Advances in Food and Nutrition Research Volume 53

Advances in

FOOD AND NUTRITION RESEARCH VOLUME

53

ADVISORY BOARDS KEN BUCKLE University of New South Wales, Australia

MARY ELLEN CAMIRE University of Maine, USA

ROGER CLEMENS University of Southern California, USA

HILDEGARDE HEYMANN University of California, Davis, USA

ROBERT HUTKINS University of Nebraska, USA

RONALD JACKSON Quebec, Canada

HUUB LELIEVELD Global Harmonization Initiative, The Netherlands

DARYL B. LUND University of Wisconsin, USA

CONNIE WEAVER Purdue University, USA

RONALD WROLSTAD Oregon State University, USA

SERIES EDITORS GEORGE F. STEWART

(1948–1982)

EMIL M. MRAK

(1948–1987)

C. O. CHICHESTER

(1959–1988)

BERNARD S. SCHWEIGERT (1984–1988) JOHN E. KINSELLA

(1989–1993)

STEVE L. TAYLOR

(1995–

)

Advances in

FOOD AND NUTRITION RESEARCH VOLUME

53 Edited by

STEVE L. TAYLOR University of Nebraska, Lincoln

AMSTERDAM • BOSTON • HEIDELBERG • LONDON NEW YORK • OXFORD • PARIS • SAN DIEGO SAN FRANCISCO • SINGAPORE • SYDNEY • TOKYO Academic Press is an imprint of Elsevier

Academic Press is an imprint of Elsevier 84 Theobald’s Road, London WC1X 8RR, UK Radarweg 29, PO Box 211, 1000 AE Amsterdam, The Netherlands Linacre House, Jordan Hill, Oxford OX2 8DP, UK 30 Corporate Drive, Suite 400, Burlington, MA 01803, USA 525 B Street, Suite 1900, San Diego, CA 92101-4495, USA First edition 2007 Copyright # 2007 Elsevier Inc. 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 without the prior written permission of the publisher. Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford, UK: phone (+44) (0) 1865 843830; fax (+44) (0) 1865 853333; email: [email protected]. Alternatively you can submit your request online by visiting the Elsevier web site at http://elsevier.com/locate/permissions, and selecting Obtaining permission to use Elsevier material. Notice No responsibility is assumed by the publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made. ISBN: 978-0-12-373729-8 ISSN: 1043-4526 For information on all Academic Press publications visit our website at books.elsevier.com Pinted and bound in USA 07 08 09 10 11 10 9 8 7 6 5 4 3 2 1

CONTENTS

Contributors

1. Influence of Processing on Functionality of Milk and Dairy Proteins

vii

1

Mary Ann Augustin and Punsandani Udabage Introduction Physical Modification Processes Enzymatic Modification Processes Chemical Modification Processes Emerging Processes Conclusion Acknowledgment References I. II. III. IV. V. VI.

2. Central Nervous System Tissue in Meat Products: An Evaluation of Risk, Prevention Strategies, and Testing Procedures

2 4 18 23 27 29 30 30

39

M.B. Bowling, K.E. Belk, K.K. Nightingale, L.D. Goodridge, J.A. Scanga, J.N. Sofos, J.D. Tatum, and G.C. Smith Introduction Prevalence as an Evaluator of BSE Food Safety Risks Carcass Contamination with Potentially Infectious Tissues Methods of Detection of CNS Tissue in Meat Products Conclusion and Future Trends References I. II. III. IV. V.

3. Functional Genomics of Wine Yeast Saccharomyces cerevisiae

40 42 45 51 60 61

65

Linda F. Bisson, Jonathan E. Karpel, Vidhya Ramakrishnan and Lucy Joseph Introduction Challenges in the Investigation of Native Yeast Strains Functional Genomic Analysis of Wine Yeast Conclusions Acknowledgments References I. II. III. IV.

66 66 91 108 109 109

v

vi

Contents

4. Monascus Rice Products

123

Tseng-Hsing Wang and Tzann-Feng Lin I. II. III. IV. V.

The Taxonomy of Monascus spp. History of Using Monascus Rice Products in Asia Production Methods Evidence for Health Benefits Safety Acknowledgments References

124 126 128 137 148 151 151

5. Designer Milk

161

Latha Sabikhi Introduction Milk ‘‘Designing’’: The Prospects Milk Fat Modification Milk Sugar (Lactose) Modification Milk Protein Modification Designer Milk for Infant Health Milk with Human Therapeutic Proteins Designer Milk for Animal Growth and Health Assorted Advantages The Future References I. II. III. IV. V. VI. VII. VIII. IX. X.

6. The Sweet Taste Receptor: A Single Receptor with Multiple Sites and Modes of Interaction

162 163 165 174 176 179 183 187 190 191 193

199

Pierandrea Temussi Introduction Indirect Mapping of Active Sites Sweet Macromolecules The Sweet Taste Receptor Mechanisms of Interaction Beyond the Sweet Receptor Acknowledgments References I. II. III. IV. V. VI.

Index See Color Insert in the back of this book

200 202 209 218 221 231 232 232 241

CONTRIBUTORS

Numbers in parentheses indicate the pages on which the authors' contributions begin.

Mary Ann Augustin School of Chemistry, Monash University, Clayton, VIC 3800, Australia and Food Science Australia, Werribee, VIC 3030, Australia (1) K.E. Belk Center for Red Meat Safety, Department of Animal Sciences, Colorado State University, Fort Collins, Colorado 80525 (39) Linda F. Bisson Department of Viticulture and Enology, University of California, Davis, California 95616 (65) M.B. Bowling Center for Red Meat Safety, Department of Animal Sciences, Colorado State University, Fort Collins, Colorado 80525 (39) L.D. Goodridge Center for Red Meat Safety, Department of Animal Sciences, Colorado State University, Fort Collins, Colorado 80525 (39) Lucy Joseph Department of Viticulture and Enology, University of California, Davis, California 95616 (65) Jonathan E. Karpel Department of Viticulture and Enology, University of California, Davis, California 95616 (65) Tzann-Feng Lin Liquor Research Institute, Taipei 106, Taiwan, Republic of China (123) K.K. Nightingale Center for Red Meat Safety, Department of Animal Sciences, Colorado State University, Fort Collins, Colorado 80525 (39) Vidhya Ramakrishnan Department of Viticulture and Enology, University of California, Davis, California 95616 (65)

vii

viii

Contributors

Latha Sabikhi Dairy Technology Division, National Dairy Research Institute, Karnal 132001, Haryana, India (161) J.A. Scanga Center for Red Meat Safety, Department of Animal Sciences, Colorado State University, Fort Collins, Colorado 80525 (39) G.C. Smith Center for Red Meat Safety, Department of Animal Sciences, Colorado State University, Fort Collins, Colorado 80525 (39) J.N. Sofos Center for Red Meat Safety, Department of Animal Sciences, Colorado State University, Fort Collins, Colorado 80525 (39) J.D. Tatum Center for Red Meat Safety, Department of Animal Sciences, Colorado State University, Fort Collins, Colorado 80525 (39) Pierandrea Temussi Dipartimento di Chimica, Universita` di Napoli Federico II, Via Cinthia, Napoli I-80126, Italy; and National Institute for Medical Research, The Ridgeway, London NW7 1AA, United Kingdom (199) Punsandani Udabage Food Science Australia, Werribee, VIC 3030, Australia (1) Tseng-Hsing Wang Liquor Research Institute, Taipei 106, Taiwan, Republic of China (123)

CHAPTER

1 Influence of Processing on Functionality of Milk and Dairy Proteins Mary Ann Augustin*,† and Punsandani Udabage†

Contents

I. Introduction II. Physical Modification Processes A. Heat treatment B. Acidification C. Addition of mineral salts D. Homogenization and shear E. Dehydration III. Enzymatic Modification Processes A. Renneting B. Hydrolysis C. Transglutamination IV. Chemical Modification Processes A. Use of chemical agents B. Maillard reaction V. Emerging processes VI. Conclusion Acknowledgment References

Abstract

The inherent physical functionality of dairy ingredients makes them useful in a range of food applications. These functionalities include their solubility, water binding, viscosity, gelation, heat stability, renneting, foaming, and emulsifying properties. The suitability of

2 4 4 9 13 15 16 18 18 20 22 23 23 25 27 29 30 30

* School of Chemistry, Monash University, Clayton, VIC 3800, Australia {

Food Science Australia, Werribee, VIC 3030, Australia

Advances in Food and Nutrition Research, Volume 53 ISSN 1043-4526, DOI: 10.1016/S1043-4526(07)53001-9

#

2007 Elsevier Inc. All rights reserved.

1

2

Mary Ann Augustin and Punsandani Udabage

dairy ingredients for an application can be further tailored by altering the structure of the proteins using appropriate processes. The processes discussed include physical modification (heat treatment, acidification, addition of mineral slats, homogenization, and shear), enzymatic modification (renneting, hydrolysis, and transglutamination), and chemical modification (use of chemical agents and the Maillard reaction). Emerging food processes (high pressure and ultrasound) are also discussed. The challenges for using dairy ingredients for the delivery of nutrients and bioactive components, while maintaining physical functionality, are also highlighted. There is a need for continued research into the fundamental aspects of milk proteins and their responses to various stresses for further differentiation of milk products and for the delivery of ingredients with consistent quality for target applications.

I. INTRODUCTION Milk and dairy ingredients are used in a range of food applications. Their value as food ingredients stems from their ability to impart a range of desirable attributes to food. They contribute to nutritional quality as they are a good source of nutrients. They have roles in influencing the textural and sensory characteristics of the food because of their physical functional properties. These include the ability of milk proteins to hold water and impart viscosity, to form gels, foams, and emulsions and to remain stable during exposure to heating under appropriate conditions. The milk fat component also contributes to the properties of food as it possesses a desirable delicate flavor and can influence the textural properties of food. Although dairy ingredients have traditionally been used for their nutritional and physical functional properties, there is now an increasing interest in the bioactivity of milk components and their potential to have physiological functional roles. This makes them also attractive as ingredients in functional foods that have a role beyond normal nutrition. Although dairy ingredients can potentially provide a range of functionalities, the requirements of dairy ingredients vary with their application. Hence, matching the functionality of the dairy ingredient to their end-use is of paramount importance for their successful incorporation into foods as each application may require one or several functional properties. While dairy ingredients inherently possess several functionalities, their suitability for an application can be further tailored by appropriate processing of milk or ingredients separated from milk and/or modification of the protein (e.g., whey proteins, casein) and nonprotein components (e.g., fat, mineral salt) of a dairy stream (Augustin, 2004; Augustin and Versteeg, 2006).

Dairy Protein Functionality

3

Both the protein and fat components in milk influence the properties of food, but the ability of the milk to impart desirable properties to food is mostly influenced by the physical functional properties of the milk protein components (Kinsella, 1984; Mulvihill and Fox, 1989). The inherent functionality of milk proteins is related to the structural/ conformational properties of protein, which is influenced by both the intrinsic properties of the protein and extrinsic factors. Modification of the protein composition or structure and the organization of the proteins within the dairy ingredient through the application of physical, chemical, or enzymatic processes, alone or in combination, enable the differentiation of the functionality of the ingredient and designing the required functionality for specific applications (Chobert, 2003; Foegeding et al., 2002). This chapter discusses the influence of processing on the physical functional properties of the milk and milk protein components. The modification of the physical properties of milk, milk powders, and milk protein-based products by the application of various unit processes is the focus of this chapter. Examples are given to demonstrate the effects of various physical, chemical, and enzymatic processes on the structure and functionality of the dairy ingredients. The functionalities that will be covered include solubility, water binding, viscosity, gelation, heat stability, renneting, foaming, and emulsifying. These are the major functionalities that contribute to the physical properties of food (Tables 1 and 2). The potential for the application of emerging food processing TABLE 1

Functional properties of milk proteins

Functionality

Attributes

Water binding

Ability to bind water and swell Dependent on water–protein interactions through peptide bonds or side chains

Solubility

Ability to dissolve Prerequisite for most other desired properties Dependent on pH Proteins are least soluble at their pI

Heat stability

Ability to withstand heat without thickening Essential attribute in many food product applications

Viscosity and gelling

Ability to thicken and form a gel Related to hydration properties and ability to form a network

Emulsifying and foaming

Ability to stabilize interfaces Dependent on the amphiphilic properties of proteins and their ability to unfold at an interface

4

Mary Ann Augustin and Punsandani Udabage

TABLE 2

Desirable Functional Properties of Dairy Proteins for Food Applications

Application

Major Desirable Functionalities

UHT milk and evaporated milk Sweetened condensed milk Cheese Yoghurt Ice cream Confectionery Bakery Manufactured meat and fish products Chocolate

Heat stability, emulsifying Viscosity Rennetability Water binding, viscosity, gelling Foaming, emulsifying Water binding, foaming, emulsifying Water binding, foaming, emulsifying Water binding, foaming, emulsifying High ‘‘free-fat’’

technologies for modification of dairy ingredient functionality and challenges for using dairy ingredients for the delivery of nutrients and bioactive components are highlighted.

II. PHYSICAL MODIFICATION PROCESSES Heat treatment of milk has been one of the most common methods used to alter its functionality. Other processing treatments such as the alteration of pH, mineral adjustment, or homogenization or a combination of these can affect the physical functionality of milk. Processes used in the production of dried dairy ingredients also can influence their functional properties, particularly in the manufacture of powders with high protein content.

A. Heat treatment 1. Influence on milk components The primary purpose of heat treatment is to destroy harmful microorganisms. However, heat treatments induce many other changes in milk, including inactivation of enzymes, denaturation of whey proteins, alteration of the states of association of the casein micelles, chemical modification of amino acid side chains, and changes in the equilibria of the milk salts (Fox, 1989; Holt, 1995). The consequence of heat treatment is the altered functionality of the milk and the dairy ingredient. The extent of the change in functionality depends on the time and temperature of the treatment and the original composition of the dairy stream and the degree of reversibility of the heat-induced changes.

Dairy Protein Functionality

5

a. Whey protein denaturation Heat treatment of milk above 60  C causes denaturation of whey proteins. The extent of denaturation depends on the temperature, and pH at the time of heating, with increasing pH above the natural pH of milk increasing the rate of denaturation of b-lactoglobulin and a-lactalbumin (Law and Leaver, 2000). At temperatures up to 90  C, unfolding of the protein is rate-limiting but further increases in the heating temperature result in only small increases in the rate of denaturation as aggregation of the proteins becomes rate-limiting (Tolkach and Kulozik, 2005). Increasing pH above the natural pH of milk markedly accelerates the rate of denaturation of b-lactoglobulin. The denatured whey proteins associate with the casein micelles or remain in the serum phase as complexes of denatured whey proteins or denatured whey protein in association with k-casein. Generally a decrease in the pH of milk systems prior to heating results in more association of the denatured whey proteins to the casein micelle (Corredig and Dalgleish, 1996; Oldfield et al., 2000; Vasbinder and de Kruif, 2003). Even small changes in pH can shift the distribution of the association of the denatured whey protein with the casein micelle. For example, at a level of 95% whey protein denaturation, there is 70% of the denatured whey proteins associated with the casein micelle at pH 6.55 and this is decreased to 30% when the pH of milk prior to heating was 6.7. This was reflected in the larger increase in the casein micelle size when milk was heated at the lower pH (Anema and Li, 2003). Characterization of the aggregates in heated milk revealed that the serum aggregates are mainly disulphide-linked complexes of whey protein and k-casein (Jean et al., 2006). Increasing the pH of milk from 6.5 to 7.2 produces smaller aggregates with a higher content of k-casein (Renan et al., 2006). In contrast to the whey proteins, caseins are more stable to heat. However, at high temperatures for long times (120–150  C up to 60 min), there is aggregation, fragmentation, and dephosphorylation of casein and destruction of some amino acids (Guo et al., 1989). The heat-induced changes to proteins and their states of association have significant consequences for the functionality of proteins. Heat treatment of milk and dairy streams is often used to manipulate the physical functionality of these ingredients.

2. Effects on heat stability Milk and dairy streams need to be stable during heat processing as it is an integral step in the manufacture of dairy products. The heat stability of milk, which is the ability of milk to withstand heat treatment without excessive thickening or coagulation, has therefore been a subject that has attracted a lot of interest over many years.

6

Mary Ann Augustin and Punsandani Udabage

a. Milk and concentrated milks Single strength milk is heat stable at its natural pH (6.7). Concentration of milk decreases heat stability and shifts the pH of maximum heat stability to lower pH. For example, skim milk (20% solids) has significantly lower heat stability than normal strength milk even at its pH of maximum heat stability at pH 6.4–6.6 (Singh, 2004). Most milk has maximum heat stability at pH 6.7 and minimum heat stability at pH 6.9. When the pH is increased above 6.9, there is an increase in heat stability. The maintenance of micellar integrity is important for heat stability. The decrease in heat stability of single strength milk at pH 6.9 is often linked to the dissociation of k-casein from the micelle (Singh and Fox, 1985). This is because the k-casein-depleted micelles are more susceptible to calcium-induced precipitation. The stabilizing effect of b-lactoglobulin on the heat stability of milk at the pH of minimum heat stability has been related to its ability to reduce the dissociation of k-casein (Singh and Fox, 1987). O’Connell and Fox (2001) suggested that the role b-lactoglobulin at the pH of minimum heat stability may be linked to its effect on sensitizing casein micelles to heat-induced precipitation of calcium phosphate by increasing the hydrophobicity of the micelle. In addition, they found that the heat-induced precipitation of calcium phosphate as a function of pH appears to be inversely related to the heat stability of milk. The change in heat stability of milk as a function of pH and concentration has been related to the states of the association of the milk proteins and the equilibria between the milk proteins. Hence, it is not surprising that the common treatments used for improving heat stability of concentrated milks have involved giving milk a preheat treatment prior to concentration. Newstead and Baucke (1983) investigated the effect of different preheat treatments (10–240 s at 90–140  C and from 10–2700 s at 90  C) of raw skim milk on heat stability of concentrated milks. The treatment at 110–120  C for 120–240 s was found to be most effective for improving heat stability. Another lever that has been used to improve the heat stability of concentrated milk is the pH adjustment of the skim milk prior to preheating. The heat stability of recombined concentrated milk was generally improved by lowering the pH of the skim milk by 0.05–0.10 units prior to processing (Newstead and Conaghan, 1978). An alternative strategy for improving heat stability of concentrated milk is to alter the mineral balance of the system. This is discussed in Section II.C.2. b. Modified milks Modification of the ratio of casein to whey protein or protein standardization of milk with other milk fractions expands the range of dairy products. The addition of whey protein concentrate or individual whey proteins (a-lactalbumin or b-lactoglobulin) to milk generally causes a reduction in heat stability (Rattray and Jelen, 1997). Standardization of the protein content of skim milk with ultrafiltered permeates from

Dairy Protein Functionality

7

various dairy sources (skim milk, sweet whey, and acid whey) produced standardized milk with different heat stability. This depended on the source used for preparation of the permeate. Increased heat stability was obtained with addition of permeate from skim milk or sweet whey, whereas heat stability decreased when permeate from acid whey was used. The detrimental effects on heat stability were related to the higher level of soluble calcium in permeate from acid whey (Rattray and Jelen, 1996).

c. Whey Whey proteins are inherently unstable to heating. The stability of whey to heat-induced precipitation can be reduced by the addition of phosphates or citrates. These complexing agents mask the effect of calcium on precipitation of proteins (de Rham and Chanton, 1984). The presence of caseins in combination with whey proteins at the time of heating protects the whey protein against precipitation. Dickinson and Parkinson (2004) found that substitution of 10% of the b-lactoglobulin with caseinate improved the heat stability of emulsions (10 vol % oil, 2 wt % protein). O’Kennedy and Mounsey (2006) showed that as1/b-casein improved the stability of whey protein isolate suspensions (0.5 wt %, pH 6.0) heated at 85  C for 10 min. The ratio of whey protein isolate: as1/b-casein required for complete suppression of heat-induced aggregation was 1:0.1 (w/w). The presence of the caseins did not alter the extent of denaturation but inhibited denatured whey protein aggregation. Micellar casein could also be used to stabilize whey proteins against heat-induced aggregation. Although whey protein denaturation was promoted in the presence of micellar casein, aggregation of the denatured whey proteins was controlled down to pH 5.4. Thus, although both as1/b-casein mixtures and micellar casein can stabilize whey proteins, the mechanisms of their actions are different.

3. Effects on gelling properties at neutral pH Heat treatment of whey protein dispersions results in the formation of precipitates, gels, or soluble polymer dispersions depending on pH, ionic strength, protein concentration, and the presence of salts (Donovan and Mulvihill, 1987; Foegeding et al., 2002). By controlling the conditions of the heat treatment, whey-based gels of varying textures may be obtained. Most research on gels of dairy proteins at approximately neutral pH has been done on whey proteins. The literature on the effects of gelling of whey proteins suggests that the states of protein created by the heat treatment, which can be manipulated by a variety of factors including pH, time and temperature of heating, presence of salts, concentration of protein at time of heating, and order of processing, dictate the properties of heat-induced whey protein gels. As whey proteins are heated, they denature. A gel is formed when there is a sufficient interaction between the denatured protein molecules. A precipitate, gel, or soluble whey aggregates is obtained depending on

8

Mary Ann Augustin and Punsandani Udabage

protein concentration, temperature and time of heating, pH, and ionic strength. The gels obtained can be particulate, fine-stranded, or have a mixed network (Bottcher and Foegeding, 1994; Mangino, 1992; Mulvihill and Donovan, 1987). Where there is a strong electrostatic repulsion between the protein particles, fine-stranded gels are formed (i.e., when the protein is at a pH much above the isoelectric point and at low ionic strength). Particulate gels are formed when the pH is nearer the isoelectric point and in high ionic strength environments where the electrostatic charge is screened. Particulate gels possess a loose network of large protein particles and these gels have low water-binding capacity and are usually opaque. Fine-stranded gels have a network caused by the association of strands and these have good water-binding activity and are usually translucent (Langton and Hermansson, 1992). The texture, opacity, and water-holding properties of the whey gels can be further modulated by salts as they influence protein–protein interactions. When heated, whey protein isolate dispersions (10% w/w; pH 6.9; 80  C/15 min) were made with different concentrations of NaCl, the heated dispersions did not form a gel and remained transparent at low NaCl concentration (<60 mM NaCl). However, they formed gels with increasing elastic modulus and high water-binding capacity with increasing NaCl concentration (60–150 mM NaCl). At higher salt concentrations (150–200 mM), gels had decreased elastic modulus and lower water-binding capacity (Chantrapornchai and McClements, 2002). An alternative approach to the formation of whey gels is to predenature whey proteins under carefully controlled conditions to form aggregates, followed by cold gelation in the presence of salts (Bryant and McClements, 2000). The formation of these cold-set gels involves a heat treatment of native whey dispersions at about neutral or alkaline pH under conditions of low ionic strength and at low enough protein concentration to avoid gelation during heating. This causes unfolding of the native whey proteins and the formation of disulphide cross-linked aggregates. The conditions of heating (e.g., pH and whey protein concentration) as well as the rate of cold gelation can be manipulated to give cold-set gels with different properties (Alting et al., 2003; Bryant and McClements, 2000). Mleko and Foegeding (2000) prepared whey protein polymers by heating whey protein isolate dispersions (4% w/v; pH 8) at 80  C and adjusting pH (6.0–8.0). Weak gels were formed at pH 6.0 and 6.5, while highly viscous solutions were obtained at pH 7.0–7.5. Yet another approach has been the development of cold gelling-derivatized whey protein isolate powders. The preparation of these powders involved heating whey proteins to form gels and subsequent drying of the preparations. Reconstitution of the powders in water at ambient or refrigeration temperatures results in viscous solutions or weak gels (Firebaugh and Daubert, 2005; Hudson et al., 2000; Resch and Daubert, 2002). Other approaches that have been examined to manipulate gelling properties

Dairy Protein Functionality

9

include a two-stage heat-induced polymerization and aggregation process. Whey protein isolate dispersions subjected to a double heating step (30 min at 80  C at pH 8 followed by 30 min at 80  C at pH 7.0) were compared to those subjected to a single heating step (30 min at 80  C and pH 7.0). Significant increases in the elastic modulus of gels could be obtained by using a double heating process compared to a single heating step (Glibowski et al., 2006). There are novel uses for gelling dairy ingredients as demonstrated by the use of cold-set b-lactoglobulin gels for delivery of iron. This has been achieved by the addition of Fe2+ to preheated b-lactoglobulin dispersions to form gels with entrapped iron with different microstructures and iron: protein ratios (Remondetto et al., 2002).

4. Effects on surface properties The effects of heat treatment on functional properties other than heat stability or gelling have received less attention. Heat treatment and pH at the time of heat treatment have an impact on the surface properties of proteins. The amount of denaturation and aggregation induced by various heat treatments has to be controlled to optimize the surface properties of proteins. An increase in surface hydrophobicity of whey protein concentrates with heat treatments has been correlated with an improvement in emulsifying and foaming capacity (Moro et al., 2001). A heat treatment is required to expose buried hydrophobic groups of proteins. This is required for improving surface properties. However, if the heat treatment is too harsh, aggregation can occur and this decreases surface hydrophobicity leading to a decrease in surface properties. Zhu and Damodaran (1994) suggested that the ratio of monomeric to polymeric protein in whey protein isolates exposed to heat treatments (70 or 90  C) influenced their foaming properties. When heat treatment of whey protein concentrates was carried out at 84  C for 30 s (pH 6.0, 6.5, and 7.0), the improvement in emulsion stability was greatest when pH was 6.0 (Moon and Mangino, 2004). However, when the heat treatment was applied to the whey prior to ultrafiltration for preparation of whey protein concentrates, the emulsifying properties of the resultant whey protein concentrate were improved by heat treatment (70  C for 2 min) when pH was increased from pH 6 to 7 but a higher heat treatment (80  C for 2 min) decreased emulsifying properties at pH 7 (Fachin and Viotto, 2005). The different effects of heating, depending on the dairy stream and heating conditions, highlight the need to control the protein species in each of the heated streams.

B. Acidification Acidification results in an alteration of the protein and mineral equilibria with consequent effects on the physical, chemical, and functional properties of milk and ingredients.

10

Mary Ann Augustin and Punsandani Udabage

1. Influence on milk components Acidification of milk results in the formation of a gel. Gelation of milk is primarily due to the charge neutralization of the protein particles in milk as the isoelectric point of the milk protein particles is approached. The micelles maintain their original size until the first signs of gelation. The nature of the casein particles in milk is changed as a consequence of lowering pH because of the complex equilibrium that dictates the distribution of minerals and caseins between the serum and colloidal phases of milk. Acidification causes the solubilization of colloidal calcium phosphate, resulting in an increase in Ca2+ and phosphate activity (Dalgleish and Law, 1989; Van Hooydonk et al., 1986a). The solubilization of colloidal calcium phosphate is accompanied by a release of caseins from the micelles. Approximately about 14% of the calcium is still present in the micellar phase even after virtually all the colloidal inorganic phosphate is solubilized at pH 5.3 and 30  C (Van Hooydonk et al., 1986a). The calcium that remains is regarded as the calcium directly bound to casein through phosphoseryl and carboxylate residues. The amount of caseins released at a given pH decreases with increasing temperature. The percentage of (b + g)-casein released at a given pH is greater than that of other caseins at a fixed temperature (Dalgleish and Law, 1988). The pH at which the maximum dissociation of casein occurs also depends on the temperature, with the pH of the maximum decreasing with decreasing temperature. This occurs at pH 5.1 at 4  C, 5.4 at 20  C and 5.5–5.6 at 30  C (Dalgleish and Law, 1988; Van Hooydonk et al., 1986a). The gelation pH increases with increasing temperature, occurring at pH 5.0 at 15  C and pH 5.1 at 20  C. Heat treatment of milk also increases gelation pH (Banon and Hardy, 1991). For example, milk heated at 80–90  C has a gelation pH of 5.4. The gelation pH has been related to the percentage of total b-lactoglobulin in the heated milk (Vasbinder et al., 2003).

2. Acidified milk gels and yoghurts Heating of milk prior to addition of cultures is known to increase the firmness of yoghurt and reduce syneresis (Augustin et al., 1999; Dannenberg and Kessler, 1988a,b). Heat treatment of milk prior to acidification with glucono-d-lactone has been shown to increase the firmness of the acid gels, to increase the pH at which gelation occurs and to reduce syneresis (Lucey et al., 1997, 1998). The higher pH at gelation of heated milks was considered to be due to the alteration of the casein micelle surface because of the attachment of denatured whey protein to the casein micelle. However, by using confocal scanning electron microscopy with separate staining of the proteins in milk, it has been shown that the denatured whey proteins that remain in the serum phase also gel at the

Dairy Protein Functionality

11

same time as the whey-coated casein micelles (Vasbinder et al., 2004). Guyomarc’h et al. (2003) suggested that the soluble aggregates containing denatured whey proteins in heated milk have a greater effect on increasing gelation pH than whey proteins bound to casein micelles.

a. Effects of milk solids concentration The properties of acid and yoghurt gels may be altered by changing the concentration and distribution of proteins in the milk. However, the effects obtained depend on the milk composition and conditions used during the heat treatment. This is because both these factors affect the association of the milk proteins. Yoghurts made with higher milk solids generally have improved properties, but these effects are dependent on the source of milk solids used. Yoghurts made from milks fortified with ultrafiltered milk solids were firmer that those fortified with skim milk powder (Becker and Puhan, 1989). b. Effects of alteration of the casein:whey ratio Changing the ratio of casein:whey protein in yoghurt milk results in yoghurts with different textures. Yoghurts had increased gel strength and reduced syneresis when the protein content of yoghurt milk was kept constant and the ratio of whey protein to casein was increased. This was attributed to differences in microstructure of the yoghurts where increasing whey protein led to a finer structure and a denser network of protein aggregates (Puvanenthiran et al., 2002). Augustin et al. (2003) showed that increasing the protein content of yoghurt milk by partial substitution (20%) of skim milk solids with whey protein concentrates increased the firmness of yoghurts. However, the increase was obtained only in cases where the yoghurt milk was stable to the heat treatment applied prior to the addition of cultures. When there was excessive thickening during the heat treatment of yoghurt milk, the properties of the yoghurt were compromised, suggesting that the state of aggregation of the proteins in the heated milk affected yoghurt properties. c. Effect of heating conditions Altering the pH of the heat treatment prior to acidification, which alters the states of association of whey proteins with the casein micelles, may be used to manipulate the properties of milk gels. The change in gelation properties of milks heated at different pH prior to acidification has been related to the differences in proportions of denatured whey proteins associated with the casein micelle and soluble complexes in the serum. Adjustment of milk pH (6.9–6.35) prior to heat treatment and readjustment to pH 6.7 followed by acidification markedly change the gelation properties of milk. The differences in gelation properties are related to the

12

Mary Ann Augustin and Punsandani Udabage

pH dependence of casein–whey protein interactions, which results in different protein structures being formed at the time of heating. At pH > 6.6, there is partial coverage of casein micelles and separate whey protein aggregates, whereas at pH < 6.6, the whey proteins are attached to the micelles. The different protein structures formed during heating at various pH influence their gelation properties on acidification (Vasbinder and de Kruif, 2003). Anema et al. (2004) found that increasing the pH at heating from 6.5 to 7.1, which increases the amount of aggregates in the serum, followed by readjustment to pH 6.7 prior to addition of glucono-d-lactone, increased the pH of gelation, decreased gelation time, and increased the firmness (elastic modulus) of the gels. Large differences in gel properties were obtained, even though there were small changes in the extent of denaturation of whey proteins. Their results suggest that the decreased association of the denatured whey proteins with casein micelles and increased levels of soluble whey protein complexes obtained with increase in pH at heating improved gelation properties. The importance of the soluble complexes in heated milk on the structure of acid gels has been confirmed by others who showed that increasing the amounts of soluble complexes in milk by increasing the pH of heat treatment of milk from 6.5 to 7.1 gave rise to stronger acid gels. However, heat treatment of milk at higher pH (7.2), which further increases the amount of soluble material, weakens the acid gels formed from these milks (Rodriguez del Angel and Dalgleish, 2006). Schorsch et al. (2001) examined the effects of denaturation of whey proteins in the presence and absence of casein micelles on gel properties. Heat treatment sequence was found to influence the acid gelation properties of casein–whey mixtures. Denaturation of whey proteins in the absence of casein micelles induced more rapid gelation on addition of acid. Gels made from these milks had a more particulate gel structure than gels made from casein–whey mixtures which were heated without prior denaturation of the whey proteins.

3. Acidified whey gels These types of gels are prepared by heating whey proteins and acidifying the heated solutions. The presence of calcium ions in solutions at the time of heating affects the acid gels subsequently formed. Britten and Giroux (2001) made gels by preheating whey at pH 6.5–8.5 in the absence or presence (up to 4 mM) of calcium at 90  C for 15 min prior to acidification. Opaque particulate gels were formed and their gel strength was dependent on the type of whey polymers formed. Where polymers with high intrinsic viscosity were produced on heating, these generally resulted in strong gels. The ability to manipulate acid gelation properties of whey

Dairy Protein Functionality

13

polymers enables their incorporation into yoghurt formulations (Britten and Giroux, 2001).

C. Addition of mineral salts The addition of mineral salts to alter protein and mineral equilibria in milk is a strategy that has been used to manipulate milk functionality, either alone or in combination with other processing treatments, such as alteration of pH, ultrafiltration, diafiltration, heating and cooling, or static high-pressure treatment.

1. Influence on milk components In milk, the caseins and minerals are in dynamic equilibrium between the micellar (colloidal) and the serum phase. When the native environment of caseins and minerals is altered by the addition of mineral salts, the partitioning of minerals and caseins between the serum and colloidal phases is altered. Altering the composition of milk with the addition of mineral salts at constant pH induces shifts in the mineral and casein partition, causing the establishment of new positions of equilibria. In general, addition of calcium or inorganic phosphate causes transfer of serum calcium or inorganic phosphate into the colloidal phase. The Ca2+ activity is also affected, increasing on the addition of calcium and decreasing on the addition of inorganic phosphate (Rose, 1968; Tessier and Rose, 1958; Udabage et al., 2000; Van Hooydonk et al., 1986c). The content of colloidal calcium phosphate is also changed on addition of salts. Disintegration of casein micelles is observed with addition of calcium chelating agents (EDTA or citrate) as a result of the solubilization of the colloidal calcium phosphate and the micellar casein (Griffin et al., 1988; Holt, 1982; Lin et al., 1972; Rollema and Brinkhuis, 1989). Solubilization of colloidal calcium phosphate beyond a critical level causes the disintegration of casein micelles and a loss of micellar integrity (Udabage et al., 2000). The alteration of mineral and casein equilibria is reflected in changes to the physical properties of milk. The addition of citrate and different types of phosphates (ortho-, pyro-, or hexameta) to milk protein concentrate solutions, which alters the distribution of calcium and inorganic phosphate between the colloidal and serum phases of milk, affects its turbidity and buffering capacity (Mizuno and Lucey, 2005). The turbidity is affected because dissolution of colloidal calcium phosphate is accompanied by release of caseins into the serum. The changes in the equilibria affect many of the functional properties including heat and ethanol stability, renneting, solubility, foaming, and

14

Mary Ann Augustin and Punsandani Udabage

emulsifying. The effects of altering mineral equilibria on functionality of milk have been previously reviewed (Augustin, 2000).

2. Effects on heat stability It is well known that the addition of soluble calcium salts reduces the heat stability of milk, whereas the addition of calcium complexing agents with the appropriate control of pH improves heat stability. Phosphates and citrates have often been used to increase the heat stability of concentrated milks (Augustin and Clarke, 1990; Pouliot and Boulet, 1991; Sweetsur and Muir, 1982a). A reduction in Ca2+ activity by the addition of these salts contributes to the improved heat stability of concentrated milks, but the effects of salts on the equilibrium of caseins between the serum and micellar phases of milk also affect heat stability. Evidence for the importance of mineral-protein equilibria was seen by comparing the heat stability-pH profiles of concentrated milk with added EDTA or phosphates. Although EDTA caused a similar reduction of the Ca2+ activity in recombined concentrated milks compared to those with added phosphate, milks with added EDTA had reduced heat stability. This was attributed to the difference in the level of colloidal calcium phosphate in micelles which changes the partitioning of the caseins between the serum and colloidal phases (Augustin and Clarke, 1990). The interest in mineral fortification of milk for the production of milks with higher nutritional value is a challenge. This is because the introduction of minerals upsets the mineral-protein equilibria in milk which will affect their stability. Philippe et al. (2004) showed that supplementation of skim milk with calcium gluconate, calcium lactate, or calcium chloride (up to 16 mmole added Ca/kg) decreased the heat stability. The addition of MgCl2 or FeCl3 (at a level of 8 mmole/kg) also reduced the heat stability of casein micelles (Philippe et al., 2005). However, by manipulating the mineral equilibria of milk with the use of a combination of soluble calcium salts and orthophosphates, it is possible to produce milks (with up to 20 mmole added Ca/kg) that are stable to heating (Williams et al., 2005). O’Kennedy et al. (2001) showed that denatured whey proteins could be used as a carrier for calcium phosphate and further that adequate heat stability at 130  C of whey protein-calcium phosphate suspensions could be achieved by appropriate adjustment of pH.

3. Effects on surface properties Superior foaming properties of milk have been obtained by addition of calcium complexing agents. Kelly and Burgess (1978) demonstrated that addition of sodium hexametaphosphate to milk protein concentrate solutions prepared by ultrafiltration improved foam volume and stability on whipping. The addition of EDTA to milk, which causes dissociation of the casein micelle, improved the foaming properties of milk (Ward et al., 1997).

Dairy Protein Functionality

15

The increase in soluble casein for interaction with the interface during whipping accounts for the increase in the foaming capacity. It is possible that the increase in the viscosity of milks with increasing states of disaggregation of the casein micelle contributed to this stability. Citrate salts have long been used in the processed cheese industry as ‘‘emulsifying salts,’’ and there is still interest in the mechanism of their action. Shirashoji et al. (2006) examined the effects of trisodium citrate on the properties of processed cheese. Increasing concentration of sodium citrate decreased the size droplets of the cheese. This effect is typical when emulsifying properties of a system are improved. This is expected as the complexation of calcium by citrate causes dissociation of the casein micelle, making the casein more available for emulsifying fat droplets. This possibly contributed to the reinforcement of the structure of the processed cheese.

D. Homogenization and shear The main purpose of homogenization in the dairy industry is for the emulsification of fat. Homogenization results in the creation of smaller fat globules with altered interfaces. A more stable emulsion that is resistant to creaming is usually obtained on homogenization, and this has benefits for fluid milks and dairy products. Homogenization can also have other effects on the functionality of dairy products. For example, heat in combination with shear has been used for the microparticulation of globular proteins.

1. Effects on heat stability Homogenization (up to 20.7 MPa) of whole milk decreases heat stability, with the effect being greater at increasing homogenization pressure. Homogenization of skim milk (up to 31 MPa) has only a negligible effect on skim milk (Sweetsur and Muir, 1983). The position of the homogenization process in the manufacture of concentrated milks in relation to the stage of addition of stabilizing salts (phosphates) influences heat stability. The stabilization of milk to heat by added phosphate was more effective when the phosphates were added prior to homogenization (Sweetsur and Muir, 1982b).

2. Effects on gelling properties The effects of shear on the properties of gels are influenced by the presence of fat in the dairy systems. Homogenized fat droplets can act as active fillers in milk gels. Xiong et al. (1991) found that the addition of emulsified fat into skim milk increased the gelation rate and shear modulus of acid-induced milk gels and that decreasing the fat droplet size at the same fat content resulted in firmer gels. An increase in the fat content

16

Mary Ann Augustin and Punsandani Udabage

of gels at the same solids nonfat:water ratio also increases the complex modulus of acid and heat-induced milk gels (Underwood and Augustin, 1997). Blends of whey protein isolate and denatured whey protein isolates were microparticulated using a microfluidizer prior to the formation of heat-set gels. Increasing the number of passes in the microfluidizer increased the hardness of the gels, an effect attributed in part to the more homogenous gelation of smaller aggregates (Sanchez et al., 1999).

3. Effects on microparticulated whey proteins A combination of heat and shear has been used to create whey protein particles with controlled particle size and properties. A well-known example of the use of microparticulation of thermally denatured whey protein is for the production of SimplesseÒ 100, a whey-based fat replacer (Lieske and Konrad, 1993). Shear can be used to modulate gel properties of whey protein isolate gels. Spiegel and Huss (2002) controlled pH and calcium levels during heat treatment and shearing in a scraped surface heat exchanger to produce whey protein aggregates of between 0.5 and 10 mm, which give a smooth mouthfeel. Heating whey protein concentrate dispersions at 110  C and a low pH (<5.5) produced the small aggregates. Oestergaard (2005) obtained particles of 1–12 mm with creamy consistency by ultrafiltering whey to obtain a protein concentrate (60% protein) and then heating the concentrate under controlled shear rates in a scraped surface heat exchanger.

E. Dehydration The conversion of liquid dairy streams into powders is an important processing operation in the dairy industry. The removal of water from a dairy stream during concentration and drying can influence the functional properties of the resultant powders. This depends on factors such as the composition of the stream to be dried, the type of driers used, and the conditions of drying. Both the physical characteristics of the powders (e.g., particle size, bulk density, occluded air) and their functional properties can be affected by drying (Tong, 2001). Of all the major unit processes in conventional milk powder manufacture such as preheating, concentration, and drying, the preheating step has the major effect on milk powder functionality. In fact, skim milk powders are still classified on the basis of the heat treatment applied to the milk during powder manufacture (American Dairy Products Institute, 1990). The heat classification, based on the amount of undenatured whey protein in the milk powder, is still being used as a general guide to the physical functionality of milk powders and the selection of powders for

Dairy Protein Functionality

17

specific food applications. However, this guide should be used with caution as the initial content of whey protein and the nonprotein nitrogen content of the milk used in the powder production can affect this value. The process for spray-drying of conventional skim and full-cream milk powders is routine in the dairy industry. However, there is still research on the drying of specialized and newer dairy powders (Kelly, 2006).

1. Milk powders for chocolate A desirable property of milk powders intended for chocolate manufacture is a high level of unencapsulated fat (free-fat). The traditional method for manufacture of these powders has been roller drying as the method of drying gives rise to high levels of unencapsulated fat (>85% w/w of fat in powder). This contrasts with the low level of unencapsulated fat (<4% w/w of fat in powder) in milk powder (26% fat) made using the traditional processes of preheating, concentration, homogenization, and spray-drying. Altering the composition of the milk, the process variables, and the order of unit operations during the manufacture of spray-dried powder can affect the level of unencapsulated fat. Increasing the content of total fat in powder (from 26% to 70% w/w) increases the level of unencapsulated fat in powders (Kelly et al., 2002). Clarke and Augustin (2005) found that by separating full-cream milk into cream and skim milk fractions, pasteurizing the cream fraction, then cooling the cream and recombining it with a skim milk concentrate prior to drying increased ‘‘free-fat’’ in powder. Another method of increasing ‘‘free-fat’’ is homogenizing the cream fraction at high temperature and pressure prior to combining it with a skim milk concentrate. These methods enabled the production of whole milk powders (30% total fat in powder) with high levels of unencapsulated fat (up to 40% w/w of total fat). Another factor that influences the level of ‘‘free-fat’’ is the solid-fat content of the milk fat. Twomey et al. (2000) found that high-fat powders (56% total fat in powder) had increased the level of unencapsulated fat when there was an increase in the solid-fat content of milk fat. The particle size of the powders, which is affected by milk composition and the solid-fat content of the milk fat, also affects the suitability of powders for chocolate manufacture (Keogh et al., 2002). By increasing the spray nozzle size for the concentrate and increasing the air outlet temperature of the dryer, the particle size of milk powders can be increased (Keogh et al., 2004). This makes milk powders more suitable for chocolate manufacture.

2. High-protein milk powders Milk protein concentrate (MPC) powders (>50% protein) are typically made by ultrafiltration/diafiltration prior to drying. The drying of these highprotein concentrates is known to cause a loss of functionality. This is typically

18

Mary Ann Augustin and Punsandani Udabage

exemplified by poor hydration properties, loss in solubility, and poor reconstitutability. Good solubility of powders is usually required to enable the functionality of the protein ingredient to be fully realized. This is because solubility is a prerequisite for many other functional properties of proteins such as their ability to build viscosity, gel, and stabilize foams and emulsions. High-protein powders are generally more difficult to reconstitute compared to conventional skim milk powder (34% protein), and the problems of reconstitution are worse with increasing protein content. The solubility of the MPC powders deteriorates further during storage. MPC powders with improved solubility are made by the addition of monovalent salts to the ultrafiltered retentate (Carr, 2002) or by the removal of calcium ions prior to drying (Bhaskar et al., 2003). The effect of removing water beyond a critical level, as can happen during concentration and drying, can lead to aggregation and irreversible denaturation of protein species at an interface. The rate of water removal can induce changes in protein structure. The rate of dehydration is influenced by the mineral environment of dairy concentrates (Schuck et al., 1999). Any change in protein structure can potentially lead to an altered functionality of the powder on reconstitution. The application of a preheat treatment (above 72  C) to the milk protein retentate or increasing the inlet temperature from 200 to 250  C impaired the hydration properties of MPC powders with high protein content (>70% protein) (De Castro-Morel and Harper, 2003). The insoluble particles obtained on reconstitution of high-protein MPC powders have been ascribed primarily to the hydrophobic association of casein micelles in the powders (Havea, 2006).

III. ENZYMATIC MODIFICATION PROCESSES Enzymatic processes such as renneting, hydrolysis, and cross-linking with transglutaminase change the integrity of the casein micelles, resulting in physicochemical and functional changes to milk and milk-derived ingredients. The resultant properties of milk and milk-derived ingredients are largely dependent on the condition in which these enzymatic processes were carried out. An enzymatic route has the benefits of being less harsh compared to the modifications which use chemical agents.

A. Renneting Renneting is the most used enzymatic process in the dairy industry. When milk is treated with rennet, a selective cleavage of the Phe(105)-Met(106) bond of k-casein (hairy layer) occurs due to the action of chymosin. On cleavage, k-casein is split into two polypeptides with very different

Dairy Protein Functionality

19

properties, a hydrophilic caseinomacropeptide containing residues (106– 169) which diffuses into the serum and a hydrophobic para-k-casein (para-casein; residues 1–105) which remains (Dalgleish, 1992). The cleavage of k-casein removes the steric layer and part of the electrostatic repulsion, which stabilizes casein micelles. This results in a decrease in casein micellar size (10–14 nm decrease in diameter) and a decrease in zeta potential by 40%. Both these factors reduce micelle–micelle repulsion and promote the aggregation of the para-casein (Horne and Davidson, 1992; Walstra and Jenness, 1984; Walstra et al., 1981). The paracasein can either be allowed to coagulate forming a cheese curd or precipitated to obtain rennet casein. This type of casein is commonly used as dairy-based cheese analogues.

1. Rennet gels Many factors affect the renneting process. The gel development, the structural changes, and the final strength of the gel are influenced by several factors. Some of these factors include the amount of material capable of forming the gel, both the amount of protein and colloidal calcium phosphate (Casiraghi et al., 1987; Green, 1987; McMahon et al., 1993; Storry and Ford, 1982; Udabage et al., 2001; Zoon et al., 1988), the rate of gel formation, and the concentration of rennet used (Lomholt and Qvist, 1997; Okigbo et al., 1985; Zoon et al., 1988). The rate of the enzymatic cleavage can proceed at temperatures as low as 4  C, although increasing the temperature increases the rate of the reaction (Dalgleish, 1979). Another way of affecting the enzymatic process is to decrease the pH. The pH of maximum velocity is pH 6.0 (Van Hooydonk et al., 1986b). Differences in the coagulation rate of renneted micelles arise from the differences in the neutralization of negative charge within the micelles, a decrease in repulsion promoting the closest approach of micelles and allowing hydrophobic interactions (Dalgleish, 1992). Increasing the concentration of calcium at a fixed pH (Udabage et al., 2001; Van Hooydonk et al., 1986c), reducing the pH (Sharma et al., 1994; Van Hooydonk et al., 1986a), ultrafiltering milk (Sharma et al., 1994), and increasing the temperature from 31  C at pH 6.6 (Fox and Mulvihill, 1990) all promote the aggregation process. Heat treatment of milk above 60  C, which promotes whey protein denaturation and its complexation with k-casein at normal milk pH (6.6), also affects renneting properties. An increase in rennet coagulation time and a decrease in gel firmness were observed with increased heat treatment of milk (Menard and Camier, 2005). Ultra-high temperature (UHT) treated milk failed to coagulate completely but the coagulation properties were restored by threefold concentration of the UHT milk (McMahon et al., 1993).

20

Mary Ann Augustin and Punsandani Udabage

2. Acid and rennet gels Milk gels can be made by the combined action of rennet and acid. With the combined action of acid and rennet, gels can be made over a broader pH and temperature range than by acidification alone, with both the pH and rennet action influencing the resulting gel properties (Roefs et al., 1990). The properties of the mixed gels are different from rennet gels or gels made by acidification. Milk gels formed on acidification are relatively less viscous than rennet gels over timescales longer than 1 s (Van Vliet et al., 1989). The aggregating species in milk renneted at pH 6.7 contains colloidal calcium phosphate, whereas those in acid gels are depleted in colloidal calcium phosphate.

B. Hydrolysis Hydrolysis of proteins results in a cleavage of peptide bonds. This has the effect of reducing the molecular weight of the protein. The original structure and conformation of the protein are lost. Depending on the sites cleaved by enzymes, a range of peptides with altered ratios of hydrophobic to hydrophilic groups are obtained. All these changes will have significant effects on the functionality of the protein. The digestibility of protein is altered and allergenicity of the protein can be reduced. Physical functional properties that are affected by hydrolysis include heat stability, gelling properties, foaming, and emulsification. Hydrolysis is now commonly used to make physiologically functional dairy ingredients. Most of the current interest in hydrolysis of milk proteins is directed at the production of bioactive peptides. This aspect is not covered here but reviews provide an update of these interests (Korhonen and Pihlanto, 2006). Hydrolysis of proteins for modification of functionality has been also covered by reviews (Chobert, 2003; Foegeding et al., 2002; Kilara and Panyam, 2003).

1. Effects on surface properties As hydrolyzed proteins are smaller than unhdyrolyed proteins, they can move to an interface and stabilize it more rapidly than intact proteins. However, in comparison to intact proteins, the smaller peptides in hydrolysates form a less cohesive film at the interface and this can affect the stability of the emulsions and foams. The effects of hydrolysis on surface properties depend on the type of milk protein and the conditions of hydrolysis. Hydrolysis of globular proteins results in the exposure of buried hydrophobic groups. This enhances surface hydrophobicity that improves surface properties. The degree of hydrolysis needs to be optimized for good surface properties. This is governed by the type of protein used, the extent of hydrolysis, and the enzymes used.

Dairy Protein Functionality

21

Foaming properties are affected by hydrolysis of proteins. Limited hydrolysis (4–10%) of whey protein concentrate (WPC80) by a protease from Bacillus licheniformis results in improved foaming properties (Chen, 2003). Partial hydrolysis (up to 6.5%) of whey protein concentrate (WPC35) by pepsin also improved foaming properties but increasing hydrolysis (>6.5%) impaired foaming and emulsification properties. This was considered to be due to the destabilizing effects of small peptides (Konrad et al., 2005). Giardina et al. (2004) showed that hydrolysis of caseinate diminished its ability to foam. The differences may be related to the different structure of the intact proteins, differences in enzymes used, and conditions of the reaction. Hydrolysis of proteins has marked effects on their emulsifying properties. Hydrolyzed whey protein with a degree of hydrolysis of between 10% and 20% had good emulsifying properties (Dalgleish and Singh, 1998). Euston et al. (2001) found that whey protein concentrates with low degree of hydrolysis (4–10%) impaired the emulsifying capacity of whey protein concentrate but increasing the degree of hydrolysis to 10–27% improved emulsifying capacity. However, further increases in the degree of hydrolysis reduced emulsion stability and heat stability of emulsions. In analyzing the observations, factors that control surface activity and stability of foams and emulsions should be considered. Analysis of sequences of peptides and their properties from enzymatic digests may provide a more rational approach to the development of protein hydrolysates with superior emulsifying properties (Panyam and Kilara, 2004). A peptide must be surface-active for it to lower the surface tension. For it to contribute to stability, it should be able to form cohesive films. Both these properties are required for the formation of foams and emulsions. Rahali et al. (2000) suggested that an alternative distribution of hydrophobic and hydrophilic sites on the peptides is necessary for emulsification. van der Ven et al. (2001) suggested that the emulsifying capacity appears to be unrelated to molecular weight or degree of hydrolysis but that emulsion stabilization properties were related to molecular weight, with peptides of >2 kDa being required to impart stability. In the case of foam formation, hydrolysates with few large hydrophobic peptides were required for rapid diffusion to the interface and for stabilization of the bubble (Rahali and Gueguen, 2000). For foam stability, whey protein hydrolysates required a sufficient amount of >3 kDa peptides while for casein hydrolysates peptides of >7 kDa were desirable (van der Ven et al., 2002).

2. Effects on gelling properties Hydrolysis of proteins can be used to manipulate gel properties of whey proteins (Foegeding et al., 2002). The hydrolysis of whey proteins (>18%) can lead to gel formation (Doucet et al., 2001). Gelation was attributed to

22

Mary Ann Augustin and Punsandani Udabage

the small molecular weight peptides that were held together by noncovalent interactions (Doucet et al., 2003). The treatment of whey proteins with a protease from B. licheniformis has been shown to induce gelation of both unheated and heated whey proteins. Increasing the degree of hydrolysis resulted in earlier gelation and increases in gel firmness (Ju et al., 1997). As with intact whey protein gels, the properties of gels made from whey protein hydrolysates varied with pH, but this was dependent on whether the whey proteins were denatured. Gels made with hydrolysates were also sensitive to the presence of salts, with increasing salt concentration leading to more coagulum-like gels (Otte et al., 1999).

C. Transglutamination Transglutaminase (EC2.3.2.13) catalyzes an acyl-transfer reaction and this results in the formation of cross-links between glutamine and lysine residues. The introduction of new cross-links has important consequences for the functionality of proteins. Many aspects have been covered in a review on the use of transglutaminase in milks and dairy products, which show that it can be used to improve various functional properties (Jaros et al., 2006). Selected aspects are highlighted below.

1. Effects on heat stability Transglutaminase treatment of milk offers a novel way to improve the heat stability of milks without the use of chemical additives. Transglutaminase-treated milk had markedly improved heat stability at pH > 6.5 compared to untreated milk. This may be related to the effect of intramolecular cross-links formed in transglutaminase-treated milk, which prevents the dissociation of caseins from the micelles under conditions where it would have otherwise occurred (e.g., when colloidal calcium phosphate is removed). This was considered to be the mechanism by which the enzyme-treated milk was stabilized to heat treatment (O’Sullivan et al., 2002a,b). The treatment of micellar casein dispersions altered pH-heat stability profiles. At a pH up to 6.45, there was a negligible effect on heat stability, but stability to heat treatment at 140  C was markedly improved when pH was increased to 7.1. In these systems, there was minimal intermolecular cross-linking between micelles (Mounsey et al., 2005).

2. Effects on water-binding and gelling properties Transglutaminase-treated cross-linked milk that was subsequently acidified with glucono-d-lactone formed significantly firmer gels, which had a finer protein network than untreated acid milk gels (Faergemand and Qvist, 1997). Treatment of casein micelles with transglutaminase by

Dairy Protein Functionality

23

addition of the enzyme to milk containing glucono-d-lactone results in firmer gels with superior water-binding properties compared to acid gels made from untreated micellar casein. Different gel structures in transglutaminase-treated micellar dispersions can be manipulated by controlling the extent of intra- and intermolecular cross-linking of micelles and the degree of disaggregation of the micelles (Schorsch et al., 2000a,b). The use of transglutaminase increases the strength of acidified whey gels. This was achieved by first cross-linking whey proteins at high pH (7–8) using transglutaminase, followed by cold-set acidification with glucono-d-lactone to low pH (4). Although the cold-set gels made with enzyme-treated whey proteins were less homogenous than those made with untreated whey proteins, the enzyme-treated gels were much firmer. This was attributed to the formation of additional cross-links between enzyme-treated whey proteins (Eissa and Khan, 2005; Eissa et al., 2004).

IV. CHEMICAL MODIFICATION PROCESSES It is well established that chemical modification of proteins, such as acylation, succinylation, esterification, chemical hydrolysis, and phosphorylation, cause changes in the physical properties of proteins and their digestibility. Chemical agents have generally been used for the synthesis of chemical-modified proteins. However, there are opportunities to use the Maillard reaction, a natural reaction that occurs on heat treatment of food, for covalently attaching sugars and polysaccharides with reducing sugar groups to a protein.

A. Use of chemical agents Most chemical agents used for studying the chemical modification of proteins are not suitable for food applications. These studies nevertheless demonstrate how changes in amino acid side chains, and both the structure and conformation of proteins can impact on functionality. A comprehensive review of chemical modification of milk proteins has been carried out (Chobert, 2003). Only some highlights and more recent work on modification with chemical agents are covered here.

1. Acylation Acylation (e.g., acetylation and succinylation) modifies the charge of proteins. When succinylation is carried out, positive amino groups are replaced by negative succinyl groups, inducing a greater increase in negative charge compared to acetylation where the amino groups are replaced by neutral acetyl groups.

24

Mary Ann Augustin and Punsandani Udabage

The use of the modified proteins has shown the importance of electrostatic interactions in the formation of milk protein-based gels. Succinylation of milk affects the rennet coagulation time and the rate of firming of the coagulum. While some have ascribed the effects to the impaired rate of the primary stage of renneting process, as reflected in the slower release of the caseinomacropeptide (Lieske et al., 2000), others have considered that succinylation affected the secondary stage of aggregation by increasing electrostatic repulsions (Vidal et al., 1998). The pH at which cold-set acid whey gels are formed decreases with succinylation. Unmodified b-lactoglobulin starts to aggregate and gel at a pH of 5.1 (near its pI), while succinylated forms of the protein only gel at pH 2.5. This contrasts with the effects of methylation of carboxylic acid groups which removes negative charge, resulting in gelation at alkaline pH (Alting et al., 2002). Acylation affects the casein micelles of milk. The main effects are increased dissolution of the calcium and phosphate from the micelle and increased solubilization of caseins as a consequence of acylation (Vidal et al., 2002). As the equilibria of caseins between the micellar and serum phases are known to affect a number of functional properties (e.g., gelation, emulsification), it may be expected that acylation will affect functionality.

2. Esterification with alcohols As natural milk proteins have acidic isoelectric points, they have low solubility at acid pH and this compromises many of their functional properties in acidic environments. Esterification of proteins increases the net negative charge and raises the isoelectric point of proteins, making them more functional at acidic pHs. Esterification of milk proteins (b-lactoglobulin, a-lactalbumin, b-casein) with methanol, ethanol, or propanol improves their solubility in the pH range 3–6 and emulsifying activity and stability at low pH (3–5). The extent of improvement was dependent on the degree of esterification, the type of ester group attached, and the nature of the milk protein (Sitohy et al., 2001a). Esterification also resulted in reduced digestibility of the proteins by trypsin (Sitohy et al., 2001b).

3. Phosphorylation Caseins have natural ester-bound phosphate and this gives caseins some of their unique properties. Whey proteins do not naturally contain phosphate ester groups. Studies have shown that the phosphorylation of caseins and whey proteins creates novel functionality in these proteins. Van Hekken and Strange (1997) phosphorylated whole caseins using POCl3. Solutions (0.2–0.7% protein) containing the superphosphorylated caseins, with higher amounts of bound phosphorus (9- to 12.5-mmole

Dairy Protein Functionality

25

bound P/mmole casein) compared to unmodified caseins (5.6-mmole bound P/mmole casein), were more resistant to thickening in high Ca2+ solutions at low protein concentrations. At higher protein concentrations (1–4% protein), they formed gels at high Ca2+ concentrations (20–30 mM). Li et al. (2005) phosphorylated whey protein isolate by dry heating in the presence of pyrophosphate. Phosphorylation improved the stability of the whey protein to heat at pH 7. Gels made with phosphorylated proteins were firmer, more resilient, and had better water-holding capacity compared to untreated whey protein isolate gels.

B. Maillard reaction The Maillard reaction is a complex series of reactions that begins with the interaction of an amino group with a reducing sugar group. The reaction is well known for its effects on the physical properties of food, particularly the color and flavor of foods (Fayle and Gerrard, 2002; Nursten, 2005). The Maillard reaction can occur naturally as in the production or storage of milk powders (Guyomarc’h et al., 2000). Under these conditions, the extent of the Maillard reaction is not controlled and is considered to have detrimental effects on powder quality. However, under controlled conditions, it has potential to be used for production of tailored dairy ingredients. The changes in the structure of the protein on conjugation of sugars or polysaccharides under controlled conditions give rise to the development of differentiated functionalities, which are useful when the modified protein is used in ingredient applications (Kato, 2002; Oliver et al., 2006a). There has been much recent interest in the use of glycation for modification of proteins as it is a naturally occurring reaction in foods. It is viewed as an attractive alternative to modification of proteins compared to the use of chemical agents. The type of carbohydrates and proteins used and the conditions of the reaction have to be controlled to optimize functionality while minimizing excessive browning and formation of other undesirable products which can be obtained in the final stages of the Maillard reaction. Various researchers have examined the effects of type of protein and sugar or carbohydrate, amounts of reactants, and conditions of reaction on a range of functional properties of milk proteins (Chevalier et al., 2001a,b).

1. Conditions for preparation of Maillard conjugates The Maillard reaction can occur under wet conditions in solutions or in the powdered state in humidified atmospheres [typically 60–80% relative humidity (RH)]. Studies have shown that when the Maillard reaction is carried out in the powdered state in humidified atmospheres (‘‘dry’’ reaction, 65% RH, 50  C, 2–48 hour), the structure of the whey proteins

26

Mary Ann Augustin and Punsandani Udabage

are not significantly altered, whereas there are significant structural changes in proteins when a reaction is carried out in aqueous systems (60  C, 6–130 hour, pH 7.2) (Morgan et al., 1999).

2. Effects on solubility and heat stability The covalent attachment of a sugar or a carbohydrate with a reducing sugar end to the free amino groups of a protein causes a loss in positive charge. This results in a change in the solubility profile of the protein as a function of pH and heat treatment. Improved solubility at low pH was obtained on conjugation with casein with maltodextrin under ‘‘dry’’ conditions (Shepherd et al., 2000). Compared to unreacted protein, b-lactoglobulin that was glycated with sugars (arabinose, galactose, glucose, lactose, rhamose, or lactose) at 60  C (aqueous systems, pH 7.2, 72 hour, anaerobic conditions) was more soluble at acidic pH and more stable to heating at pH 5 (Chevalier et al., 2001a). These studies demonstrate the usefulness of the Maillard reaction for enabling dairy proteins to have differentiated properties compared to the unmodified proteins.

3. Effects on surface properties There has been a significant interest in the use of the Maillard reaction for improving the emulsifying properties of proteins. The introduction of a sugar or polysaccharide group changes the charge on the protein. This has an impact on its emulsifying capacity and solubility. When a polysaccharide is used, it has the added advantage of imparting increased stability. This is because of the coupling of the steric stabilizing influence of the polysaccharide to the surface-activity of the proteins. Shepherd et al. (2000) showed that conjugation of caseins with maltodextrins improved the emulsifying capacity and stability of caseins at low pH. Darewicz and Dziuba (2001) observed improved emulsifying capacity and stability in glycated b-casein in aqueous systems (37  C, pH 7.4, 24 hour). This was related to better solubility of the glycated protein and to its ability to form thicker layers around the oil droplets. However, there was no change in emulsifying properties when b-casein was glycated in aqueous systems at lower pH and temperature for longer times (60  C, pH 6.5, 72 hour) (Groubet et al., 1999). Chevalier et al. (2001a) glycated b-lactoglobulin in aqueous systems and obtained improvement in the foaming and emulsifying properties, but the improvement obtained depended on the type of sugar used. This was attributed to the differences in the site of glycation with the different sugars used (Chevalier et al., 2001b). The conjugation of protein with polysaccharides has been examined. The emulsifying properties of whey protein isolate that was conjugated to low methoxy pectins under ‘‘dry’’ heat conditions had superior emulsion

Dairy Protein Functionality

27

stabilization properties at pH 5.5. This modification allows whey protein isolates to be used in acidic conditions (Neirynck et al., 2004). Others have compared the effects of type of milk protein and pectin. Einhorn-Stoll et al. (2005) formed milk protein (casein or whey protein)–pectin (low or high methoxy pectin) conjugates under ‘‘dry’’ reaction conditions (50–60  C, 65–80% RH, pH 5.8–7.0, up to 15 days). In the systems they examined, the conjugate made with whey protein isolate and high methoxy pectin was the best emulsifier. Caseinate was found to be an unsuitable substrate for effective conjugation and it was suggested that the thermodynamic incompatibility between caseins and pectins contributed to poor conjugate formation. Maillard conjugates made by interaction of milk proteins with sugars have been shown to enhance the delivery of omega-3 oils. In this application, the good emulsifying properties of the Maillard conjugate and the inbuilt antioxidant activity of the Maillard products enable the production of high-fat tuna oil powders (50% fat) with improved shelf life stability (Augustin et al., 2006).

4. Effects on viscosity and gelation Another consequence of the Maillard reaction on functionality is a modification of viscosity. Oliver et al. (2006b) found that glycoconjugates of casein with inulin and reducing sugars had higher viscosity compared to unmodified casein. Although high viscosity was obtained when reducing sugars (ribose or glucose) were conjugated with casein, this was accompanied by excessive browning. With the use of inulin in combination with fructose, it was possible to markedly increase viscosity without either gelation or excessive browning. As the Maillard reaction progresses, there can be cross-linking between protein species and the formation of polymeric species. The formation of these species is expected to further modify the viscosity and gelation properties of the proteins.

V. EMERGING PROCESSES There are emerging food processing technologies that have the potential for altering the functionality of milk and dairy products. These include static high-pressure processing, dynamic high-pressure processing, ultrasound, pulsed-electric field, and microwave heating. The technology that has attracted the most interest in the dairy industry to date is static high-pressure processing. Many studies have examined the use of high pressure processing for inactivation of microflora. However, it has the potential to alter the physical and technological properties of milk, making it an alternative to other processing methods

28

Mary Ann Augustin and Punsandani Udabage

for altering the functionality of milk proteins. Much fundamental research has been carried out to show that high pressure (100–600 MPa) causes significant changes to the milk. Notable among these, which can impact on milk functionality, are the disruption of casein micelles, the denaturation of whey proteins, and pressure-induced changes to the mineral equilibria of milk. Some of the consequences of the high-pressure treatment include a reduced rennet coagulation time, an increased cheese yield, and an increase in the firmness of yoghurts (Huppertz et al., 2002, 2006; Lo´pez-Fanˇdino, 2006). Dynamic high pressure, at pressures higher than those used in conventional dairy processing, has also been examined. The use of higher pressure homogenizers (e.g., Microfluidizer or Emulsiflex equipment) in place of conventional homogenizers results in a smaller size distribution of fat globules and changes to the organization of milk protein components (Dalgleish et al., 1996; Paquin, 1999). Recent work has shown that high homogenization pressures (41–186 MPa) cause changes to the structural properties of casein micelles. There was a decrease in micelle size and an increase in the amount of nonsedimentable caseins in the serum (Sandra and Dalgleish, 2005). The structural changes in the casein micelle are expected to have consequences for some of the functional properties of milks. Further work is needed to ascertain the nature and extent of these effects. Hardham et al. (2000) showed that UHT milk treated by microfluidization has adequate heat stability and further that creaming of the milk on storage was reduced. Whiteley and Muir (1996) found that microfluidization was effective for reducing particle size of concentrated milks. Surprisingly, the heat stability of the microfluidized milk was also markedly improved. Further work is required to understand the effects of microfluidization on heat stability. The use of microfluidization as a means to improve the heat stability of whey proteins has been investigated (Iordache and Jelen, 2003). These authors found that microfluidization (150 MPa) of heated whey protein concentrate suspensions disintegrated the insoluble particles in these solutions to nonsedimenting particles. However, these particles that were resolubilized by microfluidization were still sensitive to secondary heat-induced coagulation. Ultrasound may be used for disruption of fat globules in milk and is an alternative to homogenization for this purpose. It also has the potential to alter the functionality of milk, as demonstrated by its effects on the properties of yoghurt (Vercet et al. 2002). These authors showed that the simultaneous application of heat and ultrasound (12 s at 20 kHz) under moderate pressure (2-kg pressure) improved the textural properties of yoghurts. The effects of high pressure and ultrasound as well as other emerging processing technologies such as pulse-electric field and microwave heating

Dairy Protein Functionality

29

on the properties of milk and their impact on the functional properties of milk need to be examined in more detail. All these processing treatments result in a stress being applied to the milk system, which will no doubt result in changes to functionality. More research is required to examine the extent to which these processing technologies can be used as an alternative to or in combination with traditional processes (e.g., heating, homogenization, and acidification) to alter functionality of milk and dairy ingredients.

VI. CONCLUSION Milk and dairy ingredients are valued ingredients in the market place. The functionality of the protein components in milk makes them useful in a range of food applications. Understanding the inherent properties of the individual milk proteins, their interactions with other milk components and also with other components in the final food matrix during processing, is essential for the design of ingredients. The need to deliver nutrients while maintaining physical functionality represents an added challenge to the design of dairy systems for the delivery of bioactives. The new functionalities that can be achieved by various processing treatments will need to be reexamined when dairy foods are used as delivery vehicles for bioactives. This is because the balance of the components in the system is changed, and this adds a layer of complexity to both formulation and processing. The ability to expand and combine the traditional processes and emerging technologies is an opportunity to meet the demands of providing dairy foods that have sensory appeal and also deliver the health benefits of the added nutrients. Continued research into the fundamental aspects of milk proteins and the responses to various stresses is necessary for further differentiation of milk products and the delivery of ingredients with consistent quality for target applications. New approaches are required to meet the challenges of designing fitness-for-purpose dairy ingredients. A more complex science approach to complement the traditional reductionist approach to dairy ingredient development may provide further insights into the interactions of ingredients in food formulations and processing environments. A multidisciplinary approach, bringing together traditional dairy and food scientists with scientists from other disciplines such as materials science, molecular science, nanotechnology, and the science of complex systems, is desirable. Methods that enable an examination of dairy and food systems in real time and on different length scales are expected to provide new information about the relationship between functionality and organization of food components into supramolecular and higher hierarchical structures. This information may then be used as the basis for

30

Mary Ann Augustin and Punsandani Udabage

new processing and formulation strategies to engineer step-changes in the development of new dairy ingredients.

ACKNOWLEDGMENT Ms. Christine Margetts is gratefully acknowledged for assistance in sourcing information and helpful comments on the chapter.

REFERENCES American Dairy Products Institute (1990). ‘‘Standards for Grades of Dry Milks Including Methods of Analysis’’. Bulletin 916, ADPI, Chicago. Anema, S.G., and Li, Y. (2003). Effect of pH on the association of denatured whey proteins with casein micelles in heated reconstituted skim milk. J. Agric. Food Chem. 51, 1640–1646. Anema, S.G., Lee, S.K., Lowe, E.K., and Klostermeyer, H. (2004). Rheological properties of acid gels prepared from heated pH-adjusted skim milk. J. Agric. Food Chem. 52, 337–343. Alting, A.C., de Jongh, H.H., Visschers, R.W., and Simons, J.W.F.A. (2002). Physical and chemical interactions in cold gelation of food proteins. J. Agric. Food Chem. 50, 4682–4689. Alting, A.C., Hamer, R.J., Kruif de, C.G., Paques, M., and Visschers, R.W. (2003). Number of thiol groups rather than the size of the aggregates determines the hardness of cold set whey protein gels. Food Hydrocolloids 17, 469–479. Augustin, M.A. (2000). Mineral salts and their effect on milk functionality. Aust. J. Dairy Technol. 55, 61–64. Augustin, M.A. (2004). Scientific insights leading to innovative dairy ingredients. NZ Food J. 4(5), 23–29. Augustin, M.A., and Clarke, P.T. (1990). Effects of added salts on the heat stability of recombined concentrated milk. J. Dairy Res. 57, 213–226. Augustin, M.A., and Versteeg, C. (2006). Milkfat: Physical, chemical and enzymatic modification. In ‘‘Advances in Dairy Chemistry—Lipids’’ (P.F. Fox and P. McSweeney, eds.), 2nd Ed., pp. 293–325. Springer, New York. Augustin, M.A., Cheng, L.J., and Clarke, P.T. (1999). Effects of preheat treatment of milk powder on the properties of reconstituted set skim yogurts. Int. Dairy J. 9, 415–416. Augustin, M.A., Sanguansri, L., and Bode, O. (2006). Maillard reaction products as encapsulants for fish oil powders. J. Food Sci. 71(2), E25–E32. Augustin, M.A., Cheng, L.J., Glagovskaia, O., Clarke, P.T., and Lawrence, A. (2003). Use of blends of skim milk and sweet whey protein concentrates in reconstituted yogurt. Aust. J. Dairy Technol. 58, 30–35. Banon, S., and Hardy, J. (1991). Study of acid milk coagulation by an optical method using light reflection. J. Dairy Res. 58, 75–84. Becker, T., and Puhan, Z. (1989). Effect of different processes to increase the milk solids nonfat content on the rheological properties of yoghurt. Milchwissenschaft 44, 626–629. Bhaskar, G.V., Singh, H., and Blazey, N.D. (2003). Milk protein products and processes. US Patent US 2003/0096036-A1. Bottcher, S.R., and Foegeding, E.A. (1994). Whey protein gels: Fracture stress and strain and related microstructural properties. Food Hydrocolloids 82, 113–123. Britten, M., and Giroux, H.J. (2001). Acid-induced gelation of whey protein polymers: Effects of pH and calcium concentration during polymerization. Food Hydrocolloids 15, 609–617. Bryant, C.M., and McClements, D.J. (2000). Optimizing preparation conditions for heatdenatured whey protein solutions to be used as cold-gelling ingredients. J. Food Sci. 65, 259–263.

Dairy Protein Functionality

31

Carr, A.J. (2002). Milk protein concentrate products and the uses thereof. International Patent Application WO02/196208A2. Casiraghi, E.M., Peri, C., and Piazza, L. (1987). Effect of calcium equilibria on the rate of syneresis and on the firmness of curds obtained from milk UF retentates. Milchwissenschaft 42, 232–235. Chantrapornchai, W., and McClements, D.J. (2002). Influence of NaCl on optical properties, large-strain rheology and water holding capacity of heat-induced whey protein isolate gels. Food Hydrocolloids 16, 467–476. Chen, S.A. (2003). Modification of foaming properties of proteins. United States Patent US651636B2. Chevalier, F., Chobert, J.-M., Poineau, Y., Nicolas, M.G., and Haertle, T. (2001a). Improvement of functional properties of b-lactoglobulin glycated through the Maillard reaction is related to the nature of the sugar. Int. Dairy J. 11, 145–152. Chevalier, F., Chobert, J.-M., Molle, D., and Haertle, T. (2001b). Maillard glycation of betalactoglobulin with several sugars: Comparative study of the properties of the obtained polymers and of the substituted sites. Lait 81, 651–666. Chobert, J.-M. (2003). Milk protein modification to improve the functional and biological properties. Adv. Food Nutr. Res. 47, 1–71. Clarke, P.T., and Augustin, M.A. (2005). Manipulation of solvent-extractable fat content of milk powders. Lait 84, 261–268. Corredig, M., and Dalgleish, D.G. (1996). Effect of temperature and pH on the interactions of whey proteins with casein micelles in skim milk. Food Res. Int. 29, 49–55. Dalgleish, D.G. (1979). Proteolysis and aggregation of casein micelles treated with immobilized or soluble chymosin. J. Dairy Res. 46, 653–661. Dalgleish, D.G. (1992). The enzymic coagulation of milk. In ‘‘Advanced Dairy Chemistry’’ (P.F. Fox, ed.), Vol. 1, pp. 579–619. Elsevier Applied Science, London. Dalgleish, D.G., and Law, A.J.R. (1988). pH-Induced dissociation of bovine casein micelles. I. Analysis of liberated caseins. J. Dairy Res. 55, 529–538. Dalgleish, D.G., and Law, A.J.R. (1989). pH-Induced dissociation of bovine casein micelles. II. Mineral solubilization and its relation to casein release. J. Dairy Res. 56, 727–735. Dalgleish, D.G., and Singh, A.M. (1998). The emulsifying properties of hydrolyzates of whey proteins. J. Dairy Sci. 81, 918–924. Dalgleish, D.G., Tosh, S.M., and West, S. (1996). Beyond homogenization: The formation of very small emulsion droplets during the processing of milk by a microfluidizer. Neth. Milk Dairy J. 50, 135–148. Dannenberg, F., and Kessler, H.G. (1988a). Effect of denaturation of b-lactoglobulin on texture properties of set-style nonfat yoghurt. I. Syneresis. Milchwissenschaft 43, 632–635. Dannenberg, F., and Kessler, H.G. (1988b). Effect of denaturation of b-lactoglobulin on texture properties of set-style nonfat yoghurt. II. Firmness and flow properties. Milchwissenschaft 43, 700–704. Darewicz, M., and Dziuba, J. (2001). The effect of glycosylation on emulsifying and structural properties of bovine beta-casein. Nahrung 45, 15–20. de Castro-Morel, M., and Harper, W.J. (2003). Effect of retentate heat treatment and spray dryer inlet temperature on the properties of milk protein concentrates (MPC’s). Milchwissenschaft 58, 13–15. de Rham, O., and Chanton, S. (1984). Role of ionic environment in insolubilization of whey protein during heat treatment of whey products. J. Dairy Sci. 67, 939–949. Dickinson, E., and Parkinson, E.L. (2004). Heat-induced aggregation of milk protein-stabilized emulsions: Sensitivity to processing and composition. Int. Dairy J. 14, 635–645. Donovan, M., and Mulvihill, D.M. (1987). Thermal denaturation and aggregation of whey proteins. Ir. J. Food Sci. Technol. 11, 87–100.

32

Mary Ann Augustin and Punsandani Udabage

Doucet, D., Gauthier, S.F., and Foegeding, E.A. (2001). Rheological characterization of a gel formed during extensive enzymatic hydrolysis. J. Food Sci. 66, 711–715. Doucet, D., Otter, D.E., Gauthier, S.F., and Foegeding, E.A. (2003). Enzyme-induced gelation of extensively hydrolyzed whey proteins by alcalase: Peptide identification and determination of enzyme specificity. J. Agric. Food Chem. 51, 6300–6308. Einhorn-Stoll, U., Ulbrich, M., Sever, S., and Kunzek, H. (2005). Formation of milk proteinpectin conjugates with improved emulsifying properties by controlled dry heating. Food Hydrocolloids 19, 329–340. Eissa, A., and Khan, S.A. (2005). Acid-induced gelation of enzymatically modified, preheated whey proteins. J. Agric. Food Chem. 53, 5010–5017. Eissa, A.S., Bisram, S., and Khan, S.A. (2004). Polymerization and gelation of whey protein isolates at low pH using transglutaminase enzyme. J. Agric. Food Chem. 52, 4456–4464. Euston, S.R., Finnigan, S.R., and Hirst, R.L. (2001). Heat-induced destabilization of oil-inwater emulsions formed from hydrolyzed whey protein. J. Agric. Food Chem. 49, 5576–5583. Faergemand, M., and Qvist, K.B. (1997). Transglutaminase: Effect on rheological properties, microstructure and permeability of set style acid skim milk gel. Food Hydrocolloids 11, 287–292. Fachin, L., and Viotto, W.H. (2005). Effect of pH and heat treatment of cheese whey on solubility and emulsifying properties of whey protein concentrate produced by ultrafiltration. Int. Dairy J. 15, 325–332. Fayle, S.E., and Gerrard, J.A. (2002). ‘‘The Maillard Reaction’’. Royal Society of Chemistry, Cambridge, 120 pages. Firebaugh, J.D., and Daubert, C.R. (2005). Emulsifying and foaming properties of a derivatized whey protein ingredient. Int. J. Food Prop. 8, 243–253. Foegeding, E.A., Davis, J.P., Doucet, D., and McGuffey, M.K. (2002). Advances in modifying and understanding whey protein functionality. Trends Food Sci. Technol. 13, 151–159. Fox, P.F. (1989). Heat-induced changes in milk. In ‘‘Bulletin of the International Dairy Federation No. 238’’ (P.F. Fox, ed.). International Dairy Federation, Brussels. Fox, P.F., and Mulvihill, D.M. (1990). Casein. In ‘‘Food Gels’’ (P. Harris, ed.), pp. 121–171. Elsevier Applied Science, London. Giardina, C., Pelizzola, V., Avalli, A., Iametti, S., and Cattaneo, T.M.P. (2004). Functional properties of milk protein hydrolysates obtained by controlled enzymatic hydrolysis. Milchwissenschaft 59, 476–479. Glibowski, P., Mleko, S., and Wesolowska-Trojanowska, M. (2006). Gelation of single heated vs. double heated whey protein isolate. Int. Dairy J. 16, 1113–1118. Green, M.L. (1987). Effect of manipulation of milk composition and curd-forming conditions on the formation, structure and properties of milk curd. J. Dairy Res. 54, 303–313. Griffin, M.C.A., Lyster, R.L.J., and Price, J.C. (1988). The disaggregation of calcium- depleted casein micelles. Eur. J. Biochem. 174, 339–343. Groubet, R., Chobert, J.-M., and Haertle, T. (1999). Functional properties of milk proteins glycated in mild conditions. Sci. Aliments 19, 423–438. Guo, M., Fox, P.F., Flynn, A., and Mahammad, K.S. (1989). Heat-induced changes in sodium caseinate. J. Dairy Res. 56, 503–512. Guyomarc’h, F., Warin, F., Muir, D.D., and Leaver, J. (2000). Lactosylation of milk proteins during the manufacture and storage of skim milk powders. Int. Dairy J. 10, 863–872. Guyomarc’h, F., Queguiner, C., Law, A.J.R., Horne, D.S., and Dalgleish, D.G. (2003). Role of the soluble and micelle-bound heat-induced protein aggregates on network formation in acid skim milk gels. J. Agric. Food Chem. 51, 7743–7750. Hardham, J.F., Imison, B.W., and French, H.M. (2000). Effect of homogenisation and microfluidisation on the extent of fat separation during storage of UHT milk. Aust. J. Dairy Technol. 55, 16–22.

Dairy Protein Functionality

33

Havea, P. (2006). Protein interactions in milk protein concentrate powders. Int. Dairy J. 16, 415–422. Holt, C. (1982). Inorganic constituents of milk. III. The colloidal calcium phosphate of cow’s milk. J. Dairy Res. 49, 29–38. Holt, C. (1995). Effect of heating and cooling on the milk salts and their interaction with casen. In ‘‘Heat-Induced Changes in Milk’’ (P.F. Fox, ed.), pp. 105–133. International Dairy Federation, Brussels. Horne, D.S., and Davidson, C.M. (1992). Direct observation of decrease in size of casein micelles during the initial stages of renneting of skim milk. Int. Dairy J. 3, 61–71. Hudson, H.M., Daubert, C.R., and Foegeding, E.A. (2000). Rheological and physical properties of derivitized whey protein powders. J. Agric. Food Chem. 48, 3112–3119. Huppertz, T., Kelly, A.L., and Fox, P.F. (2002). Effects of high pressure on constituents and properties of milk. Int. Dairy J. 12, 561–572. Huppertz, T., Smiddy, M.A., Upadhyay, V.K., and Kelly, A.L. (2006). High-pressure-induced changes in bovine milk: A review. Int. J. Dairy Technol. 59, 58–66. Iordache, M., and Jelen, P. (2003). High pressure microfluidization treatment of heat denatured whey proteins for improved functionality. Innovtive Food Sci. Emerging Technol. 4, 367–376. Jaros, D., Partschefeld, C., Henle, T., and Rohm, H. (2006). Transglutaminase in dairy products: Chemistry, physics, applications. J. Texture Stud. 37, 113–155. Jean, K., Renan, M., Famelart, M.H., and Guyomarc’h, F. (2006). Structure and surface properties of the serum heat-induced protein aggregates isolated from heated skim milk. Int. Dairy J. 16, 303–315. Ju, Z.Y., Otte, J., Zakora, M., and Qvist, K.B. (1997). Enzyme-induced gelation of whey proteins: Effect of protein denaturation. Int. Dairy J. 7, 71–78. Kato, A. (2002). Industrial applications of Maillard-type protein-polysaccharide conjugates. Food Sci. Technol. Res. 8, 193–199. Kelly, J., Kelly, P.M., and Harrington, D. (2002). Influence of processing variables on the physicochemical properties of spray dried fat-based milk powders. Lait 82, 401–412. Kelly, P.M. (2006). Innovation in milk powder technology. Int. J. Dairy Technol. 59, 70–75. Kelly, P.M., and Burgess, K.J. (1978). Foaming properties of milk protein concentrate prepared by ultrafiltration. Ir. J. Food Sci. Tech. 2, 93–104. Keogh, K., Twomey, M., O’Kennedy, B., and Mulvihill, D. (2002). Effect of milk composition on spray-dried high-fat milk powders and their use in chocolate. Lait 82, 531–539. Keogh, K., Murray, C., Kelly, J., and O’Kennedy, B. (2004). Effect of the particle size of spraydried milk powder on some properties of chocolate. Lait 84, 375–384. Kilara, A., and Panyam, D. (2003). Peptides from milk proteins and their properties. Crit. Rev. Food Sci. Nutr. 43, 607–633. Kinsella, J.E. (1984). Milk proteins: Physicochemical and functional properties. Crit. Rev. Food Sci. Nutr. 21, 197–262. Konrad, G., Kleinschmidt, T., Rohenkohl, H., and Reimerdes, E.H. (2005). Peptic partial hydrolysis of whey protein concentrate for modifying the surface properties of whey protein. II. Effects on the emulsifying and foaming properties. Milchwissenschaft 60, 195–198. Korhonen, H., and Pihlanto, A. (2006). Bioactive peptides: Production and functionality. Int. Dairy J. 16, 945–960. Langton, M., and Hermansson, A.M. (1992). Fine-stranded and particulate gels of betalactoglobulin and whey protein at varying pH. Food Hydrocolloids 5, 523–539. Law, A.J.R., and Leaver, J. (2000). Effect of pH on the thermal denaturation of whey proteins in milk. J. Agric. Food Chem. 48, 672–679. Li, C.P., Enomoto, H., Ohki, S., Ohtomo, H., and Aoki, T. (2005). Improvement of functional properties of whey protein isolate through glycation and phosphorylation. J. Dairy Sci. 88, 4137–4145.

34

Mary Ann Augustin and Punsandani Udabage

Lieske, B., and Konrad, G. (1993). Functional changes in whey protein caused by microparticulation—using Simplesse 100 as example. Deutsche Milchwirtschaft 44, 1252–1256. Lieske, B., Konrad, G., and Faber, W. (2000). Effects of succinylation on the renneting properties of raw milk. Milchwissenschaft 55, 71–74. Lin, S.H.C., Leong, S.L., Dewan, R.K., Bloomfield, V.A., and Morr, C.V. (1972). Effect of calcium ion on the structure of native bovine casein micelles. Biochemistry 11, 1818–1821. Lomholt, S.B., and Qvist, K.B. (1997). Relationship between rheological properties and degree of k-casein proteolysis during renneting of milk. J. Dairy Res. 64, 541–549. Lo´pez-Fanˇdino, R. (2006). Functional improvement of milk whey proteins induced by high hydrostatic pressure. Crit. Rev. Food Sci. Nutr. 46, 351–363. Lucey, J.A., Teo, C.T., Munro, P.A., and Singh, H. (1997). Rheological properties at small (dynamic) and large (yield) deformations of acid gels made from heated milk. J. Dairy Res. 64, 591–600. Lucey, J.A., Tamehana, M., Singh, H., and Munro, P.A. (1998). Effect of interactions between denatured whey proteins and casein micelles on the formation and rheological properties of acid skim milk gels. J. Dairy Res. 65, 555–567. Mangino, M.E. (1992). Gelation of whey protein concentrates. Food Technol. 46(1), 114, 116–117. McMahon, D.J., Yousif, B.H., and Kalab, M. (1993). Effect of whey protein denaturation on structure of casein micelles and their rennetability after ultra-high temperature processing of milk with or without ultrafiltration. Int. Dairy J. 3, 239–256. Menard, O., and Camier, B.G.F. (2005). Effect of heat treatment at alkaline pH on the rennet coagulation properties of skim milk. Lait 85, 515–526. Mizuno, R., and Lucey, J.A. (2005). Effects of emulsifying salts on the turbidity and calcium phosphate-protein interactions in casein micelles. J. Dairy Sci. 88, 3070–3078. Mleko, S., and Foegeding, E.A. (2000). pH induced aggregation and weak gel formation of whey protein polymers. J. Food Sci. 65, 139–143. Moon, B., and Mangino, M.E. (2004). The effect of preheating on functionality of whey protein concentrates. Milchwissenschaft 59, 294–297. Morgan, F., Leonil, J., Molle, D., and Bouhallab, S. (1999). Modification of bovine betalactoglobulin by glycation in a powdered state or in an aqueous solution: Effect on association behavior and protein conformation. J. Agric. Food Chem. 47, 83–91. Moro, A., Gatti, C., and Delorenzi, N. (2001). Hydrophobicity of whey protein concentrates measured by fluorescence quenching and its relation with surface functional properties. J. Agric. Food Chem. 49, 4784–4789. Mounsey, J.S., O’Kennedy, B.T., and Kelly, P.M. (2005). Comparison of re-micellised casein prepared from acid casein with micellar casein prepared by membrane filtration. Lait 85, 419–430. Mulvihill, D.M., and Donovan, M. (1987). Whey proteins and their thermal denaturation— A review. Ir. J. Food Sci. Tech. 11, 43–75. Mulvihill, D.M., and Fox, P.F. (1989). Physico-chemical and functional properties of milk proteins. In ‘‘Developments in Dairy Chemistry—4—Functional milk proteins’’ (P.F. Fox, ed.), pp. 131–172. Elsevier Applied Science, London. Neirynck, N., van der Meeren, P., Bayarri Gorbe, S., Dierckx, S., and Dewettinck, K. (2004). Improved emulsion stabilizing properties of whey protein isolate by conjugation with pectins. Food Hydrocolloids 18, 949–957. Newstead, D.F., and Baucke, A.G. (1983). Heat stability of recombined evaporated milk and reconstituted concentrated skim milk: Effects of temperature and time of preheating. N. Z. J. Dairy Sci. Technol. 18, 1–11. Newstead, D.F., and Conaghan, E.F. (1978). Zur Hitzestabilitaet rekombinierter Kondensmilch—Einfluss der pH-Aenderung in der Magermilch vor der Verarbeitung. [Heat stability of recombined evaporated milk—effect of pH change in skim-milk before processing.]. Deutsche Molkerei Zeitung 99, 1688–1690.

Dairy Protein Functionality

35

Nursten, H. (2005). ‘‘The Maillard Reaction: Chemistry, Biochemistry and Implications’’. Royal Society of Chemistry, Cambridge, 214 pages. O’Connell, J.E., and Fox, P.F. (2001). Effect of beta-lactoglobulin and precipitation of calcium phosphate on the thermal coagulation of milk. J. Dairy Res. 68, 81–94. O’Kennedy, B.T., and Mounsey, J.S. (2006). Control of heat-induced aggregation of whey proteins using casein. J. Agric. Food Chem. 54, 5637–5642. O’Kennedy, B.T., Halbert, C., and Kelly, P.M. (2001). Formation of whey protein particles using calcium phosphate and their subsequent stability to heat. Milchwissenschaft 56, 625–628. O’Sullivan, M.M., Kelly, A.L., and Fox, P.F. (2002a). Effect of transglutaminase on the heat stability of milk: A possible mechanism. J. Dairy Sci. 85, 1–7. O’Sullivan, M.M., Kelly, A.L., and Fox, P.F. (2002b). Influence of transglutaminase treatment on some physico-chemical properties of milk. J. Dairy Res. 69, 433–442. Oestergaard, B. (2005). Microparticulation of whey. Deutsche Milchwirtschaft 56, 218–219. Okigbo, L.M., Richardson, G.H., Brown, R.J., and Ernstrom, C.A. (1985). Effects of pH, CaCl2, and chymosin concentration on coagulation properties of abnormal and normal milk. J. Dairy Sci. 68, 2527–2533. Oldfield, D.J., Singh, H., Taylor, M.W., and Pearce, K.N. (2000). Heat-induced interactions of beta-lactoglobulin and alpha-lactalbumin with the casein micelle in pH-adjusted skim milk. Int. Dairy J. 10, 509–518. Oliver, C.M., Melton, L.D., and Stanley, R.A. (2006a). Creating proteins with novel functionality via the Maillard reaction: A review. Crit. Rev. Food Sci. Nutr. 46, 337–350. Oliver, C.M., Melton, L.D., and Stanley, R.A. (2006b). Functional properties of caseinate glycoconjugates prepared by controlled heating in the dry state. J. Sci. Food Agric. 86, 732–740. Otte, J., Schumacher, E., Ipsen, R., Ju, Z.Y., and Qvist, K.B. (1999). Protease-induced gelation of unheated and heated whey proteins: Effects of pH, temperature, and concentrations of protein, enzyme and salts. Int. Dairy J. 9, 801–812. Panyam, D., and Kilara, A. (2004). Emulsifying peptides from the tryptic hydrolysis of casein. J. Food Sci. 69, FCT154–FCT163. Paquin, P. (1999). Technological properties of high pressure homogenizers: The effect of fat globules, milk proteins, and polysaccharides. Int. Dairy J. 9, 329–335. Philippe, M., Gaucheron, F., and Graet le, Y. (2004). Physico-chemical characteristics of calcium supplemented milk: Comparison of three calcium salts. Milchwissenschaft 59, 498–502. Philippe, M., Graet le, Y., and Gaucheron, F. (2005). The effects of different cations on the physicochemical characteristics of casein micelles. Food Chem. 90, 673–683. Pouliot, Y., and Boulet, M. (1991). Seasonal variations in the heat stability of concentrated milk: Effect of added phosphates and pH adjustment. J. Dairy Sci. 74, 1157–1162. Puvanenthiran, A., Williams, R.P.W., and Augustin, M.A. (2002). Structure and visco-elastic properties of set yoghurt with altered casein to whey protein ratios. Int. Dairy J. 12, 383–391. Rahali, V., and Gueguen, J. (2000). Foaming characteristics of chemical and enzymatic hydrolysates of bovine beta-lactoglobulin. Nahrung 44, 309–317. Rahali, V., Chobert, J.-M., Haertle, T., and Gueguen, J. (2000). Emulsification of chemical and enzymatic hydrolysates of beta-lactoglobulin: Characterization of the peptides adsorbed at the interface. Nahrung 44, 89–95. Rattray, W., and Jelen, P. (1996). Thermal stability of skim milk with protein content standardized by the addition of ultrafiltration permeates. Int. Dairy J. 6, 157–170. Rattray, W., and Jelen, P. (1997). Thermal stability of skim milk/whey protein solution blends. Food Res. Int. 30, 327–334. Remondetto, G.E., Paquin, P., and Subirade, M. (2002). Cold gelation of b-lactoglobulin in the presence of iron. J. Food Sci. 67, 586–595.

36

Mary Ann Augustin and Punsandani Udabage

Renan, M., Mekmene, O., Famelart, M.H., Guyomarc’h, F., Arnoult-Delest, V., Paquet, D., and Brule, G. (2006). pH-dependent behaviour of soluble protein aggregates formed during heat-treatment of milk at pH 6.5 or 7.2. J. Dairy Res. 73, 79–86. Resch, J.J., and Daubert, C.R. (2002). Rheological and physicochemical properties of derivatized whey protein concentrate powders. Int. J. Food Prop. 5, 419–434. Rodriguez del Angel, C., and Dalgleish, D.G. (2006). Structures and some properties of soluble protein complexes formed by the heating of reconstituted skim milk powder. Food Res. Int. 39, 472–479. Roefs, S.P.F.M., Van Vliet, T., Van Den Bijgaart, H.J.C.M., de Groot-Mostert, A.E.A., and Walstra, P. (1990). Structure of casein gels made by combined acidification and rennet action. Neth. Milk Dairy J. 44, 159–188. Rollema, H., and Brinkhuis, J.A. (1989). A 1H-NMR study of bovine casein micelles; influence of pH, temperature and calcium ions on micellar structure. J. Dairy Res. 56, 417–425. Rose, D. (1968). Relation between micellar and serum caseins in bovine milk. J. Dairy Sci. 51, 1897–1902. Sanchez, C., Pouloit, M., Renard, D., and Paquin, P. (1999). Uniaxial compression of thermal gels based on microfluidized blends of WPI and heat-denatured WPI. J. Agric. Food Chem. 47, 1162–1167. Sandra, S., and Dalgleish, D.G. (2005). Effects of ultra-high-pressure homogenization and heating on structural properties of casein micelles in reconstituted skim milk powder. Int. Dairy J. 15, 1095–1104. Schorsch, C., Carrie, H., Clark, A.H., and Norton, I.T. (2000a). Cross-linking casein micelles by a microbial transglutaminase conditions for formation of transglutaminase-induced gels. Int. Dairy J. 10, 519–528. Schorsch, C., Carrie, H., and Norton, I.T. (2000b). Cross-linking casein micelles by a microbial transglutaminase: Influence of cross-links in acid-induced gelation. Int. Dairy J. 10, 529–539. Schorsch, C., Wilkins, D.K., Jones, M.G., and Norton, I.T. (2001). Gelation of casein-whey mixtures: Effects of heating whey proteins alone or in the presence of casein micelles. J. Dairy Res. 68, 471–481. Schuck, P., Briard, V., Mejean, S., Piot, M., Famelart, M.H., and Maubois, J.L. (1999). Dehydration by desorption and by spray drying of dairy proteins; influence of the mineral environment. Drying Technol. 17, 1347–1357. Sharma, S.K., Mittal, G.S., and Hill, A.R. (1994). Effect of milk concentration, pH and temperature on k-casein hydrolysis at aggregation, coagulation and curd cutting times of ultrafiltered milk. Milchwissenschaft 49, 450–453. Shepherd, R., Robertson, A., and Ofman, D. (2000). Dairy glycoconjugate emulsifiers: Caseinmaltodextrins. Food Hydrocolloids 14, 281–286. Shirashoji, N., Jaeggi, J.J., and Lucey, J.A. (2006). Effect of trisodium citrate concentration and cooking time on the physicochemical properties of pasteurized process cheese. J. Dairy Sci. 89, 15–28. Singh, H. (2004). Heat stability of milk. Int. J. Dairy Technol. 57, 111–119. Singh, H., and Fox, P.F. (1985). Heat stability of milk: pH-dependent dissociation of micellar kappa-casein on heating milk at ultra high temperatures. J. Dairy Res. 52, 529–538. Singh, H., and Fox, P.F. (1987). Heat stability of milk: Role of beta-lactoglobulin in the pH-dependent dissociation of micellar kappa-casein. J. Dairy Res. 54, 509–521. Sitohy, M., Chobert, J.-M., and Haertle, T. (2001a). Improvement of solubility and of emulsifying properties of milk proteins at acid pHs by esterification. Nahrung 45, 87–93. Sitohy, M., Chobert, J.-M., and Haertle, T. (2001b). Susceptibility to trypsinolysis of esterified milk proteins. Int. J. Biol. Macromol. 28, 263–271.

Dairy Protein Functionality

37

Spiegel, T., and Huss, M. (2002). Whey protein aggregation under shear conditions-effects of pH-value and removal of calcium. Int. J. Food Sci. Tech. 37, 559–568. Storry, J.E., and Ford, G.D. (1982). Some factors affecting the post clotting development of coagulum strength in renneted milk. J. Dairy Res. 49, 469–477. Sweetsur, A.W.M., and Muir, D.D. (1982a). Natural variation in heat stability of concentrated milk before and after homogenization. J. Soc. Dairy Technol. 35, 120–126. Sweetsur, A.W.M., and Muir, D.D. (1982b). Manipulation of the heat stability of homogenized concentrated milk. J. Soc. Dairy Technol. 35, 126–132. Sweetsur, A.W.M., and Muir, D.D. (1983). Effect of homogenization on the heat stability of milk. J. Dairy Res. 50, 291–300. Tessier, H., and Rose, D. (1958). Calcium ion concentration in milk. J. Dairy Sci. 41, 351–359. Tolkach, A., and Kulozik, U. (2005). Effect of pH and temperature on the reaction kinetic parameters of thermal denaturation of b-lactoglobulin. Milchwissenschaft 60, 249–252. Tong, P.S. (2001). Milk powder functionality. Manufacturing Confectioner 81(9), 103–112. Twomey, M., Keogh, M.K., O’Kennedy, B.T., Auty, M., and Mulvihill, D.M. (2000). Effect of milk composition on selected properties of spray-dried high-fat and skim-milk powders. Ir. J. Agric. Food Res. 39, 79–94. Udabage, P., McKinnon, I.R., and Augustin, M.A. (2000). Mineral and casein Equilibria in Milk: Effects of added salts and calcium chelating agents. J. Dairy Res. 67, 361–370. Udabage, P., McKinnon, I.R., and Augustin, M.A. (2001). Effects of mineral salts and calcium chelating agents on the gelation of renneted skim milk. J. Dairy Sci. 84, 1569–1575. Underwood, J.W., and Augustin, M.A. (1997). Rheological characteristics of acid and heatinduced gels made from reconstituted or recombined concentrated milk. Aust. J. Dairy Technol. 52, 88–91. van der Ven, C., Gruppen, H., de Bont, D.B.A., and Voragen, A.G.J. (2001). Emulsion properties of casein and whey protein hydrolysates and the relation with other hydrolysate characteristics. J. Agric. Food Chem. 49, 5005–5012. van der Ven, C., Gruppen, H., de Bont, D.B.A., and Voragen, A.G.J. (2002). Correlations between biochemical characteristics and foam-forming and -stabilizing ability of whey and casein hydrolysates. J. Agric. Food Chem. 50, 2938–2946. Van Hekken, D.L., and Strange, E.D. (1997). Rheology and microstructure of chemically superphosphorylated whole casein. J. Dairy Sci. 80, 2740–2750. Van Hooydonk, A.C.M., Boerrigter, I.J., and Hagedoorn, H.G. (1986a). pH-induced physicochemical changes of casein micelles in milk and their effect on renneting. 1. Effect of acidification on physico-chemical properties. Neth. Milk Dairy J. 40, 281–296. Van Hooydonk, A.C.M., Boerrigter, I.J., and Hagedoorn, H.G. (1986b). pH-induced physicochemical changes of casein micelles in milk and their effect on renneting. 2. Effect of pH on renneting of milk. Neth. Milk Dairy J. 40, 297–313. Van Hooydonk, A.C.M., Hagedoorn, H.G., and Boerrigter, I.J. (1986c). The effect of various cations on the renneting of milk. Neth. Milk Dairy J. 40, 369–390. Van Vliet, T., Roefs, S.P.F.M., Zoon, P., and Walstra, P. (1989). Rheological properties of casein gels. J. Dairy Res. 56, 529–534. Vasbinder, A.J., and de Kruif, C.G. (2003). Casein-whey protein interactions in heated milk: The influence of pH. Int. Dairy J. 13, 669–677. Vasbinder, A.J., Alting, A.C., and de Kruif, K.G. (2003). Quantification of heat-induced casein-whey protein interactions in milk and its relation to gelation kinetics. Colloids Surf. B Biointerfaces 31, 115–123. Vasbinder, A.J., van de Velde, F., and de Kruif, C.G. (2004). Gelation of casein-whey protein mixtures. J. Dairy Sci. 87, 1167–1176.

38

Mary Ann Augustin and Punsandani Udabage

Vercet, A., Oria, R., Marquina, P., Crelier, S., and Lopez-Buesa, P. (2002). Rheological properties of yoghurt made with milk submitted to manothermosonication. J. Agric. Food Chem. 50, 6165–6171. Vidal, V., Gastaldi, E., Lefebvre-Cases, E., Laguade, A., Marchesseau, S., Tarodo de la Fuente, B., and Cuq, J.L. (1998). Effect of succinylation on the rennet coagulation of milk. J. Dairy Sci. 81, 69–75. Vidal, V., Marchesseau, S., and Cuq, J.L. (2002). Physicochemical properties of acylated casein micelles in milk. J. Food Sci. 67, 42–47. Xiong, Y.L.L., Aguilera, J.M., and Kinsella, J.E. (1991). Emulsified milkfat effects rheology of acid-induced milk gels. J. Food Sci. 56, 918–925. Walstra, P., and Jenness, R. (1984). ‘‘Dairy Chemistry and Physics’’. Wiley-Interscience, New York. Walstra, P., Bloomfield, V.A., Wei, G.J., and Jenness, R. (1981). Effect of chymosin action on the hydrodynamic diameter of casein micelles. Biochim. Biophys. Acta 669, 258–259. Ward, B.R., Goddard, S.J., Augustin, M.A., and McKinnon, I.R. (1997). EDTA-induced dissociation of casein micelles and its effect on foaming properties of milk. J. Dairy Res. 64, 495–504. Whiteley, A.J., and Muir, D.D. (1996). Heat stability of homogenised concentrated milk. 1. Comparison of microfluidiser with a valve homogenizer. Milchwissenschaft 51, 320–323. Williams, R.P.W., D’Ath, L., and Augustin, M.A. (2005). Production of calcium fortified milk powders using soluble calcium salts. Lait 85, 369–381. Zhu, H., and Damodaran, S. (1994). Heat-induced conformational changes in whey protein isolate and its relation to foaming properties. J. Agric. Food Chem. 42, 846–855. Zoon, P., Van Vliet, T., and Walstra, P. (1988). Rheological properties of rennet-induced skim milk gels. 1. Introduction. Neth. Milk Dairy J. 42, 249–269.

CHAPTER

2 Central Nervous System Tissue in Meat Products: An Evaluation of Risk, Prevention Strategies, and Testing Procedures M.B. Bowling,* K.E. Belk,* K.K. Nightingale,* L.D. Goodridge,* J.A. Scanga,* J.N. Sofos,* J.D. Tatum,* and G.C. Smith*

Contents

Abstract

I. Introduction II. Prevalence as an Evaluator of BSE Food Safety Risks III. Carcass Contamination with Potentially Infectious Tissues A. Stunning B. Carcass splitting C. Removal of CNS tissue IV. Methods of Detection of CNS Tissue in Meat Products A. Histological staining and IHC B. Immunochemical assays and quantification of cholesterol C. Gas chromatography-mass spectrometry D. Polymerase chain reaction V. Conclusion and Future Trends References

40 42 45 46 48 50 51 52 53 58 59 60 61

Since the outbreak of bovine spongiform encephalopathy (BSE) in the United Kingdom in 1986 and its subsequent link to the human neurological disorder variant Creutzfeldt–Jakob disease (vCJD),

* Center for Red Meat Safety, Department of Animal Sciences, Colorado State University, Fort Collins, Colorado 80525 Advances in Food and Nutrition Research, Volume 53 ISSN 1043-4526, DOI: 10.1016/S1043-4526(07)53002-0

#

2007 Elsevier Inc. All rights reserved.

39

40

M.B. Bowling et al.

presence of tissues from the central nervous system (CNS) in meat products has been considered a public health concern and, thus, has been banned from entering the human food chain in many countries. Despite this, potential can exist during harvesting to contaminate or cross-contaminate edible meat products with CNS tissue that is designated as a specified risk material (SRM) in many countries. Methods used to detect CNS tissue in meat products vary greatly in their sensitivity, specificity, cost, labor and expertise needed, ease of completion, and type of results given (qualitative vs quantitative) and, within these constraints, appropriate testing methods must be selected to monitor or verify that meat products system controls are effective in removing CNS tissue from the human food chain. The extent to which monitoring procedures are needed should be based on the public health risk of CNS tissue in meat products as determined by each sovereign nation and/or third-party international organizations such as the World Organization for Animal Health (OIE). Risk associated with consumption of CNS tissue should be estimated by sovereign nations by establishing prevalence of BSE within their borders. Using this information, science-based decisions may guide international policy and trade. Using available scientific information, appropriate testing methods for monitoring or verification, and prevalence information, nations can estimate and reduce, to the extent deemed necessary, the public health risk of vCJD.

I. INTRODUCTION Bovine spongiform encephalopathy (BSE) was first identified in the United Kingdom in 1986 and became a reportable disease in 1987 (Wells et al., 1987). Since the BSE outbreak in the United Kingdom, the disease has spread to 24 other countries including the United States, Canada, Japan, and most countries of the European Union (EU) (OIE, 2007). BSE is a fatal neurodegenerative disorder that affects the central nervous system (CNS) of adult cattle (DeArmond and Prusiner, 2003). While CNS tissue is not the only specified risk material (SRM) associated with BSE, it has historically been discussed with the most trepidation because of the high tissue infectivity titers it displays (personal communication with Danny Matthews of the United Kingdom Department for Environment Food and Rural Affairs and Gerald Wells of Fulmer Consulting Ltd.). Consumption of CNS tissue (and other SRM) that is infected with BSE is thought to cause the human neurological disease, variant Creutzfeldt–Jakob disease (vCJD), and consumption of BSE-infected brains, spinal cord, tonsils,

Central Nervous System Tissue in Meat Products

41

distal ileum, dorsal root ganglia, trigeminal ganglia, and eyes has resulted in transmission of BSE to other cattle (Hill et al., 1997; Wells et al., 1998). Currently, regulations are in place requiring the removal of SRM from the food and feed chain in countries affected with BSE (Table 1). During slaughter, potential exists for the contamination and/or cross-contamination of meat with CNS tissue by animal stunning, improper removal of SRM, carcass splitting, carcass washing, and exsanguination. Due to risk associated with cross-contamination of meat during harvest of BSE-infected cattle, testing methods have been developed and implemented to identify presence of CNS tissue in meat products. This chapter investigates the potential risk of a public health threat due to consumption of BSEinfected meat products based on prevalence information, potential routes of CNS tissue contamination of beef carcasses, and current testing methods employed to identify CNS tissue; and, it discusses governmental, industrial, and scientific methods that have been identified to reduce or eliminate CNS tissue cross-contamination onto meat and meat products.

TABLE 1 Age at which tissues are designated as SRMs in Australia, Japan, the United Kingdom, and the United States

a b c d

Tissue

Australiaa

Japanb

United Kingdomc

United Statesd

Tonsils Intestine (duodenum to rectum) Skull Brain Eyes Spinal cord Trigeminal ganglia Dorsal root ganglia Vertebral column Mesentery

Not an SRM Not an SRM

All ages All ages

All ages All ages

All ages All ages

Not an SRM Not an SRM Not an SRM Not an SRM Not an SRM

All ages All ages All ages All ages All ages

>12 >12 >12 >12 Not an SRM

>30 >30 >30 >30 >30

Not an SRM

All ages

>24

>30

Not an SRM

All ages

>24

>30

Not an SRM

Not an SRM

All ages

Not an SRM

DAFF (2007). MHLW (2005). Food Standards Agency (2007). USDA-FSIS (2004).

42

M.B. Bowling et al.

II. PREVALENCE AS AN EVALUATOR OF BSE FOOD SAFETY RISKS In order to determine the food safety risk from consumption of BSEinfected meat products, many factors must be taken into consideration. First and foremost, the infective dose of prions in humans must be quantified. To date, no scientific research has determined the infectious dose of BSE prions in humans, and, therefore, it is impossible to exactly quantify the risk of human consumption of beef products that may have small amounts of CNS tissue present. However, through use of known tissue infectivity for transmission to cattle, surveillance at all animal production sectors and at harvest, and slaughter control processes, risk can be estimated and greatly reduced. Therefore, each country, and each facility within those countries, must evaluate the likelihood of a BSEinfected animal entering their slaughter process and, subsequently, their food chain, undetected, and must identify the control measures needed to prevent or reduce (to the extent needed) that risk. Countries that conduct a surveillance program and determine BSE to be a low food safety risk would not, in theory, need to implement as many controls as countries that have a high prevalence of the disease. Countries that have a high prevalence of the disease and, thus, a higher food safety risk would, in theory, implement more controls to ensure that SRM does not enter the human food chain. Currently, the key factor in trade of beef and beef products between nations is whether BSE has been found in the indigenous cattle population of the exporting country. More important, however, than the actual number of BSE cases are the surveillance techniques implemented by each country and the control measures in place to prevent, to the greatest degree possible, a food safety threat. The World Animal Health Organization (OIE) outlines two different types of surveillance for BSE in the indigenous cattle population in their Terrestrial Animal Health Code (OIE, 2006b). The goal of each category of surveillance is to determine (using a 95% confidence interval) the BSE prevalence in a country by testing cattle subpopulations. Each subpopulation is assigned a point value (lower-risk cattle being worth fewer points and higher-risk cattle being worth more points) and, in order to meet OIE standards, each country must accumulate enough points to satisfy the requirements of the OIE surveillance program with which they attempt to comply. The first type of OIE surveillance program (denoted as ‘‘Type A’’ surveillance by OIE) is conducted by countries to determine prevalence of BSE and allows for detection of BSE if there is one BSE case (with a 95% confidence interval) per 100,000 adult cattle (OIE, 2006b). Countries implementing Type A surveillance must accumulate points based on the points system outlined in the Terrestrial Animal Health Code (OIE, 2006b).

Central Nervous System Tissue in Meat Products

43

In this system, a young animal tested at slaughter showing no symptoms is worth 0.1 point on the OIE point scale, whereas an animal that is between 4 and 7 years old showing clinical signs of BSE is worth 750 points (OIE, 2006b). The United States and Japan have implemented surveillance programs that meet or exceed OIE Type A surveillance. In 2006, Japan tested over 6,000,000 cattle for BSE including every animal at harvest, regardless of age, and found 10 animals to be positive for BSE (OIE, 2007). Through the duration of their enhanced BSE surveillance program, the United States reported test results for over 735,000 adult cattle and found 2 positive animals (USDA, 2006b), and reported a total OIE point accumulation of over 2,900,000—nearly 10 times the OIE recommended level of testing to determine prevalence of BSE (USDA, 2006a). Through Type A surveillance, countries establish the prevalence of BSE in their country. On completion of Type A surveillance (i.e., when the prevalence of BSE is well established within a country), each individual country may continue surveillance on a smaller scale using OIE Type B surveillance (OIE, 2006b). Countries designated as being at negligible risk for having BSE in their indigenous cattle herd (see below), and countries that have completed a Type A surveillance program, may implement a Type B surveillance to monitor their indigenous cattle population for BSE. Type B surveillance is designed similarly to Type A surveillance in that it uses the same point system and scale for at-risk animals (OIE, 2006b). Type B surveillance is designed, however, to detect BSE if there is 1 BSE-positive animal in every 50,000 adult animals (with a 95% confidence interval) (OIE, 2006b). Australia is a country that uses Type B surveillance. In 2005, Australia tested 501 animals, enough to meet the requirements of OIE Type B surveillance (Australia National Health Information System, 2007). It is important for all countries to conduct Type A surveillance and determine the prevalence of BSE within their borders. Furthermore, nations must continue to diligently monitor for BSE in their cattle population to prevent the inadvertent spread of the disease. Even though the factors that contribute to the spread of BSE are well known, the origin of the disease is still unknown (ILC, 2006). It is possible that BSE occurred sporadically and, therefore, all nations must actively monitor their cattle population and continue to implement risk mitigation factors such as ruminant-to-ruminant feed bans (Brown et al., 2006). In addition to creating standards for surveillance of BSE, in 2006 the OIE stipulated three categories of BSE risk in a cattle population and, thus, in a country, based on results of the surveillance plans described above (OIE, 2006a). To date, OIE has not given any country a designation based on the new criteria. However, many countries, including the United States, have presented OIE with the results of their Type A surveillance and have petitioned for OIE to determine their status, and many countries

44

M.B. Bowling et al.

will likely be designated during the voting session in May 2007 (personal communication, Dr. Chuck Lambert, the Under Secretary for Marketing and Regulatory Programs, USDA). Determination of status by a thirdparty international organization may allow for increased understanding of the risk posed by importing beef and beef products. The first category of BSE risk as stipulated by OIE is countries with ‘‘negligible BSE risk’’ (OIE, 2006a). These are countries that meet the OIE surveillance guidelines and have not detected indigenous BSE (OIE, 2006a). Australia is an example of a country that, theoretically, would be designated as having negligible BSE risk and, as such, has not implemented stringent domestic SRM removal and disposal laws (Table 1). In countries with negligible BSE risk, prevention and surveillance are of paramount importance. Stringent live-animal importation and feeding laws must be observed to prevent the introduction of the disease into the population. The second category of BSE risk includes countries with a ‘‘controlled BSE risk’’ (OIE, 2006a). Countries in this category have detected BSE in their indigenous cattle herd and have implemented necessary control measures to ensure food safety (OIE, 2006a). The United States, the United Kingdom, and Japan are examples of countries that could be designated as controlled BSE risk countries; tissues designated as SRM in these countries are listed in Table 1. Each of these countries that, hypothetically, would be designated within this category has had drastically different experiences with controlling BSE. Because BSE was discovered in the United Kingdom, control measures were not immediately available and prevalence of the disease reached very high levels. Once control measures were in place and the proper amount of time elapsed for them to take effect, cases of BSE in the United Kingdom declined (OIE, 2007). The United States, on learning of BSE and the experiences of the United Kingdom, implemented preventive measures to minimize transmission of the disease in its cattle herd and has had only two cases of indigenous BSE (USDA, 2006a). In contrast, Japan did not implement BSE control measures until the disease was found in its cattle herd and Japan is still experiencing new cases of BSE, including 10 cases in 2006 (OIE, 2007). In order to be considered for the controlled BSE risk category, countries must meet the requirements of Type A surveillance and complete a risk assessment (OIE, 2006b). The third category of BSE risk includes countries that have an ‘‘undetermined BSE risk’’ (OIE, 2006a). Countries in this category do not perform BSE surveillance, do not report any BSE surveillance results to the OIE, or do not meet the requirements of either of the other two categories (OIE, 2006a). Any country that does not meet the requirements of the ‘‘negligible BSE risk’’ or ‘‘controlled BSE risk’’ categories would

Central Nervous System Tissue in Meat Products

45

hypothetically be designated as an ‘‘undetermined BSE risk’’ country by OIE. Testing methods and procedures for BSE are outside the scope of this chapter; however, without internationally accepted and practiced methods of BSE testing for prevalence determination, all countries are not parallel and the risk of consuming meat products from countries that do not conform to international standards cannot be estimated. This problem is further exacerbated by countries that consider themselves free of BSE but do not test for the disease based on international standards, orat all. Once prevalence is established through an accepted surveillance program, countries can be categorized and risk can be more closely estimated. Once risk can be estimated, control measures can be implemented and that risk can be reduced at measurable amounts until an acceptable level of safety for all parties concerned is achieved. Each individual sovereign country has the right and responsibility to protect their population from public health threats. Currently, the food safety threat to public health of BSE is unknown because the infectious dose is unknown. Nevertheless, countries can make judgments based on available scientific knowledge concerning tissue infectivity, prevalence information provided by each country to third-party international agencies such as OIE, and countermeasures (to prevent BSE and vCJD amplification) in place in each country from the farm to the consumer. If countries determine that BSE is a hazard of private or public health concern, there are specific countermeasures that can be used to prevent the presence of potentially infective tissue in meat products.

III. CARCASS CONTAMINATION WITH POTENTIALLY INFECTIOUS TISSUES Currently, there is no test (other than histology, possibly) that is capable of identifying the presence of SRM on a beef carcass. As such, this chapter will hereafter concentrate on the contamination of carcasses by CNS tissue and the diagnostic testing for such contamination. The authors do, however, acknowledge that other potentially infective tissues, including SRM, exist. Therefore, if, during their risk analysis, a country or facility determines that CNS tissue (that could potentially harbor BSE and, thus, vCJD-causing prions) in meat products is a food safety risk, routes of contamination with CNS tissue from the carcass to itself and to other carcasses must be identified and reduced to the lowest possible amount. Additionally, the effect of abattoir personnel coming into contact with CNS tissue and spreading it throughout the production plant must be considered. The following is a discussion of potential routes of carcass contamination with CNS tissue.

46

M.B. Bowling et al.

A. Stunning In almost all countries, cattle are rendered unconscious by a device that delivers blunt force trauma to the forehead of the animal. There are several types of stunning devices that have been used in the past, including penetrative and nonpenetrative captive bolt stunners powered by gun powder, pneumatic captive bolt stunning devices, and air-injection penetrative captive bolt stunners. Some pneumatic captive bolt stunning devices inject air intracranially and, combined with the invasive action of the captive bolt, have been shown to dislodge brain and spinal cord material such that CNS tissue may enter the bloodstream and be transported into the lungs or heart of animals. Because of this, the use of airinjection captive bolt stunning devices is prohibited in the United States (USDA-FSIS, 2004) and in many other countries. In some countries, the practice of inserting a rod through the captive bolt stunning aperture and destroying the brain stem and the spinal cord in order to completely stop nerve firing and subsequent animal jerking (especially leg kicking), known as pithing, is used to augment worker safety. Garland et al. (1996) reported grossly visible brain tissue varying in size from several millimeters to 14 cm in the lungs of 2.5–5.0% of cattle at slaughter when a pneumatic air-injection captive bolt stunner was used. Similarly, Schmidt et al. (1999) evaluated presence of blood clots in hearts of cattle in 15 packing plants in the United States. In plants where airinjection stunning devices were used, 33% (n ¼ 1050) of hearts evaluated contained large clots in the right ventricle. Additionally, in plants that used pneumatic (non-air-injecting) and captive bolt stunning devices, 12% and 1% of hearts, respectively, contained clots in the right ventricle. Schmidt et al. (1999) also reported presence of large (10 to 13-cm long) pieces of spinal cord in the hearts of two animals that were stunned using pneumatic air-injection stunning devices. They noted that, in cow/bull harvest facilities (as opposed to steer and heifer facilities) due to the animal’s old age, the captive bolt remained inside the skull longer and, thus, more air was injected into the cranial cavity. Therefore, severe disruption of the brain and spinal cord ensued which was then transferred in large pieces via venous blood to the heart (Schmidt et al., 1999). Anil et al. (1999) stunned 60 animals with one of four different types of stunning devices, including a penetrative captive bolt stunner used with and without air-injection, a nonpenetrative captive bolt stunner, and a pneumatic air-injection stunning device. Blood samples were collected for 60 s following stunning, and the buffy coat of the blood was assayed using an enzyme-linked immunosorbent assay (ELISA) for presence of Syntaxin 1-B and Annexin V (Anil et al., 1999). They reported that CNS tissue was present in the jugular venous blood of 4 of 15 animals stunned using a pneumatic air-injection captive bolt stunner and 1 of 16 animals stunned

Central Nervous System Tissue in Meat Products

47

by a captive bolt stunner followed by pithing. No CNS tissue was found in venous jugular blood of animals stunned by use of a penetrative captive bolt stunner without pithing or nonpenetrative captive bolt stunning (Anil et al., 1999). Prendergast et al. (2003) collected Syntaxin 1-B and glial fibrillary acidic protein (GFAP) swab samples from 21 different locations in a beef packing abattoir, including slaughtered animals, slaughter equipment, the abattoir environment, and personnel. Samples collected after stunning indicated that the aperture created by the use of a penetrating captive bolt stunner allowed for CNS tissue to drain from the animal and onto the floor, equipment, and personnel of the abattoir. Furthermore, Prendergast et al. (2003) reported that CNS tissue remained on the captive bolt stunner, providing a vehicle for cross-contamination with CNS tissue. Rovira et al. (2007) evaluated two different methods of stunning cattle. The authors rendered 10 animals insensible in a controlled laboratory setting with a non-air-injection penetrative captive bolt stunning device. Following stunning, five of the animals were immediately exsanguinated, while the other five were administered an electric shock by a hands-free heart defibrillator in an attempt to stop blood circulation. Blood samples were collected from jugular catheters before stunning and at 90-s intervals for 6 min. Scientists did not find any CNS tissue in the whole or buffy coat of any of the blood samples, regardless of the stunning method. To further investigate potential CNS tissue spread during stunning, the researchers collected blood samples from 360 animals, immediately after sticking, at 12 commercial beef packing facilities that used non-airinjection pneumatic captive bolt stunning devices. Only 1 of the 360 samples collected was positive for CNS tissue. Finally, Rovira et al. (2007) evaluated blood samples from 30 cattle collected during Kosher slaughter (because animals harvested in this manner are not rendered unconscious before exsanguination) and found no CNS tissue contamination. These researchers concluded that (1) because the heart functions normally after stunning, the interval between stunning and sticking is the period of highest risk for dissemination of CNS tissue; and (2) non-airinjection penetrating captive bolt stunning devices are safe, if used properly, and do not create or perpetuate CNS tissue cross-contamination hazards (Rovira et al., 2007). In summary, stunning of animals creates the potential for CNS tissue contamination, slaughter equipment, and abattoir personnel, depending on the type of stunning and whether pithing is used. Further investigation is needed to determine methods that may be implemented to control this contamination. Air-injection stunning devices used to render animals unconscious may dislodge brain and spinal cord tissue and could allow for dissemination of CNS tissue through the bloodstream. The threat of facility personnel coming into contact with CNS tissue draining from the

48

M.B. Bowling et al.

stunning aperture is of particular concern as human movement throughout facilities and outside of the facility could potentially spread CNS tissue to areas thought to be free of CNS tissue. In countries or facilities which determine that BSE is a food safety risk, efforts should be undertaken to promote methods of stunning that do not penetrate the skull of the animal, to limit the draining of CNS tissue from the stunning aperture (perhaps by ‘‘corking’’) and to minimize employee contact with CNS tissue. Furthermore, countries that use pithing and/or air-injection stunning devices should be encouraged to cease those practices because they could contribute to the spread of CNS tissue to meat products.

B. Carcass splitting In abattoirs around the world, beef carcasses are normally split laterally down the center of the vertebral column, usually by use of a circular band saw, to separate them into ‘‘sides.’’ Carcass splitting disrupts, severs, and spreads the spinal cord tissue along the vertebral column of a carcass. Additionally, carcass splitting saws accumulate spinal cord tissue inside the saw housings during the splitting process and spread that CNS tissue to the split surfaces of subsequent carcasses. Helps et al. (2002) compared the CNS tissue contamination from use of a common commercial carcass splitting saw (Jarvis Buster VI) to that from use of an experimental oval-shaped saw designed to remove a portion of the vertebral column without disrupting the spinal cord. Samples were collected from five areas of the carcass, from saw operators’ aprons, and from aerosol screens placed near the site of splitting. Results indicated that use of the experimental oval-shaped saw to remove the spinal cord and surrounding vertebral column resulted in significantly less CNS tissue contamination of the carcass and the saw operators’ apron than use of the commonly used carcass splitting saw. No CNS tissue was found in any of the aerosol screens near the carcass splitting environment (Helps et al., 2002). Helps et al. (2004) slaughtered two female cattle, followed by one male, and then four female cattle and collected swab samples from the split vertebral-column surfaces. Real-time polymerase chain reaction protocols were followed to determine the extent to which tissue from the male carcass accumulated in the splitting saw and was disseminated to subsequent female carcasses. Under simulated abattoir conditions (i.e., washing the saw for 5 s between carcasses and washing the carcasses before collecting samples), these researchers reported that 0.01% of the tissue recovered from the split vertebral-column surface of the final female carcass in the sequence was from the male carcass and that 10% of the tissue remaining in the housing of the saw was from the male carcass. It was concluded from that study that ‘‘should a BSE-positive carcass be

Central Nervous System Tissue in Meat Products

49

identified, significant contamination of carcasses further down the line cannot be ruled out’’ (Helps et al., 2004). Bowling et al. (2006) evaluated cross-contamination of carcasses with CNS tissue via splitting saws at commercial beef packing facilities and identified the split surface of the aitch (pelvic) bone as a viable testing site at which to measure cross-contamination with CNS tissue. They reported that (1) the carcass splitting saw completely severs the aitch bone before beginning to split the vertebral column and, therefore, CNS tissue on the aitch bone would likely result from cross-contamination derived from the carcass splitting saw; (2) this was confirmed by collecting swab samples from the split surface of the aitch bone and from the cutaneous omobrachialis muscle of the same carcass and having the DNA of the samples compared; and (3) in each of five cases, the DNA from the excised muscle samples did not match the DNA from the material collected from the split aitch bone surface of the same animal (Bowling et al., 2006). In a subsequent trial, Bowling et al. (2006) collected samples from the aitch bone after carcass splitting and after carcass washing in five commercial beef packing plants. Samples were analyzed for the presence of GFAP, using the procedures of Reddy et al. (2006). They reported that 5.6% of 320 samples collected after carcass splitting were positive for GFAP, while 2.5% of samples collected after carcass washing were positive for GFAP (Bowling et al., 2006), indicating that carcass splitting saws cross-contaminate CNS tissue and that final carcass washing cabinets do not completely remove that contamination. In another trial, Bowling et al. (2006) evaluated four different saw-washing procedures performed with three different water temperatures circulating within the saw. Results indicated that tissue accumulation inside the saw housings and on the saw blade was not different among different water washing temperatures. The authors noted that, under normal slaughter conditions, hot (60
C) water was circulated within the carcass splitting saw to reduce microbiological cross-contamination and was not likely to be changed to reduce CNS tissue contamination. Finally, Bowling et al. (2006) evaluated two different carcass splitting-saw models and determined that both harbor CNS tissue in the saw housings and on the saw blade, and that neither one demonstrated any comparative advantages over the other with respect to preventing cross-contamination with CNS tissue. In addition to investigating the effects of carcass splitting and alternative methods of spinal cord removal on CNS tissue dissemination, researchers have also investigated methods for removing meat products from the carcass without splitting the carcass. Rotterud et al. (2005) investigated hot boning of carcasses, the process of removing prerigor meat from the carcass as a procedure to prevent CNS tissue contamination due to splitting. The researchers laterally split carcasses with a circular saw to simulate conventional carcass splitting (although carcass splitting is

50

M.B. Bowling et al.

more conventionally accomplished with a band saw) and horizontally split intact carcasses between the 10th and 11th vertebrae; intact carcasses were never split laterally, thus limiting dissemination of CNS tissue. Results from carcass sampling sites indicated that GFAP ranged from ‘‘not detectable’’ to 932.0 ng/mg total protein on conventionally split carcasses and from ‘‘not detectable’’ to 5.8 ng/mg total protein on intact carcasses. Results from minced (ground) beef samples indicated that there was no difference in GFAP concentrations between splitting methods. However, GFAP concentration on surfaces of tables on which conventionally split carcasses had been fabricated was nearly 100 times higher than the GFAP concentration on surfaces of tables on which intact carcasses had been fabricated. These researchers concluded that (1) boning of intact carcasses split horizontally rather than laterally resulted in significantly lower amounts of CNS tissue on carcasses and fabrication tables, (2) a cost-benefit analysis must be completed to determine the risk of minute amounts of CNS tissue in meat products, and (3) the cost of shifting to hot boning of intact carcasses may be high (Rotterud et al., 2005). In summary, scientific evidence indicates that carcass splitting could potentially contribute to CNS tissue cross-contamination if not controlled. In addition to spreading CNS tissue from the carcass being split, carcass splitting saws may harbor and disseminate CNS tissue to subsequent carcasses. Alternative methods for carcass splitting, spinal cord removal before carcass splitting, and carcass boning/fabrication have been shown to be more effective than traditional carcass splitting at reducing the spread of CNS tissue. However, factors such as cost, time the process takes, existing facility design, effectiveness, and the likelihood of a public health risk must be taken into account when use of these alternative methods is considered. Additionally, the risk of minute amounts of CNS tissue in meat products must be quantified in order to determine if drastic changes to beef slaughter and fabrication practices are needed.

C. Removal of CNS tissue SRM which include CNS tissue are tissues from animals that are known to carry, transfer, or perpetuate infectivity of the BSE causative agent, prions. Because BSE is an adult-onset disease, identification of tissues as SRM is dependent on the age of the animal at slaughter, and definitions of SRM vary by country, based on differences in interpretation of scientific evidence and the amount of risk allowed (Table 1). Care must be taken when removing SRM (including CNS tissue) to prevent contamination of products due to improper or incomplete removal of SRM and to prevent cross-contamination of SRM to meat products via personnel or equipment. To quantify potential CNS tissue cross-contamination during the slaughter and fabrication processes, Prendergast et al. (2003) collected samples

Central Nervous System Tissue in Meat Products

51

from (1) the hands and aprons of facility workers immediately after head removal and during transfer of heads from the chain to an SRM bin, (2) the wash water draining from the skull after washing, (3) the knives and aprons of employees on the fabrication floor, and (4) tables and conveyor belts on the fabrication floor. While both hands and aprons of workers that removed heads and workers who placed the heads into SRM bins were contaminated with SRM, the hands were significantly more contaminated. Wash water draining out of the skull after head washing was also contaminated with CNS tissue. On the fabrication floor, CNS tissue contamination occurred on the aprons and knives of workers, on samples collected from the striploin on the fabrication tables after 2 hours of operation, and in the carcass separation saws and on the carcass conveyor belts after 6 hours of operation (Prendergast et al., 2003). One strategy that has been implemented to prevent SRM tissue crosscontamination during SRM removal involves use of dedicated knives and splitting saws where SRM could be potentially handled. This strategy could be used during carcass splitting (and where saws are used on the fabrication floor), during head removal, and to remove the spinal cord. A ‘‘dedicated’’ splitting saw can be used for all carcasses over the age limit at which spinal cord is designated as an SRM. Many facilities have implemented a standard operating procedure for ‘‘dedicated’’ knife SRM removal whereby edible tissues are removed by a knife with a handle of a specified color, while SRM is removed with a knife having a handle of a different color. When the latter procedure is followed, care must be taken not to allow cross-contamination among the two kinds of knives. Proper removal of SRM (and, thus, CNS tissue) from beef carcasses is of paramount importance for processors that export, as different trading partners consider BSE differently, as a food safety risk; international standards should be followed to determine acceptable practices in trade. In international trade, presence of CNS tissue in meat products (from countries that have had indigenous cases of BSE) is unacceptable and, thus, must be completely prevented.

IV. METHODS OF DETECTION OF CNS TISSUE IN MEAT PRODUCTS Many methods have been investigated and are in use today for detection of CNS tissue in meat products, but not necessarily for all SRMs. In order to identify which test is the most efficacious, the factors of subjectivity versus objectivity, labor, the need for quantitative versus qualitative results, cost, training of qualified personnel, sensitivity, and specificity must be considered. In addition, effects of processing (including grinding, heating, and the addition of other ingredients that may interfere with the assay) must be

52

M.B. Bowling et al.

considered. Methods of testing for presence of CNS tissue in meat products include tissue dissection and visual inspection, histological staining, immunohistochemistry (IHC), immunochemical assays such as ELISA and Western blot analysis, quantification of cholesterol, reverse transcription polymerase chain reaction (RT-PCR), and gas chromatography-mass spectrometry (GC-MS). The following is a description of these methods and specific CNS tissue markers used by each method.

A. Histological staining and IHC Wenisch et al. (1999) investigated histological staining and immunohistochemistry (IHC) as a means to detect CNS tissue in sausage products. They added bovine brain tissue to normal sausage formulations at concentrations of 0.0%, 7.4%, and 33.3% and heated the mixtures at 80–100
C for 1 hour. Results indicated that, due to homogenization, histological identification of CNS tissue was not possible, regardless of staining method or amount of brain in the mixture. Immunostaining with mouse antibody for human neuron-specific enolase (NSE) revealed the presence of brain tissue, and varying degrees of staining could be determined based on brain tissue concentration in the sausage mixture. These researchers concluded that the histological procedure was an unreliable method for detecting presence of CNS tissue in processed meat products due to homogenization and high-pressure heating, while IHC with NSE is an effective method for detecting CNS tissue in such products (Wenisch et al., 1999). Kelley et al. (2000) evaluated hematoxylin and eosin (HE) histochemistry, IHC, and polarization microscopy for determining presence of CNS tissue in ground meat products produced by advanced meat recovery (AMR) systems. They collected ground beef samples from establishments that used vertebrae (as a raw material) in their AMR systems and 64 control samples of ground beef from AMR systems in establishments that hand-deboned the vertebral column before generating their products. Of the 196 samples collected from establishments that used vertebrae in their AMR systems, 19 were not subjected to desinewing, a process of pressing the product through fine screens to reduce fragment size to 2–3 nm. Of those 19 samples, CNS tissue was detectable by HE histological staining in only 2 samples, while no CNS tissue was found by HE histological staining in any of the other 177 samples subjected to desinewing, or in any of the 64 control samples. The researchers next investigated use of neurofilament and GFAP antibodies for immunohistochemical staining and found CNS tissue in 7 of 17 samples with both methods. Results indicated that peripheral nervous tissue (rather than CNS tissue) was present and detected by both GFAP and neurofilament staining. The researchers also investigated synaptophysin, a transmembrane glycoprotein localized in the CNS and not found in peripheral nervous tissue.

Central Nervous System Tissue in Meat Products

53

They found it to be a useful marker for CNS tissue and, more specifically, a useful marker for differentiating between CNS tissue and tissues of the peripheral nervous system in AMR products (Kelley et al., 2000). Tersteeg et al. (2002) evaluated the immunostaining ability of four antibodies in minced and intact meat products with CNS tissue contamination at 0%, 5%, 10%, and 20% levels and heat treatments of 0, 70, and 115
C. The antibodies evaluated were anti-neurofilament (anti-NF), antimyelin basic protein (anti-MBP), anti-NSE, and anti-GFAP. Results from this study indicated that anti-MBP was the most effective staining method due to its ability to properly stain target tissue even after heating and mincing treatments. Anti-NF was able to detect CNS tissue in all raw meat products; however, when heat treatments were applied, anti-NF staining diminished. Anti-GFAP staining was effective at creating a strong staining reaction with minced and intact products; however, background staining diminished the researchers’ ability to identify CNS tissue, and heating of samples produced a background staining that made interpretation impossible. Anti-NSE stained effectively in raw and minced unheated meat products; but when heat treatments were applied, the anti-NSE was undetectable in all products. Furthermore, at the 1% CNS tissue level, staining with anti-NSE was ineffective. These researchers concluded that anti-MBP was the most effective IHC staining antibody due to its ability to effectively stain CNS tissue in raw, minced, and heated meat products (Tersteeg et al., 2002). In summary, histological staining has been shown to be an ineffective method of CNS tissue determination in processed meat products subjected to homogenization even at CNS tissue concentrations of up to 33.3%. Additionally, histology is a qualitative method that requires highly trained personnel and is subject to misleading results due to the small proportion of the product that can be viewed microscopically. Similarly, IHC methods require highly trained personnel, expensive equipment, a long period of time to complete, and are limited by the amount of sample that is viewed for detection. As such, histological staining and IHC methods are poor CNS tissue screening tests and are more applicable in confirmatory/reference assay roles. Due to the seemingly poor sensitivity of these assays (reported to be 10 times lower than the immunochemical methods described by Hossner et al., 2006), other methods may be more effective.

B. Immunochemical assays and quantification of cholesterol There are many immunochemical methods that have been implemented to detect CNS tissue in meat products. Additionally, different CNS tissue markers have been used to investigate, qualitate, and quantitate CNS tissue presence in meat products. Immunochemical methods offer

54

M.B. Bowling et al.

highly sensitive and specific CNS tissue detection capabilities and are not subjective. Lucker et al. (1998) evaluated histological and immunochemical methods of detection of CNS tissue in meat products and reported that because the cholesterol content of tissues of the CNS is 2000 mg/100 g compared to 100 mg/100 g in other tissues, cholesterol could be used as a screening marker for CNS tissue presence in meat products. Their data suggested that cholesterol content increased by 26 mg/100 g of fresh substance for each percentage of brain tissue added to a reference product produced in their laboratory, and that normal cholesterol content of emulsion-type cooked sausages and liver sausages was 115 mg/100 g and 181 mg/100 g of product, respectively. The researchers indicated that any amount of cholesterol over those reference cutoff values should be further evaluated with a more specific CNS tissue marker (Lucker et al., 1998). In an assessment of various histological staining methods specific to CNS tissue, the researchers were unable to identify CNS tissue in any of the samples. As a result, the researchers then investigated immunohistochemical staining of products containing 0%, 7.4%, and 33.3% of brain tissue by use of monoclonal anti-NSE antibodies and were able to detect CNS tissue due to the increased staining intensity associated with increased brain content of the samples. Finally, Lucker et al. (1998) evaluated immunochemical detection of NSE tissue by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting. NSE is a dimeric protein composed of three immunologically distinct subunits (gg and ag) which are present in CNS tissue in greater concentrations than in non-CNS tissues (Kato et al., 1982). Results indicated that brain was detected at a 1% concentration using the Western blotting method in emulsion-type sausages, and at a <4% concentration in liver-type sausages. The researchers concluded that the cholesterol assay offered an inexpensive and rapid screening method for CNS tissue in meat products, particularly for processed products such as sausages, and that NSE immunochemistry is a highly specific marker for CNS tissue presence in meat products (Lucker et al., 1998). Lucker et al. (1999) evaluated 402 retail sausage samples from four different categories of sausages including (1) cooked emulsion-type sausages, (2) cooked blood sausages, (3) cooked sausages of the fat or gel type, and (4) heat-treated meat products such as hamburgers or meatballs. The upper limit of acceptable cholesterol content in each product was first determined using methods previously reported (Lucker et al., 1998), and then products were divided into two groups based on cholesterol levels. The first group consisted of all products except liver sausage, and they ranged in cholesterol levels from 119 to 132 mg/100 g product. For liver sausage, the upper limit of cholesterol was 201 mg/100 g due to the higher endogenous levels of cholesterol in liver tissue. Using the

Central Nervous System Tissue in Meat Products

55

upper limits for cholesterol concentration to detect CNS tissue described by Lucker et al. (1998), scientists determined that 16 of 402 retail sausage samples could contain CNS tissue and that those samples required further testing using immunochemical detection of NSE (by the methods of Lucker et al., 1998). Positive NSE immunochemical results were obtained in 7 of the 16 samples that had cholesterol levels above the acceptable limit. These researchers concluded that CNS tissue was present in German retail sausage products and reiterated the effectiveness of their previously reported testing method (Lucker et al., 1999). Lucker et al. (2001) evaluated 126 liver sausages for presence of CNS tissue using an immunoassay that detects presence of NSE. They detected CNS tissue in 5 of 126 samples assayed using NSE immunochemistry, and concluded that NSE immunochemistry was a highly specific and moderately sensitive method for detection of CNS tissue in non-heat-treated meat products (Lucker et al., 2001). GFAP is an antigen that is largely restricted to occurrence in CNS. It is a specific marker for differentiated astrocytes and is the cytoskeletal protein providing structural support to CNS tissue cells (Eng and Ghirnikar, 1994). Schmidt et al. (1999) first reported a method for using GFAP in a colorimetric ELISA to detect CNS tissue in meat products. These researchers quantified presence of GFAP in bovine brain, cerebral cortex, spinal cord, sciatic nerve, diaphragm, blood clots, skeletal muscle, and ground beef. They reported large amounts of GFAP in tissues from the CNS, small amounts of GFAP in the sciatic nerve, and were unable to detect any GFAP in samples from skeletal muscle, ground beef, or blood with the exception of a sample from a neck muscle, which the researchers concluded was likely cross-contaminated during the slaughter process. A limit for detection of GFAP using the colorimetric ELISA was reported to be 1.0 ng, and intra-assay variation ranged from 3.25% to 4.0% (Schmidt et al., 1999). Schmidt et al. (2001b) admitted having made mistakes in calculating the levels of GFAP in tissues in their previous study (Schmidt et al., 1999) and reported that GFAP is present at 2 ng/mg in spinal cord, 600 ng/mg in brain, 12 ng/mg in sciatic nerve, and 2 ng/mg in cervical ganglia. They also reported development of a more sensitive fluorescent ELISA (FELISA) for detecting GFAP in meat products that increased sensitivity of the assay, lowering the detection limit to 0.2-ng GFAP. In another study, these scientists reported that GFAP was detectable after the addition of normal sausage ingredients and after being heated up to 80
C; only after heating sausage samples to 115
C for 100 min did GFAP become undetectable (Schmidt et al., 2001a). Additionally, Schmidt et al. (2001a) reported that the F-ELISA was more sensitive than was the Syntaxin 1-B ELISA. Schmidt et al. (2001b) compared the F-ELISA to a commercial ELISA (R-Biopharm, Inc., Darmstadt, Germany) test kit based on their

56

M.B. Bowling et al.

colorimetric ELISA (R-ELISA) and reported that the F-ELISA was more sensitive than was the R-ELISA, but that their results were deemed inconclusive due to the poor homogenization of spinal cord/ground beef mixtures analyzed. Agazzi et al. (2002) compared the commercially available Brainostic (ScheBo-Biotech, Marietta, Georgia, United States) NSE Western blot test kit to the R-ELISA using heat-treated sausages (0, 80, and 120
C) that were spiked with CNS tissue concentrations of 0%, 0.5%, 1.0%, and 2.0%. After sausage samples were sent to 29 laboratories for determination of CNS tissue presence using both the BrainosticTM test kit and the R-ELISA, results indicated that sensitivity of both tests in nonheated and moderately heated samples was 0.5% CNS tissue (raw weight basis). In strongly heated samples, the R-ELISA was more specific than the BrainosticTM test kit, which exhibited a low level of false-negative results. As such, the detection limit for the BrainosticTM test kit in strongly heated products was determined to be 2%, whereas the detection limit for the R-ELISA remained at 0.5% (raw weight) (Agazzi et al., 2002). Hajmeer et al. (2003) compared the commercial BrainosticTM test kit to the R-ELISA by mixing spinal cord with a ground product from carcasses of three grades (utility, select, and choice) to yield samples containing CNS tissue concentrations of 0.0%, 0.0125%, 0.025%, 0.05%, 0.1%, 0.2%, 0.4%, 0.8%, and 1.6% for products from each grade of beef. For ground beef of each grade, five samples were assayed using each diagnostic test for presence of CNS tissue. Results from the BrainosticTM diagnostic test kit indicated that the limit of detection of NSE was 0.25%, regardless of quality-grade origin of the beef, and the limit of detection of CNS tissue for the R-ELISA was 0.025%, regardless of quality grade of beef. They concluded that the BrainosticTM diagnostic kit was 10 times less sensitive than the R-ELISA, took 30 hours to complete compared to 2 hours for the R-ELISA, and costs 4 times more than the R-ELISA to perform (Hajmeer et al., 2003). Findings of this latter study agreed with the cautions of Agazzi et al. (2004) with regard to the R-ELISA, indicating that the test is an effective commercial tool for qualitative analysis of CNS tissue presence in meat products, but that caution should be used when attempting to quantify the amount of CNS tissue present in samples because there is a mixture of brain and spinal cord in the R-ELISA standards (Hajmeer et al., 2003). Agazzi et al. (2004) prepared products, spiked with bovine brain tissue at concentrations of 0.0%, 0.5%, 1.0%, and 2.0% that were subjected to three differing heat treatments (none, 80
C for 20 min, and 120
C for 20 min), coded them, and sent samples of each product to each of 19 laboratories for analysis by the R-ELISA method. They reported that all samples with 0% CNS tissue were correctly identified as negative for presence of CNS tissue and that, at the 0.5% level of CNS tissue

Central Nervous System Tissue in Meat Products

57

concentration, the number of false negatives exceeded 5%. Therefore, 1.0% CNS tissue was determined to be the limit of detection of the R-ELISA, irrespective of heat treatment applied; R-ELISA is an effective qualitative detection tool for CNS tissue above those levels (Agazzi et al., 2004). Hossner et al. (2006) compared the GFAP F-ELISA developed by Schmidt et al. (2001a), the R-ELISA, and immunohistochemical (IHC) methods used by USDA-FSIS (USDA-FSIS, 1998) to detect CNS tissue in meat products by spiking ground beef samples with brain, spinal cord, or dorsal root ganglia, and then homogenizing samples to obtain CNS tissue concentrations varying from 0.05% to 0.5%. In addition to laboratory analysis of spiked samples, these researchers collected samples from the AMR systems of five commercial beef packing facilities. Results indicated that the F-ELISA was able to detect 0.3-ng GFAP while the R-ELISA was only able to detect 1.2-ng GFAP per well; and, in addition to being more sensitive, the F-ELISA was also more specific than the R-ELISA. Results indicated that the intraassay coefficients of variation were 0.61–13.20% and 3.0–36.0% for the F-ELISA and the R-ELISA, respectively, and that the interassay variation was 6–26% and 18–32% for the F-ELISA and the R-ELISA, respectively. None of the methods tested could routinely detect dorsal root ganglia at any of the concentrations tested. These researchers reported that (1) the F-ELISA was able to detect brain and spinal cord at 0.05% raw weight, whereas the R-ELISA was only able to detect spinal cord at 0.3% and was unable to detect brain tissue at any level; (2) the IHC method was able to detect 0.3% spinal cord and 0.5% brain (raw weight) with 100% of samples testing positive; (3) when lower concentrations of brain and spinal cord were used, the IHC method was able to detect CNS tissue in only 70–90% of the samples; and (4) of the AMR samples evaluated by all three methods, 17.2%, 3.2%, and 2.1% of samples assayed by the F-ELISA, the R-ELISA, and IHC methods, respectively, were positive for CNS tissue contamination (Hossner et al., 2006). Hossner et al. (2006) concluded that (1) the GFAP F-ELISA is the most effective method to detect CNS tissue in meat products, but not dorsal root ganglia (Hossner et al., 2006); (2) the R-ELISA should be used with caution because results were highly variable, the assay was not as sensitive as the F-ELISA, the assay was highly sensitive to room temperature, and the assay could not detect brain tissue at any of the concentrations tested; and (3) the IHC method was less sensitive and more time consuming than was the F-ELISA and required trained personnel and specialized equipment to perform. In summary, immunochemical testing is a highly sensitive, specific, and objective method of determining CNS tissue presence. Furthermore, immunochemical methods are relatively rapid, easy to complete with little training, and equipment costs are low. All of these traits lead to the conclusion that immunochemical assays are well suited for screening of

58

M.B. Bowling et al.

meat products for presence of CNS tissue. Drawbacks to use of immunochemical methods for detection of CNS tissue include (1) tissue-specific results that are not capable of differentiating species or age of animals; (2) unquantifiable loss of protein and, thus, sensitivity due to heating or further processing; and (3) a lack of ability to quantify results, especially in heat-treated or processed products (Biedermann et al., 2004).

C. Gas chromatography-mass spectrometry Another method that has been evaluated for its ability to detect CNS tissue in meat products is GC-MS. Niederer and Bollhalder (2001) reported that fatty acids could be detected at a sensitivity of 0.01% by solid phase extraction, acidic methanolysis to methylated esters (FAME), and evaluation via GC-MS. Biedermann et al. (2002) evaluated the brain-specific fatty acids, docosahexaenoic acid (DHA; C 22:6), lignoceric acid (C 24:0), nervonic acid (C 24:1), and cerebronic acid (C 24oh) as potential markers for CNS tissue in meat products. After creating standardized meat products with known CNS tissue concentrations, fatty acid concentrations were determined and, by modifying the protein extraction methods of Niederer and Bollhalder (2001), they were able to achieve a tenfold increase in sensitivity which resulted in a detection limit of 0.01% CNS tissue (Biedermann et al., 2002). By modifying the GC-MS methods of Niederer and Bollhalder (2001), Biedermann et al. (2002) were able to correctly identify 60 samples with varying CNS tissue amounts as positive or negative. Use of GC-MS detection of CNS tissue in meat products could serve as a valid reference method for immunochemical or immunohistochemical determination of CNS tissue in meat products (Biedermann et al., 2002). Lucker et al. (2004) evaluated the ability of GC-MS using several brainspecific fatty acids to identify and quantify CNS tissue in meat products. The researchers determined specific fatty acid content of brains from cattle, calves, sheep, pigs, turkeys, as well as muscle and adipose tissue. They determined that species and age characterizations could be made based on the concentration of specific fatty acids present in a sample. Sensitivity of GC-MS CNS tissue detection was reported to be 0.01% raw weight, but the practical sensitivity was 0.1–0.5% raw weight CNS due to the fatty acid baseline content in muscle and adipose tissue. Detection of CNS tissue by GC-MS is a highly sensitive assay. The ability of GC-MS to differentiate fatty acids between species and ages of animals is very strong, and it is a useful analytical tool with respect to addressing regulatory issues concerning SRM removal. However, detection of CNS tissue by GC-MS is likely impractical on a large scale, everyday basis due to its high cost, the length of time required to conduct the assay, and the technical expertise required of technicians (Biedermann et al., 2004). Furthermore, GC-MS detection of CNS tissue, thus far, has

Central Nervous System Tissue in Meat Products

59

only accounted for fatty acids that are found in the brain and not those found in other CNS tissues (i.e., spinal cord). A method that is incapable of detecting tissues from the spinal cord could result in false-negative results and inaccurately low numbers of positives when CNS tissue is quantified in meat products. Research is needed to identify fatty acids present in spinal cord and brain in order for GC-MS to serve as a reference assay for CNS tissue in meat products.

D. Polymerase chain reaction Detection of GFAP mRNA in minced meat and meat products by an RT-PCR was described by Seyboldt et al. (2003). These researchers collected samples of liver, kidney, spleen, lung, lymph node, heart, skeletal muscle, and spinal cord of adult cattle and, using RT-PCR methods, detected GFAP mRNA in heart, skeletal muscle, and spinal cord samples (Seyboldt et al., 2003). The unexpected presence of GFAP in heart and skeletal muscle samples was explained by the scientists as likely peripheral nervous system tissue; however, the possibility exists that, due to stunning and other slaughter processes, CNS tissue could have crosscontaminated tissues from other parts of the animal and/or carcass. They further reported that CNS tissue could be detected at concentrations of 0.5% in a brain/minced meat homogenate stored for up to 35 days and heated up to 70
C; nevertheless, sensitivity of the RT-PCR assay is likely below 0.5% CNS tissue on a wet weight basis (Seyboldt et al., 2003). These researchers explored the use of restriction fragment length polymorphisms (RFLP) as a means to identify the animal species from which the CNS tissue originated that was found in meat products. Using brain tissue from bovine, equine, ovine, porcine, turkey, chicken, and deer sources, these researchers found that species identification was possible, except that deer and bovine species could not be differentiated (Seyboldt et al., 2003). More research is needed to determine the sensitivity of the aforementioned assay. By using an RT-PCR assay as a reference, a speciesspecific assay could be implemented. However, due to equipment costs, the amount of technical expertise required to run the assay, and the time needed, RT-PCR is not likely to become a broad-based screening method in commercial applications. CNS tissue in meat products can be detected in a variety of ways and with varying degrees of sensitivity and specificity. For industrial application, immunochemical detection of NSE or GFAP as a screening method, followed by a confirmatory species- and/or age-specific assay (if needed), seems most appropriate and efficient. Each assay described herein has specific limitations that preclude the declaration of one single assay as the most appropriate for every type of testing. Researchers and industry scientists must evaluate specific objectives of their sampling and testing

60

M.B. Bowling et al.

program against capabilities of each assay and determine which assay or combination of assays is most prudent for their work. Methods that require highly trained personnel, long periods of time to complete, expensive equipment, and those that are subjective, and/or that do not evaluate all CNS tissues (i.e., brain, DRG, and spinal cord) are likely to be used only as reference assays and not as screening methods. Further research is needed to increase the sensitivity of rapid, inexpensive assays to a level that would allow companies to identify, distinguish, and quantify peripheral nervous system tissue and dorsal root ganglia in meat products in production plants.

V. CONCLUSION AND FUTURE TRENDS Meat products containing SRM (and possibly CNS tissue) tissue were banned from human consumption because of knowledge that the infective agent of both BSE and vCJD, the prion, if present, can be contained within such tissues, as well as others. Since BSE and vCJD were linked, fewer than 200 people worldwide have died of vCJD, suggesting intuitively that the infective dose of BSE prions in humans is relatively high compared to the infective dose needed to transmit the disease in cattle. During cattle slaughter, there are many potential sites or sources of CNS tissue contamination, and numerous efforts have been investigated and implemented to control them. For countries that consider BSE a food safety concern that is reasonably likely to occur, further research is needed to identify an effective method of carcass splitting that does not disrupt the spinal cord or, alternatively, that removes the spinal cord and DRG before carcass splitting. Currently, in the EU and Japan, a system is used that vacuums the spinal cord longitudinally from the spinal foramen before splitting. However, this system is incapable of removing the spinal cord at the high production speeds used in the United States and some other countries. Considerations such as cost, time needed to perform, and potential damage of high-value edible meat items near the vertebral column must be taken into consideration when developing a new carcass splitting technology. In addition, research is needed to create rapid testing methods that are highly sensitive and specific to CNS tissue for commercial use in plants. Currently, many tests exist that are able to detect minute amounts of tissues from the brain and spinal cord in meat products. Development of a test able to detect dorsal root ganglia at low concentrations, or other SRM (e.g., distal ileum and tonsils), would be beneficial. Before the first case of BSE occurred in the United States, beef exports were worth approximately $3 billion/year. In the years since BSE was detected in the United States, beef exports plummeted to extremely low levels, resulting in a $165–$190 reduction in the value of each beef animal

Central Nervous System Tissue in Meat Products

61

harvested; comparable losses in carcass value were experienced in countries throughout the world when BSE was detected in those countries. There is consensus that BSE-infected tissues, when consumed by humans at high enough doses, can cause susceptible humans to contract the neurological disorder, vCJD. In the absence of knowledge regarding the infective dose of BSE prions to humans, CNS and other tissues were banned for human consumption. It is now believed that the infective dose must be very high. In addition to this information, prevalence information collected by many countries indicates that BSE prevalence may vary tremendously, and that worldwide prevalence is declining due to implementation of successful control measures. Furthermore, there is scientific evidence indicating that CNS tissue presence in meat products usually occurs at extremely low levels—in those plants, at least, rates of cross-contamination are low. In the future, as knowledge about pathogenicity of BSE becomes clearer, cattle slaughter practices can be more effectively directed toward more complete removal of CNS and other SRM tissues from edible products. As testing methods become more sensitive and specific, it is conceivable that the outright ban on all CNS tissue will be removed and replaced with standards for acceptable limits of CNS tissue content based on risk as determined by individual countries or, perhaps, by an international body such as OIE or Codex Alimentarius. No country or process is 100% effective at removing CNS tissue from meat products. However, because of a low or decreasing prevalence of BSE, the use of best practices for SRM removal, good manufacturing practices (e.g., ‘‘dedicated’’ tools and equipment), and standard operating procedures (SOP) for detection and removal of CNS tissue from meat products, the risk of humans contracting vCJD from eating meat products is extremely low. Therefore, a more prudent policy of risk mitigation as it pertains to humans contracting vCJD from eating meat products contaminated with CNS tissue and, in exceptionally rare cases with BSE prions, would be to allow international bodies, countries, and individual facilities to address risk based on hazard analysis and critical control points (HACCP) principles of risk analysis and reduction. Agreement on, and implementation of, international testing policies for prevalence of BSE in the indigenous cattle herd of countries and on methods for determination of CNS tissue presence (even in the smallest of amounts) would allow for trade to occur based on knowledge of risk, as opposed to zero tolerance policies.

REFERENCES Agazzi, M.E., Moreno, J.B., von Holst, C., Lucker, E., and Anklam, E. 2002. Performance comparison of two analytical methods for the detection of tissues of the central nervous system in sausages: Results of an interlaboratory study. Eur. Food Res. Technol. 215, 334–339.

62

M.B. Bowling et al.

Agazzi, M.E., Moreno, J.B., von Holst, C., Lucker, E., and Anklam, E. 2004. Quantitative analysis of tissues of the central nervous system in food products by GFAP-ELISA test kit. Results of an interlaboratory study. Food Control 15, 297–301. Anil, M.H., Love, S., Williams, S., Shano, A., Mckinstry, J.L., Helps, C.R., WatermanPearson, A., Seghatchian, J., and Harbour, D.A. 1999. Potential contamination of beef carcasses with brain tissue at slaughter. Vet. Rec. 145, 460–462. Australia National Health Information System 2007. National TSE Surveillance Programcattle data. Accessed on 1/13/07 athttp://www.animalhealthaustralia.com.au/nahis/ public.php?page ¼ viewNTSEdata. Biedermann, W., Lucker, E., and Hensel, A. 2002. Detection of central nervous system (CNS) as specified risk material (SRM) in meat products by means of gas chromatography-mass spectrometry. Berl. Munch. Tierarztl. Wschr. 115, 131–134. Biedermann, W., Lucker, E., Porschmann, J., Lachaab, S., Truyen, U., and Hensel, A. 2004. Structural characterization of some fatty acids from the brain as biomarkers of BSE risk material. Anal. Bioanal. Chem. 379, 1031–1038. Bowling, M.B., Yemm, R.S., Belk, K.E., Scanga, J.A., Sofos, J.N., and Smith, G.C. 2006. An evaluation of central nervous system cross-contamination due to carcass splitting in commercial beef packing plants. Unpublished data. Center for Red Meat Safety, Colorado State University, Fort Collins, CO. Brown, P., McShane, L.M., Zanusso, G., and Detwiler, L. 2006. On the question of sporadic Bovine Spongiform Encephalopathy and Creutzfeldt-Jacob Disease. Emerging Infectious Disease 12, 1816–1821. DAFF (Department of Agriculture, Fisheries and Forestry, Canberra, Australia). 2007. Accessed on 2/18/07 at http://www.daff.gov.au/content/output.cfm?ObjectID ¼ FA29A970-95A3-442E-9D7908A00B409A59. DeArmond, S. and Prusiner, S. 2003. Perspectives on prion biology, prion disease pathogenesis, and pharmacologic approaches to treatment. Clin. Lab. Med. 23, 1–41. Eng, L.F. and Ghirnikar, R.S. 1994. GFAP and Astrogliosis. Brain Pathol. 4, 229–237. Food Standards Agency. 2007. BSE controls explained: Main controls on beef production. Accessed on 1/13/07 at http://www.food.gov.uk/bse/what/beef/controls. Garland, T., Bauer, N., and Bailey, M., Jr. 1996. Brain emboli in the lungs of cattle after stunning. Lancet 348, 610. Hajmeer, M., Cliver, D.A., and Provost, R. 2003. Spinal cord tissue detection in comminuted beef: Comparison of two immunological methods. Meat Sci. 65, 757–763. Helps, C.R., Hindell, P., Hillman, T.J., Fisher, A.V., Anil, H., Knight, A.C., Whyte, R.T., O’Neill, D.H., Knowles, T.G., and Harbour, D.A. 2002. Contamination of beef carcasses by spinal cord tissue during splitting. Food Control 13, 417–423. Helps, C.R., Fisher, A.V., Harbour, D.A., O’Neill, D.H., and Knight, A.C. 2004. Transfer of spinal cord material to subsequent bovine carcasses at splitting. J. Food Prot. 67, 1921–1926. Hill, A.F., Desbruslais, M., Joiner, S., Sidle, K.C., Gowland, I., Collinge, J., Doey, L.J., and Lantos, P. 1997. The same prion strain causes vCJD and BSE. Nature 389, 448–450. Hossner, K.L., Yemm, R.S., Sonnenshein, S.E., Mason, G.L., Cummings, B.A., Reddy, M.C.S., Sofos, J.N., Scanga, J.A., Tatum, J.D., Smith, G.C., and Belk, K.E. 2006. Comparison of immunochemical (enzyme-linked immunosorbent assay) and immunohistochemical methods for the detection of central nervous system tissue in meat products. J. Food Prot. 69, 644–650. ILC (International Livestock Congress). 2006. Bovine Spongiform Encephalopathy. In ‘‘Proceedings of the 2006 International Livestock Congress’’ Calgary, Alberta, Canada. pp. 7–10. Kato, K., Ishiguro, Y., Suzuki, F., Ito, A., and Semba, R. 1982. Distribution of nervous systemspecific forms of enolase in peripheral tissues. Brain Res. 237, 441–448. Kelley, K.C., Hafner, S., McCaskey, P.C., Sutton, M.T., and Langheinrich, K.A. 2000. An evaluation of methods for the detection of spinal cord derived from advanced meat recovery systems. J. Food Prot. 63, 1107–1112.

Central Nervous System Tissue in Meat Products

63

Lucker, E., Eigenbrodt, E., Wenisch, S., Failing, K., Leiser, R., and Bulte, M. 1998. Development of an integrated procedure for the detection of central nervous tissue in meat products using cholesterol and neuron-specific enolase as markers. J. Food Prot. 62, 268–276. Lucker, E., Eigenbrodt, E., Wenisch, S., Leiser, R., and Bulte, M. 1999. Identification of central nervous system tissue in retail meat products. J. Food Prot. 63, 258–263. Lucker, E., Horlacher, S., and Eigenbrodt, E. 2001. Brain in human nutrition and variant Creutzfeldt—Jakob disease risk (vCJD): Detection of brain in retail liver sausages using cholesterol and neuron specific enolase. Br. J. Nutr. 86, S115–S119. Lucker, E., Biedermann, W., Lachhab, S., Truyen, U., and Hensel, A. 2004. GC-MS detection of central nervous tissues as TSE risk material in meat products: Analytical quality and strategy. Anal. Bioanal. Chem. 380, 866–870. MHLW (Ministry of Health, Labour and Welfare). 2005. Enforcement regulation for the law on special measures against Bovine Spongiform Encephalopathy under jurisdiction of the Ministry of Health, Labour and Welfare. Ministry of Health, Labour and Welfare Ordinance No. 110, Article 2. Tokyo, Japan. Niederer, M. and Bollhalder, R. 2001. Identification of species-specific central nervous system tissue by gas chromatography-mass spectrometry (GC-MS)—A possible method for supervision of meat products and cosmetics. Mitt. Lebansm. Hyg. 92, 133–144. As reported by Biedermann et al. (2002). Berl. Munch. Tierarztl. Wschr. 15, 131–134. OIE (World Animal Health Organization). 2006a. Bovine Spongiform Encephalopathy. Terrestrial Animal Health Code: Article 2.3.13.1. Paris, France. OIE (World Animal Health Organization). 2006b. Surveillance for Bovine Spongiform Encephalopathy. Terrestrial Animal Health Code: Article 3.8.4. Paris, France. OIE (World Animal Health Organization). 2007. Number of reported cases of Bovine Spongiform Encephalopathy (BSE) in farmed cattle worldwide. Accessed on 1/13/17 at http://www.oie.int/eng/info/en_esbmonde.htm. Prendergast, D.M., Sheridan, J.J., Daly, D.J., McDowell, D.A., and Blair, I.S. 2003. Dissemination of central nervous system tissue from the brain and spinal cord of cattle after captive bolt stunning and carcass splitting. Meat Sci. 65, 1201–1209. Reddy, M.C.S., Hossner, K.L., Belk, K.E., Scanga, J.A., Yemm, R.S., Sofos, J.N., and Smith, G.C. 2006. Detection of central nervous system tissue on meat and carcass-splitting band saw blade surfaces using modified fluorescent glial fibrillary acidic protein enzymelinked immunosorbent assay sampling and extraction procedures. J. Food Prot. 69, 1966–1970. Rotterud, O.J., Helps, C.R., Hillman, T.J., Fisher, A.V., Harbour, D., Anil, H., and Nesbakken, T. 2005. Hot boning of intact carcasses: A procedure to avoid central nervous system self-contamination in beef and beef products. J. Food Prot. 69, 405–411. Rovira, P., Scanga, J.A., Grandin, T., Hossner, K.L., Yemm, R.S., Belk, K.E., Tatum, J.D., Sofos, J.N., and Smith, G.C. 2007. Central nervous system tissue contamination of the circulatory system following humane cattle stunning procedures. Food Prot. Trends. In Press. Schmidt, G.R., Hossner, K.L., Yemm, R.S., Gould, D.H., and O’Callaghan, J.P. 1999. An enzyme-linked immunosorbent assay for glial fibrillary acidic protein as an indicator of the presence of brain or spinal cord in meat. J. Food Prot. 62, 394–397. Schmidt, G.R., Yemm, R.S., Childs, K.D., O’Callaghan, J.P., and Hossner, K.L. 2001a. The detection of central nervous system tissue on beef carcasses and in comminuted beef. J. Food Prot. 64, 2047–2052. Schmidt, G.R., Yemm, R.S., Childs, K.D., O’Callaghan, J.P., and Hossner, K.L. 2001b. Verification of different glial fibrillary acidic protein (GFAP) analyses as accurate detectors of central nervous system tissue in advanced meat recovery (AMR) products. Meat Sci. 62, 79–84. Seyboldt, C., John, A., Mueffling, T.V., Nowak, B., and Wenzel, S. 2003. Reverse transcription polymerase chain reaction assay for species-specific detection of bovine central nervous system tissue in meat and meat products. J. Food Prot. 66, 644–651. Tersteeg, M.H.G., Koolmees, P.A., and van Knapen, F. 2002. Immunohistochemical detection of brain tissue in heated meat products. Meat Sci. 61, 67–72.

64

M.B. Bowling et al.

USDA-FSIS. 1998. Proposed rules: Meat produced by advanced meat/bone separation machinery and recovery systems. Federal Register 63, 1795–1796. USDA-FSIS. 2004. Prohibition of the use of specified risk materials for human food and requirements for the disposition of non-ambulatory disabled cattle; meat produced by advanced meat/bone separation machinery and meat recovery (AMR) systems; prohibition of the use of certain stunning devices used to slaughter; Bovine Spongiform Encephalopathy surveillance program; interim final rules and notice. Federal Register 9CFR 310.22. USDA. 2006a. Bovine spongiform encephalopathy (BSE) ongoing surveillance plan. Accessed on 2/2/07 at http://www.aphis.usda.gov/newsroom/hot_issues/bse/downloads/ BSE_ongoing_surv_plan_final_71406%20.pdf. USDA. 2006b. BSE test results. Accessed on 1/13/07 athttp://www.aphis.usda.gov/lpa/ issues/bse_testing/test_results.html. Wells, G.A., Hopkins, S.A., Green, R.B., Austin, A.R., Dexter, I., Spencer, Y.I., Chaplin, M.J., Stack, M.J., and Dawson, M. 1998. Preliminary observations on the pathogenesis of experimental Bovine Spongiform Encephalopathy (BSE): An update. Vet. Rec. 142, 103–106. Wells, G.A.H., Scott, A.C., Johnson, C.T., Gunning, R.F., Hancock, R.D., Jeffrey, M., Dawson, M., and Bradley, R. 1987. A novel progressive spongiform encephalopathy in cattle. Vet. Rec. 121, 419–420. Wenisch, S., Lucker, E., Eigenbrodt, E., Leiser, R., and Bulte, M. 1999. Detection of central nervous tissue in meat products—An immunohistochemical approach. Nutr. Res. 19, 1165–1172.

CHAPTER

3 Functional Genomics of Wine Yeast Saccharomyces cerevisiae Linda F. Bisson,* Jonathan E. Karpel,* Vidhya Ramakrishnan* and Lucy Joseph*

Contents

Abstract

66

I. Introduction II. Challenges in the Investigation of Native Yeast Strains A. The grape juice environment B. Wine yeast strain diversity C. The genomic tool chest III. Functional Genomic Analysis of Wine Yeast A. Transcript profiling B. Proteomics C. Metabolomics IV. Conclusions Acknowledgments References

66 67 73 75 91 91 104 106 108 109 109

The application of genomic technologies to the analysis of wine strains of Saccharomyces cerevisiae has greatly enhanced our understanding of both native and laboratory strains of this important model eukaryote. Not only are differences in transcript, protein, and metabolite profiles being uncovered, but the heritable basis of these differences is also being elucidated. Although some challenges remain in the application of functional genomic technologies to commercial and native strains of S. cerevisiae, recent improvements, particularly in data analysis, have greatly extended the utility of these tools. Comparative analysis of laboratory and

* Department of Viticulture and Enology, University of California, Davis, California 95616 Advances in Food and Nutrition Research, Volume 53 ISSN 1043-4526, DOI: 10.1016/S1043-4526(07)53003-2

#

2007 Elsevier Inc. All rights reserved.

65

66

Linda F. Bisson et al.

wine isolates is refining our understanding of the mechanisms of genome evolution. Genomic analysis of Saccharomyces in native environments is providing evidence of gene function to previously uncharacterized open reading frames and delineating the physiological parameters of ecological niche specialization and stress adaptation. The wealth of information being generated will soon be utilized to construct commercial stains with more desirable phenotypes, traits that will be designed to be genetically stable under commercial production conditions.

I. INTRODUCTION The completion of the sequence of the genome of the yeast Saccharomyces cerevisiae heralded a new era in research on this important industrial and experimental organism (Goffeau et al., 1996). A host of new tools were developed that allow physiological interrogations of this yeast to occur at a remarkable breadth and depth. S. cerevisiae is involved in the production of bread, beer, and wine, is used as a food flavorant, and is an additive enhancing the nutritional content of numerous food products. This yeast is also being used and developed for a host of other processes: nutra- and pharmaceutical production and delivery, probiotics, bioremediation, biofuel and bioelectricity generation, and biosensing. It is estimated that Saccharomyces has been cultivated by man for several thousand years and can be considered a domesticated yeast (Cavalieri et al., 2003; Fay and Benavides, 2005a; Mortimer, 2000). This yeast is commonly found in association with mankind and civilization, but its true origin in the natural environment has not been determined. It is a minor resident of high-sugar plant surfaces and has evolved to dominate batch fermentations (Boulton et al., 1996). Functional genomic analysis of this organism in its native environments will allow a better understanding of the activities of yeast in important food and beverage production systems and assist in the development of novel yeast-based industrial processes, as well as improve our understanding of this model eukaryote. This review will describe the challenges in the investigation of industrial yeast strains, present a synopsis of the functional genomic technologies available for analysis of S. cerevisiae and the issues with their application and interpretation, and finally summarize what has been learned to date about the biology of this yeast in natural and commercial environments as a consequence of the exploitation of these novel experimental tools.

II. CHALLENGES IN THE INVESTIGATION OF NATIVE YEAST STRAINS There are two main goals for the genomic analysis of wine yeast. The first aim is to understand how wine yeasts differ from laboratory strains and from each other at the genomic level, and to exploit these differences to

Functional Genomics of Wine Yeast Saccharomyces cerevisiae

67

broaden our understanding of the fundamental biology of this important model eukaryote. The second aim is to investigate the biological activities of yeast in environments in which they evolved in order to manipulate strain behavior to produce different or more highly valued products (Stephanopoulos et al., 2004). The application of genomic technologies to native yeast in wild or production settings poses considerable difficulty. This difficulty arises largely from the widespread differences in genome architecture known to exist among native isolates and laboratory strains and the complexity of the interaction between these yeast and their native environments. The batch fermentation of grape juice exposes S. cerevisiae to a wide array of different biotic and abiotic stressors (Bisson, 1999), which is a driving force in the generation of naturally arising strain diversity. There are also technical challenges in the application of genomic technologies to commercial and native strains of S. cerevisiae. Specific analytical platforms have not been adequately validated by laboratories employing these technologies. This lack of validation is increasingly being recognized as a major problem in the field (Draghici et al., 2006; Grunenfelder and Winzeler, 2002; Lian and Kelemen, 2006; Quackenbush, 2004, 2005; Shields, 2006). Currently applied statistical methods that seek to limit the number of false positives in a data set may be too restrictive instead increasing the level of false negatives and impacting true trait discovery. Methods aimed instead at optimizing the false discovery rate will be of more utility in the analysis of expression profile differences across strains and growth conditions (Storey and Tibshirani, 2003). There are several additional important factors that must be considered in designing a genomic analysis of industrial yeast including the experimental design itself, the biases that may have been inadvertently introduced, and the choice of strain and assay conditions. As detailed in the following sections, challenges posed to the interpretation of genomic data include the complexity of the nutritional environment, strain diversity, varying physical parameters of growth, the difficulty of controlling critical variables in production situations, and, in the special case of wine yeasts, the inability to reproduce the complexities of the diverse microbial populations present normally during fermentation. In addition, brewers, bakers, winemakers, and researchers employ different fermentation management strategies that can have a striking impact on yeast metabolic activities, transcript, and protein expression profiles. It is important to interpret results judiciously making sure that broad conclusions reached from genomic analyses are not in fact restricted to the specific conditions and strains used in the study.

A. The grape juice environment Grape juice can host a diverse microbial community in a multifaceted and variable growth environment. These growth conditions present a variety of biological challenges to yeast including high sugar, high ethanol,

68

Linda F. Bisson et al.

extremes of temperature, low pH, microbial competition, variable nutrient availability, suboptimal oxygen levels, presence of phenolic compounds and the impact on cellular redox status, and maintenance of metabolic rates under nonproliferative conditions. This variability in chemical and physical environmental parameters obviously poses an experimental design challenge for functional genomics. The observations made and the conclusions reached will be dependent on the growth conditions of the yeast, factors that must be taken into consideration in data interpretation. Typical grape juice composition is given in Table 1 (Amerine et al., 1980). Grape juice is initially high in sugar content, containing between 200 and 280 g/liter of sugar as an equimolar mixture of glucose and fructose. During fermentation, this high osmolarity decreases as sugar is consumed and is replaced by an equally high concentration of ethanol, decreasing the specific gravity of the environment below that of water. The yeast must contend with this wide range of changes in medium density. In addition, yeast fermentative metabolism produces significant heat as an end product. For every 100 g of sugar consumed, an increase in temperature of 1.3
C is obtained (Boulton et al., 1996). Depending on the type of fermentation vessel, ambient temperature, or the use of refrigeration, temperature increases of 12–15
C or higher are common. In addition, unless mechanical mixing is applied, stratification of temperature develops in batch fermentation conditions. This is particularly true in red wine production where the skins float to the surface once fermentation is initiated due to the presence of carbon dioxide. The ‘‘cap,’’ as the layer of skins is called, can retain heat and result in higher metabolic activity. Temperatures can be much higher in this area than in the rest of the tank. Reproducing such nonisothermal conditions in a laboratory in order to study the impact on yeast activities is not a trivial problem. Thus, the yeast must contend with high levels of sugar and ethanol as well as high and variable temperature. Other environmental factors may also be a source of stress. The pH of grape juice is generally between 3.0 and 4.0, but can vary in this range depending on the metabolic activities of yeast and the other microbes present. In general, wine strains are tolerant to a pH as low as 2.8, and so may be at the edge of this limit during production conditions. As the pH rises above 3.6, a multitude of bacteria that were inhibited at lower pH values can begin to grow (Boulton et al., 1996). At that point, the yeast faces not only competition for nutrients but also must handle the inhibitory effects of the end products of other microbes. Oxygen is frequently limiting during grape juice fermentation. The lack of oxygen as a participant in biochemical reactions negates the use of some metabolic options for the organism. Indeed, nutrient starvation under anaerobic conditions has been shown to be fundamentally different

Functional Genomics of Wine Yeast Saccharomyces cerevisiae

69

TABLE 1 The composition of grape juice (Amerine et al., 1980) Concentration Component

(g/100 ml)

Carbohydrates Glucose Fructose Inositol Pectin Pentoses

8–16 8–16 0.02–0.08 0.01–0.1 0.08–0.2

Organic Acids Citrate Malate Tartrate

0.01–0.05 0.1–0.8 0.2–1.0

Nitrogenous compounds Amino N Amide N Ammonia Protein

0.017–0.110 0.001–0.004 0.001–0.012 0.001–0.01

Minerals and salts Aluminum Boron Calcium Chlorine Copper Iron Magnesium Manganese Phosphate Potassium Rubidium Sodium Sulfate

Trace–0.003 Trace–0.007 0.004–0.025 0.001–0.01 Trace–0.0003 Trace–0.003 0.01–0.025 Trace–0.005 0.02–0.05 0.15–0.25 Trace–0.0001 Trace–0.02 0.003–0.035

Vitamins Biotin Folic acid Nicotinic acid Pantothenate Pyridoxine Riboflavin Thiamin

(mg/liter)

0.01–0.06 Trace–0.05 0.3–8.8 0.25–10.5 0.1–2.9 Trace–1.5 0.1–1.2

70

Linda F. Bisson et al.

from starvation under aerobiosis (Thomsson et al., 2005). Under aerobic conditions, yeasts tolerate carbon limitation better than nitrogen limitation, but under anaerobic conditions, the opposite is true (Thomsson et al., 2005). Carbon limitation anaerobically leads to a large drop in ATP levels that is not observed aerobically where respiratory metabolism can be employed to boost ATP production. Growth conditions strongly influence the ability of yeast to adapt to a changing environment and delineate the nature of what that adaptation may be. How the yeasts are pregrown can also have a dramatic impact on the response to stress in the environment. The basal levels of expression of stress response genes will affect the tolerance to specific stress conditions encountered by the yeasts and impact the observations detected at the genomic level (Davidson and Schiestl, 2001; Gasch, 2003; Ivorra et al., 1999; Zuzuarregui and del Olmo, 2004). Strains that show a low-fold induction of stress genes often are more tolerant than those showing a high-fold induction (Nugent, Mangahas, and Bisson, unpublished observations) because the difference between basal level and maximally expressed level is not as important as the basal level itself (Siderius and Mager, 2003). Grape juice also contains a plethora of phenolic compounds (Table 2), many of which have been shown to be bioactive in humans. An early study suggested that the presence of these compounds dramatically influenced yeast metabolic activities (Cantarelli, 1989). The impact of these compounds on the redox status of the yeast cells, and the impact of redox conditions in the grape milieu on wine yeast, has gone largely unexplored. One of the largest gene families in Saccharomyces is the multidrug-resistant transporter family (Goffeau et al., 1997). The true role of these genes may be in maintaining the redox status of the cell. Alternately, the growth, metabolic activities, and stress tolerance of many types of yeast are increased in the presence of phenolic compounds. The roles of these bioactive molecules in yeast may be varied and either beneficial or deleterious. In any event, they are present in the environment and therefore must be dealt with by the organisms. TABLE 2

The phenolic composition of grape juice (Amerine et al., 1980) Concentration (mg/liter)

Component

Red grapes

White grapes

Benzoic acids Cinnamic acids Flavonols Anthocyanidins Flavan-3-ols Flavan-3,4-diols

50–100 50–100 10–15 20–500 50–5000 Trace

1–5 2–10 Trace 0 0–100 Trace

Functional Genomics of Wine Yeast Saccharomyces cerevisiae

71

The chemical composition of grape juice is likewise highly variable. In the native environment, yeasts frequently encounter nutrient excess or starvation, and with the potential for rapid shifts in between. Industrial yeasts have been selected over time that can withstand these swings in nutritional status. The response to sudden starvation is dependent on previous growth conditions. This has important implications in the design of genomic experiments and in the comparison of results across laboratories. Direct application of observations made using laboratory yeast in laboratory conditions is challenging. Much of the ‘‘dogma’’ surrounding our understanding of the biology of S. cerevisiae based on exclusive examination of laboratory strains is not valid in the interpretation of behavior of native and commercial strains in the absence of an appreciation of the differences in environmental forces affecting metabolic activities of these organisms. For example, the current model of cell cycles and growth based on laboratory strains in laboratory media poses nutrient uptake in G1, with accumulation of sufficient nutrients signaling start of a new cell cycle. At the end of that cycle, nutrients are again accumulated for the next cycle, and so on. Vacuoles accumulate in mass at the point of glucose exhaustion in stationary phase. In contrast, in wine yeast under wine production conditions, stationary phase-type vacuoles are present during active growth. They serve an important purpose as sites of storage of medium nitrogen. Yeasts efficiently deplete nitrogen from the medium, storing it internally, and can undergo several rounds of replication without further nutrient addition. This divergence is likely due to the vast differences between sugar levels in laboratory media and in the natural environment. Laboratory yeasts have largely been examined under a single condition: limitation for a carbon and energy source while in the wild, other nutrients may be limiting initially with growth eventually being impacted by the accumulation of ethanol. We found that in the absence of other stress factors, laboratory industrial and native isolates of Saccharomyces grow quite well even at ethanol concentrations of 10% if sufficient other nutrients have been provided and a sugar substrate is available. This is also supported by analyses of gene expression in isolated native strains of S. cerevisiae versus those that have been cultivated in laboratories (Kuthan et al., 2003; Palkova, 2004). Strains rapidly lose some phenotypes associated with growth in the wild (Palkova, 2004). One of the main challenges for yeast during fermentation is to maintain cellular parameters within specific restrictions to attain and maintain optimal conditions for metabolic activity. The limiting nutrient in grape juice is most often nitrogen. Numerous factors impact the ability to use nitrogen in the environment: the presence and complexity of nitrogen sources, cellular storage levels at the onset of fermentation, metabolic demands for nitrogen, the presence of ethanol and oxygen, the medium pH, and potassium levels.

72

Linda F. Bisson et al.

It is also important to consider the possible effects of the other microbes present in grape juice that can impact the metabolic activities of yeast (Renouf et al., 2006). Several yeast genera: Brettanomyces, Candida, Debaryomyces, Hanseniaspora, Kloeckera, Kluyveromyces, Metschnikowia, Pichia, Schizosaccharomyces, Torulaspora, and Zygosaccharomyces, have all been reported to occur in grape juice (Fleet, 1993; Fleet and Heard, 1993). The levels of yeast found vary quite dramatically, depending on winery practices and the use of antimicrobial agents. Lactic and acetic acid bacteria are also present, the specific genera and species are largely dependent on grape juice pH, the temperature of fermentation, and the sensitivity of the strain to the metabolic activities of Saccharomyces (Fleet, 1993; Fleet and Heard, 1993). Yeast fermentation behavior has been difficult to monitor, given the number of parameters involved and the varying composition of grape juice (Cramer et al., 2002). Glucose is consumed more quickly than fructose, and cell viability is rapidly lost on sugar depletion (Figure 1). Although nitrogen is most often the limiting parameter, the kinetics of carbon utilization during most of the fermentation is not well correlated with nitrogen levels, especially toward the end of fermentation (Insa et al., 1995;

Synthetic grape juice fermentation 100 Glucose fermentation rate Fructose fermentation rate Glucose Fructose Cell mass Viable cells

120 100 80

10

1

60 40

0.1

20 0 0

2

4

6

−20

8

Absorbance (580 nm) − colony forming units (x10E6)

Sugar (g/liter) - fermentation rate (g/liter/day)

140

10 0.01

Time (days)

FIGURE 1 Yeast cell growth, viability, sugar consumption, and ethanol production patterns of a typical fermentation. In a synthetic grape juice medium, Triple M (Spiropoulos et al., 2000) was used and inoculated with a commercial strain of S. cerevisiae. Glucose and fructose concentrations were determined by enzymatic assay, viable cell counts by plating on YPD medium, and cell mass by absorbance at 580 nm.

Functional Genomics of Wine Yeast Saccharomyces cerevisiae

73

Manginot et al., 1998). Fermentation rates likewise are not well correlated with cell number as the fermentation capacity of cells can vary. Energy reserves at the point of implementation of stress appear to be a critical factor with higher storage levels of glycogen and trehalose associated with improved survival (Thomsson et al., 2005). Finally, the bulk of the fermentation in grape juice production as well as in beer and during bread production is conducted by nongrowing cells. Our understanding of nonproliferative metabolically active states is limited. In the case of wine, the limitation on growth is caused by the attainment of terminal cell density. We have observed that cells immediately resume growth with no appreciable lag if the cell number in the nonproliferative condition is reduced by centrifugation. Thus, though not growing these cells are primed to grow as soon as biomass levels are reduced providing conditions are still permissive for growth. We have observed growth at ethanol levels considered to be inhibitory based on studies conducted with laboratory yeasts (Kumar, Goyashiki, Karpel, Ramakrishnan, and Bisson, unpublished data).

B. Wine yeast strain diversity Genomic analyses have already revealed that many of the currently used commercial strains have acquired altered signaling properties (Verstrepen et al., 2004). Commercial and native yeast isolates display greater genomic and genetic instability than laboratory strains (Ambrona et al., 2005). Furthermore, trisomy and tetrasomy for some chromosomes are common in industrial strains (Bakalinsky and Snow, 1990). Wild strains are generally homothallic and tend to show low sporulation rates, poor spore viability, high levels of heterozygositiy and chromosomal polymorphisms, rearrangements, and karyotype instability (Carro and Pina, 2001; Codon et al., 1998; Landry et al., 2006a,b; Longo and Vezinhet, 1993; Myers et al., 2004). This dynamic instability is the foundation of the ‘‘genome renewal’’ hypothesis for Saccharomyces (Mortimer et al., 1994). Hauser et al. (2001) proposed that chromosome imbalances must pose some selective advantage in the environment because of their persistence in wild populations and the frequency with which they arise. These genetic differences give rise to differences in phenotype. Various methods, including molecular and more traditional techniques, have been used to compare the genetic relatedness among wine isolates of Saccharomyces. Initial studies of the differences in the genomes of wine yeast were done using traditional genetic tools. Crosses were conducted with strains that showed characteristics of interest, and spore segregation ratios were determined by tetrad (Cummings and Fogel, 1978; Takahashi, 1978; Thornton and Eschenbruch, 1976) or random spore progeny analysis (Bakalinsky and Snow, 1990; Spencer et al., 1980). The initial studies

74

Linda F. Bisson et al.

indicated that there was a high level of diversity in the wine yeast strains. These methods measured only chromosome numbers but gave more information than that provided by measuring total DNA content (Leusch et al., 1985). Methods that require sporulation are limited when sporulation is poor and spore viability is low. Analysis shows that spore viability is highly variable in wine yeast with numbers as low as 0% and as high as 100% (Guijo et al., 1997; Johnston et al., 2000; Thornton, 1986; Thornton and Eschenbruch, 1976). Methods that were not dependant on sporulation ability were necessary to look at genetic diversity in large numbers of wine strains (Schuller et al., 2005). Methods that targeted metabolic products rather than direct genetic analysis, such as fatty acid analysis with gas chromatography (Ockert and Kock, 1989; Tredoux et al., 1987), were developed to investigate strain diversity. This method and the related fatty acid methyl ester (FAME) analysis (Peltrouche-Llacsahuanga et al., 2000) have been used successfully but rely on only a single class of metabolic products from yeast. The availability of molecular genetic tools for Saccharomyces in the 1980s led to a proliferation of techniques applicable to the examination of the genetic diversity of wine yeasts. These methods have the advantage of being more random than fatty acid analysis and are not dependant on sporulation. Currently, restriction fragment length polymorphism (RFLP) of mitochondrial DNA is a popular method, while pulse field gel electrophoresis is often used as the standard of comparison for new methods. Polymerase chain reaction (PCR)-based techniques are also popular. These include interdelta sequence analysis, intron splice analysis, random amplification of polymorphic DNA (RAPD), multi-locus sequence typing (MLST), amplified fragment length polymorphism (AFLP), and simple sequence repeats (SSRs) or microsatellite sequencing. Karyotype analysis using pulse field electrophoresis (Briones et al., 1996; Izquierdo Canas et al., 1997) and DNA fingerprinting using RFLP of various regions as well as microsatellite PCR (Baleiras Couto et al., 1996) were some of the first techniques to be used. More recently, comparative microarray and proteome analyses have also been used to characterize differences in and classify strains of S. cerevisiae used in winemaking. Several comparisons of these techniques were carried out with similar results. Baleiras Couto et al. (1996) indicated that a combination of analyses, including RAPD, SSR, and RFLP, were required to give good discrimination between S. cerevisiae strains. Schuller et al. (2004) concluded that any single method they tried: SSR, interdelta sequence analysis, or RFLP of mitochondrial DNA, gave adequate discrimination between strains. However, the SSR was the most sensitive method. Gallego et al. (2005) compared RAPD, AFLP, and SSR and determined that, while all were adequate, SSR again gave the highest level of discrimination. In all of these analyses, the majority of strains were distinguishable and only

Functional Genomics of Wine Yeast Saccharomyces cerevisiae

75

a small percentage appeared to be identical strains. In the Baleiras Cuoto study, 13 out of 16 isolates (81%) were determined to be unique strains and the 4 identical strains were isolated from the same location. In the Schuller study, 23 commercially available isolates gave 21 or 22 strain types (91–96% of those tested) depending on the technique used. Of the 27 strains tested by Gallego et al. (2005), 81–91% were determined to be unique strains, again depending on the method. Strains that were not unique were isolated from the same must or different musts within the same winery. It is now possible to rapidly type both commercial and wild wine yeasts to determine if they are unique strains. With the availability of microarray analysis, it may be possible to begin to determine in what genes and characteristics these differences lie. A direct comparison between a laboratory and a wine yeast (Hauser et al., 2001) found more than 40 genes that showed different expression patterns under the same conditions. On careful analysis, these differences were attributable to gene copy number and small variations in promoter regions. A study compared four wine strains with the fully sequenced laboratory strain (S288C) in a microarray karyotype analysis (Dunn et al., 2005). In this study, comparison between the wine strains and the laboratory strains showed differences primarily in transport and permease genes, particularly those involved in drug resistance. There were small but significant differences detected between the four wine strains studied. These differences were enough to distinguish the strains from one another and to give each of the strains a microarray karyotype ‘‘signature.’’ While the study looked at relative copy numbers of genes compared to the sequenced laboratory strain (S288C), it was not possible to determine if there were genes that were present in the wine strains and not in the sequenced strain. Determining how the genetic differences in the strains account for the differences observed in fermentation characteristics and variations in finished wines remains a daunting task.

C. The genomic tool chest The yeast community has benefited from the development of several elegant genomic tools that can be applied to the study of yeast biology (reviewed in Lockhart and Winzeler, 2000). These tools have changed analysis from a focus on single genes and proteins to examination of the comprehensive transcript or protein composition of yeast. Global transcript analysis has been termed transcript profiling or ‘‘transcriptome’’ analysis and the companion analysis of total cellular proteins termed ‘‘proteomics.’’ In addition, techniques in chemical constituent analysis are also being applied in order to complete the picture and display the

76

Linda F. Bisson et al.

‘‘metabolome’’ of yeast. The analysis of the complete set of mutant phenotypes has been termed ‘‘phenomics’’ (Bader et al., 2003; Warringer et al., 2003). Additional experimental tools for genome-wide analysis have been developed and tested on yeast, such as methods allowing global searches for functional elements in promoter regions that can be used by any researcher (Cliften et al., 2003; David et al., 2006; Kellis et al., 2003). Arrays developed for mRNA transcript analysis are also being used to analyze DNA directly (Dunn et al., 2005; Gresham et al., 2006; Winzeler et al., 2003). This provides important information on the global differences between strains and is a critical first step for comparative analyses of different strains (Dunn et al., 2005). Analytical tools have been developed that allow genotyping directly from mRNA arrays (Ronald et al., 2005). Transcript profiling can also be used on populations following adaptive evolution (Ferea et al., 1999). In this case, a population of isogenic cells is exposed to some consistent unique environment for multiple generations. At the end of the adaptive period, both transcriptome and DNA analyses can be performed on DNA chips. This can reveal both the transcriptional changes as well as possible nucleotide polymorphisms that have arisen in this culture as compared to the control culture not exposed to the same environmental conditions. Oligonucleotide arrays have also been used to map the topography of replication (Raghuraman et al., 2001). In addition to these indispensable technologies, several other tools have been developed to facilitate genome-wide research on Saccharomyces. Two types of comprehensive collections of mutants have been generated, based either on insertional or deletional mutagenesis (reviewed in Vidan and Snyder, 2001). Large-scale insertional mutagenesis was used to disrupt open reading frames (ORFs) (Kumar et al., 2002, 2004). This study also used a transposable element gene trap based on detection of expression of a LacZ fusion protein. The appearance of b-galactosidase activity indicated that a fusion to a functional ORF had occurred and that a true expressed gene had been identified or trapped. This analysis resulted in the identification of ORFs that had not been previously annotated in the genome. The second comprehensive mutant collection was generated directly via systematic and specific disruption of every putative ORF in S. cerevisiae (Giaever et al., 2002; Winzeler et al., 1999). Each putative ORF has been deleted, and the set of nonlethal null mutations is commercially available from Open Biosystems (http://www.openbiosystems.com/Geneexpression/Yeast/YKO/). Each deletion also contains a pair of unique sequence tags or ‘‘bar codes’’ that allow that mutation to be identified in a population of cells (Figure 2). This important set of mutants allows highthroughput screening for desired phenotypes (Tong et al., 2001; Warringer and Blomberg, 2003). These screens, when used in combination

Functional Genomics of Wine Yeast Saccharomyces cerevisiae

77

Chromosome KANR

Region of homology Common upstream and downstream flanking PCR priming sites Tag sequence unique to gene KANR

Selectable kanamycin resistance gene

FIGURE 2 Depiction of the strategy used for the creation of the genomic deletion set of strains for S. cerevisiae. Each disruption contains a selectable marker, kanamycin (G418) resistance, and a tag or specific sequence unique to each construct. The PCR primers flanking the unique tag or ‘‘bar code’’ can be used to quantify the persistence of that tag and therefore of that specific mutation in a population of cells.

with transcriptome or proteome data, can be used to confirm the functional or regulatory role of a gene. Fitness studies can also be conducted under specific growth conditions, with analysis for each specific tag of the DNA isolated from the final population. Tags that are over- or underrepresented suggest a more or less fit genotype for the conditions under study. High-throughput microscopy techniques screen for various morphological aberrations in the yeast deletion set (Narayanaswamy et al., 2006). Other tools include a comprehensive set of fusions of fluorescent or affinity tags to each gene in the yeast genome (Andrews et al., 2003; Gelperin et al., 2005). The availability of a set of fluorescently tagged genes has allowed global analysis of protein localization in yeast (Huh et al., 2003). The use of affinity tags has allowed better visualization and quantitation of protein species. The tandem affinity purification (TAP) tag allows immunodetection of tagged proteins as well as immunopurification (Ghaemmaghami et al., 2003). The set of deletion strains is also being used for metabolic footprinting, a comprehensive analysis of the metabolites released into the medium to be correlated with the specific knockout mutation borne by the strain (Allen et al., 2003). The use of excreted metabolites avoids the issues of differential extraction and sample preparation if internal components were being evaluated. Comprehensive studies to define gene function have also been undertaken (Wu et al., 2002). An issue has arisen with respect to both types of sets of comprehensive mutations. Several studies have found that aneuploidy or other

78

Linda F. Bisson et al.

compensating mutations can arise in these mutant strains on cultivation (Hughes et al., 2000; Oshiro and Winzeler, 2000; Vidan and Snyder, 2001). The appearance of these types of genetic modifiers can arise at surprisingly high frequencies, and can be influenced by experimental design. As a consequence, all genes identified as positively affecting a specific phenotype need to be confirmed, but false negatives, strains with modifiers suppressing the original phenotype, may be more difficult to identify. Genetic footprinting using Ty1 insertional mutagenesis has also been undertaken (Smith et al., 1995). Genes containing insertions in a population can be identified by PCR screening. After growth or application of selective pressure, the population can be rescreened and those genes that have been enriched or selected against identified. The advantage of this technique is that it can be applied relatively quickly to other yeast strains. This technique can also be used on a new population minimizing the risks of accumulation of compensatory secondary mutations or aneuploidies. Many studies have also been undertaken to define the complete set of functional and physical interactions among proteins (Bader and Hogue, 2002; Gavin et al., 2002; Hazburn and Fields, 2001; Ho et al., 2002; Ito et al., 2001; Link et al., 1999; Tong et al., 2004; Uetz and Finley, 2005; Uetz et al., 2000; Von Mering et al., 2002), providing an important reference data set for the rest of the research community, even if the different studies have not reached the same conclusions (Grigoriev, 2003). Gene annotation services in Saccharomyces (Dwight et al., 2002; Gollub et al., 2003) serve as a model for other organisms and provide a critical database for researchers in the field. The ongoing gene ontology project continues to provide and catalog important molecular function of gene products and their cellular roles. Much of the thrust of this work has been to provide tools or resources for the scientific community. Although this work has been undertaken virtually exclusively in laboratory strains, much of the information gained should be applicable to wine yeast strains. As beneficial as these technologies are, there are some shortcomings in how data are being analyzed and interpreted. In some cases, the platform chosen for transcriptome or proteome analysis limits the information to be obtained (reviewed in Draghici et al., 2006). A published comparison of mouse transcriptome analyses concluded that there was more crossplatform variation than cross-laboratory variation for the same platform (Kuo et al., 2006). Not surprisingly, they found the greatest disagreement among low expression values than for medium and high expression values. They also found that the variability was greater for noncommercial arrays. To address the variability in microarray data and so that data sets across laboratories and platforms may be compared, several initiatives aimed at improving array reproducibility have been launched: MIAME [Minimum Information About a Microarray Experiment (http://www. mged.org/Workgroups/MIAME/miame.html)], ERCC [the External RNA

Functional Genomics of Wine Yeast Saccharomyces cerevisiae

79

Controls Consortium (Baker et al., 2005)], and MAQC [Microarray Quality Control Project (http://www.fda.gov/nctr/science/centers/toxicoinformatics/maqc/)]. However, these guidelines tend to focus on documentation of some analytical details or use control nucleotides that do not broadly represent the full dynamic range of the array and fall short of really addressing the technical issues in these types of analyses (Draghici et al., 2006; Shields, 2006). One common use of these technologies is in the analysis of genes of unknown function, and it is a belief among many that coexpression implies coregulation and that the appearance of a transcript under certain stressful conditions means that the gene product is required for tolerance to that condition. Although there is some experimental evidence to support the relationship between coexpression and coregulation, the combination of transcriptome analysis with high-throughput screening of the deletion set of strains indicates that genes that may be highly expressed in response to a specific stress often, when mutated, do not impact stress sensitivity (Birrell et al., 2002). Coregulation in some cases is merely circumstantial and gene expression is not functionally linked (Heyer et al., 1999). Finally, it is important to note that the most important gadgets of the yeast tool kit are the excellent gene annotation and database resources that are available such as the Saccharomyces Genome Database maintained by Stanford (http://www.yeastgenome.org/)—which includes and maintains up-to-date information on gene annotation, scientific literature, and is a resource for tools and other information of importance to the research community—and the Stanford Microarray Database (http://genome-www5. stanford.edu/). Also important is the companion Comprehensive Yeast Genome Database (http://mips.gsf.de/genre/proj/yeast/). Databases of protein identification from two-dimensional sodium dodecyl sulfate polyacrylamide gel electrophoresis (2D SDS–PAGE) also exist in the Yeast Protein Map project (http://www.ibgc.u-bordeaux2.fr/YPM/), SWISS 2D PAGE (http://www.expasy.org/ch2d/), the Yeast Proteome Database (http:// www.biobase-international.com/pages/index.php?id=139), and the Yeast Resource Center for analysis of protein interactions and complexes (http:// www.yeastrc.org/unknown_orfs/complexes.html). This is only a partial list highlighting the abundant resources available to yeast researchers.

1. Transcriptome analyses There are four main types of trancriptome analyses utilized in yeast mRNA profiling, each with different benefits and deficiencies (Table 3). The first two methods rely on arraying complementary nucleotide information on a grid with subsequent hybridization of a sample preparation. These methods are based on scaling up Northern blot technology of complementary nucleic acid hybridization to assess multiple transcripts

80

Linda F. Bisson et al.

TABLE 3

Transcript profiling methods

Method

Process

Quantification

Hybridization

Probe: cDNA Labeled mRNA cRNA Target: PCR fragment Oligonucleotide

Intensity equated with relative level of expression

mRNA fragment analysis (SAGE)

Capture PolyA regions DNA sequence analysis of concatemers

Frequency of sequence equated with level of expression

Quantitative reverse transcription PCR

PCR reaction

Intensity of labeling during PCR reaction equated with level of expression

simultaneously. There are two basic types of hybridization methods being employed, those that use arrays based on PCR products specific for each gene or ORF and those using oligonucleotides (reviewed in Draghici et al., 2006). The oligonucleotides may be short (25–30 base pairs) or long (60–70 base pairs), and contact spotted, inkject deposited, or synthesized directly on the array. In the first of these methods, specific PCR fragments of yeast ORFs are arrayed on a glass slide or grid (DeRisi et al., 1997; Lashkari et al., 1997; Schena et al., 1995), or on a nylon membrane (Alberola et al., 2004). The mRNA is then purified from reference and experimental conditions, labeled with a fluorescent tag directly (biotinylated) or after conversion to cDNA [Cy3 (green) or Cy5 (red)], or during synthesis to cRNA, and the tagged mRNA/cDNA/cRNA is then hybridized against the PCR fragment array (reviewed in Lockhart and Winzeler, 2000). Samples may be hybridized to the array or grid singly, or after mixture of samples labeled with two different dyes. In the double dye-binding method, differential fluorescence is scanned and used to calculate the differences in ratios of expression between the reference and experimental sample. Similar types of strategies have been employed using radioactively labeled cDNA preparations (Rep et al., 2000; Zuzuarregui and del Olmo, 2004). The advantages of this technique are its relatively inexpensive cost, ability to be utilized in individual laboratories, and relative ease of data manipulation. The disadvantages center on the spotting technology and the inability to achieve uniform

Functional Genomics of Wine Yeast Saccharomyces cerevisiae

81

spots with the associated difficulty in determining the actual strength of the signal. The comparative analysis of array platforms by Kuo et al. (2006) found that these double dye-binding methodologies were not as reproducible as single signal platforms. However, other studies have reached the opposite conclusion, that relative comparisons of expression are more robust than attempts to quantify individual expression patterns (Draghici et al., 2006). Other issues concern the quality of the mRNA preparation and the ability to biologically reproduce the pattern. Replication of the same sample generally yields acceptable levels of reproducibility (Lee et al., 2000). There is greater variation if true biological replication is utilized, that is, a sample is completely replicated in the growth conditions, harvesting and preparation rather than simply analyzing the same mRNA preparation twice (Quackenbush, 2005). The costliness of many commercial array technologies has forced a compromise on replication. Specialized arrays, with a more limited set of genes represented, have also been employed (Rodriguez-Pena et al., 2005). In the second type of hybridization-based analysis, complimentary oligonucleotides are used to identify the specific cDNA species present (Lockhart et al., 1996; Schadt et al., 2000; Wodicka et al., 1997). Each transcript is represented by a few to several oligomers providing independent signals for each gene. As a control for nonspecific hybridization, the Affymetrix design also includes a mismatch of each oligonucleotide. The strength of the signal is then estimated both on the absolute values as well as on the difference between the perfect match and mismatch signals across the gene. The Affymetrix version also includes control chips, which we have found to be invaluable in the analysis of replication error. These control chips contain 30 and 50 regions of three genes. A 30 to 50 ratio of 1 indicates that the 30 and 50 regions are equally represented in the population of labeled cDNA. We have found that if the Affymetrix platform is used and the test chips are employed, the differences between biological replications is greatly diminished if high-quality standards (nearness to a 30 to 50 ratio of 1 for all three marker genes) are utilized. In yeast, the detection limit or dynamic range of mRNA for these technologies is on the order of 1–10 copies of mRNA per cell (Draghici et al., 2006; Holland, 2002). Other studies have shown that the majority of mRNA species in S. cerevisiae are at or below this limit (Varela et al., 2005) and are therefore below the sensitivity of current array technologies (Shields, 2006). A single mRNA molecule on average can produce 4800 protein molecules, so it is not surprising that the majority of genes would be expressed in this low range (Wohlschlegel and Yates, 2003). In addition to obstacles imposed by low transcript abundance, accurate quantitation of absolute expression levels can be difficult to achieve (Draghici et al., 2006). At issue is the need to select single-hybridization conditions for these technologies at the same time that specificity of probe design is

82

Linda F. Bisson et al.

maintained. Cross-hybridization can also be a major impediment in accurate quantitation and has been shown to lead to compression of expression ratio differences where the observed ratio is less than the true ratio due to the high background of nonspecific binding (Draghici et al., 2006). Probe selection influences signal strength, and it is important that imaging software, signal detection threshold, and quantitation take this into account (Wang et al., 2006). The third class of transcript analysis is serial analysis of gene expression or SAGE (Kal et al., 1999; Velculescu et al., 1995). In this technology, the ends of the mRNA transcript are harvested, formed into concatemers, and then the concatenated molecules are sequenced. The higher the number of times a particular sequence appears, the higher the level of that transcript in the population. SAGE analysis works well for highly expressed genes, but has limitations for genes that are not that highly expressed, as they may be represented only once or twice in the sequenced pool. The limitation here, as with the other array platforms, is the expense both in cost and in time, of sequencing a much larger set of concatemers to have statistically valid information for the messages with low expression levels. The final type of transcript analysis is quantitative reverse transcription PCR (QRT-PCR) (Holland, 2002). This technology can be scaled to be genome wide, although its major use seems to be in confirming array data generated by other means. QRT-PCR has a broader dynamic range than array technologies but is subject to other types of limitations, such as the nature of the primers used (Freeman et al., 1999). The comparative analysis by Kuo et al. (2006) indicated good agreement among the array and QRT-PCR methodologies for highly expressed genes. Although these study centered on mouse genomics, the conclusions reached are likely broadly applicable to genomic analyses in general.

2. Proteome analyses The term ‘‘proteome’’ was coined as a companion to transcriptome to define tools and analyses directed at dynamically profiling the protein complement of a cell (reviewed in Graves and Haystead, 2002). The aim of proteomic analyses is the separation, identification, and quantitation of all proteins in a mixture. Proteomic analyses have been as valuable as transcript profiling in the analysis of yeast physiology. These technologies are not as robust as transcript profiling since proteins are far more heterogeneous than RNA. Currently, it is not possible to know with any degree of certainty that all proteins are being visualized and that relative ratios of proteins in the cytoplasm are being maintained through sample preparation and protein separation (Gygi et al., 1999a,b, 2000; Moseley, 2001; Peng et al., 2003; Ravichandran and Sriram, 2005). However, as with proteomics, some important observations have been made using subsets of components that could be accurately quantified. When these limitations

Functional Genomics of Wine Yeast Saccharomyces cerevisiae

83

are taken into account, these technologies can provide significant insights into yeast biology, especially when used in combination with other types of genomic analysis. There are several techniques currently being employed for proteome analysis of yeast, ranging from traditional 2D SDS-PAGE to more sophisticated separations technologies coupled to mass spectrometry. Proteome analysis in the form of electrophoresis (2D SDS-PAGE) (Garrels, 1983) actually preceded mRNA profiling. In this technology a protein mixture is separated in two dimensions electrophoretically. In the first dimension, a pH gradient is employed to separate proteins according to their isoelectric point. The second dimension separates proteins according to molecular weight as influenced by the Stoke’s radius of the protein. Different staining techniques are employed to visualize proteins. 2D SDSPAGE allows visualization of proteins modified either by charge (change in isoelectric point) or by size and therefore has the potential to readily detect key protein modifications. Individual proteins can then be identified following in-gel digestion and analysis of peptide fragments by mass spectrometry (Muddiman et al., 1997) or by Edman degradation and peptide sequencing (Edman and Begg, 1967). Tandem mass spectrometry (MS/MS) allows the initial peptides and accurate mass determination followed by selection of peptide masses of interest and further fractionation (Shevchenko et al., 1996a,b; reviewed in Chalmers and Gaskell, 2000). Mass identification of the peptides obtained is then used to determine the sequence of the original peptide and ultimately of the protein (Figure 3). The completion of a genome of an organism greatly enhances this technology in providing a database of possible protein sequences allowing more rapid identification of the protein. Peptide mass fingerprinting, the enzymatic fractionation of a protein followed by accurate mass determination of the peptides produced, provides a robust mechanism of protein identification without the need for sequencing (Shevchenko et al., 1996a,b). Two analytical methods are utilized to ionize proteins or peptides without damage: matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS) (Larsson et al., 1997; Sagliocco et al., 1996) and electrospray ionization tandem mass spectrometry (ESI MS/MS) (Shevchenko et al., 1996a,b). Both methods are ideal for analysis of complex mixtures of proteins or peptides, but suffer from interferences from the protein matrix that may require sample manipulations that could affect the proteins obtained. Although the most commonly used method for proteome analysis remains 2D gel electrophoresis, there are several important limitations of this technology (Gygi et al., 2000; Olineka et al., 2005). First, large proteins and hydrophobic (membrane) proteins have difficulty entering the gel and are typically lost from the protein profile as are small proteins that may migrate too rapidly. Proteins that are highly basic or acidic are

84

Linda F. Bisson et al.

Band isolation

In gel digestion

Peptide fragments

Peptide selection

Mass spectrometry

Peptide mass fingerprinting

Peptide fragmentation

Peptide sequence determination

Database search and protein identification

FIGURE 3 Protein separation and identification using 2D SDS-PAGE electrophoresis. Spots of interest are identified, excised from the gel, and subjected to in-gel digestion with a protease with well-characterized cleavage sites such as trypsin. The peptide fragments generated are then subjected to separation and analysis by mass spectrometry. The array of peptide fragments produced, or peptide mass fingerprint, may be used to identify the protein directly by searching databases consisting of predicted cleavage sites of proteins predicted from genomic sequencing. Alternately, tandem mass spectrometry may be used to directly determine the sequence of the peptides. The peptide sequence can then be used to identify the protein.

poorly represented in the gel. A third problem is the limited dynamic range of protein detection. Protein concentrations are estimated to vary from a few copies per cell to over a million of a single protein species. Highly expressed proteins can be readily visualized but there is no technique equivalent to PCR that allows amplification of minor protein species. Simply loading more protein is not an option as trailing on the part of the major species would obscure detection. A fourth problem is the inherent variability in the separations of proteins in replicate samples. This is mitigated somewhat by using commercially prepared gels, but the actual running conditions of the gel can lead to differences in temperature gradients across the gel as a function of how many gels are run concurrently, ambient room temperature, and other factors. Comparison of replicate gels requires use of sophisticated morphing or warping software that can be time consuming. An alternate approach is to use a double dyebinding method. In this methodology, proteins in two samples of interest to be compared are each labeled with a different fluorescent dye or isotope (Jiang and English, 2002; Smith et al., 2002; Zhou et al., 2002);

Functional Genomics of Wine Yeast Saccharomyces cerevisiae

85

the protein samples are then mixed and run on the same gel. Direct comparisons of differences in fluorescence can be used to determine the ratio of expression of the protein between the two samples following the same logic as the double dye-binding transcriptome technologies (reviewed in Graves and Haystead 2002). The final major problem with 2D SDS-PAGE-based methodologies is that they are labor intensive and time consuming and challenging to automate. To address some of these limitations, nongel technologies are being developed and utilized for protein analysis. These methods rely on other types of separation protocols, liquid chromatography (LC) or capillary electrophoresis (CE). These methods can also be adapted to a multidimensional format. Multidimensional liquid chromatography tandem mass spectrometry (MudPIT) employs a strong cation exchange and reverse phase separation technology and has been successfully applied to yeast proteomics (Link et al., 1999; Washburn et al., 2001) Capillary isoelectric focusing coupled to capillary reversed phase electrophoresis has been used to separate yeast proteins and peptides (Chen et al., 2003; Haynes et al., 1998). In this case, proteins can be focused or concentrated in the first phase, then separated in the second phase, allowing detection of a much wider set of proteins from the mixture. Breci et al. (2005) compared several methods of protein separation and detection using the same yeast extract: nanoflow liquid chromatography-liquid chromatography tandem mass spectrometry [nano LC/LC MS/MS (MudPIT)], nano LC MS/MS with gas phase fractionation by mass range selection (the same technique but with gas-phase fractionation by ion abundance selection), 2D SDS-PAGE coupled to in-gel digestion and nano LC MS/MS of gel slices, and isoelectric focusing followed by nano LC MS/MS of gel slices. The 2D SDS-PAGE method led to the identification of the greatest number of proteins and allowed higher sequence coverage of individual proteins (more peptides visualized per protein), making protein identification more robust. It is also possible to digest a complex mixture of protein prior to separation and identification of peptides using a tandem MS/MS protocol. This allows identification of proteins in a specific mixture but is not directly quantitative (Patterson and Aebersold, 2003). The intensity of a peptide does not reflect its relative concentration as different peptides may produce different signal intensities depending on their composition or the matrix in which they are separated. Various technologies to allow comparative analysis of peptide abundance in two samples have been developed. These methods rely on the use of different isotopes used as label for each sample, and then peptides from the same protein can be identified from the discrete change in mass resulting from the particular label used for that mixture. In one specific example, termed ICAT analysis, a cysteine reactive molecule, labeled with either light or heavy deuterium, is used creating a mass tag for all cysteine-containing

86

Linda F. Bisson et al.

peptides (Gygi et al., 1999a). Peptides containing the tag can be isolated by avidin affinity chromatography, reducing the number of peptides analyzed. These newer methodologies are not ideal either and suffer from their own limitations. For many of these techniques, some fractionation or reduction in the number of proteins or peptides identified is necessary prior to or during analysis. These methods also may not detect all species in a mixture, depending on the nature of the separation and the biochemical properties of the protein. Another method using protein ‘‘chips’’ has also been developed (Haab et al., 2001; Zhao et al., 2001; Zhu et al., 2001). These chips contain ordered arrays of protein-binding molecules with the proteins then bound or enriched on a segment of the chip to be analyzed via various technologies. Antibodies recognizing specific protein sequences or modifications can be used, for example, to separate proteins with a specific activity from a complex mixture. A different version of a protein chip has been created by Zhu et al. (2001) and by Martzen et al. (1999). In these cases, the proteins are arrayed on the chip and the proteome can then be screened for interaction with particular substrates and ligands, and for certain classes of biochemical properties and activities. Often it is of interest to compare the relative levels of proteins across strains or conditions of growth. There are several methods available for these types of analyses (Table 4) (reviewed in Graves and Haystead, 2002). Individual 2D SDS-PAGE gels can be run with the gels and then directly compared using spot identification software and gel morphing or TABLE 4 Comparative analyses of proteome profiles Method

Process

Quantification

2D SDS-PAGE gel

Morph/warp-stained gels for constellation identification and spot matching

Relative intensities of spots (stained or labeled proteins)

Protein chip

Proteins labeled and blotted against chip

Relative intensity at a specific chip locale correlates with relative expression level

Differential protein targeting

Protein species tagged using tags of differing molecular weight; protein samples then mixed

Mass spectrometry analysis of peaks separated by the precise molecular weight of the tag

Functional Genomics of Wine Yeast Saccharomyces cerevisiae

87

warping tools based on constellation (characteristic subclusters of proteins) matching. Alternatively, protein samples may be differentially labeled in two extracts, the two extracts mixed, and the proteins blotted against a protein chip. The relative intensities of the two labels are correlated with the relative concentrations of the two proteins. Mixed extracts from two differently labeled samples can also be subjected to protein digestion and their peptide fragments analyzed directly by mass spectrometry. Peptides derived from the identical proteins in the two extracts will be separated from each other by the precise molecular weight of the tag. The sequence of peptides of interest can then be determined if quadrupole mass spectrometry is used. An obvious question concerns the degree of correlation between transcriptome and proteome data. Several yeast studies have addressed this topic with different conclusions having been reached. Gygi et al. (1999b) and Griffin et al. (2002) found insufficient correlation between 2D SDS-PAGE and SAGE data (both of which are biased toward abundant proteins and mRNA species) to allow prediction of protein content from transcriptome data. Other studies have found a statistically significant correlation between protein and mRNA abundance (Futcher et al., 1999). These studies are not actually in conflict. There does seem to be a correlation between protein and mRNA abundance, just not one that allows prediction of protein level from mRNA level. Belle et al. (2006) explored this issue from the perspective of including changes in protein turnover rates. These researchers took advantage of an epitope-tagged library for quantification of protein half-lives and found that proteins generally fell into one of two categories: optimized for efficient production or optimized for efficient regulation. They found a correlation between half-life and transcriptional regulation in some sets of coregulated genes. Transcriptional regulation may be used to buffer the consequences of changes in protein half-life so that overall protein content remains the same (Belle et al., 2006). Changes in levels of highly abundant proteins may be difficult to detect over a short time course, so there is a dynamic interaction between mRNA and protein stability. Obviously, how a study is conducted will influence the conclusion regarding the relationship between protein and transcript abundance. If cell growth has been arrested, for example, both processes may be in a state of flux as the cell adapts to the new environmental conditions and correlations may be difficult to detect.

3. Metabolome analyses The third level of genome-wide analyses of cells following the transcriptome and proteome is that of the metabolome. The ‘‘metabolome’’ refers to the entire complement of all the low- and intermediate–molecularweight metabolites inside a cell suspension of interest, but for practical reasons, subsets of metabolites (metabolic profiles) are more typically

88

Linda F. Bisson et al.

obtained with each screening technique. There is a growing consensus that knowledge of transcript and protein profiles is insufficient to predict actual metabolite levels, and without this information, whole cell metabolic reconstruction and modeling will not be possible. Similar to proteomics, issues of analyte complexity confront the field of metabolomics (Fiehn, 2002). The chemical diversity of components to be evaluated makes selection of a single profiling technology difficult if not impossible. Differential extraction, chemical reactivity, and sample preparation impact the relative levels of compounds in a given sample; so with current techniques, it is not possible to display all cellular components. Several technologies, as described below, are being developed for metabolomic investigations. Metabolites produced in microorganisms (referred to as secondary metabolites) are also an invaluable source of useful compounds, including pharmaceuticals, toxins, and other chemicals (Peric-Concha and Long, 2003). Only recently have analytical techniques progressed to a level that broad metabolic profiling has become a reality and has evolved into the science of metabolomics. For instance, over 700 different biochemical compounds have been identified within a single bacterial species alone (Nobeli et al., 2003). How to measure these compounds, how to identify the molecular structure of each individual compound, how to place each individual compound in the relevant biosynthesis pathway, and how each compound relates to functional properties within an organism or for human/animal health and nutrition are all different aspects of metabolomics. To monitor, in parallel, hundreds or even thousands of metabolites, high-throughput techniques are required that enable screening for relative changes rather than absolute concentrations of compounds. Most analytical techniques for profiling small molecules consist of a high-performance liquid chromatograph (HPLC) or a gas chromatograph (GC) coupled to a mass spectrometer (MS). These techniques include GC-time-of-flight-MS (GC-TOF MS) (Fiehn, 2002), LC-MS (Lenz et al., 2004), direct-electrospray MS (Allen et al., 2003), CE-MS (Soga et al., 2003), multidimensional chromatographies linked to MS (Blumberg, 2003), and the newly emerging Fourier transform ion cyclotron resonance MS (FTICR-MS) (Aharoni et al., 2002). Other techniques such as Fourier transform infrared (FTIR) (Oliver et al., 1998) and nuclear magnetic resonance (NMR) spectrometries (Raamsdonk et al., 2001) are also popular. Somewhat in contrast to transcriptomics and proteomics, the great chemical heterogeneity of the metabolome means that no single method will realistically capture all metabolites. Mass spectrometers are generally more sensitive and more selective than any other types of detectors. When coupled with the appropriate sampleintroduction and ionization techniques, mass spectrometers can selectively analyze both organic and inorganic compounds (Baldwin, 2005).

Functional Genomics of Wine Yeast Saccharomyces cerevisiae

89

Nevertheless, prior to detection, the metabolites have to be separated by chromatographic techniques that are coupled to the mass detector. GC is used to separate compounds on the basis of their relative vapor pressures and affinities for the material in the chromatography column, but is restricted to compounds that are volatile and heat stable (Littlewood, 1970). Most biological compounds, such as sugars, amino acids, and organic acids, are not sufficiently volatile to be separated by GC in their native state and must therefore be derivatized prior to GC separations (Gehrke et al., 1987). This can be time consuming and adds to the difficulty of analyzing these compounds. HPLC separations are better suited for the analysis of labile and high-molecular-weight compounds and for the analysis of nonvolatile polar compounds in their natural form (Unger, 2002). Although GC- and HPLC-based profiling techniques are not truly quantitative, the compounds detected and their relative amounts may be compared between studies by employing the proper standards. This is in contrast to NMR techniques that can provide truly quantitative measurements, yet NMR cannot provide the sensitivity offered by MS. The high-throughput screening with GC- and HPLC-MS techniques also generates large volumes of analytical data that require advanced informatics technologies to organize vast amounts of information. NMR techniques offer several advantages for metabolomics data acquisition. This technique requires minimal or no sample preparation, is nondestructive, and can be implemented in a noninvasive manner. Therefore, NMR is useful for studies of biofluids, for cell extracts, and for cell cultures and tissues in vitro or in vivo (Serber et al., 2005). The multinuclear capabilities of NMR provide various means to observe different chemicals. Metabolomics work in NMR has mainly used 1H NMR since no labeling is necessary, but other nuclides (e.g., 13C, 31P, 15N, 19F, 23 Na, and 2H) may provide additional information about various metabolite pools in microbiology (Grivet et al., 2003). The limited availability and high cost of labeled compounds can offset the usefulness of some of these nuclides. The other major limitations of NMR relate to spectral resolution and sensitivity, both of which are improved by experimentation at high magnetic field strengths. Broadened NMR spectral lines can degrade resolution and the ability to differentiate metabolite signals (Wang et al., 2003). Factors influencing the NMR line widths are due to molecular dynamics and include sample viscosity, macromolecules, binding of small molecules, compartmentalization, and sample heterogeneity, which are inherent problems with cells and tissues. Finally, metabolic analysis results in large collections of data. Datamining techniques reduce complexity by focusing on the information content of a given data set (Fiehn et al., 2000). Clustering or principal component analyses are among the main approaches that are used for data analysis (Glassbrook et al., 2000). The integration of data from

90

Linda F. Bisson et al.

metabolomics to data from transcriptomics and proteomics still represents an intriguing challenge. Private companies like Beyond Genomics and Paradigm Genetics are trying to develop comprehensive bioinformatics systems to analyze and interpret these large sets of data (Harrigan, 2002).

4. Systems biology Systems biology endeavors to integrate diverse types of genomic, proteomic, phenomic, metabolomic, and biochemical data to create a comprehensive description of a biological system (reviewed in Alberghina et al., 2005; Fo¨rster et al., 2003; Patterson and Aebersold, 2003; Van Speybroeck et al., 2005). The ultimate goal of systems biology is the accurate prediction of cellular behavior using computational or noncomputational methodologies. When system simulations match system observations, it can be concluded that the existing data set accurately explains the operations of the organism. Of course, attaining this perfect match is a futuristic goal (reviewed in Alberghina et al., 2005; Patterson and Aebersold, 2003; Van Speybroeck et al., 2005). Some success has been attained as in the detailed description of the regulatory and metabolic response to saline stress from a systems approach (Warringer et al., 2003). The limitations of existing data sets and technologies, such as the persistence of genes of unknown function, are only now being realized and impede accurate simulation (Wu et al., 2002). Even more comprehensive studies, such as the systematic evaluation of multiple deletants, are being undertaken (Tong et al., 2001), but these endeavors are quite tedious. Depending on the regulatory mechanisms in play, there are cases with good correlation between transcriptome, proteome, and metabolome data; while in other growth conditions, this is not the case (Lafaye et al., 2005). The importance of noise in gene expression and population variation is becoming more appreciated. General principles of metabolic gene regulation are emerging from the integration of global transcript data with transcriptional regulatory mechanisms (Ihmels et al., 2004). Chromatin remodeling may be a slow process (Paulsson, 2004) leading to variation in mRNA levels across a population. Increased translational efficiency, as would occur under active growth conditions, when coupled to variation in timing of transcriptional initiation, leads to an amplification of the noise in protein production patterns (Blake et al., 2003). Similarly, conditions that negatively impact translational capacity, as occurs during glucose depletion in yeast (Ashe et al., 2000), would also impact the influence of noise in transcript data. Often, these factors were not taken into account in experimental design. Nonetheless, this integrative systems approach is vital to understanding the deficiencies of existing analytical tools and will spur their improvement or further development.

Functional Genomics of Wine Yeast Saccharomyces cerevisiae

91

III. FUNCTIONAL GENOMIC ANALYSIS OF WINE YEAST Functional genomic technologies developed with laboratory strains have been applied to both commercial and native isolates of S. cerevisiae. In addition, these tools are being employed to better understand yeast in the baking and brewing process, and for the production of biofuels. Although in most cases no analysis of the differences in genomic structure of the analyzed strain versus the popular laboratory strain (S288C and its derivatives) has been conducted, some useful information can be culled from these studies. This is especially true if the analysis is focused on overall changes in pathway or process behavior and physiological profiling, and not on a comparative quantitative assessment of single proteins or the expression levels of individual genes. It is also important to have a solid understanding of the limits of current technologies as discussed above, in the interpretation of data across platforms and strains.

A. Transcript profiling Transcript profiling studies of wine strains of S. cerevisiae have largely focused on three areas: the comparison of wine strains to laboratory strains and to each other under various growth and environmental conditions, the profiling of a ‘‘normal’’ grape juice fermentation in both synthetic media and actual juices, and the analysis of the impact of normally occurring stress conditions on the wine yeast transcriptome. In the natural environment, S. cerevisiae encounters several types of stress but the most common and universal are high osmolarity, nitrogen limitation, and high ethanol concentrations.

1. Comparisons of laboratory and wine strains Functional genomic technologies have been utilized to profile genetic variation in domesticated and wild populations of S. cerevisiae (Fay and Benavides, 2005a,b). The oldest lineages and the greater variation were found among strains from sources not related to grapes or wine. There was surprisingly little variation among grape isolates, suggesting that these yeasts are highly related evolutionary in spite of being found at great geographical distances. Genomic analysis suggests that the wine strains are derived from wild populations of yeast (Fay and Benavides, 2005a). Variation in gene expression was also evaluated in a subset of these strains by Fay et al. (2004). In this study, the expression profile of nine strains grown in the presence of copper sulfate was investigated. Copper is commonly used in winemaking to remove sulfides that have been formed as a consequence of yeast metabolic activity. Over 600 genes showed variation in gene expression among these strains, with only a small subset varying in response to copper addition (Fay et al., 2004).

92

Linda F. Bisson et al.

A detailed analysis of a genetic cross between a laboratory strain and a vineyard isolate has revealed crucial information on the role of genetic architecture in naturally arising variation in gene expression (Brem et al., 2002, 2005; Ronald et al., 2005; Storey et al., 2005). This series of studies used oligoarray transcript profiling to identify pairs of alleles displaying allele-specific differences in expression. Allelic differences were detectable by changes in the coding sequence. By using oligonucleotide arrays, it was possible to differentiate changes in expression due to regulatory effects by using tools developed to model the energy of probe-target sequence duplex formation (Zhang et al., 2003). Both cis- and trans-factors were responsible for the observed variation, with cis-regulatory variation being the most common facet (Brem et al., 2002). The occurrence of transregulatory differences was rarer, but with more widespread consequences for the organism (Brem et al., 2002). The differential level of gene expression can then be used as a heritable trait in genetic analyses. This subsequent segregation analysis revealed that many gene expression traits are linked to two or more genes (Storey et al., 2005). It was also found that the progeny of a cross of two parents may display a wider range of expression profiles than that of either parent, termed transgressive segregation (Brem et al., 2002). Thus, the underlying cause of an observed difference in expression between laboratory and wine yeast, or among wine yeasts, may be complex in nature. A variety of strains have been used in the comparative analyses of wine strains of S. cerevisiae to laboratory strains, most commonly to S288C as the genome of this particular strain was sequenced. In addition to strain variation, the media and growth conditions used as well as the platform differ, making comparisons across studies challenging. Only rarely have differences been confirmed using a technique such as QRT-PCR. In one study, Northern blot analysis of 49 genes found on Chromosome III was conducted comparing the common laboratory strain S288C to a commercial strain, V 5 (Rachidi et al., 2000). In standard laboratory cultivation conditions, the strains were very similar in expression profiles, but in synthetic grape juice media, the commercial strain altered expression of several genes, particularly the PAU stress response genes, not altered in expression in the laboratory strain. Hauser et al. (2001) reported that sequence homology between laboratory and wine strains is extensive, but that the differences seem to have profound effects on gene expression. They found that over 40 genes were consistently changed in expression pattern in standard complex laboratory growth media, and that changes in promoter regions, number, and location of transposable elements and gene copy number were likely responsible. Another study focused on a comparison of ‘‘flor’’ strains of S. cerevisiae. These strains have adapted to a distinct environment and form a film or ‘‘flor’’ on the air interface of wine during sherry production (Infante et al., 2003).

Functional Genomics of Wine Yeast Saccharomyces cerevisiae

93

Multiple differences in gene copy number were found, affecting 38% of the genome. Dunn et al. (2005) compared four commercial wine yeast strains to S288C. The four commercial strains displayed common as well as unique differences in expression profiles as compared to S288C. Some of these differences were again attributable to differences in gene copy number. The wine strains were very similar to each other, in spite of having been originally isolated from very different regions. The similarity may reflect the fact that commercial strains are selected for similar attributes or that wine strains are indeed quite similar to each other genomically. Many of the differences observed between the commercial strains were for expression of transporter genes. This family of genes may be particularly important in adaptive evolution to specific microenvironments. Zuzuarregui and del Olmo (2004) examined transcript profiles in strains displaying differences in fermentative ability. Those strains with more severe fermentation problems had both higher and maintained levels of mRNA as compared to strains that were able to completely consume available sugar. The strains capable of total sugar metabolism showed an intermediate consumption of nitrogen. Additionally, strains that consumed nitrogen more quickly or more slowly had reduced expression of stress genes. Two strains were not able to adapt to the high osmolarity of the synthetic grape juice media. In this study, strains were not nutrient limited, and presumably entered stationary phase because maximal cell density was attained. The expression profile on entry into nonpermissible growth conditions showed some similarity to nutrient arrest, but represents a distinct physiological state. Another observation in this study (Zuzuarregui and del Olmo, 2004) was that the appearance of aneuploidy or polyploidy may lead to altered basal or expressed levels of gene expression that prevent certain adaptive responses from occurring. These studies all indicate that, although there are differences between wine and laboratory strains of S. cerevisiae, those differences are explainable by normal processes of genetic modification and chromosomal rearrangements. It appears that the basic elements of the genome are highly conserved among S. cerevisiae strains with the differences largely explainable by known adaptive and evolutionary mechanisms. One study that analyzed genetic variation in a native vineyard strain of S. cerevisiae found far more differences among spores arising from a single isolate (Cavalieri et al., 2000). The strain was found to contain multiple heterozygosities, and over 6% of the genome showed a significant change in expression pattern across the different spore types obtained. The major differences occurred in protein degradation and amino acid and sulfur metabolism (Townsend et al., 2003). These observations suggest that genetic variation in strains in the wild may occur to a greater extent than in strains in a controlled environment, as would occur in a laboratory or in a production facility for commercial yeast products. Higher

94

Linda F. Bisson et al.

spontaneous mutagenic rates or genome tweaking may be advantageous for the natural environment by allowing more rapid rates of adaptive evolution. Investigation of the impact of change in growth environments among wine strains on global transcript patterns has also been analyzed. Rossignol et al. (2006) explored changes in the transcript profile on rehydration of commercial strains and inoculation into synthetic grape juice media. The changes observed were all predictable from studies of laboratory strain adaptations to these two types of growth environments. The rehydrated strain displays transcript profiles consistent with limitation for nitrogen and carbon under aerobic conditions (the conditions of commercial strain preparation) with a shift to a fermentative mode of metabolism on introduction into grape juice. Similarly, Roberts and Hudson (2006) examined the switch from aerobic growth on glycerol to fermentative conditions and observed similar kinds of adaptations. The conclusions reached in these studies and patterns of regulation and expression observed are in concert with known physiological adaptations observed in laboratory strains. In some cases, the details of expression of individual genes may vary, but the overall depiction of physiological activity is predictable. Of course, the danger here is that there is too strong of a reliance on what is already known about yeast biology in the interpretation of the data. Models of metabolic and regulatory behavior generated from analysis of laboratory strains appear remarkably consistent with the behavior of native and commercial isolates.

2. Assessing gene expression in the wine environment

Since S. cerevisiae is a domesticated microbe, grape juice fermentation represents an important adaptive environment for this yeast. In order to understand the physiological responses of this yeast to native conditions, several investigators have profiled yeast expression patterns in natural grape juices or in synthetic juice media with the goal of defining a ‘‘typical’’ profile (Backhus et al., 2001; Marks et al., 2003; Puig and Perez-Ortin, 2000; Riou et al., 1997; Rossignol et al., 2003; Varela et al., 2005; Zuzuarregui et al., 2006). Global transcript profiling has also been undertaken for brewing strains under brewing conditions (James et al., 2003). In spite of the fact that different transcript profiling platforms and a range of commercial and native isolates were used, a consistent picture of the typical gene expression changes during fermentation of synthetic or actual grape juices has emerged. A typical fermentation is defined as one in which the majority of yeast growth occurs within the first 48–72 hours, following a variable lag phase. At this point, the cells are at terminal cell density and have consumed most of the available nutrients. The cells enter a nonproliferative phase, but one during which high levels of metabolic activity are maintained. As ethanol accumulates in the environment,

Functional Genomics of Wine Yeast Saccharomyces cerevisiae

95

metabolic levels are reduced presumably due to the inhibitory effects of high ethanol. Metabolic activity continues albeit at a reduced rate until all sugar has been consumed. After exhaustion of sugar, significant loss of culture viability occurs, again most likely due to the need for sustained ATP production to counter the inhibitory effects of ethanol. Transcript profiling of each of these stages has been undertaken in order to identify key physiological markers of normal progression through fermentation. Transcript profiling has revealed that on entry into the nonproliferative metabolically active state, there is a global remodeling of ribosomal composition, translation, and mRNA processing to adapt to the new conditions. These responses likely signal exit from active growth and occur regardless of the cause of growth cessation. As fermentation progresses, ethanol stress increases, activating a stress response. This response appears to be a graded response with a gradual decrease in the expression of genes involved in biosynthesis, and global changes in transport proteins. It is clear that the cells are undergoing a gradual and continual adaptation to the disruptive effects of ethanol. There is also an increased expression of genes involved in oxidative stress response. This may appear paradoxical, given that these fermentations are largely anaerobic. However, acetaldehyde, an oxidizing agent, is an intermediate in ethanol production and may be responsible for the need to induce these pathways. The phenolic compounds found normally in grape juice can react with oxygen to produce hydrogen peroxide. Thus, even in the absence of respiration, reactive oxygen species may be present. It is also interesting that genes in the multidrug resistance pathway are also expressed late in fermentation, perhaps the true substrate of these proteins is the phenolic constituents of grape juice. Increases in expression of genes known to be involved in ethanol tolerance are observed, as is a change in the isoforms of many of the proteins of glycolysis. Genes involved in glycogen, trehalose, and glycerol metabolism also increase in expression, and these components have been shown to be important in survival of ethanol stress. It has also been suggested that they form a futile cycle with glucose degradation to allow fine-tuning of the ATP status of the cells (Rossignol et al., 2003; Varela et al., 2005). The similarities observed in these studies establish a consistent profile of changing expression patterns throughout fermentation and provide an excellent base on which to further compare strains to associate strain behavior with strain expression profile. Although different transcriptome platforms were used, many of these studies used the same commercial yeast, EC1118. Rossignol et al. (2003) examined transcript profiles for EC1118 during growth and fermentation in a synthetic grape juice media. Over 2000 genes showed a significant change in expression. The authors observed that genes in specific pathways behaved in a highly coordinated manner. Entry into stationary phase in this case was caused by nitrogen depletion of the medium.

96

Linda F. Bisson et al.

Significant changes in the transcriptome accompanied arrest of growth; however, only 30% of the induced genes corresponded to genes reported to be induced in previous reports of the stationary phase response (Rossignol et al., 2003). Of the 367 genes found in common in analyses of the common stress response genes (Causton et al., 2001; Gasch et al., 2000; Gasch and Werner-Washburne, 2002), 213 were expressed during fermentation of the synthetic grape juice. This difference may be due to the simultaneous presence of ethanol stress in addition to arrest of growth. The yeasts were still metabolically active and continued to ferment, increasing ethanol levels. Thus although the initial stress factor was nutrient limitation, as fermentation continues, ethanol stress became an equally critical factor. Although the nitrogen level used in this study was considered low by the authors, it is in fact sufficient to allow complete utilization of substrate within 5 days (Rossignol et al., 2003). All of the nitrogen was consumed within 48 hours in this study. It is not clear that the yeasts were in fact starving as this pattern of nitrogen consumption may be normal for these conditions. Studies with laboratory yeast under laboratory conditions of high nitrogen levels and low energy source concentrations suggest that in the typical cell cycle, sufficient nitrogen is consumed during the cell cycle to meet the biosynthetic needs for production of a daughter cell. When sufficient nitrogen is present within the cell, a new cell cycle is initiated. At the end of this cell cycle, nutrients must then be accumulated for a succeeding cell cycle to occur. Our observations of yeast in juice and juicelike conditions are inconsistent with this view. We have found similar kinetics of nitrogen consumption (Monteiro and Bisson, 1991a,b, 1992a,b); however, if the cell number is reduced via mild centrifugation, the cells that remain regrow to the same terminal cell density without further addition of nitrogen. This observation suggests that when an energy source is plentiful, indeed in great excess over the metabolic needs of the cell for growth, nitrogen accumulation may be uncoupled from initiation of a new cell cycle. Instead, nitrogen is consumed until depleted from the medium and stored inside of the cells, allowing subsequent multiple generations until cellular pools are depleted. These observations are consistent with earlier reports that nitrogen limitation alone does not lead to a quiescent state in S. cerevisiae (Granot and Snyder, 1993). Our conclusions regarding laboratory yeast physiology under laboratory conditions may be overly biasing interpretation of the results obtained from studies of wine yeast in juice environments. A similar study, profiling transcripts of EC1118 using the same synthetic medium as used by Rossignol et al. (2003) but a different platform, SAGE, for transcript quantification was conducted by Varela et al. (2005). Since SAGE was used, this study was able to quantify messages not represented on commercial yeast arrays. The authors found expressed

Functional Genomics of Wine Yeast Saccharomyces cerevisiae

97

sequences from intragenic regions as well as messages that did not match any known sequence in the S288C genomic sequence, particularly in stationary phase. However, they did not compare their analysis to the study by Kumar et al. (2002) that identified several similar types of nonannotated ORFs in a laboratory strain. Two other possibilities exist. Either the genome of EC1118 is not pure S. cerevisiae and instead represents a hybrid between species, or the laboratory strain S288C has lost a significant component of its genome, perhaps as a consequence of laboratory cultivation. Three independent commercial preparations of EC1118 were compared to S288C by Dunn et al. (2005), and these authors concluded that although they could not completely rule out the presence of additional genomic DNA in this strain, their analysis suggests that it is unlikely. Resolution of this issue will require sequencing of the wine yeast genome. For the genes that were represented in the S288C genome, the authors found that the majority (88.6%) were expressed at 10 copies per cell or less. This value may be below the sensitivity of array technologies, depending on how they are conducted. Many of the same gene families were identified in these two studies. Both Rossignol et al. (2003) and Varela et al. (2005) used the identical synthetic grape juice medium MS300 in addition to using the same commercial yeast strain. Varela et al. (2005) modified the medium by increasing the sugar concentration from 200 to 240 g/liter. Rossignol et al. (2003) used microarray analysis for studying the transcription profiles at six different stages during the fermentation, and Varela et al. (2005) used SAGE to profile transcript levels at mid-log phase, early-stationary phase, and late-stationary phase during the yeast growth and fermentation. SAGE analysis developed by Velculescu et al. (1995) allows identification and quantification of both known and novel gene transcripts since it is independent of the genome sequence. Both groups observed growth arrest coinciding with the depletion of assimilable nitrogen, and an increase in the expression of stress-responsive genes was observed by both methods. Similar profiles were also observed for genes involved in carbohydrate metabolism. On comparing the data, one noticeable observation is that many of the transcripts detected as significantly different by microarray were not detected using SAGE analysis. Rossignol et al. (2003) found a decrease in expression of genes involved in protein, nucleotide, and amino acid biosynthesis whereas Varela et al. (2005) saw a decrease in transcript levels of biosynthetic pathways as fermentation progressed, but were not able to detect most of the transcripts for genes involved in amino acid biosynthesis. Similarly, genes involved in putative cell wall proteins were found to be induced over time in the microarray study but were undetected by SAGE analysis. MET30, a key regulator in the methionine biosynthesis pathway, was upregulated during late-stationary phase by both the methods.

98

Linda F. Bisson et al.

The expression profiles of some of the hexose transporters were found to be inverse by the two papers. While the Rossignol group found an increase in expression of HXT3 and HXT7 over the time course of the fermentation, Varela et al. (2005) found the transcript levels of HXT3, HXT6, and HXT7 to decrease over time. Expression of genes involved in the reserve carbohydrate biosynthetic pathways also differed in the two methods. Many of these differences reflect differences in the dynamic range of the two methods. SAGE is better able to detect abundant mRNA species. Other differences are more difficult to explain, and an independent method would need to be employed to validate the observations. However, these two methods do provide similar conclusions on the physiology of yeast at different stages of fermentation. It is clear that a native yeast strain spends most of its life in a nonproliferative stage, as is thought to be true for most microbes (Palkova, 2004). In addition, environmental conditions rarely are conducive to growth at maximal rates as can be created in the laboratory environment. It is a common concern that analyses of growth and metabolic behavior in laboratory growth conditions have little to no bearing on understanding the physiological status of cells in their native environments. This is only true if one takes a restricted view of laboratory growth. As noted above, laboratory growth is typically energy source limited but macronutrient rich, in contrast to the natural environment of grape juice. As long as this difference is appreciated and its impact on the physiology of the cell understood, laboratory studies do indeed provide a valid framework for the analysis of strains in native environments. The nonproliferative phase of fermentation has been studied in detail by several authors. It has been compared to energy source depletion in laboratory strains, and many of the same genes are observed to be expressed. The nonproliferative phase of juice fermentation has been likened to stationary phase in laboratory strains. Indeed, many of the same genes are expressed under both conditions. Conventional wisdom holds that the absence of growth or limitation of the ability to grow is defined as stress. Transcript profiling has revealed many features of the nonproliferative nonquiescent fermentation stage of S. cerevisiae. On attainment of maximal cell density, further growth ceases and fermentation rate is at its maximum (Rossignol et al., 2003). As fermentation continues, the fermentation rate gradually decreases. Genes associated with cell growth and amino acid biosynthesis also are increasingly downregulated as fermentation progresses, with the exception of the methionine pathway. Since this pathway is required for the synthesis of pathway intermediates, factors needed for stress tolerance, S-adenosyl methionine, and cysteine needed for glutathione, it is not surprising that expression of these proteins is maintained. Interestingly, the expression of genes required for sterol biosynthesis also gradually decreases, explaining the failure of late

Functional Genomics of Wine Yeast Saccharomyces cerevisiae

99

oxygen additions to enhance ethanol tolerance as this is largely mediated by sterol production. There is a shift in the isoforms expressed for enzymes of glycolysis. The different isoforms may have altered function or substrate specificity, as is the case in the change of hexokinase P2 for the more fructophilic hexokinase P1, or may reflect the need for a different subcellular localization or complex. Alternately the isoforms may be more resistant to the denaturing effects of ethanol or the oxidative damage from acetaldehyde. Both Rossignol et al. (2003) and Backhus et al. (2001) found that genes involved in vitamin biosynthesis show an increased level of expression suggesting that these compounds also play a role in stress tolerance and increased expression of genes involved in nitrogen recycling. Certain heat shock proteins were dramatically induced. Interestingly, in the nitrogen-limited synthetic juice conditions (Backhus et al., 2001); a decrease in expression of genes involved in growth was not seen. This is again consistent with the observation that nitrogen limitation does not lead to a quiescent state (Granot and Snyder, 1993). Several gene candidates as markers for normal fermentation progression have been proposed from these studies. The heat shock protein encoding genes: HSP12, HSP26, HSP30, and HSP82, show specific increases in expression at specific times during the fermentation as ethanol increases. HSP12 and HSP26 are expressed late in a normal fermentation that is accumulating ethanol (Backhus et al., 2001). Strains with higher basal and induced levels of HSP12 were found to resist stress more effectively (Ivorra et al., 1999). HSP12 and HSP26 expression increased during low-temperature stress, in contrast to many other heat shock genes, and may therefore represent generic markers of cellular stress response (Sahara et al., 2002). HSP30 appears to be expressed to a greater extent in nitrogen limited than in nitrogen sufficient fermentations (Backhus et al., 2001). It may be challenging however to develop absolute measures of expression of these genes that would indicate either normal or aberrant stress response was occurring due to differences in basal levels of expression.

3. Analysis of stress responses in wine yeast

One of the major areas of research interest in wine strains of S. cerevisiae is the analysis of response to stress. This focus is motivated by both practical and fundamental interests. The production of wine imposes both biotic and abiotic stresses on the yeast. The principal stresses encountered are high osmolarity, high ethanol, extremes of temperature, nutrient limitation, and presence of inhibitory metabolites (Bisson, 1999). Genomic analysis of the response to each of these types of stress has been conducted (Alexandre et al., 2001; Aranda and del Olmo, 2004; Backhus et al., 2001; Erasmus et al., 2003; Kuhn et al., 2001; Marks et al., 2003; Rep et al., 2000; Rossignol et al., 2006; Sahara et al., 2002). Several

100

Linda F. Bisson et al.

excellent reviews on the yeast stress response have been published (Gasch, 2003; Gasch and Werner-Washburne, 2002; Gray et al., 2004; Siderius and Mager, 2003). The existing environmental growth conditions can profoundly influence the stress response (Siderius and Mager, 2003). Plasma membrane composition at the time of stress can impact detection of stress and signal transduction, and the availability of nutrients can be important for the synthesis of stress response factors. Even under permissive growth conditions, the stress response on rich media (YPD) varies from that on minimal media (YNB) (Siderius and Mager, 2003). Environmental conditions may also mitigate a stress response. Davidson and Schiestl (2001) found that cells were more heat tolerant under anaerobic conditions than under aerobiosis. This observation suggests that a primary consequence of heat exposure is the release of a reactive oxygen species that accompanies disruption of metabolically active mitochondrial membranes. Studies of arrest of growth in laboratory strains, typically due to energy source depletion, has led to the identification of a set of genes expressed in stationary phase. Many of these genes are expressed on exposure to stress as well as a DNA element, the stress response element (STRE), which has been identified in the coding regions of many of these genes (Kobayashi and McEntee, 1990, 1993; Marchler et al., 1993). Many, but not all, of the genes with STREs are expressed in the later stages of fermentation (Puig and Perez-Ortin, 2000). Gasch et al. (2000) in a detailed study of responses to a series of stress situations, including temperature shock, osmotic shock, nitrogen depletion, and the presence of various drugs and inhibitors, defined elements of the ‘‘environmental stress response’’ (ESR). Approximately 900 genes displayed a similar response, either increase or decrease in expression, across all stress conditions. Approximately 600 genes decreased in transcript level, either due to repression or increased message turnover rates. The majority of these genes (70%) are involved in protein synthesis (Gasch, 2003). The remaining genes are associated with cell growth. Of the roughly 300 genes showing an increase in transcript level, either due to increased expression or message stabilization, 45% are genes of unknown function (Gasch, 2003). The remaining genes impact carbohydrate metabolism (primarily in the change of isozyme species produced), detoxification of reactive oxygen species and mitigation of oxidative damage, metabolite transport, protein folding and degradation, DNA repair, and cytoskeletal reorganization. In addition, there were specific patterns of expression unique to each type of stress. Other studies of stress responses in laboratory and wine strains have largely found identical results (Causton et al., 2001; reviewed in Gasch, 2003; Gasch and Werner-Washburne, 2002). In general, the imposition of stress results in a transient adaptive phase to the new growth conditions or results in transit of the yeast from growth to

Functional Genomics of Wine Yeast Saccharomyces cerevisiae

101

a nonproliferative state. When the cells are returned to permissive conditions, the transcript profile returns to the nonstressed state (Gasch and Werner-Washburne, 2002). The principal transcriptional regulators involved in the general ESR are Msn2p and Msn4p and the heat shock factor Hsf1p (Boy-Marcotte et al., 1998, 1999). The ESR is a graded response. The primary goal of the response is to allow adaptation to the new growth conditions to maintain optimal cellular performance or, failing that, to equip the cell for entry into a true resting stationary phase (Figure 4). Thus, the overlap with changes observed during the progression of fermentation is to be expected. In this case, there is a gradual, yet steady, increase in the level of stress and the cell’s ability to acquire tolerance and continue growth or, at later stages, metabolism (Figure 4). Regardless of the nature of the stress, some of the cellular consequences are identical, so it is not surprising that there is a common set of gene expression changes in the response to stress. The imposition of stress results in the inability to continue to function in the existing physiological state. The first phase of this response would then be to mitigate the effects of the stress by ceasing transcription and translation, DNA synthesis, arresting protein localization and organelle biogenesis, and redirecting available resources. The initial aim of the stress response is to repair damage; to restabilize cellular structures, membranes, metabolite gradients, and pools; and to acquire tolerance of the new condition in

Entry into stationary state

Stress

Stress detection

Modification of transcript protein and metabolite profile

Resumption of growth/metabolic activity

Repair of damage

Acquisition of tolerance

FIGURE 4 Cyclical nature of the adaptation to stress during wine fermentation. Cells continually monitor and adapt to changing environmental conditions with the goal of restoring a permissive growth or metabolic state. When acquisition of tolerance is no longer possible, cells will enter a resting or quiescent state.

102

Linda F. Bisson et al.

order to recommence growth and/or metabolism. Alternately if the stress is severe, the aim is to attain a resting or hibernating state that protects cellular functionality and viability at the expense of growth. There appears to be a delicate balance between loss of ability to proliferate and expression of ESR genes with varying consequences. For example, one key component of the ESR is to express genes involved in synthesis of glycogen, glycerol, and trehalose. These components may serve as energy reserves or alternately fine-tune ATP levels based on a dynamic fluctuation between glycolysis (ATP generation) and energy reserve compound synthesis (ATP consumption). These compounds, however, have also been shown to play roles in stabilization of cellular structures due to the change in temperature (stabilize membranes), loss of water (osmotolerance), or loss of water of hydration (ethanol tolerance). Trehalose binding to proteins has been shown to protect these molecules against oxidative damage (Benaroudj et al., 2001). Such interactions would also be predicted to limit function of the macromolecule affected. Thus, restoration of growth may require elimination of these protective factors in order to increase rates of cellular activity. The observation that both enzymes of trehalose and glycogen synthesis and degradation are induced simultaneously may not indicate the need for fine-tuning of the ATP pool, but for maintenance of a dynamic state between a protected (sugar bound) and fully active (no sugar) state of proteins and macromolecular structures. Conditions that lead to enhanced tolerance of stress factors tend to reduce growth capacity. Different types of stationary phase have been identified in yeast (Drebot et al., 1990). These authors found, by using a specific mutation that resulted in death in true stationary phase, that there is a distinct difference between arrest of growth and entry into the classically described stationary phase. The physiological properties of arrested cells are a function of the conditions leading to arrest. The genotype may influence the type of response that can be mounted by the cell. How the cell responds to stress is also dependent on previous growth conditions and available nutrients and substrates. For example, in the presence of oxygen, nitrogen limitation leads to the expression of genes involved in respiration, suggesting that nitrogen limitation places a greater demand on the cell for ATP or limits glucose fermentation such that respiration is needed to provide for cellular energy needs. Our analysis of nitrogen limitation in a synthetic grape juice medium also demonstrated an increase in expression of genes involved in respiration in spite of the very high glucose concentrations remaining in the medium, but there was no general relief of glucose repression (Backhus et al., 2001). Stress response genes were elevated in both the nitrogen-sufficient and nitrogen-limited cultures, in the former perhaps because of ethanol stress and the latter due to nutrient limitation. In general, the response to ethanol observed in the nitrogen-sufficient culture was not seen under

Functional Genomics of Wine Yeast Saccharomyces cerevisiae

103

nitrogen-limitation. This may be because the cells had not attained a high enough ethanol concentration, or that the absence of nitrogen prevented adaptation to ethanol. In any event, respiration appears to be more nitrogen conserving than fermentation. In a parallel study, nitrogen in the form of diammonium phosphate was added during fermentation of white grape juice at the point of entry into stationary phase and the transcript profile changes evaluated (Marks et al., 2003). Approximately 350 genes changed in expression with roughly half increasing and half decreasing in expression. Not surprisingly, many of the genes increasing in expression were associated with active growth while those that decreased were associated with alternate nitrogen source use and the stress response. Two critical stressors that have been examined in industrial yeasts are the response to osmotic stress and to ethanol. Osmotic stress leads to an increase in expression of the glycolytic and pentose phosphate pathways and a decrease in expression of genes involved in biosynthesis (Erasmus et al., 2003; Rep et al., 2000; Zuzuarregui et al., 2005). In one study, osmostress tolerance was shown to be affected by mutations in the genes required for adenine biosynthesis (Ando et al., 2005). However, expression of these genes decreases in the transcriptome analysis. This discrepancy is likely due to the need to make ATP and assure sufficient energy balance for the cell under these conditions such that inability to synthesize adenine would be problematic. In contrast, since the growth rate of the cell was reduced, high levels of expression of these genes were not required. Another study involving the comparison of transcriptional response to saline stress between laboratory (FY834) and brewing yeast (IFO2347) strains found that an addition of up to 1M NaCl to complete rich media (YPD) led to an increased expression of genes involved in stress response, carbohydrate metabolism, and energy metabolism (Hirasawa et al., 2006). In addition, the brewing strain displayed increased expression of genes in vitamin metabolism, glycerol synthesis, sodium ion efflux pump, and copper binding as compared to the laboratory strain. These genes are all speculated to be responsible for the quick adaptation of the brewing strain to high concentrations of NaCl as compared to the laboratory strain (Hirasawa et al., 2006). In wine strains, the response to increasing ethanol is quite complex. This is because ethanol rarely increases in the absence of other stress factors, such as nutrient limitation, and is accompanied by acetaldehyde production which is itself a stress factor (Aranda and del Olmo, 2004). Short-term ethanol stress has been evaluated in an attempt to dissect the responses specific to this compound (Alexandre et al., 2001). Approximately 3.1% of the transcripts analyzed increased in expression in response to exposure to ethanol. Of these genes, 49.4% belong to the ESR genes; however, this represents only 73 of the 300 known ESR genes.

104

Linda F. Bisson et al.

An additional 14.4% of the genes increasing in expression have other known roles in response to stress. Genes involved in energy production, protein localization, and ion homeostasis also increase in expression. Genes decreasing in expression are associated with growth and biosynthesis. We have examined the effect of simultaneous imposition of heat and ethanol stress, as this is a common occurrence in commercial wine production, and yeast strains differ in their ability to tolerate both stressors simultaneously. In spite of crossover tolerance seen for ethanol and heat (Gasch, 2003), the presence of ethanol adversely affects temperature tolerance under winemaking conditions (reviewed in Bisson, 1999). We evaluated the effect of a moderate temperature shift, from 25 to 30
C, at 8% and 11% ethanol for three commercial strains, UCD904, UCD522, and UCD2032 (Nugent, 2004). Two strains, UCD522 and UCD2032, completed fermentation regardless of the shift in temperature. UCD904 was able to complete fermentation in the absence of a temperature shift, but arrested sugar utilization at the higher temperature. Interestingly, nearly 70% of the differences that we observed between strains were in genes of unknown function. All three strains were less transcriptionally active at 11% ethanol. In a parallel study (Mangahas, 2003) using two different strains, UCD905 and UCD2031, there were clearer differences between the temperature-tolerant and less-tolerant strains. The more tolerant strain had both higher basal levels of expression of stress genes and a higher level of a stress response. One of the important issues in interpretation of these types of studies is to understand the reason for the observed changes. Strains with high basal levels of expression of some stress genes show a stronger stress tolerance than strains with lower basal levels; yet if stress is imposed, the strains with the low basal levels will show the strongest induction, but not necessarily a stronger tolerance. It is likewise important to consider if the change in expression is needed to maintain a steady state of protein rather than relative protein concentration changing. In this case, the change in expression may appear to be coregulated with the stress response, but this is coincidental. Protein turnover rates and efficiencies of translation may also differ between environmental conditions, and changes in expression may occur simply to counter these effects (Kuhn et al., 2001; Sahara et al., 2002; Stahl et al., 2004).

B. Proteomics Proteomic analysis has been applied to laboratory, commercial, and native wine yeast strains predating completion of the sequence of the yeast genome, but not to the same extent as transcriptome profiling (Norbeck and Blomberg, 1997; Pardo et al., 1999). The studies conducted

Functional Genomics of Wine Yeast Saccharomyces cerevisiae

105

can be divided similar to transcript profiling: comparison of strains, analysis of the normal proteome of fermentation, and the presence of stressors in the environment. In this latter case, one major goal has been to identify key proteins that could then be used in a rapid assay, such as an immunochemical assay, to monitor the status of fermentation. An alternate goal is to use protein or peptide profiles to monitor the microbial complexity of fermentation, although this has been more strongly pursued in brewing strains (Dowhanick et al., 1990). The lack of studies on wine proteomics is likely due to the fact that the majority of proteins detectable on the gels are involved in glycolysis and that the major focus in yeast proteomics remains technique development and optimization (Joubert et al., 2001; Kobi et al., 2004). Comparisons of wine and laboratory yeast strains has identified electrophoretic protein variants unique to wine strains (Brousse et al., 1985). Comparative profiling of two wine strains with different fermentation properties has also been undertaken (Zuzuarregui et al., 2006). This study included transcript profiling as well. The two strains selected had similar nitrogen consumption profiles, growth rates, and attained similar levels of biomass. However, one strain, ICV27, was unable to complete the fermentation leaving high residual levels of sugar while the other, ICV16, had no difficulty finishing fermentation. At the point of entry into stationary phase, ICV27 displayed a decrease in fermentation capacity and glucose consumption eventually leading to arrest of sugar uptake. Although different sets of genes and proteins were observed in the transcript and protein profiling, the overall conclusions reached from each analysis were similar. The ICV16 strain showed higher expression of genes and proteins involved in carbohydrate metabolism, telomere maintenance, cell wall and sterol metabolism, and increased expression of stress factors associated with protein degradation. These changes are all predicted from a strain that is normally progressing in fermentation. In contrast, the ICV27 strain showed greater levels of expression of genes involved in gluconeogenesis, ATPase activity, oxidative and osmotic stress, and transporter proteins, particularly those for amino acids. This profile is quite informative. It suggests that this strain is struggling to deal with excess proton entry into the cell, a hallmark of ethanol intolerance, and increased oxidative damage likely due to increased levels of acetaldehyde. Although there was no congruence between the transcripts showing the most dramatic effects and the proteins, independent analysis of both sets of data would lead to the same physiological conclusions. Trabalzini et al. (2003) also profiled changes in the proteome as cells entered stationary phase under enological conditions. They observed three classes of proteins: repressed, induced, and autolyzed. They most frequently observed that protein spots were either present or absent, not reduced or increased in expression. The largest changes in expression were for proteins of glycolysis. In our studies

106

Linda F. Bisson et al.

of wine strain proteomics (Cooney, 2003; Olineka et al., 2005; Weiss and Bisson, unpublished observations), we have observed a similar trend. As fermentation progressed, there was a change between isoforms of glycolytic enzymes as well as the appearance of specific degradation products. Our analysis suggested that these degradation products can be quite stable over several days of the fermentation, suggesting that continued protein degradation has been impaired or that these protein fragments play some other role in the cell, perhaps as a storage form of nitrogen to augment the vacuole. We found specific degradation fragments from the C-terminal end, the N-terminal end, and both ends using MALDI-TOF MS (Weiss and Bisson, unpublished observations). Larsen et al. (2001) investigated the appearance of multiple forms of enolase during fermentation and found identical results to that of our own. They reported 11 forms of enolase appearing during fermentation with modifications other than common degradation. Imposition of sulfur stress through the inclusion of cadmium in the medium has also been explored (Fauchon et al., 2002). In this case, the proteome was modified to switch between glycolytic protein isozymes to produce versions with less S-containing amino acids. This study suggests that one role of isozymes may be to maintain metabolic rates in the presence of nutrient limitation. Proteome analysis has also been used to evaluate the response to sorbic acid stress (De Nobel et al., 2001), a compound used to stabilize wine against further yeast growth postfermentation. These authors also conducted transcript profiling and concluded that both analyses provided unique insights into the physiological processes disrupted in the yeast strains.

C. Metabolomics One of the most seminal metabolomics papers concerning laboratory yeast was published a few years ago in Nature Biotechnology. Raamsdonk et al. (2001) established a technique to reveal the phenotype of silent mutations in the yeast genome. They have developed functional analysis by coresponses in yeast (FANCY), which is based on the measurement of the steady-state response of two variables to the change of a system parameter, in this case, the change in concentrations of two metabolites in response to a mutation. The lack of an observable phenotype, such as a change in growth rate, due to a mutation can be explained by the presence of another gene (or genes) in the genome that can substitute for its function. The premise for this approach is that similar responses can be produced due to mutations in the cell that have the same functional response. Therefore, matching the metabolic profile of known gene mutations with those associated with genes of unknown function can reveal the function of unknown genes.

Functional Genomics of Wine Yeast Saccharomyces cerevisiae

107

The authors used two deletion mutants in the glycolysis pathway to demonstrate this powerful technique. The two mutants without growthrate phenotypes showed metabolic phenotypes when concentrations of six metabolites were measured (Raamsdonk et al., 2001). To analyze all the metabolites in the cell at the same time, they used NMR spectroscopy and found that changes in the concentrations of metabolites were similar in these two mutants. The authors concluded that mutant strains containing defective genes involved in similar pathways displayed metabolite profiles that could be clustered together using principal component analysis. Another approach for the identification of individual metabolites in yeast was undertaken by Mashego et al. (2006). The authors used specialized perturbation experiments to elaborate the kinetics of metabolism and identify targets for metabolic engineering in laboratory yeast, but used classic techniques to differentiate between intracellular and extracellular metabolites. The authors found that when the chemostat conditions were varied from aerobic to anaerobic, there was a large effect on the glycolytic flux. In addition, the authors critically evaluated several different sampling techniques in a previous paper (Mashego et al., 2003) and found that quenching metabolism by exposing the cells to precooled stainless steel beads worked the best. There has been a growing application of metabolomics to wine yeast and fermentation as more people have recognized the power of these techniques. One application of metabolomics has been to study the interactions of yeast during fermentation. Howell et al. (2006) used metabolic profiling to study how multiple strains of Saccharomyces spp. grown together in grape juice affect the flavor and aroma compounds in the fermented wine. Another interesting application of metabolomics to wine yeast was the investigation of the central carbon metabolism in yeast during fermentation (Camarasa et al., 2003). Alcoholic fermentation takes place in almost complete absence of oxygen and in the presence of a large glucose concentration. As a consequence, many enzymes are repressed and the carbon flux through the tricarboxylic acid (TCA) pathway is severely reduced. Using [3–13C]-aspartate as the sole nitrogen source, these authors showed that the TCA cycle is in fact interrupted. Malate and succinate were both labeled on C2 and C3, consistent with a synthesis by the reductive branch of the TCA cycle (oxaloacetate to malate and succinate). When [3–13C]glutamate is used as a substrate, only labeled succinate is produced by the oxidative branch of the cycle (through ketoglutarate). NMR evidence alone did not pinpoint the site of interruption, but comparison with mutants proved that the succinate dehydrogenase complex is in fact inactive (Camarasa et al., 2003). The metabolomic techniques described so far are even easier when applied to extracellular or secondary metabolites to gain information on

108

Linda F. Bisson et al.

yeast from the medium. Allen et al. (2003) employed noninvasive, mass spectrometric monitoring to measure metabolites in spent culture medium. The authors found that this ‘‘metabolic footprinting’’ allowed them to distinguish between different physiological states of yeast growth and even between different yeast single-gene deletion mutants. This line of research has also been applied to the analysis of musts for brewing and wine yeast fermentations. Secondary metabolites play a large role in the quality of the final fermented product. For example, bitter acids are ubiquitous in beer and contribute to astringency, while esters and higher alcohols are desirable metabolites in wine that markedly influence flavor and depend on the presence of precursor compounds in the grape must. Kaukovirta-Norja et al. (2004) studied the effects of germination of oat and barley on the metabolite content of grains used in beer fermentation. One of the main goals of germinating, or ‘‘malting,’’ these grains is to produce nutrients for brewing yeast, but this process also releases secondary metabolites that have flavor- and color-enhancing effects. Metabolomics can be used in this capacity to analyze beer worts in preparation of brewing and to glean information on how the wort affects the final product.

IV. CONCLUSIONS Functional genomic technologies have been productively applied to the investigation of both commercial and native wine yeasts under a variety of conditions. Although the specific details of which genes or proteins may be changing in concentration and relative levels may differ, the physiological depiction that emerges is consistent. Wine strains are similar in gene regulation and protein expression patterns to laboratory strains, as would be predicted from the similarity of their genomes. Differences that have been observed are due to the natural processes of mutation and genome rearrangement, phenomena that underlie adaptive evolution to restricted microenvironments. The importance of cis- and trans-regulatory variation in gene expression has been documented suggesting that a ‘‘typical’’ expression profile will vary depending on the specific changes in genetic architecture inherited by a given strain. The behavior of yeast in commercial environments is largely predictable from the vast data available for laboratory strains, a fact that is really not all that surprising. The physiological profile during fermentation of grape juice changes from an initial adaptive response to high-sugar conditions, active growth, arrest of cell proliferation, and a graded stress response as nutrients become limiting and ethanol accumulates in the environment. Strains that show a deficiency in growth or fermentation in the native environment largely fail to undergo this adaptive regime. In general, strains with higher basal levels of expression of genes associated

Functional Genomics of Wine Yeast Saccharomyces cerevisiae

109

with the stress response survive and tolerate stress in the environment more effectively than strains with lower levels of expression. However, strains with high basal levels of stress gene expression do not initiate growth quickly, thus explaining the persistence of such genetic diversity in the wild. Functional genomic technologies will greatly aid in further study of the biological properties of wine yeast. Of the 6604 ORFs in the Saccharomyces genome, about 20% of the gene products are of uncharacterized function (www.yeastgenome.org). Many of these genes of unknown function appear to be expressed during growth and metabolism of wine strains, and wine strains may therefore be exploited to define the role of these genes in the biology of the organism. Genes important to flavorant production will be able to be identified, and metabolism will be understood at a more detailed level. Equally important will be the understanding of the ecology and evolution of Saccharomyces, and the forces that drive genetic change and exchange.

ACKNOWLEDGMENTS The authors thank Rebekah Karpel for assistance in editing and critical reading of the chapter and in drawing the figures. We also acknowledge the support of the American Vineyard Foundation, Maynard A. Amerine Endowment, and the California Competitive Grant Program for Research in Viticulture and Enology for support of the unpublished research described herein.

REFERENCES Aharoni, A., Ric de Vos, C.H., Verhoeven, H.A., Maliepaard, C.A., Kruppa, G., Bino, R., and Goodenowe, D.B. 2002. Nontargeted metabolome analysis by use of Fourier Transform Ion Cyclotron Mass Spectrometry. Omics 6, 217–234. Alberghina, L., Hohmann, S., and Westerhoff, H.V. 2005. Systems biology: Necessary developments and trends. Top. Curr. Genet. 13, 389–402. Alberola, T.M., Garcia-Martinez, J., Antunez, O., Viladevall, L., Barcelo, A., Arino, J., and Perez-Ortin, J.E. 2004. A new set of DNA macrochips for the yeast Saccharomyces cerevisiae: Features and uses. Int. Microbiol. 7, 199–206. Alexandre, H., Ansanay-Galeote, V., Dequin, S., and Blondin, B. 2001. Global gene expression during short-term ethanol stress in Saccharomyces cerevisiae. FEBS Lett. 498, 98–103. Allen, J., Davey, H.M., Broadhurst, D., Heald, J.K., Rowland, J.J., Oliver, S.G., and Kell, D.B. 2003. High-throughput classification of yeast mutants for functional genomics using metabolic footprinting. Nat. Biotechnol. 21, 692–696. Ambrona, J., Vinagre, A., and Ramirez, M. 2005. Rapid asymmetrical evolution of Saccharomyces cerevisiae wine yeasts. Yeast 22, 1299–1306. Amerine, M.A., Berg, H.W., Kunkee, R.E., Ough, C.S., Singleton, V.L., and Webb, A.D. 1980. In ‘‘The Technology of Wine Making’’, 4th Ed., p. 794. AVI Publishing Co, Westport, CO. Ando, A., Tanaka, F., Murata, Y., Takagi, H., and Shima, J. 2005. Identification and classification of genes required for tolerance to high-sucrose stress revealed by genome-wide screening of Saccharomyces cerevisiae. FEMS Yeast Res. 6, 249–267.

110

Linda F. Bisson et al.

Andrews, B.J., Bader, G.D., and Boone, C. 2003. Playing tag with the yeast proteome. Nat. Biotechnol. 21, 1297–1299. Aranda, A. and del Olmo, M. 2004. Exposure of Saccharomyces cerevisiae to acetaldehyde induces sulfur amino acid metabolism and polyamine transporter genes, which depend on Met4p and Haa1p transcription factors, respectively. Appl. Environ. Microbiol. 70, 1913–1922. Ashe, M.P., De Long, S.K., and Sachs, A.B. 2000. Glucose depletion rapidly inhibits translation in yeast. Mol. Biol. Cell 11, 833–848. Backhus, L.E., DeRisi, J., and Bisson, L.F. 2001. Functional genomic analysis of a commercial wine strain of Saccharomyces cerevisiae under differing nitrogen conditions. FEMS Yeast Res. 1, 111–125. Bader, G.D. and Hogue, C.W. 2002. Analyzing yeast protein-protein interaction data obtained from different sources. Nat. Biotechnol. 20, 991–997. Bader, G.D., Heilbut, A., Andrews, B., Tyers, M., Ughes, T., and Boone, C. 2003. Functional genomics and proteomics: Charting a multidimensional map of the yeast cell. Trends Cell Biol. 13, 344–356. Bakalinsky, A.T. and Snow, R. 1990. The chromosomal constitution of wine strains of Saccharomyces cerevisiae. Yeast 6, 367–382. Baker, S.C., Bauer, S.R., Beyer, R.P., Brenton, J.D., Bromley, D., Burrill, J., Causton, H., Conley, M.P., Elespuru, R., Fero, M., Foy, C., Fuscoe, J., et al. 2005. The external RNA controls consortium: A progress report. Nat. Methods 2(10), 731–734. Baldwin, M. 2005. Mass spectrometers for the analysis of biomolecules. Methods Enzymol. 402, 3–48. Baleiras Couto, M.M., Eijsma, B., Hofstra, H., Huis in’t Veld, J.H.J., and vander Vossen, J.M.B.M. 1996. Evaluation of molecular typing techniques to assign genetic diversity among Saccharomyces cerevisiae strains. Appl. Environ. Microbiol. 62, 41–46. Belle, A., Tanay, A., Bitincka, L., Shamir, R., and O’Shea, E.K. 2006. Quantification of protein half-lives in the budding yeast proteome. Proc. Natl. Acad. Sci. USA 103, 13004–13009. Benaroudj, N., Lee, D.H., and Goldberg, A.L. 2001. Trehalose accumulation during cellular stress protects cells and cellular proteins from damage by oxygen radicals. J. Biol. Chem. 276, 24261–24267. Birrell, G.W., Brown, J.A., Wu, H.I., Giaever, G., Chu, A.M., Davis, R.W., and Brown, J.M. 2002. Transcriptional response of Saccharomyces cerevisiae to DNA-damaging agents does not identify genes that protect against these agents. Proc. Natl. Acad. Sci. USA 99, 8778–8783. Bisson, L.F. 1999. Stuck and sluggish fermentations. Am. J. Enol. Vitic. 50, 107–119. Blake, W.J., Kaem, M., Cantor, C.R., and Collins, J.J. 2003. Noise in eukaryotic gene expression. Nature 422, 633–637. Blumberg, L.M. 2003. Comprehensive two-dimensional gas chromatography: Metrics, potentials, limits. J. Chromatogr. A 985, 29–38. Boulton, R.B., Singleton, V.L., Bisson, L.F., and Kunkee, R.E. 1996. In ‘‘Principles and Practices of Winemaking’’, p. 604. Chapman & Hall, New York. Boy-Marcotte, E., Perrot, M., Bussereau, F., Boucherie, H., and Jacquet, M. 1998. Msn2p and Msn4p control a large number of genes induced at the diauxic transition which are repressed by cyclic AMP in Saccharomyces cerevisiae. J. Bacteriol. 180, 1044–1052. Boy-Marcotte, E., Lagniel, G., Perrot, M., Bussereau, F., Boudsocq, A., Jacquet, M., and Labarre, J. 1999. The heat shock response in yeast: Differential regulations and contributions of the Msn2p/Msn4p and Hsf1p regulons. Mol. Microbiol. 33, 274–283. Breci, L., Hattrup, E., Keeler, M., Letarte, J., Johnson, R., and Haynes, P.A. 2005. Comprehensive proteomics in yeast using chromatographic fractionation, gas phase fractionation, protein gel electrophoresis, and isoelectric focusing. Proteomics 5, 2018–2028.

Functional Genomics of Wine Yeast Saccharomyces cerevisiae

111

Brem, R.B., Yvert, G., Clinton, R., and Kruglyak, L. 2002. Genetic dissection of transcriptional regulation in budding yeast. Science 296, 752–755. Brem, R.B., Storey, J.D., Whittle, J., and Kruglyak, L. 2005. Genetic interactions between polymorphisms that affect gene expression in yeast. Nature 436, 701–703. Briones, A.I., Ubeda, J., and Grando, M.S. 1996. Differentiation of Saccharomyces cerevisiae strains isolated from fermenting musts according to karyotype patterns. Int. J. Food Microbiol. 28, 369–377. Brousse, M., Bataille, N., and Boucherie, H. 1985. Identification of glycolytic enzyme polypeptides on the two-dimensional protein map of Saccharomyces cerevisiae and application to the study of some wine yeasts. Appl. Environ. Microbiol. 50, 951–957. Camarasa, C., Grivet, J., and Dequin, S. 2003. Investigation by 13C-NMR and TCA deletion mutant analysis of pathways for succinate formation in S. cerevisiae during anaerobic fermentation. Microbiol. 149, 2669–2678. Cantarelli, C. 1989. Phenolics and yeast: Remarks concerning fermented beverages. In ‘‘Proceedings of the Seventh International Symposium on Yeasts’’ (A. Martini and A. Vaughn-Martini, eds.), pp. S51–S52. John Wiley & Sons, Ltd., New York. Carro, D. and Pina, B. 2001. Genetic analysis of the karyotype instability in natural wine yeast strains. Yeast 18, 1457–1470. Causton, H.C., Ren, B., Koh, S.S., Harbison, C.T., Kanin, E., Jennings, E.G., Lee, T.I., True, H.L., Lander, E.S., and Young, R.A. 2001. Remodeling of yeast genome expression in response to environmental changes. Mol. Biol. Cell 12, 323–337. Cavalieri, D., Townsend, J.P., and Hartl, D.L. 2000. Manifold anomalies in gene expression in a vineyard isolate of Saccharomyces cerevisiae revealed by DNA Microarray analysis. Proc. Natl. Acad. Sci. USA 97, 12369–12374. Cavalieri, D., McGovern, P.E., Hartl, D.L., Moritmer, R., and Polsinelli, M. 2003. Evidence for S. cerevisiae fermentation in ancient wine. J. Mol. Evol. 57, S226–S232. Chalmers, M.J. and Gaskell, S.J. 2000. Advances in mass spectrometry for proteome analysis. Curr. Opin. Biotech. 11, 384–390. Chen, J., Balgley, B.M., DeVoe, D.L., and Lee, C.S. 2003. Capillary isoelectric focusing-based multidimensional concentration/separation platform for proteome analysis. Anal. Chem. 75, 3145–3152. Cliften, P., Sudarsanam, P., Desikan, A., Fulton, L., Fulton, B., Majors, J., Waterston, R., Cohen, B.A., and Johnston, M. 2003. Finding functional features in Saccharomyces genomes by phylogenetic footprinting. Science 301, 71–76. Codon, A.C., Benitez, T., and Korhola, M. 1998. Chromosomal polymorphism and adaptation to specific industrial environments of Saccharomyces strains. Appl. Microbiol. Biotech. 49, 154–163. Cooney, C.D. 2003. ‘‘Proteomic analysis of yeast strain UCD932 during an alcoholic fermentation’’, p. 86. MS Thesis, UC Davis. Cramer, A.C., Vlassides, S., and Block, D.E. 2002. Kinetic model for nitrogen-limited wine fermentations. Biotech. Bioeng. 77, 49–60. Cummings, J. and Fogel, S. 1978. Genetic homology of wine yeasts with Saccharomyces cerevisiae. J. Inst. Brewing 84, 267–270. David, L., Huber, W., Granovskaia, M., Toedling, J., Palm, C.J., Bofkin, L., Jones, T., Davis, R.W., and Steinmetz, L.M. 2006. A high-resolution map of transcription in the yeast genome. Proc. Natl. Acad. Sci. USA 103, 5320–5325. Davidson, J.F. and Schiestl, R.H. 2001. Cytotoxic and genotoxic consequences of heat stress are dependent on the presence of oxygen in Saccharomyces cerevisiae. J. Bacteriol. 183, 4580–4587. De Nobel, H., Lawrie, L., Brul, S., Kils, F., Davis, M., Alloush, H., and Coote, P. 2001. Parallel and comparative analysis of the proteome and transcriptome of sorbic acid-stressed Saccharomyces cerevisiae. Yeast 18, 1413–1428.

112

Linda F. Bisson et al.

DeRisi, J.L., Iyer, V.R., and Brown, P.O. 1997. Exploring the metabolic and genetic control of gene expression on a genomic scale. Science 278, 680–686. Dowhanick, T., Sobczak, J., Russell, I., and Stewart, G. 1990. The rapid identification by protein fingerprinting of yeast and bacterial brewery contaminants. Am. Soc. Brew. Chem. 48, 75–79. Draghici, S., Khatri, P., Eklund, A.C., and Szallasi, Z. 2006. Reliability and reproducibility issues in DNA microarray measurements. Trends Genet. 22, 101–109. Drebot, M.A., Barnes, C.A., Singer, R.A., and Johnston, G. 1990. Genetic assessment of stationary phase for cells of the yeast Saccharomyces cerevisiae. J. Bacteriol. 172, 3584–3589. Dunn, B., Levine, R.P., and Sherlock, G. 2005. Microarray karyotyping of commercial wine yeast strains reveals shared, as well as unique, genomic signatures. BMC Genomics 6(53), 1–21. Dwight, S.S., Harris, M.A., Dolinski, K., Bell, C.A., Binkley, G., Christie, K.R., Fisk, D.G., IsselTarver, L., Schroeder, M., Sherlock, G., Sethuraman, A., Weng, S., et al. 2002. Saccharomyces genome database (SGD) provides secondary gene annotation using the Gene Ontology (GO). Nucleic Acids Res. 30, 69–72. Edman, P. and Begg, G. 1967. A protein sequenator. Eur. J. Biochem. 1, 80–91. Erasmus, D.J., van der Merwe, G.K., and van Vuuren, H.J.J. 2003. Genome-wide expression analyses: Metabolic adaptation of Saccharomyces cerevisiae to high sugar stress. FEMS Yeast Res. 3, 375–399. Fauchon, M., Lagniel, G., Aude, J.-C., Lombardia, L., Soularue, P., Petat, C., Marguerie, G., Sentenac, A., Werner, M., and Labarre, J. 2002. Sulfur sparing in the yeast proteome in response to sulfur demand. Mol. Cell 9, 713–723. Fay, J.C. and Benavides, J.A. 2005a. Evidence for domesticated and wild populations of Saccharomyces cerevisiae. PLoS Genet. 1, e5. Fay, J.C. and Benavides, J.A. 2005b. Hypervariable noncoding sequences in Saccharomyces cerevisiae. Genetics 170, 1575–1587. Fay, J.C., McCullough, H.L., Sniegowski, P.D., and Eisen, M.B. 2004. Population genetic variation in gene expression is associated with phenotypic variation in Saccharomyces cerevisiae. Genome Biol. 5, R26.1–R26.14. Ferea, T.L., Botstein, D., Brown, P.O., and Rosenzweig, R.F. 1999. Systematic changes in gene expression patterns following adaptive evolution in yeast. Proc. Natl. Acad. Sci. USA 96, 9721–9726. Fiehn, O. 2002. Metabolomics: The link between genotypes and phenotypes. Plant Mol. Biol. 48, 155–171. Fiehn, O., Kopka, J., Dormann, P., Altmann, T., Trethewey, R.N., and Willmitzer, L. 2000. Metabolite profiling for plant functional genomics. Nat. Biotechnol. 18, 1157–1161. Fleet, G.H. 1993. The microorganisms of winemaking – Isolation, enumeration and identification. In ‘‘Wine Microbiology and Biotechnology’’ (G. H. Fleet, ed.), pp. 1–26. Harwood Academic Publishers, Camberwell, Victoria, Australia. Fleet, G.H. and Heard, G.M. 1993. Yeasts – Growth during fermentation. In ‘‘Wine Microbiology and Biotechnology’’ (G. H. Fleet, ed.), pp. 27–54. Harwood Academic Publishers, Camberwell, Victoria, Australia. Fo¨rster, J., Famili, I., Fu, P., Palsson, B.., and Nielssen, J. 2003. Genome-scale reconstruction of the Saccharomyces cerevisiae metabolic network. Genome Res. 13, 244–253. Freeman, W.M., Walker, S.J., and Vrana, K.E. 1999. Quantitative RT-PCR: Pitfalls and potential. Biotechniques 26, 112–125. Futcher, B., Latter, G.I., Monardo, P., McLaughlin, C.S., and Garrels, J.I. 1999. A sampling of the yeast proteome. Mol. Cell Biol. 19, 7357–7368. Gallego, F.J., Perez, M.A., Nunez, Y., and Hildago, P. 2005. Comparison of RAPDs, AFLPs and SSR marker for genetic analysis of yeast strains of Saccharomyces cerevisiae. Food Microbiol. 22, 561–568.

Functional Genomics of Wine Yeast Saccharomyces cerevisiae

113

Garrels, J.I. 1983. Quantitative two-dimensional gel electrophoresis of proteins. Meth. Enzymol. 100, 411–423. Gasch, A.P. 2003. The environmental stress response: A common yeast response to diverse environmental stresses. In ‘‘Topics in Current Genetics’’ (S. Hohmann and P. W. H. Mager, eds.), Vol. 1, pp. 11–70. Springer-Verlag, Heidelberg, Germany. Gasch, A.P. and Werner-Washburne, M. 2002. The genomics of yeast responses to environmental stress and starvation. Funct. Integr. Genomics 2, 181–192. Gasch, A.P., Spellman, P.T., Kao, C.M., Carmel-Harel, O., Eisen, M.B., Stortz, G., Botstein, D., and Brown, P.O. 2000. Genomic expression programs in the response of yeast cells to environmental changes. Mol. Biol. Cell 11, 4241–4257. Gavin, A.-C., Bosche, M., Krause, R., Grandi, P., Marzioch, M., Bauer, A., Schultz, J., Rick, J.M., Michon, A.M., and Cruciat, C.M. 2002. Functional organization of the yeast proteome by systematic analysis of protein complexes. Nature 425, 141–147. Gehrke, C., Wall, L., Absheer, J., Kaiser, F., and Zumwalt, R. 1987. Sample preparation for chromatography of amino acids. In ‘‘Amino Acid Analysis by Gas Chromatography’’ (R. Zumwalt, K. Kuo, and C. Gehrke, eds.), Vol. 1, pp. 1–21. CRC Press, Inc., Florida. Gelperin, D.M., White, M.A., Wilkinson, M.L., Kon, L.A., Kung, Y., Wise, K.J., LopezHoyo, N., Jiang, L., Piccirillo, S., Yu, H., Gerstein, M., Dumont, M.E., et al. 2005. Biochemical and genetic analysis of the yeast proteome with a movable ORF collection. Genes Dev. 19, 2816–2826. Ghaemmaghami, S., Huh, W.K., Bower, K., Howson, R.W., Belle, A., Dephoure, N., O’Shea, E.K., and Weissman, J.S. 2003. Global analysis of protein expression in yeast. Nature 425, 737–741. Giaever, G., Chu, A.M., Ni, L., Connelly, C., Riles, L., Veronneau, S., Dow, S., LucauDanila, A., Anderson, K., Andre, B., Arkin, A.P., Astromoff, A., et al. 2002. Functional profiling of the Saccharomyces cerevisiae genome. Nature 418, 387–391. Glassbrook, N., Beecher, C., and Ryals, J. 2000. Metabolic profiling on the right path. Nat. Biotechnol. 18, 1142–1143. Goffeau, A., Barrell, B.G., Bussey, H., Davis, R.W., Dujon, B., Feldmann, H., Galibert, F., Hoheisel, J.D., Jacq, C., Johnston, M., Louis, E.J., Mewes, H.W., et al. 1996. Life with 6000. Genes Sci. 274, 546–567. Goffeau, A., Park, J., Paulsen, I.T., Jonniaux, J.-L., Dinh, T., Mordant, P., and Saier, M.H., Jr. 1997. Multidrug-resistant transport proteins in yeast: Complete inventory and phylogenetic characterization of yeast open reading frames within the major facilitator superfamily. Yeast 13, 43–54. Gollub, J., Ball, C.A., Binkley, G., Demeter, J., Finkelstein, D.B., Hebert, J.M., HernandezBoussard, T., Jin, H., Kaloper, M., Matese, J.C., Schroeder, M., Brown, P.O., et al. 2003. The Stanford microarray database: Data access and quality assessment tools. Nucleic Acids Res. 31, 94–96. Granot, D. and Snyder, M. 1993. Carbon source induces growth of stationary phase yeast cells, independent of carbon source metabolism. Yeast 9, 465–479. Graves, P.R. and Haystead, T.A.J. 2002. Molecular biologist’s guide to proteomics. Microbiol. Mol. Biol. Rev. 66, 39–63. Gray, J.V., Petsco, G.A., Johnston, G.C., Ringe, D., Singer, R.A., and Werner-Washburne, M. 2004. ‘‘Sleeping beauty’’: Quiescence in Saccharomyces cerevisiae. Microbiol. Mol. Biol. Rev. 68, 187–206. Gresham, D., Ruderfer, D.M., Pratt, S.C., Schacherer, J., Dunham, M.J., Botstein, D., and Kruglyak, L. 2006. Genome-wide detection of polymorphisms at nucleotide resolution with a single DNA microarray. Science 311, 1932–1936. Griffin, T.J., Gygi, S.P., Ideker, T., Rist, B., Eng, J., Hood, L., and Aebersold, R. 2002. Complementary profiling of gene expression at the transcriptome and proteome levels in Saccharomyces cerevisiae. Mol. Cell Proteomics 1, 323–333.

114

Linda F. Bisson et al.

Grigoriev, A. 2003. On the number of protein-protein interactions in the yeast proteome. Nucleic Acid Res. 31, 4157–4161. Grivet, J., Delort, A., and Portais, J. 2003. NMR and Microbiology: From Physiology to Metabolomics. Biochimie 85, 823–840. Grunenfelder, B. and Winzeler, E.A. 2002. Treasures and traps in genome-wide data sets: Case Examples from yeast. Nat. Rev. 3, 653–661. Guijo, S., Mauricio, C., Salmon, J.M., and Ortega, J.M. 1997. Determination of the relative ploidy in different Saccharomyces cerevisiae strains used for fermentation and ‘flor’ film ageing of dry sherry-type wines. Yeast 13, 101–117. Gygi, S.P., Rist, B., Gerber, S.A., Turecek, F., Gelb, M.H., and Aebersold, R. 1999a. Quantitative analysis of complex protein mixtures using isotope-coded affinity tags. Nat. Biotechnol. 17, 994–999. Gygi, S.P., Rochon, Y., Franza, B.R., and Aebersold, R. 1999b. Correlation between protein and mRNA abundance in yeast. Mol. Cell Biol. 19, 1720–1730. Gygi, S.P., Rist, B., and Aebersold, R. 2000. Measuring gene expression by quantitative proteome analysis. Curr. Opin. Biotech. 11, 396–401. Haab, B.B., Dunham, M.J., and Brown, P.O. 2001. Protein microarrays for highly parallel detection and quantitation of specific proteins and antibodies in complex solutions. Genome Biol. 2, RESEARCH0004. Harrigan, G. 2002. Metabolic profiling: Pathways in drug discovery. Drug Discov. Today 7, 351–352. Hauser, N.C., Fellenberg, K., Gil, R., Bastuck, S., Hoheisel, J.D., and Perez-Ortin, J.E. 2001. Whole genomes of a wine yeast strain. Comp. Funct. Genomics 2, 69–79. Haynes, P.A., Gygi, S.P., Figeys, D., and Aebersold, R. 1998. Proteome analysis: Biological assay or data archive? Electrophoresis 19, 1862–1871. Hazburn, T.R. and Fields, S. 2001. Networking proteins in yeast. Proc. Natl. Acad. Sci. USA 98, 4277–4278. Heyer, L.J., Kruglyak, S., and Yooseph, S. 1999. Exploring expression data: Identification and analysis of co-expressed genes. Genome Res. 9, 1106–1115. Hirasawa, T., Nakakura, Y., Yoshikawa, K., Ashitani, K., Nagahisa, K., Furusawa, C., Katakura, Y., Shimizu, H., and Shioya, S. 2006. Comparative analysis of transcriptional responses to saline stress in the laboratory and brewing strains of Saccharomyces cerevisiae with DNA microarray. Appl. Microbiol. Biotechnol. 703, 346–357. Ho, Y., Gruhler, A., Heilbut, A., Bader, G.D., Moore, L., Adams, S.L., Millar, A., Taylor, P., Bennett, K., Boutilier, K., Yang, L., Wolting, C., et al. 2002. Systematic identification of protein complexes in Saccharomyces cerevisiae by mass spectrometry. Nature 415, 180–183. Holland, M.J. 2002. Transcript abundance in yeast varies over six orders of magnitude. J. Biol. Chem. 277, 14363–14366. Howell , K.S., Cozzolino, D., Bartowsky, E.J., Fleet, G.H., and Henschke, P.A. 2006. Metabolic profiling as a tool for revealing Saccharomyces interactions during wine fermentation. FEMS Yeast Res. 6(1), 91–101. Hughes, T.R., Roberts, C.J., Dai, H., Jones, A.R., Meer, M.R., Slade, D., Burchard, J., Dow, S., Ward, T.R., Kidd, M.J., Friend, S.H., and Marton, M.J. 2000. Widespread aneuploidy revealed by DNA microarray expression profiling. Nat. Genet. 25, 333–337. Huh, W.-K., Falvo, J.V., Gerke, L.C., Carroll, A.S., Howson, R.W., Weissman, J.S., and O’Shea, E.K. 2003. Global analysis of protein localization in budding yeast. Nature 425, 686–691. Infante, J.J., Dombek, K.M., Rebordinos, L., Cantoral, J.M., and Young, E.T. 2003. Genomewide amplifications caused by chromosomal rearrangements play a major role in the adaptive evolution of natural yeast. Genetics 165, 1745–1759. Ihmels, J., Levy, R., and Barkai, N. 2004. Principles of transcriptional control in the metabolic network of Saccharomyces cerevisiae. Nat. Biotechnol. 22, 86–92.

Functional Genomics of Wine Yeast Saccharomyces cerevisiae

115

Insa, G., Sablayrolles, J.-M., and Douzal, V. 1995. Alcoholic fermentation under oenological conditions. Bioproc. Engr. 13, 171–176. Ito, T., Chiba, T., Ozawa, R., Yoshida, M., Hattori, M., and Sakaki, Y. 2001. A comprehensive two-hybrid analysis to explore the yeast protein interactome. Proc. Natl. Acad. Sci. USA 98, 4569–4574. Ivorra, C., Perez-Ortin, J.E., and del Olmo, M. 1999. An inverse correlation between stress resistance and stuck fermentations in wine yeasts. A molecular study. Biotech. Bioeng. 64, 698–708. Izquierdo Canas, P.M., Ubeda Iranzo, J.F., and Briones Perez, A.I. 1997. Study of the karyotype of wine yeasts isolated in the region of Valdepenas in two consecutive vintages. Food Microbiol. 14, 221–225. James, T.C., Campbell, S., Donnelly, D., and Bond, U. 2003. Transcription profile of brewery yeast under fermentation conditions. J. Appl. Microbiol. 94, 432–448. Jiang, H. and English, A.M. 2002. Quantitative analysis of the yeast proteome by incorporation of isotopically labeled leucine. J. Proteome Res. 1, 345–350. Johnston, J.R., Baccari, C., and Mortimer, R.K. 2000. Genotypic characterization of strains of commercial wine yeasts by tetrad analysis. Res. Microbiol. 151, 583–590. Joubert, R., Strub, J.-M., Zugmeyer, S., Kobi, D., Carte, N., Van Dorsselaer, A., Boucherie, H., and Jaquet-Gutfreund, L. 2001. Identification by mass spectrometry of two-dimensional gel electrophoresis-separated proteins extracted from lager brewing yeasts. Electrophoresis 22, 2969–2982. Kal, A.J., van Zonneveld, A.J., Benes, V., van den Berg, M., Koerkamp, M.G., Albermann, K., Strack, N., Ruijter, I.M., Richter, A., Dujon, B., Ansorge, W., and Tabak, H.F. 1999. Dynamics of gene expression revealed by comparison of Serial analysis of gene expression transcript profiles from yeast grown on two different carbon sources. Mol. Biol. Cell 10, 1859–1872. Kaukovirta-Norja, A., Wilhelmson, A., and Poutanen, K. 2004. Germination: A means to improve the functionality of oat. Agric. Food Sci. 13, 100–112. Kellis, M., Patterson, N., Endrizzi, M., Birren, B., and Lander, E.S. 2003. Sequencing and comparison of yeast species to identify genes and regulatory elements. Nature 423, 241–254. Kobayashi, N. and McEntee, K. 1990. Evidence for a heat shock transcription factor– independent mechanism for heat shock induction of transcription in Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. USA 87, 6550–6554. Kobayashi, N. and McEntee, K. 1993. Identification of cis and trans components of a novel heat shock stress regulatory pathway in Saccharomyces cerevisiae. Mol. Cell Biol. 13, 248–256. Kobi, D., Zugmeyer, S., Potier, S., and Jaquet-Gutfreund, L. 2004. Two-dimensional protein map of an ‘‘ale’’-brewing yeast strain: Proteome dynamics during fermentation. FEMS Yeast Res. 5, 213–230. Kuhn, K.M., DeRisi, J.L., Brown, P.O., and Sarnow, P. 2001. Global and specific translational regulation in the genomic response of Saccharomyces cerevisiae to a rapid transfer from a fermentable to a nonfermentable carbon source. Mol. Cell Biol. 21, 916–927. Kumar, A., Harrison, P.M., Cheung, K.-H., Lan, N., Echols, N., Bertone, P., Miller, P., Gerstein, M.B., and Snyder, M. 2002. An integrated approach for finding overlooked genes in yeast. Nat. Biotechnol. 20, 58–63. Kumar, A., Seringhaus, M., Biery, M.C., Sarnovsky, R.J., Umansky, L., Piccirillo, S., Heidtman, M., Cheung, K.-H., Dobry, C.J., Gerstein, M.B., Craig, N.L., and Snyder, M. 2004. Large-scale mutagenesis of the yeast genome using a Tn7-derived multipurpose transposon. Genome Res. 14, 1975–1986. Kuo, W.P., Liu, F., Trimarchi, J., Punzo, C., Lombardi, M., Sarang, J., Whipple, M.E., Maysuria, M., Serikawa, K., Lee, S.Y., McCrann, D., Kang, J., et al. 2006. A sequence-oriented

116

Linda F. Bisson et al.

comparison of gene expression measurements across different hybridization-based technologies. Nat. Biotechnol. 24, 832–840. Kuthan, M., Devaux, F., Janderova, B., Slaninova, I., Jacq, C., and Palkova, Z. 2003. Domestication of wild Saccharomyces cerevisiae is accompanied by changes in gene expression and colony morphology. Mol. Microbiol. 47, 745–754. Lafaye, A., Junot, C., Pereira, Y., Lagniel, G., Tabet, J.-C., Ezan, E., and Labarre, J. 2005. Combined proteome and metabolite-profiling analyses reveal surprising insights into yeast sulfur metabolism. J. Biol. Chem. 280, 24723–24730. Landry, C.R., Oh, J., Hart, D.L., and Cavalieri, D. 2006a. Genome-wide scan reveals that genetic variation for transcriptional plasticity in yeast is biased towards multi-copy and dispensable genes. Gene 366, 343–351. Landry, C.R., Townsend, J.P., Hartl, D.L., and Cavalieri, D. 2006b. Ecological and evolutionary genomics of Saccharomyces cerevisiae. Mol. Ecol. 15, 575–591. Larsen, M.R., Larsen, P.M., Fay, S.J., and Roepstorff, P. 2001. Characterization of differently processed forms of enolase 2 from Saccharomyces cerevisiae by two-dimensional gel electrophoresis and mass spectrometry. Electrophoresis 22, 566–575. Larsson, T., Norbeck, J., Karlsson, H., Karlsson, K.-A., and Blomberg, A. 1997. Identification of two-dimensional gel electrophoresis resolved yeast proteins by matrix-assisted laser desorption ionization mass spectrometry. Electrophoresis 18, 418–423. Lashkari, D.A., DeRisi, J.L., McCusker, J.H., Namath, A.F., Gentile, C., Hwang, S.Y., Brown, P.O., and Davis, R.W. 1997. Yeast microarrays for genome wide parallel genetic and gene expression analysis. Proc. Natl. Acad. Sci. USA 94, 13057–13062. Lee, M.-L.T., Kuo, F.C., Whitmore, G.A., and Sklar, J. 2000. Importance of replication and microarray gene expression studies: Statistical methods and evidence from repetitive cDNA hybridizations. Proc. Natl. Acad. Sci. USA 97, 9834–9839. Lenz, E.M., Bright, J., Knight, R., Wilson, I.D., and Major, H. 2004. A metabonomic investigation of the biochemical effects of mercuric chloride in the rat using 1H NMR and HPLCTOF/MS: Time dependent changes in the urinary profile of endogenous metabolites as a result of nephrotoxicity. Analyst 129, 535–541. Leusch, H.G., Hoffman, M., and Emeis, C.C. 1985. Fluorometric determination of the total DNA content of different yeast species using 40 6-diamidino-2-phenylindole dihydrochloride. Can. J. Microbiol. 31, 1164–1166. Lian, Y. and Kelemen, A. 2006. Associating phenotypes with molecular events: Recent statistical advances and challenges underpinning microarray experiments. Funct. Integr. Genomics 6, 1–13. Link, A.J., Eng, J., Schieltz, D.M., Carmack, E., Mize, G.J., Morris, D.R., Garvik, B.M., and Yates, J.R., III 1999. Direct analysis of protein complexes using mass spectrometry. Nat. Biotechnol. 17, 676–682. Littlewood, A.B. 1970. ‘‘Gas Chromatography: Principles, Techniques, and Applications’’, 2nd Ed. Academic Press, New York. Lockhart, D.J. and Winzeler, E.A. 2000. Genomics, gene expression and DNA arrays. Nature 405, 827–836. Lockhart, D.J., Dong, H., Byrne, M.C., Follettie, M.T., Gallow, M.V., Chee, M.S., Mittmann, M., Wang, C., Kobayashi, M., Horton, H., and Brown, E.L. 1996. Expression monitoring by hybridization to high-density oligonucleotide arrays. Nat. Biotechnol. 14, 1675–1680. Longo, E. and Vezinhet, F. 1993. Chromosomal rearrangements during vegetative growth of a wild strain of Saccharomyces cerevisiae. Appl. Environ. Microbiol. 59, 322–326. Mangahas, J.L. 2003. ‘‘An analysis of the adaptive response in fermenting wine yeast using genechips’’, p. 87. MS Thesis, UC Davis.

Functional Genomics of Wine Yeast Saccharomyces cerevisiae

117

Manginot, C., Roustan, J.L., and Sablayrolles, J.-M. 1998. Nitrogen demand of different yeast strains during alcoholic fermentation. Importance of the stationary phase. Enzyme Microb. Technol. 23, 511–517. Marchler, G., Schuller, C., Adam, G., and Ruis, H. 1993. A Saccharomyces cerevisiae UAS element controlled by protein kinase A activates transcription in response to a variety of stress conditions. EMBO J. 12, 1997–2003. Marks, V.D., van der Merwe, G.K., and van Vuuren, H.J.J. 2003. Transcriptional profiling of wine yeast in fermenting grape juice: Regulatory effect of diammonium phosphate. FEMS Yeast Res. 3, 269–287. Martzen, M.R., McCraith, S.M., Spinelli, S.L., Torres, F.M., Fields, S., Grayhack, E.J., and Phizicky, E.M. 1999. A biochemical genomics approach for identifying genes by the activity of their products. Science 286, 1153–1155. Mashego, M., Van Gulik, W., Vinke, J., and Heijnen, J. 2003. Critical evaluation of sampling techniques for residual glucose determination in carbon-limited culture of Saccharomyces cerevisiae. Biotech. Bioeng. 83, 395–399. Mashego, M., Van Gulik, W., and Heijnen, J. 2006. Metabolome dynamic responses of Saccharomyces cerevisiae to simultaneous rapid perturbations in external electron acceptor and electron donor. FEMS Yeast Res. 7(1), 48–66. Monteiro, F.F. and Bisson, L.F. 1991a. Biological assay of nitrogen content of grape juice and prediction of sluggish fermentations. Am. J. Enol. Vitic. 42, 47–57. Monteiro, F.F. and Bisson, L.F. 1991b. Amino acid utilization and urea formation during vinification fermentations. Am. J. Enol. Vitic. 42, 199–208. Monteiro, F.F. and Bisson, L.F. 1992a. Nitrogen supplementation of grape juice. I. Effect on amino acid utilization during fermentation. Am. J. Enol. Vitic. 43, 1–10. Monteiro, F.F. and Bisson, L.F. 1992b. Nitrogen supplementation of grape juice II. Effect on amino acid and urea release following fermentation. Am. J. Enol. Vitic. 43, 11–17. Mortimer, R.K. 2000. Evolution and variation of the yeast Saccharomyces genome. Genome Res. 10, 403–409. Mortimer, R.K., Romano, P., Suzzi, G., and Polsinelli, M. 1994. Genome renewal: A new phenomenon revealed from a genetic study of 43 strains of Saccharomyces cerevisiae derived from natural fermentation of grape musts. Yeast 10, 1543–1552. Moseley, M.A. 2001. Current trends in differential expression proteomics: Isotopically coded tags. Trends Biotech. 19, S10–S16. Muddiman, D.C., Bakhtiar, R., Hofstadler, S.A., and Smith, R.D. 1997. Matrix-assisted laser desorption/ionization mass spectrometry: Instrumentation and applications. J. Chem. Edu. 74, 1288–1292. Myers, C.L., Dunham, M.J., Kung, S.Y., and Troyanskaya, O.G. 2004. Accurate detection of aneuploidies in array CGH and gene expression microarray data. Bioinformatics 20, 3533–3543. Narayanaswamy, R., Niu, W., Scuras, A.D., Hart, G.T., Davies, J., Ellingotn, A.D., Iyer, V.R., and Marcotte, E.M. 2006. Systematic profiling of cellular phenotypes with spotted cell microarrays reveals mating-pheromone response genes. Genome Biol. 7, R6.1–R6.9. Nobeli, I., Ponstingl, H., Krissinel, E., and Thornton, J. 2003. A structure-based anatomy of the E. coli. metabolome. J. Mol. Biol. 334, 697–719. Norbeck, J. and Blomberg, A. 1997. Two-dimensional electrophoretic separation of yeast proteins using a non-linear wine range pH 3–10 immobilized pH gradient in the first dimension; reproducibility and evidence for isoelectric focusing of alkaline pI > 7 proteins. Yeast 13, 1519–1534. Nugent, S.L. 2004. ‘‘The effect of heat stress at 8 and 11% ethanol on genomic gene expression in three wine yeast strains’’, p. 79. MS Thesis, UC Davis.

118

Linda F. Bisson et al.

Ockert, P.H.A. and Kock, J.L.F. 1989. Differentiation of yeast species, and strains within a species, by cellular fatty acid analysis. 1. Application of an adapted technique to differentiate between strains of Saccharomyces cerevisiae. J. Microbiol. Methods 10, 9–23. Olineka, T.L., Spiropoulos, A., Mara, P.A., and Bisson, L.F. 2005. Optimization of proteome analysis for wine yeast strains. In ‘‘Microbial Products and Biotransformations’’ ( J.-L. Barredo- Fuente, ed.), pp. 345–368. Humana Press, Totowas, NJ. Oliver, S.G., Winson, M.K., Kell, D.B., and Baganz, F. 1998. Systematic functional analysis of the yeast genome. Trends Biotechnol. 16, 373–378. Oshiro, G. and Winzeler, E.A. 2000. Aneuploidy- it’s more common than you think. Nat. Biotechnol. 18, 715–716. Palkova, Z. 2004. Multicellular microorganisms: Laboratory versus nature. EMBO Rep. 5, 470–476. Pardo, M., Monteoliva, L., Pla, J., Sanchez, M., Gil, C., and Nombela, C. 1999. Two-dimensional analysis of proteins secreted by Saccharomyces cerevisiae regenerating protoplasts: A novel approach to study the cell wall. Yeast 15, 459–472. Patterson, S.D. and Aebersold, R.H. 2003. Proteomics: The first decade and beyond. Nat. Genet. 33, 311–323. Paulsson, J. 2004. Summing up the noise in gene networks. Nature 427, 415–418. Peltrouche-Llacsahuanga, H., Schmidt, S., Lutticken, R., and Haase, G. 2000. Discriminative power of fatty acid methyl ester FAME analysis using the microbial identification system MIS for Candida, Torulopsis glabrata and Saccharomyces cerevisiae. Diagn. Microbiol. Infect. Dis. 38, 213–221. Peng, J., Elias, J.E., Thoreen, C.C., Licklider, L.J., and Gygi, S.P. 2003. Evaluation of multidimensional chromatography coupled with tandem mass spectrometry (LC/LC-MS/MS) for large-scale protein analysis: The yeast proteome. J. Proteome Res. 2, 43–50. Peric-Concha, N. and Long, P. 2003. Mining the microbial metabolome: A new frontier for natural product lead discovery. Drug Discov. Today 8, 1078–1084. Puig, S. and Perez-Ortin, J.E. 2000. Stress response and expression patterns in wine fermentations of yeast genes induced at the diauxic shift. Yeast 16, 139–148. Quackenbush, J. 2004. Data standards for ‘omic’ science. Nat. Biotechnol. 22, 613–614. Quackenbush, J. 2005. Using DNA microarrays to assay gene expression. In ‘‘Bioinformatics: A Practical Guide to the Analysis of Genes and Proteins’’ (A. D. Baxevanis and B. F. F. Ouellette, eds.), 3rd Ed., pp. 410–444. John Wiley & Sons, Ltd., New York. Raamsdonk, L.M., Teusink, B., Broadhurst, D., Zhang, N., Hayes, A., Walsh, M., Berden, J.A., Brindle, K.M., Kell, D.B., Rowland, J.J., Westerhoff, H.V., Van Dam, K., et al. 2001. A functional genomics strategy that uses metabolome data to reveal the phenotype of silent mutations. Nat. Biotechnol. 19, 45–50. Rachidi, N., Barre, P., and Blondin, B. 2000. Examination of the transcriptional specificity of an enological yeast. A pilot experiment on the chromosome-III right arm. Curr. Genetics 37, 1–11. Raghuraman, M.K., Winzeler, E.A., Collingwood, D., Hunt, S., Wodicka, L., Conway, A., Lockhart, D.J., Davis, R.W., Brewer, B.J., and Fangman, W.L. 2001. Replication dynamics of the yeast genome. Science 294, 115–121. Ravichandran, V. and Sriram, R.D. 2005. Toward data standards for proteomics. Nat. Biotechnol. 23, 373–376. Renouf, V., Perello, M.C., Strehaiano, P., and Lonvaud-Funel, A. 2006. Global survey of the microbial ecosystem during alcoholic fermentation in winemaking. J. Int. Sci. Vigne. Vin. 40, 101–116. Rep, M., Krantz, M., Thevelein, J.M., and Hohmann, S. 2000. The transcriptional response of Saccharomyces cerevisiae to osmotic shock: Hot1p and Msn2p/Msn4p are required for the

Functional Genomics of Wine Yeast Saccharomyces cerevisiae

119

induction of subsets of high osmolarity glycerol pathway-dependent genes. J. Biol. Chem. 275, 8290–8300. Riou, C., Nicaud, J.-M., Barre, P., and Gaillardin, C. 1997. Stationary-phase gene expression in Saccharomyces cerevisiae during wine fermentation. Yeast 13, 903–915. Roberts, G.C. and Hudson, A.P. 2006. Transcriptome profiling of Saccharomyces cerevisiae during a transition from fermentative to glycerol-based respiratory growth reveals extensive metabolic and structural remodeling. Mol. Gen. Genomics 276, 170–186. Rodriguez-Pena, J.M., Perez-Diaz, R.M., Alvarez, S., Bermejo, C., Garcia, R., Santiago, C., Nombela, C., and Arroyo, J. 2005. The ‘yeast cell wall chip’ – a tool to analyse the regulation of cell wall biogenesis in Saccharomyces cerevisiae. Microbiology 151, 2241–2249. Ronald, J., Akey, J.M., Whittle, J., Smith, E.N., Yvert, G., and Kruglyak, L. 2005. Simultaneous genotyping, gene expression measurement, and detection of allele-specific expression with oligonucleotide arrays. Gen. Res. 15, 284–291. Rossignol, T., Dulau, L., Julien, A., and Blondin, B. 2003. Genome-wide monitoring of wine yeast gene expression during alcoholic fermentation. Yeast 20, 1369–1385. Rossignol, T., Postaire, O., Storai, J., and Blondin, B. 2006. Analysis of the genomic response of a wine yeast to rehydration and inoculation. Appl. Microbiol. Biotechnol. 71, 699–712. Sagliocco, F., Guillemot, J.-C., Monribot, C., Capdevielle, J., Perrot, M., Ferran, E., Ferrara, P., and Boucherie, H. 1996. Identification of proteins of the yeast protein map using genetically manipulated strains and peptide-mass fingerprinting. Yeast 12, 1519–1533. Sahara, T., Goda, T., and Ohgiya, S. 2002. Comprehensive expression analysis of timedependent responses of yeast cells to low temperature. J. Biol. Chem. 51, 50015–50021. Schadt, E.E., Li, C., Su, C., and Wong, W.H. 2000. Analyzing high-density oligonucleotide gene expression array data. J. Cell. Biochem. 80, 192–202. Schena, M., Shalon, D., Davis, R.W., and Brown, P.O. 1995. Quantitative monitoring of gene expression patterns with a complementary DNA microarray. Science 270, 467–470. Schuller, D., Valero, E., Dequin, S., and Casal, M. 2004. Survey of molecular methods for the typing of wine yeast strains. FEMS Microbiol. Lett. 231, 19–26. Schuller, D., Alves, H., Dequin, S., and Casal, M. 2005. Ecological survey of Saccharomyces cerevisiae strains from vineyards in the Vinho Verde region of Portugal. FEMS Microbiol. Ecol. 51, 167–177. Serber, Z., Corsini, L., Durst, F., and Dotsch, V. 2005. In-Cell NMR Spectroscopy. Methods Enzymol. 394, 17–41. Shevchenko, A., Jensen, O.N., Podtelejnikov, A.V., Sagliocco, F., Wilm, M., Vorm, O., Mortensen, P., Shevchenko, A., Boucherie, H., and Mann, M. 1996a. Linking genome and proteome by mass spectrometry: Large-scale identification of yeast proteins from two dimensional gels. Proc. Natl. Acad. Sci. USA 93, 14440–14445. Shevchenko, A., Wilm, M., Vorm, O., and Mann, M. 1996b. Mass spectromic sequencing of proteins from silver-stained polyacrylamide gels. Anal. Chem. 68, 850–858. Shields, R. 2006. MIAME, we have a problem. Trends Genet. 22, 65–66. Siderius, M. and Mager, W.H. 2003. Conditional response to stress in yeast. Monatsheft fur Chemie 134, 1433–1444. Smith, V., Botstein, D., and Brown, P.O. 1995. Genetic footprinting: A genomic strategy for determining a gene’s function given its sequence. Proc. Natl. Acad. Sci. USA 92, 6479–6483. Smith, R.D., Anderson, G.A., Lipton, M.S., Pasa-Tolic, L., Shen, Y., Conrads, T.P., Veenstra, T.D., and Udseth, H.R. 2002. An accurate mass tag strategy for quantitative and high-throughput measurements. Proteomics 2, 513–523. Soga, T., Ohashi, Y., Ueno, Y., Naraoka, H., Tomita, M., and Nishioka, T. 2003. Quantitative metabolome analysis using capillary electrophoresis mass spectrometry. J. Proteome Res. 2, 488–494.

120

Linda F. Bisson et al.

Spencer, J.F.T., Laud, P., and Spencer, D.M. 1980. The use of mitochondrial mutants in the isolation of hybrids involving industrial yeast strains. Mol. Gen. Genet. 178, 651–654. Spiropoulos, A., Tanaka, J., Flerianos, I., and Bisson, L.F. 2000. Characterization of hydrogen sulfide formation in commercial and natural wine isolates of Saccharomyces. Am. J. Enol. Vitic. 51, 233–248. Stahl, G., Ben Salem, S.N., Chen, L., Zhao, B., and Farabaugh, P.J. 2004. Translational accuracy during exponential, postdiauxic and stationary growth phases in Saccharomyces cerevisiae. Euk. Cell 3, 331–338. Stephanopoulos, G., Alper, H., and Moxley, J. 2004. Exploiting biological complexity for strain improvement through systems biology. Nat. Biotechnol. 22, 1261–1267. Storey, J.D. and Tibshirani, R. 2003. Statistical significance for genomewide studies. Proc. Natl. Acad. Sci. USA 100, 9440–9445. Storey, J.D., Akey, J.M., and Kruglyak, L. 2005. Multiple locus linkage analysis of genomewide expression in yeast. PLoS Biol. 3, e267. Takahashi, T. 1978. Genetic analysis of a German wine yeast. Bull. Brew. Sci. 24, 39–47. Thomsson, E., Gustafsson, L., and Larsson, C. 2005. Starvation response of Saccharomyces cerevisiae grown in anaerobic nitrogen- or carbon-limited chemostat cultures. Appl. Environ. Microbiol. 71, 3007–3013. Thornton, R.J. 1986. Genetic characterization of New Zealand and Australian wine yeasts, occurrence of killer systems and homothallism. J. Microbiol. 52, 97–104. Thornton, R.J. and Eschenbruch, R. 1976. Homothallism in wine yeast. Antonie van Leeuwenhoek 42, 503–509. Tong, A.H.Y., Evangelista, M., Parsons, A.B., Xu, H., Bader, G.D., Page, N., Robinson, M., Raghibilizadeh, S., Hogue, C.W.V., Bussey, H., Andrews, B., Tyers, M., et al. 2001. Systematic genetic analysis with ordered arrays of yeast deletion mutants. Science 294, 2364–2367. Tong, A.H.Y., Lesage, G., Bader, G.D., Ding, H., Xu, H., Xin, X., Young, J., Berriz, G.F., Brost, R.L., Chang, M., Chen, Y.Q., Cheng, X., et al. 2004. Global mapping of the yeast genetic interaction network. Science 303, 808–813. Townsend, J.P., Cavalieri, D., and Hartl, D.L. 2003. Population genetic variation in genomewide gene expression. Mol. Biol. Evol. 20, 955–963. Trabalzini, L., Paffetti, A., Scalioni, A., Talamo, F., Ferro, E., Coratza, G., Bovalini, L., Lusini, P., Martelli, P., and Santucci, A. 2003. Proteomic response to physiological fermentation stresses in a wild type wine strain of Saccharomyces cerevisiae. Biochem. J. 370, 35–46. Tredoux, H.G., Kock, J.L.F., Lategan, P.M., and Muller, H.B. 1987. A rapid identification technique to differentiate between Saccharomyces cerevisiae strains and other yeast species in the wine industry. Am. J. Enol. Vitic. 38, 161–164. Uetz, P. and Finley, Jr., R. L. 2005. From protein networks to biological systems. FEBS Lett. 579, 1821–1827. Uetz, P., Giot, L., Cagney, G., Mansfield, T.A., Judson, R.S., Knight, J.R., Lockshon, D., Narayan, V., Srinivasan, M., Pochart, P., Qureshi-Emili, A., Li, Y., et al. 2000. A comprehensive analysis of protein-protein interactions in Saccharomyces cerevisiae. Nature 403, 623–627. Unger, K. 2002. Silica as a Support in HPLC of Biomacromolecules. In ‘‘HPLC of Biological Macromolecules’’ (K. Gooding and F. Rognier, eds.), 2nd Ed., pp. 1–16. Marcel Dekker, Inc., New York. Van Speybroeck, L., De Backer, P., VanPucke, J., and De Waele, D. 2005. The conceptual challenge of systems biology. Bioessays 27, 1305–1307. Varela, C., Cardenas, J., Melo, F., and Agosin, E. 2005. Quantitative analysis of wine yeast gene expression profiles under winemaking conditions. Yeast 22, 369–383. Velculescu, V.E., Zhang, L., Vogelstein, B., and Kinzler, K.W. 1995. Serial analysis of gene expression. Science 270, 484–487. Verstrepen, K.J., Iserentant, D., Malcorps, P., Derdelinckx, G., Van Dijck, P., Winderickx, J., Pretorius, I.S., Thevelein, J.M., and Delvaux, F.R. 2004. Glucose and sucrose: Hazardous fast-food for industrial yeast. Trends Biotechnol. 22, 531–537.

Functional Genomics of Wine Yeast Saccharomyces cerevisiae

121

Vidan, S. and Snyder, M. 2001. Large-scale mutagenesis: Yeast genetics in the genome era. Curr. Opin. Biotech. 12, 28–34. Von Mering, C., Krause, R., Snel, B., Cornell, M., Oliver, S.G., Fields, S., and Bork, P. 2002. Comparative assessment of large-scale data sets of protein-protein interactions. Nature 417, 399–403. Wang, Y., Bollard, M., Keun, H., Antti, H., Beckonert, O., Ebbels, T., Lindon, J., Holmes, E., Tang, H., and Nicholson, J. 2003. Spectral editing and pattern recognition methods applied to high-resolution magic-angle spinning 1H NMR spectroscopy of liver tissues. Anal. Biochem. 323, 26–32. Wang, Y., Barbacioru, C., Hyland, F., Xiao, W., Hunkapiller, K.L., Blake, J., Chan, F., Gonzalex, C., Zhang, L., and Samaha, R.R. 2006. Large scale real-time PCR validation on gene expression measurements from two commercial long-oligonucleotide microarrays. BMC Genomics 759, 1–16. Warringer, J. and Blomberg, A. 2003. Automated screening in environmental arrays allows analysis of quantitative phenotype profiles in Saccharomyces cerevisiae. Yeast 20, 53–67. Warringer, J., Kericson, E., Fernandez, L., Nerman, O., and Blomberg, A. 2003. High-resolution yeast phenomics resolves different physiological features in the saline response. Proc. Natl. Acad. Sci. USA 100, 15724–15729. Washburn, M.P., Wolters, D., and Yates, J.R., Jr. 2001. Large-scale analysis of the yeast proteome by multidimensional protein identification technology. Nat. Biotechnol. 19, 242–247. Winzeler, E.A., Shoemaker, D.D., Astromoff, A., Liang, H., Anderson, K., Andre, B., Bangham, R., Benito, R., Boeke, J.D., Bussey, H., Chu, A.M., Connelly, C., et al. 1999. Functional characterization of the S. cerevisiae genome by gene deletion and parallel analysis. Science 285, 901–906. Winzeler, E.A., Castillo-Davis, C.I., Oshiro, G., Liang, D., Richards, D.R., Zhou, Y., and Hartl, D.L. 2003. Genetic diversity in yeast assessed with whole-genome oligonucleotide arrays. Genetics 163, 79–89. Wodicka, L., Dong, H., Mittmann, M., Ho, M.-H., and Lockhart, D.J. 1997. Genome-wide expression monitoring in Saccharomyces cerevisiae. Nat. Biotechnol. 15, 1359–1367. Wohlschlegel, J.A. and Yates, J.R. 2003. Where’s Waldo in yeast? Nature 425, 671–672. Wu, L.F., Hughes, T.R., Davierwala, A.P., Robinson, M.D., Stoughton, R., and Altschuler, S.J. 2002. Large-scale prediction of Saccharomyces cerevisiae gene function using overlapping transcriptional clusters. Nat. Genetics 31, 255–265. Zhang, L., Miles, M.F., and Aldape, K.D. 2003. A model of molecular interactions on short oligonucleotide microarrays. Nat. Biotechnol. 21, 818–821. Zhou, H., Roy, S., and Natan, M.J. 2001. Solution and chip arrays in protein profiling. Trends Biotechnol. 19, 534–539. Zhou, H., Ranish, J.A., Watts, J.D., and Aebersold, R. 2002. Quantitative proteome analysis by solid-phase isotope tagging and mass spectrometry. Nat. Biotechnol. 19, 512–515. Zhu, H., Bilgin, M., Bangham, R., Hall, D., Casamayor, A., Bertone, P., Lan, N., Jansen, R., Bidlingmaier, S., Houfek, T., Mitchell, T., Miller, P., et al. 2001. Global analysis of protein activities using proteome chips. Science 293, 2101–2105. Zuzuarregui, A. and del Olmo, M. 2004. Expression of stress response genes in wine strains with different fermentative behavior. FEMS Yeast Res. 4, 699–710. Zuzuarregui, A., Carrasco, P., Palacios, A., Julien, A., and del Olmo, M. 2005. Analysis of the expression of some stress induced genes in several commercial wine yeast strains at the beginning of vinification. J. Appl. Microbiol. 98, 299–307. Zuzuarregui, A., Monteolivia, L., Gil, C., and del Olmo, M. 2006. Transcriptomic and proteomic approach for understanding the molecular basis of adaptation in Saccharomyces cerevisiae to wine fermentation. Appl. Environ. Microbiol. 72, 836–847.

This page intentionally left blank

CHAPTER

4 Monascus Rice Products Tseng-Hsing Wang* and Tzann-Feng Lin*

Contents

Abstract

I. The Taxonomy of Monascus spp. II. History of Using Monascus Rice Products in Asia III. Production Methods A. Traditional production B. Production of polyketide metabolites by Monascus spp. C. Pigments production D. Monacolin K production E. GABA production IV. Evidence for Health Benefits A. Cholesterol-lowering effect B. Other effects V. Safety Acknowledgments References

124 126 128 128 129 133 135 137 137 137 140 148 151 151

The fermentation products of Monascus, especially those produced by solid-state fermentation of rice, have been used as food and health remedies for over 1000 years in China. Monascus rice products (MRPs) are currently being used as health foods in the United States and many Asian countries such as Japan, Taiwan, China, Korea, Thailand, the Philippines, and Indonesia. Many studies have shown that Monascus spp. produce commercially viable metabolites, including food colorants, cholesterol-lowering agents, and antibiotics. The most important bioactive compound isolated from Monascus is monacolin K, which is identical to the potent cholesterol-lowering, antiatherosclerotic drug lovastatin, a 3-hydroxy-3-methylglutaryl

* Liquor Research Institute, Taipei 106, Taiwan, Republic of China Advances in Food and Nutrition Research, Volume 53 ISSN 1043-4526, DOI: 10.1016/S1043-4526(07)53004-4

#

2007 Elsevier Inc. All rights reserved.

123

124

Tseng-Hsing Wang and Tzann-Feng Lin

coenzyme A (HMG-CoA) reductase inhibitor. Several species of the genus Monascus also produce citrinin, a mycotoxin harmful to the hepatic and renal systems. Monacolin K and citrinin are polyketide fungal metabolites. The biosynthetic pathways leading to the formation of polyketides, including monacolin K and citrinin, have been elucidated in Aspergillus and Monascus. The concern for safety is, therefore, high for the development of MRPs as health foods. Other attractive applications for MRPs are likely, as supported by recent studies that indicate that MRPs contain other substances (flavonoids, polyunsaturated fats, phytosterols, pyrrolinic compounds, and others) with a wide variety of biological activities and pharmacological potentials. Their effects in lowering blood sugar and triacylglycerol while raising HDL-C are more pronounced than those of monacolin K alone. Beyond cholesterol lowering, MRP may also be an ideal candidate for the treatment of metabolic syndrome.

I. THE TAXONOMY OF MONASCUS SPP. Monascus spp. have been used as foods and medicines in the Orient for over 1000 years (Wong, 1982). In China and Taiwan, it has been called ‘‘Hong Qu,’’ ‘‘Hon-Chi,’’ ‘‘Anka,’’ or ‘‘Ang-kak’’ using the Chinese or Taiwanese phonetic alphabet. The Japanese use the name ‘‘Beni Koji’’ or ‘‘red Koji.’’ In the United States and Europe, it has been called ‘‘red rice,’’ ‘‘red-mold rice,’’ or ‘‘red Chinese rice.’’ Many publications and commercial products use ‘‘red yeast rice,’’ which is not an appropriate name for filamentous fungi. In taxonomy, the genus Monascus belongs to the family Monascaceae and to the order Eurotiales. The so-called yeast Saccharomyces belongs to the family Saccharomycetaceae and to the order Endomycetales. The genus was first suggested by van Tieghem (1884) over a century ago, when its species became known to Westerners as contaminants of cereals, starch, silage, and other agricultural products (Iizuka and Lin, 1980; Young, 1930). Some strains of Monascus are characterized by their economic importance, being involved in the fermented foods industry in the Orient (Hesseltine, 1965; Lin, 1975). Species of Monascus have frequently been found in fermented food, foodstuffs rich in starch, moldy high-moisture fruits, moldy silages, and soil. For example, seven strains of Monascus spp. have been isolated from the starters of Kaoliang in Taiwan and Kinmen by Lin (1975). It was found that there were at least six species of Monascus in the starters of Kaoliang Brandy, two from Kinmen and four from Taiwan. After the isolation of the first species by van Tieghem in 1884, a long debate followed concerning the nomenclature. Morphological, physiological, and biochemical characteristics, such as the shape of the colony, length of conidial chain, and production of pigment, have been considered suitable keys to the classification of

Monascus Rice Products

125

Monascus. Hawksworth and Pitt (1983) revised the genus to cover three species, namely M. pilosus K. Sato, M. purpureus Went, and M. ruber van Tieghem. Monascus sp. is a filamentous fungus producing single-cell conidia for asexual reproduction, or cleistothecium for sexual reproduction (Figure 1). Ascomata, a stalked cleistothecium, arises singly at the tip of stalk-like hyphae scattered on the mycelium. The ascomatal wall is composed of two distinct layers: the inner wall which results from the swelling of the tips of the stalk-like hyphae forming a vesicle-like structure and the outer layer consisting of hyphal branches growing out from A

B

FIGURE 1 The morphology of Monascus spp. observed under optical microscope (Panel A) and scanning electron microscopy (SEM) (Panel B).

126

Tseng-Hsing Wang and Tzann-Feng Lin

the base and fusing with the inner vesicle; asci evanescent at an early stage; hyaline, single-celled, ellipsoidal ascospores (Barker, 1903; Carels and Shepherd, 1975). Based on physiological and morphological characteristics, there are six major species: M. pilosus, M. purpureus, M. ruber, M. floridanus, M. pallens, and M. sanguineus. The phylogenetic relationships of the Monascus spp. have been determined by the sequences of the D1/D2 region of the large subunit (LSU) rRNA genes (Park and Jong, 2003). Their results showed M. ruber and M. pilosus could not be differentiated using these sequences. Park et al. (2004) reported the same results using the internal transcribed spacer (ITS) and the partial b-tubulin gene as molecular markers. Although these two species have been recognized as separate species before, recent molecular information has strongly indicated that they are actually the same. Many species of Monascus fungi are readily available to the public from several institutes having culture collections such as the American Type Culture Center (ATCC) in the United States, the National Institute of Technology and Evaluation Biological Resource Center (NBRC) in Japan, and the Food Industry Research and Development Institute in Taiwan assigned with ‘‘CCRC’’ prefix.

II. HISTORY OF USING MONASCUS RICE PRODUCTS IN ASIA The Monascus rice products (MRPs) that are produced especially by solidphase fermentation on rice have been used for over 1000 years. The use of MRP in China was first documented in Song Dynasty (sixteenth century), as ‘‘Jiuqu’’ to make rice wine (Lin et al., 2005b). Chinese brewers had used ‘‘Jiuqu’’ for thousands of years, but did not realize that its innate characteristics were based on microorganisms and enzymes. ‘‘Jiuqu,’’ or ‘‘Qu’’ in the Chinese phonetic alphabet, are molded cereals that are sources of enzymes necessary for the breakdown of carbohydrates and proteins in the grains. Many strains of Rhizopus, Mucor, Aspergillus, and Monascus genera and yeasts with superior hydrolyzing or fermenting power have been isolated from spent grains. MRP (Hong Qu) and its use in alcoholic drinks and in the food industry first appeared in the literature of the Song Dynasty. The traditional techniques of making MRP were recorded in ‘‘Ben Cao Gang Mu’’ of Li Shi-zhen (1578) and Song Yin-xing’s ‘‘Tian Gong Kai Wu’’ (1637). Traditionally, MRP is cultivated on steamed rice until the mycelium totally covers the whole surface of the grains and the product is used directly (Lotong, 1985; Su and Wang, 1983). Because large quantities of secondary metabolites and hydrolytic enzymes, such as a-amylase, b-amylase, glucoamylase, protease, and lipase, are produced by Monascus spp., MRP is widely used as a preservative for meat and fish

Monascus Rice Products

127

and as a colorant or flavor in foods. It is also used for brewing red rice wine and liquor in many Asian countries such as Japan, Taiwan, China, the Philippines, and Indonesia. In China, MRP is legally classified into four types (based on the differences in production methods and appearance of raw materials): ‘‘Ku Qu,’’ ‘‘Qing Qu,’’ ‘‘Se Qu,’’ and ‘‘Wu Yi Hong Qu.’’ Among them, ‘‘Ku Qu’’ is mainly used for making rice wine. ‘‘Qing Qu’’ may be used either for making rice wine or as a colorant for foods. ‘‘Se Qu’’ is mainly used as a colorant for foods, and its weight per unit volume is the lightest, due to the longest fermentation time. ‘‘Wu Yi Hong Qu’’ is a mixed culture of a Monascus sp. with Aspergillus niger. ‘‘Wu Yi Hong Qu’’ is used in making rice wine with higher amylase activity. The traditional areas of MRP production were centered in South China, in areas such as Fujian Province, Zhejiang Province, Jiangsu Province, Jiangxi Province, and Taiwan. Gutian, a town in Fujian Province, has been one of the most famous red rice production centers. For the use of MRP in Chinese cuisine, ‘‘Hong Cao’’ is often the first choice. It is used to stir-fry or steam meat such as roasted red pork ‘‘Hong Cao Pork,’’ ‘‘Hong Cao Chicken,’’ and ‘‘Hong Cao Fish.’’ Another traditional meat product is Chinese-style red sausage ‘‘La Chang.’’ The advantages of using Monascus spp. to give food color are that they are considered nontoxic and remain stable even when exposed to high temperature. These foods are important mainly because of their color, especially when they are used in the celebration of Chinese New Year as the red color implies ‘‘prosperity.’’ In addition to use in meat cuisine, MRP has been used in the fermentation industry for the preparation of red rice wines and foods such as sufu (or ‘‘Dou-Fu-Ru’’ in Chinese, a mold-fermented soybean curd product), fish sauce, fish paste, and red soybean curd (a cheese-like product used as a spice). Red rice wine is called ‘‘Hong Qu Jiu’’ in Chinese. This kind of rice wine has a long history and the MRP-producing areas associated with production of this rice wine are broad and scattered mainly over the Jiangsu, Jiangxi, Fujian, and Zhejiang provinces in China. This rice wine is brewed from polished glutinous rice with MRP and wheat Qu or rice Qu as saccharifying and fermenting agents. The wine is bright goldenyellow in color, has mellow aromas and elegant flavors, and leaves a relaxing and pleasant aftertaste. Kinetic studies of the headspace components from rice and agar with (experimental) and without (control) inoculation with a Monascus sp. have been carried out (Chung et al., 2004). The results showed that five alcohols, four esters, two ketones, and one furan, with odor activity values (OAV) > 1 dominated the overall flavor of the product. The application of MRP is not restricted to the above-mentioned products. For example, we have used MRP to invent methods for producing a vegetarian lactic acid beverage similar to yogurt (drinking yogurt) (Wang et al., 2003b) and a beer-like, alcohol-free fermented beverage (Lin et al., 2006a).

128

Tseng-Hsing Wang and Tzann-Feng Lin

MRP is not only utilized directly for food, but is also used indirectly such as to produce low-cholesterol eggs (Wang and Pan, 2003) and Arbor Acres broiler chickens (Wang et al., 2006). In addition to its value in preparing delicious dishes, MRP has also been used in traditional Chinese medicine for centuries to help maintain a healthy heart and circulatory system. A health-promoting effect is mentioned in an ancient Chinese pharmacopoeia of medicinal foods and herbs, ‘‘Ben Cao Gang Mu’’ of Li Shi-zhen (1578), where it is proposed to be a mild aid for gastric complication problems (indigestion, diarrhea, and others), blood circulation, and spleen and stomach health. Among the species of Monascus, M. purpureus has been identified and used in traditional Chinese medicine for the treatment of blood stasis, a disorder related to dyslipidemia and atherosclerosis (Journoud and Jones, 2004). The MRP prepared with the M. pilosus IFO 4520 strain produced by Gunze, Ltd., effectively reduces elevated blood pressure and was approved as a food supplement for specified health use in Japan (Himeno, 1997).

III. PRODUCTION METHODS A. Traditional production MRPs are produced by solid-phase fermentation on rice. A detailed description of making MRPs is found in the ancient Chinese book, ‘‘Tian Gong Kai Wu,’’ published in 1637. These traditional methods for manufacturing MRPs are very complicated and time consuming. Since the Monascus strain is slow growing, MRPs prepared with this strain are often contaminated with other fast-growing microorganisms during MRP making in the open environment. To prevent contamination, it is necessary to prepare ‘‘seed Koji’’ as a starter before making MRPs. The techniques for preparing ‘‘seed Koji’’ are proprietary and conducted by ‘‘masters’’ with special training. Nonglutinous varieties of rice are most suitable for preparing MRP, since kernels of glutinous varieties tend to stick together and thus reduce the surface to volume ratio of solid material which is critical to the growth of Monascus. The best raw material is longshaped, non-glutinous rice. It is first washed, soaked in water for about 1 day or more, and drained thoroughly. The moist rice is then cooked. On cooling, the steamed rice is mixed with a diluted vinegar solution or a solution of alum to acidify the raw material (because the Monascus spp. are acidophilic), then is inoculated with ‘‘seed Koji.’’ The inoculated rice is thoroughly mixed and then incubated at an appropriate temperature in the range of 33–42
C. During the first few days, the rice would have taken on a pink color and been stirred and shaken to redistribute the moisture

Monascus Rice Products

129

and kernels with respect to depth from the surface of the fermenting mass which should be spread out and piled up in turn. It may have been necessary to add an adequate amount of water to compensate for the moisture lost during incubation. Within about 2 weeks, the rice would take on a deep purplish red color with no observable clumping of the kernels.

B. Production of polyketide metabolites by Monascus spp. Many secondary metabolites with complex chemical structures, including pigments (Figure 2) and monacolins (Figure 2), are synthesized from the polyketide pathway in Monascus spp. (Simpson, 1986). Several effectors controlling the polyketide synthesis of Monascus have been reported by using submerged culture systems (Lin, 1991). Considerable research has been conducted on the industrial production of Monascus in complex liquid media (Shepherd and Carels, 1983). A

O

R O

O

O O R

rtn Rubropunctatin n-C5H11 mbn Monascorubrin n-C7H15 O R O

O

O O

R mnc Monascin n-C5H11 ank Ankaflavin n-C7H15 O R1 O

N O

R2

O R1

R2

rtm Rubropunctamine

n-C5H11

H

mbm Monascorubramine

n-C7H15

H

FIGURE 2

(continued)

130

Tseng-Hsing Wang and Tzann-Feng Lin

B

HO

8

R1

11 O H 10

9

R2 8

15 O

13

H

1

6

CH3 2

7 5

4a

3 4

R1

R2

CH3

O Mevinolin (monacolin K)

CH3

Monacolin L

CH3

H

Monacolin J

CH3

OH

Monacolin X

CH3

O

CH3

O

Monacolin M

CH3

O

O CH3

O

OH CH3

O

O

Compactin (ML-236B)

H

ML-236A

H

OH

ML-236C

H

H

O

C

OH

HOOC

O O

CH3 CH3

CH3

Citrinin

FIGURE 2 Chemical structures of selected second metabolites from Monascus. (A) Monascus pigments, (B) monacolins, (C) citrinin.

Monascus Rice Products

131

Based on the works of Hadfield et al. (1967) and Turner (1971), a scheme of the hypothetical routes for the biosynthesis of these pigments was proposed by Hajjaj et al. (2000). The condensation of 1 mol of acetylCoA with 5 mol of malonyl-CoA leads to the formation of a hexaketide skeleton by the polyketide synthase. Then a medium-chain fatty acid such as octanoic acid, likely produced by the fatty acid biosynthetic pathway, is bound to the hexaketide by a transesterification reaction to generate the orange pigment monascorubrin (or rubropunctatin on transesterification with hexanoic acid). Based on observations from an isotope-labeling experiment, it is probable that the Monascus pigments and monacolin K are made by similar polyketide-forming enzymes (Turner and Aldridge, 1983). The biosynthetic pathway of lovastatin in A. terreus has been investigated by nuclear magnetic resonance (NMR) and mass spectroscopy (Yoshizawa et al., 1994). These studies concluded that lovastatin is composed of two distinct polyketide chains joined through an ester linkage. Proof that these two polyketides are assembled by two discrete polyketide synthases came from the cloning and partial characterization of the lovastatin biosynthetic gene cluster from A. terreus (Hendrickson et al., 1999; Kennedy et al., 1999). Hajjaj et al. (1999b) have demonstrated that the biosynthesis of citrinin (Figure 2) by Monascus originates from a tetraketide instead of a pentaketide as has been found in A. terreus and Penicillium citrinum. Since the pigments are produced from a hexaketide, this suggests the existence of a branch point at the tetraketide level, which can account for differential production routes of pigments and citrinin in M. ruber (Hajjaj et al., 1999a). Several nutrient effectors in controlling polyketide biosynthesis are described separately below.

1. Carbon source Within the range from 4% to 10%, higher concentrations of glucose support higher production of monacolin K (Buckland et al., 1989). Furthermore, addition of glucose during the fermentation increases production to a greater extent than addition of the slowly utilized glycerol. One possible reason for the stimulation of monacolin K production by glucose may be its catabolite repression of nicotinamide adenine dinucleotide phosphate hydrogen (NADPH) generation as well as repression of tricarboxylic acid (TCA) cycle enzymes (Buchanan and Lewis, 1984; Buchanan et al., 1985). A decline in TCA cycle flux may lead to the accumulation of acetyl-CoA for the synthesis of polyketides (Demain, 1968). Utilization of carbon sources for growth appears to be strain specific. Yoshimura et al. (1975) have reported that 5% ethanol is a very good carbon source for pigment formation but Lin (1982) claimed that ethanol, when its concentration is higher than 2%, inhibits both growth and pigment production. Hajjaj et al. (2001) have shown that carbon source starvation is required for lovastatin biosynthesis by A. terreus.

132

Tseng-Hsing Wang and Tzann-Feng Lin

Cultured on glucose and glutamate, lovastatin synthesis is initiated when glucose consumption starts to level off.

2. Nitrogen source Nitrogen source regulation of polyketide biosynthesis is well known, for example, nitrate repression of aflatoxin formation (Bennett et al., 1979). However, utilization of different nitrogen sources often causes pH change during fermentation. This independently affects growth and pigment production (Carels and Shepherd, 1977, 1978, 1979; Shepherd and Carels, 1983; Wong et al., 1981) and has resulted in confusion with respect to nitrogen control of pigment synthesis. MOPS (3-[N-morpholino]propanesulfonic acid) buffer was used to overcome this problem (Lin, 1991). Nitrogen metabolism can lead to the formation of intermediates of the TCA cycle and thus influences the cycle (Berg et al., 2006). Bhatnagar et al. (1986) have shown that the NAD-requiring glutamate dehydrogenase (catalyzing the conversion of glutamate to a-ketoglutarate) is more active in a medium with ammonium as the sole nitrogen source than in a medium with ammonium plus asparagines as nitrogen sources. The high activity of the NAD-requiring glutamate dehydrogenase during the exponential growth phase results in the accumulation of a-ketoglutarate, which inhibits the TCA cycle, thus minimizing acetyl-CoA oxidation and making it available for increased aflatoxin synthesis.

3. Other Factors Other factors reported to have significant effects on the synthesis of polyketide mycotoxins and Monascus pigments include metals, oxygen, and temperature (Demain, 1986). Metals have important effects on secondary metabolism (Weinberg, 1989). The growth of M. purpureus is inhibited, but the production of certain pigments is promoted, by the addition of zinc sulfate (Johnson and McHan, 1975; Wong, 1982; Wong and Bau, 1977). In the case of the production of Monascus pigments, Su (1978) found that the agitation of fermentation broths at 500 and 700 rpm yielded the same mycelial form, but pigment formation was higher at 500 rpm than at 700 rpm. The lower production of Monascus pigments at the higher agitation rate may be caused by the inhibition of pigment formation by oxygen and/or shear stress on the mycelia (Yoshimura et al., 1975). An additional problem is that polyketide formation requires acetyl-CoA, malonyl-CoA, and NADPH generated by primary metabolic pathways. These precursors and the cofactor are also used for fatty acid biosynthesis. An inverse relationship between the synthesis of fatty acids and polyketide compounds has been found in the mevinolin (lovastatin)-producing species of Aspergillus (Dutton, 1988; Greenspan and Yudkovitz, 1985). Thus, any regulatory factor that substantially alters the rate or extent of formation of these precursors and cofactor may affect polyketide formation.

Monascus Rice Products

133

C. Pigments production There are six well-known Monascus pigments (azaphilones) that are produced and are divided into three pairs. Rubropunctatin (C21H22O5) and monascorubrin (C23H26O5) are orange pigments with different aliphatic side chains on the b-ozo-lactone ring. The two corresponding red pigments, rubropunctamine (C21H23NO4) and monascorubramine (C23H27NO4), are nitrogen analogues of the orange pigments. The yellow pigments, monascin (C21H26O5) (syn. monascoflavin) and ankaflavin (C23H30O5), are the reduced forms of the orange pigments (Kurono et al. 1963; Sweeny et al., 1981). Carels and Shepherd (1977) have proposed that the orange pigments are initially synthesized and the yellow and red pigments are derived from the orange counterparts. Monascus pigments are stable in the pH range of 2–10. They are also heat stable and can be autoclaved (Francis, 1987). Mutation is a useful technique for the enhancement of pigment production. For example, Lin and Iizuka (1982) have derived a hyperpigmentproducing mutant, M. kaoliang R-10847, through a series of mutagenesis steps. This mutant, derived from an intracellular parent, produces extracellular pigment. The productivity of pigment is about 100-fold greater than that of the wild type. Pigment production can also be improved by optimizing the culture condition of Monascus. Shin et al. (1998) showed that when a Monascus isolate was cocultured with either Saccharomyces cerevisiae or A. oryzae in a solid sucrose medium, there were significant morphological changes in the Monascus culture. Cocultures exhibited cell mass increases of 2 times and pigment yield increases of 30–40 times compared to monocultures of Monascus. The hydrolytic enzymes produced by S. cerevisiae, such as amylase and chitinase, are thought to be the effectors. Using transformation systems, Campoy et al. (2003) were able to manipulate the natural pigment producers, M. purpureus and M. ruber. The high-level pigment-producing Monascus strain IBCC1 was characterized by random amplification of polymorphic DNA as M. purpureus. Their results showed that 60% of the Monascus transformants were found to be stable in mitosis and retained the plasmid inserted in the chromosome after repeated sporulation cycles. The transformants obtained by Agrobacterium-mediated DNA transfer remained fully stable (98%) after four sporulation rounds and showed bands of hybridization corresponding to integration of the plasmid in different sites of the genome. A characterization of a non-pigment-producing mutant, M. purpureus M12, compared with its parental strain, M. purpureus Went CBS 109.07, has been performed aiming to investigate the relation between pigment biosynthesis and other characteristics (Rasheva et al., 2003). In a selected albino mutant, the growth rate, metabolic activity, and capacity for polyketide production typical for Monascus secondary metabolites were reduced considerably. The mutant strain produced C17, C20, and C22 fatty acids but did not produce citrinin. By immobilized repeated-batch

134

Tseng-Hsing Wang and Tzann-Feng Lin

processes, extracellular production of M. purpureus C322 pigment was studied by Fenice et al. (2000). Using Ca-alginate as an immobilizing carrier, pigment production reached a plateau while the cell leakage was negligible and mechanical stability of the Ca-alginate bead was good. By the addition of 20 amino acids, the red derivative pigments were produced (Jung et al., 2003). Liquid chromatography-mass spectroscopy (LC-MS) and 1H and 13C NMR structural analyses confirmed that the derivative pigments contained the moieties of the added amino acids. These pigments showed enhanced photostability (Jung et al., 2005). Under sunlight, the half-life of derivatives was increased to 1.45–5.58 hours, corresponding to a 6- to –25-fold improvement over a control red pigment (0.22 hours). MRP has been cultured traditionally on rice and other cereals by solidstate fermentation. However, for large-scale cultivation, solid-state cultures were associated with some problems such as contamination and scale-up requirements. There have been reports that Monascus could be cultured in submerged culture systems; however, the pigment production in submerged culture was reduced to one-tenth of that achievable in the solid-state fermentation (Lin, 1973). For the submerged culture studies with rice particles, a stirred-tank fermentor was not suitable as the impeller tended to break the particles into small pieces. A conventional bubble column was also unsuitable since its mixing capability was poor. Wu et al. (2000) developed a modified bubble column with wire-mesh draft tubes for the cultivation of M. purpureus. The production of pigments using the proposed column was 80% higher than that achieved using the conventional bubble column. A process combining solid-state and submerged cultivations, intermittently rinsing the rice with monosodium glutamate (MSG) solutions every 12 h, shows advantage. Following an adsorptive extraction of the red pigment dissolved in the rinsing solution, the Monascus red pigment production was found to increase by 24% as compared with that realized by the plain fixed-bed cultivation (Hsu et al., 2002). Considerable research has been conducted on the industrial production of Monascus pigments in complex liquid media (Campoy et al., 2005; Krairak et al., 2000). In general, high broth viscosity is a key factor to be considered in a submerged fermentation of filamentous fungi. The resultant high viscosity induced heterogeneity inside the fermentor, poor oxygen transfer, and low pigment yield. However, these problems could be overcome by reducing fungal growth rate by culturing at low temperature (25
C). As a result, the pigment yield at 25
C was 10 times greater than at 30
C (Ahn et al., 2006). Light is another factor. In nature, light is one of the most crucial environmental signals for developmental and physiological processes. Miyake et al. (2005) have found that both red and blue lights affect development in Monascus, influencing the processes of mycelium and spore formation, and the production of secondary metabolites such as g-aminobutyric acid (GABA), red pigments, monacolin K, and citrinin.

Monascus Rice Products

135

Although toxicity studies show that Monascus pigments are safe for human consumption (Su and Wang, 1983), low solubility in water and the sensitivity to decoloration by sunlight restricted the wide use of the Monascus pigments in the beverage and confectionary industries (Sweeny et al., 1981). Thus, many patents have focused on the improvement of extraction and solubilization of Monascus pigments (Wong, 1982). Several chemical processes have been patented for semisynthetic watersoluble red pigments (Moll and Farr, 1976). Direct production of water-soluble pigments by fermentation offers a more acceptable alternative to the total and semisynthetic processes, since it avoids the use of ‘‘chemical additives’’ in foods (Spears, 1988). We have developed a chemically defined medium and a resting-cell system, and discovered new water-soluble red pigments (Lin, 1991). These were isolated and characterized as glutamate derivatives of the orange pigments (rubropunctatin and monascorubrin). These new pigments showed superior properties such as higher water solubility, higher absorption coefficient (e value), and greater resistance to decoloration by light.

D. Monacolin K production Initially identified from Monascus spp., monacolin K (C24H36O5) is a polyketide, which is structurally identical to lovastatin (Endo, 1979, 1980). In addition to monacolin K, there are several other minor monacolins (Ma et al., 2000). At least six structurally related monacolins have been identified from the genus Monascus, namely monacolin J, K, L, and X, dihydromonacolin K, and dihydromonacolin L (Endo et al., 1985a,b). Many analytical procedures based on high-performance liquid chromatography (HPLC) and LC-MS have been developed for the determination of lovastatin and other statins in biological samples. Li et al. (2004) developed a chemical fingerprint-profiling method using HPLC with a photodiode array (PDA) detector and tandem mass spectrometry (MS/MS). A fingerprint profile containing 14 monacolin compounds, including monacolin K (mevinolin), J, L, M, and X, and their hydroxyl acid forms, as well as dehydromonacolin K, dihydromonacolin L, compactin, and a-hydroxy-3, 5-dihydromonacolin L in Monascus spp., can be obtained using that method. Another detection method for the assay of the monacolin series compounds was established by Li et al. (2005c). By using reversed phase HPLC (RPHPLC) with PDA, well-resolved peaks of seven main compounds of the monacolin family were profiled. The method was established with a C18 reverse-phase column using a linear gradient of 0.1% trifluoroacetic acid and acetonitrile as the mobile phase, and the detection wavelength was set at 237 nm. Li et al. (2005c) used this method to study the stability of monacolin K under different storage conditions. The results showed that the monacolins in MRP powder are light sensitive and thermal sensitive. Monacolins

136

Tseng-Hsing Wang and Tzann-Feng Lin

decomposed significantly under the conditions of high humidity at high temperature (75% RH, 60
C) and sunlight. Monacolin K and its hydroxyl acid form would be dehydrated and converted to dehydromonacolin K at high temperature (80
C) while the monacolin K, J, and L would be transformed into their corresponding hydroxyl acid forms under the condition of high humidity (92.5% RH, 25
C). Only a few species of Monascus can produce monacolin K (Table 1). Since Monascus pigments and monacolin K are made by the same or similar polyketide-forming enzymes (Turner and Aldridge, 1983), the ability of various species of Monascus to produce monacolin K may be predicted based on its mycelia color. Response surface methodology (RSM) was employed by Chang et al. (2002) to study the effect of the composition of the rice–glycerol complex medium on the production of monacolin K by M. ruber in mixed solid– liquid (or submerged) cultures at 25
C. The best composition (in g/liter) derived from RSM regression was rice powder 34.4 g, peptone 10.8 g, glucose 129 g, KNO3 8.0 g, MgSO4 7H2O 4.0 g, and glycerol 36.4 ml per liter. With this composition, the monacolin K production was 157 mg/liter after 10 days of cultivation. In a study by Miyake et al. (2006), it was found that M. pilosus required a suitable concentration of organic peptone for high monacolin K production. They had developed a glucose-glycerol-peptone (GGP) medium which contained 3% glucose, 7% glycerol, 3.8% peptone, 0.1% MgSO47H2O, and 0.2% NaNO3; and with this medium, M. pilosus MK-1 produced the highest level of monacolin K. For the production of monacolin K-containing MRP on an industrial scale especially, we have successfully developed a fermentation process by cultivating a low-citrinin, high-monacolin K-producing strain of M. purpureus on rice in an aseptic rotary vessel to minimize microbial contamination due to the slow growth of Monascus. In brief, steamed rice was inoculated with M. purpureus with an inoculum of 100 spores/kg raw rice. The inoculated rice was cultivated at 30
C for 5 days until the steamed rice turned a deep red color. The colored rice was used as ‘‘seed Koji.’’ The ‘‘seed Koji’’ was uniformly mixed with steamed rice at a ratio of 3% based TABLE 1 The species of Monascus that produce monacolin K Species

Monacolin K production

References

M. ruber M. purpureus NTU 601 M. ruber M. pilosus M. purpureus NTU 301 M. pilosus MK-1

17.4 mg/liter 0.53 mg/g 2.5–3.0 mg/g 2.52 mg/g 2.584 mg/g 725 mg/liter

Endo, 1979 Wang et al., 2003 Wei et al., 2004 Chen and Hu, 2005 Lee et al., 2006 Miyake et al., 2006

Monascus Rice Products

137

on the weight of steamed rice. The resulting mixture was soaked in water twice and cultivated for 6 days at a temperature of 35
C and a relative humidity below 95% to produce Monascus rice with a bright red color. Since the formation of the secondary metabolites of the Monascus spp. is affected by cultivation conditions, Lee et al. (2006) used sweet potato (Ipomoea batatas), potato (Solanum tuberosum), cassava (Manihot esculenta), and dioscorea (Dioscorea batatas) as the substrates to identify the best choice for monacolin K production. The results showed that M. purpureus NTU 301, with dioscorea as the substrate, could produce monacolin K at 2584 mg/kg, which is 5.37 times more than that resulted when rice is used as the substrate. For storage, monacolins in MRP powder decreased significantly under the conditions of high humidity at high temperature (75% RH, 60
C) and sunlight. Therefore it has been suggested that the preparations containing monacolins be stored in a cool and lightproof place (Li et al., 2005c).

E. GABA production GABA is produced by the decarboxylation of glutamic acid by glutamate decarboxylase. In the process of making ‘‘Jiuqu’’ (mold-containing MRP for making rice wine), glutamic acid is produced from steamed rice by an acid protease and an acid carboxypeptidase that are secreted by the mold. GABA has several physiological functions, including neurotransmitting, hypotensive, and diuretic effects (Keisuke et al., 1992; Matheson et al., 1986). The changes in GABA content during the preparation of MRP have been examined by Kono and Himeno (2000). When prepared with M. pilosus IFO 4520, the production of GABA in MRP peaked on the fifth day and thereafter declined. Su et al. (2003) showed that the amounts of GABA produced by different strains varied greatly. In their study, solid-state cultivation always produced more GABA than submerged cultivation did. The addition of sodium nitrate during the solid-state fermentation of M. purpureus improved the productivity of GABA to 1267.6 mg/kg. GABA productivity increased further to 1493.6 mg/kg when dipotassium hydrophosphate was added to the medium. Wang et al. (2003a) showed that addition of 0.5% ethanol increased production of GABA from 1060 to 7453 mg/kg.

IV. EVIDENCE FOR HEALTH BENEFITS A. Cholesterol-lowering effect 1. The inhibitor of HMG-CoA reductase Hypercholesterolemia, especially elevated plasma low-density lipoprotein cholesterol (LDL-C), is a key risk factor leading to the pathogenesis of atherosclerosis (Steinberg, 2002). The identification of compactin

138

Tseng-Hsing Wang and Tzann-Feng Lin

(ML-236B) from P. citrinum (Endo et al., 1976) and P. brevicompactum (Brown et al., 1976) and lovastatin (mevinolin) from A. terreus (Alberts et al., 1980) and M. purpureus as the inhibitors of 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase has greatly advanced the development of cholesterol-lowering drugs. Inhibition of hepatic HMG-CoA reductase, the rate-limiting enzyme in the cholesterol biosynthetic pathway, stimulates the expression of LDL receptor (also called ApoB/E receptor) (Brown and Goldstein, 1986). Increased uptake of LDL through a receptor-mediated pathway reduces plasma LDL-C. Many clinical studies have shown that lovastatin and other statin drugs reduce plasma total cholesterol (TC) and LDL-C. Treatment with lovastatin also reduces plasma triacylglycerols (TG) and increases high-density lipoprotein cholesterol (HDL-C) to an extent less than the magnitude of TC and LDL-C lowering. MRP is used for dietary supplements with the health claim that the product will lower plasma lipids, especially plasma TC and LDL-C. The MRP supplements contain a family of naturally occurring statins (monacolins) including monacolin K. There are many commercial products of MRP worldwide. Among them ‘‘Cholestin’’ (2.4 g/day contained 9.6 mg total monacolins, Pharmanex, Inc., Simi Valley, CA), ‘‘Xuezhikang’’ (1.2 g/day contained 13.5 mg total monacolins, WBL Peking University Biotech Co., Ltd., China), and ‘‘Unchole’’ (1.0 g/day contained 8.0 mg total monacolins, Taiwan Tobacco & Liquor Corp., Taiwan, ROC) have demonstrated a cholesterol-inhibiting effect similar to statins in animal studies and clinical trials.

2. Animal studies The effects of ‘‘Xuezhikang’’ on plasma cholesterol and functions of endothelial cells in cholesterol-fed rabbits have been studied by Wu et al. (2003). After a 12-week feeding experiment, serum TC, LDL-C, TG, and plasma endothelin-1(ET-1) decreased and serum NO level increased in the ‘‘Xuezhikang’’ group as compared with those of the hypercholesterol-fed control group (p < 0.05). The areas of lipid deposition on the intimal surface of the aorta and coronary arteries were reduced and the ultrastructural injuries of endothelial cells were milder in the ‘‘Xuezhikang’’ group. Apolipoprotein E-deficient [ApoE (/)] mice are the common animal model and display high similarity to human atherosclerosis. Using ApoE (/) mice, Zheng et al. (2003) have shown that ‘‘Xuezhikang’’ lowers the serum TC, TG, and LDL-C and reduces the atherosclerotic lesions after a 14-week feeding. For animal studies, hamsters are considered to be one of the most suitable animal models for human lipid and lipoprotein metabolism (Harris, 1997; Ntanios and Jones, 1999). Using this animal model fed with products from a Monascus strain, M. purpureus NTU568, similar results with decreases in TC, TG, and LDL-C levels were shown by Lee et al. (2005). Long-term feeding effects of M. purpureus-fermented rice (Cholestin) on serum lipids and the severity of atherosclerosis were examined in

Monascus Rice Products

139

rabbits fed for 200 days on a semipurified diet containing 0.25% cholesterol (Wei et al., 2003). Total serum cholesterol was 25% and 40% lower, respectively, in rabbits fed 0.4 or 1.35 g/kg/day of MRP compared to those of controls. This treatment also lowered serum LDL-C, serum TG, and the atherosclerotic index (ratio of non-HDL-C to HDL-C). Although similar reductions of TC, LDL-C, and TG were observed, a parallel group of rabbits fed with lovastatin (0.0024 g/kg/day) did not have a significantly reduced atherosclerotic index. We have used hamsters as an animal model to elucidate the effects of a monacolin K-containing M. purpureus rice product (‘‘Unchole,’’ 1.0 g/ day containing 8.0 mg total monacolins, Taiwan Tobacco & Liquor Corp.) on serum lipid and lipoproteins (Lin et al., 2005b; Table 2). Results showed that MRP treatment lowered total serum cholesterol and LDL-C. The MRP treatment also increased the secretion of fecal cholesterol, which was not found in the lovastatin-treated group (Lin et al., 2005b; Table 3). Treatment with ‘‘Unchole’’ also lowered serum TG, although its TG-lowering mechanism is unclear. TABLE 2 Lipid and glucose concentrations of hyperlipidemic hamster after 31 days feeding study Basline (n ¼ 8)

Control (n ¼ 8)

‘‘Lovastatin’’ (n ¼ 8)

‘‘Unchole’’ (n ¼ 8)

277.0  33.9 274  94 222  34 79.5  26.2 135.5  6.4 62.0  1.4

235.0  33.9 268  76 174  27 49.5  2.1 106.0  29.7 79.5  2.1

184.5  24.7 156  64 114  32 14.5  3.5 88.0  12.7 82.0  8.5

(mg/dl) TC TG Glucose VLDL-C LDL-C HDL-C

98.5  0.7 239  46 192  42 17.0  8.5 31.0  7.1 50.5  16.3

Baseline, chow diet (Purina 5001); Control, chow diet plus 0.25% cholesterol (w/w) and 5% soybean oil; ‘‘Lovastatin,’’ Control diet plus Lovastatin (50 mg/kg); ‘‘Unchole,’’ Control diet plus ‘‘Unchole’’ (25 g/kg). HDL-C was determined by the following formula: (HDL-C) ¼ TC  (LDL-C)  ( VLDL-C). Results were shown as mean  S.D. Data with same superscript in the same row were not significantly different ( p > 0.05).

TABLE 3 Fecal cholesterol

Cholesterol (mg/g dry feces)

Basline (n ¼ 8)

Control (n ¼ 8)

‘‘Lovastatin’’ (n ¼ 8)

‘‘Unchole’’ (n ¼ 8)

5.8  0.4

11.6  0.2

13.4  0.6

19.4  0.1

Fecal cholesterol was recovered by saponification and extraction with n-hexane. The recovered cholesterol was determination by a modified Liebermann-Burchard method.

140

Tseng-Hsing Wang and Tzann-Feng Lin

The most plausible explanation of monacolin K-containing MRP reducing plasma cholesterol and LDL-C, beyond the contribution of its monacolin K content, may be due to the additive and/or synergistic effects of monacolin K with other monacolins and substances (such as phytosterols) in MRP. In addition to the phytosterols (b-sitosterol, campesterol, stigmasterol, and sapogenin), MRP has been found to contain isoflavones and isoflavone glycosides, and monounsaturated fatty acids (Heber et al., 1999). Based on the observation that the solid-phase fermentation is a prolonged and extensive process (which is a slow process of over 10 days), we propose that the resulting Monascus preparation is a product of extensive utilization of starch and significant enrichment of phytosterols. Many studies have demonstrated the cholesterol-lowering effects of plant sterols (Ostlund, 2002). Plant sterols may competitively inhibit the intestinal absorption of cholesterol and therefore lower its level in the plasma.

3. Clinical studies

Several pilot studies (Cicero et al., 2005; Liu et al., 2003; Sumioka et al., 2006) and clinical trials (Table 4) have demonstrated that the intake of MRP significantly decreases TC, LDL-C, and TG in subjects, without causing clinically adverse effects in the liver and muscle tissue. Since doubleblind, randomized, placebo-controlled prospective studies have been performed on subjects with ‘‘Xuezhikang,’’ ‘‘Cholestin,’’ and other MRPs, it suggests that MRP can be a safe and efficacious agent for subjects at risk for cardiovascular diseases. Although one proprietary strain of MRP has been demonstrated to lower cholesterol levels significantly in clinical trials, not all strains being sold as dietary supplements have undergone similar evaluation. In order to determine whether the results of a clinical trial conducted with one strain of MRP could be extended to other preparations of MRP, nine different commercially available dietary supplements were purchased and tested for chemical constituents by Heber et al. (2001). They concluded that standardized manufacturing practices should be established for MRP as a dietary supplement. It can ensure equivalence of content of active ingredients in preparations being sold to the public and to limit the production of unwanted by-products of fermentation such as citrinin. There is still a need for further study on the long-term safety and efficacy of MRP as a dietary supplement in a larger population. The effects in lowering TG and increasing HDL-C also need to be elucidated in future studies.

B. Other effects Besides monacolin K (lovastatin, once the world’s largest selling class of cholesterol-lowering drugs), the Monascus products also contain many other substances (flavonoids, polyunsaturated fats, pyrrolinic compounds, and so on) with a wide variety of actions. Their effects may be

TABLE 4 Preliminary clinical data of MRP in colesterol-lowering effect References

Design

Dosage

Patients(n)

Results

Kou et al., 1997

R

Xuezhikang (1.2 g/day, 13.5 mg total monacolins) Zocor (10 mg/day) 8 weeks

108 patients with primary hyperlipidemia

Xuezhikang

Xuezhikang (1.2 g/day) Gemfibrozil (1.2 g/day) 8 weeks

91 patients with hyperlipidemia

comparative study of Xuezhikang and Simvastatin (Zocor)

Jing et al., 1999

R comparative study of Xuezhikang and Gemfibrozil

TC: 23% (p < 0.001) TG: 28.1% (p < 0.001) LDL-C: 28% (p < 0.001) HDL-C: þ5% (p > 0.05) Zocor TC: 23.3% (p < 0.001) TG: 29.5% (p < 0.001) LDL-C: 29.5% (p < 0.001) HDL-C: þ14.3% (p < 0.01) Xuezhikang TC: 21.6% (p < 0.01) TG: 23.3% (p < 0.01) LDL-C: 33.3% (p < 0.01) HDL-C: þ33.7% (p < 0.01) LP(a): 28.2% (p < 0.01) TXB2: 34.2% (p < 0.01) 6-keto-PGF1a: þ65.4% (p < 0.01) Gemfibrozil TC: 20.4% (p < 0.01) TG: 40.3% (p < 0.01) LDL-C: 24.8% (p < 0.01) HDL-C: þ26.9% (p < 0.01) LP(a): 4.9% (p < 0.01) TXB2: 8.4% (p < 0.01) 6-keto-PGF1a: þ11.7% (p < 0.01) (continued)

TABLE 4 (continued)

142

References

Design

Dosage

Patients(n)

Results

Heber et al., 1999

PC; R; DB

Cholestin (2.4 g/day, 9.6 mg total monacolins) 8 and 12 weeks

83 healthy sunjects with hyperlipidemia

Lin et al., 2003

PC; R

Xuezhikang (1.2 g/day) 8 weeks

60 CHD patients

Zhao et al., 2004

PC; R

Xuezhikang (1.2 g/day) 6 weeks

50 CHD patients

8 weeks TC: 16.8% (p < 0.05) TG: 11.3% (p < 0.05) LDL-C: 22.3% (p < 0.05) HDL-C: 0% 12 weeks TC: 16.1% (p < 0.05) TG: 6.7% (p < 0.05) LDL-C: 21.9% (p < 0.05) HDL-C: 0% TC: 21% (p < 0.05) TG: 25% (p < 0.05) LDL-C: 30% (p < 0.05) HDL-C: þ16% (p < 0.05) LP(a): 23% (p < 0.05) TC: 18.8% (p < 0.001) TG: 31.1% (p < 0.001) LDL-C: 28.3% (p < 0.001) HDL-C: þ17.4% (p < 0.001) Hs-CRP: 50% (p < 0.001)

Li et al., 2005b

R

Xuezhikang (1.2 or 2.4 g/day)

48 patients with stable angina

79 patients with a mean baseline LDL-C level of 203.9 mg/dl 4870 CHD patients

1.2 g/day, median CRP: 28.6%; TC: 13%; LDL-C: 23%, (p < 0.05) 2.4 g/day, median CRP: 30.4%; TC: 22%; LDL-C: 32%, (p < 0.01) LDL-C: 27.7% TC: 21.5% TG: 15.8% ApoB: 26.0% Incidence of nonfatal MI was reduced by 60.8% (p < 0.0000)

591 CHD patients with diabetes

Incidence of death from CHD reduced by 31.0% (p < 0.0048) Total mortality was lowered by 33.0% (p ¼ 0.0003) Incidence of CHD events reduced by 50.8% (p ¼ 0.0008)

14 days

Lin et al., 2005a

PC; R; DB

M. purpureus Went rice (600 mg) 8 weeks

Lu and Fu, 2005

PC; R; DB

Xuezhikang (0.6 g, bid) þ conventional therapy 4 years

Lu et al., 2005

PC; R; DB

Xuezhikang (0.6 g, Bid) þ conventional therapy 4 years

Incidence of death from CHD reduced by 44.1%(p ¼ 0.0246) Incidence of nonfatal MI was reduced by 63.8% (p ¼ 0.0151) Total mortality was lowered by 44.1% (p ¼ 0.0097)

143

(continued)

TABLE 4 (continued) References

Design

Dosage

Patients(n)

Results

Du et al., 2006

PC

Xuezhikang (1.2 g/day)

2135 patients with MI history of 28 days to 3 months (group A) 2735 patients with MI history of 3–60 months (group B)

Group A: reduced the risk of CHD events by 56.7% (p < 0.0001)

8 weeks

Group B: decreased the risk of CHD events by 5.3% (p ¼ 0.0008)

PC ¼ Placebo-controlled; R ¼ randomized; DB ¼ double-blind; ACS ¼ acute coronary syndrome; Hs-CRP ¼ high sensitivity-C reactive protein; MMP-9 ¼ matrix metalloproteinase-9; CHD ¼ coronary heart disease; MI ¼ myocardial infarction; LP(a) ¼ lipoprotein a; ApoB ¼ apolipoprotein B; TXB2 ¼ thromboxane B2; PGF1a ¼ prostaglandins F1a.

Monascus Rice Products

145

more extensive and complex than those of statins alone. It also makes MRP an ideal candidate for the treatment of the metabolic syndrome (Bianchi, 2005). The most recent studies are described briefly as follows and are summarized in Table 5.

1. Antihypertensive effect The vasodilatory effects of an aqueous extract of MRP fermented with M. ruber IFO 32318 were examined on the isolated rat aorta by Rhyu et al. (2000). The results showed that MRP-induced aortic relaxation involved the release of NO from endothelium. It seems that an unknown factor(s), other than acetylcholine (Ach) and GABA, in the aqueous extract of MRP, may stimulate vascular endothelial cells to produce and/or release NO. Hsieh and Tai (2003) showed that the intragastric loading of fructosefed rats with M. purpureus M9011 containing GABA(1 mg/kg/day) prevented the development of fructose-induced hypertension. Additionally, they tested the reverse effect. After fructose-induced hypertension had been established, intragastric loading of M. purpureus M9011 reversed the elevated blood pressure to a normal level. However, administration of pure GABA at the same dose as that contained in M. purpureus M9011 failed to prevent or reverse hypertension due to high fructose consumption. Prolonged M. purpureus M9011 treatment significantly suppressed the fructose-induced elevation in plasma TC and improved the HDL-C:TC ratio.

2. Antioxidant effect

The antioxidant and hepatoprotective actions of M. anka against acetaminophen (AAP)-induced liver toxicity have been investigated (Aniya et al., 1998). Their results show that M. anka prevents AAP-induced liver toxicity by both antioxidant action and the inhibition of AAP metabolism. Further antioxidant action of M. anka was studied in vitro and in vivo TABLE 5 Summary of the biological activities and pharmacological potentials (see text in detail) reported to MRP

1. Cholesterol-lowering effect 2. Antihypertensive effect 3. Antioxidant effect 4. Antihyperglycemic activity 5. Antiproliferate effect 6. Suppression of adipogenesis 7. Antimicrobial activity 8. Macrophage-stimulating activity

146

Tseng-Hsing Wang and Tzann-Feng Lin

(Aniya et al., 1999). Antioxidant activity was evaluated by scavenging stable free radical 1,1-diphenyl-2-picrylhydrazyl (DPPH) and lipid peroxidation of rat liver microsomes. M. anka was shown to have the strongest action. When galactosamine (GalN, 400 mg/kg) or GalN plus lipopolysaccharide (LPS, 0.5 mg/kg) was given intraperitoneally to SpragueDawley rats, aspartate aminotransferase (AST) and glutathione (GSH) S-transferase (GST) activities in serum were significantly increased. Hepatotoxicity marked by an increase in serum enzyme levels was reduced when the extract prepared from M. anka was given 1 and 15 hours before the toxic insult. In further studies, Aniya et al. (2000) isolated and identified the antioxidant component of M. anka as dimerumic acid. When the dimerumic acid (12 mg/kg) was given to mice prior to a carbon tetrachloride (CCl4, 20 ml/kg, ip) treatment known to elicit liver toxicity in mice, the elevation of serum AST and alanine aminotransferase (ALT) activities was decreased, suggesting a hepatoprotective action of dimerumic acid. The antioxidant mechanism of dimerumic acid is due to one electron donation of the hydroxamic acid group in the dimerumic acid molecule toward oxidants, resulting in formation of nitroxide radical (Taira et al., 2002).

3. Antihyperglycemic activity

MRPs produced by fermentation with M. pilosus and M. purpureus were used for antihyperglycemic activity screening in streptozotocin-induced diabetic rats (STZ-diabetic rats) (Chang et al., 2006). Single oral administration of MRP decreased plasma glucose in STZ-diabetic rats in a dosedependent manner from 50 to 350 mg/kg. Moreover, mRNA levels of phosphoenolpyruvate carboxykinase (PEPCK) in liver from STZ-diabetic rats were reversed in a dose-dependent manner by the repeated oral treatment of MRP 3 times daily for 2 weeks. These results suggest that oral administration of MRP could decrease hepatic gluconeogenesis to lower plasma glucose in diabetic rats with insulin deficiency. The hypoglycemic effect of MRP was also studied by another group, Chen and Liu (2006). Oral administration of MRP, fermented with M. pilosus and M. purpureus for 90 min to fasting Wistar rats resulted in a decrease in plasma glucose in a dose-dependent manner. In parallel to the reduction of plasma glucose, an increase in the plasma level of insulin and C-peptide was also observed. The study also suggests that MRP has an ability to stimulate the release of acetylcholine from nerve synapses, which in turn stimulates muscarinic M(3) receptors in pancreatic cells and augments insulin release to result in a plasma glucose-lowering action.

4. Antiproliferative effect Using a cell-based cytotoxicity assay, a cytotoxic compound was found in Monascus by Su et al. (2005). Ankaflavin, but not monascin, was found to be toxic to human cancer cell lines (Hep G2 and A549) with a similar IC50

Monascus Rice Products

147

value of 15 mg/ml, but posed no significant toxicity to normal cells (MRC-5 and WI-38) at the same concentration. From a morphological observation of the chromatin, Su et al. proposed that apoptosis should be the possible mechanism. To elucidate the molecular mechanisms responsible for the antiproliferative effect of monacolin K in cancer cells, Lin et al. (2006b) have used proteomic analysis by two-dimensional gel electrophoresis, matrix-assisted laser desorption ionization time-of-flight/time-of-flight mass spectrometry (MALDI-TOF/TOF MS), MS/MS, and database interrogation to separate and identify the proteins of Caco-2 cells treated with monacolin K. Their results showed that monacolin K inhibited the proliferation of Caco-2 cells in a dose-dependent manner. The proteins identified in the proteomic analysis included antioxidation enzymes related to reactive oxygen species stress, cytoskeleton proteins, glycolytic enzymes, and enzymes involved in mediating protein interactions.

5. Suppression of adipogenesis

Jeon et al. (2004) demonstrated that MRP extracts, which were extracted from embryonic rice fermented with M. ruber, significantly decreased glycerol-3-phosphate dehydrogenase (GPDH) activity and lipid accumulation, a marker of adipogenesis, in a dose-dependent manner. Moreover, MRP extracts significantly decreased gene expression of adipocyte fatty acid binding protein (aP2) and leptin, two adipogenic marker proteins and C/EBPa and PPARg target genes. Their results suggested that the inhibitory effect of MRP extracts on adipocyte differentiation might be mediated through the downregulated expression of adipogenic transcription factors and other specific genes.

6. Antimicrobial activity MRP has been used as a food preservative from old times. The antibacterial activity of M. purpureus was demonstrated by Wong and Bau (1977). The active compound(s) was named monascidin, and is a potent but not a broad-spectrum antimicrobial agent. It is effective against Bacillus spp. tested, especially B. megaterium. In Japan, several species of Monascus, such as M. paxii (Matsumoto et al., 1989), Monascus sp. ATCC 16775 (Araki et al., 1998), and M. pilosus IFO 4520 (Kono and Himeno, 1999), are also reported to have demonstrated antimicrobial activity. The pigments from the mycelium of M. purpureus display significant antimicrobial activity against B. subtilis and Candida pseudotropicalis (Martinkova et al., 1999). The antimicrobial compounds have been identified as rubropunctatin and monascorubrin.

7. Macrophage-stimulating activity

The pigments, monascin and ankaflavin, from the mycelium of M. purpureus have immunosuppressive activity on mouse T-splenocytes (Martinkova et al., 1999). Yu et al. (2005) developed a liquid medium for the production of an

148

Tseng-Hsing Wang and Tzann-Feng Lin

exobiopolymer with macrophage-stimulating activity in a submerged culture of M. pilosus. The highest amount of the exobiopolymer (20.1 mg/ml) was obtained on the fourth day of cultivation in the optimal medium having the following composition (g/liter): 20 g of rice bran, 5 g of peptone, and 1 g of KH2PO4. The optimal culture pH and temperature for mycelial growth and exobiopolymer production was pH 5.0 and 25
C, respectively. The exobiopolymer, a crude polysaccharide fraction, mainly contained neutral sugar (81.8%) with a considerable amount of uronic acid (18.2%).

V. SAFETY Although MRP is now used as a natural colorant and a dietary supplement all over the world, the discovery of citrinin in MRP has led to a controversy about the safety. Citrinin is a fungal metabolite known since 1931, when it was isolated from P. citrinum by Hetherington and Raistrick (1931). Citrinin has been associated with yellow rice disease in Japan (Saito et al., 1971). It has also been implicated as a contributor to porcine nephropathy. Citrinin acts as a nephrotoxin in all animal species tested, but its acute toxicity varies in different species (Carlton and Tuite, 1977). Citrinin was characterized as an antibacterial compound (Betina, 1984). Citrinin was tested for activities against bacteriophages, sarcomas, protozoa, animal cells, and plant cells (Betina, 1984). Citrinin was identified in over a dozen species of Penicillium (e.g., P. camemberti), several species of Aspergillus (e.g., A. terreus, A. niveus, and A. oryzae), and Monascus spp. (Bennett and Klich, 2003). Blanc et al. (1995) characterized the antimicrobial compound, monascidin A, from Monascus as citrinin using qualitative methods, mass spectra, and NMR from M. purpureus and M. ruber. With an acidic character, citrinin is practically insoluble in water. It is soluble in hot alcohol, dioxane, and other nonpolar solvents. Due to its conjugated double bonds, citrinin absorbs light in the visible wavelength range. Its color varies from lemon yellow at pH 4.6 to cherry red at pH 9.9. Its absorption maxima are in the UV range: 250–331 nm. It has a melting point of 175
C and molecular mass 250.25 g/mol. The 50% lethal dose (LC50) is 57 mg/kg (body weight) for ducks, 95 mg/kg for chickens, and 134 mg/kg for rabbits (Hanika and Carlton, 1994). Citrinin can act synergistically with ochratoxin A to suppress RNA synthesis in murine kidney (Sansing et al., 1976). The cytotoxic effects of citrinin have been extensively studied by Liu et al. (2005). Using human embryonic kidney cells (HEK293) as a cellular model, the concentrations causing 50% cell death by the lipophilic extracts of Monascus were in the range of 1.8–4.7 mg/ml. The aqueous extract showed a lower cytotoxicity. Incubation of HEK293 cells with 60-mM pure citrinin for 72 hours caused cell viability to fall to 50% of control levels. In addition, coadministration of pure citrinin and

Monascus Rice Products

149

the lipid extracts from Monascus samples significantly enhanced citrinin cytotoxicity for HEK293 cells using the MTT assay. For analysis of citrinin in Monascus, several methods were developed including thin layer chromatography (TLC), RP-HPLC, and immunoassays (Chu, 1991). A highly sensitive determination (1 ppb) of citrinin in Monascus by GC-selected ion-monitoring mass spectrometry was developed by Shu and Lin (2002). Pisareva et al. (2005) chose 16 Monascus strains to monitor the biosynthesis of citrinin and pigments quantitatively. The results showed that the formation of citrinin appeared to be strain-specific and did not correlate with the pigment formation. In China, researchers from the Institute for Nutrition and Food Safety, Chinese Center for Disease Control and Prevention, have also screened 35 Monascus strains used in the food industry to investigate the effect of cultivation conditions and the medium composition on citrinin production (Li et al., 2003). The results indicated that all strains produced citrinin during fermentation on rice with the levels ranging from 0.28 to 2458.80 mg/kg (201.60 mg/kg for the average and 61.99 mg/kg for the median, respectively). The citrinin level resulting from fermentation on rice was higher than in a liquid medium. The survey reported by Li et al. (2005a) showed that 68 of 114 (59.65%) samples from either solid or liquid phases and collected from either markets or food processing facilities were positive for citrinin with the levels between 0.18 and 1739.23 mg/kg. In Taiwan, HPLC was used to analyze citrinin levels in commercialized MRPs including capsule, colorant, and daily MRP food products (Hsieh and Pan, 2002). The results showed that the amount of citrinin ranges from undetectable to 122.09 mg/kg in the MRPs. However, citrinin is very low or undetectable in the daily food products with Monascus additives, such as bread, salad, Monascus sauce, fermented glutinous rice, Chinese cheese, and wines. Liu et al. (2005) showed that citrinin was detected in lipid extracts of all examined commercialized Monascus products at concentrations ranging between 0.28 and 6.29 mg/kg, but was not found in aqueous extracts. Thus, domestic MRPs may be contaminated by citrinin, which would result in consumer exposure to this mycotoxin. Since citrinin is a mycotoxin and possesses nephrotoxic and hepatotoxic effects, it has a negative impact on the acceptance of red mold rice by consumers. Studies on MRP with a high concentration of monacolin K and a low concentration of citrinin have been conducted by several laboratories (Chen and Hu, 2005; Wang et al., 2004). Since citrinin possesses antibacterial activity for B. subtilis, a simple and quick selection method for mutant strains with low citrinin production was developed based on the formation of an inhibition zone around the colony of the Monascus strain (Wang et al., 2004). Their results showed that mutant strain, M. purpureus N 301, only produced 0.23 mg/kg citrinin, which was 50% less than that of the parent strain. Chen and Hu (2005) obtained

150

Tseng-Hsing Wang and Tzann-Feng Lin

a mutant strain of M. pilosus by treating a wild strain collected in China with mutagenic agents. At the optimum conditions, the concentrations of monacolin K and citrinin were 2.52 mg/g and 0.13 ng/g, respectively. In exploring the effects of a nanoparticulate dispersion of MRP after wet-milling technology treatment, Yu et al. (2006) have shown that monacolin K was reduced to 50–92% of its base level and citrinin was reduced to 48–74% of its base level. Further experimentation will be needed to evaluate the safety and verify the functionality of this nanoparticulate dispersion. The polyketide pathway is a major route for the formation of secondary metabolites including various mycotoxins (including citrinin) in filamentous fungi (e.g., Monascus) (Chandle et al., 1992). The incorporation of 13 C isotope into citrinin from M. ruber incubated with 13C acetate revealed that citrinin is biogenetically originated from a tetraketide, instead of a pentaketide, as has been shown for Penicillium and Aspergillus spp. The production of polyketide red pigments and citrinin by M. ruber may therefore be regulated at the level of the tetraketide branch point (Hajjaj et al., 1999a). Hajjaj et al. (2000) investigated the effects of medium- and long-chain fatty acids on pigment and citrinin production. Their results show that the synthesis of pigments is barely affected whereas the production of citrinin is strongly inhibited, likely by a hydrogen peroxide-mediated degradation of the toxin due to fatty acid-induced peroxisome proliferation. A full-length polyketide synthase (PKS) gene (pksCT) of 7838 bp from M. purpureus has been cloned. It encodes a 2593-amino acid protein that contains putative domains for ketosynthase, acyltransferase, acyl carrier protein (ACP), and a rare methyltransferase (Shimizu et al., 2005). Using a truncated disruption construct resulted in a pksCT-disrupted strain of M. purpureus. Shimizu et al. (2005) have shown that the disruptant does not produce citrinin, but a pksCT revertant generated by successive endogenous recombination events in the pksCT disruptant restores citrinin production. These observations indicate that pksCT encoding the PKS is responsible for citrinin biosynthesis in M. purpureus. Subsequently, these investigators used the gateway system to facilitate the introduction of 7.8 kbp DNA fragments into M. purpureus (Shimizu et al., 2006). The transformants showed 1.5-fold higher production of citrinin than the wild-type strain. Wild et al. (2003) detected two compounds with identical UV absorption at 306–307 nm wavelength. They were purified by HPLC and structurally elucidated by mass spectrometry and NMR spectroscopy. Among them, monascopyridine A contains a g-lactone, propenyl group, hexanoyl side chain, and a pyridine ring. The more lipophilic monascopyridine B is a higher homologue of monascopyridine A with a more lipophilic octanoyl instead of a hexanoyl side chain. The toxicological properties

Monascus Rice Products

151

of monascopyridine A and B, which were the dehydrogenated derivatives of the corresponding red pigments rubropunctamine and monascorubramine, were studied using immortalized human kidney epithelial cells (Knecht and Humpf, 2006; Knecht et al., 2006). The results show that these two compounds are cytotoxic in the micromolar range with median effective concentration values between 20.7 and 43.2 mM. These effects indicate an aneuploidic potential and that monascopyridines may contribute to tumor formation. Several cases of anaphylaxis or asthma may have been caused by MRP (Hipler et al., 2000; Vandenplas et al., 2000; Wigger-Alberti et al., 1999). Hipler et al. (2002) reported the first case of M. purpureus as a possible allergic agent by means of a prick-to-prick test, ‘‘cellular antigen stimulation test’’ (CAST), and different immunoblots. In their study, a 26-year-old butcher experienced a severe anaphylactic reaction with sneezing, rhinitis, conjunctivitis, generalized pruritus, followed by widespread urticaria, Quincke’s edema, and dyspnea after starting to prepare sausages containing MRP.

ACKNOWLEDGMENTS Research in our laboratory was supported by Taiwan Tobacco & Liquor Corp., Taipei, Taiwan, Republic of China. The animal study of the product, Unchole, was performed under the supervision of Professor Ming-Shi Shiao, Department of Life Science, Chang Gung University, Taiwan. We also thank Professor. Ming-Shi Shiao for reviewing the chapter.

REFERENCES Ahn, J., Jung, J., Hyung, W., Haam, S., and Shin, C. 2006. Enhancement of Monascus pigment production by the culture of Monascus sp. J101 at low temperature. Biotechnol. Prog. 22(1), 338–340. Alberts, A.W., Chen, J., Kuron, G., Hunt, V., Huff, J., Hoffman, C., Rothrock, J., Lopez, M., Joshua, H., Harris, E., Patchett, A., Monaghan, R., et al. 1980. Mevinolin: A highly potent competitive inhibitor of hydroxymethylglutaryl-coenzyme A reductase and a cholesterol-lowering agent. Proc. Natl. Acad. Sci. USA 77, 3957–3961. Aniya, Y., Yokomakura, T., Yonamine, M., Nagamine, T., and Nakanishi, H. 1998. Protective effect of the mold Monascus anka against acetaminophen-induced liver toxicity in rats. Jpn. J. Pharmacol. 78(1), 79–82. Aniya, Y., Yokomakura, T., Yonamine, M., Shimada, K., Nagamine, T., Shimabukuro, M., and Gibo, H. 1999. Screening of antioxidant action of various molds and protection of Monascus anka against experimentally induced liver injuries of rats. Gen. Pharmacol. 32(2), 225–231. Aniya, Y., Ohtani, I. I., Higa, T., Miyagi, C., Gibo, H., Shimabukuro, M., Nakanishi, H., and Taira, J. 2000. Dimerumic acid as an antioxidant of the mold, Monascus anka. Free Radic. Biol. Med. 28(6), 999–1004. Araki, J., Yamaguchi, F., and Nakadai, T. 1998. Application of Monascus sp. For preservation of less salt soy sauce. J. Jpn. Soy Sauce Res. Inst. 24, 127–132 (in Japanese). Barker, B. T. P. 1903. The morphology and development of the ascocarp in Monascus. Ann. Bot. 17, 167–236.

152

Tseng-Hsing Wang and Tzann-Feng Lin

Bennett, J. W. and Klich, M. 2003. Mycotoxins. Clin. Microbiol. Rev. 16(3), 497–516. Bennett, J. W., Rubin, P. L., Lee, L. S., and Chen, P. N. 1979. Influence of trace elements and nitrogen sources on versicolorin production by a mutant strain of Aspergillus parasiticus. Mycopathologia 69, 161–166. Berg, J. M., Tymoczko, J. L., and Stryer, L. 2006. Protein turnover: Amino acid degradation and the urea cycle. In ‘‘Biochemistry,’’ 6th Ed., Chapter 23. WH. Freeman and Company, New York. Betina, V. 1984. Mycotoxins—Production, Isolation, Separation and Purification. In ‘‘Developments in Food Science’’ (V. Betina, Ed.), Vol. 8, pp. 217–236. Elsevier, Amsterdam. Bhatnagar, R. K., Ahmad, S., Mukerji, K. G., and Venkitasubramanian, T. A. 1986. Nitrogen metabolism in Aspergillus parasiticus and Aspergillus flavus in relation to aflatoxin production. J. Appl. Bacteriol. 60, 203–211. Bianchi, A. 2005. Extracts of Monascusus purpureus beyond statins—profile of efficacy and safety of the use of extracts of Monascus purpureus. Chin. J. Integr. Med. 11(4), 309–313. Blanc, P. J., Laussac, J. P., Le Bars, J., Le Bars, P., Loret, M. O., Pareilleux, A., Prome, D., Prome, J. C., Santerre, A. L., and Goma, G. 1995. Characterization of monascidin A from Monascus as citrinin. Int. J. Food Microbiol. 27(2–3), 201–213. Brown, A. G., Smale, T. C., King, T. J., Hasenkamp, R., and Thompson, R. H. 1976. Crystal and molecular structure of compactin, a new antifungal metabolite from Penicillium brevicompactum. J. Chem. Soc. Perkin I 11, 1165–1170. Brown, M. S. and Goldstein, J. L. 1986. A receptor-mediated pathway for cholesterol homeostasis. Science 232, 34–47. Buchanan, R. L. and Lewis, D. F. 1984. Regulation of aflatoxin biosynthesis: Effect of glucose on the activities of various glycolytic enzymes. Appl. Environ. Microbiol. 48, 306–310. Buchanan, R. L., Federowicz, D., and Stahl, H. G. 1985. Activities of tricarboxylic acid cycle enzymes in an aflatoxigenic strain of Aspergillus parasiticus after a peptone to glucose carbon source shift. Trans. Br. Mycol. Soc. 84, 267–275. Buckland, B., Gbewonyo, T., Hallada, T., Kaplan, L., and Masurekar, P. 1989. Production of lovastatin, an inhibitor of cholesterol accumulation in humans. In ‘‘Novel Microbial Products for Medicine and Agriculture’’ (A. L. Demain, G. A. Somkuti, J. C. HunterCevera, and H. W. Rossmore, eds.), pp. 161–169. Elsevier, Amsterdam. Campoy, S., Perez, F., Martin, J. F., Gutierrez, S., and Liras, P. 2003. Stable transformants of the azaphilone pigment-producing Monascus purpureus obtained by protoplast transformation and Agrobacterium-mediated DNA transfer. Curr. Genet. 43(6), 447–452. Campoy, S., Rumbero, A., Martin, J. F., and Liras, P. 2005. Characterization of an hyperpigmenting mutant of Monascus purpureus IB1: Identification of two novel pigment chemical structures. Appl. Microbiol. Biotechnol. 7, 1–9. Carels, M. and Shepherd, D. 1975. Sexal reproductive cycle of Monascus in submerged shaken culture. J. Bacteriol. 122(1), 288–294. Carels, M. and Shepherd, D. 1977. The effect of different nitrogen sources on pigment production and sporulation of Monascus species in submerged, shaken culture. Can. J. Microbiol. 23(10), 1360–1372. Carels, M. and Shepherd, D. 1978. The effect of pH and amino acids on conidiation and pigment production of Monascus major ATCC 16362 and Monascus rubiginosus ATCC 16367 in submerged shaken culture. Can. J. Microbiol. 24(11), 1346–1357. Carels, M. and Shepherd, D. 1979. The effect of changes in pH on phosphate and potassium uptake by Monascus rubiginosus ATCC 16367 in submerged shaken culture. Can. J. Microbiol. 25(12), 1484–1488. Chandle, M., McIntyre, C. R., and Simpson, T. J. 1992. Biosynthesis of LL-D253a, a polyketide chromanone metabolite of Phoma pigmentivora: Incorporation of 13C, 2H and 18O labelled precursors. J. Chem. Soc. Perkin Trans. I 1992, 2285–2293.

Monascus Rice Products

153

Chang, J. C., Wu, M. C., Liu, I. M., and Cheng, J. T. 2006. Plasma glucose-lowering action of Hon-Chi in Streptozotocin-induced diabetic rats. Horm. Metab. Res. 38(2), 76–81. Chang, Y. N., Lin, Y. C., Lee, C. C., Liu, B. L., and Tzeng, Y. M. 2002. Effect of rice—glycerol complex medium on the production of Lovastatin by Monascus rubber. Folia Microbiol. (Praha) 47(6), 677–684. Chen, C. C. and Liu, I. M. 2006. Release of acetylcholine by Hon-Chi to raise insulin secretion in Wistar rats. Neurosci. Lett. 404(1–2), 117–121. Chen, F. and Hu, X. 2005. Study on red fermented rice with high concentration of monacolin K and low concentration of citrinin. Int. J. Food Microbiol. 103(3), 331–337. Chu, F. S. 1991. Immunoassays for trace chemical analysis: Monitoring toxic chemicals in humans, food, and the environment. In ‘‘Current Immunochemical Methods for Mycotoxin Analysis’’ (M. Vanderlaan, L. H. Stanker, B. E. Watkins, and D. W. Roberts, eds.), pp. 140–157. American Chemical Society, Washington, DC. Chung, H. Y., Ma, W. C., Kim, J. S., and Chen, F. 2004. Odor-active headspace components in fermented red rice in the presence of a Monascus species. J. Agric. Food Chem. 52(21), 6557–6563. Cicero, A. F., Brancaleoni, M., Laghi, L., Donati, F., and Mino, M. 2005. Antihyperlipidaemic effect of a Monascus purpureus brand dietary supplement on a large sample of subjects at low risk for cardiovascular disease: A pilot study. Complement. Ther. Med. 13(4), 273–278. Demain, A. L. 1968. Regulatory mechanisms and the industrial production of microbial metabolites. Lloydia 31, 395–418. Demain, A. L. 1986. Regulation of secondary metabolism in fungi. Pure Appl. Chem. 58, 219–226. Du, B. M., Lu, Z. L., Chen, Z., Wu, Y. F., Zhao, W. D., and Huang, T. G. 2006. China coronary secondary prevention study: Analysis of patients with different myocardial infarction history. Zhonghua Nei Ke Za Zhi. 45(1), 21–24 (in Chinese). Dutton, M. F. 1988. Enzymes and aflatoxin biosynthesis. Microbiol. Rev. 52, 274–295. Endo, A. 1979. Monacolin K, a new hypocholesterolemic agent produced by a Monascus species. J. Antibiot. 32, 852–854. Endo, A. 1980. Monacolin K, a new hypocholesterolemic agent that specifically inhibits 3-hydroxy-3-methylglutaryl coenzyme A reductase. J. Antibiot. 33, 334–336. Endo, A., Kuroda, M., and Tsujita, Y. 1976. ML-236A, ML-236B, and ML-236C, new inhibitors of cholesterogenesis produced by Penicullium citrinum. J. Antibiot. 29, 1346–1348. Endo, A., Hasumi, K., Nakamura, T., Kunishima, M., and Masuda, M. 1985a. Dihydromonacolin L and monacolin X, new metabolites which inhibit cholesterol biosynthesis. J. Antibiot. (Tokyo) 38(3), 321–327. Endo, A., Hasumi, K., and Negishi, S. 1985b. Monacolins J and L, new inhibitors of cholesterol biosynthesis produced by Monascus rubber. J. Antibiot. (Tokyo) 38(3), 420–422. Fenice, M., Federici, F., Selbmann, L., and Petruccioli, M. 2000. Repeated-batch production of pigments by immobilised Monascus purpureus. J. Biotechnol. 80(3), 271–276. Francis, F. J. 1987. Lesser-known food colorants. Food Technol. 41, 62–68. Greenspan, M. D. and Yudkovitz, J. B. 1985. Mevinolinic acid biosynthesis by Aspergillus terreus and its relationship to fatty acid biosynthesis. J. Bacteriol. 162, 704–707. Hadfield, J. R., Holker, J. S. E., and Stanway, D. N. 1967. The biosynthesis of fungal metabolites. Part II. The b-oxo-lactone equivalents in rubropunctatin and monascorubrin. J. Chem. Soc. Sect. C 10, 751–755. Hajjaj, H., Blanc, P. J., Groussac, E., Goma, G., Uribelarrea, J. L., and Loubiere, P. 1999a. Improvement of red pigment/citrinin production ratio as a function of environmental conditions by Monascus rubber. Biotechnol. Bioeng. 64(4), 497–501.

154

Tseng-Hsing Wang and Tzann-Feng Lin

Hajjaj, H., Klaebe, A., Goma, G., Blanc, P. J., Barbier, E., and Francois, J. 2000. Medium-chain fatty acids affect citrinin production in the filamentous fungus Monascus rubber. Appl. Environ. Microbiol. 66(3), 1120–1125. Hajjaj, H., Klaebe, A., Loret, M. O., Goma, G., Blanc, P. J., and Francois, J. 1999b. Biosynthetic pathway of citrinin in the filamentous fungus Monascus rubber as revealed by 13C nuclear magnetic resonance. Appl. Environ. Microbiol. 65(1), 311–314. Hajjaj, H., Niederberger, P., and Duboc, P. 2001. Lovastatin biosynthesis by Aspergillus terreus in a chemically defined medium. Appl. Environ. Microbiol. 67(6), 2596–2602. Hanika, C. and Carlton, W. W. 1994. Toxicology and pathology of citrinin. In ‘‘Biodeterioriation Research’’ (G. C. Llewellyn, W. V. Dashek, and C. E. O’Rear, eds.), Vol. 4, pp. 41–63. Plenum Press, New York. Harris, W. S. 1997. n-3 Fatty acids and serum lipoproteins: Animal studies. Am. J. Clin. Nutr. 65(Suppl.), 1611S–1616S. Hawksworth, D. L. and Pitt, J. I. 1983. A new taxonomy for Monascus based on cultural and microscopical characters. Aust. J. Bot. 31, 51–61. Heber, D., Yip, Y., Ashley, J. M., Elashoff, D. A., Elashoff, R. M., and Go, V. L. W. 1999. Cholesterol-lowering effects of a proprietary Chinese red-yeast rice dietary supplement. Am. J. Clin. Nutr. 69, 231–236. Heber, D., Lembertas, A., Lu, Q. Y., Bowerman, S., and Go, V. L. 2001. An analysis of nine proprietary Chinese red yeast rice dietary supplements: Implications of variability in chemical profile and contents. J. Altern. Complement. Med. 7(2), 133–139. Hendrickson, L., Davis, C. R., Roach, C. D., Nguyen, K., Aldrich McAda, T. P. C., and Reeves, C. D. 1999. Lovastatin biosynthesis in Aspergillus terreus: Characterization of blocked mutants, enzyme activities and a multifunctional polyketide synthase gene. Chem. Biol. 6, 429–439. Hesseltine, C. W. 1965. A millennium of fungi and fermentation. Mycologia 57, 149–197. Hetherington, A. C. and Raistrick, H. 1931. Studies in the biochemistry of microorganisms. Part XIV. On the production and chemical constitution of a new yellow colouring matter, citrinin, produced from glucose by Penicillium citrinum Thom. Phil. Trans. R. Soc. London Ser. B 220B, 269–295. Himeno, K. 1997. Function of beni-koji and its utilization for food. Shokuhinkogyo 40, 49–55. (in Japanese). Hipler, U. C., Wigger-Alberti, W., Bauer, A., and Elsner, P. 2000. Monascus purpureus: A new fungus of allergenic relevance. Mycoses 43(Suppl. 2), 29–31. Hipler, U. C., Wigger-Alberti, W., Bauer, A., and Elsner, P. 2002. Case report. Monascus purpureus—a new fungus of allergologic relevance. Mycoses 45(1–2), 58–60. Hsieh, P. S. and Tai, Y. H. 2003. Aqueous extract of Monascus purpureus M9011 prevents and reverses fructose-induced hypertension in rats. J. Agric. Food Chem. 51(14), 3945–3950. Hsieh, Y. T. and Pan, T. M. 2002. The analytic methods and assays for secondary metabolites of Monascus products. J. Biomass Energy Soc. China 21(1–2), 63–71 (in Chinese). Hsu, F. L., Wang, P. M., Lu, S. Y., and Wu, W. T. 2002. A combined solid-state and submerged cultivation integrated with adsorptive product extraction for production of Monascus red pigments. Bioprocess Biosyst. Eng. 25(3), 165–168. Iizuka, H. and Lin, C. F. 1980. On the genus Monascus of Asia and its specific characteristics. In ‘‘Advances in Biotechnology’’ (M. Moo-Young, C. W. Robison, and C. Vezina, eds.), Vol. 2, pp. 555–561. Pergamon Press, Toronto. Jeon, T., Hwang, S. G., Hirai, S., Matsui, T., Yano, H., Kawada, T., Lim, B. O., and Park, D. K. 2004. Red yeast rice extracts suppress adipogenesis by down-regulating adipogenic transcription factors and gene expression in 3T3-L1 cells. Life Sci. 75(26), 3195–3203. Jing, J. B., Hao, X. Y., and Deng, C. H. 1999. The effects of Xuezhikang on serum lipid profile, thrombozane A2 and prostacyclin in patients with hyperlipidemia. Chinese J. Int. Med. 38(8), 517–519 (in Chinese).

Monascus Rice Products

155

Johnson, G. T. and McHan, F. 1975. Some effect of zinc on the utilization of carbon sources by Monascus purpureus. Mycologia 67(4), 806–816. Journoud, M. and Jones, P. J. H. 2004. Red yeast rice: A new hypolipidemic drug. Life Sci. 74, 2675–2683. Jung, H., Kim, C., Kim, K., and Shin, C. S. 2003. Color characteristics of Monascus pigments derived by fermentation with various amino acids. J. Agric. Food Chem. 51(5), 1302–1306. Jung, H., Kim, C., and Shin, C. S. 2005. Enhanced photostability of monascus pigments derived with various amino acids via fermentation. J. Agric. Food Chem. 53(18), 7108–7114. Keisuke, T., Tomio, I., Nobukazu, T., Hiroshi, O., Shirou, A., Shouichi, T., and Yasue, N. 1992. Extraction of hypotensive substance from wheat beni-koji. Nippon Shokuhin Kogyo Gakkaishi 39, 913–918. Kennedy, J., Auclair, K., Kendrew, S. G., Park, C., Vederas, J. C., and Hutchinson, C. R. 1999. Modulation of polyketide synthase activity by accessory proteins during lovastatin biosynthesis. Science 284, 1368–1372. Knecht, A. and Humpf, H. U. 2006. Cytotoxic and antimitotic effects of N-containing Monascus metabolites studied using immortalized human kidney epithelial cells. Mol. Nutr. Food Res. 50(4–5), 406–412. Knecht, A., Cramer, B., and Humpf, H. U. 2006. New Monascus metabolites: Structure elucidation and toxicological properties studied with immortalized human kidney epithelial cells. Mol. Nutr. Food Res. 50(3), 314–321. Kono, I. and Himeno, K. 1999. Antimicrobial activity of Monascus pilosus IFO 4520 against contaminant of Koji. Biosci. Biotechnol. Biochem. 63(8), 1494–1496. Kono, I. and Himeno, K. 2000. Changes in g-aminobutyric acid content during beni-koji making. Biosci. Biotechnol. Biochem. 64(3), 617–619. Kou, W., Lu, Z., and Guo, J. 1997. Effect of xuezhikang on the treatment of primary hyperlipidemia. Chinese J. Int.l Med. 36(8), 529–531 (in Chinese). Krairak, S., Yamamura, K., Irie, R., Nakajima, M., Shimizu, H., Chim-Anage, P., Yongsmith, B., and Shioya, S. 2000. Maximizing yellow pigment production in fedbatch culture of Monascus sp. J. Biosci. Bioeng. 90(4), 363–367. Kurono, M., Nakanishi, K., Shindo, K., and Tada, M. 1963. Biosynthesis of monascorubrin and monascoflavin. Chem. Pharm. Bull. 11, 359–362. Lee, C. L., Tsai, T. Y., Wang, J. J., and Pan, T. M. 2005. In vivo hypolipidemic effects and safety of low dosage Monascus powder in a hamster model of hyperlipidemia. Appl. Microbiol. Biotechnol. 11, 1–8. Lee, C. L., Wang, J. J., Kuo, S. L., and Pan, T. M. 2006. Monascus fermentation of dioscorea for increasing the production of cholesterol-lowering agent-monacolin K and antiinflammation agent-monascin. Appl. Microbiol. Biotechnol. 72(6), 1254–1262. Li, F., Xu, G., Li, Y., and Chen, Y. 2003. Study on the production of citrinin by Monascus strains used in food industry. Wei Sheng Yan Jiu 32(6), 602–605 (in Chinese). Li, F. Q., Xu, G. R., Li, Y. W., Chen, Y., and Ji, R. 2005a. Natural occurrence of citrinin in Monascus products. Wei Sheng Yan Jiu 34(4), 451–454 (in Chinese). Li, J. J., Hu, S. S., Fang, C. H., Hui, R. T., Miao, L. F., Yang, Y. J., and Gao, R. L. 2005b. Effects of xuezhikang, an extract of cholestin, on lipid profile and C-reactive protein: A short-term time course study in patients with stable angina. Clin. Chim. Acta 352(1–2), 217–224. Li, Y. G., Liu, H., and Wang, Z. T. 2005c. A validated stability-indicating HPLC with photodiode array detector (PDA) method for the stress tests of Monascus purpureusfermented rice, red yeast rice. J. Pharm. Biomed. Anal. 39(1–2), 82–90. Li, Y. G., Zhang, F., Wang, Z. T., and Hu, Z. B. 2004. Identification and chemical profiling of monacolins in red yeast rice using high-performance liquid chromatography with photodiode array detector and mass spectrometry. J. Pharm. Biomed. Anal. 35(5), 1101–1112.

156

Tseng-Hsing Wang and Tzann-Feng Lin

Lin, C. C., Li, T. C., and Lai, M. M. 2005a. Efficacy and safety of Monascus purpureus Went rice in subjects with hyperlipidemia. Eur. J. Endocrinol. 153(5), 679–686. Lin, C. F. 1973. Isolation and cultural conditions of Monascus sp. for the production of pigment in a submerged culture. J. Ferment. Technol. 51, 407–414. Lin, C. F. 1975. Studies on the Monascus isolated from the starter of kaoliang brandy. Chinese J. Microbiol. 8, 152–160 (in Chinese). Lin, C. F. and Iizuka, H. 1982. Production of extracellular pigment by a mutant of Monascus kaoliang sp. nov. Appl. Environ. Microbiol. 43(3), 671–676. Lin, T. F. 1982. Investigation and improvement on the process of making anka. Ann. Rept. Res. Inst. Wines 151–155 (in Chinese). Lin, T. F. 1991. Studies on the formation of Monascus red pigments. Ph. D. Thesis, Massachusetts Institute of Technology, USA. Lin, T. F., Chen, Y. L., and Wang, T. H. 2006a. Method for Producing Beer-like and Alcoholfree Fermented Beverage. UK patent GB2,414,486. Lin, T. F., Yu, H. C., Huang, C. T., Wang, H. C., Jen, W. P., and Shiao, M. S. 2005b. Development of Monascus Products as Functional Foods: Fermentation, Quality Control, and Health Beneficial Effects The 230th ACS National Meeting, in Washington, DC, Aug 28–Sept 1. Lin, W. Y., Song, C. Y., and Pan, T. M. 2006b. Proteomic analysis of Caco-2 cells treated with monacolin K. J. Agric. Food Chem. 54(17), 6192–6200. Liu, B. H., Wu, T. S., Su, M. C., Chung, C. P., and Yu, F. Y. 2005. Evaluation of citrinin occurrence and cytotoxicity in Monascus fermentation products. J. Agric. Food Chem. 53(1), 170–175. Liu, L., Zhao, S. P., Cheng, Y. C., and Li, Y. L. 2003. Xuezhikang decreases serum lipoprotein (a) and C-reactive protein concentrations in patients with coronary heart disease. Clin. Chem. 49(8), 1347–1352. Lotong, N. 1985. Koji. In ‘‘Microbiology of Fermented Food’’ (B. J. B. Wood, ed.), Vol. 2, pp. 252–254. Elsevier, New York. Lu, Z. L. and Fu, W. 2005. China coronary secondary prevention study (CCSPS). Zhonghua Xin Xue Guan Bing Za Zhi 33(2), 109–115 (in Chinese). Lu, Z. L., Du, B. M., Chen, Z., Wu, Y. F., Yu, X. H., and Zhao, Y. C. 2005. China coronary secondary prevention study (CCSPS): Outcomes from analysis of coronary heart disease patients with diabetes. Zhonghua Xin Xue Guan Bing Za Zhi 33(12), 1067–1070 (in Chinese). Ma, J., Li, Y., Ye, Q., Li, J., Hua, Y., Ju, D., Zhang, D., Copper, R., and Chang, M. 2000. Constituents of red yeast rice, a traditional Chinese food and medicine. J. Agric. Food Chem. 48, 5220–5225. Martinkova, L., Patakova-Juzlova, P., Krent, V., Kucerova, Z., Havlicek, V., Olsovsky, P., Hovorka, O., Rihova, B., Vesely, D., Vesela, D., Ulrichova, J., and Prikrylova, V. 1999. Biological activities of oligoketide pigments of Monascus purpureus. Food Addit. Contam. 16(1), 15–24. Matheson, G. K., Freed, E., and Tunnicliff, G. 1986. Novel GABA analogues as hypotensive agent. Neuropharmacology 25, 1191–1195. Matsumoto, N., Mori, A., Imai, C., Ogama, H., Murakawa, S., and Endo, A. 1989. Antibacterial substance produced by Monascus paxii (Abstr.). Annual Meeting of Soc. Ferment. Technol., Japanp 69. Miyake, T., Mori, A., Kii, T., Okuno, T., Usui, Y., Sato, F., Sammoto, H., Watanabe, A., and Kariyama, M. 2005. Light effects on cell development and secondary metabolism in Monascus. J. Ind. Microbiol. Biotechnol. 32(3), 103–108. Miyake, T., Uchitomi, K., Zhang, M. Y., Kono, I., Nozaki, N., Sammoto, H., and Inagaki, K. 2006. Effects of the principal nutrients on lovastatin production by Monascus pilosus. Biosci. Biotechnol. Biochem. 70(5), 1154–1159. Moll, H. R. and Farr, D. R. 1976. Red pigment and process. United States Patent 3,993.

Monascus Rice Products

157

Ntanios, F. Y. and Jones, P. J. H. 1999. Dietary sitosterol reciprocally influences cholesterol absorption and biosynthesis in hamsters and rabbits. Atherosclerosis 143, 341–351. Ostlund, R. E. 2002. Phytosterols in human nutrition. Annu. Rev. Nutr. 22, 533–549. Park, H. G. and Jong, S.-C. 2003. Molecular characterization of Monascus strains based on the D1/D2 regions of LSU rRNA genes. Mycoscience 44, 25–32. Park, H. G., Stamenova, E. K., and Jong, S.-C. 2004. Phylogenetic relationships of Monascus species inferred from the ITS and the partial b-tubulin gene. Bot. Bull. Acad. Sin. 45, 325–330. Pisareva, E., Savov, V., and Kujumdzieva, A. 2005. Pigments and citrinin biosynthesis by fungi belonging to genus Monascus. Z. Naturforsch [C] 60(1–2), 116–120. Rasheva, T. V., Nedeva, T. S., Hallet, J. N., and Kujumdzieva, A. V. 2003. Characterization of a non-pigment producing Monascus purpureus mutant strain. Antonie Van Leeuwenhoek 83(4), 333–340. Rhyu, M. R., Kim, D. K., Kim, H. Y., and Kim, B. K. 2000. Nitric oxide-mediated endotheliumdependent relaxation of rat thoracic aorta induced by aqueous extract of red rice fermented with Monascus ruber. J. Ethnopharmacol. 70(1), 29–34. Saito, M., Enomoto, M., and Tatsuno, T. 1971. Yellowed rice toxins: Luteroskyrin and related compounds, chlorine-containing compounds and citrinin. In ‘‘Microbial Toxins’’ (A. Ciegler, S. Kadis, and S. J. Ajl, eds.), Vol. VI: fungal toxins, pp. 299–380. Academic Press, New York. Sansing, G. A., Lillehoj, E. B., and Detroy, R. W. 1976. Synergistic toxic effect of citrinin, ochratoxin A and penicillic acid in mice. Toxicon 14, 213–220. Shepherd, D. and Carels, M. 1983. Product formation and differentiation in fungi. In ‘‘Fungal Differentiation’’ (J. E. Smith, ed.), pp. 515–535. Marcel Dekker, New York. Shimizu, T., Kinoshita, H., Ishihara, S., Sakai, K., Nagai, S., and Nihira, T. 2005. Polyketide synthase gene responsible for citrinin biosynthesis in Monascus purpureus. Appl. Environ. Microbiol. 71(7), 3453–3457. Shimizu, T., Kinoshita, H., and Nihira, T. 2006. Development of transformation system in Monascus purpureus using an autonomous replication vector with aureobasidin a resistance gene. Biotechnol. Lett. 28(2), 115–120. Shin, C. S., Kim, H. J., Kim, M. J., and Ju, J. Y. 1998. Morphological change and enhanced pigment production of monascus when cocultured with Saccharomyces cerevisiae or Aspergillus oryzae. Biotechnol. Bioeng. 59(5), 576–581. Shu, P. Y. and Lin, C. H. 2002. Simple and sensitive determination of citrinin in Monascus by GC-selected ion monitoring mass spectrometry. Anal. Sci. 18, 283–287. Simpson, T. J. 1986. Studies of polyketide chain-assembly processes, mycotoxins and phycotoxins. In ‘‘Mycotoxins and Phycotoxins’’ (P. S. Steyn and R. Vleggaar, eds.), pp. 85–96. Elsevier, Amsterdam, the Netherlands. Spears, K. 1988. Developments in food colorings: The natural alternatives. Trends Biotechnol. 6, 283–288. Steinberg, D. 2002. Atherosclerosis in perspective: Hypercholesterolemia and inflammation as partners in crime. Nat. Med. 8, 1211–1217. Su, N. W., Lin, Y. L., Lee, M. H., and Ho, C. Y. 2005. Ankaflavin from Monascus-fermented red rice exhibits selective cytotoxic effect and induces cell death on Hep G2 cells. J. Agric. Food Chem. 53(6), 1949–1954. Su, Y. C. 1978. The production of Monascus pigments. Food Sci. 5, 4A–17A (in Chinese). Su, Y. C. and Wang, W. H. 1983. Chinese red rice: Anka. In ‘‘Handbook of Indigenous Fermented Foods’’ (K. H. Steinkraus, R. E. Cullen, C. S. Pederson, L. F. Nellis, and B. K. Gavitt, eds.), pp. 547–553. Marcel Dekker, New York. Su, Y. C., Wang, J. J., Lin, T. T., and Pan, T. M. 2003. Production of the secondary metabolites gamma-aminobutyric acid and monacolin K by Monascus. J. Ind. Microbiol. Biotechnol. 30(1), 41–46.

158

Tseng-Hsing Wang and Tzann-Feng Lin

Sumioka, I., Hayama, M., Shimokawa, Y., Shiraishi, S., Tokunaga, A., and Hiroshima, J. 2006. Lipid-lowering effect of Monascus garlic fermented extract (MGFE) in hyperlipidemic subjects. Med. Sci. 55(2), 59–64. Sweeny, J. G., Estrada-Valdes, M. C., Iacobucci, G. A., Sato, H., and Sakamura, S. 1981. Photoprotection of the red pigments of Monascus anka in aqueous media by 1,4, 6-trihydroxynaphthalene. J. Agric. Food Chem. 29, 1189–1193. Taira, J., Miyagi, C., and Aniya, Y. 2002. Dimerumic acid as an antioxidant from the mold, Monascus anka: The inhibition mechanisms against lipid peroxidation and hemeproteinmediated oxidation. Biochem. Pharmacol. 63(5), 1019–1026. Turner, W. B. 1971. ‘‘Fungal Metabolites,’’ pp. 55–223. Academic Press, London. Turner, W. B. and Aldridge, D. C. 1983. ‘‘Fungal Metabolites II,’’ pp. 55–223. Academic Press, New York. van Tieghem, P. 1884. Monascus, genre nouvau de l’ordre des Ascomycetes. Bull. Soc. Bot. France 31, 226–231. Vandenplas, O., Caroyer, J. M., Cangh, F. B., Delwiche, J. P., Symoens, F., and Nolard, N. 2000. Occupational asthma caused by a natural food colorant derived from Monascus rubber. J. Allergy Clin. Immunol. 105(6 Pt. 1), 1241–1242. Wang, J. J. and Pan, T. M. 2003. Effect of red mold rice supplements on serum and egg yolk cholesterol levels of laying hens. J. Agric. Food Chem. 51(16), 4824–4829. Wang, J. J., Lee, C. L., and Pan, T. M. 2003a. Improvement of monacolin K, gammaaminobutyric acid and citrinin production ratio as a function of environmental conditions of Monascus purpureus NTU 601. J. Ind. Microbiol. Biotechnol. 30(11), 669–676. Wang, J. J., Lee, C. L., and Pan, T. M. 2004. Modified mutation method for screening low citrinin-producing strains of Monascus purpureus on rice culture. J. Agric. Food Chem. 52(23), 6977–6982. Wang, J. J., Pan, T. M., Shieh, M. J., and Hsu, C. C. 2006. Effect of red mold rice supplements on serum and meat cholesterol levels of broilers chicken. Appl. Microbiol. Biotechnol. 71(6), 812–818. Wang, T.-H., Lin, H.-S., and Lin, T.-F. 2003b. Method of producing vegetarian lactic acid and non-alcoholic beverages with koji-saccharified high-sugar syrup. United States Patent 6,641,852. Wei, W., Lim, C., Wang, Y., Su, H., Zhu, J., and Kritchevsky, D. 2003. Hypolipidemic and anti-atherogenic effects of long-term Cholestin (Monascus purpureus-fermented rice, red yeast rice) in cholesterol fed rabbits. J. Nutr. Biochem. 14(6), 314–318. Wei, P. L., Zhou, L. P., and Cen, P. L. 2004. Solid-state Fermentation for lovastatin production by Monascus rubber. Food and Fermentation Industries 30(4), 32–35(in Chinese). Weinberg, E. D. 1989. Roles of micronutrients in secondary metabolism of actinomycetes. In ‘‘Regulation of Secondary Metabolism in Actinomycetes’’ (S. Shapiro, ed.), pp. 239–261. CRC Press, Boca Raton, FL. Wigger-Alberti, W., Bauer, A., Hipler, U. C., and Elsner, P. 1999. Anaphylaxis due to Monascus purpureus-fermented rice (red yeast rice). Allergy 54(12), 1330–1331. Wild, D., Toth, G., and Humpf, H. U. 2003. New Monascus metabolites with a pyridine structure in red fermented rice. J. Agric. Food Chem. 51(18), 5493–5496. Wong, H. C. 1982. Antibiotic and pigment production by Monascus purpureus. Ph. D. Thesis, University of Georgia. Wong, H. C. and Bau, Y. S. 1977. Pigmentation and antibacterial activity of fast neutron- and x-ray-induced strains of Monascus purpureus Went. Plant Physiol. 60, 578–581.

Monascus Rice Products

159

Wong, H. C., Lin, Y. C., and Koehler, P. E. 1981. Regulation of growth and pigmentation of Monascus purpureus by carbon and nitrogen concentration. Mycologia 73, 649–654. Wu, W. T., Wang, P. M., Chang, Y. Y., Huang, T. K., and Chien, Y. H. 2000. Suspended rice particles for cultivation of Monascus purpureus in a tower-type bioreactor. Appl. Microbiol. Biotechnol. 53(5), 542–544. Wu, W., Zheng, K., Chen, Y., Xing, F. Q., and Zeng, D. Y. 2003. Effects of xuezhikang on cholesterol and activity substances in vascular endothelial cell. Chin. J. Arterioscler. 11(5), 419–422 (in Chinese). Yoshimura, M., Yamanada, S., Mitsugi, K., and Hirose, Y. 1975. Production of Monascus pigment in a submerged culture. Agric. Biol. Chem. 39, 1789–1795. Yoshizawa, Y., Witter, D. J., Liu, Y., and Vederas, J. C. 1994. Revision of the biosynthetic origin of oxygens in mevinolin (lovastatin), a hypocholesterolemic drug from Aspergillus terreus MF 4845. J. Am. Chem. Soc. 116, 2693–2694. Young, E. M. 1930. Physiological studies in relation to the taxonomy of Monascus species. Trans. Wis. Acad. Sci. Arts and Lett. 25, 227–244. Yu, K. W., Kim, Y. S., Shin, K. S., Kim, J. M., and Suh, H. J. 2005. Macrophage-stimulating activity of exo-biopolymer from cultured rice bran with Monascus pilosus. Appl. Biochem. Biotechnol. 126(1), 35–48. Yu, C. C., Lee, C. L., and Pan, T. M. 2006. A novel formulation approach for preparation of nanoparticulate red mold rice. J. Agric. Food Chem. 54(18), 6845–6851. Zhao, S. P., Liu, L., Cheng, Y. C., Shishehbor, M. H., Liu, M. H., Peng, D. Q., and Li, Y. L. 2004. Xuezhikang, an extract of cholestin, protects endothelial function through antiinflammatory and lipid-lowering mechanisms in patients with coronary heart disease. Circulation 110(8), 915–920. Zheng, G. J., Zhang, W. G., Zhang, Y. T., Zhang, Y. F., and Ma, X. S. 2003. Xuezhikang’s effect in preventing the Arteriosclerosis of apolipoprotein E deficient mice. Chin. J. Arterioscler. 11(5), 408–410 (in Chinese).

This page intentionally left blank

CHAPTER

5 Designer Milk Latha Sabikhi*

Contents

I. Introduction II. Milk ‘‘Designing’’: The Prospects III. Milk Fat Modification A. Altering the fatty acid chain length and level of saturation in milk fat B. Increasing CLA levels in milk fat C. The omega fatty acids D. Reducing fat content in milk E. Type of fatty acids versus product quality IV. Milk Sugar (Lactose) Modification A. Preharvest methods of lactose reduction V. Milk Protein Modification A. Modifying the major milk proteins B. Modifying the minor milk proteins C. Targeting the proteinase-cleavage sites VI. Designer Milk for Infant Health A. Lactoferrin B. Lysozyme C. Cow milk allergy D. Lactose intolerance VII. Milk with Human Therapeutic Proteins VIII. Designer Milk for Animal Growth and Health IX. Assorted Advantages X. The Future References

162 163 165 165 167 170 172 173 174 175 176 176 178 178 179 179 180 181 182 183 187 190 191 193

* Dairy Technology Division, National Dairy Research Institute, Karnal 132001, Haryana, India Advances in Food and Nutrition Research, Volume 53 ISSN 1043-4526, DOI: 10.1016/S1043-4526(07)53005-6

#

2007 Elsevier Inc. All rights reserved.

161

162

Abstract

Latha Sabikhi

Dairy biotechnology is fast gaining ground in the area of altering milk composition for processing and/or animal and human health by employing nutritional and genetic approaches. Modification of the primary structure of casein, alteration in the lipid profile, increased protein recovery, milk containing nutraceuticals, and replacement for infant formula offer several advantages in the area of processing. Less fat in milk, altered fatty acid profiles to include more healthy fatty acids such as CLA and o-fats, improved amino acid profiles, more protein, less lactose, and absence of b-lactoglobulin (b-LG) are some opportunities of ‘‘designing’’ milk for human health benefits. Transgenic technology has also produced farm animals that secrete in their milk, human lactoferrin, lysozyme, and lipase so as to simulate human milk in terms of quality and quantity of these elements that are protective to infants. Cow milk allergenicity in children could be reduced by eliminating the b-LG gene from bovines. Animals that produce milk containing therapeutic agents such as insulin, plasma proteins, drugs, and vaccines for human health have been genetically engineered. In order to cater to animal health, transgenic animals that express in their mammary glands, various components that work against mastitis have been generated. The ultimate acceptability of the ‘‘designer’’ products will depend on ethical issues such as animal welfare and safety, besides better health benefits and increased profitability of products manufactured by the novel techniques.

I. INTRODUCTION Reports of prolific and successful research in the areas of biotechnology and genetic engineering have unleashed potential ideas that were previously inconceivable in the subject of dairying. It is now firmly established that novel value-added products can be derived from milk and milk products with nutritional and biotechnological interventions. While until recently, breeding policies have aimed at producing more milk, attempts are now directed toward enhancing the value of milk and studying its health implications. This has found more support with clinically established epidemiological linkages between diet and chronic diseases that encourage search for new links between food and disease. The extranutritional therapeutic attributes of milk and milk products have also been brought into this broad network of research. Milk composition can be altered by nutritional management or through the manipulation of naturally occurring genetic variation among cattle. The possible

Designer Milk

163

channels of influencing milk composition to suit specific needs can be investigated with the help of a thorough comprehension of the biochemistry, genetic traits, and factors in the animal diet that affect milk synthesis and composition. By an intelligent combination of the two approaches— nutritional and genetic—a milk designed to suit consumer preferences can be developed. This ‘‘designer milk’’ may be rich in specific milk components that may have influence on well-being or on processing. This chapter examines the potential that exists in altering milk composition by nutritional and genetic approaches in order to achieve specific health benefits and/or processing opportunities.

II. MILK ‘‘DESIGNING’’: THE PROSPECTS Man has been taming and manipulating other species for his own benefits for thousands of years. Several breeds of cattle that produce large quantities of milk exist today as a result of selective breeding adopted by farmers over centuries. The global appeal of milk as a healthy beverage that is good for adults as well as infants has prompted much investigation on the commodity. Research on animal breeding, husbandry, and feeding conditions has always had a profound impact on the quality of milk, its constituents, and the subsequently manufactured products. Altering the composition of milk in a manner that suits health and processing needs forms the basis of the current research interests in the area. For example, a greater proportion of unsaturated fatty acids in milk fat, reduced lactose content in milk for lactose-intolerant people, and/or milk free from b-lactoglobulin (b-LG) would benefit human diet and health. From a technological point of view, there exist vast opportunities in altering the primary structure of casein to improve the technological properties of milk and producing milk high in protein content. Engineering milk that clots in less time leads to increased yield and/or more protein recovery during cheese manufacture. Milk that contains nutraceuticals and replacement ingredients for infant formula are other interesting avenues. Genetic manipulation (GM) also offers the prospect of healthier animals with improved resistance to diseases such as mastitis or to the ticks that can infest cattle, thus reducing the need for antibiotics and pesticides. Medicines may be produced in the milk of cows. For example, GM cows could produce milk with a clotting factor for hemophiliacs, milk containing human serum albumin for blood transfusions, or milk with a hepatitis vaccine. Several of these medicines could be produced much more efficiently than with the technologies currently used. Some of the potential changes that can be brought about in milk are listed in Table 1.

164

Latha Sabikhi

TABLE 1 Selected reports on opportunities for ‘‘designing’’ milk No.

Modification

A. Fat modification A.1 Remove/reduce fat

A.2

Alter the fatty acid chain length

A.3

Increase CLA levels in milk

A.4

Alter proportion of o-6 to o-3 fatty acids

Benefits

Referencesa

Low-fat milk and products, caters to the health-conscious consumers Increased nutrition, better manufacturing properties, better product quality Anticarcinogenic and other therapeutic properties Several health benefits

Wall et al., 1997

B. Carbohydrate modification B.1 Overexpress Better lactose b-galactosidase digestibility, caters to enzyme the lactose-intolerant customers B.2 Remove a-LA, Reduced synthesis of produce lactase lactose by transgenic technology C. Protein modification C.1 Increase amino acids content, casein C.2 C.3

Genetically engineer casein Remove b-LG

C.3

Modify bovine milk to simulate human milk

C.4

Introduce human therapeutic proteins

Increased protein, better processing properties, better nutrition Better manufacturing properties Less milk allergies, better processing properties Better infant health, less mortality, less problems due to milk allergy

CSIRO, 1999; O’Donnell, 1993; Mason, 2001 Pszczola et al., 2000; Stanton, 2000 Dhiman et al., 1999; Kao et al., 2006

Bremel et al., 1989

Jost et al., 1999; Karatzas and Turner, 1997

www.agresearch.co. nz, 2001

Bleck et al., 1998a,b; Brophy et al., 2003 www.agresearch.co. nz, 2001 Lonnerdal, 1996; Maga et al., 2006

AgResearch Now, 2005; Morgan, 2006; Pettus, 2006

Designer Milk

165

TABLE 1 (continued) No.

Modification

D. Miscellaneous D.1 Produce in milk antibodies, antimicrobials against pathogens D.2 Produce spider silk in milk a

Benefits

Referencesa

Safer food, prevention of mastitis and other diseases

Margawati, 2003; Wall et al., 2005

Industrial applications

Anonymous, 2002; Dove, 2000; Lazaris et al., 2002

An indicative and partial list.

III. MILK FAT MODIFICATION The quantity of milk fat is a determinant of its value and hence a major indicator of the revenue accrued from milk. As a recent trend, advanced knowledge about the chemistry of milk fat and human physiology encourages product developers to modify the milk fat to counter the changing functional and nutritional challenges. Dairy products provide less than 15% of the total fat available in the diet (O’Donnell, 1993). Milk fat provides 25% of the saturated fat, which is still not as high as that in the two groups of fats and oils (29%) and meat, poultry, and fish (39%). The contribution of dairy products to the total cholesterol is 16%, much less than that in eggs (39%) and meat, poultry, and fish (43%). Modifications of the composition and quality of fodder result in different milk fat compositions and influence the nutritional and technological value of fats. A sophisticated trend in the ‘‘health market’’ today is to modify the milk fat composition by either adopting suitable feeding strategies or by genetic modes. It is now almost possible to achieve the ideal composition of milk fat for human health and well-being recommended by O’Donnell (1989), after the Wisconsin Milk Board 1988 Milk Fat Roundtable. The combination suggested at the meeting was less than 10% polyunsaturated fatty acids (PUFA), less than 8% saturated fatty acids (SFA), and more than 82% monounsaturated fatty acids (MUFA).

A. Altering the fatty acid chain length and level of saturation in milk fat The long-chain fatty acids of milk fat are derived from the diet via blood. The short-chain fatty acids (C10 and below) of milk fat are first synthesized in the mammary gland and then elongated to C12–C16. If the

166

Latha Sabikhi

mechanism for elongation is blocked by genetic technology, the ratio of medium-chain fatty acids (C12–C16) to short-chain fatty acids in milk fat should reduce. Since the C12–C16 fatty acids are generally regarded by nutritionists as less desirable, milk fat with reduced content of mediumlength fatty acid chains would garner more value due to greater consumer demand. There is ample experimental evidence to suggest that nutritional modifications can cause significant changes in milk fat composition. The degree of unsaturation of the serum lipids, tissue fat, and milk fat may be increased promptly by feeding unsaturated fats in an encapsulated or protected form to lactating animals (Ashes et al., 1997). It is established that MUFA (C18:1) content can be increased by 50–80% and may approach 50% of milk fatty acids by feeding lipids rich in 18-carbon fatty acids (Grummer, 1991). Feeding low-roughage diets increases the proportion of MUFA in milk fat, the effects of feeding low-roughage diets and lipid being additive. The SFA content (palmitic acid, C16:0) of milk fat can also be reduced by 20–40% unless the supplemented lipid is rich in palmitic acid. SFA particularly palmitic and other medium-chain fatty acids tend to increase levels of blood cholesterol (O’Donnell, 1993). Feeding highly unsaturated oils (e.g., soybean oil) caused depression in milk fat, but increased the proportion of unsaturated fatty acids to SFA in milk (www.extension.iastate.edu). A study at the University of Alberta (Mason, 2001) revealed that feeding canola oil in the encapsulated form (to protect it from biohydrogenation by the rumen microorganisms) led to higher increases in linoleic (18:2) and linolenic (18:3) acids than while feeding unprotected oil seeds. As the melting point of milk fat containing unsaturated fatty acids is more, the spreadability of butter made from this milk improved tremendously. An Australian study involving the feeding of a special blend of canola and soybean meal in the protected form resulted in doubling the spreadability of butter (CSIRO, 1999). When taken out of a refrigerator at 5
C, the butter was nearly as spreadable as margarine, without losing its special eating qualities. Clinical trials revealed that consumption of dairy products made from this milk led to decrease in lowdensity lipoprotein (LDL) levels in the blood of the consumers. Chouinard et al. (1998) compared the results of feeding to Holstein cows, a control total mixed ration (TMR) with TMR supplemented with calcium salts of three fatty acids from oils with progressive degree of unsaturation— canola oil, soybean oil, or linseed oil. The digestibility of nutrients was higher for rations containing calcium salts than for the control ration. The milk yield increased in proportion to the degree of unsaturation in the feed supplement. The fat content in milk reduced in all the experimental diets as compared to the control. The addition of calcium salts to the ration decreased the proportions of SFA that contained C6–C16 and increased the proportions of C18:0, cis-9-C18:1, and trans-11-C18:1 in milk fat. These

Designer Milk

167

findings were confirmed later by Aigster et al. (2000) who reported that feeding calcium salts of high-oleic sunflower oil (HOSO) containing more than 86% oleic acid at the rate of 7.5% of diet dry matter weight to Holstein cows increased the oleic acid content of milk fat from 26% to over 40% and decreased the cholesterol-raising saturates from 41% to 33%. Reports by Lee et al. (2004) elucidated the feeding of goats with different kinds of protein-oil supplements to alter the milk fat composition. Combinations of HOSO with keratin (KN), casein (CN), and dry casein (DCN) were fed to lactating goats. Oleic acid levels increased to 19.0% on the KN-oil supplement diet, 19.2% on the CN-oil supplement diet, and 25.2% on DCN-oil supplements diet compared to 12.5% in milk fat from goats on normal diet. Feeding the protein-oil supplements also decreased solid content in milk fat at 10
C from 34.8% to 23.0% for DCN, to 26.5% for CN, and to 29.6% for KN. It was suggested that the DCN-oil supplement diet might increase the spreadability of butter at refrigerated temperatures. Studies at the University of California (Davis) are focused on the desaturase gene to produce milk with decreased levels of SFA (CDRF, 2004). The researchers targeted the stearoyl-CoA desaturase enzyme that converts specific medium- and long-chain SFA to their monounsaturated forms. The overall fatty acid composition of the milk of a transgenic line of goats that expressed a bovine b-LG promoter-rat stearoyl-CoA desaturase gene tilted in favor of a less saturated and more MUFA profile at the seventh day of lactation (Reh et al., 2004). Efforts are under way to determine if genetic differences among breeds and individual animals are translated into ratios of SFA and unsaturated fatty acids.

B. Increasing CLA levels in milk fat Milk fat is a good source of the putative anticancer agent, conjugated linoleic acid (CLA), a product synthesized in the rumen during the biohydrogenation of linoleic acid (LA). Table 2 lists the CLA content in selected dairy products. Research has shown that it is possible to influence the extent of ruminal biohydrogenation and the concentration of CLA absorbed and incorporated into milk fat. There is evidence that the concentration of CLA in milk influences its pharmaceutical properties (Kelly and Bauman, 1996). The level of CLA could, therefore, influence the value of the milk as a commodity, although it is not at present a criterion for deciding the price of milk. CLAs reportedly suppress carcinogens, inhibiting proliferation of leukemia and cancers of the colon, prostate, ovaries, and breast. They are the only natural fatty acids accepted by the National Academy of Sciences of United States as exhibiting consistent antitumor properties at levels as low as 0.25–1.0% of total fats (Eynard and Lopez, 2003). The other reported

168

Latha Sabikhi

TABLE 2

a

CLA content in selected dairy productsa

Dairy product

Total CLA (mg/g fat)

Buffalo milk Cow Milk Homogenized milk Butter Cultured buttermilk Ice cream Yoghurt Low fat Nonfat Plain Cheese American processed Cottage Mozzarella Ricotta Sharp Cheddar Romano

6.1 5.5 4.5 6.0 5.4 3.6 4.4 1.7 4.8 5.0 4.5 4.9 5.6 3.6 2.9

Compiled from Muller and Delahoy (1988), National Dairy Council (2000), and Tyagi et al. (2004).

beneficial health effects of CLA as supported by biomedical studies with animal models are antiatherogenic effect, altered nutrient partitioning, improved lipid metabolism, antidiabetic action (type II diabetes), immunity enhancement, and improved bone mineralization (Bauman et al., 2001; Bell and Kennelly, 2001). Reports suggest that feeding lipid sources rich in linoleic and linolenic acids either as seeds or free oil increases the CLA content of milk when oil is accessible to the rumen microorganisms for biohydrogenation (Dhiman et al., 2000). The scientists found that supplementing the dietary dry matter with 2% or 4% soybean resulted in a 237% or 314% increase in CLA content of milk compared with the control. Stanton (2000) and her team worked on the supplementation of cow’s diet with ingredients such as full fat rapeseed, full fat soybean, and pulp-n-brew (by-product of brewers’ grains rich in LA) to study their effect on the CLA levels in milk. When diets of pasture-fed cows were supplemented with full fat rapeseed and full fat soybean, the CLA levels in milk fat increased by 53% and 34%, respectively, after 18 days of feeding when compared to the unsupplemented group of cows on pasture which served as control. The yield and proximate composition of milk were unaffected by the supplementation.

Designer Milk

169

Milk from a grass-fed cow can have five times as much CLA as milk from a grain-fed animal (Robinson, 2003). An experiment supplementing either silage, autumn grass, or spring grass over three periods with pulp-n-brew revealed that the CLA levels increased in case of supplementation of silage and autumn grass, but was less effective in the case of spring grass (Stanton, 2000). Spring grass feeding led to a 2.1-fold increase in CLA content of milk. The CLA-enriched milk fat exhibited cytotoxicity toward mammary and colon cancer cells. Incorporating CLA along with soy oil in the diet of cows increased the CLA levels, simultaneously decreasing the SFA in milk fat (Pszczola et al., 2000). In an attempt to increase the CLA content in milk via the cow’s diet, Bell and Kennelly (2001) divided 28 Holstein cows into 4 groups and fed them different diets—control diet (CTD), low-fat diet (LFD), high-fat diet 1 (HF1), and high-fat diet 2 (HF2). The animals were kept on CTD for 8 days before starting the different diet regimen. All experimental diets resulted in lower fat percentage in the milk when compared to CTD, whereas other parameters such as milk yield, protein, and lactose were unaffected. The CLA concentration in milk fat was 0.49%, 0.56%, 3.7%, and 5.63% in the group fed CTD, LFD, HF1, and HF2, respectively. Thus, increasing the fat content in the diet increased the CLA content up to 9–12 times, despite lower total fat content. AbuGhazaleh et al. (2003) found that feeding lactating dairy cows a blend of fish oil and MUFA and PUFA resulted in an increase in the concentrations and yields of CLA in milk, the greatest increase being with a blend of a high LA source (e.g., regular sunflower seeds). Beaulieu and Drackley (2004) reported similar results where a diet rich in LA led to increasing the CLA levels in milk fat twofold. Supplementing nonluminous green fodder with mustard cake in the feed of buffaloes resulted in 6.18-mg CLA per gram of fat as compared to 6.05 mg/g when the supplement was groundnut cake (Tyagi et al., 2004). The total CLA in buffalo milk and milk products increased significantly when the animals were fed berseem and wheat straw in the ratio 87:13. Tsiplakou et al. (2006) examined the CLA content in the milk fat of sheep and goat milk segregated into two groups. Animals in Group 1 were totally on pasture from April onward with supplementary feeding during winter, whereas animals in Group 2 served as the control group and were kept indoors without grazing. The study revealed that the CLA content in milk fat of Group 1 increased in April and May, during the availability of early grass and declined thereafter, whereas that in Group 2 remained more or less constant. The CLA content (cis-9, trans-11) in sheep milk was 2% of the total fatty acids fat content and was much higher than that in goat milk (0.62%). Animal variation is also a major source of differences in the CLA content of milk fat. Bauman and Perfield (2002) discovered that the 9,11

170

Latha Sabikhi

isomer of CLA in milk fat is synthesized by the cow and not rumen bacteria as had earlier been reported. Synthesis involves a mammary enzyme, delta-9 desaturase, which acts on a trans-fatty (vaccenic) acid produced by rumen bacteria. Several genetic factors that regulate the expression of the delta-9 desaturase gene have been identified. In a line of transgenic goats that contained a rat stearoyl-CoA desaturase gene targeted at converting medium- and long-chain SFA to their monounsaturated forms, Reh et al. (2004) found that the desaturase enzyme also converted the rumen-derived MUFA C18:1 trans-11 to the C18:2 cis-9 trans-11 isomer (CLA) in the milk fat of one of these animals.

C. The omega fatty acids Omega-6 and omega-3 are essential fatty acids, but the body requires them in a ratio that is not normally achieved by the typical diet of today’s developed nations. It is reported that the current average intakes of essential fatty acids expressed as ratios of o-6 to o-3 fatty acids are 8:1 in United Kingdom, 10:1 in United States, and 12:1 in Australia (www. omega-3info.com). Health bulletins indicate that the proportion of o-6 to o-3 fatty acids should be equal or close to 5 for cardiovascular health (Simopoulos, 1999). At present, the average PUFA content in modern diets (nearly 30% of calories) is too high. It is suggested that our PUFA intake should not be much greater than 4% of the caloric total, in approximate proportions of 2% o-3 linolenic acid and 2% o-6 linoleic acid (Fallon and Enig, 2000). The intake of total o-3 fatty acids in the United States is 1.6 g/day (KrisEtherton et al., 2002). Of this, a-linolenic acid (ALA) accounts for 1.4 g/ day, whereas eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) together account only for 0.1–0.2 g/day. DHA is required by the brain and nerve cells and is essential for normal visual and neurological development in infants (Tomlinson, 2003). The major food sources of ALA are vegetable oils, principally canola and soybean oils. Oily fish are the richest source of EPA and DHA. EPA and DHA can be made by the body from ALA, but sometimes this capacity is impaired, so oily fish remains the best source. The recommendations for intake of o-3 fatty acids range from 0.5 to 2 g/day. ISSFAL (International Society for the Study of Fatty Acids and Lipids) recommend 0.65-g EPA and DHA per day (Willumsen, 2006). Of this, the content of each should be at least 0.22 g. Omega-6 is the essential fatty acid that is in ample supply in oils, nuts, and seeds. Table 3 lists the o-6 and o-3 fatty acid contents in commonly available food ingredients. Too much o-6 in the diet creates an imbalance that can disrupt the production of prostaglandins leading to increased tendency to form blood clots, inflammation, high blood pressure, irritation of the digestive tract,

Designer Milk

TABLE 3 The o-3 and o-6 fatty acids content in some common food ingredientsa Food ingredient

o-3 FA

o-6 FA

Vegetable oils (g/100g) Almond oil Canola/rapeseed oil Corn oil Flax/linseed oil Grapeseed oil Olive oil Palm oil Safflower oil Sesame oil Soybean oil Sunflower oil Walnut oil Wheat germ oil

0 9 0.7 58 – 0.60 0.2 – 0.3 7 – 11.5 7

17 20 58 14 68 7.90 9 74 41 51 63 58 55

Fish oils (g/100g) Cod-liver oil Salmon oil Sardine oil

20.5 36 26

1.9 4.5 5

Nuts (g/100g edible portion) Almonds Brazilnuts Cashew nuts Hazelnuts Peanuts Pine nuts Pistachios Walnuts

Trace Trace Trace Trace Trace 1 0.254 9

10 23 8 4 16 25 13 37

Seeds (g/100g edible portion) Flax/linseeds Pumpkin seeds Safflower seeds Sesame seeds Sunflower seeds

15–25 7–10 0.111 Trace Trace

6 20 28 25 30

Meat and fish (EPA þ DHA g/100g edible portion) Poultry 0.05 Oily fish 1.8–1.9 Bacon and ham 0.008–0.009 a

Compiled from: www.annecollins.com, www.longevinst.org, www.nutraingredients.com

– – –

171

172

Latha Sabikhi

depressed immune function, sterility, cell proliferation, cancer, water retention, and weight gain. On the other hand, deficiency in o-3 is associated with asthma, heart disease, and learning deficiencies. It is established that o-3 fatty acids have a hypolipidaemic action in human, reducing harmful cholesterol levels, particularly plasma triglycerides (Tomlinson, 2003). It also has an anti-inflammatory action and helps to reduce platelet aggregation. Essential fatty acids have proven to be effective in the treatment of several other ailments including eczema, rheumatoid arthritis, asthma, Alzheimer’s disease, and Attention Deficit Hyperactivity Disorder (ADHD). There are reports that approximately equal amounts of these two fats in the diet will result in lower risk of cancer, cardiovascular disease, autoimmune disorders, allergies, obesity, diabetes, dementia, and some mental disorders (www.flax.com/newlibrary/ESSENT.html). Dietary manipulation in cows is a practical way to maintain a desired ratio of o-6 to o-3 fatty acids in milk. Milk from pastured cows contains an ideal ratio of essential fatty acids. Dhiman et al. (1999) reported equal quantities of the omega fatty acids (16.5 mg/g fat) in the milk of cows entirely on pasture. Reducing the proportion of grass to two-third of the ration increased the o-6 fatty acids to 31.4 and decreased o-3 fatty acids to 13.5 mg/g milk fat. Further reduction in the dietary proportion of grass to one-third resulted in 42.7 and 8.2 mg/g fat of o-6 and o-3 fatty acids, respectively. There are reports that organic milk contains almost 70% more o-3 fatty acid than nonorganic milk (Cheek, 2006). Mammals are dependent on dietary sources of essential fatty acids as they lack the desaturase enzymes necessary to synthesize them. Kao et al. (2006) engineered transgenic mice expressing the o-3 fatty acid desaturase enzyme from the nematode Caenorhabditis elegans, which synthesizes a wide range of PUFA and possesses the only known example of an o-3 desaturase enzyme in the animal kingdom. The milk from these mice had more o-3 and less o-6 PUFA, and hence had showed an overall decrease in the o-6:o-3 PUFA ratio in the milk. The milk phospholipids from the transgenic mice had an o-6:o-3 ratio of 1.78 as compared to 9.82 in the control animals. The authors anticipate that this may be a suitable method to improve the nutritional profile of dairy-based diets.

D. Reducing fat content in milk It has long been recognized that the yield of milk fat can be altered through nutritional interventions. Several workers have reported that supplementing normal diet with fats in different forms and concentrations decreases the yield of fat in milk (Baumgard et al., 2000; Bell and Kennelly, 2001; Chouinard et al., 1999; Peterson et al., 2002). Genetic studies also pointed to the power of hereditary traits in influencing the quality of milk. Genetic markers for milk quality of dairy cattle were

Designer Milk

173

discovered and reported by the Iowa State University in the United States in 1996 (www.biotech.iastate.edu/biotech_update). Laboratory experiments with the marker revealed that animals with the ability to produce low-fat milk could be accurately identified. Such genetic testing was aimed at improving dairy herd performance by identifying animals with the potential to produce low-fat milk. A herd producing low-fat milk was seen as a means to reduce milk-processing costs as the low-fat milk eliminated the necessity to separate fat from milk. As a variation to altering the fat composition, Wall et al. (1997) suggested that modifying the cow’s genetic makeup to enable it to produce milk with 2% fat would reduce the cost of feed per kilogram milk by 22%. In changing the fat composition, targeting enzymes that influence the synthesis of fat is important. As an example, reduction of acetyl-CoA carboxylase that regulates the rate of fat synthesis within the mammary gland would translate to a drastic reduction in the fat content of milk and reduce the energy required by the animal to produce milk (Ntambi et al., 1999).

E. Type of fatty acids versus product quality The type of fatty acids present in milk fat can influence the flavor and physical properties of dairy products. There are reports that butter produced from cows fed high-oleic sunflower seeds and regular sunflower seeds were equal or superior in flavors to the control butter (Middaugh et al., 1988). The experimental butter was softer, more unsaturated and exhibited acceptable flavor, manufacturing, and storage characteristics. Other workers (CSIRO, 1999; Mason, 2001) have also reported the increase in the unsaturated fatty acids content in milk fat, leading to an improvement in the spreadability of butter even at refrigerated temperatures (Section III.A). Extruded soybean and sunflower diets yielded a Cheddar cheese that had higher concentrations of unsaturated fatty acids while maintaining flavor, manufacturing, and storage characteristics similar to that of control cheese (Lightfield et al., 1993). It is also beneficial from a safety point of view as the accumulation of fatty acids, namely C12, C14, C18:1, and C18:2, enhanced the safety of cheeses against Listeria monocytogenes and Salmonella typhimurium (Schaffer et al., 1995). Increasing the oleic acid content of milk fat from 26% to over 40% by feeding calcium salts of HOSO containing more than 86% oleic acid at the rate of 7.5% of diet dry matter weight to Holstein cows did not affect the sensory and physicochemical properties of Latin American white cheese (Queso Blanco). There was also no difference (as a result of the modified and improved fatty acid profile) between the firmness of the product from modified milk and that made from normal milk (Aigster et al., 2000).

174

Latha Sabikhi

IV. MILK SUGAR (LACTOSE) MODIFICATION Lactose, the major milk sugar, is also responsible for the osmotic regulation of lactation, thus causing the movement of water into milk. This carbohydrate is synthesized in the secretory vesicles of the mammary glands by the lactose synthase complex. As lactose cannot diffuse out of the vesicles, it draws water into the vesicles by osmosis. Thus, the volume of milk produced is directly dependent on the amount of lactose synthesized. Lactose cannot be transported to the bloodstream directly. It can be absorbed only after its enzymatic hydrolysis to the monosaccharides glucose and galactose by intestinal lactase (b-galactosidase). For many human beings, the level of b-galactosidase declines early in life to the point of virtual absence in adulthood, making them lactose intolerant. It is reported that more than 75% of the human adult population suffers from deficiency of b-galactosidase (Vilotte, 2002). When such individuals ingest milk or milk products, the lactose remains undigested and malabsorbed in the gut, where it causes retention of water by its osmotic action. This water retention coupled with the bacterial production of large volumes of carbon dioxide leads to intestinal upset and dehydration (Vesa, 1999). One management tactic suggested for such patients is the avoidance of dairy products. However, as milk is a major component in the human diet, this deprives them from the use of a valuable nutritional source. In addition, since milk can provide much of the required calcium for maintaining bone health, lactose intolerance can also be associated with osteopaenia in old people (Corazza et al., 1995). A report suggests that by 2020, half of all American citizens older than 50 will have low bone mass and be at risk for fractures from osteoporosis if appropriate dietary and other precautions are not followed (Carmona, 2006). Therefore, excluding milk from diet has adverse effects on health. The consequences of lactose intolerance can also be limited through the use of b-galactosidase-replacement (preharvest) or hydrolyzed lowlactose (postharvest) products. Besides the obvious nutritional advantage, a reduction in milk lactose content could also benefit agricultural and industrial purposes with less volume to transport, better milk coagulation, and less effluent production. The complete removal of lactose from milk creates milk that is extremely viscous, containing very little water. It is extremely difficult to extract this milk from the mammary gland, making the milking process difficult and painful for the animal. However, research has shown that with controlled reduction in the lactose content of milk, it is possible to decrease the water, increase the percentage of total solids, and reduce the lactose yield of the milk while keeping fluidity intact.

Designer Milk

175

A. Preharvest methods of lactose reduction a-Lactalbumin (a-LA) is one of the major milk proteins present in almost all mammalian milks. It interacts with b-1,4-UDP-galactosyl transferase (UDPgal) to modify substrate specificity of this enzyme, virtually creating a unique binding site for glucose and leading to the synthesis of lactose (Vilotte, 2002). The preharvest methodologies of reducing lactose involve either the introduction of b-galactosidase enzyme into milk via mammary gland-specific expression or the removal of a-LA and gene ‘‘knockout’’ methodologies. Although these successful approaches provide valuable tools to address milk physiology, they reduce the overall sugar content of the milk, resulting in highly viscous milk. Studies on mice have revealed that reduction of lactose via a-LA deletion was inappropriate because it impaired milk volume regulation. The milk of such mice was highly viscous with very high protein (88%) and fat (60%), no a-LA and no lactose (Karatzas and Turner, 1997). Knocking out the UDP-gal gene in mice also produced milk with no lactose but very high viscosity (Vilotte, 2002). An alternative to produce low-lactose milk is overexpression of b-galactosidase in milk. However, the monosachharides produced within the formed milk increases the osmotic pressure within the alveolar lumen, thereby drawing more water and resulting in further dilution of other milk components (Bremel et al., 1989). Jost et al. (1999) explained an in vivo technique for low-lactose milk production. They generated transgenic mice that selectively produced a biologically active b-galactosidase in their milk. In these transgenic mice, the lactose content of the milk is at least halved, even though the b-galactosidase expression levels were relatively low. The authors claim that it is likely that at least twofold greater levels of lactose reduction could be achieved. In contrast to the previous studies by Bremel et al. (1989) and Karatzas and Turner (1997), these experiments led to reduction in the lactose content while retaining most of the monosaccharide content of the milk. b-galactosidase synthesis in the mammary gland caused a significant decrease in milk lactose (50–85%) without obvious changes in fat and protein concentrations. It thus helped to maintain a balanced nutrient supply as reflected in the similar growth curve reared on transgenic or control milk. It is likely that transgenic low-lactose milk production could offer a more balanced approach to managing lactose intolerance than postharvest or lactose-replacement products. It is also technically feasible to produce transgenic livestock carrying this transgene and probably similar or better expression levels could be achieved. However, more detailed analysis on several aspects such as the effect of splitting the lactose into glucose and galactose on the osmotic balance in the milk in the gut after ingestion and also the economic viability of the technology need to be investigated (Whitelaw, 1999).

176

Latha Sabikhi

V. MILK PROTEIN MODIFICATION One of the major products of the mammary glands being protein, exciting opportunities in research and technology extend the benefits of better protein supplementation. One of the most obvious changes in milk is the selective increase of a component that is already present. For example, an increase in one of the casein components in milk might provide a method of increasing the value of milk for the production of cheese. It was estimated that an increase of 20% in the as1-casein (as1-CN) would increase the revenue of the cheese industry by almost $200 million annually (Hennighausen et al., 1990). Similarly, improvement in the amino acid profiles by increasing the amounts of L-taurine, L-leucine, and L-phenylalanine offers additional nutritional benefits. Protein modification in milk started with experiments on laboratory animals two decades ago. The success of these experiments on small animals prompted researchers to extend the work on cattle and other farm animals with considerable success. The Dairy Cooperative Research Centre (Dairy CRC) in Australia reported the cloning of 14 calves with an extra copy of the cow’s own gene for casein besides the 4 normal ones found in cattle (CRC Factsheet, 2006). This increased the quantity of protein secreted in the milk, thus increasing the nutritional quality and the value of milk.

A. Modifying the major milk proteins The four bovine casein genes lie within a single, multigene locus of 200 kb in length. Zuelke (1998) worked on the hypothesis that this multigene locus contains all of the DNA sequences required to regulate the coordinated expression of all four individual casein genes. A bacterial artificial chromosome (BAC) library of genomic DNA from elite dairy cattle was prepared in his laboratory and tested in mice with the hope that transgenic calves that possess this BAC casein construct could be produced. Jeng et al. (1997) characterized and partially purified bovine b-casein (b-CN) from the milk of transgenic mice. The approximate expression of the protein was 3.0 mg/ml of milk. The workers reported that phosphorylation of the bovine b-CN in the milk of transgenic mice was the same as that of native bovine b-CN. If the modification and/or enhancement in the b-CN could be extrapolated into farm animals, there would be several other potential advantages to processing. The extra b-CN would increase the cheese yield besides improving the curd strength. Bleck and Wheeler (1998) also reported the generation of b-CN in the milk of transgenic mice. While murine milk normally contains about 30% total solids, the experimental animals produced milk that had 40–50%

Designer Milk

177

total solids (10–20% higher). This increase in total solids obviously caused a decrease in the amount of water in the milk and was accompanied by the concomitant decrease in total volume of milk. Thus, the milk was very viscous, lacked fluidity, and could not be removed easily from the mammary gland. There has also been an attempt to form a glycosylated b-CN in milk (Bleck et al., 1998a). This has the potential to increase the solubility of b-CN and modify other functional properties such as viscosity, water-holding capacity, foaming, and emulsification. k-Casein (k-CN) is responsible for micelle formation and establishes micelle size and function, thus influencing many of the physical characteristics of milk. Gutierrez-Adan et al. (1996) generated transgenic mice bearing the bovine k-CN gene. They found that the milk from transgenic mice with high bovine k-CN had a significantly smaller micelle size than did control milk. Although there was no effect on the rennet coagulation time, the milk of transgenic lines had stronger curd in gels produced by rennet. Brophy et al. (2003) introduced additional copies of the genes encoding bovine b-CN and k-CN into female bovine fibroblasts. The transgenic cows secreted elevated levels of b-CN (8–20%) and k-CN (twofold) and had a considerably modified k-CN to total casein ratio. b-CN, which is the most abundant milk protein, is involved in binding calcium phosphate and thus controlling milk calcium levels. Higher k-CN content in milk is linked to smaller micelles, better heat stability, and improved cheesemaking properties. In the transgenic animals engineered by Brophy et al. (2003), the total milk protein increased by 13–20% and total milk casein by 17–35% compared to nontransgenic control cows. This has obviously a positive influence on the cheese yield and also the casein and milk protein concentrate industry. Edible casein is used in vitamin tablets, instant drinks, and infant formulas, whereas acid casein is used for paper coatings, cosmetics, button making, paints, and textile fabrics (Karatzas, 2003). There was a measurable variation in the concentration of both b-CN and k-CN among the eight transgenic genetic clones generated by Brophy et al. (2003). In addition to embryonic cell-derived nuclear DNA, the transgenic animals contained oocyte-derived mitochondrial DNA. St. John (2002) postulated that this mitochondrial DNA may also have been derived from donor cell. As several metabolic reactions responsible for lactation are located in the mitochondria, it is possible that differences in the source of the genetic bloodline of the mitochondria in these transgenic animals may account for differences in the physiology of lactation and ultimately milk production. A2 MilkTM from commercial dairy herds is being marketed in New Zealand and Australia at a small premium over regular or A1 milk. A2 Corporation scientists claim that as A2 MilkTM has only negligible amounts of the A1 b-CN in it, the perceived risks associated with the

178

Latha Sabikhi

consumption of this type of casein (such as autism or Asperger’s syndrome, child diabetes, schizophrenia, and coronary heart disease) are effectively removed (Lacefield, 2003). They further maintain that A2 b-CN was the original b-CN gene, whereas subsequent genetic mutation generated A1 b-CN. Hence, all milk produced by cattle thousands of years ago, before the large-scale domestication of cows, was A2 MilkTM (www.a2corporation. com). The animals producing A2 MilkTM are not genetically modified. They have been selected with the help of genetic markers that indicate cows that naturally produce in their milk, the original form of casein protein (A2b) rather than the altered A1 b-form (Goel, 2005).

B. Modifying the minor milk proteins The first attempts at modification of milk proteins through genetic engineering techniques started with the minor milk proteins. Simons et al. (1987) generated transgenic mice carrying the sheep b-LG gene. The b-LG was specifically and plentifully expressed in the mammary gland of mice during lactation, though the protein is not naturally present in rodent milk. Bleck and Bremel (1994) produced transgenic mice to study the production of bovine a-LA in their milk. Milk of multiple mice from the second, third, and fourth generation from each of the three transgenic lines was analyzed for the presence of bovine a-LA. The protein was present at concentrations up to 1.5 mg/ml of mouse milk. Bovine a-LA from the milk of transgenic mice was characterized, partially purified, and quantified as 1.0 mg/ml by Jeng et al. (1997). The N-terminal amino acid sequence of HPLC-purified bovine a-LA from mouse milk was identical to native bovine a-LA. In addition, the calcium-binding properties of this protein were also similar to the native protein. More details on the modification of minor milk protein fractions are enumerated in other segments (Section VI.A–C) of this chapter.

C. Targeting the proteinase-cleavage sites Caseins, particularly the b-, as1-, and as2-caseins being easily digestible, are quite sensitive to plasmin, a serine protease occurring naturally in milk along with plasminogen. Plasmin activity leads to limited proteolysis in milk. This offers a dual disadvantage of decreasing the curd yield, besides causing bitterness in cheese and inducing organoleptic defects and gelation in ultra high temperature-treated milk. Milk augmented with specific inhibitor of either plasmin or plasminogen activator would therefore be a boon for the process industry (Murthy and Kanawjia, 2002). Bleck et al. (1998a) modified b-CN to remove the plasmin-cleavage site. They also report the removal of the chymosin-cleavage site from b-CN, thus positively influencing the cheese yield.

Designer Milk

179

VI. DESIGNER MILK FOR INFANT HEALTH It is said that breast milk is the ultimate designer food for babies. Nature has designed human milk for optimal nourishment and growth during infancy and also for supplying certain bioprotective factors that afford protection against commonly occurring infections. However, in certain situations such as lactation failure, insufficient milk secretion, and where mothers suffer from transmittable diseases, human milk substitutes serve as precious lifesavers during vulnerable stages of infancy. Then it becomes imperative to have infant formulas, which closely imitate human milk so as to provide comparable nutritional and health benefits. The composition of these formulas could be greatly improved to suit the needs of the infant by incorporating ingredients that resemble those of human milk, thereby ‘‘humanizing’’ the bovine milk.

A. Lactoferrin Lactoferrin (LF) is a single-chain, metal-binding glycoprotein of 77 kDa and is a component of the intrinsic host defense of mammals. It has antibacterial, antifungal, anti-endotoxin, and antiviral activities. It is an iron-binding protein and may also mediate some effects of inflammation and have a role in regulating various components of the immune system. LF in milk might play a role in iron absorption and/or excretion in newborns, as well as in promotion of intestinal cell growth. Its level in human milk is about 1 g/liter and in human colostrums, about 7 g/liter. As the levels of LF in cow milk is only about one-tenth of that in human milk, this has caught the attention of those involved in designing human milk replacement formulas. Oral feeding of bovine LF (1 mg/ml) led to an increase in the probiotic species bifidobacteria in infant gut (Roberts et al., 1992). Several such infant formulas are marketed in Japan under brand names such as Hagukumi, Chilmil Ayumi, Non-Lact, E-Akachan, GP-P, and New-NA-20Morinaga. The consumption of such formulas may result in anti-infection, improvement of orogastrointestinal microflora, immunomodulation, anti-inflammation, and antioxidation (Wakabayashi et al., 2006). Researchers (Nuijens et al., 1997) at the Leiden University (the Netherlands) in collaboration with Pharming, NV (Leiden, the Netherlands) compared recombinant human lactoferrin (rhLF) expressed in the milk of transgenic mice with natural human milk-derived lactoferrin (hLF). They concluded that the unsaturated rhLF and natural hLF had comparable properties, indicating that hLF produced in bovine milk will exert similar, if not identical, antibacterial and anti-inflammatory activities in vivo. Pharming also developed the first transgenic bull in the late

180

Latha Sabikhi

1980s and a line of transgenic cows to produce several proteins including hLF (Subramanian, 2004). The company believes that as receptors in the human gut have better affinity to a human protein than a bovine one, the ingredient would be more effective in boosting gut health. Four lines of transgenic cows that harbor the rhLF were developed (van Berkel et al., 2002). The milk of these animals had 0.4, 0.8, 2, and 3 g/liter of the rhLF in their milk. These levels of expression remained constant throughout the lactation period of 280 days. The milk volume, cell counts, and proximate composition were not altered by the genetic transformation. The recombinant protein was structurally and functionally comparable to natural hLF and had similar iron binding and release and antibacterial activities. The authors further postulate that with such expression levels and an assumed milk yield of 8000 liters of milk per cow annually, one cow can produce about 24-kg rhLF in a year. Thus, a herd of a few hundred animals could produce enormous quantities of this biological protein in a year.

B. Lysozyme Lysozyme (LZ) is an enzyme that is abundantly present in the mucosal membranes that line the human nasal cavity and tear ducts. It can also be found in high concentration in egg white. LZ destroys bacterial cell walls by hydrolyzing the polysaccharide component of the cell wall. Human milk contains 0.4 g/liter of LZ, an enzyme that contributes to antibacterial activity in human milk. Active human lysozyme (hLZ) has been produced in the milk of transgenic mice at the concentrations of 0.78 g/liter (Maga et al., 1995). Milk from these transgenic lines had the same antibacterial activity as human milk LZ. The researchers found a zone of clearance in the gel containing the test organism Micrococcus lysodeikticus and the recombinant protein indicating that the hLZ in the mouse milk was active. On the processing front, the expression of LZ in milk results in the reduction of rennet clotting time and greater gel strength in the clot. In the transgenic line of mice generated by Maga et al. (1995), the milk exhibited a 35% decrease in rennet clotting time, a smaller casein micelle size (157 nm as against 172 nm in the nontransgenic animals) and a 2.5- to 3-fold greater gel strength than control milk. A group of researchers in China also developed two lines of transgenic mice that expressed fully active recombinant hLZ in the mammary gland (Yu et al., 2006). The antibacterial activity of the LZ from the transgenic lines (480.4 and 301.6 U/ml) was 18 and 11 times greater than that of the nontransgenic mice (25.9 U/ml). Maga et al. (2006) designed a line of transgenic goats that expressed hLZ in the mammary gland. On characterizing the milk from five transgenic goats of this line, they found that the hLZ content in the milk was

Designer Milk

181

270 mg/ml or 68% of the level found in human milk. Milk from these transgenic animals had a lower somatic cell count, which may influence udder health positively. It also had a shorter rennet clotting time and increased curd strength (17 min and 26.3 Pa, respectively, in the experimental samples as against 20 min and 20.7 Pa in control samples). The aim now is to produce cows that will produce LZ in their milk. Such LZfortified milk has the potential to reduce udder infections in dairy cows and intestinal ailments in humans who drink milk (Bailey, 2001). A double transgenic cow that coexpresses both hLF and hLZ in milk may also reduce the incidence of intramammary infection or mastitis. Feeding young goats and pigs with this LZ-enriched milk produced by transgenic goats altered their intestinal bacterial profile (www.eurekalert. org, 2006). Pigs were chosen owing to the similarity of their digestive system to that of humans. The choice of goats extended the study to ruminant models. The young pigs fed the LZ-rich milk from transgenic goats had lower levels of coliform bacteria in the small intestine, including fewer Escherichia coli, than did the control group. In contrast, the kid goats fed LZ-rich goat’s milk had higher levels of coliform bacteria and roughly the same level of E. coli, compared to control group. The researchers attributed this variation to the difference in the respective digestive systems and the bacterial profile of the systems. Despite the difference, both animal groups were healthy and exhibited normal growth patterns. The researchers anticipate that these results will pave the way for protection of infants and children against diarrheal illnesses through milk-feeding programs.

C. Cow milk allergy An allergic reaction to cow milk is a complex disorder involving an abnormal immunological response to one or more of milk’s proteins and more than one immunological mechanism. Both casein and whey proteins are reported to be responsible for these allergic responses. Although the reasons for cow milk allergy are not well understood, genetic and environmental factors and their interaction are thought to be responsible (Crittenden and Bennett, 2005; Halken, 2004; Wal, 2002). Comparatively few infants develop cow milk protein allergy. Usually, infants and young children (2%) suffer from this ailment and outgrow it by the age of five (Host, 2002). It is rare in adults (0.1–0.5%). Cow milk protein allergy can be diagnosed by one or more of cutaneous (e.g., eczema, rashes), gastrointestinal (e.g., nausea, vomiting, diarrhea), or respiratory (e.g., asthma, rhinitis, wheezing) symptoms. Severe symptoms would need special prescription medications such as antihistamines and epinephrine (Nuble, 2006). The only effective management strategy

182

Latha Sabikhi

for cow milk protein allergy is avoidance of cow milk and its products, which in turn negatively influences nutritional management through diet. Cow milk allergenicity in children is often caused by the presence of b-LG, which is absent in human milk. Although b-LG has been implicated most often in allergic reactions to cow milk, the caseins, a-LA, serum albumin, and immunoglobulins and digests of these proteins are also allergenic in infants and children. Elimination of b-LG by knocking out its gene from cow milk is unlikely to have any detrimental effects on either cow or human formula and might actually overcome many of the major allergy problems associated with cow milk. AgResearch (New Zealand) is field-testing dairy cattle that have been genetically modified to eliminate the b-LG gene (www.agresearch.co.nz, 2001). Further, as milk protein allergenicity studies demonstrate that all food proteins are potential allergens and that allergenic structures are widely spread throughout the protein molecule, milk is a good model in the search for means of characterizing allergenic structures in food (Wal, 1998). Therefore, while developing strategies for the identification and evaluation of potential allergenicity in novel foods, many of the technological practices used in the assessment of milk protein allergenicity can be adapted.

D. Lactose intolerance Lactose intolerance is a distinct entity from cow milk protein sensitivity and causes abdominal pain, diarrhea, nausea, flatulence, and/or bloating. While avoidance of milk and other dairy products will bring relief in children suffering from lactose intolerance, it may cause problems in optimal bone mineralization owing to lack of calcium in diet. Several lactose-free and lactose-reduced milks are now available in markets to cater to such infants. The scope of transgenic technology to reduce the lactose content in the milk of small animals has been reviewed elsewhere in this chapter (Section IV.A). The extension of this technique to include farm animals is targeted in the future. Inability to digest milk is not exclusively due to lactose intolerance. From a study involving African-Americans between the ages 12 and 40 years, Johnson et al. (1993) concluded that cause of milk intolerance in as many as one-third of the subjects claiming symptoms after ingestion of a moderate amount of milk was not its lactose content. Many infants, especially those born before term, have low lipase activity. One potential application of transgenic technology could be to produce the human lipase, which is stimulated by bile salt in the milk of bovines. The lipase thus produced could be used as a constituent of infant formulas to increase the digestibility of milk lipids, particularly in premature infants who have low lipase activity (Lonnerdal, 1996).

Designer Milk

183

VII. MILK WITH HUMAN THERAPEUTIC PROTEINS The preparation of high-value, low-volume therapeutic proteins in the milk of domestic animals through transgenic technology is becoming a reality. Several high-affinity and high-specificity monoclonal antibodies for in vivo therapy in human beings using transgenic mice have been reported earlier (Gorman and Clark, 1990; Little et al., 2000; Thomas, 2001; Yang et al., 2001). Antibodies are used for a number of human clinical applications such as treatment of infectious diseases, cancer, transplanted organ rejection, autoimmune diseases and also as antitoxins. Statistics suggest that at least 33 different drugs in clinical testing and in pivotal trials contain variable regions encoded by human sequences from transgenic mice (Lonberg, 2005). Progress in research may make it possible to extend this technology to use transgenic farm animals to directly generate and produce human proteins. There were several bottlenecks envisaged in realizing this hypothesis a decade earlier. First, the unpredictability of the expression level of the genes of interest associated with milk protein gene control regions was recognized as a challenge. Then, the recombinant proteins secreted in milk are not always in a satisfactory biochemical form. It was also observed that cleavage and glycosylation are not always carried out correctly. The problem of the possible presence of agents pathogenic for humans in proteins extracted from milk was also a major worry (Houdebine, 1995). Despite these disadvantages, the major benefit of transgenic technology offers a means to produce proteins at a very low cost. Mammalian cell culture systems are often used for expression of recombinant human proteins (rHP), as the latter can only be obtained in a biologically active conformation when produced in such cells. However, this approach has limited production capacity and is expensive. In contrast, the production of rHP in milk of transgenic cattle is a safe and less-expensive alternative with the advantage of better protein output (Brink et al., 2000). Ebert et al. (1991) reported the generation of two transgenic goats that expressed a variant of human tissue plasminogen activator (htPA). The milk from one of these contained enzymatically active longer acting tissue plasminogen activator (LAtPA) at a concentration of 3 mg/ml. Economic comparison of production costs of htPA through bacterial fermentation, mammalian cell culture, and cow transgenic technology estimates the cost per gram of htPA to be 20,000, 10,000, and 10 US$, respectively (Karatzas and Turner, 1997). Shani et al. (1992) tested the feasibility of producing large quantities of human serum albumin, which is used in blood transfusions, in the milk of transgenic livestock by generating transgenic mice as a model system. Charlie and George were two calves created from fetal cells in the US in 1998 to produce the human serum albumin in their milk (Johnston, 2006).

184

Latha Sabikhi

The first report on expression of recombinant human fibrinogen (rHF) to the mammary gland of transgenic mice appeared a decade ago (Prunkard et al., 1996). Fibrinogen is a complex plasma protein composed of two each of three different polypeptide chains. The workers coinjected three expression cassettes, each containing the genomic sequence for one of the three human fibrinogen chains controlled by sheep whey protein b-LG promoter sequences into fertile mouse eggs. Analysis by PAGE revealed that the milk from the highest producing founder animal contained human fibrinogen subunits at concentrations of 2000 mg/ml. Incubation of the transgenic milk with thrombin and factor XIII produced a cross-linked fibrin clot, demonstrating that a major portion of the secreted fibrinogen was functional. Coleman (1996) reported that the concentration of rHF produced in mouse milk was more than 2 g/liter, whereas the amount generated by the cell culture method was only 0.002 g/liter. Because a-LA has a well-balanced amino acid composition, increasing the amount of a-LA in milk at the expense of b-LG may, besides lowering the risk of cow milk allergies, improve the nutritional quality of the milk. The technique of custom-designing amino acids in a protein to obtain special foods with therapeutic properties offers a ray of hope for patients of phenylketonuria (PKU). This is a congenital disease occurring in those without the enzyme that metabolizes phenylalanine. Although products are now available with no or low phenylalanine content, they are unappetizing. The knowledge that a-LA contains only four phenylalanine residues in its amino acid makeup and their position can be determined easily makes this whey protein a possible target for treatment of PKU (Coleman, 1996). Replacing these four phenylalanine residues with other amino acids by site-directed mutagenesis followed by the subsequent expression of the modified protein in milk and its purification would offer a logical sequence to the preparation of an improved dietary formula for PKU patients. Transgenic animals can also secrete proteins such as blood clotting factors needed by human hemophilia sufferers in their milk (Suraokar and Bradley, 2000). On these lines, Polly, a genetically altered sheep was created at the Roslin Institute in Scotland to produce milk that contained the protein used to treat human hemophilia (Pettus, 2006). Wright et al. (1991) described the generation of five transgenic sheep (four female and one male) for a fusion of the ovine b-LG gene promoter to the human a-1 antitrypsin (AAT) genomic sequences. AAT is a singlechain glycoprotein secreted by the liver and its main function is the inhibition of the enzyme elastase. The absence of active AAT in the system leads to emphysema (loss of elasticity of the lung tissue) and/or other lung-related ailments such as cystic fibrosis and adult respiratory distress syndrome. Milk of three of the ewes generated by Wright et al. (1991)

Designer Milk

185

expressed human AAT (hAAT) at levels greater than 1 g/liter. In one case, initial levels of hAAT exceeded 60 g/liter and stabilized at 35 g/ liter as lactation progressed. hAAT purified from the milk of these animals was fully N-glycosylated and had a biological activity indistinguishable from human plasma-derived material. These animals and their progeny exhibited stable transmission of the transgene. The founder animals yielded very similar levels of hAAT protein in milk continuously over several lactations. A flock of seven first generation ewes (derived from a founder male) yielded comparable levels of hAAT protein in first and second lactation milk. Two second generation ewes of this line also produced equivalent quantities of the human protein (Carver et al., 1993). Several animals from this group are reported to secrete milk containing 35–47 g/liter without affecting the production of other proteins in milk (Coleman, 1996). With successful advances in research, several other recombinant proteins of pharmaceutical interest have been developed from the milk of transgenic animals. In this context, some human proteins have already been expressed with success. Products such as insulin and growth hormone have also been obtained from the milk of transgenic cows, sheep, or goats (Margawati, 2003). GTC Biotherapeutics (Framingham, MA) uses both goats and cows to produce more than 60 therapeutic proteins, including plasma proteins, monoclonal antibodies, and vaccines. A recombinant human antithrombin III—an anticoagulant protein found in blood—which was produced in goat milk and was in the last stage of testing a couple of years ago (Subramanian, 2004) is now almost ready to be marketed. The company claims that the antithrombin (ATrynÒ ) has been recommended for market authorization for the prophylaxis of venous thromboembolism in surgery of patients with congenital antithrombin deficiency (www.transgenics. com/news.html). Besides being the first antithrombin product approved by the European Medicines Agency for use in all 25 countries of the European Union, ATrynÒ will also be the only available antithrombin product that is produced by recombinant biotechnology and is not derived from the human blood supply. This drug is envisaged to be a boon for those suffering from a deficiency in antithrombin, a condition which becomes dangerous during surgery or childbirth, when they cannot take conventional blood-thinning pills. The Scientific American (www.sciam.com, 2001) reported a study at the National Institute of Allergy and Infectious Disease (United States) in which two mouse strains were genetically engineered to produce large quantities of a malarial parasite surface protein from Plasmodium falciparum. The malaria vaccine secreted in their milk was able to contain the disease in monkeys vaccinated with the same. This study has now been extrapolated to target livestock as the source animals. GTC is currently

186

Latha Sabikhi

working on a project to develop the malaria vaccine from goat milk. It is understood that a liter of goat milk can contain up to 9 g of the transgenic protein and that eight goats can produce enough vaccine to inoculate 20 million people. The cost to produce a transgenic protein in goat milk can thus be 3–30 times cheaper than the current method using mammalian cell culture. In addition to these, GTC is also working on the development of a recombinant hAAT, a recombinant human albumin, and a CD137 antibody to stimulate the immune system as a potential treatment for solid tumors. PPL Therapeutics (Edinburgh, United Kingdom, and Blacksburg, VA) is working with rabbits and sheep to produce AAT, fibrinogen, and a lipase to treat pancreatic insufficiency in digesting dietary lipids. They are also attempting to engineer sheep to produce in milk a protein that reduces lung damage, thus providing hope to cystic fibrosis sufferers (Morgan, 2006). The Environment Risk Management Authority (ERMA) of New Zealand has admitted an application from AgResearch, New Zealand, for field testing of some cattle which are genetically modified with copies of genes or nucleic acids derived from humans or cattle (www.ermanz. govt.nz, 2001). The first of these involves inserting additional copies of two cattle milk casein genes to increase the protein content of milk (www. agresearch.co.nz, 2001). The second one is to disrupt the b-LG in order to decrease causes of milk allergies. The field tests also involve the insertion of a copy of the human myelin basic protein (MBP) gene in cattle. The protein, when secreted in their milk, may be purified and tested for its efficacy in the treatment of multiple sclerosis. Multiple sclerosis is a chronic demyelinating disease of the human central nervous system connected with clinical neurological signs of paralysis and histopathological changes. It is expected that secretion of MBP in cattle milk will allow the generation of large amounts of human MBP and will ultimately facilitate as a drug for the treatment of multiple sclerosis. Scientists at AgResearch have also patented a technology to produce bovine milk with enhanced quantities of immunoglobulin (IgA) antibodies (AgResearch Now, 2005). IgA is the dominant immunoglobulin in human milk and provides infants with essential protection against pathogens. IgA also contributes to adult health by helping to protect human mucosal surfaces like the stomach, intestinal tract, lungs, nose, ears, and eyes. One application of the technology being pursued is the prevention of human fungal infections including thrush, caused mostly by Candida albicans. These antibodies can also be tailored to protect against specific gut or oral diseases. In this context, the milk is undergoing trials at Otago’s school of dentistry to assess its potential for protecting teeth against decay-causing bacteria (New Zealand Herald, 2006). It is reported that about 10,000 cows

Designer Milk

187

immunized to protect against fungal infections would be sufficient to provide a protein powder for a mouthwash or lozenge that would form a barrier on the teeth, tongue, and cells lining the mouth and hence repel oral thrush. Pharming, NV (Leiden, the Netherlands) has obtained a US Patent for the production and composition of pharmaceuticals containing human a-glucosidase (Biotech Patent News, 2000). They have developed transgenic cattle capable of producing this enzyme in their milk. The company claims that this can help in treating Pompe’s disease, a hereditary, lethal muscle disorder that annually affects almost 5000–10,000 people living in the West. Genzyme Transgenics (Framingham, MA) have also succeeded in producing human a-glucosidase from transgenic animals (Dove, 2000). The Pharming Group maintains a transgenic dairy herd in the Dane County town of Vienna and claims that over the next decade their 13-strength herd will procreate 200 or more Holstein and Brown Swiss cows. The cattle will produce milk from which medicinal proteins can be turned into drugs that fight human illnesses such as hemophilia, hereditary angiodema, and gastrointestinal infections besides Pompe’s disease (Millard, 2000). Nexia, a US company, has engineered a herd of goats with milk containing an antidote for the nerve agents sarin and VX (Morgan, 2006). Experiments are under way at Hematech, South Dakota to create a transgenic cow (Transchromic, Tc cow) that will have an immune system that is half human and half bovine. Its cells will contain an additional artificial chromosome, with the genes for human antibodies (Morgan, 2006). The firm claims that when the Tc cow is immunized with an infectious agent, such as inactivated botulinum toxin, it will produce human polyclonal antibodies. These can then be purified from the cow’s blood and given to patients who cannot fight these infections owing to defects in their own immune systems. Hematech has already created Tc calves, which carry the human chromosome. The remaining work is to knock out the equivalent bovine antibody genes, so the Tc cow produces purely human antibodies in its blood. The firm anticipates that the technology will be more efficient and less expensive than present methods for generating human antibodies—cell culture—which can make only monoclonal (single variant) antibodies.

VIII. DESIGNER MILK FOR ANIMAL GROWTH AND HEALTH Vaccines, antibiotics, and the natural immune system of the animal have been used to cure diseases in the cow till recently. The use of gene transfer technologies to produce dairy cows that resist several infections and diseases is a novel development in biotechnology.

188

Latha Sabikhi

The volume of water drawn into milk by osmotic forces is related to lactose synthesis and therefore, a-LA is a deciding factor in the ultimate milk volume. Noble et al. (2002) used transgenic gilts expressing bovine a-LA in their milk to study whether the presence of the transgene influences lactation. They reported that lactose concentrations and milk production increased in the experimental group and that the piglets reared on the transgenic animals exhibited enhanced growth rates. In the area of pig husbandry, the trend of reduced lactation lengths increases the number of pigs born per sow annually, but also creates the need for sows that produce more milk in early lactation, to obtain maximal pig growth during the short lactation period. This becomes difficult in pigs as the maximum milk production does not occur until between the 21st and 28th days of lactation. Also, the higher number of pigs born per litter increases the demand for milk production. With a view to provide better energy intake through higher lactose levels in sow milk, Bleck et al. (1998b) attempted overexpression of bovine a-LA in porcine milk. The two lines of transgenic pigs so produced had 3.8% lactose in their milk as compared to 2.6% in the milk of control animals. The tracheal mucosa of the cow is the source of tracheal antimicrobial peptide (TAP), a member of the b-defensin family of antibiotic peptides. TAP protects the upper airway of bovines from infection. The limited availability of bovine TAP (bTAP) prompted researchers to create transgenic mice expressing bTAP (Yarus et al., 1996) in their milk. They purified bTAP from milk by acid precipitation, reverse-phase HPLC, and ion-exchange chromatography. This milk-derived bTAP had antimicrobial activity against E. coli. The work may herald the expression of the bTAP in bovine milk and its evaluation or use as an antibiotic in agriculture and medicine. Mastitis, an inflammatory reaction of the mammary gland, usually resulting from a microbial infection, is a widespread disease seen in cattle throughout the world. The major bacterial species that are responsible for bovine mastitis are Staphylococcus aureus, Streptococcus dysgalactiae, Streptococcus agalactiae, Streptococcus uberis, and E. coli. Of these, the first three cause a contagious route of transmission, whereas Streptococcus uberis and E. coli are considered to be environmental agents. While stringent disease control plans have eradicated Streptococcus dysgalactiae and Streptococcus agalactiae from many herds (Kerr and Wellnitz, 2003), controlling S. aureus has been difficult. Transgenic technology to control mastitis would involve the production of antibacterial enzymes by the mammary epithelial cells. These being degraded along with other milk proteins would not pose a health risk to the consumer like regular antibiotics. Initial studies on the application of transgenic technology involved the generation of mice that produced milk containing hLZ (Maga et al., 1994) and human lactoferrin

Designer Milk

189

(Platenburg et al., 1994). While the former has limited effect on mastitiscausing organisms, no conclusive study on the efficacy of the latter is reported. Lysostaphin is a potent peptidoglycan hydrolase naturally secreted by S. simulans. This antimicrobial protein has a potent antistaphylococcal activity and its secretion into milk offers considerable resistance to infection caused by S. aureus. Kerr et al. (2001) developed three lines of transgenic mice that produced varying levels of lysostaphin (100 mg/ml in two and 1 mg/ml in the third) in their milk. Transgenic as well as control mice were challenged with intramammary infusion of a strain of S. aureus. None of these glands from the transgenic animals were visibly infected. These glands contained less than 10% of the bacterial load observed in the heavily infected controls. Wall et al. (2005) engineered and introduced into Jersey cows, a transgene that includes the genetic code for producing lysostaphin. The gene for secreting lysostaphin was introduced from a nonpathogenic species of Staphylococcus that uses the protein to deter the pathogenic counterpart, S. aureus. The lysostaphin was secreted into milk of the modified animals to the order of 0.9–14 mg/ml. The milk destroyed the causative organisms in vitro. Ten nontransgenic cows that were given intramammary infusions of S. aureus showed positive signs of mastitis infection, whereas three transgenic cows under the same treatment remained unaffected. A small dose of 3 mg/ml of lysostaphin in milk was sufficient to provide protection against S. aureus. Although milk containing other natural antimicrobial proteins such as lactoferrin and LZ has been approved for human consumption, milk containing lysostaphin would need permission from regulatory agencies (Bliss, 2005). Coliforms account for about 40–50% of mastitis cases in the United States. One-tenth of the affected animals become useless for milk production and several die from shock induced by the bacterial toxin, or endotoxin, causing an estimated loss of $1.4 billion annually for farmers in terms of medical expenses and cost of milk that cannot be sold (McBride, 2002). Vaccines, while having limited success in reducing clinical symptoms, do not remove the organisms. Wang et al. (2002) report the identification and characterization of the gene for soluble CD14, which binds and neutralizes endotoxins responsible for mastitis. This soluble protein which can be found set in the membranes of white blood cells in cows was also discovered in cow milk. Just as the protein increases during coliform infections in humans and laboratory animals, it was shown to tone down the severe reaction of the bovines to coliform endotoxin as well as initiate a suitable response to the infiltrating bacteria. The gene was cloned and the recombinant bovine CD14 protein (rb-CD14) produced was evaluated. Intraperitoneal injection of rb-CD14 together with endotoxin reduced fatality in mice. Preliminary trials showed that intramammary injection

190

Latha Sabikhi

of soluble rb-CD14 is 100% effective in preventing mastitis by E. coli in lactating dairy cows. The workers have filed a patent application on rb-CD14 which promises both effective treatment for infected cows and prevention in future cows genetically engineered with the gene. They claim that the gene for CD14 can be designed and inserted into the modified animals so that they can produce the protein only in their mammary cells.

IX. ASSORTED ADVANTAGES The use of molecular biology to reduce the presence of pathogenic organisms in milk is a potentially advantageous prospect and has been reviewed in the previous section (Section VIII). It is clear that it might now be possible to produce specific antibodies in the mammary gland that are capable of preventing mastitis infection or those that aid in preventing human diseases. Thus, one can foresee antibodies against Salmonella, Listeria, or other pathogens that will produce safer milk products. Active recombinant immunoglobulin capable of neutralizing Coronavirus was produced in mouse milk (Castilla et al., 1998). Increasing the concentration of IgA receptors in mammary cell may potentially lead to the accumulation of the protective antibodies in milk (De Groot et al., 1999). While recombinant immunoglobulins have been expressed in mammalian transgenic milk (Gavilondo et al., 2000), a calf with a gene that promotes the growth of red cells in humans has been produced by transgenesis (www.publicscan.fi). Research is also under way to manufacture milk through transgenesis for treatment of diseases such as PKU, hereditary emphysema, and cystic fibrosis (Margawati, 2003). In an interesting application of transgenic technology combining sericulture and dairying, goats that produce spider silk in milk have been engineered (Dove, 2000). Spider silk is made of protein-based polymer filaments or threads secreted by specialized epithelial cells. These fibers are flexible and lightweight and have extraordinary strength and toughness comparable to those of synthetic high-performance fibers. Spider silk is, thus, one of the strongest and most versatile naturally occurring materials in nature (Samson, 2004). Although it appears delicate and is one-tenth the width of a human hair, it is stronger than steel and stretches more than nylon (www.sciam.com, 2002). Several attempts have been made to synthesize spider silk for industrial and medical applications. Unfortunately, the high shearing forces in conventional fermentations cause the spidroin protein to aggregate, making it useless for manufacturing fibers. So, these spider genes were introduced into the cells of lactating goats and they secreted silk in tiny strands along with their milk. These polymer strands could be woven into threads after extracting them

Designer Milk

191

from the milk and used for applications such as military uniforms, medical microsutures, and tennis racket strings (Anonymous, 2002). Nexia, a US company, has developed a strain of fast-maturing Breed Early Lactate Early (BELE) transgenic goats that secretes the spider silk in milk (Dove, 2000). Nexia is already working with the US army to develop bulletproof vests and surgical suture material from spider silk. The team produced soluble recombinant spider silk proteins with molecular masses of 60–140 kDa (Lazaris et al., 2002). They were able to wet spin the silk monofilaments derived from a concentrated aqueous solution of soluble recombinant spider silk protein under conditions of low shear and coagulation. The spun fibers were water-insoluble with diameters ranging from 10 to 40 mm and exhibited toughness and modulus values comparable to those of native dragline silks but had lower tenacity. They anticipate that the manufacturing processes for these products would be more environmental friendly than the production of conventional plastics.

X. THE FUTURE Novel research in genetic engineering attempts to alter and control the genetic makeup of animals in several ways. Some of these attempts may target only individual animals and not generations together. One example is that of somatic cell therapy, wherein specific cells of an individual animal are modified to produce desired characteristics without changing its genetic characters. On the other hand, transgenesis involves the modification of the genetic line wherein the altered traits can be inherited by the progeny. Both techniques are useful in the enhancement of animal productivity, faster growth, improved feed conversion, better quality of animal products, and improved resistance to diseases. However, the technology presents several hurdles to its ultimate development. The most daunting challenge in producing transgenic animals is producing healthy adults. Cattle, for example, often suffer at birth from immature respiratory systems. They are also frequently born late, leading to problems with delivery. The survival rate in cloning attempts also is low. Generally only 1–5% of embryos typically survive to term (Kling, 2001). Dolly, the sheep and the first animal to be cloned from an adult cell in 1996, was the lone success out of the 277 attempts (Taylor, 2006). Successful transgenic studies have generated mice that produce milk with 33% more total solids (40–50% TS) and 17% less lactose than normal mouse milk. As the increase in the total solids is associated with a decrease in total milk volume, the net result is the same quantity of fat and protein being produced in a lesser total milk volume. If this technology could be extrapolated in dairy animals, milk that contains

192

Latha Sabikhi

6.5% protein, 7% fat, 2.5% lactose, and 50% less water is not an improbable accomplishment. The advantages in terms of animal health would include less stress on the cow and on her udder since she would be producing onehalf her normal volume of milk and decrease in mastitis owing to less lactose availability for the causative organisms. The processing industry would gain in terms of (1) skim milk with twice protein content and half the lactose content of normal milk, (2) easier to produce low lactose or lactose-free dairy products, (3) better product yields due to concentration, (4) reduction in total whey output because of low milk volume and lactose content, and (5) direct economic benefits in terms of 50% reduction in the cost of milk transportation owing to reduced volumes. The future of biotechnologically derived foods is, however, uncertain even after three decades of positive results. Improved and sophisticated equipment for milk processing may be more acceptable to consumers than genetic and transgenic technology for altering milk composition. Consumer acceptance will always contribute to and guide decisions about biotechnological manipulations aimed at increasing milk production or altering milk composition. Various ethical, legal, and social aspects of biotechnological research would need to be addressed in the current economic and social climate before designer transgenic herds thrive just as their counterparts in organic herds do. Animal welfare, demonstrable and sustained safety of the product, improved health properties of the product, and enhanced profitability as compared with conventional practices would be the key factors that would eventually decide the future of designer foods. The natural human tendency to resist change, especially those that trouble their feeling and instincts, is an important point to consider. As all consequences of biological research involving animal studies could be classified under this category, there is bound to be tremendous resistance to topics such as transgenic technology. The perception among various groups of human beings is also different. While farmers interpret animal welfare in terms of health and production, consumers interpret it in terms of freedom to move and fulfill natural desires. That a human being controls these natural desires of animals itself is a reason for dissent. The topics of genetic engineering and transgenic technology have been under debate ever since the idea was first conceived. In this context, animal bio- and moral ethics are new keywords that have found entry into the dictionary of animal agriculture. The activists argue that the transformation of animals according to human needs smudges the clear demarcation between man and animal in ethical, moral, and biological perspectives. Thus, the moral principles, which were behind the treatment of animals in the past, are no more valid or adequate. Researchers while considering human requirements also need to find new methods of dealing with animals with their interests, suffering, and welfare in mind.

Designer Milk

193

REFERENCES AbuGhazaleh, A.A., Schingoethe, D.J., Hippen, A.R., and Kalscheur, K.F. (2003). Milk conjugated linoleic acid response to fish oil supplementation of diets differing in fatty acid profiles. J. Dairy Sci. 86, 944–953. AgResearch Now (2005). Your pasture, our pasture. 30% more milk. Interested? pp. 1–20 www.agresearch.cri.nz/publications/nowJune, Issue 3. Aigster, A., Sims, C., Staples, C., Schmidt, R., and O’keefe, S.F. (2000). Comparison of cheeses made from milk having normal and high oleic fatty acid compositions. J. Food Sci. 65, 920–924. Anonymous (2002). The scientific American 50: Celebrating the year’s top technology leaders. Sci. Am. 12, 48. Ashes, J.R., Gulati, S.K., and Scott, T.W. (1997). Potential to alter content and composition of milk fat through nutrition. J. Dairy Sci. 80, 2204–2212. Bailey, P. (2001). Why milk? UC Davis Magazine Online. Spring. Vol. 18. No. 3. www.ucdmag. ucdavis.edu/sp01/feature_1.html. Bauman, D.E., and Perfield, J.W., II (2002). CLA: The milk fat wonder. Pro-dairy North East Dairy Bus. 6, 21. www.dairybusiness.com. Bauman, D.E., Corl, B.A., Baumgard, L.H., and Griinari, J.M. (2001). Conjugated linoleic acid (CLA) and the dairy cow. In ‘‘Recent Advances in Animal Nutrition’’ (P. C. Garnsworthy and J. Wiseman, eds.), pp. 221–250. Nottingham University Press, Nottingham, UK. Baumgard, L.H., Corl, B.A., Dwyer, D.A., and Bauman, D.E. (2000). Identification of the conjugated linoleic acid isomer that inhibits milk fat synthesis. Am. J. Physiol. Regul. Integr. Comp. Physiol. 278, R179–R184. Beaulieu, A.D., and Drackley, J.K. (2004). Can milk fat be beneficial to your health? www. traill.uiuc.edu/dairynet/paperDisplay?ContentID ¼ 168. Bell, J.A., and Kennelly, J.J. (2001). Conjugated Linoleic Acid enriched milk: A designer milk with potential. Adv. Dairy Technol. 13, 213–228. Biotech Patent News (2000). September Issue. www.allbusiness.com. Bleck, G.T., and Bremel, R.D. (1994). Variation in expression of a bovine a-lactalbumin transgene in milk of transgenic mice. J. Dairy Sci. 77, 1897–1904. Bleck, G.T., and Wheeler, M.W. (1998). Increasing total solids and reducing lactose and water content of milk. www.traill.uiuc.edu. Bleck, G.T., Choi, B.K., Wheeler, M.W., and Jime´nez-Flores, R. (1998a). Modification of bovine beta-casein to improve the characteristics and manufacturing properties of cow’s milk. www.traill.uiuc.edu/dairynet/paperDisplay?ContentID ¼ 245. Bleck, G.T., White, B.R., Miller, D.J., and Wheeler, M.W. (1998b). Production of bovine a-lactalbumin in the milk of transgenic pigs. J. Anim. Sci. 76, 3072–3078. Bliss, R.M. (2005). Transgenic cows resist mastitis-causing bacteria. www.ars.usda.gov/is/ pr/2005/050404.htm. Bremel, R.D., Yom, H.C., and Bleck, G.T. (1989). Alteration of milk composition using molecular genetics. J. Dairy Sci. 72, 2826–2833. Brink, M.F., Bishop, M.D., and Pieper, F.R. (2000). Developing efficient strategies for the generation of transgenic cattle which produce biopharmaceuticals in milk. Theriogenology 53, 139–148. Brophy, B., Smolenski, G., Wheeler, T., Wells, D., L’Huillier, P., and Laible, G. (2003). Cloned transgenic cattle produce milk with higher levels of b-casein and k-casein. Nat. Biotechnol. 21, 157–162. Carmona, R.H. (2006). Improving bone health. J. Am. Diet. Assoc. 106, 651. Carver, A.S., Dalrymple, M.A., Wright, G., Cottom, D.S., Reeves, D.B., Gibson, Y.H., Keenan, J.L., Barrass, J.D., Scott, A.R., Colman, A., and Garner, I. (1993). Transgenic livestock as bioreactors: Stable expression of human alpha-1-antitrypsin by a flock of sheep. Biotechnology 11, 1263–1270.

194

Latha Sabikhi

Castilla, J., Pintado, B., Sola, I., Sanchez-Morgado, J.M., and Enjuanes, L. (1998). Engineering passive immunity in transgenic mice secreting virus-neutralizing antibodies in milk. Nat. Biotechnol. 16, 349–354. CDRF (2004). Investigators aim to improve milk composition for increased utilization. California Dairy Research Foundation. Preharvest Research. www.cdrf.org/newsletter. Cheek, M. (2006). Go organic. www.dailyrecord.co.uk/news. Chouinard, P.Y., Girard, V., and Brisson, G.J. (1998). Fatty acid profile and physical properties of milk fat from cows fed calcium salts of fatty acids with varying unsaturation. J. Dairy Sci. 81, 471–481. Chouinard, P.Y., Corneau, L., Barbano, D.M., Metzger, L.E., and Bauman, D.E. (1999). Conjugated linoleic acids alter milk fatty acid composition and inhibit milk fat secretion in dairy cows. J. Nutr. 129, 1579–1584. Coleman, A. (1996). Production of proteins in the milk of transgenic livestock: Problems, solutions and successes. Am. J. Clin. Nutr. 63, 639S–645S. Corazza, G.R., Benati, G., di Sario, A., Tarozzi, C., Strocchi, A., Passeri, M., Gasbarrini, G., and DiSario, A. (1995). Lactose intolerance and bone mass in postmenopausal Italian women. Br. J. Nutr. 73, 479–487. CRC Factsheet (2006). Research in Genetic Modification. www.dairycrc.com. pp. 1–2. Crittenden, R.G., and Bennett, L.E. (2005). Cow’s milk allergy: A complex disorder. J. Am. Coll. Nutr. 24, 582s–591s. CSIRO (1999). Healthy butter spreads better. Media Release. www.csiro.au. De Groot, N., van Kuik-Romeijn, P., Lee, S.H., and De Boer, H. (1999). Over-expression of the murine polymeric immunoglobulin receptor gene in the mammary gland of transgenic mice. Transgenic Res. 8, 125–135. Dhiman, T.R., Anand, G.R., Satter, L.D., and Pariza, M.W. (1999). Conjugated linoleic acid content of milk from cows fed different diets. J. Dairy Sci. 82, 2146–2156. Dhiman, T.R., Satter, L.D., Pariza, M.W., Galii, M.P., Albright, K., and Tolosa, M.X. (2000). Conjugated linoleic acid (CLA) content of milk from cows offered diets rich in linoleic and linolenic acid. J. Dairy Sci. 83, 1016–1027. Dove, A. (2000). Milking the genome for profit. Nat. Biotechnol. 18, 1045–1048. Ebert, K.M., Selgrath, J.P., DiTullio, P., Denman, J., Smith, T.E., Memon, M.A., Schindler, J.E., Monastersky, G.M., Vitale, J.A., and Gordon, K. (1991). Transgenic production of a variant of human tissue-type plasminogen activator in goat milk: Generation of transgenic goats and analysis of expression. Biotechnology 9, 835–838. Eynard, A.R., and Lopez, C.B. (2003). Conjugated linoleic acid (CLA) versus saturated fats/ cholesterol: Their proportion in fatty and lean meats may affect the risk of developing colon cancer. Lipids Health Dis. 2, 6. Fallon, S., and Enig, M.G. (2000). ‘‘Nourishing Traditions: The Cookbook that Challenges Politically Correct Nutrition and the Diet Dictocrats’’. New Trends Publishing Inc., Washington, USA. Gavilondo, J.V., Larrick, J.W., and Borrebaeck, C.A.K. (2000). Human antibodies for therapy. Immunologist 8, 58–65. Goel, R. (2005). And now designer milk! www.tribuneindia.com/2005/20050429/science.htm. Gorman, S.D., and Clark, M.R. (1990). Humanisation of monoclonal antibodies for therapy. Semin. Immunol. 2, 457–466. Grummer, R.R. (1991). Effect of feed on the composition of milk fat. J. Dairy Sci. 74, 3244–3257. Gutierrez-Adan, A., Maga, E.A., Meade, H., Shoemaker, C.F., Medrano, J.F., Anderson, G.B., and Murray, J.D. (1996). Alterations of the physical characteristics of milk from transgenic mice producing bovine k-casein. J. Dairy Sci. 79, 791–799. Halken, S. (2004). Prevention of allergic disease in childhood: Clinical and epidemiological aspects of primary and secondary allergy prevention. Pediatr. Allergy Immunol. 15(Suppl. 16), 4–5, 9–32.

Designer Milk

195

Hennighausen, L., Ruis, L., and Wall, R. (1990). Transgenic animals – Production of foreign proteins in milk. Curr. Opin. Biotechnol. 1, 74–78. Host, A. (2002). Frequency of cow’s milk allergy in childhood. Ann. Allergy Asthma Immunol. 89(6, Suppl. 1), 33–37. Houdebine, L.M. (1995). The production of pharmaceutical proteins from the milk of transgenic animals. Reprod. Nutr. Dev. 35, 609–617. Jeng, S.-Y., Bleck, G.T., Wheeler, M.W., and Jime´nez-Flores, R. (1997). Characterization and partial purification of bovine a-lactalbumin and b-casein produced in milk of transgenic mice. J. Dairy Sci. 80, 3167–3175. Johnson, A.O., Semenya, J.G., Buchowski, M.S., Enwonwu, C.O., and Scrimshaw, N.S. (1993). Correlation of lactose maldigestion, lactose intolerance, and milk intolerance. Am. J. Clin. Nutr. 57, 399–401. Johnston, I. (2006). What did Dolly do for us? The Scotsman. Science & Technology. July 5. news.scotsman.com. Jost, B., Vilotte, J.-L., Duluc, I., Rodeau, J.-L., and Freund, J.-N. (1999). Production of lowlactose milk by ectopic expression of intestinal lactase in the mouse mammary gland. Nat. Biotechnol. 17, 160–164. Kao, B.T., Lewis, K.A., DePeters, E.J., and van Eenennaam, A.L. (2006). Endogenous production and elevated levels of long-chain o-3 fatty acids in the milk of transgenic mice. J. Dairy Sci. 89, 3195–3201. Karatzas, C.N. (2003). Designer milk from transgenic clones. Nat. Biotechnol. 21, 138–139. Karatzas, C.N., and Turner, J.D. (1997). Toward altering milk composition by genetic manipulation: Current status and challenges. J. Dairy Sci. 80, 2225–2232. Kelly, M.L., and Bauman, D.E. (1996). Conjugated linoleic acid: A potent anticarcinogen found in milk fat. In ‘‘Proceedings of the Cornell Nutrition Conference for Feed Manufacturers,’’ pp. 68–74. Cornell University, Ithaca, NY, USA. Kerr, D.E., and Wellnitz, O. (2003). Mammary expression of new genes to combat mastitis. J. Anim. Sci. 81(Suppl. 3), 38–47. Kerr, D.E., Plaut, K., Bramley, A.J., Williamson, C.M., Lax, A.J., Moore, K., Wells, K.D., and Wall, R.J. (2001). Lysostaphin expression in mammary glands confers protection against staphylococcal infection in transgenic mice. Nat. Biotechnol. 19, 66–70. Kling, J. (2001). Creating healthy, long-living cloned animals. Scientist 15, 26. Kris-Etherton, P.M., Harris, W.S., and Appel, L.J. (2002). Fish consumption, fish oil, omega-3 fatty acids, and cardiovascular disease. American Heart Association. Scientific Statement. Circulation 106, 2747–2757. Lacefield, G.D. (2003). Designer milk. Forage News. www.uky.edu. Lazaris, A., Arcidiacono, S., Huang, Y., Zhou, J.-F., Duguay, F., Chretien, N., Welsh, E.A., Soares, J.W., and Karatzas, C.N. (2002). Spider silk fibres spun from soluble recombinant silk produced in mammalian cells. Science 295, 472–476. Lee, J.H., Melton, S.L., Waller, J.C., and Saxton, A.M. (2004). Modification of physicochemical characteristics of goat milk fat by feeding protected high oleic sunflower oil supplements. J. Food Sci. 69, C280–C286. Lightfield, K.K., Bear, R.J., and Schingoethe, D.J. (1993). Characteristics of milk and Cheddar cheese from cows fed unsaturated dietary fat. J. Dairy Sci. 76, 1221–1232. Little, M., Kipriyanov., S.M., Gall, F.Le, and Moldenhauer, G. (2000). Of mice and men: Hybridoma and recombinant antibodies. Immunol. Today 21, 364–370. Lonberg, N. (2005). Human antibodies from transgenic animals. Nat. Biotechnol. 23, 1117–1125. Lonnerdal, B. (1996). Recombinant human milk proteins- an opportunity and a challenge. Am. J. Clin. Nutr. 63, 622S–626S. Maga, E.A., Anderson, G.B., Huang, M.C., and Murray, J.D. (1994). Expression of human lysozyme mRNA in the mammary gland of transgenic mice. Transgenic Res. 3, 36–42.

196

Latha Sabikhi

Maga, E.A., Anderson, G.B., and Murray, J.D. (1995). The effect of mammary gland expression of human lysozyme on the properties of milk from transgenic mice. J. Dairy Sci. 78, 2645–2652. Maga, E.A., Shoemaker, C.F., Rowe, J.D., BonDurant, R.H., Anderson, G.B., and Murray, J.D. (2006). Production and processing of milk from transgenic goats expressing human lysozyme in the mammary gland. J. Dairy Sci. 89, 518–524. Margawati, E.T. (2003). Trangenic animals: Their benefits to human welfare. www. actionscience.org. Mason, S. (2001). Manipulating milk composition. ProLivestock. www.afns.ualberta.ca. McBride, J. (2002). An udder solution for Bossie’s woes. Agricultural Research. Review Publications June. pp. 18–19. www.ars.usda.gov/research/projects. Middaugh, R.P., Bear, R.J., Casper, D.P., Schingoethe, D.J., and Seas, S.W. (1988). Characteristics of milk and butter from cows fed sunflower seeds. J. Dairy Sci. 71, 3179–3187. Millard, P. (2000). Vienna Pharms calves will produce medicinal milk. Bus J. Milwaukee July 3. www.milwaukee.bizjournals.com/milwaukee/stories/2000/07/03/focus3.html. Morgan, J. (2006). The clone arrangers. The Herald. Science & Technology. June 27. www. theherald.co.uk/features/64780.html. Murthy, G.L.N., and Kanawjia, S.K. (2002). Designer milk: Nutritional and technological significance. Indian Dairyman 54, 49–58. Muller, L.D., and Delahoy, J.E. (1988). Conjugated linoleic acid (CLA): Implications for animal production and human health. DAS 04–88. pp. 1–8. www.das.psu.edu/ dairynutrition/documents. National Dairy Council (2000). Conjugated linoleic acid (CLA) content of milk and other dairy foods. Newer Knowledge of Dairy Foods. www.nationaldairycouncil.org. New Zealand Herald (2006). Designer milk to protect against tooth decay. June 26. www. nzherald.co.nz/section/story. Noble, M.S., Rodriguez-Zas, S., Cook, J.B., Bleck, G.T., Hurley, W.L., and Wheeler, M.B. (2002). Lactational performance of first-parity transgenic gilts expressing bovine a-lactalbumin in their milk. J. Anim. Sci. 80, 1090–1096. Ntambi, J.M., Choi, Y., and Kim, Y.C. Regulation of stearoyl-CoA desaturase by conjugated linoleic acid. In ‘‘Advances in Conjugated Linoleic Acid Research’’ (M. P. Yurawecz, M. M. Mossoba, J. K. G. Kramer, M. W. Pariza, and G. Nelson, eds.), Vol. 1, pp. 340–347. Nuble, C.J. (2006). Cow’s milk allergy and babies. www.market-day.net. Nuijens, J.H., vanBerkel, P.H.C., Geerts, M.E.J., Hartevelt, P.P., deBoer, H.A., vanVeen, H.A., and Pieper, F.R. (1997). Characterization of recombinant human lactoferrin secreted in milk of transgenic mice. J. Biol. Chem. 272, 8802–8807. O’Donnell, J.A. (1989). Milk fat technologies and markets: A summary of the Wisconsin Milk Marketing Board 1988 Milk Fat Roundtable. J. Dairy Sci. 72, 3109–3115. O’Donnell, J.A. (1993). Future of milk fat modification by production or processing: Integration of nutrition, food science and animal science. J. Dairy Sci. 76, 1797–1801. Peterson, D.G., Baumgard, L.H., and Bauman, D.E. (2002). Short communication: Milk fat response to low doses of trans-10, cis-12 conjugated linoleic acid (CLA). J. Dairy Sci. 85, 1764–1766. Pettus, P. (2006). Straight from the sheep’s mouth. The New York Sun. June 19. www.nysun. com/article/34616. Platenburg, G.J., Kootwijk, E.P., Kooiman, P.M., Woloshuk, S.L., Nuijens, J.H., Krimpenfort, P.J., Pieper, F.R., de Boer, H.A., and Strijker, R. (1994). Expression of human lactoferrin in milk of transgenic mice. Transgenic Res. 3, 99–108. Prunkard, D., Cottingham, I., Garner, I., Bruce, S., Dalrymple, M., Lasser, G., Bishop, P., and Foster, D. (1996). High-level expression of recombinant human fibrinogen in the milk of transgenic mice. Nat. Biotechnol. 14, 867–871.

Designer Milk

197

Pszczola, D.E., Katz, F., and Giese, J. (2000). Research trends in healthful foods. Food Technol. 54, 45–52. Reh, W.A., Maga, E.A., Collette, N.M.B., Moyer, A., Conrad-Brink, J.S., Taylor, S.J., DePeters, E.J., Oppenheim, S., Rowe, J.D., BonDurant, R.H., Anderson, G.B., and Murray, J.D. (2004). Using a stearoyl-CoA desaturase transgene to alter milk fatty acid composition. J. Dairy Sci. 87, 3510–3514. Roberts, A.K., Chierici, R., Sawatzki, G., Hill, M.J., Volpato, S., and Vigi, V. (1992). Supplementation of an adapted formula with bovine lactoferrin: 1. Effect on the infant faecal flora. Acta Paediatr. 81, 119–124. Robinson, J. (2003). Super healthy milk. www.eatwild.com. Samson, A.E.S. (2004). Gender equality and new technologies. Fact Sheet No.2. The Association for Women’s Rights in Development. Agric. Biotechnol. September, 1–4. Schaffer, S.M., Tatini, S.R., and Bear, R.J. (1995). Microbiological safety of blue and cheddar cheese containing naturally modified milk fat. J. Food Prot. 58, 132–138. Shani, M., Barash, I., Nathan, M., Ricca, G., Searfoss, G.H., Dekel, I., Faerman, A., Givol, D., and Hurwitz, D.R. (1992). Expression of human serum albumin in the milk of transgenic mice. Transgenic Res. 1, 195–208. Simons, J.P., McClenghan, M., and Clark, A.J. (1987). Alteration of the quality of milk by expression of sheep beta-lactoglobulin in transgenic mice. Nature 328, 530–532. Simopoulos, A.P. (1999). Essential fatty acids in health and chronic disease. Am. J. Clin. Nutr. 70(Suppl. 3), 560S–569S. St. John, J.C. (2002). The transmission of mitochondrial DNA following assisted reproductive techniques. Theriogenology 57, 109–123. Stanton, C. (2000). ‘‘CLA: A Health-Promoting Component of Animal and Milk Fat. End of Project Report.’’ The Dairy Products Research Centre Moorepark, Fermoy, Co. Cork. DPRC No. 26, pp. 1–13. Subramanian, S. (2004). Transgenic therapeutics: Medicines in milk. Biotechnology and Society: Part 22. www.chennaionline.com. Suraokar, M., and Bradley, A. (2000). Targeting sheep. Nature 409, 1004–1005. Taylor, A. (2006). Dolly and me: A true story. Interview with cloning scientist Ian Wilmut. The Sunday Herald July 9. www.sundayherald.com. Thomas, J. (2001). The understatement of the century: Monoclonal antibodies in the laboratory and the clinic. Harv. Sci. Rev. Spring, 9–12. Tomlinson, A. (2003). Nutrition for the brain: Promising substances in the enhancement of brain function. S. Afr. J. Nat. Med. Issue 4. Nov 5. www.naturalmedicine.co.za. Tsiplakou, E., Mountzouris, K.C., and Zervas, G. (2006). Concentration of conjugated linoleic acid in grazing sheep and goat milk fat. Livestock Sci. 103, 74–84. Tyagi, A.K., Kewalramani, N., Kaur, H., Singhal, K.K., Kaul, G., Kanawjia, S.K., and Dhiman, T.R. (2004). Dietary manipulation to enhance linoleic acid (CLA) in buffalo milk and milk products. Annual Report 2003–2004. National Dairy Research Institute, India, p. 15. van Berkel, P.H., Welling, M.M., Geerts, M., van Veen, H.A., Ravensbergen, B., Salaheddine, M., Pauwels, E.K., Pieper, F., Nuijens, J.H., and Nibbering, P.H. (2002). Large scale production of recombinant human lactoferrin in the milk of transgenic cows. Nat. Biotechnol. 20, 484–487. Vesa, T.H. (1999). Many factors affect symptoms of lactose intolerance. Food Rev. Int. 15, 235–247. Vilotte, J.-L. (2002). Lowering the milk lactose content in vivo: Potential interests, strategies and physiological consequences. Reprod. Nutr. Dev. 42, 127–132. Wakabayashi, H., Yamauchi, K., and Takase, M. (2006). Lactoferrin research, technology and applications. Int. Dairy J. 16, 1241–1251. Wal, J.M. (1998). Strategies for assessment and identification of allergenicity in (novel) foods. Int. Dairy J. 8, 413–423.

198

Latha Sabikhi

Wal, J.M. (2002). Cow’s milk proteins/allergens. Ann. Allergy Asthma Immunol. 89, 3–10. Wall, R.J., Kerr, D.E., and Bondioli, K.R. (1997). Transgenic dairy cattle: Genetic engineering on a large scale. J. Dairy Sci. 80, 2213–2224. Wall, R.J., Powell, A.M., Paape, M.J., Kerr, D.E., Bannerman, D.D., Pursel, V.G., Wells, K.D., Talbot, N., and Hawk, H.W. (2005). Genetically enhanced cows resist intramammary Staphylococcus aureus infection. Nat. Biotechnol. 23, 445–451. Wang, Y., Zarlenga, D.S., Paape, M.J., and Dahl, G.E. (2002). Recombinant bovine CD14 sensitizes the mammary gland to lipopolysaccharide. Vet. Immunol. Immunopathol. 86, 15–124. Whitelaw, B. (1999). Toward designer milk. Nat. Biotechnol. 17, 135–136. Willumsen, R. (2006). Omega-3 in food and beverages. Agro-Food Industry Hi-Tec 17, 6–7. Wright, G., Carver, A., Cottom, D., Reeves, D., Scott, A., Simons, P., Wilmut, L., Farner, I., and Colman, A. (1991). High level expression of active human a1-antitrypsin in the milk of transgenic sheep. Biotechnology 9, 830–834. Yang, X.-D., Jia, X.-C., Corvalan, J.R.F., Wang, P., and Davis, C.G. (2001). Development of ABX-EGF, a fully human anti-EGF receptor monoclonal antibody, for cancer therapy. Crit. Rev. ‘ncol. Hematol. 38, 17–23. Yarus, S., Rosen, J.M., Cole, A.M., and Diamond, G. (1996). Production of active bovine tracheal antimicrobial peptide in milk of transgenic mice. Proc. Natl. Acad. Sci. USA 93, 14118–14121. Yu, Z., Meng, Q., Yu, H., Fan, B., Yu, S., Fei, J., Wang, L., Dai, Y., and Li, N. (2006). Expression and bioactivity of recombinant human lysozyme in the milk of transgenic mice. J. Dairy Sci. 89, 2911–2918. Zuelke, K.A. (1998). Transgenic modification of cows milk for value-added processing. Reprod. Fertil. Dev. 10, 671–676.

CHAPTER

6 The Sweet Taste Receptor: A Single Receptor with Multiple Sites and Modes of Interaction Pierandrea Temussi*

Contents

200 202 202 205 209 210

I. Introduction II. Indirect Mapping of Active Sites A. Small molecular weight sweet molecules B. Early structure–activity studies III. Sweet Macromolecules A. Characterization of natural sweet proteins B. Interaction of sweet proteins with the sweet receptor IV. The Sweet Taste Receptor A. Molecular biology of taste receptors B. Computer-generated models of the sweet taste receptor V. Mechanisms of Interaction A. The ‘‘wedge model’’ mechanism for sweet proteins B. Interaction of small sweeteners with the sweet receptor C. Multiple binding sites VI. Beyond the Sweet Receptor Acknowledgments References

217 218 218 220 221 221 224 228 231 232 232

* Dipartimento di Chimica, Universita` di Napoli Federico II, Via Cinthia, Napoli I-80126, Italy; and National Institute for Medical Research, The Ridgeway, London NW7 1AA, United Kingdom Advances in Food and Nutrition Research, Volume 53 ISSN 1043-4526, DOI: 10.1016/S1043-4526(07)53006-8

#

2007 Elsevier Inc. All rights reserved.

199

200

Abstract

Pierandrea Temussi

Elucidation of the molecular bases of sweet taste is very important not only for its intrinsic biological significance but also for the design of new artificial sweeteners. Up to few years ago design was complicated by the common belief that different classes of sweet compounds, notably sweet proteins, might interact with different receptors altogether. The recent identification and functional expression of the receptor for sweet taste have shown that there is but one receptor, drastically changing our approach to the development of new sweeteners. The explanation of how the sweet receptor can bind several different classes of molecules is that rather than multiple receptors there are, apparently, multiple sites on the single sweet taste receptor. In this chapter, the mechanisms of interaction of small and macromolecular sweet molecules will be examined, with particular emphasis on sweet proteins. Systematic homology modeling yields reliable models of all possible heterodimers of the human T1R2 and T1R3 sequences with the closed (A) and open (B) conformations of one of the metabotropic glutamate receptors (mGluR1), used as template. The most important result of these studies is the ‘‘wedge model,’’ the first explanation of the taste of sweet proteins. In addition, it was shown that simultaneous binding to the A and B sites is not possible with two large sweeteners but is possible with a small molecule in site A and a large one in site B. This observation accounted for the first time for the peculiar phenomenon of synergy between some sweeteners.

I. INTRODUCTION Taste plays a key role in the selection of food. The gustatory system of all animals is primarily involved in feeding behavior, allowing them to detect useful foods and avoid toxic substances. For instance, plant-feeding insects often rapidly reject foods containing toxic plant compounds (Glendinning, 1994, 1996; Wang et al., 2004). Although there is not unanimous consensus on a sharp classification of tastes (Delwiche, 1996), the existence of five different tastes, that is sweet, bitter, salty, sour, and umami, is acknowledged by a vast majority of scientists (Lindemann, 2001). Sweet taste plays a central role for humans since most people respond positively to the sensation of sweetness, with a propensity for sweet foods that can be traced back to early life. Sweetness is also an important medical issue because there is an increasing number of people affected by diseases, like diabetes, hyperlipemia, caries, that are more or less directly linked to the secondary effects of sugar intake. New knowledge about the molecular basis of taste may suggest strategies to overcome diet-induced diseases (Mennella et al., 2005). The design of safe low-calorie sweeteners is particularly important in this context.

Active Sites of the Sweet Receptor

201

Most people associate sweet taste with sugars, but it is not generally true that sweet molecules are sugars. Several hundreds of synthetic and natural sweeteners were found either by serendipity or by targeted research. In the past, organic chemists used to taste most of the compounds they synthesized and many substances turned out to be sweet (Moncrieff, 1967). The chemical constitution of sweet molecules varies widely from sugars to amino acids, peptides, proteins, olefinic alcohols, nitroanilines, saccharin, chloroform, and many other organic compounds. In addition to diversity in chemical constitution, sweet molecules can have drastically different dimensions, the extreme example being afforded by the existence of a few very sweet proteins (Morris, 1976). The molecular volume of a typical small molecular weight sweetener, like aspartame, can be esti˚ 3 whereas that of thaumatin, a well-characterized sweet mated at 265 A ˚ 3. Such a dramatic difference poses a big protein, is of the order of 27,000 A challenge if one tries to reconcile the sweetness–activity relationship of the two classes of molecules. We shall see that the interpretation of the mechanism of action of sweet proteins plays a crucial role for the understanding of structure–activity of sweet molecules in general. When nothing was known about taste receptors, the structure–activity relationship of sweet molecules was studied mainly using an indirect mapping of the active site of the receptor, by comparing the structures and the activities of the largest possible number of sweet compounds. This approach amounted to the identification of an ideal sweetener whose shape and electronic properties reflected the nature and topological arrangement of glucophores in the majority of sweet molecules. Indirect mapping led to the development of different models of active site (Goodman et al., 1987; Iwamura, 1981; Kier, 1972; Shallenberger and Acree, 1967; Temussi et al., 1978, 1984, 1991; Tinti and Nofre, 1991), consistent with the structure–activity of most small molecular weight sweeteners but hardly explaining the enormous sweetening power of sweet proteins. The molecular biology of taste has been studied less than that of other stimuli, to the extent that Lindemann (1996) defined taste as ‘‘the Cinderella of senses’’ since its transduction mechanism was, at that time, the least well understood, but the cloning of several likely taste receptors rekindled the interest in this stimulus (Firestein, 2000) and opened brand new perspective also for the rational design of artificial sweeteners. In 2001, a major experimental effort led many groups to the identification of the sweet receptor (Bachmanov et al., 2001; Kitagawa et al., 2001; Li et al., 2001; Max et al., 2001; Montmayeur et al., 2001; Nelson et al., 2001; Sainz et al., 2001). It was found that the sweet taste receptor is a special type of G-protein–coupled receptor (GPCR), one of the so-called metabotropic or class C GPCRs. This class of GPCRs, in addition to the seven-helix transmembrane domain (7TM), has a large extracellular domain, called Venus flytrap domain (VFTD), containing the active site

202

Pierandrea Temussi

for typical ligands (Pin et al., 2003). The knowledge of the receptor provided immediately a better understanding of the molecular bases of sweet taste but left many questions unanswered. Among these, one that arises when comparing sweet molecules is whether sweet compounds of size so different as sweet proteins can interact with the same receptors as small molecular weight compounds. Since it was shown that small and large molecular weight sweet molecules do interact with the same T1R2– T1R3 receptor (Li et al., 2002), it was necessary to understand whether they use the same mechanism. We shall show how the answers to many unresolved issues on sweet taste came from modeling studies that show that the sweet taste receptor has multiple active sites. In the following sections, we shall give an overview of the history of sweet molecules with a particular emphasis on sweet proteins and the unique mechanism of their interaction with the T1R2–T1R3 receptor.

II. INDIRECT MAPPING OF ACTIVE SITES A. Small molecular weight sweet molecules Sugar, chemically better described as sucrose (a-D-glucopyranosyl-(1!2)b-D-fructofuranose) is the natural compound generally associated with sweet taste. The sugar we use everyday, commercially extracted from sugarcane or sugar beet, is also the most commonly used substance for altering the flavor of food. In addition to sucrose, there are hundreds of other ‘‘sugars,’’ that is water-soluble crystalline carbohydrates or sugar alcohols characterized by a typical sweet taste. All these compounds are characterized by the presence on their skeleton of several hydroxyl groups, but it is not easy to find a special distribution of these substituents typical only of sweet saccharides. Figure 1 shows the molecular models of some typical natural sweet molecules, including sucrose. The two monosaccharides contained in sucrose, fructose and glucose, have the same number of hydroxyl groups but have a somewhat different sweetening power: 25% and 75% as sweet as sucrose, respectively. Another class of polyol compounds that have been widely used as sugar substitutes is that of sugar alcohols. The most important of these sugar substitutes are erythritol, glycerol, mannitol (hexane-1,2,3,4,5,6hexol), and sorbitol. Figure 1 shows the molecular models of glycerol and sorbitol. It can be appreciated that sorbitol has a chemical constitution very similar to that of monosaccharides, yet it has been used as sugar substitute since the body metabolizes it slowly. Amino acids of the natural (S) configuration, that is the building blocks of proteins, are for the major part tasteless or bitter whereas glycine and some hydrophobic amino acids of R chirality are sweet (Solms et al., 1965).

Active Sites of the Sweet Receptor

OH

203

Sucrose

O OH

OH

HO

OH

HO

OH

O

O

H

CH2OH

OH

Glycerol

H

C

OH

H

C

OH

H

C

OH

H

H

C

OH

HO

C

H

H

C

OH

Sorbitol

CH2OH H H

C

COO−

NH3+

Glycine

H2C

N H

C

COO−

H O

Tryptophan

NH3+

HO O

O

H N H2N

O O

OH

CH3

OH

O NH2

Aspartame OH

N H

Monatin

FIGURE 1 Molecular formulas of some sweet molecules, representative of the major natural classes: carbohydrates (sucrose), polyols (glycerol, sorbitol), amino acids (glycine, tryptophan), peptides (aspartame, monatin).

In fact, the very name of glycine comes from the greek word for sweet (glukos). Figure 1 shows the molecular models of glycine and of R-tryptophan, the sweetest of R amino acids. In addition to simple amino acids, there are several sweet peptides; most of them, although composed of natural amino acidic residues, are not natural and are related to aspartame, l-aspartylphenylalanine methyl ester, the first sweet dipeptide, discovered

204

Pierandrea Temussi

by serendipity in the 1960s (Mazur et al., 1969). A very interesting naturally occurring peptide, monatin, has been described and fully characterized (Bassoli et al., 2005). Extracts from plants led to the discovery of several intensely sweet natural glycosides. The best known of sweet glycosides is probably stevioside, a component of Eupatorium rebaudianum (Bridel and Lavieille, 1931) 150–300 times sweeter than sucrose. Another important natural glycoside is Glycyrrhizin, the flavoring agent of licorice (Glycyrrhiza glabra),
50 times as sweet as sucrose (Tahara et al., 1971). However, the aglicone of Glycyrrhizin is a triterpene structurally similar to corticosteroids; hence, Glycyrrhizin at high doses induces hypertension. A much sweeter similar glycoside (3000 times sweeter than glucose) is Osladin, extracted from Polypodium vulgare (Tahara et al., 1971). Other small molecular weight sweeteners of natural origin are often classified as semisynthetic since the original natural substance, although not sweet, becomes sweet after minor chemical modification (Morris, 1976). The terpene perillaldehyde extracted from Perilla frutescens is not sweet but becomes 200 times sweeter than sucrose when the aldehyde functional group is changed into its syn-oxime, called perillartine (Acton and Stone, 1976). A similar relationship exists between narigin dihydrochalcone and neohesperidine dihydrochalcone and the corresponding bitter flavanone glycosides derived from citrus fruits (DuBois et al., 1981). In turn, the aglicone of neohesperidine dihydrochalcone represents an entire class of natural compounds that can be called isovanillyl sweet compounds (Bassoli et al., 2002b) since they contain the isovanillyl group (3-hydroxy-4methoxyphenyl). The most representative natural isovanillyl molecules are phyllodulcin, dihydroquercetin 3-acetate, and hematoxylin, but the sweetest isovanillyl compounds are synthetic modifications discovered in the laboratory of Merlini (Bassoli et al., 2002b). Figure 2 shows the molecular models of some larger typical natural sweet molecules (stevioside, Glycyrrhizin, neohesperidine dihydrochalcone, and phyllodulcin). In addition to many natural products, there is a huge number of unrelated organic molecules that were found to be sweet in the course of many decades of research in synthetic organic chemistry. As mentioned in Section I , in the past it was a common practice among organic chemists to taste the new compounds they synthesized, and many substances turned out to be sweet. Accordingly, there are many sweeteners totally unrelated to the classes of Figures 1 and 2. Figure 3 shows the molecular models of some representative synthetic sweeteners. It can be appreciated that all of them are considerably more hydrophobic than most natural sweeteners. None of the mentioned compounds as well as many others discovered or synthesized in the last 100 years, when taken as sweetener, is free from drawbacks. Therefore, much hope has been attached to the most unusual class of natural sweet compounds, namely sweet proteins.

205

Active Sites of the Sweet Receptor

OH HO HO

HO

O

COOH

O

O O

O

HO

Stevioside

HOOC O HO HO HOOC O O HO HO OH

OH O HO

O

Glycirrhizin

O O

HO OH

O

OH

OCH3 O

HO HO

OH

O

OMe OH

O Me HO

O

OH

O OH

HO

O

OH

O

OH

Neohesperidin dihydrochalcone

Phyllodulcin

FIGURE 2 Molecular formulas of further typical natural sweet molecules, mainly related to terpenes: stevioside, glycyrrhizin, neohesperidine dihydrochalcone, and phyllodulcin.

B. Early structure–activity studies Early attempts to understand structure–taste relationships of molecules were based on the search of specific atoms or groups of atoms (called ‘‘sapophores’’) that could impart a given taste to molecules (Cohn, 1914). In the case of sweet compounds, Cohn observed, for instance, that molecules containing several hydroxyl groups, or chlorine atoms, or the a-amino and carboxyl groups typical of amino acids are often sweet; accordingly, these groups of atoms were defined ‘‘dulcigen’’ groups (Cohn, 1914). This approach was further elaborated by Oertly and Myers (1919) who called these chemical groups ‘‘glucophores’’ and others, with the ability to increase the potency, ‘‘auxogluc.’’ It was soon clear that ‘‘glucophores’’ and ‘‘auxogluc’’ groups belonged to sweet and tasteless molecules with similar frequency, whereas other features, that is the steric disposition of groups, probably played an important role (Moncrieff, 1967). The first successful generalization can be attributed to Shallenberger and Acree (1967) who hypothesized that the main signature for sweet molecules could reside on their skeleton in the presence of a hydrogen ˚. bond donor (AH) and a hydrogen bond acceptor (B) spaced 3–4 A These two groups, by interacting with a complementary pair of hydrogen bond donor and acceptor on the receptor, would act as the main

206

Pierandrea Temussi

Cl 6Cl-saccharin

Saccharin S

O

S

O N H

O

O

N H

O N

Cyclamate S

Nitrothiophene

O

O

NH S

HO

O O O

O

HN 2

O

NH

N O

P4000

Dulcin

N N HN

SO3−

SSN

FIGURE 3 Molecular models of some representative synthetic sweeteners: saccharin, 6Cl-saccharin, cyclamate, nitrothiophene, dulcin, P4000 (ortho-propoxy-meta-nitro aniline), and SSN (3-anilino-2-styryl-3H-naphtho[1,2-d]imidazole-5-sulphonate).

anchoring points in binding. The model of Shallenberger and Acree can be regarded as a linear model. It was soon developed into a planar (triangular) geometry by Kier (1972) who introduced a third group, named the ‘‘dispersion point,’’ at a precise distance from the AH–B pair. Kier’s model accounts for the experimental observation that many, albeit not all, synthetic sweeteners are flat rigid molecules. This model enjoyed great popularity among medicinal chemists, possibly because of its simplicity, but the identification

Active Sites of the Sweet Receptor

207

of a single dispersion point represented by virtually any uncharged atom or even a point on a chemical bond is an oversimplification. One of the weakest points of this model is that, in the absence of any stereochemical information, it could not allow selection between chemically similar, or even isomeric, sweet and tasteless or bitter compounds. For instance, of the three possible nitroanilines, the meta-isomer is sweet but the ortho- is not and the para- is almost tasteless (Moncrieff, 1967). The interest was soon redirected toward the development of more general models of the receptor active site derived from the shape of conformationally rigid sweet molecules, used as molecular molds. The most exhaustive approach was that of Temussi and coworkers (Kamphuis et al., 1992; Temussi et al., 1978, 1984, 1991). They suggested a more detailed model based on an accurate superposition of rigid sweet compounds, which should reflect the overall shape of the putative receptor cavity (Kamphuis et al., 1992; Temussi et al., 1978, 1984, 1991). The combination of several observations, using also flexible compounds, notably aspartame, whose solution structure had just been determined (Lelj et al., 1976), led to a detailed quasi-planar outline of the active site. The main features of this model can be summarized as follows: (1) the active site of the receptor is a flat cavity with one side partially accessible even during the interaction with the agonist; (2) the lower part of the cavity hosts the AH–B entity complementary to that of the sweet molecule; (3) the upper part is hydrophobic and plays an important role in the case of very active sweeteners. This is often referred to as the ‘‘Temussi model’’ (Walters, 1995; Walters et al., 1986). Figure 4A shows the main contour of the active site hosting a model of aspartame in an extended conformation. Aspartame was the starting point for another model, proposed by Iwamura (1981) on the basis of QSAR analyses of dipeptide analogues. This author claimed that his receptor model is different from that proposed by Temussi et al. (1978), but the results of the calculations reflected mainly the difficulty of using conformationally flexible compounds for a QSAR calculation. Another topological model, mainly based on the conformation of aspartame and other dipeptides, was developed by Goodman et al. (1997). This model (Figure 4B), while incorporating most of the features of the Temussi model, differed in some sterical aspects. According to their model, the overall topology of a sweet tasting molecule can be described as an ‘‘L’’-shaped structure with the aspartyl moiety forming the stem of the ‘‘L’’ and the hydrophobic group X forming the base of the ‘‘L’’ (Goodman et al., 1997). The zwitterionic ring of the aspartyl residue is coplanar and essentially perpendicular to the X group. Figure 4B shows the superposition of the ‘‘L’’-shaped model and an ‘‘L’’-shaped structure for aspartame. Also the Temussi model was originally inspired by the solution conformation of aspartame (Lelj et al., 1976), but it was soon realized that aspartame is too flexible to be used as a mold (Temussi et al., 1984) and the final active site model, although consistent

208

Pierandrea Temussi

A

B X

AH

AH

B

B

C E2

XH

G

Y AH

B

D

FIGURE 4 Three of the most popular indirect models of the active site of the sweet taste receptor. (A) Main contour of the active site proposed by Temussi and coworkers (Kamphuis et al., 1992; Temussi et al., 1978, 1984, 1991), hosting a molecular model of aspartame in an extended conformation. (B) A topological model, developed by Goodman et al. (1987). The ‘‘L’’-shaped model and an ‘‘L’’-shaped conformation of aspartame are superimposed. The hydrophobic side chain of Phe is denoted X, since it corresponds to the Kier’s dispersion point. (C) 3D model of an idealized sweetener proposed by Tinti and Nofre (1991). Besides the AH–B entity, the model has six additional interaction points connected by a complex network of distances.

with the solution structure of aspartame, was built on more rigid molecules. A way to discriminate between the models of Figure 4A and B could have been to be able to predict the conformation of aspartame in the actual receptor. Experimental structural studies were not sufficient to give an unequivocal answer: the conformer found in the crystal structure of aspartame (Hatada et al., 1985) is consistent with Goodman’s model, whereas that of the more rigid and sweeter [(L-a-Me)Phe2] aspartame (Polinelli et al., 1992) is consistent with Temussi’s. The most popular model in the 1990s was that of Tinti and Nofre (1991). Following their discovery of very potent sweeteners containing a guanidinium ion (Nofre et al., 1988; Tinti and Nofre, 1991), they proposed

Active Sites of the Sweet Receptor

209

a three-dimensional (3D) model for an ideal sweetener that besides the AH–B entity has six additional interaction points connected by a complex network of distances (Figure 4C). This model may suffer from little generality since it specifically tailors the architecture of only one type of compounds, but it has the great merit of being consistent with the most powerful known sweeteners. Just before the discovery of the sweet taste receptor, Bassoli et al. (2002a) proposed a unifying model able to explain and predict semiquantitatively the sweet taste of compounds belonging to different families. An entirely different approach has been proposed by Gokulan et al. (2005). By comparing the crystal structures of synthetic supersweetener and nonsweetener compounds complexed with murine monoclonal antibody (mAB) NC6.8, they found that receptor–ligand interactions imply a complex array of hydrogen bonds, electrostatic interactions, and several hydrophobic contacts. The main conclusion was that the difference between high-potency guanidine sweeteners and related zwitterionic low-potency tastants is determined by the nature and conformation of the hydrophobic group. Their results are very interesting but suffer from the same drawbacks of indirect modeling. Since both are not based on any knowledge of the actual receptor, they may bear little relationship with actual receptor binding. The long trial and error search for ideal sweeteners via indirect mapping or intuition did produce a number of high-potency sweeteners, notably those derived from aspartame, like neotame (Prakash et al., 1999) or superaspartame (Tinti and Nofre, 1991), and the guanidinium compounds (Nofre et al., 1988; Tinti and Nofre, 1991). However, none of the existing models of the active site could explain the enormous increase in activity in going from small molecular weight compounds to proteins: monellin, for example, one of the best characterized sweet proteins, is 100,000 times sweeter than sucrose on a molar basis (Hung et al., 1999).

III. SWEET MACROMOLECULES Although not very numerous, sweet macromolecules, both natural (Morris, 1976) and synthetic (Zaffaroni, 1975), are crucial for an understanding of the mechanism of the sweet receptor. The best known among proteins with a very strong sweet taste are brazzein (Ming and Hellekant, 1994), monellin, and thaumatin (Kurihara, 1992). Figure 5 shows molecular models of these three proteins. Other two known sweet proteins are mabinlin (Kurihara, 1992) and hen egg white (HEW) lysozyme (Maehashi and Udaka, 1998), whereas miraculin and curculin, which taste sweet when combined with sour substances, can be better described as taste-modifier proteins (Kurihara, 1992).

210

Pierandrea Temussi

A

C L34 L23 B L67

FIGURE 5 Ribbon representations of the three sweet proteins of known structure. (A) Structure of MNEI (pdb entry 1fa3): the most likely sweet finger is loop L34. (B) Structure of thaumatin (pdb entry 1thw): the most likely sweet finger is loop L67. (C) Structure of brazzein (pdb entry 2brz): hairpin L23 represents the only possible sweet finger.

A. Characterization of natural sweet proteins 1. Monellin Until 1972, it was not known that a protein could taste sweet (Morris and Cagan, 1972). Monellin is one of the first two proteins with intense sweet taste unambiguously identified and characterized. Inglett and May (1969), who originally discovered it as the sweet principle of Dioscoreophyllum cumminsii, a plant taxonomically related to the sweet potato, believed it was a carbohydrate. Owing to the unexpected intensity of the sweet taste, Inglett and May called the plant ‘‘serendipity berries’’ (Morris, 1976). Later on, Morris and Cagan (1972) established that the sweet principle is a protein and named it monellin, after the Monell Chemical Senses Center where they worked. According to these authors, the sweetness of monellin relative to sucrose is 3000:1 on a weight basis, corresponding to a ratio of 90,000:1 on a molar basis. Monellin consists of two nonidentical subunits of 42- and 50-amino acid residues, called A and B respectively, that are not covalently linked but are held together only by secondary forces (Bohak and Li, 1976). The sequence of monellin bears no significant similarity to that of any of the other sweet proteins. In addition, when it was originally discovered, it was impossible to assign it to a known protein family, but Murzin (1993) on the basis of its solid state structure demonstrated that it belongs to the cystatin superfamily, albeit devoid of any activity as a protease inhibitor. The sweetness of monellin is exhibited only by the whole molecule, whereas the individual subunits are not sweet (Bohak and Li, 1976). Owing to the weak forces holding its two chains, when heated above 50 C monellin

Active Sites of the Sweet Receptor

211

dissociates into two chains and, as a consequence, it loses its sweetness altogether. Single-chain monellins in which the two chains are covalently linked retain all sweetening power but have greatly increased thermal stability (Kim et al., 1989; Tancredi et al., 1992). The first single-chain monellin (dubbed SCM) was designed by Kim et al. (1989) on the basis of the crystal structure of wild-type monellin. In SCM, the C-terminal residue of the B chain (B50E) is directly linked to the N-terminal residue of the A chain (A1R). SCM is as sweet as natural monellin, more stable on temperature or pH changes, and renatures easily even after heating to 100 C at low pH. A very similar behavior was shown by MNEI, a single-chain monellin obtained by inserting a Gly-Phe dipeptide between the B and A chains (Tancredi et al., 1992). The structures of both two-chain and single-chain forms of monellin were thoroughly characterized by X-ray and nuclear magnetic resonance (NMR) studies (Hung et al., 1998; Lee et al., 1999; Somoza et al., 1993; Spadaccini et al., 2001). The solution structure of MNEI, shown in Figure 5A, can be described as an a-helix cradled into the concave side of a five-strand antiparallel b-sheet solution (Spadaccini et al., 2001). The huge difference in size between sweet proteins and all non-proteic sweeteners led several researchers to postulate the existence, on the surface of monellin, of some kind of ‘‘sweet finger,’’ that is a protruding structural element hosting one or more glucophores similar to those of small sweeteners. ELISA tests showed cross-reactivity between antibodies raised against monellin and those raised against thaumatin (Bodani et al., 1993; Mandal et al., 1991). On this basis, the sequence TyrA13-AspA16 of native monellin and that comprising residues Tyr57-Asp59 of thaumatin were suggested as a potential sweet fingers (Kim et al., 1991). However, point mutations on synthetic monellin (Ariyoshi and Kohmura, 1994) showed that even substantial changes of residues 13 and 16 of the A chain do not affect sweetness. Actually, according to Ariyoshi and Kohmura (1994), TyrA13Gly and TyrA13Phe have a sweetening power slightly higher than wild-type monellin whereas the AspA16Abu and AspA16D-Asp have activities nearly twice as high. In addition, extensive mutagenesis studies both on SCM and wild-type monellin (Kohmura et al., 1992; Somoza et al., 1995) hinted at an area of interaction with the receptor much larger than that of a sweet finger. The residues whose mutation causes a decrease of sweetness of two or more orders of magnitude are Ile6, Asp7, Gly9 (Kohmura et al., 1992), and Arg39 (Sung et al., 2001) whereas mutations of Gln13, Lys36, Lys43, Arg72, Arg88 or deletion of Pro92–Pro96 cause a decrease of one order of magnitude (Kohmura et al., 1992; Somoza et al., 1995). The distribution of key residues (for biological activity) on a large area was confirmed without recurring to any hypothesis on the mechanism of interaction, that is, in a completely objective way, by a surface survey based on novel NMR techniques (Niccolai et al., 2001).

212

Pierandrea Temussi

The survey of the MNEI surface accessibility (Niccolai et al., 2001) was performed by means of TEMPOL, a paramagnetic probe and a direct assessment of bound water by means of ePHOGSY, a pulse sequence that allows accurate detection of NOEs between bound water and protein hydrogens (Dalvit, 1996, 1998). The result of this integrated NMR study suggested that three MNEI regions are potentially suitable for interactions with other proteins: loop L34, previously referred to as a potential sweet finger; the small N-terminal b-strand containing Ile6, Asp7, and Gly9; and a basic patch containing Arg72 and Arg88.

2. Thaumatin

The jelly-like exterior of the seeds of a West African plant, Thaumatococcus danielli, is intensely sweet. Inglett and May (1968) reported that it contains a sweet substance of ‘‘unique chemical and physical nature’’ but failed to identify it as a protein. The sweet substance that makes the seeds of T. danielli so sweet was later characterized by van der Wel and Loeve (1972) as a mixture of two proteins called thaumatin I and II endowed of a sweetening power about 1600 times higher than that of sucrose on a weight basis or 100,000 on a molar basis. So far, thaumatin is the only sweet protein that has been actually used as a sweetener: in the 1970s, Tate and Lyle began commercializing thaumatin, as extracted from T. danielli, under the trade name of Talin. Thaumatin is a single polypeptide chain of 207 residues (Iyengar et al., 1979) and, according to the SCOP classification (Murzin et al., 1995), belongs to the osmotin, thaumatin-like superfamily. The 3D structure of thaumatin, solved in the solid state by X-ray studies (de Vos et al., 1985; Ogata et al., 1992), contains three domains, mainly composed of b-sheets (Figure 5B). The structure–activity relationship of thaumatin has been studied less than that of monellin. Comparing the amino acid sequence of thaumatin with that of monellin, the other sweet-tasting protein known at the time; Iyengar et al. (1979) located five sets of identical tripeptides that might be part of a common antibody recombination site and possibly be involved in the interaction with the sweet taste receptor. However, if one runs a comparison of the sequence of thaumatin with those of other sweet tasting proteins by means of modern bioinformatics means, for example ClustalX (Thompson et al., 1997), the similarities are negligible (Tancredi et al., 2004). Mandal et al. (1991) developed a library of monoclonal antibodies that react with different surface antigenic epitopes on thaumatin and, in a few instances, also cross-react with monellin. A similar study by Slootstra et al. (1995) identified two major overlapping conformational epitopes. This region contains an aspartame-like site which is formed by Asp21 and Phe80, tips of the two extruding loops 19–29 and 77–84, which are spatially positioned next to each other. Since the aspartame-like Asp21-Phe80 site is not present in nonsweet thaumatin-like proteins, they suggested that the

Active Sites of the Sweet Receptor

213

two loops contain important sweet taste determinants. Kaneko and Kitabatake (2001), by examining in detail the role of lysines in the structure–sweetness relationship of thaumatin, found that phosphopyridoxylation of Lys78, Lys97, Lys106, Lys137, or Lys187 reduced sweetness significantly. Combination of these results with those ensuing from modifications of other charged residues led them to suggest that there is a charged side of the protein that is important for sweetness. These studies are not conclusive but seem to point to a large surface of interaction also for thaumatin.

3. Brazzein Brazzein, the smallest of sweet proteins, was discovered only in 1994 (Ming and Hellekant, 1994) in Pentadiplandra brazzeana B. This protein, whose sequence contains 54-amino acid residues, is
2000 times sweeter than sucrose when compared to a 2% sucrose aqueous solution. Its taste was described as more similar to sucrose than that of thaumatin (Ming and Hellekant, 1994). As can be seen in Figure 5C, the 3D structure of brazzein, determined by 1H NMR spectroscopy in solution at pH 5.2 (Caldwell et al., 1998), is very simple. It contains one a-helix and three strands of antiparallel b-sheet. The structure is stabilized by four disulfide bonds, three connecting the helix to the b-sheet. It does not resemble either that of monellin or that of thaumatin; instead, it resembles those of plant g-thionins and defensins and arthropod toxins. According to the SCOP classification (Murzin et al., 1995), brazzein belongs to the Scorpion toxin-like superfamily. All studies on the structure–activity relationship of Brazzein were performed by the same group that elucidated the 3D structure. Assadi-Porter et al. (2000) by introducing multiple mutations at several specific positions found that the mutations that affect most the sweetness of brazzein are localized within the tracts Asp29LysHisAlaArg33 and Tyr39AspGluLysArg43, close to the C-terminus. These data were refined by Jin et al. (2003) who investigated more mutations. Three mutants, that is Ala2ins, Asp2Asn, and Gln17Ala, were found to be as sweet as wildtype brazzein. Four mutants, that is Asp29Ala, Asp29Lys, Asp29Asn, and Glu41Lys, were found to be significantly sweeter than wild-type brazzein. In other 8 mutants the sweetness decreased significantly although they were not tasteless, whereas in 10 mutants the sweetness did not differ significantly from that of water. At about the same time, Assadi-Porter et al. (2003) proposed a very innovative approach to identify the main structural determinants for the sweetness of brazzein. Assadi-Porter et al. (2003) applied NMR methods that permit direct detection of hydrogen bonds (Cordier and Grzesiek, 1999) to screen a series of five single site mutants of brazzein with altered sweetness, looking for possible changes in backbone hydrogen bonding with respect to wild-type.

214

Pierandrea Temussi

Assadi-Porter et al. (2003) found that in the mutants, altered magnitudes of the couplings identified hydrogen bonds that were strengthened or weakened with respect to the wild type. Within the series of brazzein mutants investigated, a pattern was observed between sweetness and the integrity of particular hydrogen bonds. Assadi-Porter et al. (2003) concluded that their findings may be interpreted as supporting the hypothesis of an extensive receptor-binding surface in brazzein, involving loop 43 and the N- and C-terminal regions. The success of this approach probably reflects the fact that changes in the protein surface mirror changes of the underlying network of hydrogen bonds.

4. Mabinlin Hu and Min (1983) have isolated two new sweet proteins from the seed of Capparis masaikai Levl., a plant that grows in the subtropical region of the Yunnan Province of China and named them mabinlin I and II, after the local name of the plant (mabinlang). The sweetness of mabinlin II with respect to sucrose was estimated as 375:1 on a molar basis; it remains unchanged by more than 48-hour incubation at boiling temperature (Liu et al., 1993). Thus, although mabinlin is much sweeter than sucrose, it is considerably less sweet than monellin, thaumatin, and brazzein. Purified mabinlin II gave a single band having a molecular mass of 14kDa on SDSPAGE, but two peptide chains (A and B) were separated from reduced and S-carboxamidomethylated mabinlin II by HPLC (Liu et al., 1993). The amino acid sequences of the A chain and B chains consist of 33-amino acid and 72-amino acid residues, respectively. The A chain is mostly composed of hydrophilic amino acid residues and the B chain also contains many hydrophilic residues. High similarity was found between the amino acid sequences of mabinlin II and 2S seed storage proteins, especially 2S albumin AT2S3 in Arabidopsis thaliana (mouse-ear cress).

5. Lysozyme As it is well known, lysozyme is a small enzyme that catalyzes the hydrolysis of polysaccharides comprising the cell walls of bacteria. It is exceptionally abundant in egg whites. Lysozyme is also one of the best characterized proteins from a structural point of view, both in solution (Schwalbe et al., 2001) and in the crystal state (Strynadka and James, 1996). Figure 6A shows a ribbon representation of the tetragonal form of lyso˚ resolution (pdb entry 193L). It is interesting to note, zyme solved at 1.3-A with respect to the structures of Figure 5, that are mainly rich in b-sheets, that the structure of lysozyme is prevalently a-helical. Its inclusion among sweet proteins is quite recent. The sweetening power of HEW lysozyme corresponds to a threshold value of around 7mM (Masuda et al., 2005b), a value that is far from the nanomolar range of the three main sweet proteins but is higher than that of sucrose. Maehashi and Udaka (1998)

Active Sites of the Sweet Receptor

215

A

B NAS

NBS

FIGURE 6 Ribbon representations of lysozyme and neoculin. (A) Tetragonal form of lysozyme solved at 1.3-A˚ resolution (pdb entry 193L). (B) One of the four crystallographically independent heterodimers of neoculin (pdb entry 2d04).

claimed that HEW lysozyme has a distinct sweet taste, whereas lysozymes from other sources such as turkey and soft-shelled turtle also showed sweetness but with different tastes, heavy or light. In contrast, human lysozyme is tasteless. The amino acid sequences of the various lysozymes are similar to that of HEW lysozyme, but no lysozyme sequence shows significant homology to other sweet proteins (vide infra). Masuda et al. (2001, 2005a,b) have studied extensively the structure–activity relationship of lysozyme. The main results can be summarized as follows. Alanine substitution of lysine residues showed that two of six lysine residues, only Lys13 and Lys96, are required for lysozyme sweetness, while the remaining four lysine residues do not affect significantly the sweetness. Similarly, single alanine substitutions of arginine residues showed that three arginine residues, Arg14, Arg21, and Arg73, play significant roles in lysozyme sweetness, whereas mutation of Arg45, Arg68, and Arg125 did not affect sweetness (Masuda et al., 2005b).

6. Miraculin

The fruits of Synsepalum dulcificum have been known for more than a century to cause sour substances to taste sweet. This very unusual property earned the berries the name of miracle fruit (Morris, 1976). Theerasilp and Kurihara (1988) isolated miraculin from alkaline extracts of the

216

Pierandrea Temussi

miracle fruit and purified it with standard biochemical procedures. Miraculin is a single polypeptide chain with 191-amino acid residues (Theerasilp et al., 1989). The calculated molecular weight based on the amino acid sequence and the carbohydrate content (13.9%) was 24,600. High homology was found between the amino acid sequences of miraculin and soybean trypsin inhibitor. The primary structure of miraculin was completed with the determination of the location of disulfide bridges (Igeta et al., 1991), but no tertiary structure is yet available. It has been claimed that miraculin can have a maximum value of sweetness 400,000 times that of sucrose (Gibbs et al., 1996). However, it is difficult to compare this figure to those of the sweet tasting proteins since the mechanism of action of miraculin apparently requires preventive (nonactive) occupancy of the receptor and it becomes sweet only after acidification (Kurihara and Beidler, 1969).

7. Curculin In 1990, a new taste-modifying protein named curculin was extracted from the fruits of Curculigo latifolia and purified by ammonium sulfate fractionation, ion-exchange chromatography, and gel filtration (Yamashita et al., 1990). Curculin consists of 114 residues, but the molecular weight suggests that native curculin is a homodimer of a 12,000-Da polypeptide. Curculin itself elicits a sweet taste, albeit not very strong (equivalent to the sweetness of 0.35-M sucrose). After curculin, water elicits a sweet taste, and sour substances induce a stronger sense of sweetness. No protein with both sweet-tasting and taste-modifying activities had previously been found. Until recently, however, it proved impossible to observe the tastemodifying properties in recombinantly expressed curculin. Almost simultaneously Shirasuka et al. (2004) and Suzuki et al. (2004) isolated a gene that encodes a novel protein highly homologous to curculin. The amino acid sequence of the novel gene has 77% identity to that of curculin but, in contrast to the previously reported isoform, the new protein is acidic, with an estimated isoelectric point of 4.7. Using cDNAs of the previously known curculin (dubbed curculin1) and the novel curculin isoform (curculin2), Suzuki et al. (2004) produced a panel of homodimeric and heterodimeric recombinant curculins by Escherichia coli expression systems. They found that sweet-tasting and taste-modifying activities were exhibited solely by the heterodimer of curculin1 and curculin2. Shirasuka et al. (2004) named neoculin this heterodimer and, on the basis of the isoelectric points, neoculin basic subunit (NBS) the polypeptide corresponding to curculin1 and neoculin acid subunit (NAS) curculin2. The crystal structure of neoculin, the first of taste-modifying proteins, showed a fold quite similar to that of monocot mannose-binding lectins. Figure 6B shows a ribbon representation of one of the four crystallographically independent heterodimers of neoculin.

Active Sites of the Sweet Receptor

217

B. Interaction of sweet proteins with the sweet receptor The discovery of a few proteins with an intense sweet taste (Morris, 1976) was a great shock for all researchers studying sweetness–activity relationship. The dimensions of all sweet proteins are so different from those of typical sweeteners that it was difficult to hypothesize an interaction with the same active sites proposed for small molecular weight sweeteners. In addition, it is difficult to find commonalities among the sweet proteins. The most widely used approach for an understanding of the origin of a common function among proteins belonging to the same family is to compare their sequences, in search of corresponding parts. No sequence homology can be detected among monellin, thaumatin, brazzein, mabinlin, miraculin, and curculin. A pairwise alignment of these sequences performed by Clustal X (Thompson et al., 1997) showed that the percentages of identical residues between monellin and the other proteins are 23% between monellin and miraculin and a bare 7% between monellin and curculin (Tancredi et al., 2004). If HEW lysozyme is included in the sequence alignment, the result is a complete misalignment (Temussi, 2006). Yet, the interaction of proteins with the receptor might be explained also on the basis of the quoted indirect models of active site based on the shape of small sweeteners, provided one could identify, on the surface of the proteins, protruding features that can probe the active site, that is ‘‘sweet fingers’’ chemically similar to small sweeteners. Therefore, many efforts have been devoted to the search of possible sweet fingers on proteins. As mentioned above, on the basis of early ELISA tests the sequence TyrA13-AspA16 of native monellin and that comprising residues Tyr57-Asp59 of thaumatin were suggested as a potential sweet fingers (Kim et al., 1991). Although mutagenesis studies on monellin (Somoza et al., 1995) hinted at a much larger spread of key residues on the surface of the protein, it could not be excluded that sweet fingers play an important role in recognition. Accordingly, Tancredi et al. (2004) undertook a systematic investigation on brazzein, monellin, and thaumatin to identify possible sweet fingers. They examined in great detail the structures of brazzein, monellin, and thaumatin for the presence of common motifs. The similarity among the tertiary folds of these three proteins is very low. One of the best methods to search for structural similarities is by means of DALI (Holm and Sander, 1995): a 3D search of each of the three known structures (brazzein, monellin, and thaumatin) against the whole database by means of DALI did not even retrieve the other two proteins. However, there are structural elements common to the three proteins, in the form of single secondary traits, notably short b-sheet hairpins. Potential candidates for sweet fingers should be protruding structural features of sufficient length to enter the active site of the receptor but, in addition, they should host residues consistent with glucophores already identified in small sweeteners. Judging from all existing models of the receptor (vide infra), the

218

Pierandrea Temussi

minimum length of a candidate substructure should be of the order of ˚ , since the active site is located at the bottom of a deep cleft, 20–30 A ˚ 20 A from the surface of the protein. Likely protruding elements should also have a sufficiently stable secondary structure; thus, we can restrict our search to b-hairpins present in all three proteins shown in Figure 5 as ribbon representations. Brazzein (shown in Figure 5C) is the simplest case since its very simple structure contains only one such hairpin, loop L23. Its length and the presence of residues containing suitable glucophores are consistent with the requirements outlined above. In the case of MNEI (Figure 5A), a single-chain monellin, there is no choice as clear as the one for brazzein. Loop L23, centered around Gly51-Phe52, can be excluded right away since it is not even present in native monellin; loop L45, being an integral part of a rigid b-sheet, is not mobile enough to act as a flexible finger whereas loop L34, although not completely free, is structurally similar to the loop of brazzein and, in addition, corresponds to the original sweet finger proposed by Kim et al. (1991). In the case of thaumatin, there are numerous loops with sufficient length to probe the receptor’s cavity. However, also in this case, most of them are tightly bound to the body of bsheet that forms the architecture of this protein and thus, cannot be freed for probing the receptor interior without disrupting the structure of the protein. The only loop that is not tightly bound to the body of b-sheet is loop L56. In addition, this loop is also the one identified by DALI as the only structural element similar to corresponding ones in brazzein (L23) and monellin (L34). All three loops contain, among the side chains, an aromatic ring (e.g., belonging to either a Tyr or a Phe) in relative spatial orientation, with respect to a pair of hydrogen bond donors or acceptors, similar to that found in aspartame. Starting from the sequences of these loops, Tancredi et al. (2004) synthesized the corresponding cyclic peptides: c[C56YFDDSGSGIC66], c[C61LYVYASDKLFRAC73], and c[C37FYDEKRNLQC47], with cyclization assured by –S–S– bridges. The cyclic peptides do assume conformations consistent with the conformation of the same sequences in the parent proteins. However, none of them was able to elicit sweet taste (Tancredi et al., 2004). If, as mentioned above, we add to this result the fact that mutants affecting sweetness of monellin are distributed over a large area, rather than being concentrated on a long protruding structural entity (Somoza et al., 1995), it is fair to assume that the sweet fingers hypothesis can be abandoned.

IV. THE SWEET TASTE RECEPTOR A. Molecular biology of taste receptors Indirect studies that tried to map the sweet taste active site or to design idealized sweeteners implied the existence of specific sweet taste receptors, but it was only in the last few years that likely candidate receptors were

Active Sites of the Sweet Receptor

219

identified, expressed, and characterized. During the 1990s, the efforts devoted to the identification of the molecular components of taste transduction, using molecular biological methods, yielded a large number of proteins potentially involved sensory transduction mechanisms (Kinnamon, 2000). The first proteins identified in taste-receptor cells were components of the G-protein oligomers, that is, of the complex involved in signal transduction of many receptors, in particular a-gustducin (McLaughlin et al., 1992) and a-transducin (McLaughlin et al., 1994). From these studies, it was clear that the receptor most likely belonged to a class of GPCRs. However, although a number of candidate taste-GPCRs were proposed (Abe et al., 1993; Hoon et al., 1999; Matsuoka et al., 1993), it was difficult to establish their functional significance unequivocally (Lindemann, 1999). The first tastespecific GPCR cloned and characterized with respect to its ligand was that of the umami taste, that is the taste receptor for monosodium glutamate, a key ingredient of oriental food (Chaudhari et al., 2000). Soon after, Chandrashekar et al. (2000) reported the characterization of a large family of putative mammalian bitter taste receptors (T2Rs) and Matsunami et al. (2000) reported the identification of a family of candidate taste receptors (TRBs). At the beginning of 2001, T1R3, the first putative sweet receptor, was finally identified (Bachmanov et al., 2001; Kitagawa et al., 2001; Li et al., 2001; Max et al., 2001; Montmayeur et al., 2001; Nelson et al., 2001; Sainz et al., 2001). The approaches followed by these groups are well illustrated by Montmayeur et al. (2001). Several biochemical and electrophysiological studies had indicated that the detection of sweet, bitter, and umami taste transduction involved GPCRs (Gilbertson et al., 2000). One of the genetic loci that control sensitivity to bitter or sweet in mouse or human compounds, the Sac locus, governs the sensitivity of mice to certain sweet tastants, including sucrose and saccharin (Bachmanov et al., 1997; Blizard et al., 1999; Capeless and Whitney, 1995; Fuller, 1974; Lush, 1989; Lush et al., 1995). Montmayeur et al. (2001) followed an approach similar to that had located some T2r genes at or near genetic loci that control sensitivity to bitter taste (Adler et al., 2000; Matsunami et al., 2000). Searching the syntenic region of the human genome for genes encoding GPCRs, they identified T1R3, a gene encoding a GPCR that is expressed in a subset of taste cells in mouse, and found allelic differences in Sac taster versus nontaster strains that could result in differences in Sac phenotype. In addition, Montmayeur et al. (2001) found that in situ hybridization studies show that T1R3 is expressed in the same taste cells as T1R2, a related receptor, raising the possibility that the two receptors function as heterodimers or that these cells recognize more than one ligand. However, the common belief was that the sweet taste receptor was a (T1R3) homodimer like most metabotropic, or class C, GPCRs. Class C includes several glutamate receptors, sweet and umami (monosodium glutamate) taste receptors, the Ca2þ-sensing receptor, the g-aminobutyric acid type B receptor, and pheromone receptors (Pin et al., 2003). As all GCPRs, these

220

Pierandrea Temussi

receptors have a 7TM, but in addition they have a large extracellular domain, called VFTD, containing the active site for typical ligands, and a cysteine-rich domain. As mentioned above, several groups (Bachmanov et al., 2001; Kitagawa et al., 2001; Li et al., 2001; Max et al., 2001; Montmayeur et al., 2001; Nelson et al., 2001; Sainz et al., 2001) hypothesized almost at the same time that T1Rs (particularly T1R3 that corresponds to the Sac gene) were likely candidates for the sweet receptor. In analogy with other C receptors, it was assumed that each member of this family would form a homodimer in its active form. Only a few months later, Li et al. (2002) demonstrated that only heterodimer T1R2–T1R3 can function as a sweet receptor. The very likely presence in the sweet taste receptor of cavities similar to those hosting Glu in mGluR1, a metabotropic glutamate receptor of known structure (Kunishima et al., 2000), tells us that the sweet taste of small molecular weight sweeteners can certainly be accounted for, even if the details will remain in part obscure, at least till a receptor structure with better resolution than homology models will be available. Can the taste of sweet proteins be also explained by the knowledge of the receptor? There is no obvious answer. Let us first examine possible receptor models in detail.

B. Computer-generated models of the sweet taste receptor How can the potencies of sweet molecules span a range of five orders of magnitude? How can molecules as large as proteins interact with the same receptor as small sweeteners? Precise answers to these difficult questions would require solving the structure of complexes of the sweet receptor with representative sweet molecules or, at least, solving the structure of the receptor alone. In the meantime, crucial information was gained from modeling studies of the T1R2–T1R3 receptor. In particular, the similarity of the sweet taste receptor to mGluR1, one of the metabotropic glutamate receptors, hinted at the possible coexistence of different mechanisms for the two classes of molecules and prompted Temussi (2002) to propose the so-called ‘‘wedge model’’ for proteins. The first likely sweet taste receptor was a protein (dubbed T1R3) whose sequence has sufficient homology to several metabotropic GPCR. The sequence of T1R3 bears significant homology to several other metabotropic receptors, in particular it is 20% identical to that of mGluR1. It is a very happy coincidence that, at the time of the discovery of the sweet receptor, the structure of the N-terminal domain of mGluR1 had just been determined (Kunishima et al., 2000). The knowledge of the structure of the N-terminal domain of mGluR1 allowed homology model building. Similarly to mGluR1, that is a homodimer of a single sequence, the first homology model of the sweet receptor was built as a homodimer of two

Active Sites of the Sweet Receptor

221

T1R3 chains (Max et al., 2001). On the basis of the dimeric nature of their model, these authors postulated that taster to nontaster substitutions could affect N-linked glycosylation at N58, thus precluding correct dimerization. Almost simultaneously, another homodimeric model was used to show that the active site of T1R3 can consistently host three very sweet small molecular weight molecules (Walters, 2002). When Li et al. (2002) demonstrated that the actual taste receptor contains two similar but not identical proteins (dubbed T1R2 and T1R3) and that only the heterodimer T1R2–T1R3 can function as a sweet receptor for all classes of sweet molecules, it was necessary to revise homology models. The only possible template remains the structure of mGluR1, but it is necessary to build more than a single homology model since Kunishima et al. (2000) have shown that the extracellular N-terminal domain of mGluR1 exists in three different forms: one complexed with two molecules of glutamate and two ligand-free forms. Both the complexed receptor (Protein Data Base entry 1ewk.pdb) and the uncomplexed free form II (Protein Data Base entry 1ewv.pdb) can be called open–closed conformations and correspond to the active state of the receptor, whereas the other ligand-free form (Protein Data Base entry 1ewt.pdb) is in an open–open conformation and corresponds to a resting state of the receptor. Combining two sequences (T1R2 and T1R3) with two conformations amounts to four possible heterodimers. The first heterodimeric T1R2–T1R3 model, based on the mouse sequences and built using the complexed form as the template (1ewk.pdb), corresponded to only one of the two possible active models (Temussi, 2002). Morini et al. (2005) built all models of the human sequences and used them to identify all possible sites of interaction. Out of four T1R2–T1R3 heterodimers, two are inactive, ligand-free, open–open forms, and two are active, complexed, closed–open forms. If we model T1R2 on chain A and T1R3 on chain B of the 1ewt template, we get the two inactive dimers: Roo_AB (where R stands for resting, oo for open–open, and AB refer to the two chains of mGluR1), and Roo_BA when we model T1R3 on chain A and T1R2 on chain B of the 1ewt template. If we model T1R2 on chain A and T1R3 on chain B of 1ewk, we get Aoc_AB (where A stands for resting, oc for open–closed, and AB refer to the two chains of mGluR1), whereas when we model T1R3 on chain A and T1R2 on chain B of 1ewk, we get Aoc_BA, the two possible active dimers.

V. MECHANISMS OF INTERACTION A. The ‘‘wedge model’’ mechanism for sweet proteins The similarity between the sequences of the two chains of the T1R2–T1R3 receptor and that of the single chain of the homodimer of mGluR1 suggests that the two receptors might have the same general features,

222

Pierandrea Temussi

particularly with respect to the mechanism of activation. If the T1R2–T1R3 receptor behaves like the mGluR1, it should also exist as a mixture of three forms: a complexed form, activated by low molecular weight sweeteners, a resting ligand-free form I, and ligand-free form II, with a structure nearly identical to that of the ‘‘active,’’ complexed form. As shown by Figure 7, the resting (Roo) and active (Aoc) forms are in equilibrium even in the absence of their ligands, that is glutamate for mGluR1 and sweeteners for the sweet taste receptor. In analogy to mGluR1, the ‘‘normal’’ way to activate the receptor, shown by Figure 7A, should be the binding of a small molecular weight sweetener that transforms resting free form I into the active complexed form. However, the equilibrium between form I and form II can also be shifted if we can stabilize form II in another way. Figure 7B illustrates how stabilization can be achieved by external binding of a macromolecule on a secondary binding site on the surface of the receptor. This mechanism termed the ‘‘wedge model’’ was proposed on the basis of docking calculations of brazzein, monellin, and thaumatin to the Aoc conformation of a model receptor, built using the mouse sequences of T1R2–T1R3 (Temussi, 2002). This mechanism was soon supported by experimental and theoretical work. G16A-MNEI is a structural mutant that shows a reduction of one order of magnitude in sweetness with respect to its parent protein, MNEI, a single-chain monellin. This data was difficult to interpret since the mutation does not affect any part of the surface of MNEI but only, and slightly, its hydrophobic core. Comparison of the structures of wild-type monellin and its G16A mutant showed that

A

Aoc

Roo Small sweet molecules

B

Sweet proteins

FIGURE 7 Modes of binding of small sweeteners and sweet proteins. (A) Binding of small molecular weight ligands transforms resting (open, open) free form I (Roo, left) into the complexed form (Aoc, right), identical to active (open, closed) free form II. Small ligands in the two cavities of Aoc are shown as black balls. (B) Free form II, stabilized by protein complexation (active form, right), activates long lasting signal transduction. The ‘‘wedge’’ protein is shown in white.

Active Sites of the Sweet Receptor

223

the mutation does not affect the structure of potential glucophores but produces a distortion of the surface owing to the partial relative displacement of elements of secondary structure. These results support the hypothesis that the mechanism of interaction of sweet tasting proteins involves a large part of the sweet protein surface, as proposed in the wedge model (Spadaccini et al., 2003). The original formulation of the wedge model was based on the homology model of only one of the possible conformations of the receptor, but it has been substantiated by exhaustive modeling using the human sequences (Morini et al., 2005). After building all possible resting and active models, Morini et al. (2005) used them for docking calculations with experimental structures of brazzein, MNEI, and thaumatin, the sweet proteins of known 3D structure. The results for all three proteins are consistent with those found with the mouse receptor (Temussi, 2002). As a negative check Morini et al. (2005) calculated also the docking of sweet proteins to the inactive open–open Roo_AB and Roo_BA models. Figure 8 shows the interaction of MNEI with Aoc_AB, Aoc_BA, Roo_AB, and Roo_BA. All 10 molecules of MNEI are found, oriented in a similar albeit not identical way, in the same spot of the surface of the human Aoc_AB form, mainly belonging to the T1R3(B) chain. Efficient binding is assured mainly by shape and charge complementarity. As shown in the right-hand side of the upper panel, the results obtained by docking MNEI on Aoc_BA are similar from a structural point of view, since the MNEI molecules, also in this case, bind to a cavity on the T1R2 (B) chain. As shown in the lower panel, in the case of the inactive open– open Roo_AB and Roo_BA models, the molecules of MNEI bind to a very large area of the receptor, without any apparent regularity. The wedge model received a further validation by the design of a single-chain monellin sweeter than wild-type monellin (Esposito et al., 2006). According to the wedge model, the mechanism of interaction between a sweet protein and the sweet receptor hints at a largely positive nature of the surface of interaction of the protein with the receptor. Accordingly, three neutral residues, Met42, Tyr63, and Tyr65, comprised in a critical area of interaction hosting key residues according to previous mutagenesis studies, were changed into either acidic or basic residues. The expectation that, in order to be consistent with the wedge model, all changes of neutral residues to acidic ones ought to be more detrimental for sweetness than the corresponding ones into basic residues was met by all mutants. In addition, careful selection of the best point mutation, that is Tyr65, led to the discovery of a mutant, Y65R, that is even sweeter than MNEI itself. Tyr65 is at the center of the main interacting area predicted by the wedge model and close comparison of the surface electrostatic potential of MNEI and Y65R reveals that, indeed, the only change in going from MNEI to the mutant is an increase of the positive area of the interface at

224

Pierandrea Temussi

Aoc_AB

Roo_AB

Aoc_BA

Roo_BA

FIGURE 8 The four possible models of the human sweet receptor with bound MNEI molecules. Active models (Aoc_AB and Aoc_BA, upper panel): atoms of the T1R2 sequence are shown in green, whereas those of the T1R3 sequence are shown in dark green. Models of MNEI are represented as gold neon backbone bonds. Resting models (Roo_AB and Roo_BA, lower panel): atoms of the T1R2 sequence are shown in light green, whereas those of the T1R3 sequence are shown in blue-green. Models of MNEI are represented as gold neon backbone bonds. The models were generated by MOLMOL (Koradi et al., 1996).

the expense of a neutral patch, without significant changes in the alternation of positive and negative areas on the crucial region of interaction (Esposito et al., 2006).

B. Interaction of small sweeteners with the sweet receptor Like their template, the active conformations of the sweet taste receptor contain two cavities that can host ligands: a smaller one in the closed protomer and a larger one in the open protomer. It is natural to assume that most sweeteners interact with the sweet receptor via optimal fitting of one or both these cavities, but it is difficult to ascertain whether the

Active Sites of the Sweet Receptor

225

precise shape of the closed cavity is influenced by the binding. In fact, Gokulan et al. (2005) warned that the considerable conformational changes implied by the binding mechanism of VFTD domains might prevent accurate modeling of the active sites of the receptor. However, Morini et al. (2005) by exploiting the conservation of key residues and the similarity of some sweetener with glutamate were able to show that all active sites of the two active protomers can actually be used to account for the sweetening power of a very large number of sweet molecules. The sweet receptor can be activated by simple hydrophobic amino acids, notably D-tryptophan and synthetic dipeptides, generally derived from aspartame. These molecules have the same amino acidic moiety typical of all a-amino acids, including glutamate, that is an amino group adjacent to a carboxyl group. It is fair to assume that reliable active sites in T1R2– T1R3 receptor models should retain all the features necessary to bind this moiety. In other words, residues lining the wall of the part of the cavity that binds amino acidic moieties should be highly conserved in going from mGluR1 to T1R2–T1R3. In fact, as pointed out by Morini et al. (2005) for all their models, residues directly interacting with the a-amino acid moiety in mGluR1 are well conserved not only in T1R2–T1R3 and in other mGluRs but also in the sequences of other families of metabotropic GPCRs (Pin et al., 2003). On the other hand, residues of the other part of the cavity are expected to be more variable, since in the sweet taste receptor this part of the active site ought to accommodate molecular fragments of different size and chemical constitution. In addition, residues corresponding to those that bind the side chain of glutamate in mGluR1 should possibly turn from polar to hydrophobic in T1R2–T1R3 to accommodate molecular moieties more similar to that of tryptophan rather than that of glutamate. In agreement with these ideas, in the alignments corresponding to the four models, Morini et al. (2005) found that residues binding the glutamate side chain in mGluR1 are invariably changed to less polar or uncharged residues (Morini and Temussi, 2005). mGluR1 can bind two glutamate molecules: both closed (MOL1) and open (MOL2) protomers bind glutamate at active sites lined by the interfaces of subdomains LB1 and LB2 (Kunishima et al., 2000; Tsuchiya et al., 2002), with the only difference that, in the open protomer, the LB2 interface is not used for binding. Figure 9 illustrates these two modes of binding. Constitution and size of sweeteners can be so diverse that we cannot be sure a priori that in the sweet taste receptor both ligand-binding sites are available for sweet ligands. Even simple visual inspection of the models reveals that active sites of closed protomers, that is T1R2(A) and T1R3(A), are so small that they cannot possibly host some of the larger synthetic sweeteners (Morini et al., 2005). Owing to the large dimensions of some sweeteners, the active sites of open protomers in Aoc_AB and Aoc_BA can use both LB1 and LB2 interfaces. Figure 9B illustrates the

226

Pierandrea Temussi

A

B

Aoc_MOL1

Aoc_T1R2(A)

Aoc

Aoc_AB

Aoc_MOL2

Aoc_T1R3(B )

FIGURE 9 Mode of binding of ligands in the active sites of the protomers of metabotropic receptors. (A) Binding of two molecules of glutamate in the two protomers of mGluR1. The glutamate that binds to the closed protomer (MOL1) of the active form Aoc uses residues from both LB1 and LB2; the one that binds to the open protomer (MOL2) uses only residues from LB2. The two molecules of glutamate are represented as black balls of equal size. (B) Binding of two sweeteners of different dimensions to the active form Aoc_AB of T1R2–T1R3. The smaller sweetener, represented as a small gray ball, binds to the closed T1R2(A) site; the larger one, represented as a larger black ball, binds to the open T1R3(B) site using both LB1 and LB2 lobes.

binding of a small sweetener in site T1R2(A) and that of a larger (nonproteic) one in site T1R3(B). In order to probe semiquantitatively the fit of sweeteners in the active sites of the models, Morini et al. (2005) chose a large number of sweet molecules belonging to different families, including sugars, peptides, and supersweeteners. Their fit was evaluated by means of PrGen (Vedani et al., 1995), a program that allows the comparison of calculated binding affinity for ligands with the experimental sweetness. In the binding sites of the open protomers, it is possible to fit a large number of representative sweet compounds. In T1R2(B) and T1R3(B), 16 and 22 molecules, respectively, were used as training sets to derive the model and then to predict binding energy of other sets. Interestingly, although PrGen allows changes in the relative positions of the residues defining the site, the final active sites showed only minor changes with respect to those of the original homology models. Figure 10 shows the agreement between predicted and experimental free energies of binding for the open sites of both possible active conformations: T1R2(B) and T1R3(B). Open symbols refer to compounds used in training sets, whereas filled symbols refer to compounds of test sets. The compounds are those reported in the corresponding tables of Morini et al. (2005). Their relative sweetening power, referred to sucrose, are comprised between 200,000 (corresponding to a DG of 13.8 kcal/mol) of sucrononic acid and 0.26 (corresponding to a DG of 5.9 kcal/mol) of D-glucose. On the other hand, it proved very difficult or impossible to

227

Active Sites of the Sweet Receptor

T1R3(B)

T1R2(B)

G⬚(kcal/mol) (predicted)

−6

−10

−14

−15

−10 G⬚(kcal/mol) (experimental)

−5

−15

−10

−5

G⬚(kcal/mol) (experimental)

FIGURE 10 Correlation between calculated experimental and experimental binding affinities of sweet compounds inserted in the open sites of active conformations: T1R2 (B) and T1R3(B). Open diamonds, in the left-hand side panel, represent compounds used in the training set, whereas black diamonds represent compounds of the test set. Open circles, in the right-hand side panel, represent compounds used in the training set, whereas black circles represent compounds of the test set.

dock most of the larger ligands in the binding sites of closed protomers of the active closed–open form. In fact, it was possible to fit only four compounds in T1R2(A) with good correlation between experimental and calculated binding affinity; while in T1R3(A), although it was possible to dock the same four compounds, the correlation between experimental and calculated binding affinity was poor. Figure 11 shows the contact surface representation of the two sites of Aoc_AB with typical sweet molecules inside. The fit of saccharine in the T1R2(A) site is shown in the right-hand side of the figure and the fit of sucrononic acid (one of the guanidinium supersweeteners, 200,000 times sweeter than sucrose) in the T1R3(B) site in the left-hand side. It is easy to appreciate that large sweeteners, like sucrononic acid, can only enter the wider B sites. A special case is represented by flexible compounds, such as aspartame, which can exist in several conformations. For consistency with the PrGen calculations of Bassoli et al. (2002a), the first choice for the conformation of aspartame used by Morini et al. (2005) to derive the models was the folded conformation found in the crystal structure (Hatada et al., 1985) and used as paradigmatic in one of the indirect models (Yamazaki et al., 1994). However, our indirect model (Kamphuis et al., 1992) is consistent

228

Pierandrea Temussi

A

Y

180⬚

B

T1R3(B)

T1R2(A)

FIGURE 11 Contact surface representation of the two sites of Aoc_AB with typical sweetener molecules inside. (A) Individual protomers, T1R2(A) and T1R3(B), are represented with all heavy atoms colored in gray. The other protomer is represented by only by a line along the backbone. (B) The fit of saccharine in the T1R2(A) site is shown on the right-hand side and the fit of sucrononic acid (one of the guanidinium supersweeteners, 200,000 times sweeter than sucrose) on the T1R3(B) site is shown in the left-hand side. The two protomers, T1R2(A) and T1R3(B), are as gray contact surfaces; the sweet molecules are represented with black atoms. All models were generated by Molmol (Koradi et al., 1996).

with an extended form of aspartame, corresponding to the crystal structure of [(L-a-Me)Phe2] aspartame (Polinelli et al., 1992). Accordingly, Morini et al. (2005) checked the possibility of using [(L-a-Me)Phe2] aspartame among the test compounds in an alternative training set in lieu of folded aspartame. Both calculations showed that the folded conformation can fit both open and closed sites whereas the extended conformation of aspartame can be accommodated only in the open ‘‘B’’ cavities.

C. Multiple binding sites Early indirect models of the active site of the sweet taste receptor tried to account for the largest possible number of sweet compounds, but it was generally believed that some classes of sweet compounds, notably sweet proteins, might interact with different receptors altogether. Can we reconcile old views with the picture emerging from molecular biology and

Active Sites of the Sweet Receptor

229

homology modeling? The answer is that rather than multiple receptors there are, apparently, multiple sites on the single sweet taste receptor. The consensus feature of all indirect models was the presence of AH– B groups, in which the AH group is a hydrogen bond donor and the B group is a hydrogen bond acceptor. This feature is indeed present in both sites of the active forms of the T1R2–T1R3 receptor, as derived from homology modeling, using mGluR1 as template (Morini et al., 2005). The cavity that accepts sweet proteins (Morini and Temussi, 2005; Morini et al., 2005; Temussi, 2002) can be considered as a third independent active site. These three sites account for most observations on sweet molecules, including elusive concepts like synergy (vide infra). However, owing to the complexity of sweet compounds, it cannot be excluded that additional binding sites exist elsewhere in the heterodimeric sweet taste receptor. Two major candidates for additional sites are those proposed by Xu et al. (2004) for a site accepting both agonist cyclamate and the sweet taste inhibitor lactisole and by Jiang et al. (2004) for sweet proteins. Agonist specificity between human and rat allows rational design of specific chimeras of the receptor. Using this technique, Xu et al. (2004) mapped binding sites on the T1R2–T1R3 receptor by generating chimeras between human and rat T1Rs genes, with junctions at residues positioned at the borders of estimated transmembrane domains. These authors found that when the N-terminal domain of human T1R2 is replaced with the corresponding sequence of rat T1R2, the responses of the receptor to aspartame and neotame are abolished, showing that the N-terminal domain of human T1R2 is necessary to recognize typical sweeteners such as aspartame and neotame. However, when they replaced either the N-terminal or the C-terminal domain of human T1R2 with rat sequence, the response to cyclamate was not affected. Apparently, the transmembrane domain of human T1R3, when coexpressed with T1R2, is sufficient to recognize cyclamate. Similarly, they showed that lactisole, a human-specific sweet taste inhibitor, like cyclamate requires the human T1R3 C-terminal domain to inhibit the receptor’s response to typical sweet agonists. These observations were confirmed by Jiang et al. (2005). In addition, these authors, by means of alanine-scanning mutagenesis, identified six residues of the transmembrane domain specifically involved in the recognition of cyclamate. Contemporaneously, Winnig et al. (2005) found that a single residue, that is valine 738 on the fifth helix of the transmembrane domain of T1R3, is responsible for the lactisole insensitivity of rat sweet taste receptor. These findings do not rule out the possibility that cyclamate is recognized also by one of the two cavities of the VFT domain but suggest convincingly that the transmembrane domain of the T1R3 protomer of the sweet receptor hosts a genuine fourth site. Figure 12 summarizes all the sites described so far for T1R2–T1R3.

230

Pierandrea Temussi

Aoc_AB

T1R2(A) Small sweeteners

T1R3(B) Large nonproteic sweeteners

Cyclamate Sweet proteins

FIGURE 12 Binding sites for the T1R2–T1R3 receptor: two active sites of different dimensions for smaller and larger (non-proteic) sweeteners, the cyclamate site in the TM helices domain, and the external ‘‘wedge’’ site for proteins.

Chimera studies led Jiang et al. (2004) to propose yet another site for sweet proteins. Starting from the observation that human T1R2–T1R3 responds to brazzein whereas the chimera of human T1R2–mouse T1R3 does not, these authors argued that critical residues for this difference could be located in the cysteine-rich region of T1R3. Using human/mouse chimeras of T1R3 paired with hT1R2, they determined that, in particular, residues 536–545 of the cysteine-rich region of human T1R3 were required for responsiveness to brazzein. It is difficult at the moment to ascertain whether this is a genuine additional fifth active site since the cysteine-rich domain is a critical region that cannot be easily changed without affecting the global response of the receptor. It is fair to hypothesize that the cysteine-rich region has a crucial structural role in the conformational transitions of the sweet receptor as shown for other metabotropic receptors. For instance, in the case of the human Ca2þ receptor, Hu et al. (2000) have shown that the hCaR cysteine-rich domain plays a critical role in signal transfer from VFT to 7TM of the hCaR and for sequence specificity

Active Sites of the Sweet Receptor

231

in communication. Any mutation in this region may simply undermine the structural integrity of the sweet receptor. On the other hand, the wedge mechanism (Morini and Temussi, 2005; Morini et al., 2005; Temussi, 2002) would provide a simple explanation for the critical role played by T1R3 in the interaction with sweet proteins without invoking an additional site. Synergy between different sweeteners is a peculiar phenomenon of sweetness that has for a long time escaped a detailed interpretation at molecular level (DuBois, 2004). Sweetness synergy has been observed in several combinations of sweeteners. It had long been known that aspartame and cyclamate are synergistic in sensory experiments (Schiffman et al., 1995). The modeling of all possible conformations based on the human T1R2 and T1R3 sequences (Morini et al., 2005) suggested, among other aspects of the sweet taste, the first possible interpretation of this phenomenon. When exploring the four active sites of the heterodimers formed by human T1R2 and T1R3 sequences, using the A (closed) and B (open) chains of the ligand-binding domain of the mGluR1 glutamate receptor, it was immediately clear that both ‘‘type A sites’’ are definitely too small to host the bigger non-proteic sweeteners, but they can accommodate at least four compounds, namely saccharin, alitame, aspartame, and 6-Cl-D-tryptophan. On the other hand, both T1R2(B) and T1R3(B) can host a very large number of small molecular weight sweeteners with a good correlation between experimental and calculated binding affinity. The starting point to understand synergy is that at least three of the four compounds that, in the docking study of Morini et al. (2005), were able to bind to ‘‘type A sites,’’ aspartame, saccharin, alitame, and cyclamate, are known to be synergistic with other sweet compounds (DuBois, 2004), suggesting that, although the binding in a single subunit is sufficient for receptor activation, the additional binding of a ligand in the second subunit increases the response. The crucial point is that synergy can only be observed either with two small sweeteners or with a small and a large one, but never with two large non-proteic sweeteners. It is easy to see from Figure 9B that only the combinations small–small and small– large are consistent with the Aoc conformation of the receptor whereas two large sweeteners could stabilize the inactive Roo conformation.

VI. BEYOND THE SWEET RECEPTOR Our understanding of the structure–activity relationship of sweet molecules increases enormously with the discovery of the sweet taste receptor and with the subsequent availability of reliable homology models. Some points need a more detailed explanation, but we should probably wait for

232

Pierandrea Temussi

detailed 3D structures of the receptor and its complexes with different classes of sweeteners before we get an answer. For instance, it is not fully understood whether the T1R3 protomer has an active role in accepting large synthetic sweeteners and proteins or just a role in transmitting information to the transmembrane domain. Specificity of the T1R2–T1R3 heterodimer is assured by T1R2, since the companion T1R1–T1R3 receptor is specific for umami compounds. The docking calculations of Morini et al. (2005) are in favor of the Aoc_AB active form of the receptor, that is, with T1R2(A) and T1R3(B), but their data are not conclusive. Another interesting issue that has not yet received an explanation is the mechanism of action of substances that can suppress sweet taste, in particular gymnemic acid. However, it is fair to say that the main aspects of the interaction of sweet molecules with their receptor have been elucidated. The most unexpected findings are probably the explanations of phenomena such as sweeteners synergy and the taste of sweet proteins. As mentioned in Section I, an important motivation to study structure– activity relationships of sweet molecules is the possibility to design new sweeteners. In principle, detailed homology models of the two active sites of the active form of the human sweet receptor could indeed suggest key modifications of existing sweeteners or even entirely new scaffolds. In practice, however, the resolution of the models is not sufficient for accurate design; thus, also this problem must wait for a detailed solid state structure of the receptor. Besides, the actual use of new synthetic molecules is problematic since it requires long and costly tests before they can be introduced in the market. So far, sweet proteins have not been used as sweeteners, but they are very promising. The elucidation of their mode of action may open the way to modifications of existing proteins and even to the de novo design of new sweet proteins.

ACKNOWLEDGMENTS I wish to thank Annalisa Pastore (NIMR) for many critical readings of the chapter and sweet suggestions. Financial support from MIUR (FIRB 2003) is gratefully acknowledged.

REFERENCES Abe, K., Kusakabe, Y., Tanemura, K., Emori, Y., and Arai, S. (1993). Primary structure and cell-type specific expression of a gustatory G protein-coupled receptor related to olfactory receptors. J. Biol. Chem. 268, 12033–12039. Acton, E.M., and Stone, H. (1976). Potential new artificial sweetener from study of structuretaste relationships. Science 193, 584–586.

Active Sites of the Sweet Receptor

233

Adler, E., Hoon, M.A., Mueller, K.L., Chandrashekar, J., Ryba, N.J., and Zuker, C.S. (2000). A novel family of mammalian taste receptors. Cell 100, 693–702. Ariyoshi, Y., and Kohmura, M. (1994). Solid-phase synthesis and structure-activity relationships of analogs of the sweet protein monellin. J. Soc. Synth. Org. Chem. Jpn. 52, 359–369. Assadi-Porter, F.M., Aceti, D.J., and Markley, J.L. (2000). Sweetness determinant sites of brazzein, a small, heat-stable, sweet-tasting protein. Arch. Biochem. Biophys. 376, 259–265. Assadi-Porter, F.M., Abildgaard, F., Blad, H., and Markley, J.L. (2003). Correlation of sweetness of variants of the protein brazzein with patterns of hydrogen bonds detected by NMR spectroscopy. J. Biol. Chem. 278, 31331–31339. Bachmanov, A.A., Reed, D.R., Ninomiya, Y., Inoue, M., Tordoff, M.G., Price, R.A., and Beauchamp, G.K. (1997). Sucrose consumption in mice, major influence of two genetic loci affecting peripheral sensory responses. Mamm. Genome 8, 545–548. Bachmanov, A.A., Li, X., Reed, D.R., Ohmen, J.D., Li, S., Chen, Z., Tordoff, M.G., de Jong, P.J., Wu, C., West, D.B., Chatterjee, A., Ross, D.A., et al. (2001). Positional cloning of the mouse saccharin preference (Sac) locus. Chem. Senses 26, 925–933. Bassoli, A., Drew, M.G.B., Merlini, L., and Morini, G. (2002a). A general pseudoreceptor model for sweet compounds, a semi-quantitative prediction of binding affinity for sweet tasting molecules. J. Med. Chem. 45, 4402–4409. Bassoli, A., Merlini, L., and Morini, G. (2002b). Isovanillyl sweeteners. From molecules to receptors. Pure Appl. Chem. 74, 1181–1187. Bassoli, A., Borgonovo, G., Busnelli, G., Morini, G., and Merlini, L. (2005). Monatin, its stereoisomers and derivatives, modeling the sweet taste chemoreception mechanism. Eur. J. Org. Chem. 2005, 2518–2525. Blizard, D., Kotlus, B., and Frank, M. (1999). Quantitative trait loci associated with short-term intake of sucrose saccharin and quinine solutions in laboratory mice. Chem. Senses 24, 373–385. Bodani, U.C., Anchin, J.M., and Linthicum, D.S. (1993). Monoclonal antibodies to sweet taste proteins, II. Development of two different immunoassays for thaumatin and monellin. Hybridoma 12, 177–183. Bohak, Z., and Li, S.L. (1976). The structure of monellin and its relation to the sweetness of the protein. Biochim. Biophys. Acta 427, 153–170. Bridel, M., and Lavieille, R. (1931). The sweet principle of the leaves of Khaa-he-e (Stevia Rebaudiana Bertoni). C. R. Acad. Sci. 192, 1123–1125. Caldwell, J.E., Abildgaard, F., Dzakula, Z., Ming, D., Hellekant, G., and Markley, J.L. (1998). Solution structure of the thermostable sweet-tasting protein brazzein. Nat. Struct. Biol. 5, 427–431. Capeless, C., and Whitney, G. (1995). The genetic basis of preference for sweet substances among inbred strains of mice: Preference ratio phenotypes and the alleles of the Sac and dpa loci. Chem. Senses 20, 291–298. Chandrashekar, J., Mueller, K.L., Hoon, M.A., Adler, E., Feng, L., Guo, W., Zuker, C.S., and Ryba, N.J. (2000). T2Rs function as bitter receptors. Cell 100, 703–711. Chaudhari, N., Landin, A.M., and Roper, S.D. (2000). A metabotropic glutamate receptor variant functions as a taste receptor. Nat. Neurosci. 3, 113–119. Cohn, G. (1914). ‘‘Die Organischen Geschmackstoffe.’’ F. Siemenroth, Berlin. Cordier, F., and Grzesiek, S. (1999). Direct observation of hydrogen bonds in proteins by interresidue 3hJNC0 scalar couplings. J. Am. Chem. Soc. 121, 1601–1602. Dalvit, C. (1996). Homonuclear 1D and 2D NMR experiments for the observation of solventsolute interactions. J. Magn. Reson. B 112, 282–288. Dalvit, C. (1998). Efficient multiple-solvent suppression for the study of the interactions of organic solvents with biomolecules. J. Biomol. NMR 11, 437–444.

234

Pierandrea Temussi

de Vos, A.M., Hatada, M., van der Wel, H., Krabbendam, H., Peerdeman, A.F., and Kim, S.H. (1985). Three-dimensional structure of thaumatin I: An intensely sweet protein. Proc. Natl. Acad. Sci. USA 82, 1406–1409. Delwiche, J. (1996). Are there ‘basic’ tastes? Trends Food Sci. Technol. 7, 411–415. DuBois, G.E. (2004). Unraveling the biochemistry of sweet and umami tastes. Proc. Natl. Acad. Sci. USA 101, 13972–13973. DuBois, G.E., Crosby, G.A., and Stephenson, R.A. (1981). Dihydrochalcone sweeteners. A study of the atypical temporal phenomena. J. Med. Chem. 24, 408–428. Esposito, V., Gallucci, R., Picone, D., Tancredi, T., and Temussi, P.A. (2006). The importance of electrostatic potential in the interaction of sweet proteins with the sweet taste receptor. J. Mol. Biol. 360, 448–456. Firestein, S. (2000). Neurobiology: The good taste of genomics. Nature 404, 552–553. Fuller, J. (1974). Single-locus control of saccharin preference in mice. J. Hered. 65, 33–36. Gibbs, B.F., Alli, I., and Mulligan, C. (1996). Sweet and taste-modifying proteins: A review. Nutrition Res. 16, 1619–1630. Gilbertson, T., Damak, S., and Margolskee, R. (2000). The molecular physiology of taste transduction. Curr. Opin. Neurobiol. 10, 519–527. Glendinning, J.I. (1994). Is the bitter rejection response always adaptive? Physiol. Behav. 56, 1217–1227. Glendinning, J.I. (1996). Is chemosensory input essential for the rapid rejection of toxic foods? J. Exp. Biol. 199, 1523–1534. Gokulan, K., Khare, S., Ronning, D.R., Linthicum, S.D., Sacchettini, J.C., and Rupp, B. (2005). Cocrystal structures of NC68 Fab identify key interactions for high potency sweetener recognition, implications for the design of synthetic sweeteners. Biochemistry 44, 9889–9898. Goodman, M., Coddington, J., and Mierke, D.F. (1987). A model for the sweet taste of stereoisomeric retro-inverso and dipeptide amides. J. Amer. Chem. Soc. 109, 4712–4714. Goodman, M., Zhu, Q., Kent, D.R., Amino, Y., Iacovino, R., Benedetti, E., and Santini, A. (1997). Conformational analysis of the dipeptide taste ligand L-aspartyl-D-2-aminobutyric acid-(S)-alpha-ethylbenzylamide and its analogues by NMR spectroscopy, computer simulations and X-ray diffraction studies. J. Pept. Sci. 3, 231–241. Hatada, M., Jancarik, J., Graves, B., and Kim, S.-H. (1985). Crystal structure of aspartame, a peptide sweetener. J. Amer. Chem. Soc. 107, 4279–4282. Holm, L., and Sander, C. (1995). Dali, a network tool for protein structure comparison. Trends Biochem. Sci. 20, 478–480. Hoon, M.A., Adler, E., Lindemeier, J., Battey, J.F., Ryba, N.J., and Zuker, C.S. (1999). Putative mammalian taste receptors, a class of taste-specific GPCRs with distinct topographic selectivity. Cell 96, 541–551. Hu, J., Hauache, O., and Spiegel, A.M. (2000). Human Ca2þ receptor cysteine-rich domain Analysis of function of mutant and chimeric receptors. J. Biol. Chem. 275, 16382–16389. Hu, Z., and Min, H. (1983). Acta Bot. Yunnanica 5, 207–212. Hung, L.W., Kohmura, M., Ariyoshi, Y., and Kim, S.H. (1998). Structure of an enantiomeric protein, D-monellin at 18 A resolution. Acta Crystallogr. D Biol. Crystallogr. 54, 494–500. Hung, L.W., Kohmura, M., Ariyoshi, Y., and Kim, S.H. (1999). Structural differences in D and L-monellin in the crystals of racemic mixture. J. Mol. Biol. 285, 311–321. Igeta, H., Tamura, Y., Nakaya, K., Nakamura, Y., and Kurihara, Y. (1991). Determination of disulfide array and subunit structure of taste-modifying protein, miraculin. Biochim. Biophys. Acta 1079, 303–307. Inglett, G.E., and May, J.F. (1968). Tropical plants with unusual taste properties. Econ. Bot. 22, 326–331. Inglett, G.E., and May, J.F. (1969). Serendipity berries source of a new intense sweetener. J. Food Sci. 34, 408–411.

Active Sites of the Sweet Receptor

235

Iwamura, H. (1981). Structure-sweetness relationship of L-aspartyl dipeptide analogues A receptor site topology. J. Med. Chem. 24, 572–578. Iyengar, R.B., Smits, P., van der Ouderaa, F., van der Wel, H., van Brouwershaven, J., Ravestein, P., Richters, G., and van Wassenaar, P.D. (1979). The complete amino-acid sequence of the sweet protein thaumatin I. Eur. J. Biochem. 96, 193–204. Jiang, P., Ji, Q., Liu, Z., Snyder, L.A., Benard, L.M., Margolskee, R.F., and Max, M. (2004). The cysteine-rich region of T1R3 determines responses to intensely sweet proteins. J. Biol. Chem. 279, 45068–45075. Jiang, P., Cui, M., Zhao, B., Snyder, L.A., Benard, L.M., Osman, R., Max, M., and Margolskee, R.F. (2005). Identification of the cyclamate interaction site within the transmembrane domain of the human sweet taste receptor subunit T1R3. J. Biol. Chem. 280, 34296–34305. Jin, Z., Danilova, V., Assadi-Porter, F.M., Aceti, D.J., Markley, J.L., and Hellekant, G. (2003). Critical regions for the sweetness of brazzein. FEBS Lett. 544, 33–37. Kamphuis, J., Lelj, F., Tancredi, T., Toniolo, C., and Temussi, P.A. (1992). SAR of sweet molecules, conformational analysis of two hypersweet and two conformationally restricted aspartame analogues. QSAR 11, 486–491. Kaneko, R., and Kitabatake, N. (2001). Structure-sweetness relationship in thaumatin, importance of lysine residues. Chem. Senses 26, 167–177. Kier, L.B. (1972). Molecular theory of sweet taste. J. Pharm. Sci. 61, 1394–1397. Kim, S.H., Kang, C.-H., Kim, R., Cho, J.M., Lee, Y.-B., and Lee, T.-K. (1989). Redesigning a sweet protein, increased stability and renaturability. Protein Eng. 2, 571–575. Kim, S.H., Kang, C.-H., and Cho, J.M. (1991). Sweet proteins, biochemical studies and genetic engineering. In ‘‘Sweeteners, Discovery, Molecular Design and Chemoreception’’ (D. E. Walters, F. T. Orthofer, and G. E. DuBois, eds.), Vol. 450, pp. 28–40. ACS Symposium Series, ACS, Washington, DC. Kinnamon, S.C. (2000). A plethora of taste receptors. Neuron 25, 507–510. Kitagawa, M., Kusakabe, Y., Miura, H., Ninomiya, Y., and Hino, A. (2001). Molecular genetic identification of a candidate receptor gene for sweet taste. Biochem. Biophys. Res. Commun. 283, 236–242. Kohmura, M., Nio, N., and Ariyoshi, Y. (1992). Highly probable active site of the sweet protein monellin. Biosci. Biotechnol. Biochem. 56, 1937–1942. Koradi, R., Billeter, M., and Wu¨thrich, K. (1996). MOLMOL, a program for display and analysis of macromolecular structure. J. Mol. Graph. 14, 51–55. Kunishima, N., Shimada, Y., Tsuji, Y., Sato, T., Yamamoto, M., Kumasaka, T., Nakanishi, S., Jingami, H., and Morikawa, K. (2000). Structural basis of glutamate recognition by a dimeric metabotropic glutamate receptor. Nature 407, 971–977. Kurihara, K., and Beidler, L.M. (1969). Taste-modifying protein from miracle fruit. Science 161, 1241–1243. Kurihara, Y. (1992). Characteristics of antisweet substances, sweet proteins, and sweetnessinducing proteins. Crit. Rev. Food Sci. Nutr. 32, 231–252. Lee, S.Y., Lee, J.H., Chang, H.J., Cho, J.M., Jung, J.W., and Lee, W. (1999). Solution structure of a sweet protein single-chain monellin determined by nuclear magnetic resonance and dynamical simulated annealing calculations. Biochemistry 38, 2340–2346. Lelj, F., Tancredi, T., Temussi, P.A., and Toniolo, C. (1976). Interaction of alpha-L-aspartyl-Lphenylalanine methyl ester with the receptor site of the sweet taste bud. J. Amer. Chem. Soc. 98, 6669–6675. Li, X., Inoue, M., Reed, D.R., Hunque, T., Puchalski, R.B., Tordoff, M.G., Ninomiiya, Y., Beauchamp, G.K., and Bachmanov, A.A. (2001). High-resolution genetic mapping of the saccharin preference locus (Sac) and the putative sweet taste receptor (T1R1) gene (Gpr70) to mouse distal chromosome 4. Mamm. Genome 12, 13–16. Li, X., Staszewski, L., Xu, H., Durick, K., Zoller, M., and Adler, E. (2002). Human receptors for sweet and umami taste. Proc. Natl. Acad. Sci. USA 99, 4692–4696.

236

Pierandrea Temussi

Lindemann, B. (1996). Taste reception. Physiol. Rev. 76, 718–766. Lindemann, B. (1999). Receptor seeks ligand, on the way to cloning the molecular receptors for sweet and bitter taste. Nat. Med. 5, 381–382. Lindemann, B. (2001). Receptors and transduction in taste. Nature 413, 219–225. Liu, X., Maeda, S., Hu, Z., Aiuchi, T., Nakaya, K., and Kurihara, Y. (1993). Purification, complete amino acid sequence and structural characterization of the heat-stable sweet protein, mabinlin II. Eur. J. Biochem. 211, 281–287. Lush, I. (1989). The genetics of tasting in mice, VI. Saccharin, acesulfame, dulcin and sucrose. Genet. Res. 53, 95–99. Lush, I., Hornigold, N., King, P., and Stoye, J. (1995). The genetics of tasting in mice. VII. Glycine revisited, and the chromosomal location of Sac and Soa. Genet. Res. 66, 167–174. Maehashi, K., and Udaka, S. (1998). Sweetness of lysozymes. Biosci. Biotechnol. Biochem. 62, 605–606. Mandal, C., Shirley, F., Anchin, J.M., Mandal, C., and Linthicum, D.S. (1991). Monoclonal antibodies to sweet taste proteins. I. Analysis of antigenic epitopes on thaumatin by competitive inhibition assays. Hybridoma 10, 459–466. Masuda, T., Ide, N., and Kitabatake, N. (2005a). Effects of chemical modification of lysine residues on the sweetness of lysozyme. Chem. Senses 30, 253–264. Masuda, T., Ide, N., and Kitabatake, N. (2005b). Structure-sweetness relationship in egg white lysozyme, role of lysine and arginine residues on the elicitation of lysozyme sweetness. Chem. Senses 30, 667–681. Masuda, T., Ueno, Y., and Kitabatake, N. (2001). Sweetness and enzymatic activity of lysozyme. J. Agric. Food Chem. 49, 4937–4941. Matsunami, H., Montmayeur, J.-P., and Buck, L.A. (2000). Family of candidate taste receptors in human and mouse. Nature 404, 601–604. Matsuoka, I., Mori, T., Aoki, J., Sato, T., and Kurihara, K. (1993). Identification of novel members of G-protein coupled receptor superfamily expressed in bovine taste tissue. Biochem. Biophys. Res. Commun. 194, 504–511. Max, M., Shanker, Y.G., Huang, L., Rong, M., Liu, Z., Campagne, F., Weinstein, H., Damak, S., and Margolskee, R.F. (2001). Tas1r3, encoding a new candidate taste receptor, is allelic to the sweet responsiveness locus Sac. Nat. Genet. 28, 58–63. Mazur, R.H., Schlatter, J.M., and Goldkamp, A.H. (1969). Structure-taste relationships of some dipeptides. J. Amer. Chem. Soc. 91, 2684–2691. McLaughlin, S.K., McKinnon, P.J., and Margolskee, R.F. (1992). Gustducin is a tastecellspecific G protein closely related to the transducins. Nature 357, 563–569. McLaughlin, S.K., McKinnon, P.J., Spickofsky, N., Danho, W., and Margolskee, R.F. (1994). Molecular cloning of G proteins and phosphodiesterases from rat taste cells. Physiol. Behav. 56, 1157–1164. Mennella, J.A., Pepino, M.Y., and Reed, D.R. (2005). Genetic and environmental determinants of bitter perception and sweet preferences. Pediatrics 115, e216–e222. Ming, D., and Hellekant, G. (1994). Brazzein, a new high-potency thermostable sweet protein from Pentadiplandra brazzeana B. FEBS Lett. 355, 106–108. Moncrieff, R.W. (1967). ‘‘The Chemical Senses.’’ Leonard Hill, London. Montmayeur, J.P., Liberles, S.D., Matsunami, H., and Buck, L.B.A. (2001). Candidate taste receptor gene near a sweet taste locus. Nat. Neurosci. 4, 492–498. Morini, G., and Temussi, P.A. (2005). Micro and macro models of the sweet receptor. Chem. Senses 30, 86–87. Morini, G., Bassoli, A., and Temussi, P.A. (2005). From small sweeteners to sweet proteins, anatomy of the binding sites of the human T1R2-T1R3 receptor. J. Med. Chem. 48, 5520–5529. Morris, J.A. (1976). Sweetening agents from natural sources. Lloydia 39, 25–38.

Active Sites of the Sweet Receptor

237

Morris, J.A., and Cagan, R.H. (1972). Purification of monellin, the sweet principle of Dioscoreophyllum cumminsii. Biochim. Biophys. Acta 261, 114–122. Murzin, A.G. (1993). Sweet-tasting protein monellin is related to the cystatin family of thiol proteinase inhibitors. J. Mol. Biol. 230, 689–694. Murzin, A.G., Brenner, S.E., Hubbard, T., and Chothia, C. (1995). SCOP, a structural classification of proteins database for the investigation of sequences and structures. J. Mol. Biol. 247, 536–540. Nelson, G., Hoon, M.A., Chandrashekar, J., Zhang, Y., Ryba, N.J., and Zuker, C.S. (2001). Mammalian sweet taste receptors. Cell 106, 381–390. Niccolai, N., Spadaccini, R., Scarselli, M., Bernini, A., Crescenzi, O., Spiga, F., Ciutti, A., and Temussi, P.A. (2001). Probing the surface of a sweet protein, NMR study of MNEI with a paramagnetic probe. Protein Sci. 10, 1498–1507. Nofre, C., Tinti, J.-M., and Chatzopoulos-Ouar, F. (1988). Preparation of (phenylguanidino)and [[1-(phenylamino)ethyl]amino]acetic acids as sweeteners Eur Pat Appl EP 241,395, 1987. Chem. Abstr. 109, 190047k. Oertly, E., and Myers, R.G. (1919). A new theory relating constitution to taste [preliminary paper]. Simple relations between the constitution of aliphatic compounds and their sweet taste. J. Amer. Chem. Soc. 41, 855–867. Ogata, C.M., Gordon, P.F., de Vos, A.M., and Kim, S.H. (1992). Crystal structure of a sweet tasting protein thaumatin I, at 1.65 A resolution. J. Mol. Biol. 228, 893–908. Pin, J.P., Galvez, T., and Prezeau, L. (2003). Evolution, structure, and activation mechanism of family 3/C G-protein-coupled receptors. Pharmacol. Ther. 98, 325–354. Polinelli, S., Broxterman, Q.B., Schoemaker, H.E., Boesten, W.H.J., Crisma, M., Valle, G., Toniolo, C., and Kamphuis, J. (1992). New aspartame-like sweeteners containing L-(aMe) Phe. Bioorg. Med. Chem. Lett. 2, 453–456. Prakash, I., Bishay, I., and Schroeder, S. (1999). Neotame: Synthesis, stereochemistry and sweetness. Synth. Commun. 29, 4461–4467. Sainz, E., Korley, J.N., Battey, J.F., and Sullivan, S.L. (2001). Identification of a novel member of the T1R family of putative taste receptors. J. Neurochem. 77, 896–903. Schiffman, S.S., Booth, B.J., Carr, B.T., Losee, M.L., Sattely-Miller, E.A., and Graham, B.G. (1995). Investigation of synergism in binary mixtures of sweeteners. Brain Res. Bull. 38, 105–120. Schwalbe, H., Grimshaw, S.B., Spencer, A., Buck, M., Boyd, J., Dobson, C.M., Redfield, C., and Smith, L.J. (2001). A refined solution structure of hen lysozyme determined using residual dipolar coupling data. Protein Sci. 10, 677–688. Shallenberger, R.S., and Acree, T. (1967). Molecular theory of sweet taste. Nature 216, 480–482. Shirasuka, Y., Nakajima, K., Asakura, T., Yamashita, H., Yamamoto, A., Hata, S., Nagata, S., Abo, M., Sorimachi, H., and Abe, K. (2004). Neoculin as a new taste-modifying protein occurring in the fruit of Curculigo latifolia. Biosci. Biotechnol. Biochem. 68, 1403–1407. Slootstra, J.W., De Geus, P., Haas, H., Verrips, C.T., and Meloen, R.H. (1995). Possible active site of the sweet-tasting protein thaumatin. Chem. Senses 20, 535–543. Solms, J., Vuataz, L., and Egli, R.H. (1965). The taste of L- and D-amino acids. Experientia 21, 692–694. Somoza, J.R., Jiang, F., Tong, L., Kang, C.H., Cho, J.M., and Kim, S.H (1993). Two crystal structures of a potently sweet protein natural monellin at 275 A resolution and singlechain monellin at 17 A resolution. J. Mol. Biol. 234, 390–404. Somoza, J.R., Cho, J.M., and Kim, S.H. (1995). The taste-active regions of monellin, a potently sweet protein. Chem. Senses 20, 61–68. Spadaccini, R., Crescenzi, O., Tancredi, T., De Casamassimi, N., Saviano, G., Scognamiglio, R., Di Donato, A., and Temussi, P.A (2001). Solution structure of a sweet protein, NMR study of MNEI, a single chain monellin. J. Mol. Biol. 305, 505–514.

238

Pierandrea Temussi

Spadaccini, R., Trabucco, F., Saviano, G., Picone, D., Crescenzi, O., Tancredi, T., and Temussi, P.A. (2003). The mechanism of interaction of sweet proteins with the T1R2T1R3 receptor, evidence from the solution structure of G16A-MNEI. J. Mol. Biol. 328, 683–692. Strynadka, N.C., and James, M.N. (1996). Lysozyme, a model enzyme in protein crystallography. EXS 75, 185–222. Sung, Y.H., Shin, J., Chang, H.J., Cho, J.M., and Lee, W. (2001). Solution structure, backbone dynamics, and stability of a double mutant single-chain monellin structural origin of sweetness. J. Biol. Chem. 276, 19624–19630. Suzuki, M., Kurimoto, E., Nirasawa, S., Masuda, Y., Hori, K., Kurihara, Y., Shimba, N., Kawai, M., Suzuki, E., and Kato, K. (2004). Recombinant curculin heterodimer exhibits taste-modifying and sweet-tasting activities. FEBS Lett. 573, 135–138. Tahara, A., Nakata, T., and Ohtsuka, Y. (1971). New type of compound with strong sweetness. Nature 233, 619–620. Tancredi, T., Iijima, H., Saviano, G., Amodeo, P., and Temussi, P.A. (1992). Structural determination of the active site of a sweet protein, a 1H NMR investigation of MNEI. FEBS Lett. 310, 27–30. Tancredi, T., Pastore, A., Salvatori, S., Esposito, V., and Temussi, P.A. (2004). Interaction of sweet proteins with their receptor A conformational study of peptides corresponding to loops of brazzein, monellin and thaumatin. Eur. J. Biochem. 271, 2231–2240. Temussi, P.A. (2002). Why are sweet proteins sweet? Interaction of brazzein, monellin and thaumatin with the T1R2-T1R3 receptor. FEBS Lett. 526, 1–3. Temussi, P.A. (2006). Natural sweet macromolecules: How sweet proteins work. CMLS 63, 1876–1888. Temussi, P.A., Lelj, F., and Tancredi, T. (1978). Three-dimensional mapping of the sweet taste receptor site. J. Med. Chem. 21, 1154–1158. Temussi, P.A., Lelj, F., Tancredi, T., Castiglione-Morelli, M.A., and Pastore, A. (1984). Soft agonist-receptor interactions, theoretical and experimental simulation of the active site of the receptor site of sweet molecules. Int. J. Quantum Chem. 26, 889–906. Temussi, P.A., Lelj, F., and Tancredi, T. (1991). Structure-activity relationships of sweet molecules. In ‘‘Sweeteners, Discovery, Molecular Design and Chemoreception’’ (D. E. Walters, F. T. Orthofer, and G. E. DuBois, eds.), Vol. 450, pp. 143–161. ACS Symposium Series, ACS, Washington, DC. Theerasilp, S., and Kurihara, Y. (1988). Complete purification and characterization of the taste-modifying protein, miraculin, from miracle fruit. J. Biol. Chem. 263, 11536–11539. Theerasilp, S., Hitotsuya, H., Nakajo, S., Nakaya, K., Nakamura, Y., and Kurihara, Y. (1989). Complete amino acid sequence and structure characterization of the taste-modifying protein, miraculin. J. Biol. Chem. 264, 6655–6659. Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F., and Higgins, D.G. (1997). The CLUSTAL_X windows interface, flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res. 24, 4876–4882. Tinti, J.M., and Nofre, C. (1991). Why does a sweetener taste sweet? A new model. In ‘‘Sweeteners, Discovery, Molecular Design and Chemoreception’’ (D. E. Walters, F. T. Orthofer, and G. E. DuBois, eds.), Vol. 450, pp. 88–99. ACS Symposium Series, ACS, Washington, DC. Tsuchiya, D., Kunishima, N., Kamiya, N., Jingami, H., and Morikawa, K. (2002). Structural views of the ligand-binding cores of a metabotropic glutamate receptor complexed with an antagonist and both glutamate and Gd3þ. Proc. Natl. Acad. Sci. USA 99, 2660–2665. van der Wel, H., and Loeve, K. (1972). Isolation and characterization of thaumatin I and II, the sweet-tasting proteins from Thaumatococcus daniellii Benth. Eur. J. Biochem. 31, 221–225.

Active Sites of the Sweet Receptor

239

Vedani, A., Zbinden, P., Snyder, J.P., and Greenidge, P.A. (1995). Pseudoreceptor modelling, the construction of three dimensional receptor surrogates. J. Amer. Chem. Soc. 117, 4987–4994. Walters, D.E. (1995). Using models to understand and design sweeteners. J. Chem. Educ. 72, 680–683. Walters, D.E. (2002). Homology-based model of the extracellular domain of the taste receptor T1R3. Pure Appl. Chem. 74, 1117–1123. Walters, D.E., Pearlstein, R.A., and Krimmel, C.P. (1986). A procedure for preparing models of receptor sites. J. Chem. Educ. 63, 869–872. Wang, Z., Singhvi, A., Kong, P., and Scott, K. (2004). Taste representations in the Drosophila brain. Cell 117, 981–991. Winnig, M., Bufe, B., and Meyerhof, W. (2005). Valine 738 and lysine 735 in the fifth transmembrane domain of rTas1r3 mediate insensitivity towards lactisole of the rat sweet taste receptor. BMC Neurosci. 6, 22. Xu, H., Staszewski, L., Tang, H., Adler, E., Zoller, M., and Li, X. (2004). Different functional roles of T1R subunits in the heteromeric taste receptors. Proc. Natl. Acad. Sci. USA 101, 14258–14263. Yamashita, H., Theerasilp, S., Aiuchi, T., Nakaya, K., Nakamura, Y., and Kurihara, Y. (1990). Purification and complete amino acid sequence of a new type of sweet protein tastemodifying activity, curculin. J. Biol. Chem. 265, 15770–15775. Yamazaki, T., Benedetti, E., Kent, D., and Goodman, M. (1994). Conformational requirements for sweet-tasting peptides and peptidomimetics. Angew. Chem. Int. Ed. Engl. 33, 1437–1451. Zaffaroni, A. (1975). U.S. Patent 3 876 816, CA.

This page intentionally left blank

INDEX A AAT, see a–1 Antitrypsin Active human lysozyme, 180–181 Active sites, indirect mapping of small molecular weight sweet molecules, 202–205 structural studies, 205–209 ADHD, see Attention Deficit Hyperactivity Disorder Adipogenesis suppression, 147 Advanced meat recovery, 52 Air-injection stunning devices, 46–47 Aitch bone, role, 49 ALA, see a-Linolenic acid American Dairy Products Institute, 1990, 16 American Type Culture Center (ATCC), United States, 126 Amino acids improvements, 176 AMR, see Advanced meat recovery Ankaflavin, 129, 133, 146–147 Annexin V protein, 46 Anticancer agent, 167 Anti-GFAP antibiotic, role, 53 Anti-myelin basic protein (anti-MBP), 53 Anti-neurofilament (anti-NF), 53 Anti-NSE antibiotic, role, 53–54 Antithrombin III protein, 185 a–1 Antitrypsin, 184–185 A. oryzae, 133 Apolipoprotein E-deficient [ApoE ( / )] mice, 138 Arabidopsis thaliana, 214 Arg14, see Lysozyme Aspartame, 201, 203, 207 l-aspartylphenylalanine methyl ester, 203–204 Aspergillus niger, 127 A. terreus, lovastatin biosynthetic pathway in, 131 Attention Deficit Hyperactivity Disorder, 172 ‘‘Auxogluc’’, 205

B BAC, see Bacterial artificial chromosome Bacillus licheniformis, 21–22 Bacillus megaterium, 147 Bacillus subtilis, 147 Bacterial artificial chromosome, 176 Beef carcass carcass splitting, 48–50 CNS tissue, removal of, 50–51 contamination routes, 45 stunning devices in carcass contamination, 46–48 BELE, see Breed Early Lactate Early ‘‘Ben Cao Gang Mu’’, 126, 128 Blood, LDL levels, 166 Bovine spongiform encephalopathy, 40 controlled risk category, 44 detection, 43 testing, risk estimation and control measures, 45 Bovine TAP (bTAP), 188 Brazzein and loop L23, 218 receptor-binding surface in, 214 scorpion toxin-like superfamily, 213 Breed Early Lactate Early, 191 BSE, see Bovine spongiform encephalopathy

C Caco-2 cells, 147 Caenorhabditis elegans, 172, 173 Calcium chelating agents, see EDTA Candida albicans, 186 Candida pseudotropicalis, 147 Capparis masaikai, 214 Captive bolt stunning devices, 46 Carcass contamination, stunning devices in, 46 Carcass sampling, GFAP, 50 Carcass splitting, 48–50

241

242

Index

Casein, 167 edible, uses, 177 health risk, 177–178 micelles, EDTA and, 13–14 whey ratio, yoghurt milk and, 11 as1/b-Casein, 7 as1-Casein (as1-CN), 176 Cattle population, BSE risk, 43–44 Cellular antigen stimulation test (CAST), 151 Central nervous system, 40 Chemical fingerprint-profiling method, 135 Chinese Center for Disease Control and Prevention, 149 Cholesterol immunochemical assays and quantification, 53–54 role, CNS tissue meat products, 54 Cholesterol-lowering effect, MRP and animal studies, 138–140 clinical studies, 140 HMG-CoA reductase inhibitor, 137–138 Cholestin, 138, 140, 142 Chromatin remodeling, 90 ‘‘Cinderella of senses’’, 201 Citrate salts, in cheese industry, 15 Citrinin, 130–131, 148–150 CLA, see Conjugated linoleic acid CN, see Casein CNS, see Central nervous system CNS tissue, 51 beef carcass contamination, 45 cross-contamination, 47–49 gas chromatography-mass spectrometry, 58–59 histological staining and IHC, 52–53 immunochemical assays and quantification of cholesterol, 53–58 polymerase chain reaction, 59–60 qualitative analysis, 56–57 removal of, 50–51 Colloidal calcium phosphate solubilization, 13 Colorimetric ELISA, 56 detection of GFAP in meat products, 57 Compactin (ML-236B), 137–138 Comprehensive Yeast Genome Database, 79 Conjugated linoleic acid, 167 beneficial health effects, 167–168 content, 169–170 Control diet (CTD), 169 ‘‘Controlled BSE risk’’, 44

Coronavirus, 190 Cow dietary manipulation, 166, 168–169, 172 Cow genetic manipulation, 172–173 Cow milk allergy, 181–182 Curculigo latifolia, 216 Curculin, 209, 216 Cutaneous omobrachialis, 49

D Dairy Cooperative Research Centre (Dairy CRC), 176 Dairy industry, enzymatic processes in, see also Milk hydrolysis, 20–22 renneting, 18–20 transglutamination, 22–23 Dairy ingredients chemical modification processes chemical agents use, 23–25 Maillard reaction, 25–27 Dairy products, CLA content, 168–169 Dairy proteins, 4 DALI, 217–218 DCN, see Dry casein Delta-9 desaturase, 170 Designer milk, animal growth and health, 187–190 DHA, see Docosahexaenoic acid Dihydroquercetin 3-acetate, 204 Dimerumic acid antioxidant mechanism, 146 Dioscorea batatas, 137 Dioscoreophyllum cumminsii, 210 1,1-diphenyl-2-picrylhydrazyl (DPPH), 146 Disease prevalence and surveillance program, 42 Docosahexaenoic acid, 170 Double dyebinding method, 84–85 Dry casein, 167 ‘‘Dulcigen’’ groups, 205 Dynamic high-pressure processing technology, 28

E EC1118 profiling transcripts, 96–97 EDTA, 13–14 Eicosapentaenoic acid, 170 Electrospray ionization tandem mass spectrometry, 83 ELISA, see Enzyme-linked immunosorbent assay

Index

Energy reserve compound synthesis, 102 Environmental stress response (ESR), 100–102 Environment Risk Management Authority (ERMA), 186 Enzyme-linked immunosorbent assay, 46, 211, 217 EPA, see Eicosapentaenoic acid ePHOGSY, 212 Escherichia coli, 181, 188, 216 ESI MS/MS, see Electrospray ionization tandem mass spectrometry Eupatorium rebaudianum, 204 European Union (EU), 40

F Fatty acid chain length, alteration of, 166 Fatty acid methyl ester (FAME) analysis, 74 Fatty acids detection of CNS tissue in meat products, 58 disease treatment by, 172 versus product quality, 173 Fibrinogen protein, 184 Fluorescent ELISA (F-ELISA), 55, 57 Food Industry Research and Development Institute, Taiwan, 126 Food processing technologies, 27–28 static high-pressure processing, 27 ultrasound, 28 Fourier transform ion cyclotron resonance MS (FTICR-MS), 88 Functional analysis by coresponses in yeast (FANCY), 106

G GABA production, 137 b galactosidase, 174–175 G16A-MNEI, 222 Gas chromatography-mass spectrometry (GC-MS), 52, 58 GC-selected ion-monitoring mass spectrometry, 149 Genetic manipulation (GM), 163 ‘‘Genome renewal’’ hypothesis for Saccharomyces, 73 Genomic tool chest, 75–78 metabolome analysis, 87–89 proteome analysis, 82–87

243

systems biology, 90 transcriptome analysis, 79–82 Glial fibrillary acidic protein (GFAP), 47 in meat products, 50 restriction and detection of CNS, 55 Globular proteins hydrolysis, 20 ‘‘Glucophores’’, 205, 211, 223 D glucopyranosyl_D fructofuranose, 202 Glucose-glycerol-peptone (GGP), 136 Glycerol-3-phosphate dehydrogenase (GPDH), 147 Glycine, 202–203 Glycyrrhizin, see Natural glycoside Goats, feeding, 167 Goodman’s model, 208 G-protein–coupled receptor (GPCR), 201 Grape juice composition of, 69 environment, 67–68, 70–73 fermentation and factors affecting, 68, 70–73 phenolic composition of, 70 GTC Biotherapeutics, 185 Gustatory system, 200

H hATT, see Human ATT Health risk estimation, 42 Hematoxylin and eosin (HE), 52 Hen egg white (HEW), 209, 214–215 High-density lipoprotein cholesterol (HDL-C), 138 High-fat diet (HF), 169 High-oleic sunflower oil (HOSO), 167 High-performance liquid chromatography (HPLC), 135 High-protein milk powders, 17–18 hLF, see Human milk-derived lactoferrin hLZ, see Active human lysozyme Homogenization effects, on dairy products functionality, 15–16 Hot boning, carcasses, 49 Human ATT, 185 Human clinical applications, antibodies use, 183 Human embryonic kidney cells (HEK293), 148 Human milk-derived lactoferrin, 179 and hLZ, role, 181 Human neurological disease cause, 40–41

244

Index

Human tissue plasminogen activator (htPA), 183 Hydrogen bond acceptor (B), 205–206 Hydrogen bond donor (AH), 205–206 3-hydroxy-3-methylglutaryl coenzymeA (HMG-CoA), 138

I Immunoassay, NSE detection in liver sausages, 55 Immunochemical testing, 57–58 Immunoglobulin antibodies (IgA), 186 Immunohistochemistry (IHC) histological staining and, 52–53 method, detection of spinal cord, 57 Indirect mapping, of active sites and small molecular weight sweet molecules, 202–205 structural studies, 205–209 Infant health, designer milk for cow milk allergy, 181–182 lactoferrin, 179–180 lactose intolerance, 182 lysozyme, 180–181 Institute for Nutrition and Food Safety, China, 149 International Society for Study of Fatty Acids and Lipids (ISSFAL), 170 Ipomoea batatas, 137 Isovanillyl sweet compounds, 204

J ‘‘Jiuqu’’ in China, 126, 137

K k Casein (k CN), 5–6, 177 cleavage, 18–19 Keratin (KN), 167 Kier’s model, 206–207

L LA, see Linoleic acid Laboratory yeast, 71, 73, 96, 105–107 a Lactalbumin (a LA), 175 proteins, 5 removal, 175 Lactoferrin (LF), 179–180 b-Lactoglobulin (b LG), 163 proteins, 5–7, 9–10, 24, 26

Lactose, 174, see also Milk sugar modification intolerance, 182 reduction, preharvest methods, 175 synthesis, 174 Linoleic acid, 167 a-Linolenic acid, 170 Lipase production, 182 Listeria, 190 Longer acting tissue plasminogen activator (LAtPA), 183 Loop L23, 218 Lovastatin, 131–132, 135, 138–140 Low-density lipoprotein cholesterol (LDL-C), 137–140 Low density lipoprotein (LDL), 166 Low-fat diet (LFD), 169 ‘‘L’’-shaped model, 207–208 Lysostaphin protein, 189 Lysozyme (LZ), 180–181 hydrolysis of polysaccharides catalyzation, 214 level, in milk, 180 ribbon representations, 215

M Mabinlin, 214 Maillard reaction conjugates preparation, 25–26 solubility and heat stability, 26 surface properties, 26–27 viscosity and gelation, 27 Major milk proteins, 176–178 Malaria vaccine, 185–186 Manihot esculenta, 137 M. anka, 145–146 Mastitis, 188–189 Matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS), 83 Matrix-assisted laser desorption ionization time-of-flight/time-of-flight mass spectrometry (MALDI-TOF/TOF MS), 147 MBP, see Myelin basic protein Meat products, staining method, 53 Metabolic gene regulation principles, 90 Metabolome analysis, 87–89 Metabotropic glutamate receptors (mGluR1), 200, 220, 225 M. floridanus, 126 Microarray karyotype analysis, 75

Index

Microbiological cross-contamination, carcass splitting, 49 Micrococcus lysodeikticus, 180 Microfluidization treatment, 28 Milk fat CLA and, 167, 169 modification altering fatty acid chain and saturation, 165–167 fatty acids versus product quality, 173 increasing CLA levels, 167–170 omega fatty acids, 170–172 reducing fat content, 172–173 MUFA content, 166 oleic acid content, 167, 173 percentage, 165 Milk fuctionality and processing impacts on, 3 acidification, 9–13 acylation, 23–24 benefits and constituents, 163 chemical modification processes, 23–27 composition alteration, mineral salts and, 13 dehydration, 16–18 designing, 163–165 emerging food processing technologies, 27–29 enzymatic modification processes, 18–23 esterification with alcohols, 24 fat (see Milk fat) foaming properties and calcium complexing agents addition, 14–15 gelation, 10 heat stability and pH, 6, 14 heat treatment, 4–10 homogenization and shear, 15–16 human therapeutic proteins, 183–186 mineral-protein equilibria importance, 14 mineral salts addition in, 13–15 modification, 176–178 nutritional source and beneficial health effect, 174 physical modification processes and, 4–18 powder, 16–17 total solids, 176–177 Milk protein concentrate (MPC), 13 Milk sugar, modification, 174–175 Mineral and casein equilibria, 13 Minor milk proteins, modifying, 178 Miraculin, 209, 215–216 M. kaoliang R-10847 mutant, 133

245

Monacolin K glucose and, 131 Monascus species and, 136 MRP production process and, 136–137 production, 135–137 Monascin, 129, 133, 146–147 Monascopyridine A and B, 150–151 Monascorubramine, 129, 133, 151 Monascorubrin, 129, 131, 133, 135, 147 Monascus pigments production, 133–135 polyketide synthesis, 129–132 Monascus rice products (MRPs), 140, 146, 149 antihyperglycemic activity of, 146–147 antihypertensive effect of, 145 antioxidant effect of, 145–146 antiproliferative effect of, 146–147 clinical data in colesterol-lowering effect, 141–145 health benefits cholesterol-lowering effect, 134–140 other effects, 140, 145–148 macrophage-stimulating activity of, 147 production, traditional methods, 128–129 safety, 148–151 submerged culture system, 129, 134 types, in China, 127 use in Asia, history of, 126–128 Monascus sp. morphology, 125–126 taxanomy, 124–126 Monascus sp. ATCC 16775, 147 Monascus strain IBCC1, 133 Monatin, 203–204 Monell Chemical Senses Center, 210 Monellin, 210–212 Monosodium glutamate (MSG) solution, 134 Monounsaturated fatty acids (MUFA), 165 M. pallens, 126 M. paxii, 147 M. pilosus, 125–126, 136–137, 146–148, 150 M. pilosus IFO 4520, 128, 137, 147 M. purpureus, 125–126, 128, 132–134, 136–139, 143, 145–151, 147 M. purpureus M9011, 145 M. purpureus M12 mutant, 133 M. purpureus N 301, 149 M. purpureus NTU 301, 137 M. purpureus NTU568, 138 M. purpureus Went CBS 109.07, 133 M. purpureus Went C322 pigment, 134

246

Index

M. ruber, 125–126, 131, 133, 136, 145, 147–148, 150 M. ruber IFO 32318, 145 M. sanguineus, 126 Multidimensional liquid chromatography tandem mass spectrometry (MudPIT), 85 Multiple sclerosis disease, 186 Murine monoclonal antibody (mAB), 209 Myelin basic protein, 186

N Nanoflow liquid chromatography-liquid chromatography tandem mass spectrometry [Nano LC/LC MS/MS (MudPIT)], 85 National Institute of Technology and Evaluation Biological Resource Center (NBRC), Japan, 126 Native yeast strains investigation challenges genomic tool chest, 75–90 grape juice environment, 66–73 wine yeast strain diversity, 73–75 Natural glycoside, 204 ‘‘Negligible BSE risk’’, 44 Neoculin acid subunit (NAS), 216 Neoculin basic subunit (NBS), 216 Neohesperidine dihydrochalcone, 204 Neuron-specific enolase (NSE), 52 Nongel technologies, 85 Nuclear magnetic resonance (NMR), 211–214

O OIE, see World Animal Health Organization OIE surveillance program, role, 42–43 Oleic acid content, 167, 173 Omega fatty acids, 170–172 Open reading frames (ORFs), 76 Osladin extraction, 204

P P. brevicompactum, 138 Penicillium citrinum, 131, 138 Pentadiplandra brazzeana, 213 Perilla frutescens, 204 Perillartine, 204 Phenomics, 76 Phenylketonuria (PKU), 184 Phosphoenolpyruvate carboxykinase (PEPCK), 146 Photodiode array (PDA) detector, 135

Pithing technique, 46 Plasma endothelin-1(ET-1), 138 Plasmin, activity, 178 Plasmodium falciparum, 185 Polyketide biosynthesis, 131–132 Polyketide pathway, in Monascus sp., 129–132 Polyketide synthase (PKS) gene (pksCT), 150 Polymerase chain reaction, 59 Polypodium vulgare, 204 Polyunsaturated fatty acids, 165 Pompe’s disease, 187 PPL therapeutics, 186 Preharvest methodologies, lactose reduction, 175 Proteinase-cleavage sites, targeting, 178 Protein(s) chemical modification Maillard reaction, 25–27 use of chemical agents, 23–25 ‘‘chips’’ method, 86 emulsifying properties and hydrolysis, 21 expression, 176 foaming properties and hydrolysis, 21 hydrolysis effects, 20–22 ionization analytical methods, 83 phosphorylation, 24–25 production patterns, noise in, 90 separation and identification by 2D SDS-PAGE electrophoresis, 83–84 limitation in, 84–87 Stoke’s radius, 83 surface properties, heat treatment and pH impact on, 9 Proteome analysis differential protein targeting, 87 double dyebinding method, 84 by electrophoresis, 83 nongel technologies for, 85 protein ‘‘chips’’ method, 86 PUFA, see Polyunsaturated fatty acids

Q QSAR analyses, of dipeptide analogues, 207 Quantitative reverse transcription PCR (QRT-PCR), 82

R Real-time polymerase chain reaction (RT-PCR), 48 Recombinant bovine CD14 (rb-CD14), 189–190

Index

Recombinant human fibrinogen (rHF), 184 Recombinant human lactoferrin (rhLF), 179–180 Recombinant human proteins (rHP), 183 Red rice wine, in China, 127 R-ELISA, see Colorimetric ELISA Renneting, enzymatic process, 18–20 Response surface methodology (RSM), 136 Restriction fragment length polymorphisms (RFLP), 59 Reversed phase HPLC (RP-HPLC) with PDA method, 135 Roller drying method for chocolate powder, 17 Rubropunctamine, 129, 133, 151 Rubropunctatin, 129, 131, 133, 135, 147

S Saccharomyces cerevisiae, 133 application in genomic technologies, 67 genome sequence, 66 genomic deletion set of strains for, 77 Saccharomyces Genome Database, 79 Sac gene, 220 SAGE, see Serial analysis of gene expression Salmonella, 190 Salmonella typhimurium, 173 Saturated fatty acids (SFA), 165–166 SCOP classification, 212–213 SDS-PAGE, see Sodium dodecyl sulfate polyacrylamide gel electrophoresis 2D SDS-PAGE, see Two-dimensional sodium dodecyl sulfate polyacrylamide gel electrophoresis ‘‘Seed Koji’’preparation, 128 ‘‘Serendipity berries’’, 210 Serial analysis, of gene expression, 82 Seven-helix transmembrane domain (7TM), 201 SimplesseW 100, 16 Single-chain monellin (SCM), 211 Sodium dodecyl sulfate polyacrylamide gel electrophoresis, 54 Solanum tuberosum, 137 Solid-state and submerged cultivations combined process, 134 Solid-state fermentation, 128, 134, 137 Sorbitol, 203 Specified risk material (SRM), 40–41 removal, control measures, 50–51 ‘‘Spophores’’, 205 Stanford Microarray Database, 79

247

Staphylococcus aureus, 188 Static high-pressure processing technology, 27–28 Stearoyl-CoA desaturase enzyme, role, 167 Stevioside, see Sweet glycosides Streptococcus agalactiae, 188 Streptococcus dysgalactiae, 188 Streptococcus simulans, 189 Streptococcus uberis, 188 Streptozotocin-induced diabetic rats (STZ-diabetic rats), 146 Stress response element (STRE), 100 Sugar alcohols, 202 Sweet glycosides, 204 Sweet macromolecules and natural sweet proteins characterization curculin, 216 lysozyme, 214–215 mabinlin, 214 miraculin, 215–216 monellin, 210–212 thaumatin, 212–213 sweet proteins and receptor interaction, 217–218 Sweet proteins and receptor interaction, 217–218 role of, 201 ‘‘wedge model’’ mechanisms for, 221–224 Sweet taste receptor and computer-generated models, 220–221 and molecular biology, 218–220 multiple binding sites, 228–231 small sweetner and receptor, interaction of, 224–228 Synaptophysin protein, detection, 52–53 Synsepalum dulcificum, 215 Syntaxin 1-B antibody role, 46–47 Systems biology, 90

T Tandem affinity purification (TAP), 77 TEMPOL, 212 Temussi model, 207–208 Terpene perillaldehyde extraction, 204 Thaumatin extraction, 212–213 Thaumatococcus danielli, 212 ‘‘Tian Gong Kai Wu’’, 126, 128 Topological model, 207–208 Total cholesterol (TC), 138–139 T1R3, 219 Tracheal antimicrobial peptide (TAP), 188

248

Index

Transcriptional regulatory mechanisms, 90 Transcriptome, 75 Transcript profiling methods, 80–82 Transgenic technology, lactose reduction, 182 Transglutamination, 22–23 Tricarboxylic acid (TCA) cycle, 131, 132 T1R2–T1R3 receptor, 202, 220–222, 225, 229–230 Two-dimensional sodium dodecyl sulfate polyacrylamide gel electrophoresis, 79 Ty1 insertional mutagenesis, genetic footprinting of yeast, 78 Type A surveillance, 42–43 Type B surveillance, 43 TyrA13-AspA16, 211, 217 Tyr57-Asp59, 217

U b–1,4–UDP-galactosyl transferase (UDP-gal), 175 ‘‘Undetermined BSE risk’’, 44–45

V Variant Creutzfeldt–Jakob disease (vCJD), 41 Venus fly trap domain (VFTD), 201–202, 220, 225

W ‘‘Wedge model’’ mechanism, and sweet proteins, 221–224 Whey gels formation, 8 Whey proteins denaturation, 5 heat stability of, 7 hydrolysis and gelation, 21–22 Wine environment, gene expression, 94–99

Wine fermentation, cyclical nature of adaptation to stress, 101 Wine strains and stress factors, 103 Wine strains of S. cerevisiae ATP status of, 95 fermentation stages of, 98 impact of change in growth environments, 94 oligoarray transcript profiling and ‘‘flor’’ strains of, 92 transcript profiles in, 93 transcript profiling studies of, 91 Wine yeast functional genomic analysis, 91–99 genomic analysis, 66–67 metabolomics application, 106–108 strain diversity, 73–75 stress responses analysis, 99–104 World Animal Health Organization, 42

X Xuezhikang effects, 138, 143–145

Y Yeast cell growth of, 72 fermentation of, 71–73 gene annotation in, 78 global analysis of protein localization in, 77 metabolome of, 76 mRNA profiling by trancriptome analysis, 79–82 tool for research, 66 Yeast Protein Map project, 79 Yeast Proteome Database, 79 Yeast Resource Center for analysis of protein interactions and complexes, 79

Synthetic grape juice fermentation 100 Glucose fermentation rate Fructose fermentation rate Glucose Fructose Cell mass Viable cells

120 100 80

10

1

60 40

0.1

20 0 0

2

4

6

−20

8

Absorbance (580 nm) − colony forming units (x10E6)

Sugar (g/liter) - fermentation rate (g/liter/day)

140

10 0.01

Time (days)

Plate 1

Band isolation

In gel digestion

Peptide fragments

Peptide selection

Mass spectrometry

Peptide mass fingerprinting

Peptide fragmentation

Peptide sequence determination

Database search and protein identification

Plate 2

Aoc_AB

Aoc_BA

Roo_AB

Roo_BA

Plate 3