±5, 36, 40, 41 tocopherols 43 wax 46 liquid CO2 decaffeination 118 odorant concentration 75±7 odorant evaluation 77 organic acids 19±23 physical values 91, 100±1 robusta aroma profiles 77±8 supercritical CO2 decaffeination 117 volatile compounds 75±80, 136 zinc-chelating compounds 62±3 see also roasting processes roasters, industrial 101±7 roasting processes 90 amino acids 50, 51, 52, 53 antioxidative compounds 58 and aroma 79, 92 bean behaviour 93±6 beans' physical values 100±1 and bitter taste 53±4 cafestol/dehydrocafestol ratio 39±40 carbohydrate reactions 6±7, 9, 13±14, 23, 51±2 chlorogenic acids 54±6, 59, 158 colour development 51±2 conventional 91, 93±4 detecting optimum roast 79, 92±3 fast 92, 95, 98, 104±6 fluidized bed 91±2, 93±4, 95, 101±3, 105 heat transport in bean 96±100, 101 industrial equipment 101±7 for instant coffee 127 mass transport in bean 96±100, 101 organic acids 22±5, 29±30, 54±6, 59, 158 protein 51, 52, 53±4 zinc-chelating compounds 64 roasty odours brews 81 roast coffee 75, 76, 77, 78, 79, 80 root-knot nematodes 195 root-lesion nematodes 195 rotating bowl roasters 101, 105 rotating fluidized bed (RFB) roasters 101±3 safety, coffee consumption 165, 178±9 Secoffex decaffeination 111, 112
256
sedimentation, carbohydrate role 15 seed development 206 seed propagation 196, 212±13 seed storage protein gene 225±8 sensory characteristics 157±60 acids 27±30, 55, 147, 154, 158 espresso coffee 147, 148, 158 international standards 236 serum cholesterol 41, 171±2, 179 serum homocysteine 172 sexual organs, cancers of 166 silicas, decaffeination 112 SI units (appendix) 242±5 smoking and fertility 177 in pregnancy 176 smoky odours brews 81 roast coffee 75, 76, 77, 78, 79, 80 solid±liquid extraction, brewing 140, 141±51 solids, brew characteristics 153±4 soluble coffee see instant coffee solvent decaffeination 108, 109±10, 122 somaclonal variation 217±19 somatic embryogenesis 209±11, 213±17, 219 sourness 27±8 specific heat, coffee 96±7 spent coffee grounds, disposal 132 spicy odours brews 81 roast coffee 75, 76, 80 spinning cone columns (SCCs) 129±30 spontaneous abortion 176±7 spouting beds fast roasting 92 fluidized bed roasting 91±2 spray drying, instant coffee 130±1 squalene 44 stable isotope dilution assays 72 starch green coffee 6 soluble coffee 12 statistical services 239 steaming, instant coffee processing 130 steam roasting 106±7 sterols 33, 34, 41±2, 143, 171±2, 179 stomach cancers 167 storage of coffee 26, 79±80 Strecker degradation 82, 83, 84 succinic acid green coffee 19 roast coffee 20, 21, 22, 24
Index
storage effects 26 sucrose brews 155 green coffee 1±2 roast coffee 6, 13, 14, 23 sudden infant death syndrome 175 sulphur-containing amino acids 52±3 sulphurous odours brews 81 roast coffee 75, 76, 77, 79, 80 supercritical carbon dioxide decaffeination 108±9, 113±17, 119, 122 supercritical nitrous oxide decaffeination 114 supernatant foam 151±2 surface tension, brews 153 suspension-cultured coffee cells 208±9 sweetness, brews 158 sweet odours brews 81 roast coffee 75, 76, 77, 78, 79, 80 swelling mechanisms, beans 92, 94±5 taste acidity 27±8, 154, 158 bitterness 53±4, 158±9 brews 158±9 olfactory component 159 temperatures, bean roasting 92±5, 96, 97, 98±9, 101 temperature stresses, coffee plants 231±2 teratogenicity, caffeine 174±5 thermal concentration, instant coffee 129±30 thermal conductivity, beans 97, 100, 101 Thijssen, H.A.C. 125 thiols aroma analysis methodology 69, 70, 71, 72, 73 brews 80, 81±2 formation 84±5 heating effects 82 physical properties 136, 137 roast coffee 76, 77, 78, 79, 80 storage effects 79 tocopherols 42±4 tracheomycosis (fusarium wilt disease) 195, 231 transformation systems, coffee 230±1, 232±3 transgenic plants 230±1, 232±3 2,4,6-trichloroanisole (2,4,6-TCA) 75, 137 triglycerides 33, 34±5, 41 trigonelline 53, 54, 154, 155 Turkish coffee 143, 171, 172 Uganda robusta cloning program 217 units of measurement 242±5
Index
urinary tract cancers 166±7 vacuum coffee 144 viscosity, brews 14±15, 152±3 visual characteristics brews 157±8 international standards 236 volatile acids 26±7, 28 volatile compounds 68±9 aroma analysis methodology 69±73 brews 77±8, 80±2 green coffee 73±5 instant coffee processing 126±7, 129±30, 131±2, 133±7 odorant formation 82±5 olfaction 159 physical properties 133±7 roast coffee 75±80 Volatile Organic Compound (VOC) Directive 109 water brew characteristics 156±7 brew preparation 141, 143±6, 147±8, 149 chlorination 156 demineralizers 157
257
hardness 151, 156, 157 softening 151, 156±7 water content, roasting beans 94±5, 96±7, 101, 235±6 water decaffeination 108, 110±13, 120±1, 122 water shortages, drought tolerance 196 wax 33, 45±6 decaffeinated coffee 95±6 weather stresses 196, 231±2 World Coffee Conference 239 world coffee production 184 world germplasm collections 186 world research network 239 xanthosine-N7-methyl transferase (XMT) 228±30 Xylella fastidiosa 195 xylose green coffee 6 soluble coffee 9, 10, 12 Xylotrechus quadripes 195 zeolites decaffeination processes 111±12, 117 roasting processes 106 zinc-chelating compounds 62±5